HYDROLOGY DRIVES EVERGLADES ECOSYSTEM FUNCTION: IMPLICATIONS FOR
ECOSYSTEM VULNERABILITY TO DROUGHT, ENERGY BALANCE, CLIMATE
TELECONNECTIONS AND CLIMATE CHANGE
by
SPARKLE LEIGH MALONE
CHRISTINA L. STAUDHAMMER, COMMITTEE CO-CHAIR
GREGORY STARR, COMMITTEE CO-CHAIR
JULIA CHERRY
HENRY W. LOESCHER
MICHAEL G. RYAN
A DISSERTATION
Submitted in partial fulfillment of the requirements
for the degree of Doctors of Philosophy
in the Department of Biological Sciences
in the Graduate School of
The University of Alabama
TUSCALOOSA, ALABAMA
2014
COPYRIGHT SPARKLE LEIGH MALONE 2014
ALL RIGHTS RESERVED
ABSTRACT
Wetlands are an essential component of the terrestrial carbon pool. Hydric conditions
slow decomposition and allow for soil carbon (C) accumulation and storage for long time
periods. Although wetlands have large carbon sequestering potentials that could potentially serve
as a negative feedback to climate change, they are threatened globally by anthropogenic
pressures. In particular, water management has greatly altered the Florida Everglades, one of the
largest freshwater ecosystems in the United States. To improve degraded areas of Everglades
National Park (ENP), water management is being modified by the Comprehensive Everglades
Restoration Plan (CERP), which seeks to re-establish water levels and hydroperiods closer to
natural regimes. This study strives to understand the complex relationships between Everglades
hydrology, climate, and C dynamics at different scales (plot and ecosystem) using multiple
approaches (static chamber, eddy covariance, simulation modeling) and analysis techniques
(linear, non-linear, and time series modeling techniques). I examined the effects of hydroperiod
on the greenhouse C balance and energy balance in Everglades freshwater marsh ecosystems. I
also investigated the effect of the El Niño Southern Oscillation (ENSO) and hydrometeorological parameters on in-situ CO2 dynamics, and the potential impact of projected
climate change on ecosystem CO2 exchange rates via simulation modeling using the DAYCENT
model. Everglades hydrology was demonstrated to co-vary with changes in greenhouse warming
potentials, energy fluxes and ENSO phase, indicating that hydrology is important for creating
and maintaining conditions sufficient for wetland ecosystem structure and function.
Hydroperiods are likely to change in the future with the implementation of CERP and with
ii
climate change, making it extremely important to understand the complex relationships between
hydrology, climate, energy exchange and CO2, and how these relationships influence ecosystem
structure and function. This research contributes to the understanding of the unique hydrology of
Everglades wetland ecosystems and the complex relationships between hydrology, climate and C
dynamics.
iii
LIST OF ABBREVIATIONS AND SYMBOLYS
a.g.l
above ground level
AIC
Akaike’s information criteria
AR
autoregressive
ARIMA
autoregressive integrated moving average
BIC
Bayesian information criteria
c
carbon dioxide
C
carbon
CERP
Comprehensive Everglades Restoration Plan
CH4
methane
CH4:CO2
ratio of methane to carbon dioxide
CO2
carbon dioxide
Cp
specific heat of air at constant pressure
Cpw
specific heat of liquid water
Cps
specific heat of the soil
d
surface roughness
EC
eddy covariance
ENSO
El Niño Southern Oscillation
ENP
Everglades National Park
EPA
Environmental Protection Agency
ET
evapotranspiration
iv
FCE
Florida Coastal Everglades
G
storage flux
GEE
gross ecosystem exchange
Gw
change in energy storage associated with the water above the soil surface
Gw+s
change in energy storage in the matrix of both the water and the soil below the
soil surface
GWP
global warming potential
H
sensible heat flux
IPCC
Intergovernmental Panel on Climate Change
IRGA
infrared gas analyzer
LE
latent energy
LTER
Long Term Ecological Research
MA
moving average
Mair
molecular weights of air
Mw
molecular weights of water
N
Nitrogen
N2
Nitrogen Gas
N2O
nitrous oxide
NCDC
National Climate Data Center
NDVI
normalized difference vegetation index
NEE
net ecosystem exchange
NPQI
normalized phaeophytinization index
NOx
nitrogen oxides
NOAA
National Oceanic and Atmospheric Administration
v
ONI
Oceanic Niño Index
P
Phosphorus
p
probability of obtaining a test statistic at least as extreme as the one that was
actually observed
P
barometric pressure
PRI
photochemical reflectance index
PAR
photosynthetically active radiation
PDSI
Palmer drought severity index
Pmax
maximum ecosystem CO2 uptake rate
q
molar fraction of water vapor
r
Pearson product-moment correlation coefficient
R
ideal gas constant
R0
base respiration rate when air temperature is 0 C
R2
coefficient of determination
Reco
ecosystem respiration
REML
restricted maximum likelihood
Rh
relative humidity
Rn
net radiation
Rs
incident solar radiation
S
sulfur
SRS
Shark River Slough
Tair
air temperature
TS
Taylor Slough
Ts
sonic temperature
vi
Tsoil
soil temperature
Tw
water temperature
Tk
virtual temperature
u
longitudinal component of the wind velocity
u*
friction velocity
v
lateral component of the wind velocity
V
molar volume of air
VPD
vapor pressure deficit
VWC
soil volumetric water content
w
vertical component of the wind velocity
WI
water index
z
depth
Z
canopy height

quantum efficiency

Bowen ratio

latent heat of vaporization of water
air
air density
w
water density
s
soil bulk density
v
water vapor molar density

photosynthetically active radiation
vii
ACKNOWLEDGMENTS
I would like to express my deepest gratitude to my advisors, Christina Staudhammer and
Gregory Starr, for their encouragement, expert guidance and patience. I would also like to thank
all of my committee members, Julia Cherry, Michael Ryan and Hank Loescher for their
invaluable input, inspiring questions, and support of both the dissertation and my academic
progress. Steve Oberbauer, Paulo Olivas and Jessica Schedlbauer also provided invaluable
assistance. Financial support was provided by the United States Forest Service Rocky Mountain
Research Station, Department of Energy’s (DOE) National Institute for Climate Change
Research (NICCR) through grant 07-SC-NICCR-1059, the US Department of Education
Graduate Assistantships in Areas of National Need (GAANN) program, and the National
Science Foundation (NSF) Division of Atmospheric and Geospace Sciences (AGS) Atmospheric
Chemistry program through grant 1233006 and the Florida Coastal Everglades Long Term
Ecological Research program under Cooperative Agreements DBI-0620409 and DEB-9910514.
Finally, this research would not have been possible without the support of my friends, lab mates,
fellow graduate students and of course of my family who never stopped encouraging me to
persist.
viii
CONTENTS
ABSTRACT .................................................................................................................................... ii
LIST OF ABBREVIATIONS AND SYMBOLYS ....................................................................... iv
ACKNOWLEDGMENTS ........................................................................................................... viii
LISTS OF TABLES ...................................................................................................................... xii
LIST OF FIGURES ..................................................................................................................... xiii
CHAPTER 1: INTRODUCTION ................................................................................................... 1
Background ..................................................................................................................................... 1
Research Objectives ........................................................................................................................ 6
CHAPTER 2: EFFECTS OF SIMULATED DROUGHT ON THE CARBON BALANCE OF
EVERGLADES SHORT HYDRO-PERIOD MARSH .................................................................. 8
Abstract ........................................................................................................................................... 8
Introduction ................................................................................................................................... 10
Materials and Methods.................................................................................................................. 13
Results ........................................................................................................................................... 20
Discussion ..................................................................................................................................... 23
References ..................................................................................................................................... 40
CHAPTER 3: SEASONAL PATTERNS IN ENERGY PARTITIONING OF TWO
FRESHWATER MARSH ECOSYSTEMS IN THE FLORIDA EVERGLADES ...................... 46
Abstract ......................................................................................................................................... 46
ix
Introduction ................................................................................................................................... 48
Materials and Methods .................................................................................................................. 51
Results ........................................................................................................................................... 60
Discussion ..................................................................................................................................... 65
References ..................................................................................................................................... 86
CHAPTER 4: EL NIÑO SOUTHERN OSCILLATION (ENSO) ENHANCES CO2
EXCHANGE RATES IN FRESHWATER MARSH ECOSYSTEMS IN THE FLORIDA
EVERGLADES ............................................................................................................................ 95
Abstract ......................................................................................................................................... 95
Introduction ................................................................................................................................... 97
Materials and Methods ................................................................................................................ 100
Results ......................................................................................................................................... 111
Discussion ................................................................................................................................... 118
References ................................................................................................................................... 138
CHAPTER 5: ECOSYSTEM SENSITIVITY TO CLIMATE CHANGE: A CASE STUDY
FROM THE FRESHWATER MARSHES OF THE FLORIDA EVERGLADES .................... 144
Abstract ....................................................................................................................................... 144
Introduction ................................................................................................................................. 146
Materials and Methods ................................................................................................................ 149
Results ......................................................................................................................................... 156
Discussion ................................................................................................................................... 159
References ................................................................................................................................... 179
CHAPTER 6: CONCLUSION ................................................................................................... 191
Final Conclusion ......................................................................................................................... 191
x
References ................................................................................................................................... 196
xi
LISTS OF TABLES
1. Pearson correlation coefficients for C, reflective indices, and environmental factors......33
2. Repeated measures analysis of NEE, Reco and GEE…………………………………….34
3. Repeated measures analysis of the ratio of net CH4:CO2 and reflective indices………..35
4. Energy flux partitioning in wetland ecosystems………………………………………...75
5. Annual and seasonal ET and precipitation (2009- 2012)……………………………….76
6. Parameter estimates from ARIMA models of Rn…………………………………….....77
7. Parameter estimates from ARIMA models of H………………………………………..78
8. Parameter estimates from ARIMA models of LE………………………………………79
9. Parameter estimates from ARIMA models of the Bowen ratio………………………...80
10. Parameter estimates from ARIMA models of precipitation and temperature………….125
11. Seasonal and annual NEE, GEE, and Reco……………………………………………...126
12. Model estimates for light and temperature response curves…………………………....127
13. Parameter estimates from ARIMA models of daily NEE……………………………....128
14. Parameter estimates from ARIMA models of daily Reco………………………….........129
15. Parameter estimates from ARIMA models of daily GEE………………………….…...130
16. Parameter estimates from ARIMA models of the daily water index…………………...131
17. DAYCENT site characteristics…………………………………………………………168
18. Climate change scenarios alter cumulative NEE, Reco, and GEE……………………....169
xii
LIST OF FIGURES
1. Changes in average NEE, Reco, and GEE………………………………….….…36
2. Least square mean predicted values of the interactions. ……………………..….37
3. Changes in average CH4:CO2 …………………………….……………….…….38
4. Average leaf area, and dry weight for each drought scenario……………….......39
5. Energy budget for wetland ecosystems………………………………..………....81
6. Patterns in Tair, ET, Water level, Rh and VPD…………………………..……....82
7. Energy balance closure………………….………………………….…………....83
8. Annual patterns in Rn, storage, and surface fluxes………………………………84
9. Weekly moving average Bowen Ratios………...…………………..…………....85
10. Map of TS and SRS…………………………………………………………….132
11. Time series of β…………………………………………………….…………...133
12. Time series of precipitation, PDSI, and water levels…………………………...134
13. The U.S. Drought Monitor for the Everglades region………………………….135
14. The relationship between precipitation, season length, and NEE………………136
15. Light response curves by ENSO phase………………………………….……...137
16. Conceptual diagram of the DAYCENT ecosystem model…..………………....170
17. Long-term daily weather data………………………….………………..……...171
18. Wet season onset and length……………………………………….…………...172
19. Observed versus modeled soil temperature and soil volumetric water content...173
20. Observed versus modeled CO2 exchange rates………………………….……...174
xiii
21. Climate change scalars…………………………………………….…………....175
22. The effect of climate change on cumulative NEE, Reco and GEE at TS……......176
23. The effect of climate change on cumulative NEE, Reco and GEE at SRS……...177
24. Multifactor effect of climate change on NEE, Reco and GEE.……………..…...178
xiv
CHAPTER 1: INTRODUCTION
Background
Wetland Ecosystems
While representing just 5 to 8% of global land cover, wetlands are one of the largest
components of the terrestrial carbon (C) pool (Whiting & Chanton 1993; Zedler & Kercher 2005;
Mitsch & Gosslink 2007). Carbon is sequestered and stored in the soil, where hydric conditions
slow decomposition and C accumulates over long time periods (Whiting & Chanton 2001;
Brevik & Homburg 2004; Choi & Wang 2004). Globally wetlands are being reduced as a result
of land cover change (i.e., agriculture and development; Armentano 1980), creating great
uncertainty in the stability of this large carbon pool in the face of climate change.
The Florida Everglades
One of the largest (~700,000 ha) freshwater marsh ecosystems in North America, the
Florida Everglades has been greatly altered by water control structures and land cover change. Its
origin is ~5,000 years BP when sea level rise slowed and peat began to accumulate (Craft and
Richardson 1993). The unique mosaic of wetland ecosystems of the Everglades was shaped by
the historic hydrology of the south Florida region (Davis & Ogden 1994). Historically,
hydroperiods were a function of precipitation throughout the Kissimmee-OkeechobeeEverglades ecosystems (Davis & Ogden 1994; Redfield 2000), which varied seasonally
developing a wet and dry season (Davis & Ogden 1994). During the wet season, Lake
Okeechobee (Myers & Ewel 1992) would overflow causing surface flow to travel south through
1
the Everglades and out of the Florida Bay (Myers & Ewel 1992; Davis & Ogden 1994). Small
variations in elevation throughout the landscape determined the degree of exposure to surface
flow making the position within the landscape important in understanding seasonal patterns in
hydrology (Davis & Ogden 1994; Richardson 2008).
Currently, Everglades hydrology is heavily influenced by anthropogenic water
management (Davis & Ogden 1994; Richardson 2008). Since the early 1900s, the hydrologic
regime in this subtropical system has been altered by 2500 km of spillways, levees, and canals
(Loveless 1959; Davis & Ogden 1994) that were designed for flood protection, to make land
available for agriculture and urbanization, and to provide water to south Florida. The severe
decline in water flowing into the Everglades led to a decrease of 1.2 m in the average water table
depth in freshwater marsh ecosystems (Armentano et al., 2006) and has changed the natural
characteristics of these wetland ecosystems (Davis & Ogden 1994; Richardson 2008).
Characteristics consistent with chronically reduced water levels (i.e. peat subsidence, reduced
marl accretion rates, exotic species encroachment, altered fire regimes, higher abundance of
woody and herbaceous species) are evident in southern sections of the Everglades. As a result of
the adverse affects of critically low water flow, current water management is being modified
again under the Comprehensive Everglades Restoration Plan (CERP; Perry 2004). Water
resources are being re-distributed under the CERP to re-establish water levels and hydroperiods
closer to natural regimes and to ameliorate areas that suffer from chronically low water levels
(Perry 2004).
Energy Balance
Hydrology is the single most important factor in determining the structure and function of
wetlands, including their energy balance. The components of the ecosystem energy balance
2
provide insight into the relationship between evapotranspiration (ET), a major source of water
loss in wetlands, and ecosystem hydrology. Improving the understanding of these links, as well
as their environmental controls, is important for determining how energy balance changes
through wet and dry seasons, as well as with future climate. In wetland ecosystems net radiation
(Rn) is stored in the soil and/or water column (G), and is also partitioned as sensible heat flux (H)
and latent energy flux (LE) (Lafleur et al., 2008; Piccolo 2009; Schedlbauer et al., 2011).
Seasonality influences H and LE partitioning, which co-vary with fluctuations in hydroperiods
(Schedlbauer et al., 2011). Surface fluxes (H and LE) influence water loss as LE
(evapotranspiration; ET) (Myers & Ewel 1992; Davis & Ogden 1994), and seasonal patterns in
water level through their influence on convective rain, the main source of wet season
precipitation (Myers & Ewel 1992). Energy exchange both drives and responds to hydrology and
is therefore an import factor for describing seasonal hydrological changes in the Everglades
region. Evaluating links between hydrology and Everglades energy balance is especially
important in light of imminent changes in water management and climate change.
Carbon
Hydrology is also an extremely important factor in C cycling (Bubier et al. 2003 ab;
Smith et al., 2003; Heinsch 2004; Webster et al., 2013), which directly impacts productivity
(Childers 2006; Hao et al., 2011; Schedlbauer et al. 2012), decomposition rates, CH4 production
and oxidation (King & Skovgaard 1990; Bachoon & Jones 1992; Torn & Chapin 1993; Smith et
al., 2003; Whalen 2005), CaCO3 precipitation (Davis & Ogden 1994) and CO2 sequestration
(Jimenez et al., 2012). Within the Everglades, C is maintained as peat and marl, through
processes tightly coupled with hydroperiod (Davis & Ogden 1994) Peat accumulates in marshes
with deep water and long hydroperiods, overlying permeable limestone substrate (Myers & Ewel
3
1992). In areas with short-hydroperiods and seasonal drying, marl substrate derived from algal
carbonate precipitation in periphyton mats develops (Myers & Ewel 1992). Prior to the last 100
years, the Everglades were a net sink for organic C as peat accreted to depths of 1–3 m.
Hydroperiods have shaped soil conditions and species composition in Everglades ecosystems in
ways that have led to different seasonal patterns in CO2 exchange rates (Schedlbauer et al., 2012;
Jimenez et al., 2012). Annually and seasonally, precipitation and water levels vary substantially
with climate patterns (Davis & Ogden 1994), leading to significant shifts in ecosystem CO2
exchange rates.
El Niño Southern Oscillation
Teleconnections from the El Niño Southern Oscillation (ENSO) are known to affect
global climate (Beckage et al., 2003), global terrestrial productivity (Behrenfeld et al., 2001) and
have been associated with precipitation and water depth anomalies in the Everglades (Davis &
Ogden 1994). ENSO extremes, alternating periods of warm (El Niño phase) or cold (La Niña
phase) sea surface temperatures in the Pacific Ocean (Trenberth 1997), have occurred with
regular periodicity (3 to 7 years) over the last 130,000 years (Beckage et al., 2003). In the Florida
Everglades, changes in the long-term trends in the hydrologic cycle have been linked to ENSO
extremes (Piechota & Dracup 1996; Allan & Soden 2008). Here, El Niño phases increase dry
season rainfall, resulting in higher seasonal and annual water levels (Piechota & Dracup 1996;
Allan & Soden 2008). In contrast, La Niña phases reduce dry season rainfall, leading to extreme
drought (Piechota & Dracup 1996; Beckage et al., 2003; Allan & Soden 2008; Moses et al.,
2013). Because annual shifts in carbon dioxide (CO2) exchange rates have been linked to
changes in surface hydrology in short-term studies (Barr et al., 2010; Jimenez et al., 2012;
Schedlbauer et al., 2010; Malone et al., 2013), El Niño and La Niña phases may be a significant
4
source of seasonal-to-interannual variation in ecosystem hydrology and productivity in
Everglades freshwater marsh ecosystems.
Climate Change in the Everglades
Changes in seasonal precipitation patterns and temperature are expected with climate
change in the Everglades region (Stanton & Ackerman 2007; IPCC 2013). Although annual
precipitation is expected to rise, wet season precipitation is projected to decline by 5 to 10%
(Christensen et al., 2007), and summer months in the Everglades region will become 5 to 10%
drier (Christensen et al., 2007; Stanton & Ackerman 2007). Greater annual precipitation is
predicted to occur with warming (+1 to +4.2°C by 2100; IPCC 2013), and is associated with
larger convective storms and higher intensity hurricanes (Allan & Soden 2008; IPCC 2013).
Given that the wet season and maximum water levels occur during the summer, this could have a
significant impact on hydroperiods. Considering the strong relationship between hydrology and
wetland ecosystem structure and function, future climate could potentially change vegetation
communities and C dynamics of Everglades ecosystems (Davis & Ogden 1994; Bush et al.,
1998; Todd et al., 2010).
DAYCENT
Limited experimental capabilities exist to evaluate complicated interactive controls on
ecosystem responses to multifactor drivers (Fuhrer 2003; Luo et al., 2008), although these effects
are critical to understand how climate change impacts terrestrial ecosystems (Luo et al., 2008).
Ultimately, time and financial constraints limit multifactor experiments (Luo et al., 2008), and
thus simulation models have become a useful tool to investigate the effects of rising atmospheric
CO2 concentrations and climate change scenarios on terrestrial ecosystems (Abdalla et al., 2010).
5
The DAYCENT model (Del Grosso et al., 2001) can simulate ecosystem water, carbon and
nutrient dynamics (Parton 1987, 1988) for various native and managed systems (Del Grosso et
al., 2002; Del Grosso et al., 2009; Abdalla et al., 2010). DAYCENT has been used to
successfully simulate ecosystem responses to changes in climate (Parton et al., 1995; Lou et al.,
2008; Savage et al., 2013; Abdalla et al., 2010), and to model gas fluxes (CO2, CH4, N2O, NOx,
N2). It has also been used to model carbon and nutrient dynamics (N, P and S) in shrublands (Li
et al., 2006), forest (Hartman et al., 2007; Parton et al., 2010), crops (Del Grosso et al., 2002; Del
Grosso et al., 2005; Stehfest et al., 2007; Chang et al., 2013; Duval et al., 2003), and temperate
wetlands and savannas.
Research Objectives
This research examines how extrinsic factors such as hydroperiod and climate (in situ and
simulated) influence carbon dynamics in subtropical freshwater marsh ecosystems of Everglades
National Park (ENP). The primary objectives of this study are to: 1) detect changes in CO2 and
CH4 in freshwater marsh ecosystems in relation to experimental fluctuations in hydroperiod and
drought, 2) determine the effect of hydrology on wetland energy balance, 3) evaluate the
consequences of the El Niño Southern Oscillation (ENSO) and hydro-meteorological parameters
on in situ CO2 dynamics and 4) explore the potential impact of projected climate change on
ecosystem CO2 exchange rates via simulation modeling using the DAYCENT model (Parton et
al., 1998). I utilized different scales (plot and ecosystem), approaches (static chamber, eddy
covariance, simulation modeling), and analysis techniques (linear, non-linear, and time series
modeling techniques) to address the following four questions:
1. How do changes in hydroperiod and drought simulation influence the greenhouse carbon
balance (CH4:CO2) of freshwater marsh ecosystems of the Everglades? (Chapter 2)
2. How does hydrology influence the energy balance of wetland ecosystems? (Chapter 3)
6
3. Are climate teleconnections driving inter-annual patterns in CO2 dynamics? (Chapter 4)
4. How will climate change alter CO2 exchange rates in Everglades freshwater marsh
ecosystems? (Chapter 5)
This is the first study to quantify the greenhouse carbon balance of an Everglades
freshwater marsh ecosystem and relate it to water level fluctuations. In addition to comparing the
energy balances of short- and long-hydroperiod ecosystems in the Everglades, this is the first
study to characterize the complex interactions between energy balance components and
environmental conditions using 4 years of continuous data in the Everglades. To my knowledge
this is also the first study to link inter-annual and seasonal fluctuations in CO2 exchange rates
with climate teleconnections (ENSO) and the first to parameterize DAYCENT for Everglades
ecosystems and use it to simulate the effects of projected climate change on CO2 exchange rates.
7
CHAPTER 2: EFFECTS OF SIMULATED DROUGHT ON THE CARBON BALANCE OF
EVERGLADES SHORT HYDRO-PERIOD MARSH
Abstract
Hydrology drives the carbon balance of wetlands by controlling the uptake and release of
CO2 and CH4. Longer dry periods in between heavier precipitation events predicted for the
Everglades region may alter the stability of large carbon pools in this wetland’s ecosystems. To
determine the effects of drought on CO2 fluxes and CH4 emissions, hydro-period simulations
with three scenarios that differed in the onset rate of drought (gradual, intermediate, and rapid
transition into drought) were implemented on 18 freshwater wetland monoliths collected from an
Everglades short hydro-period marsh.
Simulated drought, regardless of the onset rate, resulted in higher net CO2 losses (NEE)
over the 22-week manipulation. Drought caused extensive vegetation dieback, increased
ecosystem respiration (Reco) and reduced carbon uptake (GEE). Photosynthetic potential
measured by reflective indices (photochemical reflectance index, water index, normalized
phaeophytinization index, and the normalized difference vegetation index) indicated that water
stress limited GEE and inhibited Reco. As a result of drought-induced dieback, NEE did not offset
methane production during periods of inundation. The average ratio of net CH4 to NEE over the
study period was 0.06, surpassing the 100-year greenhouse warming compensation point for CH4
(0.04). Drought induced diebacks of sawgrass (C3) led to the establishment of the invasive
species torpedograss (C4) when water was resupplied. These changes in the structure and
function indicate that freshwater marsh ecosystems can become a net source of CO2 and CH4 to
8
the atmosphere, even following an extended drought. Future changes in precipitation patterns
and drought occurrence/duration can change the carbon storage capacity of freshwater marshes
from sinks to sources of carbon to the atmosphere. Therefore, climate change will impact the
carbon storage capacity of freshwater marshes by influencing water availability and the potential
for positive feedbacks on radiative forcing.
9
Introduction
Wetlands have a great potential for carbon sequestration and storage as hydric conditions
slow decomposition in the soil, and carbon accumulates over long time periods at rates higher
than other ecosystems (Whiting & Chanton 2001; Brevik & Homburg 2004; Choi & Wang
2004). Globally, there is more carbon stored in soil (1,672 Gt) than in the atmosphere (738.2 Gt)
and plant biomass (850 Gt; living and dead) combined (Reddy & DeLaune 2008). Representing
just 5-8% of global land cover, wetlands are one of the largest components of the terrestrial
carbon pool, storing 535 Gt of carbon below ground as peat, which is 32% of all soil carbon
(Whiting & Chanton 1993; Zedler & Kercher 2005; Mitsch & Gosslink 2007; Adhikari et al.,
2009). The stability of this large pool of carbon is uncertain due to human influence and changes
in climate.
Globally, wetlands have been reduced as a result of land cover change (Armentano
1980). Agriculture, development, and changes in hydrology caused 50% of wetland loss in the
United States since the 1900s (Dugan 1993). One of the largest wetland ecosystems in the U.S.,
the Everglades, has had considerable human influence on its hydrological regime. Over the last
130 years, the hydrologic regime in this subtropical system has been altered by 2500 km of
spillways, levees, and canals that were designed for flood protection and to provide water to
south Florida. Due to the adverse effects of reduced water flow, current water management is
being modified again under the Comprehensive Everglades Restoration Plan (CERP) (Perry
2004). As a part of CERP, water resources will be re-distributed, changing the current water
levels and hydroperiods to levels closer to natural values in many areas of Everglades National
Park (Perry 2004).
10
Driving the greenhouse carbon balance, hydrology is an extremely important factor in
carbon cycling (Bubier et al., 2003 a, b; Smith et al., 2003; Heinsch 2004; Webster et al., 2013).
Hydroperiods directly impact productivity (Childers 2006; Hao et al., 2011; Schedlbauer et al.,
2012), decomposition rates, CH4 production and oxidation (King & Skovgaard 1990; Bachoon &
Jones 1992; Torn & Chapin 1993; Smith et al., 2003; Whalen 2005), CaCO3 precipitation (Davis
& Ogden 1994) and CO2 sequestration (Jimenez et al., 2012). Although wetland ecosystems have
large carbon pools, under certain conditions they can become a source of carbon to the
atmosphere (Jimenez et al., 2012; Webster et al., 2013). As water levels decrease, oxygen
availability increases aerobic respiration rates, decomposing the carbon stored in the soil and
causing losses to the atmosphere as CO2 (Webster et al., 2013). Wetlands also become a source
of carbon when low redox potentials initiate the production of other greenhouse gases, such as
CH4 and N2O (Smith et al., 2003; Webster et al., 2013).
The ratio of CH4 emissions to net CO2 uptake is an index for an ecosystem’s greenhouse
gas (carbon) exchange balance with the atmosphere (Whiting & Chanton 2001). In wetland
ecosystems the greenhouse gas (carbon) exchange balance is dependent on interactions between
physical conditions, microbial processes in the soil, and vegetation characteristics (King &
Skovgaard 1990; Bachoon & Jones 1992; Whiting & Chanton 2001; Smith et al., 2003). Through
heterotrophic respiration and decomposition of organic matter, CO2 is released from soil,
increasing exponentially with higher temperatures (Pearlstine et al., 2010; Bubier et al., 2003a;
Heinsch 2004; Aurela et al., 2007; Jimenez et al., 2012) and decreasing with soil saturation (Torn
& Chapin 1993; Whiting & Chanton 2001; Smith et al., 2003; Heinsch 2004; Schedlbauer et al.,
2010; Jimenez et al., 2012; Webster et al., 2013). Inundated conditions promote carbon storage
from biomass produced by photosynthesis and stimulate losses as a result of low redox potentials
11
that lead to the production of the more potent greenhouse gas CH4 in freshwater wetland
ecosystems (Mitsch & Gosslink 2007; Hao et al., 2011). Although wetlands contribute more than
10% of the annual global emissions of CH4, previous studies have shown that NEE can mitigate
the impact of CH4 efflux (Whiting & Chanton 2001; Mitra et al., 2005; Mitsch et al., 2013). Yet,
due to changes in climate, the future greenhouse gas (carbon) exchange balance of wetlands is
uncertain (Whiting & Chanton 2001).
The goal of this study was to use simulated drought to determine how the greenhouse
carbon balance of short-hydroperiod freshwater ecosystems respond to changes in hydroperiod. I
used the Everglades as my study site, as its hydrologic cycle is not only expected to change with
CERP, but with future predictions of climate change. For the south Florida region, the IPCC
(2007) projects increased occurrence of large single day rain events with higher drought
frequency and rising air temperatures. The increase in drought frequency and higher
temperatures will cause lower water availability, which will significantly influence the
greenhouse carbon balance of Everglades ecosystems (Davis & Ogden 1994; Todd et al., 2010).
Understanding the complex relationships between carbon cycling in the Everglades freshwater
marshes and environmental controls is important in determining the future carbon dynamics of
wetland ecosystems.
I hypothesize that the increased frequency and duration of droughts will increase CO2 and
decrease CH4 emissions from freshwater marshes of the Everglades’ through a reduction in soil
water availability. Reductions in water availability will result in increased soil oxygenation,
lower rates of methanogenesis, and higher CO2 production via methane oxidation and increased
aerobic respiration (Bachoon & Jones 1992). The effects of water stress will also cause a
reduction in the water content of the vegetation, light use efficiency, and chlorophyll content,
12
while chlorophyll degradation should increase. I also hypothesize that the speed of drought onset
will influence photosynthetic potentials and net exchange rates for both CO2 and CH4. A more
gradual transition to drought will produce lower net ecosystem exchange (NEE) rates (more
carbon uptake) and higher ecosystem CH4 fluxes than intermediate and rapid transition into
drought, due to the length of water availability. Drought simulation will likely decrease carbon
storage and increase atmospheric forcing. Since changes in hydrology due to climate and
management are issues for wetlands globally, results from this study can add to the
understanding of potential changes in the structure and function of these ecosystems, as well as
their contributions to the global carbon cycle and values as reservoirs for carbon.
Materials and Methods
Study Site
Intact monoliths (30 x 60 x 20 cm; n=18) of short-hydroperiod marsh (marl soils) were
collected just outside Everglades National Park (25°25'33. 55"N, 80°28'7. 37"W). Collection
occurred on sites dominated (> 90%) by sawgrass on March 18th 2011. Monoliths were cut out of
the ecosystem with their physical structure maintained (soil and roots ~20 cm deep collected,
down to the bedrock). Following collection, monoliths were placed in transportation tanks to
prevent desiccation and transported by truck to the University of Alabama. Monoliths were
removed from travel tanks and transplanted into permanent containers at the University’s flowcontrolled mesocosm facility (http://www.as.ua.edu/biolaqua/cfs/cfsmeso.htm) within 3 days of
collection, and allowed 8 weeks to recover from harvesting, travel, and transplanting.
Preliminary studies showed that at least 4 weeks was sufficient recovery time from both travel
and transplantation (Starr & Oberbauer unpublished data). During recovery, water levels were
maintained 20 cm above the soil surface to simulate inundated conditions regularly seen in the
13
short-hydroperiod marshes around Taylor Slough during the wet season (Jimenez et al., in 2012;
Schedlbauer et al., 2010, 2011).
The short hydroperiod freshwater marsh ecosystems in the Everglades region are
oligotrophic, subtropical wetlands with a year round growing season. Receiving ~1430 mm of
precipitation annually, the majority of rainfall occurs during the wet season (May to October)
with only 25% of annual precipitation falling during the dry season (November to April)
(National Climatic Data Center NCDC, http://www.ncdc.noaa.gov/). Severe drought occurs
every 4 to 6 years with La Niña events (Abtew et al., 2007). These wetlands are dominated by
the C3 species sawgrass (Cladium jamaicense), and the C4 species muhly grass (Muhlenbergia
capillaris).
Experimental Design
Using a randomized complete block design, each monolith (n=18) was randomly
assigned to one of two large flow tanks, then to a treatment (drought scenario: gradual,
intermediate, and rapid transition to simulated drought). The study was limited to two flow tanks
so that all monoliths were contained within one bay of the flow-controlled mesocosm facility to
isolate blocking variation. At the beginning of the experiment (following the 8-week recovery
period), water levels were maintained at the soil surface for 4 weeks to simulate inundated
conditions. Since water levels at Taylor Slough were ~20 cm below the surface at the peak of the
dry season in 2008 and 2010 (average years; Starr & Oberbauer unpublished data), all drought
scenarios targeted this water level. Within each tank, drought was simulated by changing
hydroperiods (raising monoliths above the water level) at 3 rates: gradual (20 cm change in 4
weeks), intermediate (20 cm change in 2 weeks), and rapid (20 cm change in 1 week). There
were 3 replicates of each treatment in each tank (n=6 per drought scenario). Once all monoliths
14
were elevated to 20 cm above the water level, blocks were maintained at this height for 3 weeks
to simulate moderate dry-season conditions (Jimenez et al., 2012). Following the dry period,
monoliths were re-inundated for 3 weeks to allow recovery from drought, and the manipulation
began again. Successive dry-downs were performed to test the additive effect of drought, a
scenario observed at Taylor Slough (2010 dry season) and a scenario likely to occur frequently in
the region as a result of climate change. The entire experiment occurred over a 22-week period
from June 2011 to October 2011.
Carbon Dynamic Measurements
CO2 Fluxes. NEE and ecosystem respiration (Reco) were measured each week of the
study with an infrared gas analyzer using a closed path chamber (LI-6200, LI-COR Inc., Lincoln,
NE) (Poorter 1993; Vourlitis et al., 1993; Oberbauer et al., 1998). Chamber temperature and
photosynthetically active radiation (PAR) were recorded using a Type-T, fine wire thermocouple
and a LI-COR LI-190S-1 quantum sensor attached to the LI-COR 6200 cuvette head.
Measurements were taken 3 times a day (sunrise: 6am-10am; noon: 11am-3pm, and dusk: 4pm8pm) at ambient light (NEE), and each measurement of NEE was followed by a dark
measurement (Reco) (Oberbauer et al., 1998). GEE was calculated as:
GEE=NEE-Reco
Eq. 1
CH4 Fluxes. Net methane flux was measured using the static chamber method (Whalen
& Reeburgh 1988; Vourlitis et al., 1993; Tsuyuzaki et al., 2001). Four static chambers (58.72 x
37.47 x126.15 cm) were used to measure methane fluxes on the 18 monoliths each week, with
two chambers randomly assigned to each tank. Headspace samples were drawn from each
chamber every 15 minutes for 45 minutes, using 3 cc syringes with needles permanently attached
with epoxy cement (Whalen & Reeburgh 1988). At each time point, duplicate samples were
15
collected. Data collection occurred during a 5-hour period (9am-2pm) during maximum CH4
fluxes (McDermitt et al., 2011). An SRI-310C (SRI International, Menlo Park, CA) FID gas
chromatograph was used to analyze samples. CH4 flux was calculated as the rate of
concentration change over the sampling period (Vourlitis et al., 1993).
Greenhouse Carbon Balance
To determine the greenhouse carbon balance of the short-hydroperiod marsh in relation to
simulated drought the greenhouse gas (carbon) exchange balance, the ratio of CH4 emissions to
net CO2 (mol/mol), and the greenhouse warming potential of methane for a 100-year time frame
(GWP=25) were used (IPCC 2007). Ratios were calculated using maximum midday CH4
emissions and midday NEE. The greenhouse carbon compensation point for a 100-year time
frame is 0.04 (1/GWP). Ratios of CH4 to CO2 greater than this value indicate CH4 emissions are
not offset by ecosystem productivity over a 100-year period, reducing the amount of carbon
stored in the ecosystem. Weekly changes in ratios over the study period for each drought
scenario were examined.
Physiological Potential
Canopy reflectance measurements were taken 30 cm above the canopy in the center of
each monolith on the same day that CO2 flux measurements were conducted, near solar noon,
using a UNI007 UniSpec-SC single channel spectrometer (PP Systems, Amesbury,
Massachusetts). Reflective indices such as the photochemical reflectance index (PRI; Gamon &
Pe uelas 1
), water index (WI; Pe uelas et al., 1993, 1997), normalized phaeophytinization
index (NPQI; Barnes et al., 1992) and the normalized difference vegetation index (NDVI;
modified from Tucker 1979; Fuentes et al., 2006; Claudio et al., 2006) were used to detect
16
changes in physiological potential and stress levels of the monoliths. PRI is generally used to
estimate light use efficiency, WI is associated with water content of the vegetation (Claudio et
al., 2006, Fuentes et al., 2006), NPQI is an estimate of chlorophyll degradation (Pe uelas &
Fiella 1998), and NVDI is an estimate of chlorophyll content and energy absorption (Fuentes et
al., 2006; Ollinger 2011).
Environmental Conditions
In addition to reflectance indices, environmental parameters within the facility bay that
are known to be relevant to carbon fluxes were monitored. Environmental conditions were
monitored continuously and data recorded half hourly with a CR1000 data logger and
multiplexer (Campbell Scientific, Logan, UT). Air and soil temperature at 5cm (type T, copperconstantan), PAR (LI 190; LI-COR Inc., Lincoln, NE), relative humidity (HMP-45C Vaisala),
temperature and relative humidity sensor (Campbell Scientific, Logan, UT), and barometric
pressure (CS106; Campbell Scientific) were measured using sensors attached to the data logger.
Soil volumetric water content (CS 615; Campbell Scientific, Logan, UT) was also monitored, in
one monolith in each treatment per tank during weeks 7-22. Measurements were made once
every 10 seconds and averaged over 30-minutes.
Biomass, Leaf Area, and Species Composition
All above ground biomass was harvested at the end of the experiment to examine dieback
and changes in species composition in response to drought simulation. Measurements of total
fresh and dry weight and total leaf area by species were made for each monolith. Initial fresh
weight was measured immediately following harvesting. Following fresh weights, leaf area was
measured by species using an LI-3000 (LI-COR Inc., Lincoln, NE). All live vegetation was
17
separated by species to determine changes in dominant species composition as compared to the
initial composition of monoliths that were dominated by sawgrass (>90%). All above ground
vegetation was dried for 48 hours at 60 in a Fisher Scientific Isotemp drying oven (Fisher
Scientific, Waltham, MA). Dry weights were recorded directly following removal from the
drying oven to prevent moisture absorption.
Data Analysis
As a preliminary step, I examined the effect of microclimatic variables on midday carbon
fluxes via correlation analysis. Pearson’s correlation coefficients were calculated for carbon
fluxes, as well as air temperature, PAR, relative humidity, and soil volumetric water content via
the SAS procedure PROC CORR.
To determine the effect of drought scenarios on carbon exchange I used repeated
measures analysis of variance methods, analyzing weekly flux data (NEE, Reco, GEE), reflective
indices (WI, NDVI, PRI, NPQI), and the ratio of CH4:CO2. I utilized mixed modeling methods,
with variance-covariance matrices explicitly formulated to appropriately account for both the
random effect of blocking (tanks) and the repeated measurement of each experimental unit over
time. Variance-covariance parameters were estimated via Restricted Maximum Likelihood
(REML) using the SAS procedure PROC MIXED. Models included fixed effects for drought
scenario and week, and a covariate was included for temperature. I also included a fixed effect to
delineate the experimental periods during the manipulation: inundation, manipulation (water
level lowering), dry periods (water levels 20cm below soil surface), and pulse events (1st day of
inundation) and drought simulations: simulation 1 (June 1st to August 10th) and simulation 2
(August 17 to October 26th). Random effects were included for tank and plot to appropriately
account for the physical design of the experiment, and week to account for the repeated measures
18
nature of the data. An additional analysis was performed for data collected during midday (9am2pm) (carbon fluxes, reflective indices, temperature, PAR, soil volumetric water content and
relative humidity). Mixed modeling methods were also used to test differences between live and
dead vegetation and/or drought scenarios for leaf area and vegetation weight at the end of the
experiment. Fixed effects included drought scenario and status (live or dead) and random effects
were included for tank and plot.
A modified backward selection method was used to determine the appropriate effects for
the final models. First, all parameters were included in the model, including all first-order
interactions between parameters. Then, the least significant effects based on the Wald chi-square
statistics were dropped one at a time until all remaining in the model were influential (p < 0.05).
Goodness of fit statistics, Akaike’s information criteria (AIC) and Bayesian information criteria
(BIC), were used to compare models. AIC and BIC are model selection statistics appropriate for
non-nested models, which measure how close fitted values are to true values, with a penalty for
the number of parameters in the model (Littell et al., 2006). At each step in model selection the
significance of each model parameter was evaluated and I ensured that the final model had the
lowest AIC and BIC values. To test for differences among levels of categorical variables, least
square means were produced, which are the marginal predicted mean values for the model
dependent variable given all other variables in the model are at their average values. Differences
among means were tested with the Tukey-Kramer multiple comparisons test. Assumptions of
normality and homoscedasticity were evaluated visually by plotting residuals.
19
Results
Carbon Dioxide
Significant changes were observed over the course of the drought experiment in NEE,
Reco, and GEE (Figure 1). Simple two-way Pearson correlations of midday NEE indicated that
the chlorophyll content of the vegetation (NDVI) increased linearly with increasing net carbon
uptake (p=0.008; r=- 0.13) (Table 1). Although temperature didn’t have a significant correlation
with midday NEE, further analysis showed a significant positive correlation with temperature for
the entire NEE dataset (sunrise, noon, dusk; p<0.001; r=-0.16). Repeated measures analysis of
net carbon uptake showed that the interaction between the time of day (diurnal: sunrise, noon, or
sunset) and experimental period (inundation, manipulation, dry period or pulse; p <0.001), and
temperature and the drought simulation (simulation 1: June 1st to August 10th and simulation 2:
August 17-October 26th; p<0.001) significantly influenced NEE (Table 2). NEE was highest
(lower net carbon uptake) during periods of manipulation at sunrise and sunset when
temperatures were lower (Figure 2d) and lowest during periods of inundation at midday (Figure
2d). The interaction between drought and temperature indicated net carbon uptake increased with
temperature during the first drought simulation while NEE remained constant as temperature
increased throughout the second drought simulation (Figure 2a).
Ecosystem respiration rates (Reco) were positively correlated with relative humidity
(p=0.005; r=0.14), and negatively correlated with water stress (WI: p=0.001; r= -0.14) and NDVI
(p=0.001, r=-0.17) (Table 1). Repeated measures analysis of ecosystem respiration (Reco) showed
that experimental period (p<0.001) and the interaction between temperature and the drought
simulation (p=0.020) were significant indicators of ecosystem respiration (Table 2; Figure 2b).
Drawdowns caused an increase in respiration and rates were highest during dry periods (Figure
20
1) until water stress limited respiration. Reco increased slightly when temperatures increased
during the first drought, but decreased slightly when temperatures increased during the second
drought period (Table 2; Figure 2b).
Time of day (diurnal; p <0.001), experimental period (p <0.001), and the interaction
between temperature and drought (p=0.015) were all significant predictors of GEE (Table 2).
GEE followed the same pattern as NEE with greater carbon uptake occurring at noon (p <0.001)
and during periods of inundation (p <0.001). Following the same pattern as Reco, during the first
drought there was a slight decrease in GEE (greater carbon uptake) as temperature increased,
while during the second drought there was a slight increase as temperature increased (Figure 2c).
All CO2 fluxes were significantly lower following the first drought simulation (Figure 2a - 2c),
and on average all drought scenarios were a net source of CO2 to the atmosphere (~0. 0 μmol m2 -1
s ).
Greenhouse Carbon Balance
Changes in the greenhouse carbon balance occurred as a result of drought simulation.
Relative humidity was positively correlated with the CH4:CO2 ratio (p=0.016; r= 0.12; Table 1),
more methane was emitted than net carbon uptake could offset as relative humidity increased.
And, as NDVI decreased due to higher water levels, methane levels rose (p= 0.010;r=-0.13).
Repeated measures analysis showed that there were no differences among drought scenarios (p
=0.497) and that NDVI was the only significant indicator of CH4:CO2 (p = 0.010; Table 3).
During periods of inundation, methane flux rates recovered faster than net carbon uptake causing
the system to become a source of carbon to the atmosphere (Figure 3). Though not significantly
higher, the gradual simulation scenario had the highest ratio of CH4:CO2 on average.
21
Physiological Potential
Physiological potential had strong correlations with environmental and flux variables
(Table 1). The chlorophyll content of the vegetation (NDVI) increased with NPQI (p<0.001;
r=0.35), PRI (p<0.001; r=0.26), temperature (p<0.001; r=0.32), pressure (p=0.001; r=0.18),
VWC (p=0.042, r=0.29), and net carbon uptake (NEE; p=0.008; r=-0.13), and decreased with
increasing ecosystem respiration (Reco; p=0.001; r=-0.17), CH4: CO2 (p=0.010; r=-0.13), WI
(p=0.001; r=-0.18), PAR (p=0.040; r=-0.10), and relative humidity (p<0.001; r=-0.27) (Table 1).
Repeated measures analysis of NDVI showed that changes in physiological potential were
significantly different over the times in the study period (Table 3). The interaction of drought
scenario and experimental period (p=0.023), NPQI (p<0.001), and WI (p<0.001) were all
significant indicators of NDVI (Table 3). NDVI was greatest during pulse events under the
gradual drought scenario and lowest during periods of manipulation under the intermediate
drought scenario (Figure 2e).
There were no significant differences between total leaf area at the end of the study for
the drought scenarios for all species combined (p = 0.282). However, total leaf area for sawgrass
was marginally significantly different by drought scenario (p= 0.157), with the gradual scenario
having the highest predicted least square mean (473.47 cm2± SE 105.24) versus the intermediate
(177.37 cm2± SE 105.24) and the rapid (263.12 cm2± SE 105.24) scenarios (Figure 4a). Prior to
the start of the study monoliths were dominated (>90%) by sawgrass. At the end of the study
sawgrass accounted for just 47% of average total leaf area while torpedograss, accounted for
35% of average total leaf area for the gradual and rapid drought scenarios (Figure 4a). Die back
as a result of drought simulation was observed under all drought scenarios and differences in dry
weight of live vegetation differed marginally by drought scenario (p=0.248) (Figure 4b). The
22
gradual transition to drought had the greatest predicted least square mean (42.43 g ± SE 10.18),
versus the intermediate (11.29 g ± SE 10.18) and rapid (12.75 g ± SE 10.18) drought scenarios.
High variability between plots led to marginal differences among drought scenarios for the
amount of vegetation and the species that sprouted following the final simulated drought (Figure
4a).
Discussion
As expected, drought turned the ecosystem into a source of carbon to the atmosphere by
increasing ecosystem respiration rates (Reco) and reducing net carbon exchange rates (NEE).
Over the entire study period, changes in NEE relative to recovering net CH4 emissions during
inundation resulted in changes in the greenhouse gas (carbon) exchange balance. These results
support my hypothesis that the increased frequency and duration of drought would increase CO2
and decrease CH4 emissions, cause a reduction in photosynthetic potential due to water stress,
and decrease the carbon storage potential of the ecosystem. Although not in situ responses to
drought, recent studies that considered the effects of hydrology on CO2 fluxes in Everglades’
freshwater marsh ecosystems (Schedlbauer et al., 2010; Jimenez et al., 2012) support these
results. Below I evaluate the effects of drought and drought onset on carbon fluxes and determine
how these changes impact the greenhouse carbon balance.
Effects of Drought on CO2 Flux Components
CO2 flux components are controlled by different biological processes, which vary in their
response to drought. Ecosystem carbon uptake is controlled by photosynthetic rates while
ecosystem respiration is the combination of both autotrophic and heterotrophic respiration.
Changes in GEE relative to changes in Reco are important for net carbon exchange rates.
Although sawgrass marshes in the Everglades region do not generally have very high
23
productivity rates, high carbon storage potential is the result of reduced decomposition rates
under anoxic conditions. In this wetland ecosystem, simulated droughts increased CO2 emissions
through a reduction in carbon uptake relative to ecosystem respiration. In response to drought
simulation, significant changes in NEE, Reco, GEE and the greenhouse carbon balance (net CH4:
NEE) were observed (Figure 1; Figure 2a - 2c; Figure 3), although drought onset rates had no
significant effect. The largest changes in carbon dynamics were in maximum photosynthetic
rates and during periods of inundation. As expected, there was greater net carbon exchange
during the midday measurements while inundated, when photosynthetic activity was greatest and
water stress was not limiting photosynthesis and respiration. During inundation, the ratio of Reco
to GEE was less than 1 (0.76), indicating GEE contributed more to net exchange rates. This
result is supported by other wetland studies that found GEE dominated NEE when water was not
limited (Alm & Walden 1999; Bubier et al., 2003a, b; Heinsch 2004; Hao et al., 2011; Webster et
al., 2013). Because photosynthetic CO2 fixation can be limited by leaf water potential, GEE is
impacted by water stress during prolonged drought (Heinsch 2004; Rocha & Goulden 2010;
Webster et al., 2013).
In wetland systems, soil respiration (Rs) dominates ecosystem respiration rates
accounting for up to 75% of Reco (Law et al., 2002). The results presented here are in line with
many other wetland ecosystems that show that Reco increases as a result of drought (Kramer &
Boyer 1995; Alm et al., 1999; Morison et al., 2000; Smith et al., 2003; Heinsch 2004; Rocha &
Goulden 2010; Webster et al., 2013). The increase in oxygen availability causes a rise in
heterotrophic respiration rates as the depth of soil oxygenation deepens and the rate of gas
diffusion into the atmosphere improves (Kramer & Boyer 1995; Morison et al., 2000; Law et al.,
2002; Bubier et al., 2003a, b; Smith et al., 2003; Heinsch 2004; Aurela et al., 2007; Barr et al.,
24
2010; Schedlbauer et al., 2010; Jimenez et al., 2012; Webster et al., 2013). An increase in
oxygen availability and diffusion rates likely caused Reco to dominate net carbon exchange
during drought simulation, indicated by a ratio of Reco to GEE greater than 1 (1.39). Reco is the
dominant flux component controlling NEE in wetland ecosystems (Jimenez et al., 2012). These
results are supported by Jimenez et al., (2012), which compared CO2 fluxes between short and
long hydroperiod marsh ecosystems in the Everglades and found that during drought, faster
diffusion and higher heterotrophic respiration rates increased ecosystem respiration. While
inundated, Reco was reduced due to lower diffusion rates out of the soil and anoxic conditions;
under drought, limitations on diffusion and oxygen stress are lifted (Jimenez et al., 2012).
Prior to drought, GEE dominated net exchange rates. During the first drought simulation,
there were no distinct changes in GEE but there was an increase in ecosystem respiration (Figure
1). Drought depressed NEE values are a result of increased Reco and not a change in GEE (Aurela
et al., 2007). However, after the first drought simulation water availability began to limit
photosynthesis, reducing carbon uptake rates and ecosystem respiration. Other wetland studies
have observed reductions in carbon uptake as a result of water stress during the dry season and
under drought conditions (Heinsch 2004; Rocha & Goulden 2010). The relationship between
carbon flux, experimental period, and the water index suggests that water stress was reducing the
capacity of this system to acquire carbon. Under normal conditions photosynthesis exceeds
respiration in plants. When photosynthesis is inhibited but respiration continues, respiration will
proceed as long as photosynthetic reserves are available in drought intolerant plants. In drought
tolerant plants like sawgrass, when photosynthesis declines, so do respiration rates (Kramer &
Boyer 1995). Eventually, dehydration inhibits respiration in all plants since water is required for
the hydrolytic reactions associated with respiration. Following the first drought when all fluxes
25
were reduced, water stress likely limited heterotrophic respiration in the soil as well (Smith et al.,
2003). As the second drought began, both photosynthesis and respiration rates were impacted
and resulted in reduced CO2 fluxes (Figure 1). Once water began to limit ecosystem productivity,
there was a great reduction in GEE and even during periods of inundation, exchange rates did not
recover.
The Greenhouse Carbon Balance
The greenhouse gas (carbon) exchange balance remained relatively constant over the
study period, yet spikes (increases in net CH4 relative to NEE) were observed during periods of
inundation. During drought, methane emissions were null. Consequently, NEE offset net CH4
rates. Following drought, inundation caused an increase in net CH4 relative to NEE resulting in a
reduction in the ratio of CH4:CO2. Prior research in wetland ecosystems has found that methane
fluxes increase with higher soil moisture levels (King & Skovgaard 1990; Bachoon & Jones
1992; Torn & Chapin 1993; Smith et al., 2003; Whalen 2005; Webster et al., 2013). However, as
the depth to oxygenation increases, soils become a sink for methane (Bachoon & Jones 1992;
Smith et al., 2003). Even during periods of inundation, sawgrass marshes are only a weak source
of methane to the atmosphere (Bachoon & Jones 1992).
In this ecosystem, methane moves out of the soil through ebullition. Ebullition is difficult
to quantify due to its stochastic nature (Tokida et al., 2007; Goodrich et al., 2011), reducing the
ability to find a significant difference between drought onset rates and the experimental period.
Although no significant difference was detected between drought scenarios, water availability
does influence methanogenesis and CH4 movement out of the soil (Bachoon & Jones 1992; Torn
& Chapin 1993; Smith et al., 2003; Whalen 2005; Goodrich et al., 2011; Webster et al., 2013).
The rapid transition to drought had the highest average ratio of net CH4 to NEE (0.16), followed
26
by the gradual transition (0.08), and finally the intermediate (0.03) transition to drought. Under
inundation CH4 production by methanogens persists, and when conditions become aerobic, both
methane oxidation rates and CH4 diffusion rates increase (Smith et al., 2003). Therefore, the
swift reduction in water levels in the rapid drought scenario may have enhanced CH4 diffusion
and ebullition out of the soil while the slow reductions in water levels may have aided greater
CH4 production and increased probabilities of methane oxidation. These results indicate drought
onset rate may actually influence CH4 diffusion out of the soil so that rapid changes in water
levels promote CH4 emission, whereas more gradual transitions allow for slower diffusion and
therefore increased rates of methane oxidation.
The 100-year greenhouse warming compensation point for methane is 0.04, indicating
that both the rapid and gradual transition to drought caused the marsh monoliths to be a source of
methane to the atmosphere over the long term. Overall the average ratio of net CH4 to NEE for
the study period was 0.06, showing that even under extended drought, freshwater marsh
ecosystems can be a source of methane to the atmosphere if inundation periods are sufficient for
CH4 production. Extended drought conditions also promote greater CO2 release and reductions in
CO2 uptake, further reducing the ratio of CH4:CO2.
Changes in Photosynthetic Potential
The effect of drought on carbon uptake rates can be explained by reductions in
photosynthetic potential, an effect of water stress. Although the water content of the vegetation
showed no significant differences between drought scenarios, the water index was a good
indicator of the onset and persistence of water stress throughout the study. Water content of the
vegetation is important in determining the response to rehydration following inundation (Kramer
& Boyer 1995). The ability to recover photosynthetic capacity when water is re-supplied is
27
influenced by the extent and duration of dehydration (Kramer & Boyer 1995; Lambers & Chapin
2008). As a result of dehydration, vascular blockages occur in the xylem from the cavitation
caused by high water tension (Lambers & Chapin 2008). More severe dehydration results in
higher tension and more frequent blockages, which can lead to cell death, further limiting
recovery when water is resupplied (Kramer & Boyer 1995; Lambers & Chapin 2008). Premature
leaf senescence is a common effect of dehydration (Kramer & Boyer 1995; Lambers & Chapin
2008). This may have also influenced the patterns observed in NDVI. NDVI showed differences
between drought scenarios with the interaction of experimental period (inundation, manipulation,
drought, and pulse events) suggesting that there were significant changes in the photosynthetic
capacity of the ecosystem in response to changes in water availability. Photosynthesis can also
be inhibited by low cellular water potentials through changes in solute concentrations (Kramer &
Boyer 1995; Lambers & Chapin 2008). Changes in the solute environment around enzymes often
lead to reductions in photophosphorylation activity (Lambers & Chapin 2008).
Reflective indices supported patterns observed in carbon flux components. Net carbon
uptake increased with NDVI and water availability. CH4:CO2, which also increased with water
levels, had a significant negative relationship with NDVI, indicating that although productivity
increased with water availability, CH4 recovered and produced greater fluxes than NEE could
account for. Previous studies have found significant relationships between NEE and CH4
production in wetland ecosystems (Whiting & Chanton 1993; Joabsson & Christensen 2001;
Mitsch et al., 2012). Normally, greater productivity relates to higher NDVI and larger methane
emissions. The decline in net carbon uptake as water was resupplied and increased
methanogenesis caused this negative relationship between CH4:CO2 and NDVI. The correlation
between NDVI and net methane flux indicates photosynthetic rates may serve as an indicator of
28
methane flux by integrating environmental variables important in methanogenesis (Whiting &
Chanton 1993; Joabsson & Christensen 2001).
Although the components of the CO2 flux did not differ significantly by drought scenario,
there was an effect of drought onset rate on the final biomass. Final biomass harvesting indicated
a reduction in the recovery of monoliths experiencing faster transitions and longer drought
durations. Drought simulation resulted in dieback in all scenarios but the amount of sawgrass
that re-sprouted following the final simulated drought was greater for the gradual drought
scenario than it was for the intermediate and rapid drought scenarios. Reductions in sawgrass
recovery may have also facilitated a community shift in the gradual and rapid transition to
drought where torpedograss rhizomes where present. Prior to the start of the study, monoliths
were dominated (>90%) by sawgrass. At the end of the study, torpedograss accounted for 35% of
average total leaf area for the gradual drought scenario. Torpedograss is an invasive species that
has taken over 70% of Florida’s public waters (Center for Aquatic and Invasive Plants
http://plants.ifas.ufl.edu/node/308). Mechanisms whereby invasive plants suppress other species
include dense rhizomes and roots that leave little space for neighbors, strong competition for
nutrients (Perry & Galatowitsch 2004), tall dense canopies that intercept light, and faster
establishment following disturbance (Herr-Turoff 2005; Zedler and Kercher 2004). The gradual
and rapid drought scenario provided conditions that facilitated establishment of torpedograss
following sawgrass dieback. Torpedograss culms originate from rhizomes, are drought tolerant,
and thrive in well drained to poorly drained soils. The rhizomes are in the short-hydroperiod
marsh ecosystem and pose a threat of invasion following disturbance.
29
Global Impacts
These results indicate that fluctuations in hydrology due to climate or water management,
an issue of great concern for wetlands globally, could cause significant changes in the structure
and function of these ecosystems. Hydrology drives the carbon sequestering capacity of wetlands
and has resulted in the accumulation of large belowground pools. This in turn makes wetland
contributions to the global carbon cycle much greater than their land area would imply. Climate
induced changes in hydrology could also distort the structure of these ecosystems by stressing
local vegetation and enhancing conditions for invasive species establishment. This study
demonstrates the complex relationship between hydrology and carbon cycling and how the value
of wetlands as reservoirs for carbon might diminish amid changes in climate and water
management.
30
Table 1. Pearson correlation coefficients (r) and their associated p-values for midday measurements of CO2 flux components,
reflective indices (NDVI, NPQI, WI), and environmental factors (Air temperature, PAR, barometric pressure, and relative humidity).
r
p-value
NEE
Reco
GEE
CH4: CO2
0.01
0.899
-0.04
0.409
0.04
0.428
NDVI
-0.13
0.008
-0.17
0.001
0.04
0.435
-0.13
0.010
PRI
-0.02
0.729
0.01
0.167
-0.07
0.148
-0.02
0.074
0.26
<0.001
NPQI
0.002
0.958
0.03
0.518
-0.03
0.577
-0.04
0.433
0.35
<0.001
0.89
<0.001
WI
-0.01
0.785
-0.14
0.001
0.11
0.030
0.04
0.451
-0.18
<0.001
-0.87
-0.89
<0.001 <0.001
Tair
0.015
0.766
0.004
0.935
0.002
0.961
0.04
0.433
0.32
<0.001
-0.29
-0.41
0.25
<0.001 <0.001 <0.001
PAR
0.08
0.118
-0.05
0.367
-0.034
0.504
-0.10
0.040
-0.003
0.946
Pressure
CH4: CO2
NDVI
PRI
NPQI
-0.08
0.114
WI
-0.06
0.274
Tair
PAR
Pressure
0.19
<0.001
0.06
0.261
-0.07
0.168
0.013
0.799
-0.03
0.508
0.18
<0.001
-0.21
-0.10
<0.001 0.053
0.17
0.13
<0.001 0.010
-0.02
0.699
Relative
Humidity
0.02
0.755
0.14
0.005
-0.12
0.013
0.12
0.016
-0.27
<0.001
0.10
0.059
0.10
0.193
-0.11
0.023
-0.03
0.528
-0.65
<0.001
-0.24
<0.001
VWC
0.22
0.139
0.08
0.457
-0.02
0.888
-0.06
0.685
0.29
0.042
0.29
0.050
0.14
0.354
-0.03
0.844
-0.28
0.058
-0.22
0.138
-0.24
0.105
33
Relative
Humidity
-0.19
0.203
Table 2. Estimated fixed effects, standard errors, and p-values for the repeated measures analysis of net ecosystem exchange (NEE),
ecosystem respiration (Reco) and gross ecosystem exchange (GEE).
Parameters
Intercept
Drought Scenario
Gradual vs. Rapid
Intermediate vs. Rapid
Diurnal
Sunrise vs. Sunset
Noon vs. Sunset
Experimental Period
Inundation vs. Pulse
Manipulation vs. Pulse
Dry period vs. Pulse
Diurnal* Experimental Period
Sunrise vs.
Inundation vs.
Sunset
Pulse
Sunrise vs.
Manipulation vs.
Sunset
Pulse
Sunrise vs.
Dry period vs.
Sunset
Pulse
Noon vs.
Inundation vs.
Sunset
Pulse
Noon vs.
Manipulation vs.
Sunset
Pulse
Noon vs.
Dry period vs.
Sunset
Pulse
Drought (First vs. Second)
Temperature
Temperature *Drought
First vs. Second
p-value
Estimate
0.2978
NEE
Std.
Errors
0.2304
0.197
0.5254
-0.0725
0.0013
0.084
0.084
0.403
0.988
-0.0588
-0.1195
0.3265
-0.2332
0.1612
0.1564
0.043
0.136
0.3181
1.1727
-0.2332
0.1728
0.1896
0.1564
0.066
<0.001
0.687
-0.7019
0.2017
0.001
-0.7821
0.2121
<0.001
-0.2555
0.2112
0.227
-0.7198
0.2012
<0.001
-0.3231
0.2075
0.120
0.3226
0.2089
0.123
1.3475
-0.0632
0.4011
0.1566
-0.0474
0.0118
Estimate
Reco
Std.
Errors
0.116
p-value
Estimate
<0.001
-0.1539
GEE
Std.
Errors
0.2385
0.668
0.388
-0.0115
0.1299
0.1421
0.1421
0.937
0.375
0.0183
-0.5321
0.0875
0.078
0.834
<0.001
p-value
0.519
0.3217
0.996
0.1528
0.1537
0.1861
0.1176
0.037
<0.001
0.195
-0.5227
-0.2495
-0.2225
0.1503
0.1802
0.1123
0.001
0.17
0.048
0.001
0.506
-0.1282
-0.0008
0.2791
0.0004
0.646
0.048
0.2362
0.0008
0.3595
0.0007
0.511
0.244
<0.001
0.0169
0.0007
0.020
-0.0251
0.0103
0.015
34
Table 3. Estimated fixed effects, standard errors and p-values for the repeated measures analysis of the ratio of net CH4 flux to net
CO2 uptake, and the normalized difference vegetation index (NDVI).
Parameters
Intercep
t
Drought Scenario
Estimate
CH4:CO2
Std.
Errors
p-value
Estimate
2.9156
0.4266
2.4586
3.1601
0.4506
0.059
0.0383
0.0224
-1.3962
3.2424
0.673
0.0336
0.0383
0.3835
0.0128
0.0009
0.0246
0.0266
0.6019
0.7325
-0.0286
0.0265
0.2798
-0.0998
-0.1184
-0.0491
-0.0611
-0.0957
-0.0389
0.035
0.0376
0.0373
0.0348
0.0376
0.0373
0.0008
0.002
0.1895
0.08
0.0114
0.299
NPQI
0.2247
0.0212
<0.0001
WI
0.0998
0.0134
<0.0001
Intermediate vs. Rapid
Experimental Period
Inundation vs. Pulse
Manipulation vs. Pulse
Dry period vs. Pulse
Drought Scenario* Experimental Period
Gradual vs. Rapid
Inundation vs. Pulse
Gradual vs. Rapid
Manipulation vs. Pulse
Gradual vs. Rapid
Dry period vs. Pulse
Intermediate vs. Rapid
Inundation vs. Pulse
Intermediate vs. Rapid
Manipulation vs. Pulse
Intermediate vs. Rapid
Dry period vs. Pulse
NDVI
-38.19
14.806
35
0.0315
p-value
2.5699
Gradual vs. Rapid
-0.016
NDVI
Std.
Errors
0.6155
0.0104
Figure 1. Changes in average NEE, Reco and GEE over the study period in response to changes in water level for all drought scenarios.
NEE and GEE were lowest (greater carbon uptake) during periods of inundation and highest (greater carbon release to the
atmosphere) during drought simulations. Reco increased during drought simulation up until water began limiting respiration rates.
36
Figure 2. Least square mean predicted values of the interactions between: (a) temperature and drought simulation for NEE, (b) Reco
and (c) GEE, and (d) diurnal (sunrise, noon, and sunset) and experimental period (Inundation, manipulation, dry period, and pulse
events) for NEE and (e) NDVI. Increasing temperatures resulted in higher fluxes in drought simulation 1 compared to drought
simulation 2 (a-c). Net carbon uptake was greatest during midday measurement when monoliths were inundated and carbon release
was greatest during manipulations (d-e). The chlorophyll content of the vegetation was highest for the gradual drought scenario during
pulse events (d).
37
Figure 3. Changes in average CH4 to net CO2 exchange over the study period in relation to average changes in water levels and the
greenhouse carbon compensation point for 100 years (0.04). This system is a small source of carbon to the atmosphere due to the lack
of recovery in NEE during periods of inundation.
38
Figure 4. (a) Average leaf area for each drought scenario by species contribution at the end of the
experiment. There were no significant differences in total leaf area however, sawgrass leaf area
had marginal differences for the drought scenarios (p = 0.157). The gradual transition to drought
had greater average sawgrass leaf area (473.47 g ± SE 105.24) compared to the intermediate
(177.37 g ± SE 105.24) and the rapid (263.11 g ± SE 105.24) drought scenarios. (b) Average dry
weight of above ground biomass for each drought scenario at the end of the experiment. There
was no significant difference in the amount of dead biomass between drought simulations.
However, there was a marginal difference in the amount of live vegetation at the end of the
experiment (p=0.248). The gradual drought scenario had the greatest predicted least square mean
(42.43 g ± SE 10.18), versus the intermediate (11.29 g ± SE 10.18) and rapid (12.75 g ± SE
10.18) drought scenarios.
39
References
Abtew, W., Huebner, R. S., & Pathak, C. (2007). Hydrology and Hydraulics of South Florida.
World Environmental and Water Resources Congress 2007 ASCE Conference in Tampa,
Florida.
Adhikari, S., Bajracharaya, R. M., & Sitaula, B. K. (2009). A Review of Carbon Dynamics and
Sequestration in Wetlands. Journal of Wetlands Ecology, 2(1).
doi:10.3126/jowe.v2i1.1855.
Alm, J., Schulman, L., Walden, J., Nykänen, H., Martikainen, P. J., & Silvola, J. (1999). Carbon
balance of a boreal bog during a year with an exceptionally dry summer. Ecology, 80(1),
161–174. doi:10.1890/0012-9658(1999)080[0161:CBOABB]2.0.CO;2.
Armentano, T. V. (1980). Drainage of Organic Soils as a Factor in the World Carbon Cycle.
BioScience, 30, 825–830. doi:10.2307/1308375.
Aurela, M., Riutta, T., Laurila, T., Touvinen, J. P., Vesala, T., Tuittila, E. S., et al. (2007). CO2
exchange of a sedge fen in southern Finland—the impact of a drought period. Tellus B,
59(5), 826–837. doi:10.1111/j.1600-0889.2007.00309.x.
Bachoon, D., & Jones, R. D. (1992). Potential rates of methanogenesis in sawgrass marshes with
peat and marl soils in the everglades. Soil Biology and Biochemistry, 24(1), 21–27.
doi:10.1016/0038-0717(92)90237-R.
Barnes, J. D., Balaguer, L., Manrique, E., Elvira, S., & Davison, A. W. (1992). A reappraisal of
the use of DMSO for the extraction and determination of chlorophylls a and b in lichens
and higher plants. Environmental and Experimental Botany, 32(2), 85–100.
doi:10.1016/0098-8472(92)90034-Y.
Barr, J. G., Engel, V., Fuentes, J. D., Zieman, J. C., O'Halloran, T. L., Smith, T. J., & Anderson,
G. H. (2010). Controls on mangrove forest‐ atmosphere carbon dioxide exchanges in
western Everglades National Park. Journal of Geophysical Research: Atmospheres
(1984–2012), 115(G2), G02020. doi:10.1002/(ISSN)1525-2027.
Brevik, E. C., & Homburg, J. A. (2004). A 5000 year record of carbon sequestration from a
coastal lagoon and wetland complex, Southern California, USA. Catena, 57(3), 221–232.
doi:10.1016/j.catena.2003.12.001.
Bubier, J., Crill, P., Mosedale, A., Frolking, S., & Linder, E. (2003a). Peatland responses to
varying interannual moisture conditions as measured by automatic CO2 chambers.
Global Biogeochemical Cycles, 17(2). doi:10.1029/2002GB001946.
Bubier, J. L., Bhatia, G., Moore, T. R., Roulet, N. T., & Lafleur, P. M. (2003b). Spatial and
Temporal Variability in Growing-Season Net Ecosystem Carbon Dioxide Exchange at a
40
Large Peatland in Ontario, Canada. Ecosystems, 6(4), 353–367. doi:10.1007/s10021-0030125-0.
Childers, D. L. (2006). A synthesis of long-term research by the Florida Coastal Everglades
LTER Program. Hydrobiologia, 569(1), 531–544. doi:10.1007/s10750-006-0154-8.
Choi, Y., & Wang, Y. (2004). Dynamics of carbon sequestration in a coastal wetland using
radiocarbon measurements. Global Biogeochemical Cycles, 18. doi:10.1002/(ISSN)15252027.
Claudio, H., Cheng, Y., Fuentes, D., Gamon, J., Luo, H., Oechel, W., et al. (2006). Monitoring
drought effects on vegetation water content and fluxes in chaparral with the 970 nm water
band index. Remote Sensing of Environment, 103(3), 304–311.
doi:10.1016/j.rse.2005.07.015.
Davis, S. M., & Ogden, J. C. (Eds.). (1994). Everglades: The Ecosystem and Its Restoration. St.
Lucie Press, Boca Raton , FL.
Dugan, P. (Ed.). (1993). Wetlands in Danger in a World Conservation Atlas. Oxford University
Press: New York City, NY.
Fuentes, D., Gamon, J., Cheng, Y., Claudio, H., Qiu, H., Mao, Z., et al. (2006). Mapping carbon
and water vapor fluxes in a chaparral ecosystem using vegetation indices derived from
AVIRIS. Remote Sensing of Environment, 103(3), 312–323.
doi:10.1016/j.rse.2005.10.028.
Gamon, J. A., Penuelas, J., & Field, C. B. (1992). A narrow-waveband spectral index that tracks
diurnal changes in photosynthetic efficiency. Remote Sensing of Environment, 41(1), 35–
44. doi:10.1016/0034-4257(92)90059-S.
Goodrich, J. P., Varner, R. K., Frolking, S., Duncan, B. N., & Crill, P. M. (2011). High‐
frequency measurements of methane ebullition over a growing season at a temperate
peatland site. Geophysical Research Letters, 38(7). doi:10.1029/2011GL046915.
Hao, Y., Cui, X. Y., Wang, Y. F., Mei, X. R., Kang, X. M., Wu, N., et al. (2011). Predominance
of Precipitation and Temperature Controls on Ecosystem CO2 Exchange in Zoige Alpine
Wetlands of Southwest China. Wetlands, 31(2), 413–422. doi:10.1007/s13157-011-01511.
Heinsch, F., Heilman, J. L., McInnes, K. J., Cobos, D. R., Zubere, D. A., & Roelke, D. L. (2004).
Carbon dioxide exchange in a high marsh on the Texas Gulf Coast: effects of freshwater
availability. Agricultural and Forest Meteorology, 125(1-2), 159–172.
doi:10.1016/j.agrformet.2004.02.007.
Herr-Turoff, A. M. (2005). Responses of an invasive grass, Phalaris arundinacea, to excess
resources. Department of Botany, Madison, WI, 128.
41
IPCC. (2007). Climate Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY,
USA.
Jimenez, K. L., Starr, G., Staudhammer, C. L., Schedlbauer, J. L., Loescher, H. W., Malone, S.
L., & Oberbauer, S. F. (2012). Carbon dioxide exchange rates from short- and longhydroperiod Everglades freshwater marsh. Journal of Geophysical Research, 117(G4),
G04009. doi:10.1029/2012JG002117.
Joabsson, A., & Christensen, T. R. (2001). Methane emissions from wetlands and their
relationship with vascular plants: an Arctic example. Global Change Biology, 7(8), 919–
932. doi:10.1046/j.1354-1013.2001.00044.x.
King, G. M., Roslev, P., & Skovgaard, H. (1990). Distribution and rate of methane oxidation in
sediments of the Florida everglades. Applied and Environmental Microbiology, 56.
Kramer, P. J., & Boyer, J. S. (1995). Water relations of plants and soils. Academic Press, San
Diego, California.
Lambers, H., Chapin, F. S., & Pons, T. L. (2008). Plant Physiological Ecology. New York, NY:
Springer New York. doi:10.1007/978-0-387-78341-3.
Law, B. E., Falge, E., Gu, L., Baldocchi, D. D., Bakwin, P., Berbigier, P., et al. (2002).
Environmental controls over carbon dioxide and water vapor exchange of terrestrial
vegetation. Agricultural and Forest Meteorology, 113(1-4), 97–120. doi:10.1016/S01681923(02)00104-1.
Littell, R. C., Milliken, G. A., Stroup, W. W., Wolfinger, R. D., & Schabenberger, O. (2006).
SAS for Mixed Models. Caryn, NC, USA: SAS Institute.
McDermitt, D., Burba, G., Xu, L., Anderson, T., Komissarov, A., Riensche, B., et al. (2011). A
new low-power, open-path instrument for measuring methane flux by eddy covariance.
Applied Physics B, 102(2), 391–405. doi:10.1007/s00340-010-4307-0.
Mitra, S., Wassmann, R., & Vlek, P. (2005). An appraisal of global wetland area and its organic
carbon stock. Current Science, 88, 25–35.
Mitsch, Gosslink. (2007). Wetlands (4 ed.). John Wiley & Sons, Inc., Hoboken, New Jersey.
Mitsch, W. J., Bernal, B., Nahlik, A. M., Mander, Ü., Zhang, L., Anderson, C. J., et al. (2013).
Wetlands, carbon, and climate change. Landscape Ecology, 28(4), 583–597.
doi:10.1007/s10980-012-9758-8.
Morison, J. I. L., Piedade, M. T. F., Müller, E., Long, S. P., Junk, W. J., & Jones, M. B. (2000).
Very high productivity of the C4 aquatic grass Echinochloa polystachya in the Amazon
42
floodplain confirmed by net ecosystem CO2 flux measurements. Oecologia, 125(3), 400–
411. doi:10.1007/s004420000464.
Oberbauer, S. F., Starr, G., & Pop, E. W. (1998). Effects of extended growing season and soil
warming on carbon dioxide and methane exchange of tussock tundra in Alaska. Journal
of Geophysical Research: Atmospheres (1984–2012), 103(D22), 29075–29082.
doi:10.1029/98JD00522.
Ollinger, S. V. (2011). Sources of variability in canopy reflectance and the convergent properties
of plants. New Phytologist, 189(2), 375–394. doi:10.1111/j.1469-8137.2010.03536.x.
Pearlstine, L. G., Pearlstine, E. V., & Aumen, N. G. (2010). A review of the ecological
consequences and management implications of climate change for the Everglades.
Journal of the North American Benthological Society, 29(4), 1510–1526. doi:10.1899/10045.1.
Penuelas, J., Pinol, J., & Ogaya, R. (1997). Estimation of plant water concentration by the
reflectance water index WI (R900/R970). International Journal of Remote Sensing,
18(13), 2869–2875.
Penuelas, J., & Filella, I. (1998). Visible and Near-Infrared Reflectance Techniques for
Diagnosing Plant Physiological Status. Trends in Plant Science, 3(4), 151–156.
Perry, W. (2004). Elements of south Florida's comprehensive Everglades restoration plan.
Ecotoxicology, 13, 185–193.
Perry, L. G., Galatowitsch, S. M., & Rosen, C. J. (2004). Competitive control of invasive
vegetation: a native wetland sedge suppresses Phalaris arundinacea in carbon‐ enriched
soil. Journal of Applied Ecology, 41(1), 151–162. doi:10.1111/j.13652664.2004.00871.x.
Poorter, L., & Oberbauer, S. F. (1993). Photosynthetic induction responses of two rainforest tree
species in relation to light environment. Oecologia, 96(2), 193–199.
doi:10.1007/BF00317732.
Rocha, A. V., & Goulden, M. L. (2010). Drought legacies influence the long‐ term carbon
balance of a freshwater marsh. Journal of Geophysical Research: Atmospheres (1984–
2012), 115(G3), G00H02. doi:10.1029/2009JG001215.
Reddy, K. R., & DeLaune, R. D. (2008). Biogeochemistry of wetlands. Boca Raton, FL: CRC
Press.
Schedlbauer, J. L., Oberbauer, S. F., Starr, G., & Jimenez, K. L. (2010). Seasonal Differences in
the CO2 Exchange of a Short-Hydroperiod Florida Everglades Marsh. Agricultural and
Forest Meteorology, 150(2010), 994–1006. doi:10.1016/j.agrformet.2010.03.005.
43
Schedlbauer, J. L., Oberbauer, S. F., Starr, G., & Jimenez, K. L. (2011). Controls on sensible
heat and latent energy fluxes from a short-hydroperiod Florida Everglades marsh. Journal
of Hydrology, 411(3-4), 331–341. doi:10.1016/j.jhydrol.2011.10.014.
Schedlbauer, J. L., Munyon, J. W., Oberbauer, S. F., & Gaiser, E. E. (2012). Controls on
ecosystem carbon dioxide exchange in short- and long-hydroperiod Florida Everglades
freshwater marshes. Wetlands, 32, 801–812. doi:10.1007/s13157-012-0311-y.
Smith, K. A., Ball, T., Conen, F., Dobbie, K. E., Massheder, J., & Rey, A. (2003). Exchange of
greenhouse gases between soil and atmosphere: interactions of soil physical factors and
biological processes. European Journal of Soil Science, 54(4), 779–791.
doi:10.1046/j.1351-0754.2003.0567.x.
Todd, M. J., Muneepeerakul, R., Pumo, D., Azaele, S., Miralles-Wilhelm, F., Rinaldo, A., &
Rodriguez-Iturbe, I. (2010). Advances in Water Resources. Advances in Water
Resources, 33(10), 1279–1289. doi:10.1016/j.advwatres.2010.04.003.
Tokida, T., Miyazaki, T., Mizoguchi, M., & Nagata, O. (2007). Falling atmospheric pressure as a
trigger for methane ebullition from peatland. Global Biogeochemical ….
doi:10.1029/2006GB002790.
Torn, M. S., & Chapin, F. S., III. (1993). Environmental and biotic controls over methane flux
from Arctic tundra. Chemosphere, 26(1-4), 357–368. doi:10.1016/0045-6535(93)90431-4
Tsuyuzaki, S., Nakano, T., Kuniyoshi, S.-I., & Fukuda, M. (2001). Methane flux in grassy
marshlands near Kolyma River, north-eastern Siberia. Soil Biology and Biochemistry,
33(10), 1419–1423. doi:10.1016/S0038-0717(01)00058-X.
Tucker, C. J. (1979). Red and photographic infrared linear combinations for monitoring
vegetation. Remote Sensing of Environment, 8(2), 127–150. doi:10.1016/00344257(79)90013-0.
Vourlitis, G. L., Oechel, W. C., Hastings, S. J., & Jenkins, M. A. (1993). A System for
Measuring in Situ CO2 and CH4 Flux in Unmanaged Ecosystems: an Arctic Example.
Functional Ecology, 7(3), 369–379.
Webster, K. L., McLaughlin, J. W., Kim, Y., Packalen, M. S., & Li, C. S. (2013). Modelling
carbon dynamics and response to environmental change along a boreal fen nutrient
gradient. Ecological Modelling, 248, 148–164. doi:10.1016/j.ecolmodel.2012.10.004.
Whalen, S. C., & Reeburgh, W. S. (1988). A methane flux time series for tundra environments.
Global Biogeochemical Cycles, 2(4), 399–409. doi:10.1029/GB002i004p00399.
Whalen, S. C. (2005). Biogeochemistry of methane exchange between natural wetlands and the
atmosphere. Environmental Engineering Science, 22(1), 73–94.
doi:10.1089/ees.2005.22.73.
44
Whiting, G. J., & Chanton, J. P. (1992). Plant‐ dependent CH4 emission in a subarctic Canadian
fen. Global Biogeochemical Cycles, 6(3), 225–231. doi:10.1029/92GB00710.
Whiting, G. J., & Chanton, J. P. (1993). Primary production control of methane emission from
wetlands. , Published Online: 26 August 1993; | Doi:10.1038/364794a0, 364(6440),
794–795. doi:10.1038/364794a0.
Whiting, G. J., & Chanton, J. P. (2001). Greenhouse carbon balance of wetlands: methane
emission versus carbon sequestration. Tellus B, 53(5), 521–528. doi:10.1034/j.16000889.2001.530501.x.
Zedler, J. B., & Kercher, S. (2010). Causes and Consequences of Invasive Plants in Wetlands:
Opportunities, Opportunists, and Outcomes. Dx.Doi.org, 23(5), 431–452.
doi:10.1080/07352680490514673.
Zedler, J. B., & Kercher, S. (2005). Wetland ResourcesS: Status, Trends, Ecosystem Services,
and Restorability. Annual Review of Environment and Resources, 30(1), 39–74.
doi:10.1146/annurev.energy.30.050504.144248.
45
CHAPTER 3: SEASONAL PATTERNS IN ENERGY PARTITIONING OF TWO
FRESHWATER MARSH ECOSYSTEMS IN THE FLORIDA EVERGLADES
Abstract
Hydrology is the single most important factor in determining the structure and function of
wetlands, including their energy balance. The components of the ecosystem energy balance
provide insight into the relationship between evapotranspiration (ET), a major source of water
loss in wetlands, and ecosystem hydrology. Improving the understanding of these links, as well
as their environmental controls, is important for determining how energy balance changes
through wet and dry seasons, as well as with future climate. This is of particular importance in
the Florida Everglades, given its unique geomorphology and climatology (Chen & Gerber 1992)
and the alteration of its hydrology through land management in the last century (Davis & Ogden
1994). To aid the understanding of the relationship between hydrology and energy fluxes, I
analyzed energy partitioning in two Everglades ecosystems that differ in hydrologic regimes. I
examined energy balance components (latent energy and sensible heat; LE and H respectively)
through several wet and dry seasons relative to climatic conditions and variable water levels to
gain better mechanistic information about the control of and changes in marsh hydrology.
Net radiation (Rn) was predominantly partitioned into LE (>50%), resulting in daytime
Bowen ratios <1, a global characteristic of wetland ecosystems. Unique to this system is the
seasonal shift in surface fluxes (H and LE) and how they feedback to the regional climate and
hydrology. In the Everglades, surface fluxes feedback to wet season precipitation and affect the
magnitude of seasonal change in water levels through water loss as LE, an estimate for ET. At
46
both sites, annual precipitation (1304 and 1207 mm yr-1, at the short- and long-hydroperiod sites,
respectively) was higher than ET (1008 and 1115 mm yr-1 at the short- and long-hydroperiod
sites, respectively), though there were seasonal differences in the ratio of ET:precipitation.
During the dry season ET was greater than precipitation, and water loss as LE (ET) caused water
levels to decline. Annual fluctuations in the timing of peak water level, H and LE suggest season
onset and length can shift substantially from one year to the next and differ by site. These results
also show that energy balance closure was within the range found at other wetland sites (60 to
80%), and was lower when sites were inundated (60 to 70%). Patterns in energy partitioning covaried with hydrology and climate, suggesting that shifts in any of these components could
disrupt current water and biogeochemical cycles throughout the Everglades region.
47
Introduction
Land use change has led to significant wetland loss (>50%; Mitsch & Gosslink 2007) and
modified terrestrial energy and water budgets enough to alter climate patterns (Pielke et al.,
1999; Chapin et al., 2002). Solar energy drives the hydrologic cycle, which regulates
biogeochemical processes (Chapin et al., 2002), making energy dynamics key to ecosystem
analysis (Odum 1968). In wetland ecosystems where hydrology strongly influences ecosystem
structure and function (Davis & Ogden 1994; Childers et al., 2006; Mitsch & Gosselink 2007),
the energy balance has a direct impact on climate and ecosystem processes. Given the vital role
of energy and hydrologic cycles on wetland ecosystem function, it is critical that we understand
their controls and the extent to which they have been modified by human actions.
Significant land cover change has occurred in the Florida Everglades (Pielke et al., 1999;
Marshall et al., 2004) where 53% of the original extent has been lost to drainage (Reddy &
Delaune 2008), development and agriculture (Davis & Ogden 1994). These ecologically
important systems have distinct wet and dry seasons that produce considerable variation in the
hydrologic cycle, affecting nutrient delivery, ecosystem primary production and ecosystem
structure, which ultimately feeds back to contribute to the water cycle (Davis & Ogden 1994).
While seasonal patterns in water cycling are heavily influenced by anthropogenic pressures (i.e.,
water control structures), implementation of the Comprehensive Everglades Restoration Plan
(CERP) and future climate projections are predicted to have significant impacts on hydrologic
patterns. To date the links between energy partitioning, climate and hydrology in short- and longhydroperiod Everglades ecosystems have not been fully explored, making it important to develop
an understanding of these interactions and what they mean for Everglades hydrology prior to
significant changes in water management and climate.
48
Patterns in energy partitioning differ among ecosystems (Twine et al., 2000; Wilson et
al., 2002; Lafleur et al., 2008) and can result from changes in environmental conditions (Lafleur
et al., 2008; Twine et al., 2000; Wilson et al., 2002; Rocha et al., 2004). In wetland ecosystems
net radiation (Rn) is stored in the soil and/or water column (G), and is also partitioned as sensible
heat flux (H) and latent energy flux (LE) (Lafleur et al., 2008; Piccolo 2009; Schedlbauer et al.,
2011). Seasonality influences H and LE partitioning, which co-vary with fluctuations in
hydroperiods (Schedlbauer et al., 2011). Surface fluxes (H and LE) influence water loss as LE
(evapotranspiration; ET) (Myers & Ewel 1992; Davis & Ogden 1994), and seasonal patterns in
water level through their influence on convective rain, the main source of wet season
precipitation (Myers & Ewel 1992). During the dry season, the Bermuda High-pressure cell
prevents convective clouds from forming thunderstorms, making continental fronts the main
source of precipitation (Chen & Gerber 1992). This switch from wet season tropical climate to
dry season temperate climate causes distinct changes in the amount of precipitation in the region
(Chen & Gerber 1992) and combined with constant water loss as LE, produces seasonal
fluctuations in water levels (Davis & Ogden 1994; Myers & Ewel 1992).
Driving and responding to changes in hydrology, H and LE are important for describing
seasonal changes in water and energy in the Everglades region. Schedlbauer et al. (2011)
examined how the components of the energy balance changed seasonally, showing that during
the dry season more available energy was partitioned as H, and LE was the dominant flux during
the wet season when water levels were typically above the soil surface. In addition to seasonal
changes in Rn, soil volumetric water content (VWC), water depth, air temperature (Tair), and
occasionally vapor pressure deficit (VPD) exhibited strong relationships with H and LE
(Schedlbauer et al., 2011). Although previous research has evaluated energy balance in
49
Everglades freshwater marsh, no study has quantified the links between climate, energy
exchanges and hydrology, or determined how these factors interact to influence seasonal
hydrologic dynamics in marshes with different hydroperiods and over multiple years of study.
At the landscape scale, Everglades hydrology is a complex function of weather patterns,
topography and water management (Davis & Ogden 1994; Richardson 2008). Historical
hydroperiods were determined by the combined effects of precipitation and runoff (inputs), and
evapotranspiration (ET) and drainage (outputs). Small variations in elevation throughout the
landscape regulate the degree of exposure to surface flow that originated in the north from Lake
Okeechobee and moved south as sheet flow. Presently, the system is subject to substantial
anthropogenic control through a complex system of canals, levees, and pumping stations
(Loveless 1959; Davis & Ogden 1994). As a result of water management activities, reduced
water flow from Lake Okeechobee changed the natural characteristics of the Everglades wetland
ecosystems (Davis & Ogden 1994; Richardson 2008). Position within the landscape is important
in understanding the seasonal patterns in hydrology, the effects of anthropogenic water
management on hydrology, and ultimately the system’s energy balance (Davis & Ogden 1994;
Richardson 2008).
Evaluating links between hydrology and Everglades energy balance is especially
important in light of imminent changes in water management and climate change. Current water
management practices will be modified under the CERP, with the goal to re-establish the
hydrology closer to natural seasonal regimes and to ameliorate areas that suffer from chronically
low water levels (Perry 2004). This could increase the amount of available energy partitioned
into LE, and potentially affect the current (albeit altered) ecosystem structure, hydrology and
local climate. What is not known is: how will changes in hydroperiod and water depth affect the
50
controls on H and LE? What are the pace, pattern and potential feedbacks of this action? The
complex interactions between environmental conditions and energy drive changes in ecosystem
function (e.g., higher LE reduces ecosystem water levels and increases the amount of exposed
leaf area), making it important to understand the links between hydrology and surface fluxes in
the Everglades region.
The goal of this research is to develop a quantitative understanding of how energy
portioning interacts with climate and hydrology in short- and long-hydroperiod wetlands of
Everglades National Park and explore if these ecosystems respond similarly to these changes. I
hypothesize that: 1) Variations in the extent and the timings of peak surface fluxes will indicate
the magnitude of the seasonal response in hydroperiods and 2) Tair, VPD and VWC will exhibit
significant positive controls on H and LE that will vary by site as a result of dissimilarities in
hydrology. Although changes in surface flow in response to precipitation patterns produce
seasonal differences in water levels, the magnitude of difference is controlled by interactions
between surface flow (inputs) and the loss of water as LE (ET) to the atmosphere. At the shorthydroperiod site, where surface water inputs are much less than at the long-hydroperiod site, I
expect that patterns (magnitude and timing of the peak) in H and LE will reflect the seasonal
differences in water levels. Additionally, I expect differences in the magnitude of H and LE will
be a result of dissimilarities in environmental controls by site.
Materials and Methods
Study Site
The Everglades are classified as subtropical wetlands with distinct wet and dry seasons
during the summer and winter months respectively. Long-term (1963 to 2012) mean maximum
51
and minimum daily temperatures were 29°C and 18°C, respectively (National Climatic Data
Center, Royal Palm Ranger Station), with the lowest daily temperatures in January and highest
daily temperatures in August (Davis & Ogden 1994). The Everglades receive approximately
1380 mm of precipitation annually (Davis & Ogden 1994). The majority of rainfall (~70%)
occurs during the wet season (May to October) as convective thunderstorms and tropical
depressions, e.g., storms and hurricanes (Davis & Ogden 1994). During the dry season
(November to April), general circulations switch from the summer Bermuda high (Wang et al.,
2010; Li et al., 2011) to a continental-based high, causing cold fronts to contribute towards
seasonal precipitation (Davis & Ogden 1994).
The study sites are two oligotrophic freshwater marsh ecosystems that are part of the
Florida Coastal Everglades (FCE) Long Term Ecological Research (LTER; TS-1 and SRS-2).
Although just 24 km apart, the two sites have contrasting hydroperiods. Taylor Slough (TS;
5° 6’16.5” N, 80°35’40.68” W) is a short-hydroperiod marsh that is flooded for 4 to 6 months
each year (June to November) and is characterized by shallow marl soils (~0.14 m) overlying
limestone bedrock (FCE-LTER, http://fcelter.fiu.edu/research/sites/). Taylor Slough has a
continuous homogeneous canopy (surface roughness, d ≈0.3 m) dominated by short-statured
(height, Z=0.73 m), emergent species, Cladium jamaicense (Crantz) and Muhlenbergia capillaris
(Lam.). Microalgae that live on submerged substrates form periphyton (Davis & Ogden 1994)
and when the site is dry, the periphyton exists as a desiccated mat between individual plants and
covering the soil surface. Shark River Slough (SRS; 25°33'6.72"N, 80°46'57.36"W) is a longhydroperiod marsh that is inundated ~12 months each year and is characterized by peat soils (~1
m thick) overlying limestone bedrock with ridge and slough microtopography (Duever et al.,
1978). For this site, Z and d are 1.02 and ~0.4 m, respectively. In ridge areas, SRS is dominated
52
by tall, dense emergent species, Cladium jamaicense, Eleocharis sp., and Panicum sp., while
short-stature, submerged species (Utricularia sp.) dominate the sloughs. Periphyton also exists
on submerged structures as floating mats at SRS (for a more detailed description of the
vegetation, see Davis & Ogden 1994).
Eddy Covariance and Micrometeorology
At TS and SRS, LE and H were measured from January 1, 2009 to December 31, 2012
using open-path eddy covariance (EC) methods (Moncrieff et al., 1996; Ocheltree & Loescher
2007). Open-path infrared gas analyzers (IRGA; LI-7500, Li-COR Inc., Lincoln, NE) were used
to measure water vapor molar density (v; mg mol-1), and sonic anemometers (CSAT3, Campbell
Scientific Inc, Logan, UT) were employed to measure sonic temperature (Ts; oK) and 3dimensional wind speed (u, v and w, respectively; m s-1). These sensors were 0.09 m apart and
installed 3.30 and 3.24 m above ground level (a.g.l.) at TS and SRS, respectively. Data were
logged at 10 Hz on a datalogger (CR1000, Campbell Scientific Inc) and stored on 2 GB
CompactFlash cards. Both IRGAs were calibrated monthly using dry N2 gas and a traceable
dewpoint generator (LI-610, LI-COR Inc.). Footprint analyses indicated that 80% of measured
fluxes were within 100 m of the tower during convective conditions at both sites (Kljun et al.,
2002, 2004). Other meteorological variables measured at 1-sec and collected as half-hourly
averages and acquired by the same datalogger included: air temperature (Tair; °C) and relative
humidity (Rh; %) (HMP45C, Vaisala, Helsinki, Finland) mounted within an aspirated shield
(43502, R.M. Young Co., Traverse City, MI), and barometric pressure (P; atm) (PTB110,
Vaisala). The Tair/Rh sensors were installed at the same height as the IRGA and sonic
anemometer.
53
At each site, additional meteorological data was measured at 15-sec, and collected as 30min averages through a multiplexer (AM16/32A Campbell Scientific Inc.) with another
datalogger (CR10X Campbell Scientific Inc.). This included photosynthetically active radiation
(PAR; mol m-2 s-1) (PAR Lite, Kipp and Zonen Inc., Delft, Netherlands), incident solar
radiation (Rs; W m-2) (LI-200SZ, LI-COR Inc.), and Rn (W m-2) (CNR2-L, Kipp and Zonen).
Precipitation measurements were made with tipping bucket rain gages (mm) (TE525, Texas
Electronics Inc., Dallas, TX). Soil volumetric water content (VWC; %) was calculated from
equations developed for peat and marl soils using methodology from Veldkamp & O’Brien
(2000) from the dielectric constant using two soil moisture sensors (CS616, Campbell Scientific
Inc.) buried between 0 and 20 cm soil depth at each site. Soil temperature (Tsoil; °C) was
measured at 5 cm, 10 cm, and 20 cm depths at two locations within each site using insulated
thermocouples (Type-T, Omega Engineering Inc., Stamford, CT). When inundated at SRS, water
temperature (Tw; °C) was measured using two pairs of insulated thermocouples, pair-located at a
fixed height 5 cm above the soil surface and another attached to shielded floats that held the
thermocouples 5 cm below the water surface. At TS, Tw was measured using insulated
thermocouples located at a fixed height 2 cm below the water surface. Water level (m) at both
sites was recorded every half-hour with a water level logger (HOBO U20-001-01, Onset,
Bourne, MA).
Governing Equations
The energy budget is illustrated in Figure 5 and defined in Eq. 1. Each term represents an
average energy flux over a half hour period,
Eq. 1
54
where, Rn is the net solar radiation (W m-2), H is the sensible heat flux (W m-2), LE is the
latent heat flux in the change of phase of water, i.e., vaporization or condensation (W m-2), Gw is
the change in energy storage associated with the water above the soil surface (Wm-2), and Gw+s is
the change in energy storage in the matrix of both the water and the soil below the soil surface
(Wm-2).
Vertical windspeed (w) was first estimated mean-to-streamline using a 2-d rotation in a
Cartesian coordinate framework (Loescher et al., 2006). The H was then determined using the
covariance of the turbulent fluctuations (noted as primes) of w and Ts (Loescher et al., 2006) and
the block average, ̅̅̅̅̅̅̅, estimated over a 30-min averaging period (noted as overbar), such that,
̅̅̅̅̅̅̅
(
̅̅̅̅̅̅ )
Eq. 2
where: air is the air density (kg m-3), Cp is the specific heat of air at constant pressure (J kg-1 ˚C1
), ̅̅̅̅̅̅ is the covariance of the turbulent fluctuations in w and the molar fraction of water vapor
calculated by unit conversion of v. Corrections for the effect of water vapor on the speed of
sound were applied (Schotanus et al., 1983). Here, actual air temperature is estimated from the
sonic temperature, Tk (ºK),
Eq. 3
Simialry, LE (W m-2) was calculated from the covariance of the turbulent fluctuations of
w and v (mg mol-1) and averaged over 30-min,
̅̅̅̅̅̅
Eq.4
where: R is the ideal gas constant (0.082 L atm K-1 mol-1),  is the heat of vaporation (J g-1), Mair
and Mw are the molecular weights of air (28.965 g mol-1) and water (18.01 g mol-1), respectively,
55
and 103 is a conversion factor (g to mg). Corrections for thermal and pressure related expansion
and/or contraction, and water dilution were applied (Webb et al., 1980).
Data (H and LE) were processed with EdiRe (v. 1.4.3.1184, Clement 1999) following
standard protocols, including despiking (Aubinet et al., 2000), and both measurements were
filtered when systematic errors in either H or LE were indicated, such as: (1) evidence of rainfall,
condensation, or bird fouling in the sampling path of the IRGA or sonic anemometer, (2)
incomplete half-hour datasets during system calibration or maintenance, (3) poor coupling of the
canopy with the external atmospheric conditions, as defined by the friction velocity, u*, using a
threshold < 0.15 m s-1 (Goulden et al., 1996; Clark et al., 1999), or (4) excessive variation from
the half-hourly mean based on analysis of standard deviations for u, v, and w wind and CO2
statistics. Quality assurance of flux data was also maintained by examining plausibility tests of H
and LE values (<-100 or >800 Wm-2), stationarity criteria, and integral turbulent statistics (Foken
& Wichura 1996). At TS 38% and 77% of the day and nighttime data were filtered, respectively.
At SRS, 34% of daytime data and 70% of nighttime data were filtered. Missing H and LE values
were then gap-filled using the linear relationship between H or LE and Rn on a monthly basis.
When R2 values were less than 70%, annual relationships between Rn and H or LE were used to
gap-fill data in that month. Less than 5% of filtered data were filled with annual equations.
Heat flux (Gw) stored within the water column was determined by measuring temperature
differences between the surface and the bottom of the water column (above the soil surface). As
a result of site hydrological characteristics, thermocouple placement within the water column
differed at SRS and TS. At SRS thermocouples measured water temperature at the surface and at
the bottom of the water column. The linear relationship between surface water temperature and
water temperature at the bottom of the water column was determined via linear regression at SRS
56
(R2>90%). At TS thermocouples were only present at the water surface, and thus a linear
relationship established at SRS was used to estimate the temperature at the bottom of the water
column at TS. The change in vertical temperature profile from the 30-minute averaging period to
the next (∂Tw/∂t) was used to determine the energy flux in the water column (Campbell &
Norman 1998):
∫
Eq.5
where: Cpw is the specific heat of liquid water (J kg-1 °C-1), w is water density (kg m-3), z is the
water depth (m), and z0 is the bottom of the water column.
Using the temperature profile for the soil, and the fraction of mineral, organic matter and
water in the soil, Gs+w was determined from the matrix of water and soil, below the soil surface
using a 2- dimensional approach:
∫
∫
Eq.6
where: a and b are the fraction of water and soil, respectively, below the soil surface, Cps is the
specific heat of the soil (J kg-1 °C-1), s is the soil bulk density (kg m-3), ∂Tsoil/∂t is the change in
vertical temperature profile of the soil and z is the soil depth (0.10 m).
Evapotranspiration (mm) was calculated from LE data with the equation
Eq. 7
where: 1800 is the number of seconds in a half hour, 103 is a conversion factor (m to mm), and 
is in J kg-1. For equations 5, 6 and 7, w was calculated via a non-linear curve (McCutcheon et
al., 1993) estimated to predict water density (kg m-3) as a function of temperature using known
water density values at specified temperatures (Campbell & Norman 1998):
Eq. 8
57
Data Analyses
Models were formulated to explore the complex relationships between Rn and surface
fluxes (H and LE), with environmental parameters (Tair, water depth, VPD and VWC). Models
for the Bowen ratio (β), which is the ratio of H versus LE, were also estimated. The β is
important to describe site hydrologic conditions. For example, when LE dominates surface fluxes
(<1), water moves from the ecosystem to the atmosphere, lowering the amount of available free
water for evaporation (i.e., water levels and soil moisture content).
As a result of autocorrelation, observations recorded at 30-minute time intervals are not
independent, violating the underlying assumptions of general linear modeling approaches
(Brocklebank & Dickey 2003). To address the serial dependence inherent in data collected over
time, a time series approach, utilizing autoregressive integrated moving average (ARIMA)
models, to identify and describe the relationship between environmental variables and energy
fluxes was used. In ARIMA models, autocorrelation in a data stream is explicitly accounted for
by incorporating its past values. These models incorporate three types of components:
autoregressive (AR) of order p, moving average (MA) of order q, and if necessary, differencing
of degree d. ARIMA models fit to time series data use AR and MA terms to describe their serial
dependence, and can also use other time series data as independent variables to explain their
dependence on outside factors.
First, both dependent and independent variables were tested for stationarity via the
augmented Dickey-Fuller test (Dickey & Fuller 1979), and differenced if necessary. Differencing
was required when time series exhibited non-stationarity (Pankratz 1983). Stationary processes
are those where the mean and standard deviation do not change over time. ARIMA models were
then fit to time series for Rn, H, LE, and  using an iterative Box-Jenkins approach: (1)
58
autocorrelation and partial autocorrelation analysis were used to determine if AR and/or MA
terms were necessary for the given time series, (2) model coefficients were calculated using
maximum likelihood techniques, and (3) autocorrelation plots of model residuals were examined
to further determine the structure of the model (Brocklebank & Dickey 2003).
Explanatory variables were selected for analysis based on the literature (Burba et al.,
1999a; Schedlbauer et al., 2011; Jimenez et al., 2012), and included: Tair, VPD, water level, and
soil VWC. A temporary intervention to examine the effect of season was also included by
incorporating an indicator series into the model. Because the presence of autocorrelation in the
explanatory series can lead to misleading conclusions about the cross-correlations between
series, autocorrelation was removed from all input series via a process called pre-whitening
(Brocklebank & Dickey 2003). ARIMA models were then fit to the dependent variables using
the pre-whitened explanatory series as predictor variables. Plots of cross-correlation functions
between each explanatory series and dependent variables were used to identify relationships at
various lags or time shifts, and autocorrelation plots of the residuals verified that the residual
series had characteristics of random error, or white noise (i.e., non-significant correlation
coefficients at non-zero time lags).
Model selection was based on minimum Akaike’s information criterion (AIC), and
models were acceptable when residual white noise was minimized (Hintze 2004). A backwards
selection method was used, removing the least significant parameter one at a time until all
regression terms in the final model were significant at the α= 0.05 level and/or no improvement
was made in the AIC. Models of Rn, H, LE, and  were estimated separately by site, but model
forms were kept the same to aid in site comparisons. Non-significant parameters remained in a
particular site’s model if they were significant in the other site and showed no effect on the final
59
model of the subject site. ARIMA model assumptions of normality and independence of
residuals were evaluated by examining residual plots. Multicollinearity between explanatory
variables was also explored to ensure models did not contain input series that were highly
correlated.
Energy Balance
Energy balance closure was analyzed daily due to lags in storage terms where energy
stored earlier in the day was released in the afternoon (Leuning et al., 2012; Gao et al., 2010).
Half-hourly values for each component of the energy balance, Rn, H, LE, Gw, and Gs were
converted to units of MJ m-2 s-1 and summed over each day. Energy balance was evaluated by
plotting the daily sum of H and LE vs. the difference, Rn - Gs (water level <0) or Rn - Gs - Gw+s
(water level >0). Linear regression was used to assess the percentage of energy balance closure
for each site. Closure was examined separately when water levels were above the soil surface
and when water levels were below the soil surface.
Results
Environmental Conditions
Major differences in hydrology were seen between the short-hydroperiod site, TS, and the
long-hydroperiod site, SRS. On average TS was inundated 211 days a year compared to 335 days
at SRS. At both sites, averages reflect drought conditions in 2009 and 2011 as measured by
Palmer Drought Severity Index (PDSI; Palmer 1965) values of -3 or lower. Drought conditions
resulted in 34 and 81 dry days at SRS and 133 and 207 dry days at TS, in 2009 and 2011,
respectively. In non-drought years, water levels at SRS were continuously above the soil surface
while at TS, days of inundation annually averaged 228 days. The amount of precipitation
60
received in drought years was not substantially less over the entire calendar year; however,
during drought years, there were fewer rainfall events during the 4 months leading up to the wet
season. During normal years (non-drought), total rainfall in the first 4 months of the year
averaged 276 mm at TS and 246 mm at SRS while in drought years, TS received 107 mm and
SRS received 82 mm. Over the study period, the magnitude and seasonal patterns in Rh, Tair, and
VPD were similar for both sites (Figure 6); however, the average annual rainfall was slightly
higher for TS (1304 mm yr-1) than for SRS (1207 mm yr-1).
Evapotranspiration and Rainfall
Annual and seasonal patterns in rainfall and ET were similar at both sites, although ET
rates were higher at SRS than at TS (averages 1115 and 1008 mm yr-1, respectively).
Precipitation peaked in July which was ~2 wks prior to peak LE (ET) except, in 2011 when the
peak in precipitation occurred much later in wet season (late July early August) and resulted in
less wet season LE compared to other years. In all years, annual rainfall was greater than annual
ET, though this pattern was not observed seasonally. Dry season ET rates surpassed rainfall
while wet season precipitation was much greater than ET (Table 5). Over the study period, 80%
of annual precipitation fell during the wet season while just 60% of annual ET occurred during
the same period. The resulting ratio of ET to precipitation was 0.61 and 1.53 during the wet and
dry season, respectively, at TS. At SRS, the ratio of ET to precipitation was 0.69 in the wet
season and 2.33 in the dry season. Although precipitation rates were lower in the dry season
during drought years, there was no change in ET rates at either site. As a result of lower
precipitation, the ratio of ET to precipitation in the dry season was higher (TS: 1.95; SRS: 3.14)
compared to the dry season of non-drought years (TS: 1.12; SRS: 1.52).
61
Seasonality in Everglades Freshwater Marshes
Seasonal oscillations in water availability followed patterns in energy partitioning (Figure
8). At both sites, rates of daily H ranged from -2 to 13 MJ m-2 day-1and increased with increasing
Rn. At both TS and SRS, H peaked at the end of the dry season (~May 4) and was followed by
the peak in Rn ~5 weeks later. Sensible heat flux accounted for just 26% of Rn at TS and 11% at
SRS annually, and the majority of energy was partitioned in the form of LE (TS: 53%; SRS:
56%). At the start of the wet season, rates of LE increased with rising Rn and estimates ranged
from -1 to 15 MJ m-2 day-1 at TS and SRS (Figure 8c, 8d). At TS the peak in LE occurred in early
August although the peak in water level was ~1 month later. At SRS the peak in LE was within 2
weeks of the peak in Rn and occurred 1 month (~Jun 27) prior to the peak in water levels
(~November 11). Although storage fluxes were large on a half-hour time scale, fluxes were small
on a daily (Figure 7b), seasonal, and annual basis at both TS and SRS. On a daily time scale,
fluxes were less than 5% of Rn. Seasonally, energy stored in the water column and soil accounted
for less than 1% of Rn and annually storage fluxes approached 0. At SRS energy stored in the
water column (3.5 MJ m-2 yr-1) was 2 times greater than at TS (1.66 MJ m-2 yr-1) and during the
dry season energy stored in the water column (average 1.2 MJ m-2 dry season-1) was half that of
the wet season (average 2.24 MJ m-2 season-1).
Stronger seasonal patterns in energy partitioning (Figure 8c) were observed at TS, where
LE dominated H fluxes during the wet season and H dominated LE fluxes during the dry season.
H fluxes were higher at TS throughout both the wet and dry season as compared to SRS (Figure
8c, 8d). At TS higher β (Figures 7), larger ranges in surface fluxes (Figure 8c), and greater
seasonal changes in hydrology were observed (Figure 6c); during the dry season the Bowen ratio
was much higher (0.64) than during the wet season (0.40). Although strong patterns were
62
observed in LE at both sites, seasonal changes in the magnitude of H exchange were limited at
SRS (Figure 8d), where energy partitioned as H was just 12% and 10% of Rn in the dry and wet
seasons, respectively. The seasonal patterns in β were also reduced at SRS (Figure 9), and the
difference between H and LE gradually expanded as Rn increased (Figure 8d).
Annual fluctuations in the timing of peak water level, H, and LE suggest there may be
significant changes in the wet and dry season onset at TS and SRS. Shifts in water level peaks
were quite variable at both TS and SRS (sd = 76 and 59 days, respectively), indicating that
seasons can shift substantially from one year to the next and the calendar-delineated season (wet
season from May to October) might not capture seasonal variability (i.e., timing and length).
Additionally, the distance between peaks in H and LE at each site differed substantially (94 and
55 days at TS and SRS, respectively), and reflected differences in hydrology.
Environmental Drivers of Energy Fluxes
Despite parallel patterns and similar magnitudes in environmental variables (Tair, VPD,
soil VWC, and Rh) observed for the two sites (Figure 6), the estimated parameters from Rn, H,
LE, and the β models differed by site. After pre-whitening, some small (<0.05) but statistically
significant autocorrelation remained in pre-whitened series. However, this sensitivity resulted
from the large number of observations available and was judged to be biologically insignificant
(Starr et al., in review). Differencing was required for water level and soil VWC time series due
to the lack of stationarity at both sites. The lack of stationarity indicates a lack of stability in the
mean of these variables over time, further suggesting that there were significant changes in
hydroperiods at both sites. By including differenced variables in models (Δwater level, ΔVWC) I
evaluated how changes in water level and soil VWC were correlated with Rn, H, ET, and β.
63
Models for Rn at both sites included a significant 24-hour lagged MA component
[MA(48)], as well as significant AR components at 0.5, 24, and 24.5 hours, reflecting daily and
half-hourly self-similarity in observations. At both sites, Tair and VPD with its half-hour lag were
significantly and positively related to Rn (p<0.001; Table 6). Net radiation was negatively
correlated with Δwater level (p=0.008) at SRS, while no significant correlation was detected at
TS (Table 6). At both sites, VPD had the strongest correlation with Rn, and the effect of VPD
lagged by one half-hour was significantly greater at TS than at SRS. When water level, VPD and
Tair were accounted for, season was not significantly correlated with Rn at either site.
Models for H at both sites included the same significant daily MA component [MA(48)]
as Rn, and had the same significant AR components at 0.5, 24, and 24.5, as well as an additional
AR component at 48 hours. VPD, Δwater level, and ΔVWC were important indicators of H
exchange (Table 7). At both sites, VPD and its half-hour lag were significant positive predictors
(p<0.001) of H, and its effect was stronger at TS. At SRS, the change in water level was
negatively correlated with H (p<0.001), while this effect was not significant at TS (p=0.4686).
The strongest driver of H was ΔVWC, which was significantly more negative at TS than at SRS
(p<0.001). Like Rn, when Δwater level, VPD and ΔVWC were accounted for, season was not
significantly correlated with H at either site.
Models for LE at both sites included the same significant MA components as those of H
and Rn, but indicated a more complex AR structure with significant components at 0.5, 1, 23.5,
24, and 24.5. Unlike models of Rn and H, the model of LE included a significant seasonal effect,
which significantly differed between sites. LE was higher during the wet season at TS
(p=0.0004), a pattern not observed at SRS where water was readily available year round. At both
sites, LE was positively correlated with Tair and synchronous VPD (p<0.001; Table 8), but these
64
effects were significantly stronger at TS (Table 8). The half-hour lag in VPD was also
significant; at SRS and positively correlated with LE (p<0.001), whereas this effect was negative
at TS (p<0.0001). Like Rn, VPD had the strongest correlation with LE at both sites, and this
effect was significantly greater at TS.
Models for β included a significant MA component at 4 hours [MA(48)] and significant
AR components at 0.5 hours, 1 hour, and 24 hours. Tair (p<0.001) was significantly positively
related to the β at both sites, with significantly larger effects at TS (Table 9). At both sites, β was
significantly lower during the wet season; however, the effect of season was only detectable at
SRS (p<0.001;Table 9).
Energy Balance
Energy balance closure was determined using daily summations of available energy
inputs and losses (Eq. 1). At both sites, the turbulent fluxes of H and LE underestimated total
available energy. Closure decreased when sites where inundated and closure was lower overall at
SRS than at TS. Closure at SRS was 61% (R2=0.77) and 66% (R2=0.87) when inundated and
dry, respectively (Figure 3). At TS, energy closure was 70% (R2=0.85) when water levels were
above the soil surface and 81% (R2=0.91) when dry.
Discussion
The unique and contrasting hydrologic attributes of Everglades freshwater ecosystems
(Davis & Ogden 1994) lead to differences in rates and drivers of energy exchange (Schedlbauer
et al., 2010). Wetland formation is related to the controls on water inputs (precipitation, sheet
flow) and outputs (ET, runoff) (Mitsch & Gosslink 2007), making water levels a key determinant
governing this ecosystem’s functions for the Greater Everglades ecosystem (Rouse 2000; Davis
65
& Ogden 1994). Here, the variance in precipitation, seasonality and controls on LE (ET) and H
in two contrasting wetlands across years that included droughts was examined. Differences in
hydrologic patterns between sites were reflected in the magnitude and timing of the peak in
surface fluxes and the effect of environmental controls. Although season was not significant in
models of LE and H, climatic variables with distinctive seasonal patterns were important and
strong predictors of LE and H. The results presented here add insight into the complicated
relationships and feedbacks between energy dynamics, hydrology, and environment controls.
Seasonality in Everglades Freshwater Marshes
In the Everglades, ET is the main source of water loss (Obeysekera et al., 1999), making
LE a major driver of both water and energy cycles. The timing and extent of water level
oscillations, which are controlled by precipitation and ET, effects the colonization and survival
of marsh vegetation (i.e., submergence; Myers & Ewel 1992). Seasonal patterns in water levels
resulted from the ratio of ET to precipitation and differences in surface flows at the study sites.
Although precipitation surpassed ET annually, dry season ET surpassed rainfall and wet season
precipitation was much greater than ET (Table 5). The positive water balance is important for
providing surface flow to wetlands downstream (Harvey et al., 2000; Sutula et al., 2001), the
quantity of freshwater and organic matter flowing into Florida Bay (Davis & Ogden 1994; Reddy
& DeLaune 2008), and for ground water recharge (Fennema et al., 1994; Harvey et al., 2000;
Sutula et al., 2001).
While precipitation is driven by climate, ET is driven by both climate and wetland
characteristics (Lafleur et al., 2008). The dominant control on ET is available energy (Piccolo
2009; Souch et al., 1998), followed by surface and vegetative resistance (Piccolo 2009). Less
66
seasonality in ET suggests that like other wetlands dominated by vascular plants, the water
availability is not substantially reduced until the water table drops below the rooting zone (Souch
et al., 1998). At SRS, hydrologic implications of ET are similar to those of regional coastal
systems (Harvey & Nuttle, 1995) where water loss through ET is minor because it is replaced by
surface runoff. Here, ET rates were not limited by water availability, but driven largely by
available energy. At TS, steady rates of ET suggest that while it is driven by energy availability,
the role of transpiration increases as water levels fall below the soil surface (Lafleur 1990;
Campbell & Williamson 1997; Souch et al., 1998).
Most wetlands have lower ET rates than would be expected from open water under the
same conditions (Idso & Anderson 1988; Lafleur 1990; Linacre 1976; Lafleur et al., 2008),
suggesting that transpiration is a very important component of ET. Similar to wetland
ecosystems with tall reed vegetation (Goulden et al., 2007; Souch et al., 1998; Lafleur et al.,
2008), ET from Everglades marshes seems to be insensitive to water level. Thus, transpiration is
likely the main source of ET, which is less affected by the presence or absence of standing water
(Lafleur et al., 2008). Wetland plants have evolved in such a way that there is an upper limit to
ET regardless of water supply and atmospheric demand (Lafleur et al., 2008), and the vascular
vegetation of the wetlands limit ET by sheltering the underlying surface from turbulence and
reducing the amount of radiant energy reaching the substrate. The most important influence of
plants on ET is their physiological control over the transpiration process through stomatal
conductance (Lafleur et al., 2008).
It has been suggested that feedbacks between wetland ET and local precipitation
perpetuate the existence of wetland areas (Lafleur et al., 2008). In the Everglades region,
precipitation patterns throughout the Kissimmee-Okeechobee-Everglades region (Davis &
67
Ogden 1994; Redfield 2000) are important for the greater Everglades hydrology. Wet season ET
feedbacks to precipitation (Douglas & Rothchild 1987; Pielke et al., 1999), which accounted for
80% of the annual precipitation during the study period. These patterns suggest that the
Everglades itself is important for perpetuating the hydrologic patterns in the region (Myers &
Ewel 1992).
Energy Balance
In wetland ecosystems, LE (ET) is the most important energy flux (Jacobs et al., 2002;
Piccolo 2009), accounting for up to 100% of precipitation losses (Linacre 1976; Lafleur 1990;
Souch et al., 1998). When standing water is present in wetland ecosystems there is very similar
flux partitioning (Souch et al., 1998). The LE accounts for approximately 50% of the Rn and
20% of the Rn go to H (Souch et al., 1998; Piccolo 2009). At both Everglades sites, energy
partitioned for H and LE were within the ranges found in other wetlands (Teal & Kanwisher
1970; Vugts & Zimmerman 1985; Rouse 2000; Beigt et al., 2008), where annual average β was <
1 (Burba et al., 1999a & b; Heilman et al., 2000; Table 9). Unique to this system was the
seasonal shift in surface fluxes and how they feedback to the regional climate and hydrology.
In the Everglades, seasonal fluctuations were driven by precipitation patterns and surface
fluxes affected the magnitude of the seasonal effect, as hypothesized. Seasonal shifts in
precipitation led to offsets in peak in H and LE. Wet season LE feedbacks to precipitation
through thunderstorm formation and the disruption of this pattern in the fall and winter months
by the Bermuda High-pressure cell (Chen & Gerber 1992) leads to a decline in surface flow and
water levels. At TS where the exposure to surface flows is lower than at SRS (Davis & Ogden
1994), water loss as LE caused a greater decline in water levels and generated larger seasonal
68
differences in surface fluxes. As a result, the distance between peak H and peak LE was ~3.1
months apart. At SRS, where surface flow exposure is greater (Davis & Ogden 1994), water
level declined less and peak H and LE were just ~1.8 months apart. Seasonal changes in H and
LE indicate that the β should reveal changes in wet and dry season strength.
The β is a useful indicator of the ecosystem energy contributions to the regional climate,
where larger H indicates a dryer climate and larger LE rates suggest a more humid climate
(Lafleur et al., 2008). Both freshwater marsh ecosystems fall within the range of wetland β (-0.11
to 1), and fall on either side of the mean growing season β for wetland ecosystems (0.3 ; Souch
et al., 1996; Price 1994; Goulden et al., 2007; Burba et al., 1999a & b; Rijks 1969; Linacre et al.,
1970). At TS, I expected and observed higher β and larger ranges in surface fluxes as a result of
greater seasonal changes in H and LE, and therefore hydrology. As a result of the strength of the
seasonal patterns at TS, the annual average β was higher than other wetland ecosystems (Table
4). Seasonality in the β was reduced at SRS as a result of the year-round dominance of LE. At
TS, β resembled those of grassland (Twine et al., 2000) and Mediterranean (Wilson et al., 2002)
ecosystems, where large seasonal differences in water availability resulted in a larger disparity
between wet and dry season β. Because SRS remained inundated throughout most normal years,
β resembled those of tropical ecosystems and energy partitioned as LE dominated H year round
(Rocha et al., 2004).
Shifts in peak precipitation, H and LE suggest that the timing and length of seasons may
fluctuate in the Everglades region (Chen & Gerber 1992) and the use of the calendar-based
seasonality does not capture changes in the wet and dry season. Future studies should take this
into consideration when determining seasonal patterns in ecosystem function in Everglades
freshwater ecosystems. The β captured the unique features of the subtropical Everglades,
69
displaying both seasonal and annual patterns in energy and water fluxes. In the future, the β
should be explored as an indicator of season onset and length.
Environmental Drivers of Energy Fluxes
Wetland functioning is intimately tied to the atmosphere by energy and mass exchanges,
which are controlled by many factors whose interactions are unique to every wetland ecosystem
(Lafleur et al., 2008). Like Rn, H and LE are controlled by surface and atmospheric properties
(Lafleur et al., 2008). Seasonal sinusoidal patterns of H and LE in subtropical wetland
ecosystems followed those in Rn although their peaks where offset. At the end of the dry season
H peaked and LE peaked soon after the peak in precipitation. In the subtropical Everglades
region, Tair and precipitation also tracked patterns in Rn, which increased in the wet summer
months and declined in the dry winter months. As found by other studies (Rouse 2000), results
show that environmental controls exerted by the atmosphere, water levels and wetland plants
interact with available energy, influencing annual and seasonal patterns in H and LE. As
hypothesized, changes in water level, Tair, VPD and ΔVWC had strong correlations with
available energy and energy partitioning, which differed by site.
The sensible heat flux is dependent on Rn and the temperature difference between the
wetland and air stream overhead (Lafleur et al., 2008). These results and other studies have
found that rising H increases Tair and VPD (Halliwell & Rouse 1987; Souch et al., 1998), which
provides a positive feedback to LE. In the Everglades, H was also negatively correlated with an
increase in water levels and soil volumetric water content. Soil properties, which play an
important role in plant composition and productivity (Adams 1963; Pennings et al., 2005; Wang
et al., 2007; Piccolo 2009) are significantly affected by H and this effect was higher at TS than at
70
SRS. Indicating that higher H is correlated with decreases in soil water content, an artifact of the
relationship between H, Tair, and VPD.
Like other wetland ecosystems, the dominant outgoing flux component, LE, is especially
important because it represents the loss of water from the ecosystem and contributes to
atmospheric moisture. Similar to the results presented here, previous research has identified two
dominating meteorological influences on LE (ET), Rn and VPD (Lafleur et al., 2008; Kellner
2001; Kim & Verma 1996; Souch et al., 1996). Despite physiological differences among wetland
species, the response in stomatal conductance to VPD is nearly universal (Lafleur et al., 2008).
The VPD is known to have a positive relationship with LE (Lafleur et al., 2008) until a threshold
is reached (Admiral & Lafleur 2007). Such behavior is an important negative feedback that
limits the upper rates of ET from wetlands (Admiral & Lafleur 2007). Time series analysis
identified VPD as a very important indicator of water and energy exchange with the atmosphere.
Extremely important for ecosystem productivity through its effect on gas exchange and on
hydrology (Burba et al., 1999a; Schedlbauer et al., 2011), VPD links energy partitioning to
ecosystem productivity (Lafleur et al., 2008), showing that water levels are responding to
fluctuations in both.
Energy Budget Closure
Energy budget closure provides a measure of how well we understand and can measure
component fluxes (Rn, H, LE), and storage fluxes (Gw+s and Gw). Using daily summations of
available energy (Rn- Gw+s - Gw) and surface fluxes (H and LE) from 2009 to 2012, surface
fluxes underestimated available energy at both sites and energy budget closure was lower at SRS
than at TS (Figure 7). Independent measurements of the major energy balance flux components
71
are not often consistent with the principle of conservation of energy (Twine et al., 2000),
resulting in the failure to close the energy budget as seen across many different FLUXNET sites
(Wilson et al., 2002). Closure in this study was within the range of values reported across
FLUXNET sites, and provided a qualitative understanding, such that rates for SRS and when
inundated at TS were lower than the FLUXNET mean of 80% closure (Wilson et al., 2002).
However, closure rates were consistent with other wetland studies that report energy balance
closure at ~70% (Mackay et al., 2007; Li et al., 2009; Schedlbauer et al., 2010). Possible
explanations for the lack of closure include (1) sampling errors associated with differences in
measurement source areas, (2) a systematic bias in instrumentation (e.g., sonic anemometry:
Frank et al., 2013; Kochendorfer et al., 2012, and condensation on net radiometers), (3)
neglected energy sinks, e.g., heat transported horizontally in the water column, (4) neglected
advection of scalars (Wilson et al., 2002; Loescher et al., 2006), and (5) the loss of low and/or
high frequency contributions to the turbulent flux. This latter explanation applies best in
heterogeneous landscapes (Foken et al., 2010), and while TS and SRS have relatively
homogenous source areas, the larger landscape at TS includes levees, canals, and agricultural
lands that are outside the flux source area.
The underestimation of total available energy was higher when water levels were above
the soil surface (Figure 6). The greater lack of closure when sites are inundated suggests storage
fluxes are being under-estimated (Lafleur et al., 1997). However, on annual time scales, energy
storage cancels out and errors in the small contributions made by storage terms (Gw and Gw+s) to
the energy budget on the daily basis would not account for a 20 to 30% lack of closure. The
previous suggested instrumentation biases may be systematically manifest in the accumulation of
condensation on net radiometers. Condensation on net radiometers causes an underestimation of
72
outgoing long-wave radiation leading to an overestimation of net radiation which may explain
the lower closure rates when sites were inundated. No correlations between the lack of closure
and the β were found, negating the notion of under-estimation of LE independent of H. The lack
of closure at both sites may be an indication that both H and LE are being underestimated, which
is important for the water balance of Everglades. Additional research on the source of error,
whether it is underestimation of the storage components or surface fluxes, is required prior to
making any modifications to the Everglade energy balance to account for the lack of energy
balance closure.
Patterns in Rn, H and LE fluctuate with climate and water levels, indicating that
hydrology drives ecosystem function and energy partitioning in the Everglades. Considering the
effect of hydrology on vegetation function and composition, changes in vegetation associated
with altered hydrology can also stimulate changes in the climate system through fluctuations in
albedo, surface roughness, soil moisture, and plant resistance to evaporation (Dickinson 1992;
Thomas & Roundtree 1992; Betts et al., 1996; Baldocchi et al., 2000).
Climate change is emerging as an important challenge for natural resource managers and
decision makers. In the southeastern United States, shifts in precipitation patterns and higher
temperatures are projected by climate models (Christensen et al., 2007; Allan & Soden 2008; Li
et al., 2011; IPCC 2013), and could have a significant effect on the timing and length of wet and
dry seasons. With the implementation of the CERP, shifts in climate that result in increases in
drought frequency and intensity (Stanton & Ackerman 2007) may be moderated by restored
hydrologic conditions. The CERP could re-establish the seasonal patterns of water depth closer
to natural levels, thereby increasing the amount of available energy partitioned into LE and
potentially affect current ecosystem structure, hydrology and climate. Feedbacks to other
73
ecological processes are also likely given this scenario, e.g., changes in species composition,
primary productivity, ratio of anaerobic: aerobic metabolism, and organic matter accumulation.
Patterns in energy partitioning co-varied with hydrology and climate, suggesting that shifts in
any of these components could disrupt current water and biogeochemical cycles throughout the
Everglades region.
74
Table 4. Energy flux partitioning in wetland ecosystems. Within wetland ecosystems energy partitioned as LE dominates the H flux.
Site
LE/ Rn
Everglades shorthydroperiod marsh,
Florida
0.53
Everglades longhydroperiod marsh,
Florida
0.56
0.67 (wet), 0.27
Nueces River Delta, Texas
(dry)
Prairie Wetland, Nebraska 0.8 - 0.9
Paynes Prarie Preserve,
Florida
0.7
H/ Rn
Storage/Rn
Β
Source
0.26
>0.05
0.49
*This study
0.11
0.30 (wet), 0.65
(dry)
-
>0.05
0.19
*This study
-
-
Heilman et al., (2000)
Burba et al., (1999a)
0.26
0.04
0.42
Boreal fen, Manitoba
0.53 - 0.76
Churchill sedge fen
0.64
Schefferville, Quebec
0.63
Hudson Bay Coast, Ontario 0.68
0.24 - 0.47
0.25 - 0.30
0.25
0.3
0.5- 0.11
0.1
0.06
0.39
Jacobs et al., (2002)
afleur et al., (1 7), Baldocchi et al.,
(2000)
Eaton et al., (2001)
Moore et al., (1994), Eaton et al., (2001)
Rouse et al., (1977), Eaton et al., (2001)
75
Table 5. Annual and seasonal ET and precipitation at Taylor Slough and Shark River Slough (2009 to 2012). Although precipitation is
greater than ET annually, ET>precipitation during the dry season and ET< precipitation during the wet season.
Taylor Slough
ET
Season/Year (mm m-2 yr-1)
wet
622.3
dry
396.3
2009
1018.6
wet
655.3
dry
347.9
2010
1003.2
wet
593.6
dry
374.7
2011
968.3
wet
660.8
dry
383.1
2012
1043.9
Average
1008.5
Precipitation
(mm m-2 yr-1)
1042.2
242.5
1284.5
1095.8
294.6
1390.4
929.6
164.8
1094.5
1089.0
361.0
1450.0
1304.8
Shark River Slough
ET
(mm m-2 yr-1)
612.3
449.5
1061.8
766.5
470.2
1236.6
587.5
515.1
1102.7
636.5
424.1
1060.6
1115.4
76
Precipitation
(mm m-2 yr-1)
934.7
155.5
1090.2
760.2
327.9
1088.1
1140.2
151.9
1292.1
1098.0
262.0
1360.0
1207.6
Table 6. Parameter estimates from ARIMA models of Rn by site. MA(48) is the estimated moving average term at a lag of 48 time
periods (24 hours), and AR(1), AR(48) and AR(49) are estimated autoregressive terms at 1-, 48-, and 49- period lags (0.5 hours, 24
hours, and 24.5 hours, respectively). Lagged values of independent variables are denoted similarly. Asterisk denotes significant
difference between sites. The averaging period for Rn and all explanatory series is 0.5 hours. Δ Water Level is the change in water
level (mm) between half-hourly measurements, Tair is air temperature (oC), and VPD is vapor pressure deficit.
Parameter
MA(48)
AR(1)
AR(48)
AR(49)
Δ Water evel
Tair
VPD
VPD(1)
Taylor Slough
Standard
Estimate
Error
0.9656
0.0011
0.7268
0.0026
0.9973
0.0003
-0.7241
0.0027
0.0005
0.0050
0.0108
0.0004
0.1626
0.0026
0.0532
0.0025
t Value
872.81
276.71
3418.72
-272.86
0.10
30.62
63.46
20.93
Approx
Pr > |t|
<.0001
<.0001
<.0001
<.0001
0.9184
<.0001
<.0001
<.0001
77
*
*
Shark River Slough
Standard
Estimate
Error
0.9615
0.0011
0.7163
0.0027
0.9975
0.0003
-0.7138
0.0027
-0.2130
0.0791
0.0092
0.0004
0.1521
0.0027
0.0260
0.0027
t Value
837.80
265.81
3582.36
-262.44
-2.69
25.70
56.26
9.78
Approx
Pr > |t|
<.0001
<.0001
<.0001
<.0001
0.0071
<.0001
<.0001
<.0001
Table 7. Parameter estimates from ARIMA models of H by site. MA(48) is the estimated moving average term at a lag of 48 time
periods (24 hours), and AR(1), AR(48), AR(49) and AR(96) are estimated autoregressive terms at 1-, 48-, 49- and 96-period lags (0.5
hours, 24 hours, 24.5 hours, and 48 hours respectively). Lagged values of independent variables are denoted similarly. Asterisk
denotes significant difference between sites. The averaging period for H and all explanatory series is 0.5 hours. Δ Water Level and Δ
VWC are the change in water level (mm) and change in volumetric water content (%), respectively, between half-hourly
measurements, and VPD is vapor pressure deficit.
Taylor Slough
Standard
Parameter
Estimate
Error
MA(48)
0.9527
0.0020
AR(1)
0.1168
0.0037
AR(48)
1.0074
0.0043
AR(49)
-0.0950
0.0039
AR(96)
-0.0315
0.0039
Δ Water evel 0.0131
0.0180
VPD
0.1039
0.0075
VPD(1)
0.0376
0.0075
Δ VWC
-3.3745
0.5133
t Value
481.67
31.13
234.65
-24.53
-8.11
0.72
13.77
4.98
-6.57
Approx
Pr > |t|
<.0001
<.0001
<.0001
<.0001
<.0001
0.4686
<.0001
<.0001
<.0001
*
*
*
*
*
*
*
78
Shark River Slough
Standard
Estimate
Error
0.9458
0.0017
0.7420
0.0026
1.0566
0.0029
-0.7337
0.0027
-0.0650
0.0027
-0.0598
0.0150
0.0107
0.0005
0.0014
0.0005
-0.75669
0.09981
t Value
559.99
289.31
364.51
-275.78
-24.04
-3.99
20.92
2.83
-7.58
Approx
Pr > |t|
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
0.0047
<.0001
Table 8. Parameter estimates from ARIMA models of LE by site. MA(48) is the estimated moving average term at a lag of 48 time
periods (24 hours), and AR(1), AR(2), AR(47), AR(48) and AR(49) are estimated autoregressive terms at 1-, 2-, 47-, 48-,49-, and 50period lags (0.5 hours, 1 hour, 23.5 hours, 24 hours, 24.5, and 25 hours respectively). Lagged values of independent variables are
denoted similarly. Asterisk denotes significant difference between sites. The averaging period for LE and all explanatory series is 0.5
hours. Tair is air temperature (oC), season is an indicator for the dry (0) and wet (1) season, and VPD is vapor pressure deficit.
Parameter
MA(48)
AR(1)
AR(2)
AR(47)
AR(48)
AR(49)
Tair
Season
VPD
VPD(1)
Taylor Slough
Standard
Estimate
Error
0.9355
0.0029
0.1121
0.0038
0.0128
0.0011
0.0185
0.0013
0.9520
0.0025
-0.0964
0.0039
0.0025
0.0005
0.0269
0.0075
0.1116
0.0076
0.0374
0.0076
t Value
327.19
29.8
11.16
14.69
378.28
-25.02
5.46
3.57
14.64
4.91
Approx
Pr > |t|
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
0.0004
<.0001
<.0001
*
*
*
*
*
*
*
*
79
Shark River Slough
Standard
Estimate
Error
0.8566
0.0040
0.6351
0.0031
0.0072
0.0014
0.0471
0.0018
0.8886
0.0039
-0.5781
0.0039
0.0028
0.0001
-0.0004
0.0034
0.0341
0.0010
-0.0083
0.0010
t Value
216.37
202.16
5.28
26.1
227.78
-149.54
21.01
-0.13
34.29
-8.44
Approx
Pr > |t|
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
0.8947
<.0001
<.0001
Table 9. Parameter estimates from ARIMA models of the Bowen ratio, β, by site. MA(48) is the estimated moving average term at a
lag of 48 time periods (24 hours), and AR(1), AR(2) and AR(48) are estimated autoregressive terms at 1-, 2-, and 48- period lags (0.5
hours, 1 hour, and 24 hours respectively). Asterisk denotes significant difference between sites. The averaging period for β and all
explanatory series is 0.5 hours. Tair is air temperature (oC), and season is an indicator for the dry (0) and wet (1) season.
Parameter
Intercept
MA(48)
AR(1)
AR(2)
AR(48)
Tair
Season
Taylor Slough
Standard
Estimate
Error
-2.1272
0.1210
0.8802
0.0045
0.0488
0.0021
0.0212
0.0017
0.9072
0.0039
0.0829
0.0042
-0.0618
0.0777
t Value
-17.58
195.25
23.59
12.14
235.32
19.55
-0.79
Approx
Pr > |t|
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
0.4268
*
*
*
*
*
80
Shark River Slough
Standard
Estimate
Error
-0.4571
0.0620
0.9101
0.0039
0.0367
0.0018
0.0221
0.0016
0.9270
0.0034
0.0149
0.0020
0.0699
0.0346
t Value
-7.37
233.61
20.02
13.84
275.13
7.41
2.02
Approx
Pr > |t|
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
0.0435
Figure 5. Energy budget for wetland ecosystems with distinct (a) wet and (b) dry seasons. LE is the dominant flux component in
wetland ecosystems. Energy partitioned to H increases during the dry season (b) when water levels are below the soil surface. Energy
fluxes in the water column (GW) and in the soil (GS) are important in these systems, as the storage potential is great and energy stored
in the standing water and soil is partitioned as H and LE flux. Energy in the water column also flows horizontally (grey dotted line);
homogeneity in the landscape makes the horizontal flux negligible.
81
Figure 6. (a) Tair, (b) ET, (c) water level, (d) Rh and (e) VPD at TS and SRS from 2009 to 2012. Patterns in Tair, ET, Rh and VPD
were similar by site.
82
Figure 7. Energy balance closure at (a) TS and (b) SRS, under inundated conditions (black circles) and dry conditions (hollow circles).
Wet conditions are defined by water levels above the soil surface while dry conditions are defined by water levels below the soil
surface. The dotted line is a one-to-one line showing that both TS and SRS underestimate available energy, and this lack of closure
increased when sites were inundated.
83
Figure 8. Annual patterns in (a) Rn and (b) storage fluxes at TS (TS) and SRS (SRS), and H and LE exchanges at (c) TS and (d) SRS,
from January 1, 2009 to December 31, 2012 (5 day moving average). Changes in H and LE exchanges followed patterns in Rn. The
shaded region highlights the wet season that extends from May to October.
84
Figure 9. Weekly moving average Bowen Ratios (β) at TS and SRS from 2009 to 2012. The β was higher during the dry season when
the amount of energy partitioned to the H flux increased. Seasonal fluctuations in the β were greater at TS, where a greater range of
water level was observed.
85
References
Adams, D. A. (1963). Factors Influencing Vascular Plant Zonation in North Carolina Salt
Marshes. Ecology, 44(3), 445. doi:10.2307/1932523.
Admiral, S. W., & Lafleur, P. M. (2007). Partitioning of latent heat flux at a northern peatland.
Aquatic Botany, 86(2), 107–116. doi:10.1016/j.aquabot.2006.09.006.
Allan, R. P. R., & Soden, B. J. B. (2008). Atmospheric warming and the amplification of
precipitation extremes. Science, 321(5895), 1481–1484. doi:10.1126/science.1160787.
Aubinet, M., Grelle, A., Ibrom, A., Rannik, J., Moncrieff, J., Foken, T., et al. (2000). Estimates
of the annual net carbon and water exchange of forests: the EUROFLUX methodology.
In Advances in Ecological Research (pp. 113–175). doi:10.1016/s0065-2504(08)60018-5.
Baldocchi, D., Kelliher, F. M., Black, T. A., & Jarvis, P. (2000). Climate and vegetation controls
on boreal zone energy exchange. Global Change Biology, 6(S1), 69–83.
doi:10.1046/j.1365-2486.2000.06014.x.
Beigt, D., Piccolo, M. C., & Perillo, G. M. E. ( 008). Surface heat budget of an estuarine tidal
flat (Bah a Blanca Estuary, Argentina) (Vol. 34).
Betts, A. K., Ball, J. H., Beljaars, A. C. M., Miller, M. J., & Viterbo, P. A. (1996). The land
surface‐ atmosphere interaction: A review based on observational and global modeling
perspectives. Journal of Geophysical Research: Atmospheres (1984–2012), 101(D3),
7209–7225. doi:10.1029/95JD02135.
Brocklebank, J. C., & Dickey, D. A. (2003). SAS System for Forecasting Time Series (1986 ed.).
SAS Institute Inc. Cary, North Carolina.
Burba, G. G., Verma, S. B., & Kim, J. (1999a). Surface energy fluxes of Phragmites australis in a
prairie wetland. Agricultural and Forest Meteorology, 94(1), 31–51. doi:10.1016/S01681923(99)00007-6.
Burba, G. G., Verma, S. B., & Kim, J. (1999b). Energy Fluxes of an Open Water Area in a midLatitude Prairie Wetland - Springer. Boundary-Layer Meteorology, 91(3), 495–504.
doi:10.1023/A:1001802821846.
Campbell, G. S., & Norman, J. M. (n.d.). An Introduction to Environmental Biophysics (2nd
ed.). Springer, NY, NY, ISBN: 978-0-387-94937-6.
Campbell, D. I., & Williamson, J. L. (1997). Evaporation from a raised peat bog. Journal of
Hydrology, 193(1-4), 142–160. doi:10.1016/S0022-1694(96)03149-6.
86
Chapin, F. S., III, Chapin, M. C., Matson, P. A., & Vitousek, P. (2002). Principles of Terrestrial
Ecosystem Ecology. Springer, NY, NY.
Chen, E., & Gerber, J. F. (1992). Climate. In R. L. Myers & J. J. Ewel (Eds.), Ecosystems of
Florida.
Childers, D. L., Iwaniec, D., Rondeau, D., Rubio, G., Verdon, E., & Madden, C. J. (2006).
Responses of sawgrass and spikerush to variation in hydrologic drivers and salinity in
Southern Everglades marshes, 569(1), 273–292. doi:10.1007/s10750-006-0137-9.
Christensen, J. H., Hewitson, B., Busuioc, A., Chen, A., Gao, X., Held, R., et al. (2007).
Regional climate projections. Climate Change, 2007: The Physical Science Basis.
Contribution of Working group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change, University Press, Cambridge, Chapter 11
(pp. 847–940).
Clark, K. L., Gholz, H., Moncrieff, J. B., Cropley, F., & Loescher, H. W. (1999). Environmental
Controls over Net Exchanges of Carbon Dioxide from Contrasting Florida Ecosystems.
Ecological Applications, 9(3), 936–948. doi:10.1890/10510761(1999)009[0936:ECONEO]2.0.CO;2Clement, R. (1999) .
Davis, S. M., & Ogden, J. C. (Eds.). (1994). Everglades: The Ecosystem and Its Restoration. St.
Lucie Press, Boca Raton , FL.
Dickey, D. A., & Fuller, W. A. (1979). Distribution of the Estimators for Autoregressive Time
Series with a Unit Root. Journal of the American Statistical Association, 74(366), 427–
431.Dickinson, R.E. (1992) Land Surface. In: Climate System Modeling (ed. Trenberth
KE), pp. 149-172. Cambridge University Press, Cambridge.
Douglas, M. S., & Rothchild, J. (1987). Marjory Stoneman Douglas: Voice of the River.
Pineapple Press.
Duever, M. J., Carlson, J. E., Riopelle, L. A., & Duever, L. C. (1978). Ecosystem analysis at
Corkscrew Swamp, in Cypress Wetlands for Water Management, Recycling, and
Conservation. The annual Report to the National Science Foundation and to the
Rockefeller Foundation, Central Wetlands (pp. 534–570). University of Florida.,
Gainesville.
Eaton, A. K., Rouse, W. R., Lafleur, P. M., Marsh, P., & Blanken, P. D. (2001). Surface energy
balance of the western and central Canadian subarctic: Variations in the energy balance
among five major terrain types. Journal of Climate, 14(17), 3692–3703.
doi:10.1175/15200442(2001)014<3692:SEBOTW>2.0.CO;2.
Eugster, W., Rouse, W. R., Pielke, R. A., Sr, Mcfadden, J. P., Baldocchi, D. D., Kittel, T. G., et
al. (2000). Land–atmosphere energy exchange in Arctic tundra and boreal forest:
87
available data and feedbacks to climate. Global Change Biology, 6(S1), 84–115.
doi:10.1046/j.1365-2486.2000.06015.x .
Fennema, R. J., Neidrauer, C. J., Johnson, R. A., MacVicar, T. K., & Persins, W. A. (1994). A
computer model to simulate natural Everglades hydrology. In S. M. Davis & J. C. Ogden
(Eds.), Everglades, the Ecosystem and its Restoration. St. Lucie Press, Delray Beach, FL,
USA.
Foken, T., & Wichura, B. (1996). Tools for quality assessment of surface-based flux
measurements. Agricultural and Forest Meteorology, 78(1-2), 83–105.
doi:10.1016/0168-1923(95)02248-1.
Foken, T., Mauder, M., Liebethal, C., Wimmer, F., Beyrich, F., Leps, J.-P., et al. (2010). Energy
balance closure for the LITFASS-2003 experiment. Theoretical and Applied
Climatology, 101(1-2), 149–160. doi:10.1007/s00704-009-0216-8.
Frank, J. M., Massman, W. J., & Ewers, B. E. (2013). Underestimates of sensible heat flux due
to vertical velocity measurement errors in non-orthogonal sonic anemometers.
Agricultural and Forest Meteorology, 171-172, 72–81.
doi:10.1016/j.agrformet.2012.11.005.
Gao, Z., Horton, R., & Liu, H. P. (2010). Impact of wave phase difference between soil surface
heat flux and soil surface temperature on soil surface energy balance closure. Journal of
Geophysical Research, 115(D16), D16112. doi:10.1029/2009JD013278 .
Goulden, M. L., Litvak, M., & Miller, S. D. (2007). Factors that control Typha marsh
evapotranspiration. Aquatic Botany, 86(2), 97–106. doi:10.1016/j.aquabot.2006.09.005.
Goulden, M. L., Munger, J. W., & FAN, S. M. (1996). Measurements of carbon sequestration by
long‐ term eddy covariance: Methods and a critical evaluation of accuracy. Global
Climate Change Biology, 2, 169–182. doi:10.1111/j.1365-2486.1996.tb00070.x .
Halliwell, D. H., & Rouse, W. R. (1987). Soil heat flux in permafrost: Characteristics and
accuracy of measurement. Journal of Climatology, 7(6), 571–584.
doi:10.1002/joc.3370070605.
Harvey, J. W., & Nuttle, W. K. (1995). Fluxes of water and solute in a coastal wetland sediment.
2. Effect of macropores on solute exchange with surface water. Journal of Hydrology,
164, 109–125–. doi:10.1016/0022-1694(94)02562-P.
Harvey, J. W., Choi, J., & Mooney, R. H. (2013). Hydrologic interactions between surface water
and ground water in Taylor Slough, Everglades National Park (pp. 24–26). U.S.
Geological Survey Open-File Report 00-449.
88
Heilman, J. L., Heinsch, F. A., Cobos, D. R., & McInnes, K. J. (2000). Energy balance of a high
marsh on the Texas Gulf Coast: Effect of water availability. Journal of Geophysical
Research: Atmospheres (1984–2012), 105(D17), 22371–22377.
doi:10.1029/2000JD900334.
Hintze, J. (2004). Number Crunching Statistical Systems. Kaysville, Utah, www.NCSS.com.
Idso, S. B., & Anderson, M. G. (1988). A comparison of two recent studies of transpirational
water loss from emergent aquatic macrophytes. Aquatic Botany, 31(1-2), 191–195.
doi:10.1016/0304-3770(88)90049-6.
IPCC (2013). Climate Change 2013: The Physical Science Basis. Contribution of Working
Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate
Change. (S. Joussaume & J. Penner, Eds.) (pp. 1–2216). Cambridge University Press,
Cambridge, UK and New York, NY, USA.
Jacobs, J. M., Mergelsberg, S. L., Lopera, A. F., & Myers, D. A. (2002). Evapotranspiration
from a wet prairie wetland under drought conditions: Paynes Prairie Preserve, Florida,
USA. Wetlands, 22(2), 374–385. doi:10.1672/02775212(2002)022[0374:EFAWPW]2.0.CO;2.
Jimenez, K. L., Starr, G., Staudhammer, C. L., Schedlbauer, J. L., Loescher, H. W., Malone, S.
L., & Oberbauer, S. F. (2012). Carbon dioxide exchange rates from short- and longhydroperiod Everglades freshwater marsh. Journal of Geophysical Research, 117(G4),
G04009. doi:10.1029/2012JG002117.
Kellner, E. (2001). Surface energy fluxes and control of evapotranspiration from a Swedish
Sphagnum mire. Agricultural and Forest Meteorology, 110(2), 101–123.
doi:10.1016/S0168-1923(01)00283-0.
Kim, J., & Verma, S. B. (1996). Surface exchange of water vapour between an open sphagnum
fen and the atmosphere. Boundary-Layer Meteorology, 79(3), 243–264.
doi:10.1007/BF00119440.
Kljun, N., Rotach, M. W., & Schmid, H. P. (2002). A Three-Dimensional Backward Lagrangian
Footprint Model For A Wide Range Of Boundary-Layer Stratifications. Boundary-Layer
Meteorology, 103(2), 205–226. doi:10.1023/A:1014556300021.
Kljun, N., Calanca, P., Rotach, M. W., & Schmid, H. P. (2004). A Simple Parameterisation for
Flux Footprint Predictions. Boundary-Layer Meteorology, 112(3), 503–523.
doi:10.1023/B:BOUN.0000030653.71031.96.
Kochendorfer, J., Meyers, T. P., Frank, J., Massman, W. J., & Heuer, M. W. (2012). How Well
Can We Measure the Vertical Wind Speed? Implications for Fluxes of Energy and Mass.
Boundary-Layer Meteorology, 145(2), 383–398. doi:10.1007/s10546-012-9738-1.
89
Lafleur, P. M. (1990). Evapotranspiration from sedge-dominated wetland surfaces. Aquatic
Botany, 37(4), 341–353. doi:10.1016/0304-3770(90)90020-L.
Lafleur, P. M., McCaughey, J. H., Joiner, D. W., Bartlett, P. A., & Jelinski, D. E. (1997).
Seasonal trends in energy, water, and carbon dioxide fluxes at a northern boreal wetland.
Journal of Geophysical Research: Atmospheres (1984–2012), 102(D24), 29009–29020.
doi:10.1029/96JD03326.
Lafleur, P. M. (2008). Connecting Atmosphere and Wetland: Energy and Water Vapour
Exchange. Geography Compass, 2(4), 1027–1057. doi:10.1111/j.17498198.2007.00132.x.
Leuning, R., van Gorsel, E., Massman, W. J., & Isaac, P. R. (2012). Reflections on the surface
energy imbalance problem. Agricultural and Forest Meteorology, 156, 65–74.
doi:10.1016/j.agrformet.2011.12.002.
Li, Y., Zhou, L., Xu, Z., & Zhou, G. (2009). Comparison of water vapour, heat and energy
exchanges over agricultural and wetland ecosystems. Hydrological Processes, 23(14),
2069–2080. doi:10.1002/hyp.7339.
Li, L., Li, W., & Kushnir, Y. (2011). Variation of the North Atlantic subtropical high western
ridge and its implication to Southeastern US summer precipitation. Climate Dynamics,
39(6), 1401–1412. doi:10.1007/s00382-011-1214-y.
Linacre, E. T. (1976). Swamps. In J. L. Monteith (Ed.), Vegetation and the Atmosphere (pp.
329–350).
Linacre, E. T., Hicks, B. B., Sainty, G. R., & Grauze, G. (1970). The evaporation from a swamp.
Agricultural Meteorology, 7, 375–386. doi:10.1016/0002-1571(70)90033-6.
Loescher, H. W., Law, B. E., Mahrt, L., & Hollinger, D. Y. (2006). Uncertainties in, and
interpretation of, carbon flux estimates using the eddy covariance technique. Journal of
Geophysical Research, 111, D21S90. doi:10.1029/2005JD006932.
Loveless, C. M. (1959). A Study of the Vegetation in the Florida Everglades. Ecology, 40(1), 1–
9. doi:10.2307/1929916.
Mackay, D. S., Ewers, B. E., Cook, B. D., & Davis, K. J. (2007). Environmental drivers of
evapotranspiration in a shrub wetland and an upland forest in northern Wisconsin. Water
Resources Research, 43. doi:10.1002/(ISSN)1525-2027.
Malone, S. L., Starr, G., & Staudhammer, C. L. (2013). Effects of simulated drought on the
carbon balance of Everglades short‐ hydroperiod marsh. Global Change ….
doi:10.1111/gcb.12211.
90
Marshall, C. H., Pielke, R. A., Sr, Steyaert, L. T., & Willard, D. A. (2004). The impact of
anthropogenic land-cover change on the Florida peninsula sea breezes and warm season
sensible weather. Monthly Weather Review, 132(1).
McCutcheon, S. C., Martin, J. L., & Barnwell, T. O. Jr. (1993). Water Quality in Maidment, D.R.
McGraw-Hill, New York, NY.
Mitsch, Gosslink. (2007). Wetlands (4 ed.). John Wiley & Sons, Inc., Hoboken, New Jersey.
Moncrieff, J. B., Malhi, Y., & Leuning, R. (1996). The propagation of errors in long-term
measurements of land-atmosphere fluxes of carbon and water. Global Change Biology,
2(3), 231–240. doi:10.1111/j.1365-2486.1996.tb00075.x.
Moore, K. E., Fitzjarrald, D. R., Wofsy, S. C., Daube, B. C., Munger, J. W., Bakwin, P. S., &
Crill, P. (1994). A season of heat, water vapor, total hydrocarbon, and ozone fluxes at a
subarctic fen. Journal of Geophysical Research: Atmospheres (1984–2012), 99(D1),
1937–1952. doi:10.1029/93JD01442.
Myers, R. L., & Ewel, J. J. (Eds.). (1992). Ecosystems of Florida. University of Central Florida
Press, Orlando, FL, USA.
Obeysekera, J., Browder, J., Hornung, L., & Harwell, M. A. (1999). The natural South Florida
system I: Climate, geology, and hydrology. Urban Ecosystems, 3(3), 223–244.
doi:10.1023/A:1009552500448.
Ocheltree, T. O., & Loescher, H. W. (2007). Design of the AmeriFlux portable eddy-covariance
system and uncertainty analysis of carbon measurements. Journal of Atmospheric and
Oceanic Technology, 24, 1389–1409.
Odum, E. P. (1968). Energy Flow in Ecosystems: A Historical Review. American Zoologist, 8,
11–18.
Palmer, W. (1965). Meteorological Drought (No. 45) (p. 58). U.S. Department of Commerce
Weather Bureau.
Pankratz, A. (1983). Forecasting With Univariate Box–Jenkins Models Concepts and Cases.
Wiley, New York.
Pennings, S. C., Grant, M. B., & Bertness, M. D. (2005). Plant zonation in low‐ latitude salt
marshes: disentangling the roles of flooding, salinity and competition. Journal of
Ecology, 93(1), 159–167. doi:10.1111/j.1365-2745.2004.00959.x.
Perry, W. (2004). Elements of South Florida's Comprehensive Everglades Restoration Plan.
Ecotoxicology, 13(3), 185–193. doi:10.1023/B:ECTX.0000023564.10311.4a.
91
Piccolo, M. C. (2009). Heat Energy Balance in Coastal Wetlands. In E. M. Gerardo, E. W.
Perillo, D. R. Cahoon, & M. M. Brinson (Eds.), Coastal Wetlands An Integrated
Ecosystem Approach (First edition. pp. 211–230). Elsevier. doi:10.1016/B978-0-44453103-2.00007-7.
Pielke, R. A., Sr, Walko, R. L., Steyaert, L. T., Vidale, P. L., Liston, G. E., Lyons, W. A., &
Chase, T. N. (1999). The influence of anthropogenic landscape changes on weather in
south Florida. Monthly Weather Review, 127(7), 1663–1673. doi:10.1175/15200493(1999)127<1663:TIOALC>2.0.CO;2.
Price, J. S. (1994). Evapotranspiration from a lakeshore Typha marsh on Lake Ontario. Aquatic
Botany, 48(3-4), 261–272. doi:10.1016/0304-3770(94)90019-1.
Reddy, K. R., & DeLaune, R. D. (2008). Biogeochemistry of wetlands. Boca Raton, FL: CRC
Press.
Redfield, G. W. (2000). Ecological research for aquatic science and environmental restoration in
south Florida. Ecological Applications, 10(4), 990–1005.
Richardson, C. (2008). The Everglades Experiments. In Lessons for Ecosystem RestorationMacrophyte Community Responses in the Everglades with an Emphasis on Cattail
(Typha domingensis) and Sawgrass (Cladium jamaicense) Interactions along a Gradient
of Long-Term Nutrient Additions, Altered Hydroperiod, and Fire. New York: Springer.
doi:10.1007/978-0-387-68923-4 .
Rijks, D. A. (1969). Evaporation from a papyrus swamp. Quarterly Journal of the Royal
Meteorological Society, 95(405), 643–649. doi:10.1002/qj.49709540515.
Da Rocha, H. R., Goulden, M. L., Miller, S. D., Menton, M. C., Pinto, L. D., de Freitas, H. C., &
e Silva Figueira, A. M. (2004). Seasonality of water and heat fluxes over a tropical forest
in eastern Amazonia. Ecological Applications, 14(sp4), 22–32.
Rouse, W. R., Mills, P. F., & Stewart, R. B. (1977). Evaporation in high latitudes. Water
Resources Research, (13), 909–914. doi:1029/WR013i006p00909.
Rouse, W. R. (2000). The energy and water balance of high‐ latitude wetlands: controls and
extrapolation. Global Change Biology, 6(S1), 59–68. doi:10.1046/j.13652486.2000.06013.x.
Schedlbauer, J. L., Oberbauer, S. F., Starr, G., & Jimenez, K. L. (2010). Seasonal Differences in
the CO2 Exchange of a Short-Hydroperiod Florida Everglades Marsh. Agricultural and
Forest Meteorology, 150(2010), 994–1006. doi:10.1016/j.agrformet.2010.03.005.
92
Schedlbauer, J. L., Oberbauer, S. F., Starr, G., & Jimenez, K. L. (2011). Controls on sensible
heat and latent energy fluxes from a short-hydroperiod Florida Everglades marsh. Journal
of Hydrology, 411(3-4), 331–341. doi:10.1016/j.jhydrol.2011.10.014.
Schedlbauer, J. L., Munyon, J. W., Oberbauer, S. F., & Gaiser, E. E. (2012). Controls on
ecosystem carbon dioxide exchange in short- and long-hydroperiod Florida Everglades
freshwater marshes. Wetlands, 32, 801–812. doi:10.1007/s13157-012-0311-y.
Schotanus, P., Nieuwstadt, F. T. M., & De Bruin, H. A. R. (1983). Temperature measurement
with a sonic anemometer and its application to heat and moisture fluxes. Boundary-Layer
Meteorology, 26(1), 81–93. doi:10.1007/BF00164332.
Soucha, C., Wolfe, C. P., & Grimmtind, C. S. B. (1996). Wetland evaporation and energy
partitioning: Indiana Dunes National Lakeshore. Journal of Hydrology, 184(3-4), 189–
208. doi:10.1016/0022-1694(95)02989-3.
Souch, Susan, C., Grimmond, B., & Wolfe, C. P. (1998). Evapotranspiration rates from wetlands
with different disturbance histories: Indiana Dunes National Lakeshore. Wetlands, 18(2),
216–229. doi:10.1007/BF03161657.
Stanton, E. A., & Ackerman, F. (2007). Florida and climate change: the costs of inaction.
Florida and climate change: the costs of inaction. Global Development and Environment
Institute, Tufts University.
Starr, G., Staudhammer, C. L., Loescher, H.W., Mitchell, R. J., Whelan, A., Hiers, K., &
O’brien, J.J. (In press). Time series analysis of Forest Carbon Dynamics: Initial Evidence
of Resiliency in Longleaf Pine Ecosystem Physiology following a Prescribed Fire. New
Forest.
Sutula, M., Jw, D., CABLE, J., & Rudnick, D. (2013). Hydrological and nutrient budgets of
freshwater and estuarine wetlands of Taylor Slough in Southern Everglades, Florida
(U.S.A.). Biogeochemistry, 56, 287–310.
Teal, J. M., & Kanwisher, J. W. (1970). Total Energy Balance in Salt Marsh Grasses. Ecology,
51(4), 690–695. doi:10.2307/1934050.
Thomas, G., & Rowntree, P. R. (1992). The Boreal Forests and Climate. Quarterly Journal of
the Royal Meteorological Society, 118(505), 469–497. doi:10.1002/qj.49711850505.
Twine, T. E., Kustas, W. P., Norman, J. M., Cook, D. R., Houser, P. R., Meyers, T. P., et al.
(2000). Correcting eddy-covariance flux underestimates over a grassland. Agricultural
and Forest Meteorology, 103(3), 279–300. doi:10.1016/S0168-1923(00)00123-4.
Vugts, H. F., & Zimmerman, J. T. F. (1985). The heat balance of a tidal flat area. Netherlands
Journal of Sea Research, 19(1), 1–14. doi:10.1016/0077-7579(85)90037-7.
93
Wang, H., Fu, R., Kumar, A., & Li, W. (2010). Intensification of Summer Rainfall Variability in
the Southeastern United States during Recent Decades. Journal of Hydrometeorology,
11(4), 1007–1018. doi:10.1175/2010JHM1229.1.
Wang, Hsieh, Harwell, Huang. (2007). Modeling soil salinity distribution along topographic
gradients in tidal salt marshes in Atlantic and Gulf coastal regions. Ecological Modelling,
201(3-4), 11–11. doi:10.1016/j.ecolmodel.2006.10.013.
Webb, E. K., Pearman, G. I., & Leuning, R. (1980). Correction of flux measurements for density
effects due to heat and water vapour transfer. Quarterly Journal of the Royal
Meteorological Society, 106(447), 85–100. doi:10.1002/qj.49710644707.
Wilson, K. B., Baldocchi, D. D., Aubinet, M., Berbigier, P., Bernhofer, C., Dolman, H., et al.
(2002). Energy partitioning between latent and sensible heat flux during the warm season
at FLUXNET sites. Water Resources Research, 30(12), 1294.
doi:10.1029/2001WR000989.
94
CHAPTER 4: EL NIÑO SOUTHERN OSCILLATION (ENSO) ENHANCES CO2
EXCHANGE RATES IN FRESHWATER MARSH ECOSYSTEMS IN THE FLORIDA
EVERGLADES
Abstract
This research examines the relationships between El Niño Southern Oscillation (ENSO),
water level, precipitation patterns and carbon dioxide (CO2) exchange rates in the freshwater
wetland ecosystems of the Florida Everglades. As the strongest internal climate mode,
teleconnections from ENSO are known to affect global climate and have been associated with
precipitation and water depth anomalies in the Everglades. Data was obtained over a 4-year
study period (2009-2012) from two freshwater marsh sites located in Everglades National Park
that differ in hydrology. At the short-hydroperiod site (Taylor Slough; TS) and the longhydroperiod site (Shark River Slough; SRS) fluctuations in precipitation patterns occurred with
changes in ENSO phase, suggesting that extreme ENSO phases alter Everglades hydrology
which is known to have a substantial influence on ecosystems carbon dynamics. I hypothesize
that El Niño and La Niña will amplify the site-specific seasonal response in CO2 fluxes.
Variations in both ENSO phase and annual net CO2 exchanges rates co-occurred with
changes in wet and dry season length and intensity. Combined with site-specific seasonality in
CO2 exchanges rates, El Niño and La Niña phases magnified both season intensity and CO2
exchange rates at both sites. At TS, net CO2 uptake rates were higher in the dry season, whereas
SRS had greater rates during the wet season. As La Niña phases were concurrent with drought
years and extended dry seasons, TS became a greater sink for CO2 on an annual basis (-11 to 95
110 g CO2 m-2 yr-1) compared to El Niño and neutral years (-5 to -43.5 g CO2 m-2 yr-1). SRS was
a small source for CO2 annually (64.5 to 80 g CO2 m-2 yr-1) except in one exceptionally wet year
that was associated with an El Niño phase (-16 g CO2 m-2 yr-1). Considering that future climate
predictions suggest a higher frequency and intensity in El Niño and La Niña phases, predicting
changes in water and carbon cycling presents an important problem. Linking ENSO phases,
precipitation, season intensity and CO2 dynamics in the Everglades region, these results suggest
that a higher frequency of extreme events as a result of climate change will significantly alter
CO2 dynamics in the Florida Everglades.
96
Introduction
Teleconnections from the El Niño Southern Oscillation (ENSO) are known to strongly
affect climate patterns across North America (Ropelewski & Halpert 1986; Piechota & Dracup
1996; Beckage et al., 2003; Moses et al., 2013). ENSO cycles are alternating periods of warm (El
Niño phase) and cold (La Niña phase) Pacific Ocean surface temperatures (Myers & O’Brien
1995; Trenberth 1997), and have occurred with regular periodicity (3 to 7 years) over the last
130,000 years (Beckage et al., 2003). Shown to influence worldwide precipitation patterns
(Beckage et al., 2003), ENSO phases are also correlated with global terrestrial productivity
(Behrenfeld et al., 2001) and climate anomalies (Davis & Ogden 1994; Beckage et al., 2003).
In the Florida Everglades, changes in the long-term hydrologic cycle have been linked to
extreme ENSO phases (El Niño and La Niña phases; Piechota & Dracup 1996; Allan & Soden
2008). Precipitation patterns in this region form wet and dry seasons, the frequency and
magnitude of which fluctuate with changing climate patterns (Davis & Ogden 1994). Here, El
Niño phases increase dry season rainfall causing higher seasonal and annual water levels
(Piechota & Dracup 1996; Allan & Soden 2008). In contrast, La Niña phases reduce dry season
rainfall, leading to extreme drought and the water table dropping below the soils surface
(Piechota & Dracup 1996; Beckage et al., 2003; Allan & Soden 2008; Moses et al., 2013).
Because annual shifts in carbon dioxide (CO2) exchange rates have been linked to changes in
surface hydrology (e.g. hydroperiods or the length of inundation) in short-term studies (Barr et
al., 2010; Jimenez et al., 2012; Schedlbauer et al., 2010; Malone et al., 2013), El Niño and La
Niña phases may be a significant driver of seasonal-to-interannual variations in hydrology and
ultimately the productivity of Everglades freshwater marsh ecosystems.
97
It is well known that wetland ecosystem structure and function is tightly coupled to
hydrology, and as such it controls wetland carbon (C) sequestration (Gorham 1991; Davis &
Ogden 1994; Whiting & Chanton 2001; Malone et al., 2013). Wetland CO2 exchange rates
respond to changes in surface hydrology (Schedlbauer et al., 2012; Jimenez et al., 2012; Malone
et al., 2013). The magnitudes of intra- and inter-annual fluctuations in surface hydrology are
sensitive to global climate cycles (Piechota & Dracup 1996), and directly affect CO2 exchange.
As a result, inter- and intra-annual fluctuations in CO2 exchange rates in the Everglades region
may be significantly influenced by El Niño and La Niña phases.
Increased atmospheric concentrations of CO2 and other greenhouse gases are expected to
alter the frequency of El Niño and La Niña phases (Timmermann et al., 1999). In addition to the
El Niño and La Niña -driven effects, climate change projections also suggest changes in the
magnitude and frequency of seasonal precipitation patterns, as well as higher dry season
temperatures (Christensen et al., 2007; Allan & Soden 2008). Precipitation projections suggest
wetter summers (wet season) and more severe drought (dry season) over the southeastern U.S.
(Li et al., 2011). Fluctuations in water availability as a result of these changes may alter
ecosystem structure and function.
Surface hydrology is managed differently among watersheds within Everglades National
Park, which provides a unique opportunity to examine ecosystem functions with differing
hydroperiods, while still experiencing similar climate. Schedlbauer et al., (2010) and Jimenez et
al., (2012) have assessed the effects of managed hydroperiods on seasonal and annual carbon (C)
dynamics for short periods (1 to 2 yr) in Everglades freshwater marsh ecosystems. However,
there has been no research to date that has assessed the effects of ENSO teleconnections on
seasonal and annual CO2 dynamics in the Everglades. Because El Niño and La Niña phases are
98
expected to alter the frequency and intensity of precipitation and temperature regimes, it is
unknown how, when, and with what magnitude ecosystem CO2 exchange rates will respond to
these fluctuations. However, this information is key to develop a prognostic understanding of
how these ecosystems will behave in the future.
The goal of this research is to understand the relationship between extreme ENSO phases
and intra- and inter-annual fluctuations in CO2 exchange rates (NEE, Reco, and GEE). I
hypothesize that El Niño and La Niña will amplify the site-specific seasonal responses in CO2
fluxes. At the short-hydroperiod site (Taylor Slough; TS) it has been shown that enhanced net
carbon uptake rates are associated with the dry season, while at the long-hydroperiod site (Shark
River Slough; SRS) greater net carbon uptake rates are associated with wet season conditions
(Schedlbauer et al., 2012; Jimenez et al., 2012). As El Niño and La Niña phases increase wet and
dry season intensity in the Everglades region (Piechota & Dracup 1996; Beckage et al., 2003;
Allan & Soden 2008; Moses et al., 2013), I expect the site-specific seasonal response to change
correspondingly with changes in seasonal intensity. Season intensity here refers to deviations
from the mean water availability, so that larger absolute numbers indicate the season intensity
and the sign indicates wetter (+) or dryer (-) conditions. I also hypothesize that the variation in
season length will explain differences in the interannual CO2 dynamics in the respective
processes of uptake and efflux. For example, longer and wetter wet seasons will increase the
capacity to uptake carbon at long hydroperiod sites, whereas longer and hotter dry seasons will
increase the capacity for carbon uptake at short hydroperiod sites. In this study I used the eddy
covariance method to estimate whole ecosystem exchanges in CO2, and a combination of linear,
non-linear and time series modeling techniques to statistically address these hypotheses.
99
Materials and Methods
Study Site
The Florida Everglades are classified as subtropical wetlands with a year-long growing
season and distinct wet and dry seasons that define annual seasonal variation (Chen & Gerber
1991; Obeysekera et al., 1999). Water enters the Everglades through local precipitation events,
which average 1380 mm annually (Davis & Ogden 1994), and through regional runoff.
Presently, water dynamics are controlled by the South Florida Water Management district, which
uses a complex system of canals, levees, and pumping stations (Loveless 1959; Davis & Ogden
1994). The majority of rainfall (~70%) occurs during the wet season, which begins in May or
early June with convective events and tropical depressions, e.g., thunderstorms and hurricanes
(Davis & Ogden 1994). Surface water levels generally increase throughout the wet season, are
highest at wet season end in October, and decline to their lowest levels by dry season end in May
(Beckage et al., 2003). Dry season precipitation results from frontal systems, which accounts for
~ 30% of annual precipitation (Davis & Ogden 1994).
The study sites are two oligotrophic freshwater marsh ecosystems that are within the
Florida Coastal Everglades (FCE) long-term ecological research (LTER) program in Everglades
National Park (FCE-LTER, http://fcelter.fiu.edu/research/sites/; Figure 10). Taylor Slough
( 5° 6’16.5” N, 80°35’40.68” W) is a short-hydroperiod marsh that is inundated for 4 to 6
months each year (~June to November) and is characterized by shallow (~0.14 m) marl soils
overlying limestone bedrock. Mean canopy height (Z) and surface roughness (d) for this site are
0.73 and ~0.3 m, respectively. Shark River Slough (25°33'6.72"N, 80°46'57.36"W) is a longhydroperiod marsh that is inundated ~12 months each year and is characterized by peat soils (~1
100
m thick) overlying limestone bedrock with ridge and slough microtopography (Duever et al.,
1978). For this site, Z and d are 1.02 and ~0.4 m, respectively. Differences in hydroperiod result
from spatial variability in elevation (Beckage et al. 2003) and exposure to surface runoff. For this
study, data from January 2010 to December 2012 was used.
In the Florida Everglades, species assemblages and dominance vary with hydrologic
patterns (Davis & Ogden 1994). At TS, the continuous homogeneous canopy is dominated by
short-statured (0.73 m) emergent species, Cladium jamaicense (Crantz) and Muhlenbergia
capillaris (Lam.). With ridge and slough microtopography (Duever et al., 1978) at SRS, tall
(1.34 m) dense, emergent species (e.g., Cladium jamaicense, Eleocharis sp. and Panicum sp.)
dominate ridge areas. Sloughs are dominated by short-statured (0.70 m), submerged species (e.g.,
Utricularia sp.). Periphyton exists at both sites on submerged structures and as floating mats at
SRS (for a more detailed description of the vegetation, see Davis & Ogden 1994). All research
was performed under permits issued by Everglades National Park (EVER-2009-SCI-0070 and
EVER-2013-SCI-0058).
Eddy Covariance and Meteorological Data
At each site, open-path infrared gas analyzers (IRGA, LI-7500, Li-COR Inc., Lincoln,
NE) were used to measure CO2 (c; mg mol-1) and water vapor molar density (v; mg mol-1), and
a paired sonic anemometer (CSAT3, Campbell Scientific Inc., Logan, UT) was employed to
measure sonic temperature (Ts; °K) and 3-dimensional wind speed (u, v and w, respectively; m s1
). These paired sensors were 0.09 m apart and installed at 3.30 and 3.24 m above ground level
(a.g.l.) at TS and SRS, respectively. Data were logged at 10 Hz on a datalogger (CR1000,
Campbell Scientific Inc.) and stored on 2 GB CompactFlash cards. Both IRGAs were calibrated
101
monthly using a trace gas standard for CO2 in air (+ 1.0%), dry N2 gas and a traceable dewpoint
generator (LI-610, LI-COR Inc.). Footprint analyses (Kljun et al., 2002, 2004) indicated that
80% of measured fluxes were within 100 m of the tower during convective conditions at both
sites. Other meteorological variables were measured at 1-sec and collected as half-hourly
averages, acquired by the same datalogger, and included: air temperature, (Tair; °C) and relative
humidity (Rh; %) (HMP45C, Vaisala, Helsinki, Finland) mounted within an aspirated shield
(43502, R.M. Young Co., Traverse City, MI), and barometric pressure (P; atm) (PTB110,
Vaisala). The Tair/Rh sensors were installed at the same height a.g.l. as the IRGA and CSAT.
At each site, additional meteorological data was measured at 15-sec, and collected as 30min averages through a multiplexer (AM16/32A Campbell Scientific Inc.) with another
datalogger (CR10X Campbell Scientific Inc.). This included photosynthetically active radiation
(PAR; mol m-2 s-1) (PAR Lite, Kipp and Zonen Inc., Delft, Netherlands), incident solar
radiation (Rs; W m-2) (LI-200SZ, LI-COR Inc.), and net radiation (Rn; W m-2) (CNR2-L, Kipp
and Zonen). Precipitation measurements were made with tipping bucket rain gages (mm)
(TE525, Texas Electronics Inc., Dallas, TX). Soil volumetric water content (VWC; %) was
calculated from equations developed for peat and marl soils using the methodology of Veldkamp
& O’Brien (2000), from the dielectric constant using two soil moisture sensors (CS616,
Campbell Scientific Inc.) buried between 0 and 20 cm soil depth at each site. Soil temperature
(Ts; °C) was measured at 5 cm, 10 cm, and 20 cm depths at two locations within each site using
insulated thermocouples (Type-T, Omega Engineering Inc., Stamford, CT). When inundated at
SRS, water temperature, (Tw; °C) was measured using two pairs of insulated thermocouples
(Type-T, Omega Engineering Inc.), each pair located at a fixed height 5 cm above the soil
surface and another attached to shielded floats that held the thermocouples 5 cm below the water
102
surface. At TS, Tw was measured using insulated thermocouples (Type-T, Omega Engineering
Inc.) located at a fixed height 2 cm below the water surface. Water level (m) at both sites was
recorded every half-hour with a water level logger (HOBO U20-001-01, Onset, Bourne, MA).
Data processing
Net ecosystem exchange (NEE) of CO2 was estimated through simplification of the
continuity equation by applying a control volume approach from the ground level to the top
measurement height (z; m; Loescher et al., 2006). Vertical windspeed (w) was first estimated
mean to streamline using a 2-d rotation in a Cartesian coordinate framework (Loescher et al.,
2006). NEE (μmol m-2 s-1) was then estimated using the covariance of the turbulent fluctuations
of the vertical rate of change of mean molar density of CO2 (cʹ), and the vertical scalar flux
divergence (wʹ), where the turbulent fluctuations are the instantaneous deviation (at 10 Hz) from
the mean block average (term I) over 30 min, and the storage flux (term II):
̅̅̅̅̅
⏟
∫
⏟
̅
Eq. 1
where, ̅̅̅̅̅ is the measured covariance (m s-1 μmol C mol-1) of the molar density of CO2
measured at a fixed plane above the plant canopy (Z), z is the vertical dimension, and the overbar
is the averaging period, in this case is 30-min. NEE was then divided by the molar volume of air,
V, (m3 mol-1) to convert the units from density to molar fraction, i.e., mol CO2 m-2 s-1, such that:
Eq. 2
where, R is the ideal gas constant (0.082 L atm K mol ), P is atmospheric pressure (1.10325
-1
-1
atm), and Tk is the actual air temperature, estimated by:
Eq. 3
103
where, q is the molar fraction of water vapor calculated by unit conversion of ρv.
Micrometeorological convention is used here, where negative NEE values indicate ecosystem
uptake of CO2.
Sensible heat (H; W m-2) was determined from the covariance of the turbulent
fluctuations of w and Ts, and w and q (noted as primes, rf. Loescher et al., 2006) estimated over a
30-min averaging period (noted as overbar), such that,
̅̅̅̅̅̅̅
̅̅̅̅̅̅ )
(
Eq.4
-3
where, ρair is the air density (kg m ) and Cp is the specific heat of air at constant pressure (J kg-1
˚C-1). Corrections for the effect of water vapor on the speed of sound were applied (Schotanus et
al., 1983).
Similarly, latent energy (LE; W m-2) was calculated from the covariance of the turbulent
fluctuations of w and ρv and averaged over 30-min,
̅̅̅̅̅̅
Eq.5
where, λ is the heat of vaporization (J g-1), and Mair and Mw are the molecular weights of air
(28.965 g mol-1) and water (18.01 g mol-1), respectively. Corrections for thermal and pressure
related expansion and/or contraction, and water dilution were applied (Webb et al., 1980).
10 Hz raw flux data were processed with EdiRe (v. 1.4.3.1184, Clement 1999), which
included despiking and air density corrections (Aubinet et al., 2000; Webb et al., 1980). Fluxes
(NEE, H, LE) were then corrected for mass transfer resulting from changes in density not
accounted for by the IRGA (Massman 2004; Webb et al., 1980), and barometric pressure data
were used to correct the fluxes to standard atmospheric pressure. All measurements were filtered
when systematic errors in either NEE, H or LE were indicated, such as: (1) evidence of rainfall,
condensation, or bird fouling in the sampling path of the IRGA or sonic anemometer, (2)
104
incomplete half-hour datasets during system calibration or maintenance, (3) poor coupling of the
canopy with the external atmospheric conditions, as defined by the friction velocity, u*, using a
threshold < 0.15 m s-1 (Goulden et al., 1996; Clark et al., 1999), and (4) excessive variation from
the half-hourly mean based on an analysis of standard deviations for u, v, and w wind and CO2
statistics. Quality assurance of the flux data was also maintained by examining plausibility tests
for implausible H (<-100 or >800 Wm-2), LE (<-100 or >800 Wm-2), and NEE (i.e., NEE < -30
or > 30 μmol m-2 s-1) values, stationarity criteria, and integral turbulent statistics (Foken &
Wichura 1996; Foken & Leclerc 2004). At TS, 38% and 77% of the day and nighttime data were
removed, respectively. At SRS, 34% of daytime data and 70% of nighttime data were removed.
Missing H and LE were then gap-filled using the linear relationship between H or LE and Rn on a
monthly basis. When R2 values were less than 70%, annual relationships between Rn and H or
LE were used to gap fill data in that month.
Missing half hourly NEE data were gap-filled using separate functions for day and night.
When PAR was ≥10 W m-2, NEE data was gap-filled using a Michaelis-Menton approach
(NEEday; Eq. 6), and when PAR was <10 W m-2, NEE data was gap-filled using an Arrhenius
approach (NEEnight; Eq. 7):
Eq. 6
where,  is the apparent quantum efficiency (
),  is PAR, Reco is ecosystem respiration (mol
CO2 m-2 s-1), and Pmax is the maximum ecosystem CO2 uptake rate (mol CO2 m-2 s-1).
Eq. 7
where, R0 is the base respiration rate when air temperature is 0 C and b is an empirical
coefficient. In equation 6, Reco is an estimated model parameter, whereas Reco measurements are
105
the dependent variable in equation 7. A bootstrap method was used for error estimation of gapfilled values of NEE. Synthetic datasets (1000) of size n (with replacement) from the original
dataset of size n for each estimated gap-filling model (Eq.6 and Eq.7) on a monthly or annual
basis where appropriate were used (Jimenez et al., 2012). Distributions of each model parameter
were constructed, which were then checked to ensure that the model parameters derived from the
original data were contained within a 95% confidence region. Following gap filling, GEE was
calculated from half hourly NEE and Reco data (Eq. 8).
GEE= NEE –Reco
Eq. 8
Gap-filled flux data for TS and SRS are made available through AmeriFlux
(http://ameriflux.ornl.gov).
Defining Seasons
Although the majority of rain in the Everglades region falls in the wet season, it is
difficult to identify the exact onset of the wet season. Previous studies define season based on the
calendar year (Malone et al., in review) or water levels (Schedlbauer et al., 2010; Jimenez et al.,
2012); however, these approaches either do not capture interannual variations or are heavily
influenced by water management activities performed by the South Florida Water Management
District. To determine the date of the shift in seasons I examined the seasonal pattern of Bowen
ratios over time,
Eq. 9
th
where, the subscript t denotes the t daily value in the time series. Similar to methods used by
Nuttle (1997) to define hydroperiods, a harmonic function (sine function) was fit to the β time
series to identify inflection points that indicate changes in the seasonal trend of the ratio of
106
energy dissipation as H and LE (Figure 11). A sine functions was fit to the β time series at each
site annually (Jan 1 to Dec 31), and the inflection point along the positively sloped portion of
each sine function was used to identify the change from dry to wet season (Figure 11). The wet
season was defined by fitting the sine function to the same set of site-specific series offset by 182 days (~6-months) and identifying the inflection point along the negatively sloped portion of
each sine function. The sine function was offset by -182 days so that the shifts in season would
not occur near the end of the time series. Previous studies show marked seasonal shifts in energy
dissipation in short and long hydroperiod marsh ecosystems (Malone et al., in review), and
therefore this method of seasonal classification should adequately capture the seasonality in both
water and energy availability.
As an indication of season intensity, the seasonal mean Palmer Drought Severity Index
was used (PDSI; Palmer 1965; Figure 12a). PDSI compares weather conditions to historical
weather data, taking into account temperature, rainfall, and the local available water content of
the soil. PDSI uses 0 to identify normal conditions, negative numbers (-1 to -6) to indicate dryer
than average conditions, and positive numbers to reflect excess rain (Figure 12). PDSI data were
retrieved from the National Climatic Data Center (http://www.ncdc.noaa.gov/temp-andprecip/drought/historical-palmers.php).
Long-term weather data
Long-term weather data were obtained from the nearest weather station, NCDC Royal
Palm Ranger Station, where NOAA surface meteorological data was available from1963 to
2012. The Oceanic Niño Index (ONI) was used to define ENSO phases and was retrieved from
the National Oceanic & Atmosphere Administrations Earth Physical System Research
107
Laboratory (http://www.esrl.noaa.gov/psd/data/climateindices). The ONI is the running 3-month
mean sea surface temperature (SST) anomaly from a 30-year mean SST for the Niño 3.4 region
(i.e., 5°N-5°S, 120°-170°W). Cold (warm) phases are defined as 5 consecutive months at or
below (above) the -0.5° (+0.5°) anomaly.
Data Analysis
Long-term patterns in monthly weather data. An intervention time series approach
was used (Brocklebank & Dickey 2003), utilizing autoregressive integrated moving average
(ARIMA) models to model three variables describing long-term weather for the site, (e.g.,
monthly precipitation, average maximum daily temperature, and average minimum daily
temperature) as a function of ENSO phase. ARIMA models incorporate three types of processes:
autoregressive (AR) of order p, moving average (MA) of order q, and if necessary, differencing
of degree d (Brocklebank & Dickey 2003). ARIMA models fit to time series data use AR and
MA terms to describe the serial dependence, and use other time series data from independent
variables to explain the dependence on outside factors (Brocklebank & Dickey 2003). To
facilitate the inclusion of independent categorical variables for ENSO, El Niño and La Niña
phases were coded as indicator variables, where a value 0 or 1 specified the absence or presence,
respectively, of a categorical effect. I then determined if there were teleconnection lags between
ENSO phase and precipitation and temperature.
In developing time series models, first, all data series were tested for stationarity via the
augmented Dickey-Fuller unit root test (Dickey & Fuller 1979). The ARIMA models were then
fit to time series data (monthly precipitation, and minimum and maximum temperature) using an
iterative Box-Jenkins approach, where: (1) autocorrelation and partial autocorrelation analysis
108
were used to determine if AR and/or MA terms were necessary for the given time series, (2)
model coefficients were calculated using maximum likelihood techniques and, (3)
autocorrelation plots of model residuals were examined to further determine the structure of the
model (Brocklebank & Dickey 2003).
Because of the presence of autocorrelation in the explanatory series, input series were
pre-whitened (Brocklebank & Dickey 2003). ARIMA models were then fit to the dependent
variables using the pre-whitened explanatory series as predictor variables. Cross-correlation
coefficient plots between the explanatory series and dependent variables were used to identify
direct and inverse relationships at various lags or time shifts, and autocorrelation plots of the
residuals verified that the residual series had characteristics of random error, or white noise
(Brocklebank & Dickey 2003). Model selection was based on minimum Akaike’s information
criterion (AIC) and models were acceptable when residual white noise was minimized (Hintze
2004). A backwards selection method was used, removing the least significant parameter one at a
time until all regression terms in the final model were significant at the α= 0.05 level and the
lowest AIC was achieved. ARIMA assumptions of normality and independence of residuals
(Brocklebank & Dickey 2003) were verified by examining residual plots.
Seasonal light and temperature response. To examine changes in the response of NEE
and Reco to light and temperature, respectively, over ENSO phase, site and season, non-linear
equations (6) and (7) were fit. Parameters for these models were fit by ENSO phase, site and
season via the SAS procedure PROC NLIN (version 9.3, SAS Institute Inc., Cary, North
Carolina). Parameter estimates were then compared to identify differences and similarities. As a
result of the high degree of autocorrelation inherent in NEE time series from half-hourly data, the
109
standard errors of parameter estimates from these models are artificially small, and statistical
tests are not valid. Therefore, this analysis is presented in a descriptive context.
Daily CO2 and water dynamics. An intervention time series approach was used to
identify and model the relationship between CO2 dynamics (NEE, GEE, and Reco) and a set of
explanatory variables over a 4-year time series of daily data (2009 to 2012). These variables
included: water level, season, ENSO phase (El Niño, La Niña and neutral), daily precipitation,
drought condition, and average air temperature. The combined effect of ENSO phase and season
(e.g. El Niño  wet season and El Niño  dry season) on CO2 fluxes were included in time series
models as predictors with indicator variables. Indicator variables were also developed to identify
sections of each season that directly followed an ENSO phase (post-La Niña and post-El Niño).
Beckage et al., (2003) found the effect of extreme ENSO phases during seasonal transitions, and
post-La Niña and post-El Niño phases capture the transition periods. To explore the effect of
precipitation on CO2 exchange rates, indicator variables were used to identify the day of a
precipitation event (Rain Day), the day after precipitation, and the quantity of precipitation
(Rain; mm). The indicator for the day after a precipitation event identified the first rain free day
following a day with precipitation. Finally, drought conditions were defined as those days where
PDSI < -2 and verified the drought extent with the National Oceanic and Atmospheric
Administration’s Drought Monitor (Svoboda 2000; Figure 12b and Figure 13). Drought Monitor
data was obtained from the National Drought Monitor Center
(http://droughtmonitor.unl.edu/DataArchive.aspx).
In addition, previous studies identified water level as one of the most important drivers
for CO2 exchange rates in the Everglades freshwater ecosystems (Jimenez et al., 2012;
Schedlbauer et al., 2012). Thus, a water index equal to the difference between each half-hourly
110
water level and its site-specific annual seasonal mean water level was computed. Using the water
index as a dependent variable in an additional analysis, time series models were estimated to
answer questions about the relationship between intraseasonal fluctuations in water levels,
precipitation, and ENSO phase.
As in the models of monthly weather data, all daily time series were tested for stationarity
and non-stationary series were made stationary by differencing (Pankratz, 1983). ARIMA
models were then fit to time series data (NEE, GEE, Reco, and the Water Index) using prewhitened explanatory series as predictor variables. Cross-correlation coefficient plots identified
relationships at various lags, and autocorrelation plots were used to verify that the residuals had
characteristics of random error. Model selection was based on minimum AIC, removing the least
significant parameters. For each dependent variable, a single model form was selected with
common predictor variables to aid site comparisons. Non-significant parameters remained in the
model only if the parameter was significant at one site and it did not affect the final model of the
other site. Multicollinearity between explanatory variables was also explored to ensure models
did not contain input series that were highly correlated.
Results
Throughout the study (2009-2012), 3 short La Niña phases (2009, 2010 and 2011), and 1
El Niño phase (2009-2010) occurred with shifts in precipitation patterns that resulted in both
wetter than average (2010) and drier than average years (2009 and 2011; Figure 12a). Season
intensity, defined by the seasonal average PDSI for the Everglades region, changed with ENSO
phases (Figure 12a). During La Niña phases, seasonal mean PDSI ranged from ~ -1 to -4 (Figure
12a). In the wet season of 2009 an El Niño phase began soon after a La Niña phase. The El Niño
111
phase extended into the dry season of 2010 where it co-occurred with wetter than average
conditions, PDSI >1.
Although TS and SRS had similar weather, hydroperiods and season length differed
annually and between sites (Figure 12). The onset of the wet season at TS lagged SRS by
approximately one month, with wet season length varying between 179 to 208 days at TS and
159 to 242 days at SRS. Wet season length was positively correlated with cumulative
precipitation from January to March (p=0.0267). During abnormally dry years (PDSI < -2), the
wet season was shortened by about 10 days at TS and 38 days at SRS compared to all other
years. In 2009 and 2011, south Florida experienced severe drought conditions resulting in 65 and
34 dry days (water levels below the soil surface) at TS and SRS, respectively (Figure 12 and 13).
Drought years were not characterized by lower annual precipitation but by lower rainfall and
fewer rain events the 3 months prior to the start of the wet season, which generated a shorter
season (Figure 14a). In 2010 and 2012, total rainfall in the first four months of the year averaged
276 mm at TS and 246 mm at SRS while drought years (2009 and 2011) received just 107 mm
and 82mm on average at TS and SRS, respectively.
Long-term weather patterns
Time series analysis of long-term monthly precipitation and minimum and maximum
daily temperatures versus ENSO phase showed that rain increased during El Niño, though not
significantly (p=0.1204), and declined synchronously during La Niña (p=0.0988) and at a lag of
1 month (i.e., at a lag of 1 month; p=0.0554) (Table 10). Monthly average maximum daily
temperatures were greater during La Niña (p<0.0001), whereas average minimum daily
temperatures were higher the month following the start of La Niña (i.e., at a lag of 1 month;
112
p=0.0789; Table 10). The month following the start of El Niño phases, monthly average
maximum and minimum temperature were also higher (p<0.0001; Table 10) than during neutral
and La Niña phases.
Seasonal and annual patterns in CO2 Fluxes
Annual net CO2 exchange rates at both freshwater marsh ecosystems co-varied with
ENSO phase, which corresponded to changes in season intensity. Although TS ranged from a
small CO2 sink to a small source on an annual basis over the four years, TS was a source for CO2
during the wet season and a sink during the dry season (Table 11). Changes in GEE relative to
Reco resulted in seasonal shifts in NEE, though there was no consistent pattern in dry season
versus wet season response in GEE or Reco. At TS, CO2 uptake rates were higher in the dry
season and during the exceptionally dry La Niña years, which corresponded to drought
conditions (Table 11). The mean annual dry season length at TS was 176 days for the four years.
During years with La Niña phases, CO2 uptake was higher, dry seasons were 10 days longer on
average, and TS was a greater sink for CO2 (Table 11). As a result of an extended drought in
2011 that occurred with two La Niña phases, TS was a sink for CO2 in both wet and dry seasons
(Table 11).
Like patterns observed at TS, annual variation in NEE corresponded to changes in ENSO
phase at SRS. Although SRS ranged from CO2 neutral to a small source of CO2 to the
atmosphere annually and seasonally, CO2 release rates increased during seasons with La Niña
phases. Ecosystem respiration was the primary control on annual ecosystem carbon balance
(Table 11) and dry season mean Reco was often higher than wet season Reco, increasing CO2
release at SRS when the dry season was extended. Similar to TS, average daily CO2 uptake rates
113
were higher in the dry season at SRS (Table 11). Even so, dry season mean daily Reco rates were
also higher than in the wet season (Table 11). Because Reco increased relative to CO2 uptake in
the dry season (Table 11), longer dry seasons were associated with greater CO2 source status at
SRS. During the exceptionally wet year that corresponded to an El Niño phase (2010), GEE
surpassed Reco and the site became a larger sink for CO2 (Table 11).
Plots of dry season length versus dry season cumulative NEE (Figure 14b and 14c)
revealed differences by site. Although there were very few observations available since the study
period included just four dry seasons, NEE at TS exhibited a negative linear relationship with dry
season length (p= 0.1476; Figure 14b), while NEE at SRS showed a positive linear relationship
(p=0.0705; Figure 14c). These results demonstrate that the seasonal response in NEE rates
differed between sites, and suggest that dry season length (and changes in dry season length)
may control the CO2 source and sink status in the future. Although, the data set is not yet large
enough to confirm the pattern between dry season length and dry season cumulative NEE as a
characteristic of each site, this relationship is an important indication of how the sites respond
differently to hydroperiods.
Seasonal light and temperature response
In addition to site differences, ENSO phase and season altered photosynthetic capacity
and ecosystem respiration (Figure 15; Table 12). At both sites, photosynthetic capacity (Pmax)
was greatest during La Niña and directly following La Niña phases. Seasonal differences in light
and temperature response curves were greater, and photosynthetic capacity and dark respiration
(Reco) were consistently higher at TS, for all ENSO phases as compared to SRS (Figure 15a and
15c; Table 12). At both sites, the effect of El Niño and La Niña phase was also greater during the
114
dry season. At SRS, photosynthetic capacity was similar for all ENSO phases during the wet
season though the effect of El Niño and La Niña phases increased during the dry season (Figure
15b and 15d; Table 12). Overall, there was a small seasonal difference in photosynthetic
capacity, which was higher on average in the wet season than in the dry season at SRS. At both
sites, differences in photosynthetic capacity by ENSO phase were greatest at higher PAR values
(> 1000 μmol m-2 s-1).
The relationship between temperature and Reco differed between ENSO phases, and Reco
showed distinct seasonal patterns in temperature sensitivity (Figure 15e - 15h; Table 12) at TS
and SRS. During the wet season at both sites, Reco was less sensitive to temperature changes and
temperature effects associated with El Niño and La Niña phases were greater at lower
temperatures (Figure 15e and 15f; Table 12). At higher temperatures Reco was more sensitive to
changes in temperature during all phases at both sites, a response that was enhanced during the
dry season (Figure 15g and 15h; Table 12). At SRS, the differences among ENSO phases were
small except at high temperatures (>24°C), while at TS the differences among ENSO phases
were consistently large at lower temperatures and converged at high temperatures (Figure 15e 15h). Similar to patterns observed in light response curves, respiration rates were higher at TS
than at SRS, and temperature patterns associated with Reco also showed greater release of CO2 at
both sites in the dry season versus the wet season (Figure 15e - 15h).
The effect of ENSO phase, precipitation, and season on daily CO2 exchange rates and water
level
After pre-whitening, some small (p<0.05) but statistically significant autocorrelation
remained in pre-whitened series; however, this sensitivity resulted from the large number of
observations available and was judged to be biologically insignificant (Starr et al, in review).
115
Differencing was required for water level and PDSI time series due to the lack of stationarity at
both sites. Non-stationarity indicates a lack of stability in the mean of these variables over time,
further suggesting that there were significant changes in hydroperiods at both sites. By including
differenced variables in models I evaluated how changes in water level and PDSI influenced
NEE, Reco, and GEE. In response to evidence of 1-month lagged teleconnections for ENSO phase
effects on precipitation and average daily maximum and minimum temperatures (Table 10),
lagged El Niño and La Niña phase indicators were included in time series models of CO2
exchange rates.
Models for NEE, had a significant lagged 1-day MA [MA(1)] component, as well as
significant AR components at a lag of 1-, and 2-days (p<0.0001;Table 13). At both sites, Δwater
level (p<0.0001) and the quantity of rain (mm day-1; p<0.0001) had significant positive
relationships with NEE, showing that as the change in water level and the quantity of daily
rainfall increased, net CO2 uptake decreased (higher NEE; Table 13). Post-La Niña phases in the
dry season were associated with significantly lower NEE (higher net CO2 uptake) at TS
(p=0.0002); however, post-La Niña phases in the dry season at SRS were associated with greater
NEE (lower net CO2 uptake; p=0.0168). Moreover, there was a significant decrease in NEE at
SRS (higher net CO2 uptake) the day after rain at SRS. At TS where hydroperiods were shorter,
the effect of rain and post-La Niña phase during the dry season were significantly stronger than
at SRS (Table 13). The quantity of rain was the strongest driver of NEE at both sites.
Models for Reco, at both sites had a significant 1-day MA [MA(1)], as well as significant
AR components at a lag of 1-, 2-, and 14-days (p<0.05; Table 14). The day after rain (p=0.0124),
and post-La Niña phases in the dry season (p=0.0452) were significantly associated with reduced
daily ecosystem respiration rates at TS (Table 14). At SRS the quantity of rain (p=0.0094) and
116
post-La Niña phases in the dry season (p<0.0001) were significantly linked to increased Reco
(Table 14), and the effect of post-La Niña phases in the dry season was larger at SRS where it
was associated with increased Reco. At SRS, seasonal ENSO phase (ENSO × Season) had the
greatest impact on Reco, while at TS Reco had the strongest association with the day following rain
events.
Similar to NEE, models for GEE had a significant 1-day lagged MA [MA(1)] and AR
[AR(1)] component at both sites (Table 15). The day of rain, day after rain, quantity of rain
(mm), and Δwater level had significant positive relationships with GEE at TS (Table 15). At
SRS, days with precipitation and the quantity of precipitation had positive relationships with
GEE (Table 15). Rain had a stronger effect at TS than at SRS, and rain had the strongest effect
on GEE at both sites (Table 15).
Models for the water index had a significant 1-day lagged MA [MA(1)] and AR [AR(1)]
component at both sites (p<0.0001; Table 16). At TS, the quantity of rain (p=0.0514), La Niña
and post-La Niña phases (wet season; p<0.0001), and post-El Niño phases (dry season;
p=0.0004) were associated with lower than average water levels (Table 16). Water levels at TS
were higher than the seasonal average the day after rain (p<0.0001), days with precipitation
(p=0.0014), during dry season EL Niño phases (p<0.0001), and throughout post-La Niña phases
that occurred at the end of the dry season (p<0.0001). At SRS, the day of (p=0.0267) and the day
after rain (p=0.0002), and post-La Niña phase at the end of the dry season (p<0.0001) were
associated with higher than average water levels (Table 16). La Niña and post-La Niña phases
(wet season; p<0.0001), and post-El Niño phases (dry season; p<0.0001) were all associated with
lower than average water levels. The effect of La Niña (wet season), El Niño (dry season), and
117
post-La Niña (wet season) phases were significantly greater at TS than at SRS, and El Niño and
La Niña phases were the strongest predictors of the water index (Table 16).
Discussion
The goal of this research was to understand the relationship between ENSO phases and
CO2 exchange rates (NEE, Reco and GEE) in Everglades freshwater marsh ecosystems. The
relationships between ENSO extremes, precipitation and hydrology in the Everglades region
suggest that El Niño and La Niña phases could be important for C dynamics (Ropelewski &
Halpert 1986, 1987; Kiladis & Diaz 1989; Hanson & Maul 1991; Sittel 1994; Livezey et al.,
1997; Mason & Goddard 2001; Schmidt et al., 2001; Beckage et al., 2003; Childers et al., 2006).
The results presented here show that climate teleconnections have significant controls on
Everglades CO2 dynamics, demonstrating that in addition to climate change and water
management, ENSO is an additional source of variation in C cycling in Everglades freshwater
ecosystems.
Annual fluctuations in season length and intensity
Interannual variability in precipitation can be large in the Everglades region (Duever et
al., 1994; Obeysekera et al., 1999). Throughout the study period, large fluctuations in seasonal
precipitation and hydrology were observed. Reduced precipitation prior to the onset of the wet
season resulted in shorter wet seasons and lower water levels, though annual precipitation was
unchanged. Cumulative precipitation in January through March was related to wet season length
(Figure 14a). Although more data is needed to validate this relationship, previous research
suggests dry season rainfall from October to April largely determines season intensity (Duever et
al., 1994; Beckage et al., 2003). Higher precipitation from January to March was also associated
118
with longer and wetter than average wet seasons. Interannual variation in the onset and length of
seasons can have a significant effect on the magnitude of ecosystem primary production
(Randerson et al., 1999), and changes in season intensity can either suppress or enhance
production (Cleland et al., 2007). Knowing that precipitation patterns were driving the variations
observed in season length and intensity, I examined the co-occurrence of ENSO phases,
previously found to alter season intensity in the Everglades region (Beckage et al., 2003;
Childers et al., 2006), to PDSI defined season intensity.
ENSO phases have been linked to climate anomalies on a global scale (Mason &
Goddard 2001), and at the study sites, abnormal precipitation patterns over the study period
coincided with El Niño and La Niña phases. Similar to results found by Beckage et al., 2003, El
Niño (La Niña) phases were correlated with increased (decreased) rainfall and water levels.
During El Niño (La Niña) phases, the equatorial (poleward) displacement of the midlatitude jet
increases (decreases) frontal precipitation in the southeastern United States (Ropelewski &
Halpert 1986, 1987; Kiladis & Diaz 1989). In Florida, El Niño is positively correlated with
winter (dry season) precipitation, explaining up to 34 % of dry season precipitation variability
(Moses et al., 2013). Here, El Niño La Niña phases have been shown to reduce seasonal
differences in rainfall without altering annual precipitation inputs (Childers et al., 2006). These
results support the previously observed patterns in ENSO phases, precipitation, and hydrology.
El Niño and La Niña phases induced fluctuations in season intensity (+ and -, respectively),
which affect season length, and have important implications for annual CO2 exchange rates.
119
Seasonal patterns in CO2 exchange rates
Hydroperiods have shaped soil conditions and species composition at each site in ways
that have led to different seasonal patterns in CO2 exchange rates. Hydroperiods alter ecosystem
production by interfering with exposed leaf area (Schedlbauer et al., 2010; Schedlbauer et al.,
2012; Jimenez et al., 2012), triggering senescence (S. Oberbauer, unpublished data, 2012), and
allowing CO2 fixation within the water column (Schedlbauer et al., 2012). At TS where
hydroperiods were shorter, net CO2 uptake was greater annually. Water levels rose during the
wet season, reducing photosynthetic capacity, and allowing Reco to dominate GEE. In the dry
season, mean GEE surpassed Reco as water level declined, exposing emergent leaf area and soil.
Although soil exposure would suggest greater soil respiration rates, the high water holding
capacity of the soil maintained soil water conditions. These patterns in CO2 exchange rates
coincide with observed changes in species activity (S. Oberbauer, personal observation).
Emergent species appear to drive CO2 exchange rates at TS (Schedlbauer et al., 2012). Rising
water levels lower the exposed leaf area of emergent species (Schedlbauer et al., 2010) and result
in the senescence of the sub-dominant Muhly grass (Muhlenbergia capilaris; S. Oberbauer,
personal observation). Here, hydroperiods may have also improved CO2 fixation by increased
periphyton biomass and enhanced CaCO3 production (Schedlbauer et al., 2010) in the water
column, though not enough to account for the reduction in emergent productivity.
Differences in seasonal CO2 exchange rates at TS and SRS are the result of varying
ecological responses to hydrology. Although both TS and SRS are dominated by sawgrass
(Cladium jamaicense), sub-dominant species can differ substantially (Davis & Ogden 1994). At
SRS, emergent species also dominate ecosystem CO2 exchange rates (Schedlbauer et al., 2012),
and sawgrass is generally taller and denser on the ridges than at TS (Davis & Ogden 1994).
120
Aquatic vegetation may also play a larger role in ecosystem productivity rates at SRS. At TS and
SRS, seasonal changes in productivity and GEE relative to Reco varied, as different modes of CO2
fixation and release became more or less active (Schedlbauer et al., 2012; Jimenez et al., 2012).
Similar to TS, CO2 uptake also improved during the dry season at SRS; however, Reco surpassed
CO2 uptake, resulting in the ecosystem being a small source of CO2 to the atmosphere. As a
result, longer dry seasons at SRS were associated with lower NEE. Although SRS was also a
small source or near neutral annually, GEE relative to Reco increased during the wet season. In
exceptionally wet periods (2010), SRS became a greater sink for CO2 (lower NEE).
Seasonal changes in photosynthetic capacity (Figure 15) and the relationship between
respiration rates and temperature support the patterns previously found in Everglade freshwater
marsh studies (Schedlbauer et al., 2012; Jimenez et al., 2012). Knowing that season intensity
changed with El Niño (+) and La Niña (-) phases (Figure 12a), I expected and saw a
magnification of the site-specific seasonal response in CO2 exchange rates during and directly
following El Niño and La Niña phases. The lag in El Niño and La Niña phase effect on CO2
exchange rates is the result of the effect of extreme ENSO phases on water levels during
transition periods. El Niño (La Niña) phases increase (decrease) surface water levels during
seasonal transitions and, major drainages often contain no water during transitions in a La Niña
phase (Beckage et al., 2003).
The sites differed in their response to El Niño and La Niña phases. At TS, CO2 exchange
rates were sensitive to El Niño and La Niña phase in both the wet and dry seasons, while at SRS
the effect of El Niño and La Niña phase on the wet season was not as strong as the effect in the
dry season. La Niña phases resulted in higher photosynthetic capacity and greater seasonal net
carbon uptake rates in both seasons at TS compared to neutral and El Niño phases. While La
121
Niña increased photosynthetic capacity at SRS, the increase in Reco reduced net exchange rates
compared to neutral and El Niño phases. As a result of changes in GEE relative to Reco, El Niño
led to greater net CO2 uptake at SRS during the dry season compared to La Niña and neutral
phases. These results indicate that there is a significant relationship between El Niño and La
Niña phase and season intensity, either creating conditions wetter (+) or dryer (-) than normal,
which magnifies CO2 exchange rates at TS and SRS.
Annual patterns in CO2 exchange rates
The results presented here show that the length and intensity of the wet and dry season in
the Everglades region varied annually with climate patterns. Considering the site-specific
response to season, these results support my hypothesis that variations in season length would
explain interannual fluctuations in NEE, Reco, and GEE. At TS where mean GEE surpassed Reco
during the dry season, an increase in dry season length and intensity amplified the site’s net CO2
uptake rates. Annually, SRS was usually a small source of CO2, although when wet season
conditions intensified during El Niño, net CO2 uptake increased. The effect of ENSO phase also
differed by site, showing that longer hydroperiods mute the effect of climate fluctuations on CO2
exchange rates, and as water levels decline the system becomes more vulnerable to climate. The
ecosystem’s sensitivity to climate fluctuations has important implications for water management
and climate change (Stanton & Ackerman 2007). The uncertainty of climate change makes it
important to understand how ecosystems respond to climate events and how these responses
aggregate to form trends in net CO2 exchange rates.
122
Effect of climate change and water management on net CO2 exchange rates
Patterns observed in ENSO phases and the co-occurrence of extreme wet and dry seasons
suggest changes in climate patterns can significantly alter ecosystem function. Equatorial Pacific
SST during the past half century show a clear warming trend that is consistent with global
warming (Latif & Keenlyside 2009), and El Niño and La Niña phases are expected to continue
increasing in severity and frequency (Timmermann et al., 1999; Beckage et al., 2003; Cleland et
al., 2007). Moreover, as a result of warming SST, ENSO amplitude may become even stronger,
intensifying feedbacks relevant to ENSO phases (Latif & Keenlyside 2009). If the frequency and
intensity of strong climatic disturbances increases beyond historical averages, altered disturbance
regimes have the capacity to significantly modify ecosystem processes (Rolim & Jesus 2005).
In sub-tropical ecosystems, phenology is less sensitive to temperature and photoperiod,
and more tuned to seasonal shifts in precipitation (Reich 1995; Morellato 2003; Sanchez et al.,
2003). Such shifts are expected to occur in concert with rising global temperatures, but both the
direction and magnitude of change vary regionally (Cubasch et al., 2001; Cleland et al., 2007).
As climate change has the potential to alter hydrologic regimes, we can expect to see significant
variations in CO2 exchange rates. Shifts in water management and land use change (e.g.,
conversion to agriculture and urban development) could also significantly alter both hydrology
and CO2 dynamics in the Everglades region, making it important to develop a baseline
understanding of how hydroperiod drives changes in CO2 dynamics and how climate alters
hydrology. With water managers striving to adjust hydroperiods closer to natural values, in the
future we might expect water levels in TS to increase, offsetting changes in climate by
maintaining current patterns in hydrology. Alternatively, we might anticipate higher water levels
to increase hydroperiods, making the system less sensitive to climate change altogether. With
123
longer hydroperiods, SRS will likely remain less sensitive to changes in climate and land
management. As a result, the two sites will likely behave more similarly in the future as SRS
remains neutral or a very small source of CO2 to the atmosphere and TS becomes more neutral.
124
Table 10. Parameter estimates from ARIMA models of monthly precipitation, and monthly average of daily maximum and minimum
temperature. MA(1) and MA(2) are the estimated moving average term at a 1- and 2-period lags (1 and 2 months, respectively) and
AR(2), AR(7), and AR(12) are the estimated autoregressive term at a 2-, 7-, and 12-period lags (2, 7 and 12 months, respectively).
Lagged values of independent variables are denoted similarly. El Niño is an indicator for El Niño phases and La Niña is an indicator
for the La Niña phase, determined by the ONI index.
Precipitation
Standard
t Value
Error
Parameter
Estimate
MA(1)
MA(2)
AR(1)
AR(2)
AR(7)
AR(12)
El Niño
El Niño (1)
La Niña
La Niña (1)
0.2427
0.0521
4.66
Approx
Pr > |t|
<.0001
0.4785
0.0404
11.85
<.0001
0.4581
11.3993
0.0352
7.3407
13.01
1.55
<.0001
0.1204
128.8878
-33.5155
17.4999
17.4971
-1.65
1.92
0.0988
0.0554
Temperature (max)
Standard
Estimate
t Value
Error
Approx
Pr > |t|
Temperature (min)
Standard
Estimate
t Value
Error
Approx
Pr > |t|
-0.1875
0.0537
6.37
0.0005
-0.1086
0.0426
-2.55
0.0107
0.7658
0.2314
0.0364
0.0364
21.02
-3.49
<.0001
<.0001
0.9725
0.0094
103.14
<.0001
-0.3361
2.0005
1.7154
0.4386
0.4350
0.3531
-0.77
4.60
4.86
0.4434
<.0001
<.0001
0.0242
2.5632
1.5680
-1.1216
0.6425
0.6352
0.6389
0.6383
0.04
4.04
2.45
-1.76
0.9700
<.0001
0.0141
0.0789
125
Table 11. Seasonal and annual NEE, GEE, and Reco (g C m-2 yr-1) at Taylor Slough and Shark River Slough (2009 to 2012). Seasons
with a La Nina or El Nino phase are marked with an * and , respectively.
TS
SRS
(7.3)
(7.1)
231.9
213.2
Season
Length
176
(8.8)
189
-456.1
(12.9)
445.1
(16.1)
(5.6)
(7.1)
(12.7)
-199.2
-219.3
-418.5
(4.2)
(6.0)
(10.2)
180.1
233.1
413.2
(5.3)
(7.0)
(12.2)
157
208
(13.3)
-302.7
-308.6
(14.0)
246.9
253.8
(11.6)
186
(15.4)
(27.0)
179
500.8
Year
Season
-30.0
19.0
(5.9)
2009
Dry
Wet
Annual
2010
2011
2012
NEE (S.E)
GEE (S.E)
(5.8)
(7.8)
-261.9
-194.2
-11.0
(13.7)
Dry
Wet
Annual
-19.1
13.8
-5.3
Dry
Wet
-55.7
-54.8
Annual
-110.5
(14.8)
(28.1)
Dry
-75.4
Wet
Annual
Reco (S.E)
-611.3
12.8)
(26.8)
(7.6)
-249.9
(7.9)
174.6
(7.7)
185
31.5
(8.3)
-123.3
(8.3)
154.9
(8.0)
181
-43.8
(15.9)
-373.2
(16.2)
329.4
(15.7)
ENSO
*

(4.5)
159
-361.3
(17.0)
441.3
(9.7)
(3.4)
(8.3)
(11.6)
-249.3
-92.3
-341.6
(3.3)
(7.4)
(10.8)
238.0
87.6
325.6
(1.9)
(5.8)
(7.7)
123
242
(12.7)
-230.9
-151.6
(7.0)
247.7
211.2
(10.8)
164
66.3
13.7
(11.3)
GEE (S.E)
(11.6)
(3.6)
-163.0
-198.3
80.0
(14.9)

*
-11.3
-4.7
-16.0
*
*
16.9
59.5
*
126
(5.2)
(5.4
229.3
211.9
Season
Length
206
NEE (S.E)
Reco (S.E)
(7.3)
(20.0)
-382.5
(7.4)
(14.4)
458.9
(7.4)
(18.1)
201
76.4
55.9
(4.7)
-115.8
(3.4)
171.7
(3.1)
172
8.5
(7.2)
-137.0
(5.0)
145.5
(4.7)
194
64.5
(11.9)
-252.8
(8.4)
317.3
(7.8)
Table 12. Model estimates from Eq. 6 (Light Response Curve) and 7 (Temperature Response Curve) for TS and SRS by ENSO phase
and season.
Site
TS
SRS
ENSO
El Niño
La Niña
Post-El Niño
Post-La Niña
El Niño
La Niña
Post-La Niña
El Niño
La Niña
Post-El Niño
Post-La Niña
El Niño
La Niña
Post-La Niña
Season
Dry
Dry
Dry
Dry
Wet
Wet
Wet
Dry
Dry
Dry
Dry
Wet
Wet
Wet
Light Response Curves
α
Pmax
Reco
-0.0077
-4.2800
1.2063
-0.0088
-5.7676
1.4031
-0.0200
-5.4761
1.8287
-0.0185
-6.0613
1.9068
-0.0122
-1.9836
0.8323
-0.0134
-3.4053
1.1470
-0.0209
-3.6654
1.2564
-0.0149
-2.0916
0.8782
-0.0138
-2.3326
0.9367
-0.0034
-1.7188
0.2540
-0.0643
-4.1691
2.9143
-0.0281
-2.2678
1.5688
-0.0217
-2.5837
1.1998
-0.0359
-3.1999
1.6757
127
Temperature Response Curves
R0
b
0.4361
0.0482
0.6134
0.0335
1.0559
0.0058
0.7596
0.0170
0.4096
0.0319
0.6350
0.0183
0.2090
0.0608
0.3408
0.0344
0.3846
0.0414
0.7255
0.2496
0.3184
0.3089
0.0198
0.0539
0.0409
0.0418
Table 13. Parameter estimates from ARIMA models of daily NEE by site. MA(1) is the estimated moving average term at a 1-period
lag (1 day), and AR (1) and AR(2) are the estimated autoregressive terms at a 1- and 2-period lags (1 and 2 days). Lagged values of
independent variables are denoted similarly. Asterisks denote significant differences between sites. Day after rain is an indicator for
the first rain free day, Rain is the quantity of precipitation (mm), ΔWater Level is the change in water level from one day to the next,
Post-La Niña (Wet Season) is an indicator for the time directly following a La Niña phase in the wet season, and Post-La Niña (Dry
Season) is an indicator for the time directly following a La Niña phase in the dry season.
Parameter
Estimate
MA(1)
AR(1)
AR(2)
Day After Rain
Rain (mm)
ΔWater Level (m)
Post-La Niña (Wet Season)
Post-La Niña (Dry Season)
0.8040
1.2829
-0.2997
0.0064
0.0068
0.9942
-0.2584
-0.2698
Taylor Slough
Standard
t Value
Error
0.0298
27.02
0.0430
29.85
0.0399
-7.51
0.0159
0.40
0.0006
10.69
0.2215
4.49
0.1358
-1.90
0.0727
-3.71
Approx
Pr > |t|
<.0001
<.0001
<.0001
0.6891
<.0001
<.0001
0.0570
0.0002
128
Estimate
*
*
0.7714
1.3315
-0.3491
-0.0271
0.0032
1.8851
0.0203
0.1831
Shark River Slough
Standard
t Value
Error
0.0365
21.15
0.0493
27.03
0.0458
-7.63
0.0137
-1.98
0.0005
5.84
0.4416
4.27
0.0986
0.21
0.0766
2.39
Approx
Pr > |t|
<.0001
<.0001
<.0001
0.0475
<.0001
<.0001
0.8372
0.0168
Table 14. Parameter estimates from ARIMA models of daily Reco by site. MA(1) is the estimated moving average term at a 1- period
lag (1 day), and AR (1), AR(2), and AR(14) are the estimated autoregressive terms at a 1-, 2-, and 14- period lags (1, 2, and 14 days).
Lagged values of independent variables are denoted similarly. Asterisks denote significant differences between sites. Day after rain is
an indicator for the first rain free day, Rain is the quantity of precipitation (mm), Post-La Niña (Wet Season) is an indicator for the
time directly following a La Niña phase in the wet season, and Post-La Niña (Dry Season) is an indicator for the time directly
following a La Niña phase in the dry season.
Parameter
Estimate
MA(1)
AR(1)
AR(2)
AR(14)
Day After Rain
Rain (mm)
Post-La Niña (Dry Season)
Post-El Niño (Dry Season)
0.9607
1.6846
-0.6707
-0.0139
-0.0183
-0.0003
-0.0833
0.1457
Taylor Slough
Standard
t Value
Error
0.0384
25.01
0.0481
35.02
0.0429
-15.62
0.0070
-2.00
0.0073
-2.50
0.0003
-1.01
0.0416
-2.00
0.0770
1.89
Approx
Pr > |t|
<.0001
<.0001
<.0001
0.0451
0.0124
0.3110
0.0452
0.0585
129
Estimate
*
*
*
*
0.7238
1.2317
-0.2589
0.0265
0.0032
0.0009
0.1850
0.0034
Shark River Slough
Standard
t Value
Error
0.0502
14.42
0.0609
20.24
0.0520
-4.98
0.0122
2.17
0.0084
0.38
0.0003
2.60
0.0431
4.30
0.0748
0.05
Approx
Pr > |t|
<.0001
<.0001
<.0001
0.0297
0.7030
0.0094
<.0001
0.9636
Table 15. Parameter estimates from ARIMA models of daily GEE by site. MA(1) is the estimated moving average term at a 1-period
lag (1 day), and AR (1) is the estimated autoregressive term at a 1-period lag (1 day). Lagged values of independent variables are
denoted similarly. Asterisks denote significant differences between sites. Rain Day is an indicator for days with precipitation >0, Day
After Rain is the first rain-free day, Rain is the quantity of precipitation (mm), ΔWater Level is the change in water level from one day
to the next, and Post-La Niña (Wet Season) is an indicator for the time directly following a La Niña phase in the wet season.
Parameter
Estimate
MA(1)
AR(1)
Rain Day
Day After Rain
Rain (mm)
ΔWater evel (m)
Post-La Niña (Wet Season)
0.6198
0.9977
0.0799
0.0328
0.0062
0.8157
-0.1922
Taylor Slough
Standard
t Value
Error
0.0210
0.0016
0.0149
0.0142
0.0006
0.1981
0.1296
29.56
643.63
5.35
2.32
9.94
4.12
-1.48
Approx
Pr > |t|
<.0001
<.0001
<.0001
0.0206
<.0001
<.0001
0.1380
130
Estimate
*
0.6435
0.9968
0.0455
-0.0149
0.0021
0.7440
0.0369
Shark River Slough
Standard
t Value
Error
0.0204
0.0020
0.0141
0.0134
0.0006
0.4320
0.0837
31.53
502.57
3.22
-1.11
3.62
1.72
0.44
Approx
Pr > |t|
<.0001
<.0001
0.0013
0.2678
0.0003
0.085
0.6597
Table 16. Parameter estimates from ARIMA models of the daily water index by site. MA(1) is the estimated moving average term at a
1-period lag (1 day), and AR (1) is the estimated autoregressive term at a 1-period lag (1 day). Lagged values of independent variables
are denoted similarly. Asterisks denote significant differences between sites. Rain Day is an indicator for days with precipitation >0,
Day After Rain is the first rain free day, Rain is the quantity of precipitation (mm), La Niña (Wet Season) is an indicator for wet
season La Niña phases, El Niño (Dry Season) is an indicator for dry season El Niño phases, Post-La Niña (Wet Season) is an indicator
for the time directly following a La Niña phase in the wet season, Post-La Niña (Dry Season) is an indicator for the time directly
following a La Niña phase in the dry season, and Post-El Niño (Dry Season) is an indicator for the time directly following a El Niño
phase in the dry season.
Parameter
Estimate
MA(1)
AR(1)
Day After Rain
Rain Day
Rain (mm)
La Niña (Wet Season)
El Niño (Dry Season)
Post-La Niña (Wet Season)
Post-La Niña (Dry Season)
Post-El Niño (Dry Season)
-0.2338
0.9917
0.0298
0.0225
-0.0005
-1.7029
0.9780
-1.7701
0.6149
-0.3638
Taylor Slough
Standard
t Value
Error
0.0263
0.0033
0.0069
0.0070
0.0002
0.0637
0.0694
0.0987
0.0507
0.1022
-8.88
301.95
4.34
3.20
-1.95
-26.74
14.09
-17.94
12.12
-3.56
Approx
Pr > |t|
<.0001
<.0001
<.0001
0.0014
0.0514
<.0001
<.0001
<.0001
<.0001
0.0004
131
Estimate
*
*
*
-0.1912
0.9883
0.0193
0.0117
0.0001
-0.7931
-0.0792
-1.2425
0.4584
-0.4192
Shark River Slough
Standard
t Value
Error
0.0262
0.0040
0.0052
0.0053
0.0002
0.0540
0.0787
0.0541
0.0425
0.0848
-7.31
248.72
3.74
2.22
0.46
-14.69
-1.01
-22.95
10.79
-4.95
Approx
Pr > |t|
<.0001
<.0001
0.0002
0.0267
0.6456
<.0001
0.3139
<.0001
<.0001
<.0001
Figure 10. Short- (TS) and long- (SRS) hydroperiod fresh water marsh study ecosystems within
Everglades National Park, Florida, U.S.A.
132
Figure 11. Time series of β for (a) TS and (b) SRS, used to determine the change from dry to wet, and wet to dry seasons. Sine
functions were fit to the β time series by year and by sites. The initiation of the dry season was determined by fitting a sine function
annually (Jan 1 to Dec 31). The inflection point along the positively sloped portion of each sine function identified the change from
dry to wet season. The initiation of the wet season was defined by fitting a sine function offset by -182 days (~6 months), and
determining the inflection point along the negatively sloped portion of each sine function. The shaded region highlights the wet
season.
133
Figure 12. Time series of precipitation, PDSI, and water levels at TS and SRS from January 2009 to December 2012: (a) shows
monthly cumulative precipitation (mm) and season intensity (as measured by seasonal average PDSI), (b) shows monthly average
Palmer Drought Severity Index (PDSI), water level (m), and El Niño and La Niña phase.
134
Figure 13. The U.S. Drought Monitor for the Everglades region 01/13/2009 to 06/16/2009 and 01/04/2011 to 09/20/2011. The
Drought Monitor is a weekly map of drought conditions that is produced jointly by the National Oceanic and Atmospheric
Administration, the U.S. Department of Agriculture, and the National Drought Mitigation Center (NDMC) at the University of
Nebraska-Lincoln. The map is based on measurements of climatic, hydrologic and soil conditions as well as reported impacts and
observations from more than 350 contributors.
135
Figure 14. (a) The relationship between cumulative precipitation (January through March) and wet season length at TS and SRS shows
that precipitation prior to the wet season is an important determinant of wet season length. The relationship between dry season length
and dry season NEE is positive at (b) TS and negative at (c) SRS.
136
Figure 15. Light response curves showing differences in photosynthetic capacity by ENSO phase during the wet season at (a) TS and
(b) SRS, and during the dry season at (c) TS and (d) SRS. Temperature response curves showing differences in the relationship
between ecosystem respiration rates and temperature by ENSO phase during the wet season at (e) TS and (f) SRS, and during the dry
season at (g) TS and (h) SRS. The shaded region highlights the wet season. At TS, there were 576 days (322 wet season; 254 dry
season) in a La Niña phase, 365 days (249 wet season; 116 dry season) in an El Niño phase, 326 days (85 wet season; 241 dry season)
in a post-La Niña phase, and 61 days (12 wet season; 49 dry season) in the post-El Niño phase. At SRS, there were 576 days (308 wet
season; 268 dry season) in a La Niña phase, 365 days (212wet season; 153 dry season) in an El Niño phase, 326 days (85 wet season;
241 dry season) in a post-La Niña phase, and 61 days (58 wet season; 3 dry season) in the post-El Niño phase.
137
References
Allan, R. P. R., & Soden, B. J. B. (2008). Atmospheric warming and the amplification of
precipitation extremes. Science, 321(5895), 1481–1484. doi:10.1126/science.1160787.
Aubinet, M., Grelle, A., Ibrom, A., Rannik, J., Moncrieff, J., Foken, T., et al. (2000). Estimates
of the annual net carbon and water exchange of forests: the EUROFLUX methodology.
In Advances in Ecological Research (pp. 113–175). doi:10.1016/s0065-2504(08)60018-5.
Barr, J. G., Engel, V., Fuentes, J. D., Zieman, J. C., O'Halloran, T. L., Smith, T. J., & Anderson,
G. H. (2010). Controls on mangrove forest‐atmosphere carbon dioxide exchanges in
western Everglades National Park. Journal of Geophysical Research: Atmospheres
(1984–2012), 115(G2), G02020. doi:10.1002/(ISSN)1525-2027.
Beckage, B., Platt, W. J., Slocum, M. G., & Panko, B. (2003). Influence of the El Nino Southern
Oscillation on fire regimes in the Florida Everglades. Ecology, 84(12), 3124–3130.
doi:10.1093/cvr/cvu015.
Behrenfeld, M. J. (2001). Biospheric Primary Production During an ENSO Transition. Science,
291(5513), 2594–2597. doi:10.1126/science.1055071.
Brocklebank, J. C., & Dickey, D. A. (2003). SAS System for Forecasting Time Series (1986 ed.).
SAS Institute Inc. Cary, North Carolina.
Chen, E., & Gerber, J. F. (1992). Climate. In R. L. Myers & J. J. Ewel (Eds.), Ecosystems of
Florida.
Childers, D. L. (2006). A synthesis of long-term research by the Florida Coastal Everglades
LTER Program. Hydrobiologia, 569(1), 531–544. doi:10.1007/s10750-006-0154-8.
Christensen, J. H., Hewitson, B., Busuioc, A., Chen, A., Gao, X., Held, R., et al. (2007).
Regional climate projections. Climate Change, 2007: The Physical Science Basis.
Contribution of Working group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change, University Press, Cambridge, Chapter 11
(pp. 847–940).
Clark, K. L., Gholz, H., Moncrieff, J. B., Cropley, F., & Loescher, H. W. (1999). Environmental
Controls over Net Exchanges of Carbon Dioxide from Contrasting Florida Ecosystems.
Ecological Applications, 9(3), 936–948.
Cleland, E. E., Chuine, I., Menzel, A., Mooney, H. A., & Schwartz, M. D. (2007). Shifting plant
phenology in response to global change. Trends in Ecology & Evolution, 22(7), 357–365.
doi:10.1016/j.tree.2007.04.003.
Clement, R. (1999). http://www.geos.ed.ac.uk/abs/research/micromet/EdiRe.
138
Cubasch, U., Meehl, G. A., Boer, G. J., & Stouffer, R. J. (2001). Projections of Future Climate
Change. In Climate Change 2001: The Scientific Basis. Contribution of Working Group I
to the Third Assessment Report of the Intergovernmental Panel on Climate Change (pp.
524–482). Cambridge University Press.
Davis, S. M., & Ogden, J. C. (Eds.). (1994). Everglades, the Ecosystem and its Restoration. St.
Lucie Press, Delray Beach, FL, USA.
Dickey, D. A., & Fuller, W. A. (1979). Distribution of the Estimators for Autoregressive Time
Series with a Unit Root. Journal of the American Statistical Association, 74(366).
Duever, M. J., Carlson, J. E., Riopelle, L. A., & Duever, L. C. (1978). Ecosystem analysis at
Corkscrew Swamp, in Cypress Wetlands for Water Management, Recycling, and
Conservation, 4th annual Report to the National Science Foundation and to the
Rockefeller Foundation, Central Wetlands (pp. 534–570). University of Florida.,
Gainesville.
Duever, M. J., Meeder, J. F., Meeder, L. C., & McCollom, J. M. (1994). The climate of south
Florida and its role in shaping the Everglades ecosystem. In S. M. Davis & J. C. Ogden
(Eds.), Everglades, the Ecosystem and its Restoration. St. Lucie Press, Delray Beach, FL,
USA.
Foken T, Leclerc MY (2004)Foken, T., & Leclerc, M. Y. (2004). Methods and limitations in
validation of footprint models. Agricultural and Forest Meteorology, 127(3-4), 223–234.
doi:10.1016/j.agrformet.2004.07.015. Agric. For. Meteorol. 127(3-4),223-234.
Foken, T., & Wichura, B. (1996). Tools for quality assessment of surface-based flux
measurements. Agricultural and Forest Meteorology, 78(1-2), 83–105. doi:10.1016/01681923(95)02248-1
Gorham, E. (1991). Northern Peatlands: Role in the Carbon Cycle and Probable Responses to
Climatic Warming. Ecological Applications, 1(2), 182–195.
Goulden, M. L., Munger, J. W., & FAN, S. M. (1996). Measurements of carbon sequestration by
long‐term eddy covariance: Methods and a critical evaluation of accuracy. Global
Climate Change Biology, 2, 169–182.
Hanson, K., & Maul, G. A. (1991). Florida Precipitation and the Pacific El Niño, 1895-1989
(Vol. 54, pp. 160–168). Florida Academy of Sciences.
Jimenez, K.L., Starr, G., Staudhammer, C.L., Schedlbauer, J.L., Loescher, H.W., Malone, S.L.,
& Oberbauer, S.F. (2012) Carbon dioxide exchange rates from short- and longhydroperiod Everglades freshwater marsh. Journal of Geophysical Research, 117(G4),
G04009, doi:10.1029/2012JG002117.
139
Kiladis, G. N., & Diaz, H. F. (1989). Global Climatic Anomalies Associated with Extremes in
the Southern Oscillation. Journal of Climate, 2(9), 1069–1090. doi:10.1175/15200442(1989)002<1069:GCAAWE>2.0.CO;2.
Kljun, N., Rotach, M. W., & Schmid, H. P. (2002). A Three-Dimensional Backward Lagrangian
Footprint Model For A Wide Range Of Boundary-Layer Stratifications. Boundary-Layer
Meteorology, 103(2), 205–226. doi:10.1023/A:1014556300021.
Kljun, N., Calanca, P., Rotach, M. W., & Schmid, H. P. (2004). A Simple Parameterisation for
Flux Footprint Predictions. Boundary-Layer Meteorology, 112(3), 503–523.
doi:10.1023/B:BOUN.0000030653.71031.96.
Latif, M. M., & Keenlyside, N. S. N. (2009). El Nino/Southern Oscillation response to global
warming. Pnas, 106(49), 20578–20583. doi:10.1073/pnas.0710860105.
Li, L., Li, W., & Kushnir, Y. (2011). Variation of the North Atlantic subtropical high western
ridge and its implication to Southeastern US summer precipitation. Climate Dynamics,
39(6), 1401–1412. doi:10.1007/s00382-011-1214-y.
Livezey, R. E., Masutani, M., Leetmaa, A., Rui, H., Ji, M., & Kumar, A. (1997). Teleconnective
response of the Pacific-North American region atmosphere to large central equatorial
Pacific SST anomalies. Journal of Climate, 10, 1787–1820. doi:10.1175/15200442(1997)010<1787:TROTPN>2.0.CO;2.
Loescher, H. W., Law, B. E., Mahrt, L., & Hollinger, D. Y. (2006). Uncertainties in, and
interpretation of, carbon flux estimates using the eddy covariance technique. Journal of
Geophysical Research, 111, D21S90. doi:10.1029/2005JD006932.
Loveless, C. M. (1959). A Study of the Vegetation in the Florida Everglades. Ecology, 40(1), 1–
9. doi:10.2307/1929916.
Malone, S.L., Starr, G., Staudhammer, C.L., Loescher, H.W., & Ryan, M.G. (In review)
Hydrological influences on energy partitioning of two freshwater marsh ecosystems in
the Florida Everglades.
Malone, S. L., Starr, G., & Staudhammer, C. L. (2013). Effects of simulated drought on the
carbon balance of Everglades short‐hydroperiod marsh. Global Change Biology,
doi:10.1111/gcb.12211.
Mason, S. J., & Goddard, L. (2001). Probabilistic precipitation anomalies associated with ENSO.
Bulletin of the American Meteorological Society, 82(4), 619–638. doi:10.1175/15200477(2001)082<0619:PPAAWE>2.3.CO;2.
140
Massman, W. J. (2004). Toward an ozone standard to protect vegetation based on effective dose:
a review of deposition resistances and a possible metric. Atmospheric Environment,
38(15), 2323–2337. doi:10.1016/j.atmosenv.2003.09.079
Meyers, S. D., & O'Brien, J. J. (1995). Pacific Ocean influences atmospheric carbon dioxide.
Eos, Transactions American Geophysical Union, 76(52), 533–533.
doi:10.1029/95EO00326.
Morellato, L.P.C. (2003). South America. In Phenology An Integrative Environmental Science
(pp. 75–92). Kluwer.
Moses, C. S., Anderson, W. T., Saunders, C., & Sklar, F. (2013). Regional climate gradients in
precipitation and temperature in response to climate teleconnections in the Greater
Everglades ecosystem of South Florida. Journal of Paleolimnology, 49(1), 5–14.
doi:10.1007/s10933-012-9635-0.
Nuttle, W. K. (1997). Measurement of wetland hydroperiod using harmonic analysis. Wetlands,
17(1), 82–89. doi:10.1007/BF03160720.
Obeysekera, J., Browder, J., Hornung, L., & Harwell, M. A. (1999). The natural South Florida
system I: Climate, geology, and hydrology. Urban Ecosystems, 3(3), 223–244.
doi:10.1023/A:1009552500448.
Palmer, W. (1965). Meteorological Drought (No. 45) (p. 58). U.S. Department of Commerce
Weather Bureau.
Pankratz, A. (1983). Forecasting With Univariate Box–Jenkins Models Concepts and Cases.
Wiley, New York.
Piechota, T. C., & Dracup, J. A. (1996). Drought and Regional Hydrologic Variation in the
United States: Associations with the El NioSouthern Oscillation. Water Resources
Research, 32(5), 1359–1373.
Randerson, J. T., Field, C. B., Fung, I. Y., & Tans, P. P. (1999). Increases in early season
ecosystem uptake explain recent changes in the seasonal cycle of atmospheric CO2 at
high northern latitudes. Geophysical Research Letters, 26(17), 2765–2768.
doi:10.1029/1999GL900500.
Reich, P. B. (1995). Phenology of tropical forests: patterns, causes, and consequences. Canadian
Journal of Botany, 73(2), 164–174. doi:10.1139/b95-020.
Rolim, S. G., Jesus, R. M., Nascimento, H. E. M., Couto, H. T. Z., & Chambers, J. Q. (2005).
Biomass change in an Atlantic tropical moist forest: the ENSO effect in permanent
sample plots over a 22-year period. Oecologia, 142(2), 238–246. doi:10.1007/s00442004-1717-x.
141
Ropelewski, C. F., & Halpert, M. S. (1986). North American precipitation and temperature
patterns associated with the El Niño/Southern Oscillation (ENSO). Monthly Weather
Review, 114(12), 2352–2362. doi:10.1175/15200493(1986)114%3C2352:NAPATP%3E2.0.CO;2.
Ropelewski, C. F., & Halpert, M. S. (1987). Global and regional scale precipitation patterns
associated with the El Nino/Southern Oscillation. Monthly WWeather Review, 115,
1606–1626. doi:10.1175/1520-0493(1987)115%3C1606:GARSPP%3E2.0.CO;2.
Sanchez-Azofeifa, A., Kalacska, M. E., Quesada, M., Stoner, K. E., Lobo, J. A., & Arroyo-Mora,
P. (2003). Tropical Dry Climates. In Phenology An Integrative Environmental Science
(Vol. 39, pp. 121–137). Dordrecht: Springer Netherlands. doi:10.1007/978-94-007-06323_9.
Schedlbauer, J. L., Oberbauer, S. F., Starr, G., & Jimenez, K. L. (2010). Seasonal Differences in
the CO2 Exchange of a Short-Hydroperiod Florida Everglades Marsh. Agricultural and
Forest Meteorology, 150(2010), 994–1006. doi:10.1016/j.agrformet.2010.03.005.
Schedlbauer, J. L., Munyon, J. W., Oberbauer, S. F., & Gaiser, E. E. (2012). Controls on
ecosystem carbon dioxide exchange in short- and long-hydroperiod Florida Everglades
freshwater marshes. Wetlands, 32, 801–812. doi:10.1007/s13157-012-0311-y.
Schmidt, N., Lipp, E. K., Rose, J. B., & Luther, M. E. (2001). ENSO influences on seasonal
rainfall and river discharge in Florida. Journal of Climate, 14(4), 615–628.
doi:10.1175/1520-0442(2001)014<0615:EIOSRA>2.0.CO;2.
Schotanus, P., Nieuwstadt, F. T. M., & De Bruin, H. A. R. (1983). Temperature measurement
with a sonic anemometer and its application to heat and moisture fluxes. Boundary-Layer
Meteorology, 26(1), 81–93. doi:10.1007/BF00164332.
Sittel, M. C. (1994). Marginal probabilities of the extremes of ENSO events for temperature and
precipitation in the southeastern Unit- ed States. Tech. Rep. - (p. 155). Center for Ocean–
Atmospheric Studies, The Florida State University, Tallahassee, FL.
Stanton, E. A., & Ackerman, F. (2007). Florida and climate change: the costs of inaction. Florida
and climate change: the costs of inaction. Global Development and Environment
Institute, Tufts University.
Svoboda, M. (2000). An Introduction to the Drought Monitor (No. 80). Drought Network News
(1994-2001). Retrieved from http://digitalcommons.unl.edu/droughtnetnews/80.
Timmermann, A., Oberhuber, J., Bacher, A., Esch, M., Latif, M., & Roeckner, E. (1999).
Increased El Ninƒ o frequency in a climate model forced by future greenhouse warming.
Nature, 398(6729), 694–697.
142
Trenberth, K. E. (1997). The Definition of El Niño. Bulletin of the American Meteorological
Society, 78(12), 2771–2777. doi:10.1175/15200477(1997)078<2771:TDOENO>2.0.CO;2.
Veldkamp, E., & O'Brien, J. J. (2000). Calibration of a Frequency Domain Reflectometry Sensor
for Humid Tropical Soils of Volcanic Origin. Soil Science Society of America Journal,
64(5), 1549–1553. doi:10.2136/sssaj2000.6451549x.
Webb, E. K., Pearman, G. I., & Leuning, R. (1980). Correction of flux measurements for density
effects due to heat and water vapour transfer. Quarterly Journal of the Royal
Meteorological Society, 106(447), 85–100.
Whiting, G. J., & Chanton, J. P. (2001). Greenhouse carbon balance of wetlands: methane
emission versus carbon sequestration. Tellus B, 53(5), 521–528.
143
CHAPTER 5: ECOSYSTEM SENSITIVITY TO CLIMATE CHANGE: A CASE STUDY
FROM THE FRESHWATER MARSHES OF THE FLORIDA EVERGLADES
Abstract
Shaped by the hydrology of the Kissimmee-Okeechobee-Everglades watershed, the
Florida Everglades is composed of a conglomerate of wetland ecosystems that have varying
capacities to sequester and store carbon. The community composition and productivity of the
greater Everglades ecosystem is driven by hydrology, which is ultimately a product of the
region’s precipitation and temperature patterns combined with the water management policy. As
shifts in both air temperature and precipitation are expected over the next 100 years as a
consequence of climate change, CO2 dynamics in the greater Everglades are expected to change.
I simulated fluxes of C among the atmosphere, vegetation, and soil using the DAYCENT model
with the goal to reduce uncertainties associated with climate change and to explore how
projected changes in atmospheric CO2 concentrations and climate can alter current CO2
exchange rates in Everglades freshwater marsh ecosystems. I explored the effects of low,
moderate, and high scenarios for atmospheric CO2 (550, 850, and 950 ppm), mean annual air
temperature (1, 2.5, and 4.2°C) and precipitation (-2, 7, and 14%), as predicted by IPCC for the
year 2100 for the region, on CO2 exchange rates in short- and long-hydroperiod wetland
ecosystems. Under 100-years of current climate and atmospheric CO2 concentrations (380 ppm),
Everglades freshwater marsh ecosystems were estimated to be CO2 -neutral. As atmospheric CO2
concentrations increased and under climate change projections, there were slight shifts in the
start and length of the wet season (-1 to +7 days) and a small enhancement in the sink capacity
144
(by -169 to -573 g C m-2 century-1) occurred at both short- and long- hydroperiod ecosystems
compared to CO2 dynamic under the current climate regime. Over 100 years, rising temperatures
increased net CO2 exchange rates (+1 to 13 g C m-2 century-1) and shifts in precipitation patterns
altered cumulative NEE by +13 to -46 g C m-2 century-1. While changes in ecosystem structure,
species composition, and disturbance regimes were beyond the scope of this research, these
results do indicate that climate change will produce very small changes in CO2 dynamics in
Everglades freshwater marsh ecosystems and suggest that the hydrologic regime and
oligotrophic conditions of Everglades freshwater marshes lowers the ecosystem sensitivity to
climate change.
145
Introduction
The Florida Everglades is composed of a conglomerate of wetland ecosystems that have
been shaped by the complex hydrology of the south Florida region (Davis & Ogden 1994).
Hydrology drives wetland community composition and structure, productivity and soil
development, and influences decomposition rates across these ecosystems (Davis & Ogden 1994;
Mitsch & Gosslink 2007). The mosaic of seasonally-fluctuating hydrologic patterns has led to a
range of capacities to sequester and store carbon (C), and may be important for ecosystem
sensitivity. We refer to ecosystem sensitivity as the magnitude of a response to change. Low
sensitivity indicates that a response to an environmental perturbation is small. Climate change
projections suggest shifts in both air temperature and precipitation over the next 50 to 100 years
(Christensen et al., 2007; Stanton & Ackerman 2007; IPCC 2013), and when coupled with water
management decisions and human population expansion in south Florida, these shifts may have
serious implications for Everglades hydrology and ultimately the region’s carbon dynamics
(Stanton & Ackerman 2007).
The Everglades is expected to be vulnerable to the impacts of climate change that, along
with anthropogenic controls, will initiate additional alterations in water levels and inundation
periods (Stanton & Ackerman 2007; IPCC 2013). Future climate is dependent on atmospheric
CO2 concentrations, which are predicted to increase in the range of 540 to 970 ppm by 2100
(Houghton et al., 2001; Stanton & Ackerman 2007). In combination, wet season precipitation is
projected to decrease by 5-10% (Christensen et al., 2007) while annual precipitation may be
altered by -2 to +14% (IPCC 2013) and temperatures may warm 1 to 4.2°C (IPCC 2013). The
change will likely include larger convective storms and greater intensity hurricanes (Allan &
Soden 2008). The uncertainty and confidence in climate change projections depends on the
146
quantity, quality, consistency of evidence (e.g., mechanistic understanding, theory, data, and
models) and the degree of agreement among models (IPCC 2013). Although climate change is
expected to alter vegetation communities and carbon dynamics within the Everglades (Davis &
Ogden 1994; Bush et al., 1998; Todd et al., 2010), the range in climate projections is wide, thus
making estimates of future carbon dynamics for the region even less certain.
Anthropogenic activities have been shown to reduce the capacity of ecosystems to cope
with disturbance and change (Jackson et al. 2001; Scheffer et al. 2001; Elmqvist et al., 2003),
which could have significant implications for ecosystem sensitivity to climate change and
ultimately the system’s C dynamics and sequestration capacity (Cao & Woodward 1
8; Riedo
et al., 2000. The greater Everglades system is highly modified by water control structures that
disconnect hydrological dynamics from precipitation patterns throughout the KissimmeeOkeechobee-Everglades watershed (Perry 2004). Efforts are currently underway to improve
water levels throughout Everglades National Park in areas suffering from chronically low water
levels (Perry 2004). Fluctuations in hydric conditions that alter ecosystem C storage or emission
rates might occur slowly, but can have a significant long lasting effect on C pools. Using the
ecosystem model DAYCENT (Del Grosso et al., 2001), we aim to explore the effects of
increasing atmospheric CO2 concentrations, temperature and altered precipitation, independently
and in combination, to determine the relative impact of each on ecosystem CO2 exchange rates.
Limited experimental capabilities exist to evaluate the complicated interactive controls on
ecosystem responses to multifactor drivers (Fuhrer 2003; Luo et al., 2008), although these effects
are critical to understand how climate change impacts terrestrial ecosystems (Luo et al., 2008).
Ultimately, time and financial constraints limit multifactor experiments (Luo et al., 2008), and
thus simulation models have been a useful tool to investigate the effects of rising atmospheric
147
CO2 concentrations and climate change scenarios on terrestrial ecosystems (Abdalla et al., 2010).
Ideally, models that simulate long-term changes in C dynamics should link plant, soil and
atmospheric processes, and account for interactions among effects (Riedo et al., 2000). The
DAYCENT model (Del Grosso et al., 2001) meets these requirements, simulating ecosystem
water, carbon and nutrient dynamics (Parton 1987, 1988) for various native and managed
systems (Del Grosso et al., 2002; Del Grosso et al., 2009; Abdalla et al., 2010).
DAYCENT simulates changes in carbon and nutrient dynamics within and through the
soil-plant-atmosphere continuum in response to changes in environmental conditions (i.e. air
temperature, precipitation and atmospheric CO2 concentrations) and management practices
(grazing, harvesting, burning, fertilizing, and irrigation; Del Grosso et al., 2000). DAYCENT has
been used to successfully simulate ecosystem responses to changes in climate (Parton et al.,
1995; Lou et al., 2008; Savage et al., 2013), and to model gas fluxes (CO2, CH4, N2O, NOx, N2).
It has also been used to model carbon and nutrient dynamics (N. P, S) in shrublands (Li et al.,
2006), forest (Hartman et al., 2007; Parton et al., 2010), crops (Del Grosso et al., 2002; Stehfest
et al., 2007; Del Grosso et al., 2009; Chang et al., 2013; Duval et al., 2013), and temperate
wetlands and savannas.
Alterations to the Everglades hydrologic cycle are expected with climate change, leaving
C pools and sequestration potential highly uncertain. The goal of these model simulations is to
determine how CO2 exchange rates in Everglades freshwater wetland ecosystems will change as
a result of rising atmospheric CO2 concentrations and associated climate projections. The model
projections are important for understanding Everglades vulnerability to climate change, and to
indicate how factors interact to influence the CO2 sequestering capacity of Everglades
ecosystems. Wetland ecosystem sensitivity in light of climate change will have a significant
148
impact on the global C cycle, considering that 535 Gt C are stored in wetland soils as peat
(Mitsch & Gosslink 2007) and this sequestered C is vulnerable to changes in hydrologic cycles.
Ecosystem sensitivity to “future” conditions can be used as an indication of the
vulnerability of Everglades C pools. Atmospheric CO2 concentrations should enhance ecosystem
productivity (Nowak et al., 2004; Ainsworth & Long 2005) by reducing nutrient limitations
(Hocking & Meyer 1991; Drake & Gonzàlez-Meler 1997; Ainsworth & Long 2005) and
photorespirations in C3 species (Bowes 1993). Holding atmospheric CO2 concentration constant,
wetland ecosystem productivity and CO2 exchange rates are driven largely by hydrology
(Schedlbauer et al., 2010), nutrient dynamics, light and temperature (Schedlbauer et al., 2010).
Both productivity and ecosystem respiration generally increase with temperature (Schedlbauer et
al., 2010). Precipitation drives hydrological patterns (e.g. water levels) and the onset and start of
seasons (Malone et al., In Preparation). Water levels above the soil surface influences soil
temperature, oxygen availability and respiratory processes, often leading to a decline in
ecosystem respiration. Productivity increases with higher water availability until water levels
interfere with exposed leaf area and oxygen availability in the soil. Low ecosystem sensitivity
would suggest that C pools are less vulnerable to climate change and the future structure and
function of ecosystems would be more heavily influenced by changes in disturbance regimes and
sea level rise.
Materials and Methods
Model description
To determine the effect of projected climate change on Everglades freshwater marshes I
used a modified version of the DAYCENT model (www.nrel.colostate.edu/projects/daycent/),
149
DAYCENT_Photosyn (Savage et al., 2013). DAYCENT_Photosyn contains the SIPNET
photosynthesis model (Braswell et al., 2005; Sacks et al., 2006, 2007), which is a simplified
Farquhar photosynthesis and respiration sub-model. In the DAYCENT net primary productivity
(NPP) sub-model carbon allocation is a function of plant phenology, water and nutrient stress
(Parton et al., 2010), soil carbon and nutrient dynamics (Parton et al., 2001), trace gas flux (Del
Grosso et al., 2000), and soil water and temperature (Parton et al., 1998; Eitzinger et al., 2000,
Figure 16). DAYCENT assumes that NPP and organic matter decomposition rates increase as
soil water content increases until optimum water content is reached, with the optimum higher for
NPP than for decomposition. Analogously, NPP and decomposition are influenced by
temperature, and the sensitivity of the temperature response is different for the processes.
Optimum, minimum, and maximum temperatures for NPP vary with vegetation type whereas no
minimum or maximum temperatures for decomposition are assumed, although the rate at which
decomposition increases declines as temperature increases. Both NPP and decomposition are
also limited by mineral N availability. DAYCENT also assumes that high CO2 increases wateruse efficiency by reducing transpiration, so that water savings are immediately reflected in soil
moisture. Increases in NPP under rising atmospheric CO2 concentrations using DAYCENT arise
from the combined direct effect of CO2 on NPP, indirect effects of soil moisture on NPP, and
reduced N-limitation to NPP due to enhanced soil- N mineralization on moister soils.
DAYCENT Parameterization
DAYCENT was parameterized for two subtropical Everglades ecosystems with
contrasting hydroperiods, Taylor Slough (TS) and Shark River Slough (SRS). The shorthydroperiod marsh, TS, is inundated 4 to 6 months each year (June to November) and is
150
characterized by shallow marl soils (~0.14 m; Duever et al., 1978) and relatively uniform
vegetation co-dominated by a C3 sedge (Cladium jamaicense Crantz) and a C4 grass
(Muhlenbergia capillaris Lam.). The long-hydroperiod marsh, SRS, is inundated ~12 months
each year and is characterized by peat soils (~1 m thick) with ridge and slough microtopography
(Duever et al., 1978). Ridges are dominated by C. jamaicense and sloughs are dominated by
Eliocharis cellulosa and Nymphae lily. Periphyton also exists on submerged structures at both
sites and as floating mats at SRS. Both sites are also P limited, have year-round growing seasons
and experience wet and dry seasons that are produced by precipitation patterns in the south
Florida region (Davis & Ogden 1994). Taylor Slough occupies ~14,398 ha while Shark River
Slough covers 88,811 ha of the 609,447 ha Everglades National Park. It is important to note that
freshwater marsh ecosystems do not dominate the entire area of TS and SRS. For a more detailed
site description, see Davis & Ogden (1994). Vegetation and soil data used in the model were
measured at the Florida Coastal Everglades Long Term Ecological Research (FCE-LTER) sites
TS-Ph1b (25°26’16.5” N, 80°35’40.68” W) and SRS-2 (25°33'6.72"N, 80°46'57.36"W) and
made available through the FCE-LTER data portal (http://fcelter.fiu.edu/data/; Table 17). Longterm weather data were obtained from the nearest weather station, NCDC Royal Palm Ranger
Station, where NOAA Daily Surface Meteorological Data were available from 1963 to 2011
(Figure 17).
To parameterize DAYCENT for TS and SRS, vegetation and water dynamics were
simplified. A dominant species at both sites, C. jamaicense, was used to characterize the
vegetation parameters used in the DAYCENT model. Emergent species dominate ecosystem
level fluxes at both sites (Starr & Oberbauer unpublished data), making this simplification
151
reasonable. Current site leaf area and P conditions were also used in the model and assumed
constant.
DAYCENT does not model water levels above the soil surface. To capture these effects
we used the program’s irrigation option, and indicated when and for how long sites were
inundated. The irrigation option keeps soils saturated, but does not include the effects of standing
water. We also adjusted optimum and maximum temperatures for production and altered the
coefficients used to calculate water stress on vegetation production. We matched historical
seasonal water levels at each site; at TS, the system was inundated throughout the wet season,
while SRS remained inundated during the entire year. The start and duration of the wet season
were determined from fluctuations in the Bowen Ratio () at TS and SRS (2009 to 2012; Malone
et al., In review). The length of the wet season was determined assuming a linear model between
cumulative precipitation (January through March) and wet season length derived from  (Figure
18a) at TS and SRS (separately) using 2009-2012 data. We also determined the linear
relationship between the wet season length and the first day of the wet season (Figure 18b).
These correlations were then used to incorporate the effect of surface flows on ecosystem
productivity and soil water availability. This method was effective in that it allowed seasons to
fluctuate with precipitation patterns, and permitted seasons to vary with climate change.
However it does not consider the effects of water depth or any changes in water levels that occur
as a result of water management activities, and assumes that relationships between precipitation
and season length and durations that occurred over a relatively short time frame (2009 - 2012)
were appropriate in the past and in the future under climate change.
This climate changes simulation will not define the future of freshwater marsh
ecosystems, but describes how the current system would respond to climate change projections
152
using simplified versions of Everglades short and long hydroperiod freshwater marsh
ecosystems. A 2000-year equilibrium simulation, under recent climate (380 ppm, and using longterm weather data) was conducted before model validations and climate projection simulations
so that simulations would start at quasi-equilibrium (Pepper et al. 2005). This approach allowed
us to attribute ecosystem responses wholly to climate change and thereby avoid any confounding
response from a non-equilibrium state.
DAYCENT validation
To examine the model’s ability to adequately characterize the study sites, I ran
DAYCENT, parameterized for each site, with weather data collected from TS-1 and SRS-2 in
2012 and compared modeled versus observed soil volumetric water content (VWC), soil
temperature (Tsoil, °C) and CO2 exchange rates (net ecosystem exchange: NEE, ecosystem
respiration: Reco, and gross ecosystem exchange: GEE). Observed soil volumetric water content
(VWC; %) is an integrated measurement from 0 to 15cm depth and average daily soil
temperature (Tsoil, °C) is measured at 5 cm depth. Observed CO2 exchange rates (NEE, Reco, and
GEE) were obtained from the eddy covariance tower sites at TS (TS-1; Schedlbauer et al., 2010;
Schedlbauer et al., 2011; Jimenez et al., 2012) and SRS (SRS-2; Jimenez et al., 2012), available
through AmeriFlux (http://ameriflux.ornl.gov). Using methods similar to Veldkamp & O’Brien
(2000), VWC was calculated from a site-specific equation for soil conditions and from the
dielectric constant using two soil moisture sensors (CS616, Campbell Scientific Inc.) buried
between 0 and 20 cm soil depth at each site. Soil temperature was measured at 5 cm depths at
two locations within each site using insulated thermocouples (Type-T, Omega Engineering Inc.,
Stamford, CT).
153
Measured and simulated outputs were evaluated using the coefficient of determination
(R2) and bias. Bias was quantified via a linear regression of simulated versus measured values.
Average bias is small when slopes are near 1 and intercepts are near 0, but a thorough
examination of modeled values must be made to appropriately evaluate patterns in over- and
under-estimation for the range of data. DAYCENT simulated daily Tsoil and VWC values did not
closely match that of the observed data, although monthly average DAYCENT values were
comparable to observed average values at TS and SRS (Figure 19). At both sites the average
difference between daily observed and modeled Tsoil and VWC was less than 0.97°C and 0.007,
respectively. Monthly average fluctuations in Tsoil were slightly underestimated at both TS (R2 =
0.98) and at SRS (R2= 0.99), likely the result of increasing water levels at each site. The soil
VWC was underestimated at TS once the site was inundated (~0.025% on average) and slightly
over-estimated at SRS in the wet season (~0.03% on average). Discrepancies in VWC at TS
were a result of differences between observed and modeled inundation. Determined by
precipitation patterns, simulated inundation occurred before the site was actually inundated at
TS, resulting in higher simulated VWC than observed (Figure 19c). Once the site actually
became inundated, model performance improved. Discrepancies in VWC and Tsoil suggest that
DAYCENT could be improved in wetland ecosystems by incorporating water depth above the
soil surface and the effect of water depth on soil water content and temperature.
DAYCENT weakly captured fluctuations in daily CO2 exchange rates, though monthly
estimates were very similar to those observed and previously reported for NEE, Reco and GEE at
TS and SRS for 2008 to 2009 (Jimenez et al., 2012). Atmospheric convention is used for CO2
exchange rates, where a positive value denotes a loss of C from the ecosystem. At TS,
DAYCENT captured fluctuations in NEE (R2= 0.80) and GEE (R2=0.94), though daily
154
fluctuations in Reco (R2=0.65) were often slightly over-estimated (Figure 20a). At SRS, NEE was
over estimated in the dry season and during transition periods as a result of small overestimations
in both Reco (R2= 0.68; Figure 20b) and GEE (R2= 0.70; Figure 20c). On a monthly basis,
DAYCENT simulated NEE, Reco and GEE (Figure 20) were realistic at both sites and
DAYCENT captured fluctuation in GEE and NEE much better than it did Reco.
Climate change simulations
Following parameterization and validation, we ran the model for 100 years (1) under
climate change projections, and (2) maintaining recent CO2 concentrations (380 ppm) and
observed air temperature and precipitation patterns (1963 to 2012; NOAA Daily Surface
Meteorological Data) in the Everglades region, for comparison. To examine the effects of rising
atmospheric CO2 concentrations and climate change projections we simulated climate change
projections by increasing atmospheric CO2 concentrations and temperature and altering
precipitation patterns individually and in combination for each of the two sites. The same
weather data used for the present-day simulation were altered by applying additive scalars to
adjust atmospheric CO2 concentrations and temperature and scalar multipliers to adjust
precipitation seasonally and annually to simulate climate change (Figure 21). Atmospheric CO2
concentrations, temperature, and precipitation were altered gradually while conserving daily
(high frequency) variability (Parton et al., 1995; McMurtrie et al., 2001). We explored a low,
moderate, and high scenario for atmospheric CO2 concentrations (550, 850, and 950 ppm,
respectively; EPA & IPCC 2007) and each climate driver: mean annual air temperature (+1, 2.5,
and 4.2°C; IPCC 2013) and precipitation (-2, +7, and +14%; Figure 21; IPCC 2013). We also
incorporated the seasonal shifts in temperature and precipitation that are expected for this region
155
(Christensen et al., 2007; Stanton & Ackerman 2007). Projected temperature change was applied
to the daily minimum temperatures in the dry season (winter months), while temperature
increases were applied to maximum temperatures in the wet season (Figure 21d and 21e;
Christensen et al., 2007). The annual distribution of precipitation was altered by reducing wet
season precipitation by 10% (Figure 21d- 21f). When applied together, atmospheric CO2
concentrations, temperature, and altered precipitation patterns approximate the IPCC 2100
climate change projections for the Everglades region. While climate projections also suggest
increased hurricane intensity (IPCC 2013) and more frequent heat waves (Stanton & Ackerman
2007), including these potentially important changes was beyond the scope of this study. We
examined the effect of elevated atmospheric CO2 concentrations and altered weather by
comparing cumulative CO2 exchange rates (NEE, Reco, and GEE) over 100 years (2001 to 2100)
to cumulative rates under current conditions.
Results
Under recent atmospheric CO2 concentrations and climate, NEE fluctuated around a null
balance with a range of ± 100 g C m-2 yr-1 at both sites. Reco ranged from ~200 to 500 g C m-2 yr1
at TS and SRS, whereas GEE ranged from -200 to -500 g C m-2 yr-1, though rates at TS were
higher on average. Cumulatively at 100 years, both TS (-30 g C m-2 century-1) and SRS (-5 g C
m-2 century-1) were very small sinks for CO2, and represent a potential accumulation of 0.33 tons
C ha-1 century-1 at TS and 0.05 tons C ha-1 century-1at SRS. These results suggest freshwater
marsh ecosystems are near neutral for CO2 loss versus gain over 100 years.
156
Main effects of climate change scenarios
Rising atmospheric concentrations of CO2 resulted in the greatest change in NEE, Reco,
and GEE at both TS and SRS (Table 18; Figure 22 and 23). At TS cumulative NEE declined
substantially with elevated CO2 concentrations compared to simulations under current climate
(380 ppm; Table 18; Figure 22a). Long-term average annual NEE was -2.0, -4.8 and -5.6 g C m-2
yr-1 in the low, intermediate and high scenarios, respectively, compared to -0.30 g C m-2 yr-1
under current climates at TS. Cumulative Reco at TS increased progressively by 1755.2, 7217.8
and 8245.8 g C m-2 century-1 under the low, moderate and high scenarios (Table 18; Figure 22d),
and similar to Reco, cumulative GEE at TS declined (higher CO2 uptake) relative to current
conditions by approximately equal and opposite amounts (Figure 22g).
At SRS elevated atmospheric concentrations also enhanced net CO2 uptake. At 550, 850
and 950 ppm, cumulative NEE decreased by 176.3, 502.1 and 573.3 g C m-2 century-1,
respectively, compared to cumulative NEE at 380 ppm over 100 years (Table 18; Figure 23a).
Under these scenarios, the long-term average annual NEE was -1.8, -5.0 and -5.7 g C m-2 yr-1 at
low, medium, and high, respectively. Similar to TS, at SRS cumulative Reco increased under the
low, medium, and high scenarios (Table 18; Figure 23d) while GEE decreased by nearly equal
and opposite amounts (Table 18; Figure 23g).
Changes in temperature alone resulted in very small changes in CO2 exchange rates of
Everglades ecosystems by causing slight shifts in the ratio of GEE: Reco. At TS, a 1, 2.5 and
4.2°C increase in mean annual temperature led to increased NEE (lower CO2 uptake) by 1.3, 3.7
and 6.6 g C m-2 century-1 (Table 18; Figure 22b). Cumulative Reco decreased progressively by
2.5, 6.7 and 11.7 g C m-2 century-1 with increased temperatures (Table 18; Figure 22e), while
GEE increased by 3.8 to 18.3 g C m-2 century-1 (Table 18; Figure 22g). At SRS, NEE decreased
157
by 3.1 to 13.0 g C m-2 century-1 as a mean annual temperature increased by 1, 2.5 and 4.2°C,
respectively (Table 18; Figure 23a). Cumulative Reco decreased by 17.2 to 73.9 g C m-2 century-1,
and GEE increased by 20.3 to 86.9 g C m-2 century-1 as a mean annual temperature increased at
SRS (Table 18; Figure 23e and 23h).
Changes in precipitation by -2, +7, and +14% altered the length of seasons by -1, 4, and 7
days on average, respectively. A 2% decrease in annual precipitation, which was weighted to
reduce wet season precipitation more than in the dry season, increased cumulative NEE by 13.0
g C m-2 century-1 at TS, while no change in cumulative NEE was found at SRS (Table 18). As
mean annual precipitation increased by 7 and 14%, net stored CO2-C increased by 20.3 and 41.8
g C m-2 century-1, respectively, at TS (Table 18; Figure 22c). A change in precipitation of -2, +7
and +14% at TS altered cumulative Reco by -1105.0, -220.1, and 432.9 g C m-2 century-1,
respectively (Table 18; Figure 22f). At TS, cumulative GEE was altered by -1105.0, -220.1, and
432.9 g C m-2 with a -2 and +7 and +14% change in mean annual precipitation (Table 18; Figure
22i). At SRS, changes in mean annual precipitation had a smaller impact on CO2 exchange rates
(Table 18; Figure 23c, 23f and 23i). A change in mean annual precipitation of -2%, 7% and 14%
shifted NEE by 0, -21.03, and -46.1 g C m-2 century-1, respectively (Table 18; Figure 23c). Reco at
SRS increased by 5.4, 147.6, and 423.3 g C m-2 century-1, and GEE shifted by 5.4, -115.8, and 469.4 g C m-2 century-1 with a -2, +7 and +14% increase in annual precipitation, respectively at
(Table 18; Figure 23f and 23i).
Interactive effects of climate change scenarios
Simultaneous changes in CO2 concentration, temperature, and precipitation modified
ecosystem CO2 exchange rates (Table 18; Figure 24). An increase in CO2 concentrations (550
158
ppm), combined with a 1°C rise in annual temperature and 2% decrease in annual precipitation
led to 156.8 g C m-2 century-1 increase in cumulative NEE (Figure 24a), 2092.4 g C m-2 century-1
increase in cumulative Reco and a 2249.2 g C m-2 century-1 decrease in GEE at TS (Table 18;
Figure 24c and 24d). At SRS, cumulative NEE decreased by 160.5 g C m-2 century-1 (Table 18;
Figure 24b), while cumulative Reco increased 2001.4 g C m-2 century-1 and GEE declined by
2162.0 g C m-2 century-1 (Table 18; Figure 24d and 24f). In the second scenario, atmospheric
concentrations rose to 850 ppm, mean annual temperatures increased by 2.5°C, and mean annual
precipitation increased by 7%. At TS, cumulative NEE, Reco, and GEE changed by -468.1,
7323.5, and -7323.5 g C m-2 century-1, respectively (Table 18; Figure 24a, 24c and 24e). At SRS,
cumulative NEE, Reco and GEE were enhanced by –509.7, 5666.8, and -6176.5 g C m-2 century-1,
respectively (Table 18; Figure 24d, 24d and 24f). The most extreme scenario explored included
an atmospheric concentration of 950 ppm, a 4.2°C increase in mean annual temperature, and a
14% increase in annual precipitation. At TS, this high scenario led to a -544.4, 9279.4, and 9279.4 g C m-2 century-1 change in cumulative NEE, GEE, and Reco respectively (Table 18;
Figure 24a, 24c and 24e). At SRS, this high scenario led to a -557.7, 6776.4, and -7334.0 g C m-2
century-1 change in cumulative NEE, GEE, and Reco respectively (Table 18; Figure 24d, 24d and
24f). Climate change projections led to a change in potential carbon sequestration of 1.7, 5.2 and
6.0 tons C ha-1 century-1 at TS and 1.8, 5.6 and 6.1 tons C ha-1 century-1 with the low,
intermediate and high scenarios, respectively.
Discussion
Climate change is considered a major threat to species survival and ecosystem integrity
(Hulme 2005; Erwin 2009). Occurring within the transition zone between aquatic and terrestrial
159
environments, wetlands ecosystems are considered to be among the most vulnerable ecosystems
to climate change (Burkett & Kusler 2000) as a result of its effects on wetland hydrology and
temperature (Ferrati et al. 2005; Erwin 2009). Everglades ecosystems have been thought to be
vulnerable to climate change, with large shifts in ecosystem structure and function projected to
occur by 2100 (Stanton &Ackerman 2007). While projections for temperature and precipitation
are within the natural range of conditions observed in the region, ecosystem CO2 exchange rates
were modified by higher winter minimum and summer temperatures, and by greater dry season
precipitation. At both sites, increasing atmospheric CO2 concentrations caused the greatest
changes in GEE and Reco exchange rates; however, because increases in these fluxes were of
similar magnitude but opposite in sign, the effect on net CO2 exchange was very small. These
results suggest that climate change will lessen seasonal differences in precipitation patterns by
reducing wet and increasing dry season precipitation and results suggest that Everglades
ecosystems are more resilient to change than previously thought. This low ecosystem sensitivity
is likely the result of hydrology and nutrient limitations.
Main and interactive effects of multifactor climate change
As an important limiting factor for the growth and productivity of many species (Vu et
al., 1997), terrestrial ecosystems are currently and have been responding to rising atmospheric
CO2 concentrations (Gifford 1980; Ciais et al., 1995; Keeling et al., 1995; Drake & GonzàlezMeler 1997). The biochemical basis of this response is well established (Farquhar et al., 1980)
and indicates that below 600 ppm atmospheric CO2 is generally limiting (Nowak et al., 2004),
and many ecosystems will respond to higher concentrations. This was the case for TS and SRS
and led to greater C uptake and release rates. Photosynthesis (Long & Drake 1992) and
160
transpiration (Heath 1948; Drake & Gonzàlez-Meler 1997) have long been known to respond to
changes in atmospheric CO2 concentrations. Elevated CO2 concentrations reduce
photorespiration rates (Bowes 1993) in C3 species and enhance light (Drake & Gonzàlez-Meler
1997; Ainsworth & Long 2005), nutrient (Hocking & Meyer 1991; Drake & Gonzàlez-Meler
1997) and water use efficiency (Drake & Gonzàlez-Meler 1997) in plants. Ecosystem respiration
rates also respond to rising atmospheric CO2 concentrations (Gonzales-Meler et al., 1996; Drake
& Gonzàlez-Meler 1997) through the direct inhibition of respiratory enzyme activity (e.g.
cytochrome c oxidase and succinate dehydrogenase; Drake & Gonzàlez-Meler 1997). Rising
atmospheric CO2 concentrations generally reduce dark respiration rates (Drake & GonzàlezMeler 1997) compared to GEE rates
A small increase in net CO2 uptake at both Everglades sites occurred in response to
elevated atmospheric CO2, suggesting that both TS and SRS will provide a negative feedback to
global warming and maintain current soil C pools. Compared to SRS, TS had greater CO2 uptake
rates, although SRS was most sensitive to changes in CO2 concentrations. Variations in site
sensitivity were due to differences in leaf area. At SRS, leaf area is greater than at TS, and
studies have shown that while rising CO2 concentrations alter productivity rates, leaf area often
remains unchanged (Drake & Gonzàlez-Meler 1997). Although both sites fluctuate between
being a very small sink, source, or near neutral annually (Jimenez et al., 2012; Malone et al., In
review), rising CO2 concentrations may cause an enhancement in the sink potential of Everglades
freshwater marsh ecosystems by improving resource use efficiencies, though the response to CO2
enrichment will be limited by low P levels.
Climate change is expected to have the strongest and most immediate effect on plant
phenology (Forrest & Miller-Rushing 2010) and physiology. Significant changes in physiology
161
have been observed in response to higher temperatures, affecting photosynthetic rates of C3
plants (Long 1991), which is a dominant mode of photosynthesis at both sites. Higher
temperatures lower the activation state of Rubisco (Kobza & Edwards 1987; Holaday et al.,
1992) and both the solubility and the specificity for CO2 relative to O2 (Jordan & Ogren 1984;
Brooks & Farquhar 1985; Long 1991). At the ecosystem level, variations of only a few degrees
centigrade are sufficient to affect gas fluxes (Hirano et al., 2009). Compared to CO2
concentrations, changes in temperature had a much smaller impact on CO2 exchange rates at
both TS and SRS and sensitivity to changes in temperature was similar at both sites. An increase
in temperature, up until optimal conditions, often leads to an increase in metabolic activity
(Medlyn et al., 2002; Lambers et al., 2008), although when temperatures rise beyond optimal
growth and activity ranges higher temperatures lead to reductions in productivity (Lambers et al.,
2008). Higher temperatures shifted the ratio of GEE to Reco and led to greater C release at TS and
SRS, though this effect was very small. In Everglades ecosystems, inundation may buffer the
ecosystem response to higher temperatures.
Precipitation affects both productivity and ecosystem respiration rates by influencing
water and oxygen availability, altering exposed leaf area through its effect on water depth, and
enhancing N inputs through wet N deposition. As water levels affect ecosystem CO2 exchange
rates differently at TS and SRS (Malone et al., In review), so do changes in precipitation
patterns. At TS a 2% decrease in precipitation led to higher net exchange rates via a smaller
decrease in ecosystem respiration relative to GEE. SRS was insensitive to the 2% reduction in
precipitation, and at both TS and SRS, precipitation increased cumulative GEE more than Reco,
leading to greater net carbon uptakes rates. In the precipitation scenarios, wet season
precipitation declined more than in the dry season, and as a result of greater sensitivity to
162
changes in water levels the short hydroperiod site, TS, was more sensitive to changes in
precipitation.
Climate change scenarios represent the range in climate that would result from changes in
greenhouse gas concentrations in the atmosphere. At both sites, NEE, Reco and GEE responded to
changes in atmospheric CO2 concentrations, temperature and precipitation, though the
magnitudes of change were small, less than a 20% increase in average annual CO2 exchange
rates. Interactive effects of atmospheric CO2 concentrations and climate change on ecosystem
CO2 exchange rates were greater than the sum of the individual effects and this enhancement was
likely due to the interaction between rising CO2 concentrations and air temperature. CO2
enrichment modifies the response to temperature (Drake & Gonzàlez-Meler 1997) in C3 species,
like the dominant species at both sites, sawgrass. Enrichment reduces photorespiration rates,
which moderates the adverse effects of high temperature on C3 photosynthesis and results in
greater net photosynthesis as growth temperatures increase (Long 1991; Vu et al., 1997). The
amount of Rubisco required also declines with increasing temperature (Drake & Gonzàlez-Meler
1997) and the degree of C3 photosynthesis enhancement by higher CO2 concentrations is
influenced by the temperature optimum for the species (Vu et al., 1997). Considerable evidence
supports the prediction that CO2 uptake will be greater in warm climates (Long 1991, McMurtrie
& Wang 1993), though this may not occur in systems with low sensitivity to climate change.
Ecosystem sensitivity
Ecosystem sensitivity may be an essential factor underlying the sustained carbon
sequestering capacities of Everglades ecosystems. Ecosystem sensitivity to climate change is
driven by water, nutrients, and the responses of C3 species relative to C4 species. At Both TS and
163
SRS, nutrient limitations should constrain the response to climate change but at TS where
hydroperiods are short, changes in water cycling and the response of C4 co-dominant species
could increase the ecosystems sensitivity to climate change. Both TS and SRS exhibit low
sensitivity to changes in rising CO2 concentrations and climate change scenarios. Previous
studies have indicated that wet systems might be more resilient to change than dry systems (Lou
et al., 2008). In the Everglades, hydroperiods may serve as a barrier to climate variation, which
links the importance of hydroperiod to ecosystem sensitivity. Water above the soil surface
dampens low temperature effects, cools the system when temperatures are high, and reduces the
temperature response of soil and macrophypte respiration by slowing gas exchange. In addition
to hydroperiods, low nutrient (P) levels may also affect ecosystem sensitivity to change (Steward
& Ornes 1975; Curtis & Wang 1998). The Everglades is an oligotrophic system (Craighead
1971), and nutrient limitations can reduce the capacity of ecosystems to respond to rising
atmospheric CO2 concentrations (Stitt & Krapp 1999). Interacting environmental stresses can
influence the response to elevated CO2 in plants (Idso & Idso 1994; Lloyd & Farquhar 1994;
Curtis 1996; Curtis & Wang 1998) and environmental stresses tend to reduce the CO2 response in
C3 species (Wand et al., 1999). Studies have shown that although vegetation may initially
respond to elevated CO2, acclimation is reported to be more pronounced when plants are N
limited (Wong 1979; Oberbauer et al., 1986; Bowes 1993; Curtis 1996; Stitt & Krapp 1999;
Ward & Strain 1999; Isopp et al., 2000). Studies have also shown that acclimation to elevated
CO2 rarely has reduced photosynthetic capacity enough to completely compensate for stimulated
photosynthetic rates (Drake & Gonzàlez-Meler 1997; Ward & Strain 1999). Although in these
simulations Everglades freshwater marsh ecosystems exhibit low sensitivity to changes in
atmospheric CO2 concentrations and climate change, shifts in disturbance regimes (e.g. fire,
164
tropical storms, drought) are also projected for the region and will significantly influence the
future condition of the Everglades.
Disturbance regimes, C3 vs C4 plants, and invasive species
To understand Everglades ecosystem vulnerability and resistance to change, future
research must consider changes in ecosystem structure, disturbance regimes, C3 versus C4 plants,
and invasive species risk in the Everglades region. Although not considered in this simulation
model, the magnitude and frequency of disturbance events may significantly alter Everglades
ecosystem structure and function (Stanton & Ackerman 2007). A rise in hurricane intensity,
more frequent heat waves, salt-water intrusion and sea-level rise are all projected for the
Everglades region (Stanton & Ackerman 2007). Disturbance and altered hydrologic regimes
promote the displacement of native vegetation by introduced or formerly restricted species
(Groves & Burdon 1986; Mooney & Drake 1986), and have already led to changes in vegetation
community composition in the Everglades (Toth 1987, 1988; Herndon et al., 1991; Urban et al.,
1993). Sawgrass communities in the northern Everglades have already been replaced by dense
stands of cattail (Typha domingensis Pers.; Richardson & Marshall 1990; Rutchey & Vilchek
1994; Jensen et al., 1995; Newman et al., 1998), which invade disturbed and nutrient rich
environments (Dykyjova & Kvet 1978; Grace & Harrison 1986; Keddy 1990). Marl prairie
landscapes comprising the outer regions of Everglades National Park (ENP) have also shown
vulnerability to woody plant expansion (Jenkins et al., 2003; Knickerbocker et al., 2009) due to
alterations in natural disturbance mechanisms (flood and fire management; Hanan et al., 2010).
Since trees have the greatest response to elevated CO2 (Ainsworth & Long 2005), woody
encroachment in TS may increase.
165
Soil disturbance has also caused encroachment by exotics (Dalrymple et al., 1993) in the
Everglades. Bahia grass (Paspalum notatum Flüggé) and torpedo grass (Panicum repens L.),
exotic species introduced for cattle forage, spread quickly and aggressively. Under current
conditions, these C4 grasses are more metabolically efficient than C3 species, allowing them to
encroach on sawgrass marshes. Invasion may increase in the future if these species are sensitive
to elevated atmospheric CO2 concentrations. It has been suggested that elevated CO2 may
preferentially increase the abundance of invasive species (Dukes & Mooney 1999; Weltzin et al.,
2003), which may have already played a stimulatory role in plant invasions (Ziska 2003) and
shifts in dominance. Although some studies have shown C4 species respond weakly to elevated
CO2 (Ainsworth & Long 2005), others have found many C4 plants to exhibit enhanced
photosynthetic and growth responses (Sionit & Patterson 1984; Ziska et al., 1990; Imai &
Okamoto-Sato 1991; Wand et al., 1999). This suggests that C4 photosynthesis is not necessarily
saturated at current CO2 levels (Sionit & Patterson 1984; Imai & Okamoto-Sato 1991). C4
grasses have shown significant changes in gas exchange, leaf area development (Wand et al.,
1999), and an average growth enhancement of ~22% (Poorter 1993) following elevated CO2
concentrations. Previous research has also shown that CO2 saturation levels in C4 species may be
altered by environmental conditions (Wand et al., 1999). Ziska et al. (1990) found that elevated
CO2 concentrations stimulated CO2 assimilation rates in a cordgrass (Spartina patens Aiton)
dominated salt marsh, though the effect was seasonal. Although enhanced carbon assimilation
rates in C3 species has been shown to decline when stressed (Wand et al., 1999), C4 species show
less negative impacts of environmental stresses (Wand et al., 1999).
166
Model limitations
Shifts in disturbance regimes and species dynamics are very important factors that will
impact the sensitivity of Everglades ecosystems. Although the DAYCENT model was
appropriate for exploring the multifactor effects of rising atmospheric CO2 concentrations and
climate change on the current ecosystems CO2 exchange rates, to better understand the future
condition of Everglades freshwater marsh ecosystems the effect of water levels above the soil
surface and changes in species cover need to be incorporated in to the model. In addition,
modeling P dynamics would be a significant improvement in our confidence in model results in
these P limited systems. Modeling water table dynamics (Dimitrov et al., 2014) will be a very
important component to determining the effects of climate change on Everglades’s ecosystems.
Climate induced alterations to disturbance regimes (i.e., salt water intrusion, sea level rise and
the frequency of hurricanes and drought) and the incorporation of data uncertainties into
DAYCENT for both data and the model output must also be considered. Further model
validation once the effects of hydrology are better incorporated into the DAYCENT model, and
comparisons with results from simulations with other ecosystem models (Sitch et al., 2003;
Gerten et al., 2004, Luo & Reynolds, 1999, Krinner et al., 2005) will be necessary before
drawing firm conclusions related to the question of how hydroperiods aid ecosystem resilience
and what thresholds for air temperature and precipitation patterns significantly alter ecosystem
sensitivity in the Florida Everglades.
167
Table 17. DAYCENT site characteristics for Taylor Slough (TS) and Shark River Slough (SRS). Site data was obtained from the
Florida coastal Everglades Long-term Ecological Research (FCE LTER sites TS-1 and SRS-2), AmeriFlux and the literature.
Parameters (Units)
Site latitude
Site Longitude
C:N ratio (AG/ BG)
TS
5° 6’16.5” N
80°35’40.68” W
Childers & Troxler 2013/ Childers & Troxler 2011/
Reddy & Connor 1999
60:40
Schedlbauer et al. 2010; Miao et al. 1998
Lignin content (AG/ BG; %)
13% AG
Percent sand, silt, and clay (%)
Bulk density (g/cm3)
Rooting depth (cm)
80-15-5
0.56
15
C in SOM mg C g-1 soil
N in SOM mg N g-1 soil
Source
http://ameriflux.ornl.gov
http://ameriflux.ornl.gov
39.2 AG/46.8 BG
Root: shoot ratio
N deposition (wet/dry; mg l-1 N)
SRS
25°33'6.72"N
80°46'57.36"W
Vegetation
34% soil
Soil
80-15-5
0.066
30
Reddy & Connor 1999
Sandy loam (Dade County Soil Survey 1996)
White and Reddy 2001
0.73 (wet)/0.07 to 0.25 (dry)
166
5.81
409
25.3
Steward 1975
Childers 2006 (TS)/ White and Reddy 2001 (SRS)
Childers 2006 (TS)/ White and Reddy 2001 (SRS)
168
Table 18. The change in cumulative NEE, Reco, and GEE (g C m-2 century-1) as a result of rising atmospheric CO2 concentrations,
increasing mean annual temperature, and variations in annual precipitation individually and in combination (IPCC 2013 projects).
Scenario
CO2 only
Temperature only
Precipitation only
IPCC
Low (550 ppm)
Moderate (850 ppm)
High (950 ppm)
Low (1°C)
Moderate (2.5°C)
High (4.5°C)
Low (-2%)
Moderate (7%)
High (14%)
Low
Moderate
High
 NEE
-169.4
-453.8
-530.2
1.3
3.7
6.6
13.0
-20.3
-41.8
-156.8
-468.1
-544.4
TS
 Reco
1755.2
7217.8
8645.8
-2.5
-6.7
-11.7
-1105.0
-220.1
432.9
2092.4
7323.5
9279.4
 GEE
-1924.6
-7671.6
-9175.9
3.8
10.4
18.3
-1105.0
-220.1
432.9
-2092.4
-7323.5
-9279.4
169
 NEE
-176.3
-502.1
-573.3
3.1
7.8
13.0
0.0
-21.0
-46.1
-160.5
-509.7
-557.7
SRS
 Reco
2270.5
5484.6
6464.1
-17.3
-43.5
-73.9
5.4
147.6
423.3
2001.4
5666.8
6776.4
 GEE
-2446.9
-5986.7
-7037.4
20.4
51.3
86.9
5.4
-115.8
-469.4
-2162.0
-6176.5
-7334.0
Figure 16. Conceptual diagram of the DAYCENT ecosystem model from Del Grosso et al.
(2008). DAYCENT incorporates a soil organic matter (SOM), net primary productivity (NPP)
and land surface sub-model. In the SOM sub-model, decomposition is microbially mediated with
an associated microbial respiration CO2 loss that depends on soil texture. The NPP sub-model
uses relationships between climatic factors and available soil nutrients to calculate plant
production and allocation of nutrients to live aboveground and belowground vegetation. The land
surface sub-model simulates water flow through the plant canopy, litter, and soil profile (Parton
et al., 1998; Eitzinger et al., 2000).
170
Figure 17. Long-term daily weather data from the NCDC Royal Palm Ranger Station from 1963 to 2012.
171
Figure 18. Wet season length and the start of the wet season at TS and SRS were estimated with data from 2009 to 2012. First, the
linear relationship between (a) wet season length and precipitation (mm m-2) January to March was estimated, and then the
relationship between (b) wet season length and the start of the wet season was determined.
172
Figure 19. Observed versus modeled soil temperature at (a) TS and (b) SRS, and soil volumetric water content (VWC) at (c) TS and
(d) SRS.
173
Figure 20. Observed (solid) versus modeled (hollow) CO2 exchange rates (NEE, Reco and GEE) at TS (a) and SRS (b). Atmospheric
convention is used here and positive numbers indicate a loss of C to the atmosphere.
174
Figure 21. Climate change scalars for (a) elevated atmospheric CO2 concentrations, (b) air temperature, and (c) precipitation. The
distributions for (d) minimum temperature, (e) maximum temperature and (f) precipitation were modified to match projected seasonal
change.
175
Figure 22. The effect of elevated atmospheric CO2 concentrations (550 ppm, 850 ppm and 950 ppm) on cumulative (a) NEE, (d) Reco
and (g) GEE, at TS. The influence of rising temperatures (1°C, 2.7°C and 4.2°C) on cumulative (b) NEE, (e) Reco and (h) GEE and
shifts in seasonal and annual precipitation patterns (-2%, +7% and +14%) on cumulative (c) NEE, (f) Reco and (i) GEE. Atmospheric
convention is used here and positive numbers indicate a loss of C to the atmosphere. All simulations were compared to current
weather and atmospheric concentrations (red line).
176
Figure 23. The effect of elevated atmospheric CO2 concentrations (550 ppm, 850 ppm and 950 ppm) on cumulative (a) NEE, (d) Reco
and (g) GEE, at SRS. The influence of rising temperatures (1°C, 2.7°C and 4.2°C) on cumulative (b) NEE, (e) Reco and (h) GEE and
shifts in seasonal and annual precipitation patterns (-2%, +7% and +14%) on cumulative (c) NEE, (f) Reco and (i) GEE. Atmospheric
convention is used here and positive numbers indicate a loss of C to the atmosphere. All simulations were compared to current
weather and atmospheric concentrations (red line).
177
Figure 24. The effect of climate change projections on cumulative CO2 exchange rates: (a) NEE,
(c) Reco and (e) GEE at TS, and on (b) NEE, (d) Reco and (f) GEE on at SRS. Atmospheric
convention is used here and positive numbers indicate a loss of C to the atmosphere.
178
References
Abdalla, M., Jones, M., Yeluripati, J., Smith, P., Burke, J., & Williams, M. (2010). Testing
DayCent and DNDC model simulations of N2O fluxes and assessing the impacts of
climate change on the gas flux and biomass production from a humid pasture.
Atmospheric Environment, 44(25), 2961–2970. doi:10.1016/j.atmosenv.2010.05.018.
Ainsworth, E. A., & Long, S. P. (2005). What have we learned from 15 years of free-air CO2
enrichment (FACE)? A meta-analytic review of the responses of photosynthesis, canopy
properties and plant production to rising CO2. New Phytologist, 165(2), 351–372.
doi:10.1111/j.1469-8137.2004.01224.x.
Allan, R. P. R., & Soden, B. J. B. (2008). Atmospheric warming and the amplification of
precipitation extremes. Science, 321(5895), 1481–1484. doi:10.1126/science.1160787.
Alongi, D. M. (2008). Mangrove forests: Resilience, protection from tsunamis, and responses to
global climate change. Estuarine, Coastal and Shelf Science, 76(1), 1–13.
doi:10.1016/j.ecss.2007.08.024.
Boesch, D. F. (1974). Diversity, stability and response to human disturbance in estuarine
ecosystems. In Proceedings of the First International Congress of Ecology (pp. 109–114).
Wageningen, The Netherlands.
Bowes, G. (1993). Facing the Inevitable: Plants and Increasing Atmospheric CO2. Annual
Review of Plant Biology, 44, 309–332. doi:10.1146/annurev.pp.44.060193.001521.
Braswell, B. H., SACKS, W. J., Linder, E., & Schimel, D. S. (2005). Estimating diurnal to
annual ecosystem parameters by synthesis of a carbon flux model with eddy covariance
net ecosystem exchange observations. Global Change Biology, 11(2), 335–355.
doi:10.1111/j.1365-2486.2005.00897.x.
Brooks, A., & Farquhar, G. D. (1985). Effect of temperature on the CO2/O 2 specificity of
ribulose-1,5-bisphosphate carboxylase/oxygenase and the rate of respiration in the light :
Estimates from gas-exchange measurements on spinach. Planta, 165(3), 397–406.
doi:10.1007/BF00392238.
Burkett, V., & Kusler, J. (2000). Climate Change Potential Impacts and Interactions in Wetlands
of the United States. Journal of the American Water Ressources Association, 36(2), 313–
320.
Bush, D. E., Loftus, W. F., & Bass, W. F. (1998). Long-Term Hydrologic Effects on Marsh Plant
Community Structure in the Southern Everglades, 18(2), 230–241.
Cao, M., & Woodward, F. I. (1998). Dynamic responses of terrestrial ecosystem carbon cycling
to global climate change. Nature, 393(6682), 249–252. doi:10.1038/30460.
179
Chang, K.-H., Warland, J., Voroney, P., Bartlett, P., & Wagner-Riddle, C. (2013). Using
DayCENT to Simulate Carbon Dynamics in Conventional and No-Till Agriculture. Soil
Science Society of America Journal, 77(3), 941–950. doi:10.2136/sssaj2012.0354.
Childers, D. & Troxler, T. (2011-08-23). Sawgrass above ground biomass from the Taylor
Slough, Everglades National Park (FCE), South Florida from August 1999 to Present.
Long Term Ecological Research Network.
http://dx.doi.org/10.6073/pasta/80f2bb1ab8c4571ba6844b00ffac5eae. Data sets were
provided by the Florida Coastal Everglades Long-Term Ecological Research (LTER)
Program. This material is based upon work supported by the National Science
Foundation under Grant No. DEB-1237517, DBI-0620409, and Grant No. DEB-9910514.
Childers, D. & Troxler, T. (2008-03-19). Sawgrass above ground biomass from the Taylor
Slough, just outside Everglades National Park (FCE), South Florida from October 1997
to December 2006. Long Term Ecological Research Network.
http://dx.doi.org/10.6073/pasta/9288a2ace33120f25b749ecc8ce2c776. Data sets were
provided by the Florida Coastal Everglades Long-Term Ecological Research (LTER)
Program. This material is based upon work supported by the National Science
Foundation under Grant No. DEB-1237517, DBI-0620409, and Grant No. DEB-9910514.
Childers, D. & Troxler, T. (2011-04-26). Sawgrass above ground biomass from the Shark River
Slough, Everglades National Park (FCE), South Florida from November 2000 to Present.
Long Term Ecological Research Network.
http://dx.doi.org/10.6073/pasta/5824e4e14b46de012d44d4b0998c0149. Data sets were
provided by the Florida Coastal Everglades Long-Term Ecological Research (LTER)
Program. This material is based upon work supported by the National Science
Foundation under Grant No. DEB-1237517, DBI-0620409, and Grant No. DEB-9910514.
Childers, D. & Troxler, T. (2011-09-21). Sawgrass Above and Below Ground Total Nitrogen and
Total Carbon from the Taylor Slough, Everglades National Park (FCE), South Florida for
March 2002 to Present. Long Term Ecological Research Network.
http://dx.doi.org/10.6073/pasta/6381ab64adcf4f68840838950870c194. Data sets were
provided by the Florida Coastal Everglades Long-Term Ecological Research (LTER)
Program. This material is based upon work supported by the National Science
Foundation under Grant No. DEB-1237517, DBI-0620409, and Grant No. DEB-9910514.
Childers, D. & Troxler, T. (2013-10-17). Sawgrass Above and Below Ground Total Phosphorus
from the Shark River Slough, Everglades National Park (FCE), from September 2002 to
Present. Long Term Ecological Research Network.
http://dx.doi.org/10.6073/pasta/656a5803beeb61f9a513735207e3397e. Data sets were
provided by the Florida Coastal Everglades Long-Term Ecological Research (LTER)
Program. This material is based upon work supported by the National Science
Foundation under Grant No. DEB-1237517, DBI-0620409, and Grant No. DEB-9910514.
180
Childers, D. & Troxler, T. (2011-09-21). Sawgrass Above and Below Ground Total Nitrogen and
Total Carbon from the Taylor Slough, Everglades National Park (FCE), South Florida for
March 2002 to Present. Long Term Ecological Research Network.
http://dx.doi.org/10.6073/pasta/6381ab64adcf4f68840838950870c194. Data sets were
provided by the Florida Coastal Everglades Long-Term Ecological Research (LTER)
Program. This material is based upon work supported by the National Science
Foundation under Grant No. DEB-1237517, DBI-0620409, and Grant No. DEB-9910514.
Childers, D. & Troxler, T. (2013-10-17). Sawgrass Above and Below Ground Total Nitrogen and
Total Carbon from the Shark River Slough, Everglades National Park (FCE), from
September 2002 to Present. Long Term Ecological Research Network.
http://dx.doi.org/10.6073/pasta/cfc20b35e784313f86b39a6e92d78c46. Data sets were
provided by the Florida Coastal Everglades Long-Term Ecological Research (LTER)
Program. This material is based upon work supported by the National Science
Foundation under Grant No. DEB-1237517, DBI-0620409, and Grant No. DEB-9910514.
Christensen, J.H., Hewitson, B,. Busuioc, A., Chen, A., Gao, X., Held, I., Jones, R., Kolli, R.K.,
Kwon, W.T., Laprise, R., Magaña Rueda, V., Mearns, L., Menéndez, C.G., Räisänen, J.,
Rinke, A., Sarr, A., & Whetton, P. (2007) Regional Climate Projections. In: Climate
Change 2007: The Physical Science Basis. Contribution of Working Group I to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon,
S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M. & Miller,
H.L. (eds). Cambridge University Press. Cambridge, United Kingdom and New York,
NY, USA. pp. 847-940.
Ciais, P., Tans, P., Trolier, M., White, J., & Francey, R. (1995). A large northern hemisphere
terrestrial CO2 sink indicated by the C13/C12 ratio of atmospheric CO2. Science, 269,
1098–1102.
Craighead, F. C. (1971). The Trees of South Florida. Vol. I. The Natural Environments and Their
Succession. Coral Gables, FL. : University of Miami Press, Coral Gables, FL.
Curtis, P. S., & Wang, X. (1998). A meta-analysis of elevated CO2 effects on woody plant mass,
form, and physiology. Oecologia, 113(3), 299–313. doi:10.1007/s004420050381.
Curtis, P. S. (1996). A Meta-Analysis of Leaf Gas Exchange and Nitrogen in Trees Grown
Under Elevated CO2 in situ. Plant, Cell & Environment, 19, 127–137.
Dalrymple, N. K., Dalrymple, G. H., & Fanning, K. A. (1993). Vegetation of Restored Rock‐
Plowed Wetlands of the East Everglades. Restoration Ecology, 1(4), 220–225.
doi:10.1111/j.1526-100X.1993.tb00030.x.
Davis, S. M., & Ogden, J. C. (Eds.). (1994). Everglades: The Ecosystem and Its Restoration. St.
Lucie Press, Boca Raton , FL.
181
Del Grosso, S. J., Parton, W. J., Mosier, A. R., Ojima, D. S., Kulmala, A. E., & Phongpan, S.
(2000). General model for N2O and N2 gas emissions from soils due to dentrification.
Global Biogeochemical Cycles, 14(4), 1045–1060.
Del Grosso, S. J., Parton, W. J., Mosier, A. R., Hartman, M. D., Brenner, L., Ojima, D. S., &
Schimel, D. S. (2001). Simulated Interaction of Carbon Dynamics and Nitrogen Trace
Gas Fluxes Using the DAYCENT Model. In M. E. A. Schaffer (Ed.), Modeling Carbon
and Nitrogen Dynamics for Soil Management (pp. 303–332).
Del Grosso, S., Ojima, D., Parton, W., Mosier, A., Peterson, G., & Schimel, D. (2002).
Simulated effects of dryland cropping intensification on soil organic matter and
greenhouse gas exchanges using the DAYCENT ecosystem model. Environmental
Pollution, 116 Suppl 1, S75–S83.
Del Grosso, S., Parton, W., Stohlgren, T., Zheng, D., Bachelet, D., Prince, S., et al. (2008).
Global potential net primary production predicted from vegetation class, precipitation,
and temperature. Ecology, 89(8), 2117–2126.
Del Grosso, S. J., Ojima, D. S., Parton, W. J., & Stehfest, E. (2009). Global scale DAYCENT
model analysis of greenhouse gas emissions and mitigation strategies for cropped soils.
Global and Planetary Change, 67(1-2), 44–50. doi:10.1016/j.gloplacha.2008.12.006.
Dimitrov, D. D., Bhatti, J. S., & Grant, R. F. (2014). Ecological Modelling. Ecological
Modelling, 274, 57–70. doi:10.1016/j.ecolmodel.2013.11.030.
Drake, B. G., & Gonzàlez-Meler, M. A. (1997). More efficient plants: a consequence of rising
atmospheric CO2? Annual Review of Plant Physiology and Plant Molecular Biology,
48(1), 609–639. doi:10.1146/annurev.arplant.48.1.609.
Duever, M.J., Carlson, J.E., Riopelle, L.A., & Duever, L.C. (1978) Ecosystem analysis at
Corkscrew Swamp, in Cypress Wetlands for Water Management, Recycling, and
Conservation, 4th annual Report to the National Science Foundation and to the
Rockefeller Foundation, Central Wetlands, pp. 534–570, University of Florida.,
Gainesville.
Dukes, J. S., & Mooney, H. A. (1999). Does global change increase the success of biological
invaders? Trends in Ecology & Evolution, 14(4), 135–139. doi:10.1016/S01695347(98)01554-7.
Duval, B. D., Anderson-Teixeira, K. J., Davis, S. C., Keogh, C., Long, S. P., Parton, W. J., &
DeLucia, E. H. (2013). Predicting Greenhouse Gas Emissions and Soil Carbon from
Changing Pasture to an Energy Crop. PLoS ONE, 8(8), e72019.
doi:10.1371/journal.pone.0072019.
Dyky ov , D., v t, J., 1 33. (1 78). Pond littoral ecosystems. New York: Springer-Verlag.
182
Eitzinger, J., Parton, W. J., & Hartman, M. (2000). Improvement and Validation of A Daily Soil
Temperature Submodel for Freezing/Thawing Periods. Soil Science, 165(7), 525.
Elmqvist, T., Folke, C., & Nyström, M. (2003). Response diversity, ecosystem change, and
resilience. Frontiers in Ecology, 1(9), 488–494. doi:10.1890/15409295(2003)001[0488:RDECAR]2.0.CO;2.
Erwin, K. L. (2009). Wetlands and global climate change: the role of wetland restoration in a
changing world. Wetlands Ecology and Management, 17(1), 71–84. doi:10.1007/s11273008-9119-1.
Farquhar, G. D., Caemmerer, von, S., & Berry, J. A. (1980). A biochemical model of
photosynthetic CO2 assimilation in leaves of C3 species. Planta, 149(1), 78–90.
doi:10.1007/BF00386231.
Ferrati, R., Canziani, G. A., & Moreno, D. R. (2005). Esteros del Ibera: hydrometeorological and
hydrological characterization. Ecological Modelling, 186, 3–15.
doi:10.1016/j.ecolmodel.2005.01.021.
Forrest, J., & Miller-Rushing, A. J. (2010). Toward a synthetic understanding of the role of
phenology in ecology and evolution. Philosophical Transactions of the Royal Society B:
Biological Sciences, 365(1555), 3101–3112. doi:10.1098/rstb.2010.0145.
Fuhrer, J. (2003). Agroecosystem responses to combinations of elevated CO2, ozone, and global
climate change. “Agriculture, Ecosystems and Environment,” 7(1-3), 1–20.
doi:10.1016/S0167-8809(03)00125-7.
Gerten, D., Schaphoff, S., Haberlandt, U., Lucht, W., & Sitch, S. (2004). Terrestrial vegetation
and water balance—hydrological evaluation of a dynamic global vegetation model.
Journal of Hydrology, 286(1-4), 249–270. doi:10.1016/j.jhydrol.2003.09.029.
Gifford, R. M. (n.d.). Carbon storage by the biosphere. In Carbon Dioxide and Climate:
Australian Research (pp. 167–181).
Gonzàlez-Meler, M. A., Ribas-CarbÓ, M., Siedow, J. N., & Drake, B. G. (1996). Direct
Inhibition of Plant Mitochondrial Respiration by Elevated CO2. Plant Physiology,
112(3), 1349–1355. doi:10.1104/pp.112.3.1349.
Grace, J. B., & Harrison, J. S. (1986). Typha latifolia L., Typha angustifolia L., and Typha
glauca Godr. Canada Journal of Plant Science, 66(2), 361–379. doi:10.4141/cjps86051doi:10.4141/cjps86-051.
Groves, R. H., & Burdon, J. J. (Eds.). (1986). Ecology of Biological Invasions. Cambridge Univ.
Press, Cambridge, MA.
183
Hanan, E. J., & Ross, M. S. (2010). Across-scale patterning of plant-soil–water interactions
surrounding tree islands in Southern Everglades landscapes. Landscape Ecology, 25(3),
463–476. doi:10.1007/s10980-009-9426-9.
Hartman, M. D., Baron, J. S., & Ojima, D. S. (2007). Application of a coupled ecosystemchemical equilibrium model, DayCent-Chem, to stream and soil chemistry in a Rocky
Mountain watershed. Ecological Modelling, 200, 493–510.
doi:10.1016/j.ecolmodel.2006.09.001.
Heath, O. V. S. (1948). Control of Stomatal Movement by a Reduction in the Normal Carbon
Dioxide Content of the Air. Nature, 161(4083), 179–181. doi:10.1038/161179a0.
Herndon, A., Gunderson, L., & Stenberg, J. (1991). Sawgrass (cladium Jamaicense) Survival in a
Regime of Fire and Flooding. Wetlands, 11(1), 12–28. doi:10.1007/BF03160838.
Hirano, T., Jauhiainen, J., Inoue, T., & Takahashi, H. (2009). Controls on the Carbon Balance of
Tropical Peatlands. Ecosystems, 12(6), 873–887. doi:10.1007/s10021-008-9209-1.
Hocking, P. J., & Meyer, C. P. (1991). Carbon dioxide enrichment decreases critical nitrate and
nitrogen concentrations in wheat. Journal of Plant Nutrition, 14(6), 571–584.
doi:10.1080/01904169109364225.
Holaday, A. S., Martindale, W., Alred, R., Brooks, A. L., & Leegood, R. C. (1992). Changes in
Activities of Enzymes of Carbon Metabolism in Leaves during Exposure of Plants to
Low Temperature. Plant Physiology, 98(3), 1105–1114. doi:10.1104/pp.98.3.1105.
Houghton, J. T., Ding, Y., Griggs, D. J., Noguer, M., van der Linden, P. J., & Xiasu, D. (Eds.).
(2001). Climate change 2001: the scientific basis. Cambridge University Press,
Cambridge.
Hulme, P. E. (2005). Adapting to climate change: is there scope for ecological management in
the face of a global threat? Journal of Applied Ecology, 42, 7840794.
Imai, K., & Sato, O. M. (1991). Effects of temperature on CO2 dependence of gas exchanges in
C3 and C4 crop plants. Japanese Journal of Crop Science (Japan), 60, 139–145.
IPCC. (2007). Climate Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY,
USA.
IPCC (2013) Climate Change 2013: The Physical Science Basis. Contribution of Working Group
I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds
Joussaume S, Penner J, Tangang F), Cambridge University Press, Cambridge, UK and
New York, NY, USA.
184
Isopp, H., Frehner, M., Long, S. P., & Nösberger, J. (2000). Sucrose‐phosphate synthase
responds differently to source‐sink relations and to photosynthetic rates: Lolium
perenne L. growing at elevated pCO2 in the field. Plant, Cell & Environment, 23(6),
597–607. doi:10.1046/j.1365-3040.2000.00583.x.
Jackson, J. B. C., Kirby, M. X., Berger, W. H., Bjorndal, K. A., Botsford, L. W., Bourque, B. J.,
et al. (2001). Historical Overfishing and the Recent Collapse of Coastal Ecosystems.
Science, 293(5530), 629–637. doi:10.1126/science.1059199.
Jenkins, J. C., Chojnacky, D. C., Heath, L. S., & Birdsey, R. A. (2003). National-Scale Biomass
Estimators for United States Tree Species. Forest Science, 49(1), 12–35.
Jensen, J. R., Rutchey, K., Koch, M. S., & Narumalani, S. (1995). Inland wetland change
detection in the Everglades Water Conservation Area 2A using a time series of
normalized remotely sensed data. Photogrammetric Engineering and Remote Sensing, 61,
199–209.
Jimenez, K.L., Starr, G., Staudhammer, C.L., Schedlbauer, J.L., Loescher, H.W., Malone, S.L.,
& Oberbauer, S.F. (2012) Carbon dioxide exchange rates from short- and longhydroperiod Everglades freshwater marsh. Journal of Geophysical Research, 117(G4),
G04009, doi:10.1029/2012JG002117.
Jordan, D. B., & Ogren, W. L. (1984). The CO2/O2 specificity of ribulose 1,5-bisphosphate
carboxylase/oxygenase. Planta, 161(4), 308–313. doi:10.1007/BF00398720.
Keddy, P. A. (1990). Competitive hierarchies and centrifugal organization in plant communities.
In J. B. Grace & D. Ž. Tilman (Eds.), Perspectives on Plant Competition (pp. 65–290).
Keeling, C. D., Whorf, T. P., Wahlen, M., & van der Plichtt, J. (1995). Interannual extremes in
the rate of rise of atmospheric carbon dioxide since 1980. Nature, 375(6533), 666–670.
doi:10.1038/375666a0.
Knickerbocker, C. M., Leitholf, S., Stephens, E. L., Keellings, D. J., Laird, H., Anderson, C. J.
R., et al. (2009). Tree Encroachment of A Sawgrass ( Cladium jamaicense) Marsh within
an Increasingly Urbanized Ecosystem. Natural Areas Journal, 29(1), 15–26.
doi:10.3375/043.029.0103.
Kobza, J., & Edwards, G. E. (1987). Influences of Leaf Temperature on Photosynthetic Carbon
Metabolism in Wheat. Plant Physiology, 83(1), 69–74. doi:10.1104/pp.83.1.
Krinner, G., Viovy, N., & de Noblet Ducoudré, N. (2005). A dynamic global vegetation model
for studies of the coupled atmosphere‐biosphere system. Global Biogeochemical
Cycles, 19, GB1015. doi:10.1002/(ISSN)1525-202710.1029/2003GB002199.
Lambers, H., Chapin, F. S., & Pons, T. L. (2008). Plant Physiological Ecology. New York, NY:
Springer New York. doi:10.1007/978-0-387-78341-3.
185
Li, X., Meixner, T., Sickman, J. O., Miller, A. E., Schimel, J. P., & Melack, J. M. (2006).
Decadal-scale Dynamics of Water, Carbon and Nitrogen in a California Chaparral
Ecosystem: DAYCENT Modeling Results. Biogeochemistry, 77(2), 217–245.
doi:10.1007/s10533-005-1391-z.
Lloyd, J., & Farquhar, G. D. (1994). 13C discrimination during CO2 assimilation by the
terrestrial biosphere. Oecologia, 99(3-4), 201–215. doi:10.1007/BF00627732.
Long, S. P., & Drake, B. G. (1992). Photosynthetic CO2 assimilation and rising atmospheric
CO2 concentrations. In N. R. Baker & H. Thomas (Eds.), Topics in Photosynthesis. Crop
Photosynthesis Spatial and Temporal Determinants (Vol. 2, pp. 69–107). Amsterdam:
Elsevier.
Long, S. P. (1991). Modification of the response of photosynthetic productivity to rising
temperature by atmospheric CO2 concentrations: Has its importance been
underestimated? Plant, Cell & Environment, 14(8), 729–739. doi:10.1111/j.13653040.1991.tb01439.x.
Luo, Y., & Reynolds, J. F. (1999). Validity of extrapolating field CO2 experiments to predict
carbon sequestration in natural ecosystems. Ecology, 80, 1568–1583. doi:10.1890/00129658(1999)080[1568:VOEFCE]2.0.CO;2.
Luo, Y., Geten, D., Le Maire, G., Parton, W. J., Weng, E., Zhou, X., et al. (2008). Modeled
interactive effects of precipitation, temperature, and CO2 on ecosystem carbon and water
dynamics in different climatic zones. Global Change Biology, 14(9), 1986–1999.
doi:10.1111/j.1365-2486.2008.01629.x.
Malone, S.L., Staudhammer, C.L., Oberbauer, S.F., Olivas, P., Ryan, M.G., Schedlbauer, J.,
Loescher, H.W., & Starr, G. (In Preparation). El Niño Southern Oscillation (ENSO)
amplifies precipitation extremes, enhancing CO2 exchange rates in freshwater marsh
ecosystems in the Florida Everglades.
McMurtrie, R. E., & Wang, Y. P. (1993). Mathematical models of the photosynthetic response of
tree stands to rising CO2 concentrations and temperatures. Plant, Cell & Environment,
16(1), 1–13. doi:10.1111/j.1365-3040.1993.tb00839.x.
McMurtrie, R. E., Medlyn, B. E., & Dewar, R. C. (2001). Increased understanding of nutrient
immobilization in soil organic matter is critical for predicting the carbon sink strength of
forest ecosystems over the next 100 years. Tree Physiology, 21, 831–839.
doi:10.1093/treephys/21.12-13.831.
Medlyn, B. E., Dreyer, E., Ellsworth, D. S., FORSTREUTER, M., Harley, P. C., Kirschbaum,
M. U. F., et al. (2002). Temperature response of parameters of a biochemically based
186
model of photosynthesis. II. A review of experimental data. Plant, Cell & Environment,
25, 1167–1179.
Miao, S. L., & Sklar, F. H. (1998). Biomass and nutrient allocation of sawgrass and cattail along
a nutrient gradient in the Florida Everglades. Wetlands Ecology and Management, 5(4),
245–264.
Mitsch, Gosslink. (2007). Wetlands (4 ed.). John Wiley & Sons, Inc., Hoboken, New Jersey.
Mooney, H. A., & Drake, J. A. Ž. (Eds.). (1 86). Ecology of Biological Invasions of North
America and Hawaii. Springer-Verlag, New York.
Newman, S., Schuette, J., Grace, J. B., & Rutchey, K. (1998). Factors influencing cattail
abundance in the northern Everglades. Aquatic Botany, 60(3), 265–280.
doi:10.1016/S0304-3770(97)00089-2.
Nowak, R. S., Ellsworth, D. S., & Smith, S. D. (2004). Functional responses of plants to elevated
atmospheric CO2- do photosynthetic and productivity data from FACE experiments
support early predictions? New Phytologist, 162(2), 253–280. doi:10.1111/j.14698137.2004.01033.x
Oberbauer S.F., Sionit N., Hastings S.J. & Oechel W.C. (1986) Effects of CO2 enrichment and
nutrition on growth, photosynthe- sis and nutrient concentration of Alaskan tundra plant
species. Canadian Journal of Botany 64, 2993–2998.
Parton, W. J., Schimel, D. S., & Cole, C. V. (1987). Analysis of factors controlling SOM levels
in Great Plains grasslands. Soil Science Society of America Journal, 51, 1173–1179.
doi:10.2136/sssaj1987.03615995005100050015x.
Parton, W. J., Stewart, J. W. B., & Cole, C. V. (1988). Dynamics of C, N, P and S in grassland
soils: a model. Biogeochemistry, 5(1), 109–131. doi:10.1007/BF02180320.
Parton, W. J., Scurlock, J., Ojima, D. S., Schimel, D. S., & Hall, D. O. (1995). Impact of climate
change on grassland production and soil carbon worldwide. Global Change Biology, 1(1),
13–22.
Parton, W.J., M. Hartman, D.S. Ojima, D.S.Schimel. 1998. "DAYCENT and its land surface
submodel: description and testing." Global and Planetary Change 19:35-48.
Parton, W. J., Holland, E. A., Del Grosso, S. J., Hartman, M. D., Martin, R. E., Mosier, A. R., et
al. (2001). Generalized model for NO x and N2O emissions from soils. Journal of
Geophysical Research: Atmospheres (1984–2012), 106(D15), 17403–17419.
Parton, W. J., Hanson, P. J., Swanston, C., Torn, M., Trumbore, S. E., Riley, W., & Kelly, R.
(2010). ForCent model development and testing using the Enriched Background Isotope
187
Study experiment. Journal of Geophysical Research: Atmospheres (1984–2012),
115(G4), G04001. doi:10.1029/2009JG001193.
Pepper, D. A., Del Grosso, S. J., McMurtrie, R. E., & Parton, W. J. (2005). Simulated carbon
sink response of shortgrass steppe, tallgrass prairie and forest ecosystems to rising [CO
2], temperature and nitrogen input. Global Biogeochemical Cycles, 19(1), n/a–n/a.
doi:10.1029/2004GB002226.
Poorter, L., & Oberbauer, S. F. (1993). Photosynthetic induction responses of two rainforest tree
species in relation to light environment. Oecologia, 96(2), 193–199.
doi:10.1007/BF00317732.
Reddy, K. R., O'Connor, G. A., & Schelske, C. L. (1999). Phosphorus Biogeochemistry of SubTropical Ecosystems. CRC Press. p 291.
Richardson, J. R., & Marshall, A. (1990). An Evaluation of Refuge Habitats and Relationship to
Water Quality, Quantity and Hydroperiod (p. 332). Report submitted to A.R.M.
Loxahatchee National Wildlife Refuge, Boynton Beach, FL.
Riedo, M., Gyalistras, D., & Fuhrer, J. (2000). Net primary production and carbon stocks in
differently managed grasslands: simulation of site-specific sensitivity to an increase in
atmospheric CO2and to climate change. Ecological Modelling, 134(2), 207–227.
doi:10.1016/S0304-3800(00)00356-2.
Rutchey, K., Vilchek, Les, & Division, S. F. W. M. D. E. S. R. (1994). Development of an
Everglades Vegetation Map Using a SPOT Image and the Global Positioning System.
Photogrammetric Engineering and Remote Sensing, 60, 767–775.
Sacks, W. J., Schimel, D. S., Monson, R. K., & Braswell, B. H. (2006). Model-data synthesis of
diurnal and seasonal CO2 fluxes at Niwot Ridge, Colorado. Global Change Biology,
12(2), 240–259. doi:10.1111/j.1365-2486.2005.01059.x.
Sacks, W. J., Schimel, D. S., & Monson, R. K. (2007). Coupling between carbon cycling and
climate in a high-elevation, subalpine forest: a model-data fusion analysis. Oecologia,
151(1), 54–68. doi:10.1007/s00442-006-0565-2.
Savage, K. E., Parton, W. J., Davidson, E. A., Trumbore, S. E., & Frey, S. D. (2013). Long-term
changes in forest carbon under temperature and nitrogen amendments in a temperate
northern hardwood forest. Global Change Biology, 19(8), 2389–2400.
doi:10.1111/gcb.12224.
Schedlbauer, J. L., Oberbauer, S. F., Starr, G., & Jimenez, K. L. (2010). Seasonal differences in
the CO2 exchange of a short hydroperiod Florida everglades marsh. Agricultural and
Forest Meteorology, 150(2010), 994–1006. doi:10.1016/j.agrformet.2010.03.005.
188
Schedlbauer, J. L., Oberbauer, S. F., Starr, G., & Jimenez, K. L. (2011). Controls on sensible
heat and latent energy fluxes from a short-hydroperiod Florida Everglades marsh. Journal
of Hydrology, 411(3-4), 331–341. doi:10.1016/j.jhydrol.2011.10.014.
Scheffer, M., Carpenter, S., Foley, J. A., Folke, C., & Walker, B. (2001). Catastrophic shifts in
ecosystems. Nature, 413. doi:10.1038/35098000.
Sionit, N., & Patterson, D. T. (1984). Responses of C4 grasses to atmospheric CO2 enrichment I.
Effect of irradiance. Oecologia, 65, 30–34.
Sitch, S., Smith, B., Prentice, I. C., Arneth, A., Bondeau, A., Cramer, W., et al. (2003).
Evaluation of ecosystem dynamics, plant geography and terrestrial carbon cycling in the
LPJ dynamic global vegetation model. Global Change Biology, 9(2), 161–185.
doi:10.1046/j.1365-2486.2003.00569.x.
Stanton, E. A., & Ackerman, F. (2007). Florida and climate change: the costs of inaction.
Stehfest, E., Heistermann, M., Priess, J. A., Ojima, D. S., & Alcamo, J. (2007). Simulation of
global crop production with the ecosystem model DayCent. Ecological Modelling, 209(24), 203–219. doi:10.1016/j.ecolmodel.2007.06.028.
Steward, K. K., & Ornes, W. H. (1975). Assessing a marsh environment for wastewater
renovation. Journal Water Pollution Control Federation, 47(7). doi:10.2307/25038914.
Stitt, M., & Krapp, A. (1999). The interaction between elevated carbon dioxide and nitrogen
nutrition: the physiological and molecular background. Plant, Cell & Environment, 22,
583–621. doi:10.1046/j.1365-3040.1999.00386.x.
Todd, M. J., Muneepeerakul, R., Pumo, D., Azaele, S., Miralles-Wilhelm, F., Rinaldo, A., &
Rodriguez-Iturbe, I. (2010). Advances in Water Resources. Advances in Water
Resources, 33(10), 1279–1289. doi:10.1016/j.advwatres.2010.04.003.
Toth, L. A. (1987). Effects of hydrologic regimes on lifetime production and nutrient dynamics
of sawgrass (No. 87-6). Techical Publication - (87 ed.). South Florida Water Management
District, West Palm Beach, FL.
Toth, L. A. (1988). Effects of hydrologic regimes on lifetime production and nutrient dynamics
of cattail. (No. 88-6,). Techical Publication .
Urban, N. H., Davis, S. M., & Aumen, N. G. (1993). Fluctuations in sawgrass and cattail
densities in Everglades Water Conservation Area 2A under varying nutrient, hydrologic
and fire regimes. Aquatic Botany, 46(3-4), 203–223. doi:10.1016/0304-3770(93)90002E.
189
Veldkamp, E., & O'Brien, J. J. (2000). Calibration of a Frequency Domain Reflectometry Sensor
for Humid Tropical Soils of Volcanic Origin. Soil Science Society of America Journal,
64(5), 1549–1553. doi:10.2136/sssaj2000.6451549x.
Vu, J., Allen, L. H., Boote, K. J., & Bowes, G. (1997). Effects of elevated CO2 and temperature
on photosynthesis and Rubisco in rice and soybean. Plant, Cell & Environment, 20(1),
68–76.
Wand, S., Midgley, G. Y., & Jones, M. H. (1999). Responses of wild C4 and C3 grass (Poaceae)
species to elevated atmospheric CO2 concentration: a meta‐analytic test of current
theories and perceptions. Global Change Biology, 5, 723–741.
Ward, J. K., & Strain, B. R. (1999). Elevated CO2 studies: past, present and future. Tree
Physiology, 19(4-5), 211–220. doi:10.1093/treephys/19.4-5.211.
Weltzin, J. F., Bridgham, S. D., Pastor, J., Chen, J., & Harth, C. (2003). Potential effects of
warming and drying on peatland plant community composition. Global Change Biology,
9(2), 141–151. doi:10.1046/j.1365-2486.2003.00571.x.
Wong, S. C. (1979). Elevated atmospheric partial pressure of CO2 and plant growth. Oecologia,
44(1), 68–74. doi:10.1007/BF00346400.
Ziska, L. H. (2003). Evaluation of the growth response of six invasive species to past, present
and future atmospheric carbon dioxide. Journal of Experimental Botany, 54, 395–404.
doi:10.1093/jxb/erg027.
Ziska, L. H., Drake, B. G., & Chamberlain, S. (1990). Long-term photosynthetic response in
single leaves of A C3 and C4 salt marsh species grown at elevated atmospheric CO2 in
situ. Oecologia, 83(4), 469–472. doi:10.1007/BF00317196.
190
CHAPTER 6: CONCLUSION
Final Conclusion
This research contributes to the understanding of the unique hydrologic attributes of
Everglades wetland ecosystems and the complex relationships between hydrology, energy
exchange, C dynamics (Jimenez et al., 2012) and climate. Everglades hydrology was shown to
co-vary with changes in ecosystems greenhouse warming potential, energy fluxes and ENSO
phase, indicating that hydrology is sensitive to climate perturbations and is important for creating
and maintaining conditions sufficient for wetland ecosystem structure and function.
Hydroperiods are likely to change in the future with the implementation of CERP and climate
change, making it extremely important to understand the complex relationships between
hydrology, climate and CO2, and how these relationships influence ecosystem structure and
function, which feedback to global warming.
Drought and Wetland Greenhouse Warming Potential
Longer dry periods in between heavier precipitation events are projected for the
Everglades region (Stanton & Ackerman 2007; IPCC 2007, 2013). As hydroperiods are
important for C dynamics, increased drought frequency and changes in hydroperiod stand to
have a significant effect on the greenhouse warming potential of Everglades ecosystems. These
results suggest that the greenhouse warming potential of short-hydroperiod freshwater marsh
ecosystems will change as a result of drought. Processes driving CO2 exchange rates and those
191
controlling CH4 production and release respond differently to changes in water levels.
Consequently, extended drought resulted in the system being a net source of carbon to the
atmosphere and increased the ecosystem vulnerability to invasive species. This experiment also
indicates that although drought occurrence and duration may increase in the future, land
managers can influence the way ecosystems respond to drought by preventing abrupt changes in
water levels, which would impede greater gas diffusion out of the soil and delay water stress at
the onset of dry periods. The increase in drought frequency and intensity in the future could
potentially turn subtropical wetland ecosystems into sources of carbon as ecosystem productivity
is reduced by water stress and C stored in the soil becomes oxic for longer periods of time.
Hydrology and Wetland Energy Balance
In addition to its role in wetland C dynamics, the hydrologic cycle is driven by energy
exchanges, which feedbacks into creating the climate sufficient for wetland maintenance. In the
Everglades, Rn, H and LE oscillate with climate and water levels, showing how hydrology drives
energy partitioning. Hydrology defines seasonality in this subtropical ecosystem through its
control on VPD, which is important for both carbon and energy exchanges. Significant
relationships between energy fluxes, VPD, water levels and soil VWC were also observed, and
have also been identified in previous research as important indicators of C dynamics
(Schedlbauer et al., 2012; Jimenez et al., 2012). Considering the effect of hydrology on
vegetation function and composition (Mitsch & Gosslink 2007), changes in vegetation associated
with altered hydrology can also stimulate changes in the climate system through fluctuations in
albedo, surface roughness, soil moisture, and plant resistance to evaporation (Thomas &
Roundtree 1992; Betts et al., 1996; Baldocchi et al., 2000). This study linked environmental
192
variables important for C dynamics and the energy balance in Everglades ecosystems, showing
there exist complex relationships between hydrology, energy exchange, and climate that is
important for creating conditions sufficient to maintain wetland ecosystems.
ENSO, hydro-meteorological parameters, and CO2 dynamics
To explain intra- and inter-annual fluctuations in in situ CO2 exchange rates, I examined
the relationship between ENSO and ecosystem CO2 dynamics in short- and long-hydroperiod
ecosystems. Results suggest that ENSO phases magnified seasonal patterns in CO2 exchange
rates, and differences in season length and intensity explain inter-annual fluctuations in NEE,
Reco, and GEE. The results presented support previously observed relationships between ENSO
phase and season intensity in the Everglades region (Piechota & Dracup 1996; Beckage et al.,
2003; Allan & Soden 2008; Moses et al., 2013), and as a result of the strong relationship between
hydrology, precipitation and CO2 exchange rates (Schedlbauer et al., 2012; Jimenez et al., 2012;
Malone et al., 2013) annual and seasonal net CO2 exchange rates also responded to changes in
ENSO phase. Linking ENSO, season intensity and CO2 dynamics in the Everglades region,
results suggest that shifts in season length and intensity as a result of climate change will likely
become one of the most important factors affecting CO2 dynamics in the Everglades region.
Water management and changes in land cover further complicate future scenarios. Although
future climate characterized by greater extremes and more erratic fluctuations would potentially
alter CO2 dynamics throughout the region, water management improvements may offset the
impact of climate change in ENP by increasing hydroperiods and reducing the system’s
sensitivity to climate.
193
Climate Change and CO2 Exchange Rates
Although short- and long- hydroperiod ecosystems are nearly neutral annually,
simulating the effect of rising atmospheric CO2 concentrations and climate change on ecosystem
CO2 dynamics suggest that carbon sequestration may improve in Everglades wetland ecosystems
with climate change. Causing the greatest increase in net carbon uptake, rising atmospheric CO2
concentrations may be important for reducing the systems vulnerability to invasive species,
disturbance regimes and changes in climate. While projections for temperature and precipitation
are within the natural range of conditions observed, ecosystem CO2 exchange rates were
modified by higher winter minimum and summer temperatures, and by changes in precipitation
patterns. Overall, water levels in the Everglades region likely enhanced the system’s low
sensitivity to change, and a future with very small increases in carbon sequestration seems likely
for the Everglades region. Shifts in species, sea level rise and changes in disturbance frequency
and intensity may also cause significant changes in Everglades ecosystems although these effects
are not considered in this study. Climate change lessened seasonal fluctuations in hydrology
leading to a small enhancement in the sink strength of Everglades ecosystems. These results
suggest that Everglades freshwater ecosystems, which were once a sink for C, may serve as a
negative feedback to global warming. The ecosystems’ hydrology and low nutrient conditions is
a barrier to climate, reducing the ecosystems exposure and vulnerability to change and protecting
soil C pools.
The Future of Everglades Wetland Ecosystems
Climate change is emerging as an important challenge for natural resource managers and
decision makers. With the implementation of the CERP, shifts in climate that result in increased
194
drought frequency and intensity (Stanton & Ackerman 2007) may be moderated by restored
hydrologic conditions. The CERP could re-establish the seasonal patterns in water depth closer
to natural levels, thereby decreasing the system’s sensitivity to climate fluctuations. Feedbacks to
other ecological processes are also likely given this scenario, e.g., changes in species
composition, primary productivity, ratio of anaerobic: aerobic metabolism and organic matter
accumulation. With water managers striving to adjust hydroperiods closer to natural values, in
the future we might expect water levels in short-hydroperiod marsh to increase, offsetting
changes in climate by maintaining current patterns in hydrology. Alternatively, we might
anticipate higher water levels to increase hydroperiods, making the system less sensitive to
climate change altogether. With longer hydroperiods, long-hydroperiod marshes will likely
remain less sensitive to changes in climate and land management. As a result, marshes from a
range of hydroperiods will likely behave more similarly in the future as those at either end of the
range of hydroperiods will become more neutral.
195
References
Abdalla, M., Jones, M., Yeluripati, J., Smith, P., Burke, J., & Williams, M. (2010). Testing
DayCent and DNDC model simulations of N2O fluxes and assessing the impacts of
climate change on the gas flux and biomass production from a humid pasture.
Atmospheric Environment, 44(25), 2961–2970. doi:10.1016/j.atmosenv.2010.05.018.
Allan, R. P. R., & Soden, B. J. B. (2008). Atmospheric warming and the amplification of
precipitation extremes. Science, 321(5895), 1481–1484. doi:10.1126/science.1160787.
Armentano, T. V. (1980). Drainage of Organic Soils as a Factor in the World Carbon Cycle.
BioScience, 30(12), 825–830. doi:10.2307/1308375.
Armentano, T. V., Sah, J. P., Ross, M. S., Jones, D. T., Cooley, H. C., & Smith, C. S. (2006).
Rapid responses of vegetation to hydrological changes in Taylor Slough, Everglades
National Park, Florida, USA. Hydrobiologia, 569(1), 293–309. doi:10.1007/s10750-0060138-8.
Bachoon, D., & Jones, R. D. (1992). Potential rates of methanogenesis in sawgrass marshes with
peat and marl soils in the everglades. Soil Biology and Biochemistry, 24(1), 21–27.
doi:10.1016/0038-0717(92)90237-R.
Baldocchi, D., Kelliher, F. M., Black, T. A., & Jarvis, P. (2000). Climate and vegetation controls
on boreal zone energy exchange. Global Change Biology, 6(S1), 69–83.
doi:10.1046/j.1365-2486.2000.06014.x.
Barr, J. G., Engel, V., Fuentes, J. D., Zieman, J. C., O'Halloran, T. L., Smith, T. J., & Anderson,
G. H. (2010). Controls on mangrove forest‐ atmosphere carbon dioxide exchanges in
western Everglades National Park. Journal of Geophysical Research: Atmospheres
(1984–2012), 115(G2), G02020. doi:10.1002/(ISSN)1525-2027.
Beckage, B., Platt, W. J., Slocum, M. G., & Panko, B. (2003). Influence of the El Nino Southern
Oscillation on fire regimes in the Florida Everglades. Ecology, 84(12), 3124–3130.
doi:10.1093/cvr/cvu015.
Behrenfeld, M. J. (2001). Biospheric Primary Production During an ENSO Transition. Science,
291(5513), 2594–2597. doi:10.1126/science.1055071.
Betts, A. K., Ball, J. H., Beljaars, A. C. M., Miller, M. J., & Viterbo, P. A. (1996). The land
surface‐ atmosphere interaction: A review based on observational and global modeling
perspectives. Journal of Geophysical Research: Atmospheres (1984–2012), 101(D3),
7209–7225. doi:10.1029/95JD02135.
196
Brevik, E. C., & Homburg, J. A. (2004). A 5000 year record of carbon sequestration from a
coastal lagoon and wetland complex, Southern California, USA. Catena, 57(3), 221–232.
doi:10.1016/j.catena.2003.12.001.
Bubier, J., Crill, P., Mosedale, A., Frolking, S., & Linder, E. (2003a). Peatland responses to
varying interannual moisture conditions as measured by automatic CO2 chambers.
Global Biogeochemical Cycles, 17(2). doi:10.1029/2002GB001946.
Bubier, J. L., Bhatia, G., Moore, T. R., Roulet, N. T., & Lafleur, P. M. (2003b). Spatial and
Temporal Variability in Growing-Season Net Ecosystem Carbon Dioxide Exchange at a
Large Peatland in Ontario, Canada. Ecosystems, 6(4), 353–367. doi:10.1007/s10021-0030125-0.
Bush, D. E., Loftus, W. F., & Bass, W. F. (1998). Long-Term Hydrologic Effects on Marsh Plant
Community Structure in the Southern Everglades. Wetlands, 18(2), 230–241.
Childers, D. L. (2006). A synthesis of long-term research by the Florida Coastal Everglades
LTER Program. Hydrobiologia, 569(1), 531–544. doi:10.1007/s10750-006-0154-8.
Choi, Y., & Wang, Y. (2004). Dynamics of carbon sequestration in a coastal wetland using
radiocarbon measurements. Global Biogeochemical Cycles, 18, GB4016.
doi:10.1002/(ISSN)1525-2027.
Craft, C. B., & Richardson, C. J. (1995). Response of everglades plant communities to nitrogen
and phosphorus additions. Wetlands, 15(3), 258–271. doi:10.1007/BF03160706.
Chang, K.-H., Warland, J., Voroney, P., Bartlett, P., & Wagner-Riddle, C. (2013). Using
DayCENT to Simulate Carbon Dynamics in Conventional and No-Till Agriculture. Soil
Science Society of America Journal, 77(3), 941–950. doi:10.2136/sssaj2012.0354.
Christensen, J. H., Hewitson, B., Busuioc, A., Chen, A., Gao, X., Held, R., et al. (2007).
Regional climate projections. Climate Change, 2007: The Physical Science Basis.
Contribution of Working group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change, University Press, Cambridge, Chapter 11
(Vol. 11, pp. 847–940).
Davis, S. M., & Ogden, J. C. (Eds.). (1994). Everglades, the Ecosystem and its Restoration. St.
Lucie Press, Delray Beach, FL, USA.
Del Grosso, S. J., Parton, W. J., Mosier, A. R., Hartman, M. D., Brenner, L., Ojima, D. S., &
Schimel, D. S. (2001). Simulated Interaction of Carbon Dynamics and Nitrogen Trace
Gas Fluxes Using the DAYCENT Model. In Modeling Carbon and Nitrogen Dynamics
for Soil Management (pp. 303–332).
197
Del Grosso, S., Ojima, D., Parton, W., Moiser, A., Peterson, G., & Schimel, D. (2002).
Simulated effects of dryland cropping intensification on soil organic matter and
greenhouse gas exchanges using the DAYCENT ecosystem model. Environmental
Pollution, 116 Suppl 1, S75–S83.
Del Grosso, S., Moiser, A., Parton, W., & Ojima, D. (2005). DAYCENT model analysis of past
and contemporary soil NO and net greenhouse gas flux for major crops in the USA. Soil
and Tillage Research, 83(1), 9–24. doi:10.1016/j.still.2005.02.007.
Del Grosso, S. J., Ojima, D. S., Parton, W. J., & Stehfest, E. (2009). Global scale DAYCENT
model analysis of greenhouse gas emissions and mitigation strategies for cropped soils.
Global and Planetary Change, 67(1-2), 44–50. doi:10.1016/j.gloplacha.2008.12.006.
Duval, B. D., Anderson-Teixeira, K. J., Davis, S. C., Keogh, C., Long, S. P., Parton, W. J., &
DeLucia, E. H. (2013). Predicting Greenhouse Gas Emissions and Soil Carbon from
Changing Pasture to an Energy Crop. PLoS ONE, 8(8), e72019.
doi:10.1371/journal.pone.0072019.t003.
Fuhrer, J. (2003). Agroecosystem responses to combinations of elevated CO2, ozone, and global
climate change. “Agriculture, Ecosystems and Environment,” 97(1-3), 1–20.
doi:10.1016/S0167-8809(03)00125-7.
Hao, Y., Cui, X. Y., Wang, Y. F., Mei, X. R., Kang, X. M., Wu, N., et al. (2011). Predominance
of Precipitation and Temperature Controls on Ecosystem CO2 Exchange in Zoige Alpine
Wetlands of Southwest China. Wetlands, 31(2), 413–422. doi:10.1007/s13157-011-01511.
Hartman, M. D., Baron, J. S., & Ojima, D. S. (2007). Application of a coupled ecosystemchemical equilibrium model, DayCent-Chem, to stream and soil chemistry in a Rocky
Mountain watershed. Ecological Modelling, 200, 493–510.
doi:10.1016/j.ecolmodel.2006.09.001.
Heinsch, F., Heilman, J. L., McInnes, K. J., Cobos, D. R., Zubere, D. A., & Roelke, D. L. (2004).
Carbon dioxide exchange in a high marsh on the Texas Gulf Coast: effects of freshwater
availability. Agricultural and Forest Meteorology, 125(1-2), 159–172.
doi:10.1016/j.agrformet.2004.02.007.
IPCC (2013). Climate Change 2013: The Physical Science Basis. Contribution of Working
Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate
Change. (S. Joussaume & J. Penner, Eds.) (pp. 1–2216). Cambridge University Press,
Cambridge, UK and New York, NY, USA.
IPCC. (2007). Climate Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate
198
Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY,
USA.
Jimenez, K. L., Starr, G., Staudhammer, C. L., Schedlbauer, J. L., Loescher, H. W., Malone, S.
L., & Oberbauer, S. F. (2012). Carbon dioxide exchange rates from short- and longhydroperiod Everglades freshwater marsh. Journal of Geophysical Research, 117(G4),
G04009. doi:10.1029/2012JG002117.
King, G. M., Roslev, P., & Skovgaard, H. (1990). Distribution and rate of methane oxidation in
sediments of the Florida everglades. Applied and Environmental Microbiology, 56, 2902–
2911.
Lafleur, P. M. (2008). Connecting Atmosphere and Wetland: Energy and Water Vapour
Exchange. Geography Compass, 2(4), 1027–1057. doi:10.1111/j.17498198.2007.00132.x.
Li, X., Meixner, T., Sickman, J. O., Miller, A. E., Schimel, J. P., & Melack, J. M. (2006).
Decadal-scale Dynamics of Water, Carbon and Nitrogen in a California Chaparral
Ecosystem: DAYCENT Modeling Results. Biogeochemistry, 77(2), 217–245.
doi:10.1007/s10533-005-1391-z.
Luo, Y., Geten, D., Le Maire, G., Parton, W. J., Weng, E., Zhou, X., et al. (2008). Modeled
interactive effects of precipitation, temperature, and CO2 on ecosystem carbon and water
dynamics in different climatic zones. Global Change Biology, 14(9), 1986–1999.
doi:10.1111/j.1365-2486.2008.01629.x.
Loveless, C. M. (1959). A Study of the Vegetation in the Florida Everglades. Ecology, 40(1), 1–
9. doi:10.2307/1929916.
Malone, S. L., Starr, G., & Staudhammer, C. L. (2013). Effects of simulated drought on the
carbon balance of Everglades short‐ hydroperiod marsh. Global Change Biology.
doi:10.1111/gcb.12211
Mitsch, Gosslink. (2007). Wetlands (4 ed.). John Wiley & Sons, Inc., Hoboken, New Jersey.
Moses, C. S., Anderson, W. T., Saunders, C., & Sklar, F. (2012). Regional climate gradients in
precipitation and temperature in response to climate teleconnections in the Greater
Everglades ecosystem of South Florida. Journal of Paleolimnology, 49(1), 5–14.
doi:10.1007/s10933-012-9635-0.
Myers, R. L., & Ewel, J. J. (Eds.). (1992). Ecosystems of Florida. University of Central Florida
Press, Orlando, FL, USA.
Parton, W. J., Schimel, D. S., Cole, C. V., & Ojima, D. S. (1987). Analysis of Factors
Controlling Soil Organic Matter Levels in Great Plains Grasslands. Soil Science Society
of America Journal, 51(5), 1173–1179. doi:10.2136/sssaj1987.03615995005100050015x.
199
Parton, W. J., Stewart, J. W. B., & Cole, C. V. (1988). Dynamics of C, N, P and S in grassland
soils: a model. Biogeochemistry, 5(1), 109–131. doi:10.1007/BF02180320.
Parton, W. J., Scurlock, J., Ojima, D. S., Schimel, D. S., & Hall, D. O. (1995). Impact of climate
change on grassland production and soil carbon worldwide. Global Change Biology, 1(1),
13–22.
Parton, W. J., Hartman, M., Ojima, D., & Schimel, D. (1998). DAYCENT and its land surface
submodel: description and testing. Global and Planetary Change, 19(1), 35–48.
doi:10.1016/S0921-8181(98)00040-X.
Parton, W. J., Hanson, P. J., Swanston, C., Torn, M., Trumbore, S. E., Riley, W., & Kelly, R.
(2010). ForCent model development and testing using the Enriched Background Isotope
Study experiment. Journal of Geophysical Research, 115(G4), G04001.
doi:10.1029/2009JG001193.
Perry, W. (2004). Elements of south Florida's comprehensive Everglades restoration plan.
Ecotoxicology, 13, 185–193.
Piccolo, M. C. (2009). Heat Energy Balance in Coastal Wetlands. In Coastal wetlands: an
integrated ecosystem approach (First edition. pp. 211–230). Elsevier. doi:10.1016/B9780-444-53103-2.00007-7.
Piechota, T. C., & Dracup, J. A. (1996). Drought and Regional Hydrologic Variation in the
United States: Associations with the El NioSouthern Oscillation. Water Resources
Research, 32(5), 1359–1373.
Redfield, G. W. (2000). Ecological research for aquatic science and environmental restoration in
south Florida. Ecological Applications, 10(4), 990–1005.
Richardson, C. (2008). The Everglades Experiments. In Lessons for Ecosystem RestorationMacrophyte Community Responses in the Everglades with an Emphasis on Cattail
(Typha domingensis) and Sawgrass (Cladium jamaicense) Interactions along a Gradient
of Long-Term Nutrient Additions, Altered Hydroperiod, and Fire. New York: Springer.
doi:10.1007/978-0-387-68923-4.
Savage, K. E., Parton, W. J., Davidson, E. A., Trumbore, S. E., & Frey, S. D. (2013). Long-term
changes in forest carbon under temperature and nitrogen amendments in a temperate
northern hardwood forest. Global Change Biology, 19(8), 2389–2400.
doi:10.1111/gcb.12224.
Schedlbauer, J. L., Oberbauer, S. F., Starr, G., & Jimenez, K. L. (2011). Controls on sensible
heat and latent energy fluxes from a short-hydroperiod Florida Everglades marsh. Journal
of Hydrology, 411(3-4), 331–341. doi:10.1016/j.jhydrol.2011.10.014.
200
Schedlbauer, J. L., Munyon, J. W., Oberbauer, S. F., Gaiser, E. E., & Starr, G. (2012). Controls
on Ecosystem Carbon Dioxide Exchange in Short- and Long-Hydroperiod Florida
Everglades Freshwater Marshes. Wetlands, 32(5), 801–812. doi:10.1007/s13157-0120311-y.
Sitch, S., Smith, B., Prentice, I. C., Arneth, A., Bondeau, A., Cramer, W., et al. (2003).
Evaluation of ecosystem dynamics, plant geography and terrestrial carbon cycling in the
LPJ dynamic global vegetation model. Global Change Biology, 9(2), 161–185.
doi:10.1046/j.1365-2486.2003.00569.x.
Stanton, E. A., & Ackerman, F. (2007). Florida and climate change: the costs of inaction.
Florida and climate change: the costs of inaction. Global Development and Environment
Institute, Tufts University.
Stehfest, E., Heistermann, M., Priess, J. A., Ojima, D. S., & Alcamo, J. (2007). Simulation of
global crop production with the ecosystem model DayCent. Ecological Modelling, 209(24), 203–219. doi:10.1016/j.ecolmodel.2007.06.028.
Thomas, G., & Rowntree, P. R. (1992). The Boreal Forests and Climate. Quarterly Journal of
the Royal Meteorological Society, 118(505), 469–497. doi:10.1002/qj.49711850505.
Trenberth, K. E. (1997). The Definition of El Niño. Bulletin of the American Meteorological
Society, 78(12), 2771–2777. doi:10.1175/15200477(1997)078<2771:TDOENO>2.0.CO;2.
Todd, Muneepeerakul, Pumo, Azaele, Rinaldo. (2010). Hydrological drivers of wetland
vegetation community distribution within Everglades National Park, Florida. Advances in
Water Resources, 33(10), 11–11. doi:10.1016/j.advwatres.2010.04.003.
Torn, M. S., & Chapin, F. S., III. (1993). Environmental and biotic controls over methane flux
from Arctic tundra. Chemosphere, 26(1-4), 357–368. doi:10.1016/0045-6535(93)904314.
Webster, K. L., McLaughlin, J. W., Kim, Y., Packalen, M. S., & Li, C. S. (2013). Modelling
carbon dynamics and response to environmental change along a boreal fen nutrient
gradient. Ecological Modelling, 248, 148–164. doi:10.1016/j.ecolmodel.2012.10.004.
Whalen, S. C. (2005). Biogeochemistry of methane exchange between natural wetlands and the
atmosphere. Environmental Engineering Science, 22(1), 73–94.
doi:10.1089/ees.2005.22.73.
Whiting, G. J., & Chanton, J. P. (1993). Primary production control of methane emission from
wetlands. , Published Online: 26 August 1993; | Doi:10.1038/364794a0, 364(6440),
794–795. doi:10.1038/364794a0.
201
Whiting, G. J., & Chanton, J. P. (2001). Greenhouse carbon balance of wetlands: methane
emission versus carbon sequestration. Tellus B, 53(5), 521–528. doi:10.1034/j.16000889.2001.530501.x.
Zedler, J. B., & Kercher, S. (2005). Wetland ResourcesS: Status, Trends, Ecosystem Services,
and Restorability. Annual Review of Environment and Resources, 30(1), 39–74.
doi:10.1146/annurev.energy.30.050504.144248.
202