Forest Ecology
and
Management
ELSEVIER
Forest Ecology and Management 74 (1995) 171-180
Seasonal changes of leaf area index (LAI) in a tropical deciduous
forest in west Mexico
Jose Manuel Maass3'*, James M. Voseb, Wayne T. Swankb, Angelina Martinez-Yrizarc
"Centra de Ecologia, Universidad National Autonoma de Mexico, Apdo. Postal 70-275, C.P. 04510, Mexico, D.F., Mexico
b
USDA Forest Service Coweeta Hydrologic Laboratory, 999 Coweeta Laboratory Road, Otto, NC 28763, USA
'Centra de Ecologia-Unidad Hermosillo, Universidad National Autonoma de Mexico, Apdo. Postal 1354, C.P. 83000,
Hermosillo, Sonora, Mexico
Accepted 6 September 1994
Abstract
Light canopy transmittance and the Beer-Lambert equation were utilized to assess monthly leaf area index
(LAI) of a tropical deciduous forest ecosystem on the west coast of Mexico. The light transmittance coefficient
(k) was obtained by analyzing vertical leaf and light distribution in the forest canopy. An independent LAI estimate was obtained using litterfall data. The calculated k value was 0.610 ±0.035 (standard error). Average maximum LAI obtained with litterfall data was 4.2 + 0.4 m2 m~ 2 . There was a significant correlation (P< 0.001, r= 0.98)
between litter-LAI estimations and those obtained with the Beer-Lambert equation. The regression explained 95%
of the variation; however, light-LAI overestimated litter-LAI by a constant of 0.87±0.12 m2 m~ 2 (the slope was
1.03 and Y intercept was 0.87). The discrepancy is partially attributed to leaf retention of the few evergreen species, and perhaps leaf retention of a few deciduous species beyond the end of the litterfall collection. Maximum
annual LAI was similar in both study years (4.5±0.3 m2 m~ 2 in 1990 and 4.9±0.4 in 1991). Minimum LAI
showed considerable variation between years with similar values in the dry seasons of 1990 and 1991 (1.0 ±0.1
m2 and 0.9 + 0.1 m2 mr2, respectively), but much higher values in 1992 (2.7 + 0.2 m2 m~ 2 ). The difference is
probably attributed to an atypical rainfall event in January 1992 (644 mm), which retarded leaf abscission.
Keywords: Canopy light transmittance; Light extinction coefficient; Litterfall; Seasonal tropical forest; Tropical dry forest
1. Introduction
Leaf area index (LAI) is an important structural characteristic of forests because the forest
canopy is the site of significant ecosystem processes such as transpiration, rainfall interception, dry deposition, and photosynthesis. As a
result, LAI has been identified as a key parame* Corresponding author: Tel. (5) 622-9005; Fax. (5) 6228995.
ter in studying and modelling ecosystem function at local, regional and global scales (Swift et
al., 1975; Parton et al., 1992). LAI varies greatly
among ecosystems, ranging from less than 1 m2
m~ 2 in arid ecosystems, up to 20 in some conifer
stands (Kozlowski et al., 1991). LAI also varies
within ecosystems depending on site conditions,
particularly water supply and soil fertility. Although LAI is well documented for temperate
forests, only a few determinations have been
made for tropical ecosystems (Medina and
0378-1127/95/S09.50 © 1995 Elsevier Science B.V. All rights reserved
SSD/0378-1127(94)03485-0
172
J.M. Maassetal. / Forest Ecology and Management 74 (1995) 171-180
Klinge, 1983). Data for tropical deciduous forests are particularly scarce (Murphy and Lugo,
1986a). In deciduous, and some evergreen ecosystems, LAI varies seasonally, having a maximum during the growing season and a minimum
in the dormant season. Quantifying this temporal variation is important for understanding
variation in the rates of many ecosystem processes (e.g. transpiration); however, data on
temporal LAI variability are extremely rare
(Lugoetal., 1978; Murphy and Lugo, 1986b).
A variety of methods have been used to assess
LAI in forest ecosystems (Chason et al., 1991).
Sampling canopy light transmittance with a
portable integrating radiometer, and applying the
Beer-Lambert Law has been used to estimate LAI
in coniferous forests (Pierce and Running, 1988;
Bolstad and Gower, 1990; Vose and Swank, 1990;
Burton et al., 1991). This approach requires an
estimate of canopy light extinction efficiency (k).
A significant advantage of light transmittance is
the ability to non-destructively quantify seasonal changes.
The objective of the present study was to
quantify total LAI and to assess seasonal LAI
variation in a tropical deciduous forest ecosystem using the light transmittance method.
the year, averaging 25 °C. Monthly mean minimum and maximum temperatures are 15.9°C
(February) and 32.2°C (August), respectively
(Bullock, 1988).
Soils are young, often shallow (0.5-1 m
depth), and poorly structured. The parent material is mostly rhyolite. Soils are predominantly
sandy loams and rocks are common in the superficial horizon. Soils have low organic matter content (mean 2.5%) and low mineral nutrient concentrations (mean exchangeable Ca2+, Mg2+,
and K+ content is 8.7 meq per 100 g of dry soil),
and pH values vary between 6.0 and 7.0 (Maass
etal., 1988).
The predominant vegetation in the area is the
tropical deciduous forest type of Rzedowski
(1978). Lott (1985) reports 758 plant species
from 107 families for the area, with Leguminosae being the most important family with 15% of
the species. According to Martinez-Yrizar et al.
(1992) above-ground phytomass is 85 Mg ha~'.
Mean canopy height is 7 m with some trees up to
15 m tall located along the river channel. The
stand basal area ranges from 14.0 to 26.1 m2 ha~'
for trees with a diameter at breast height of over
3.0 cm.
2.2. Site description
2. Methods
2.1. Study area
The study was conducted at the Estacion de
Biologia Chamela of the Universidad Nacional
Autonoma de Mexico. Chamela is located at
19°30'N, 105°03'W in the state of Jalisco on the
Pacific coast of Mexico.
The climate of the area is influenced by tropical cyclones producing a highly variable annual
rainfall regime (Garcia-Oliva et al., 1991). Mean
precipitation is 707 mm (1977-1988) with more
than a 500 mm difference between the wettest
and the driest year (897 mm in 1989 and 374
mm in 1985). Rainfall is strongly seasonal with
80% of the annual precipitation falling between
July and October, with September being the wettest month. Temperature fluctuates little during
Within the field station, five contiguous small
watersheds (12-28 ha each) have been gauged
for long-term ecosystem research (Sarukhan and
Maass, 1990). On one of the watersheds (Watershed I), three permanent plots were previously established for detailed analyses of ecosystem processes, and they were used for the present
study. The plots are distributed along the altitude gradient of the watershed, and will be referred to here as upper, middle and lower plots.
Each plot is 2400 m2 (80 m x 30 m) with the long
axis perpendicular to the stream channel. The
general characteristics of each plot are shown in
Table 1. According to L.A. Perez-Jimenez, unpublished data, 1993, the most important species on each plot are: in the upper plot, Guapira
macrocarpa Miranda, Plumeria rubra L., Lonchocarpus constrictus Pitt., Bur sera instabilis
McVaugh and Rzed. and Colubrina heteroneura
J.M. Maassetal. / Forest Ecology and Management 74(1995) 171-180
173
Table 1
Site characteristics of the permanent sampling plots of a tropical deciduous forest at Chamela, Jalisco, Mexico (Watershed I).
LAI estimated using litterfall and specific leaf area (SLA) data. Tree density and basal area were obtained from all individuals
with a diameter at breast height of over 3.0 cm
Parameter
Elevation3
Slope range3
Soil depth"
Soil organic matter3
Tree density"
Basal areab
Litterfall0
Specific leaf area (SLA) C
LAI (litter-LAI)c
Units
m above sea level
degrees
cm
%
individuals ha ~'
m 2 ha-'
Mgha~' year"'
cm2 g~ '
m 2 m- 2
Sampling plots
Lower
Middle
Upper
60-80
16-30
25-30
3.0-3.5
2113
19.77
4.5
203
5.4
120-140
8-16
15-25
2.5-3.0
3225
17.18
3.6
188
3.8
140-160
8-16
40-45
2.0-2.5
2492
11.22
3.5
154
3.3
a
Galicia(1992).
L.A. Perez-Jimenez (personal communication, 1994).
Tresent study.
b
(Griseb.) Standl.; in the middle plot, Guapira
macrocarpa Miranda, Lonchocarpus eriocarinalis Micheli, Plumeria rubra L., Piptadenia constricta (Micheli) Macbr. and Bursera instabilis
McVaugh and Rzed.; in the lower plot, Thouinidium decandrum (Humb. & Bonpl.) Radlk.,
Guapira macrocarpa Miranda, Astronium graveolens Jacq., Trichilia trifolia L. and Casearia
corymbosa HBK.
Three meteorological towers were used to
quantify vertical LAI distribution through the
canopy profile and to develop an extinction
coefficient (discussed below). One of the towers
is within the sampled area, the middle plot of
Watershed-I. The other two towers are located on
nearby sites with forests comparable to Watershed-I. One is located on Watershed II and the
other on the highest point of an access trail (Tejon) to the watersheds (about 300 m away from
the sampling plots, on a small hill of 130 m above
sea level).
2.3. LAI determinations
Seasonal LAI was determined from May 1990
to May 1992 using the light transmittance technique. The method is based on the relationship
between leaf area and light transmittance, de-
scribed by the Beer-Lambert model (Monsi and
Saeki, 1953)
where Q{ is the sub-canopy diffuse and beam radiation, Q0 is the above-canopy radiation, and k
is the radiation extinction coefficient. The model
assumes that the foliage is randomly distributed
in the canopy, and that leaf inclination angles are
spherically distributed in space. However, as discussed by Jarvis and Leverenz (1983) and Pierce
and Running (1988), the Beer-Lambert equation is fairly insensitive to all but major violations of these assumptions. To validate the light
transmittance approach, we compared light
transmittance LAI estimates (light-LAI) with
LAI estimates derived from litterfall data (litterLAI).
2.4. Litterfall collections and specific leaf area
determinations
Litterfall was collected using conical fiberglass
mesh traps, 0.5 m in diameter (area 0.1963 m2)
and 50 cm deep (for details see Martinez- Yrizar
and Sarukhan, 1990). Eight traps were located
randomly on each plot, giving a total of 24 traps.
Litter was collected during 1 year at monthly in-
174
J.M. Maassetal. /Forest Ecology and Management 74(1995) 171-180
tervals starting on June 1990. The material was
stored in paper bags and oven-dried at 80 °C to
constant mass. One-way ANOVA was performed
to test for differences between sample plots.
To determine specific leaf-area values, ten
randomly selected leaves were obtained from
each of the 24 litter fall traps (A?=240). Leaves
were sampled from the November collection,
which represents the peak litterfall during the
year. The leaves were carefully rehydrated, unfolded, pressed, and their leaf area measured using a Licor leaf area meter. The leaves were dried
at 80 °C to constant mass. To validate this procedure, 250 senescent leaves (but still attached
to branches) were randomly selected from trees
located in the permanent plots. Leaves were
pressed fresh and the leaf area measurements
were made within one week of sampling, and then
oven-dried for dry weight measurements. An
analysis of variance was conducted to compare
the two independent specific leaf-area estimates.
2.5. Light transmittance readings
Canopy transmittance of the photosynthetically active radiation was measured 1 m above
the forest floor, at the same locations as the litter
traps. At each of the 24 locations, ten measurements of canopy transmittance (QJ were taken
using a 'sunfleck ceptometer' (Decagon Devices,
1989). The ceptometer consists of a narrow (80
cm long) probe with 80 photosynthetically active radiation sensors located at 1 cm intervals
(for more details on the ceptometer description
and its theory of operation, see Pierce and Running, 1988). To obtain the ten measurements per
trap, the ceptometer was rotated 360°, and photosynthetically active radiation readings were
taken at approximately 36° intervals. Samples
were obtained on cloudless days, usually between 11:00 and 13:00 h. Total incoming PAR
((2o) was measured either above the canopy or
in an opening at the beginning and at the end of
the sampling period. The arithmetic mean of
these two readings was used as Q0 for QJQ0 calculations. Measurements were made seasonally
during May and September of 1990, and monthly
from January 1991 to May 1992.
Within-canopy photosynthetically active radiation transmittance was measured from meteorological towers. On each tower, PAR transmittance readings were taken every meter, from the
forest floor to the top of the canopy. At each level,
ten readings were taken rotating the ceptometer
at 36 ° intervals, following the same procedure as
described above.
2.6. Light extinction coefficient (k)
determination
To obtain k, vertical LAI and QJQ0 profiles
were obtained on each of the towers. In October
1991, when the canopy was fully developed, direct LAI determinations were obtained using the
following procedure. A string and plumb-bob
were placed 1 m out from the tower and lowered
vertically through the vegetation profile from the
top of the canopy to the forest floor. LAI was determined by counting the number of times the
string touched a leaf, and summarized by 1 m
height intervals. The procedure was repeated five
times in a circular fashion on each tower and the
arithmetic mean per interval was recorded. A
regression analysis was performed on cumulative LAI vs. ln((2i/(2o)> where the slope of the
regression line defines k. A total of 30 data points
was used in the analysis and each tower provided
8-12 measurements, depending on canopy
height.
An independent LAI estimate was obtained
from litterfall data (litter-LAI). Only litterfall
from mid-wet season to end of the dry season
(September-May) was considered, since maximum canopy development is obtained in September. Leaves represent 61% and 73% of the wet
and dry season total litterfall amounts, respectively (Martinez-Yrizar and Sarukhan, 1990).
Therefore, to calculate the actual leaf-fall values
(leaves only, not woody litter, fruits, flowers,
etc.), the cumulative litterfall during the September-October period (wet season) was multiplied by a factor of 0.61, and the cumulative
litterfall from the November-May period (dry
season) was multiplied by 0.73. To obtain the
litter-LAI value for each trap location, the total
J.M. Maassetal. / Forest Ecology and Management 74(1995) 171-180
cumulative leaf-fall was multiplied by the specific leaf area calculated for the trap.
Since most tree and shrub individuals in this
forest type are deciduous, almost all of the leaves
are dropped by the end of the dry season (MayJune); therefore, it was possible to calculate
monthly litter-LAI values. LAI for a particular
month was calculated by considering the cumulative leaf fall from that particular month to the
end of the dry season. Since leaves are continuously produced and dropped during the rainy
season, only those months after the end of leaf
production were considered. We assumed October as the end of the leaf production month since
it is when litterfall abruptly increased. Correlation analysis was used to compare estimated LAI
using the Beer-Lambert equation (light-LAI)
with the monthly litterfall-LAI values.
3. Results and discussion
3.1. Rainfall pattern during the study period
Two very distinctive annual rainfall patterns
occurred during the study period (Fig. 1). During 1990-1991, 557 mm of rain fell from June to
October, followed by no rainfall from November
to May. In 1991-1992, about the same amount
175
of rain fell from June to November (553 mm),
but it was followed by an extreme rainfall event
of 644 mm in January 1992. Precipitation during winter months has a strong influence on the
forest phenology by reducing leaf senescence and
abscision, and, in some cases, stimulating new
leaf production (Bullock and Solis-Magallanes,
1990; Martinez-Yrizar and Sarukhan, 1990).
3.2. Canopy photosyntheticatty active radiation
transmittance
Photosynthetically active radiation measurements were highly sensitive to canopy phenology
and vertical distribution of foliage. Fig. 2 shows
QJQo profiles measured for wet and dry seasons
on the tower located on watershed I (middle
plot). Minimum photosynthetically active radiation transmittance was observed during the
wet season in both years (October). Maximum
transmittance was observed during the 19901991 dry season (April 1991); however, the
rainfall event of January 92 delayed leaf senescence and abscission, thus reducing photosynthetically active radiation transmittance during
the 1991-1992 dry season (April 1992).
The same seasonal pattern was observed on the
other two permanent plots. Also, consistently
higher PAR transmittance was observed in the
Rainfall Distribution
-644mm
E
260
240
220
200
180
160
140
120
100
80
60
40
20
0
N
X
x
x
5?
x
x
^
X
X
x
x
1990
x
K x ^
x x x
3x
x
X X
x
x
n
x
x
x
' i '
x
q
X
x
X x
5< x
3
x
xX
3X x H
u
q
^
x
x x x
x
SI ' N ' J ' M ' M
I I i | I I i | i
•.I
x
J
1991
X
x
X
X
**
x
x
H
RH
1 1 1 1 1
X
*
1 R1 1
1
J S'N JJ.M.M
J A 0 D F A
1992
Fig. 1. Rainfall distribution during the study period. Data from the meteorological station at the Chamela Biological Reserve,
Jalisco, Mexico.
176
J.M. Maassetal. / Forest Ecology and Management 74 (1995) 171-180
Vertical Profile
file of
of Light
Light Transmittance
Tr«
(Middle
iddle plot)
l
o
q
o
1 2
3
4
5
6
7
8
9 10 11
Canopy Height (m)
Fig. 2. Light transmittance along the canopy vertical profile of a tropical deciduous forest in Chamela, Jalisco, Mexico. QJQ0 is
the sub-canopy/above-canopy PAR ratio. Sampling was conducted during the dry (April) and wet (October) seasons.
upper and middle plots, compared with the lower
plot.
composition and stand structure, resulting from
differences in topography and water availability.
For example, the Watershed-II tower site was located next to an ephemeral stream. As a result,
the Watershed-II tower site contains taller trees
and different species than those found at the
other two towers sites.
The regression analysis between the cumula-
3.3. Canopy light extinction coefficient (k)
Vertical LAI distributions determined by the
plumb method are shown in Fig. 3. Differences
between towers may reflect variation in species
Vertical Profile of Leaf Area Index
(0
o>
Q.
O
(B
O
0.0
0.4
0.8
1.2
2
2
LAI (m /m )
Fig. 3. LAI distribution along the canopy vertical profile of a tropical deciduous forest in three contrasting sites at Chamela,
Jalisco, Mexico. LAI was determined with the plumb method. Ws-I, Watershed I; Ws-II, Watershed II.
J.M. Maass el al. /Forest Ecology and Management 74 (1995) 171-180
177
LAI Vs Light Transmittance
0
-1
-2
I
O
-3
-4
ln(Qi/Qo) = -0.61 Cum LAI
12 = 0.78 (p> 0.05)
-5
-6
Cumulative LAI (m2/m2)
Fig. 4. Fitted line from the regression analysis between cumulative LAI (obtained by the plumb method) and canopy transmittance (QJQa), to obtain the light extinction coefficient (k) for a tropical deciduous forest at Chamela, Jalisco, Mexico.
tive LAI vs. ln(QJQ0) measured along the towers was highly significant (P< 0.001). About 78%
of the variation in \n(QJQ0) was explained by
the LAI (r2=0.778, Fig. 4). Small differences in
canopy structure and tree species between sampling sites (towers) may be responsible for some
variability about the regression line. Since no
tower site was a perfect match for the stands
where QJQ0 was measured, data from all towers
(three) were pooled to obtain an average canopy
value. The k (slope of the regression
linelstandard error) was -0.610±0.035. The
value is within the range of k reported by Jarvis
andLeverenz (1983) for temperate (deciduous
and evergreen) canopies (—0.40 to —0.65).
3.4. Litterfall, specific leaf area, and litter-LAI
estimations
Annual litterfall was significantly greater
(0.05 >P> 0.025) than on the lower plot (4.5 Mg
ha~') compared with the middle and upper plots
(3.6 Mg and 3.5 Mg ha"1, respectively). Maximum litterfall occurred between November and
December (Fig. 5). During these 2 months, more
than 50% of the leaves in the upper and middle
plots fell, whereas in the lower plot the percentage was 38%. Thus, leaf production was both
higher and persisted longer on the lower plot.
Annual litterfall for the watershed (averaging
all three plots) was 3.9 ± 0.2 Mg ha~'. This value
is within the 6 year (June 1992-May 1988) average range of annual litterfall values (3.4-4.7
Mg ha~') in all five watersheds at Chamela, obtained from 168 litter traps (A. Martinez-Yrizar
et al., unpublished data, 1988).
Average specific leaf area from leaves collected from the litter-traps was 182 ± 51 cm2 g~'.
This value did not significantly differ from the
average specific leaf area from senescent leaves
collected from the trees (184175 cm2 g" 1 ). A
pattern of decreasing specific leaf area with increasing altitude was observed: 203 cm2,188 cm2
and 154 cm2 g~' for the lower, middle and upper
plots, respectively. However, the differences between plots were not statistically significant
(0.25>P>0.10). Our specific leaf area values
are high, but close to the average specific leaf area
value of 195 cm2 g"1 for tropical deciduous forest trees reported by Medina and Klinge (1983).
Average annual maximum LAI obtained using
litterfall data (litter-LAI) was 4.2±0.4 m2 m"2.
It was slightly but significantly higher
(0.10>P>0.05) in the lower plot than in the
middle and upper plots (5.4±0.7 m2 m~ 2 ,
3.810.5 m2 m-2 and 3.310.5 m2 m"2, respectively). There was a significant correlation
(r=0.98, .P<0.001) between litter-LAI estimations and those obtained with the Beer-Lambert
equation (light-LAI, Fig. 6). The regression
J.M. Maassetal. / Forest Ecology and Management 74(1995) 171-180
178
Monthly Litterfall
1 -2 1
1 -0 1
I
Upper plot
Middle plot
Lower plot
to
JZ
=
0.6-
•g
to
£ 0.4
0.2-
t:
0.0
Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May
1990
1991
Fig. 5. Litterfall dynamics of a tropical deciduous forest at Chamela, Jalisco, Mexico. Sampling was conducted from June 1990
to May 1991 on three permanent 80 m x 30 m plots on Watershed I. The vertical bars indicate ± 1 SE.
Litter- Vs Light-Leaf Area Index
Light-LAI = 1.03(Litter-LAI) + 0.87
r2 = 0.95
.E
01
E,
3
I
* Upper plot
+ Middle plot
o Lower plot
£
D)
2
3
4
Litter- LAI (nWm2)
Fig. 6. Correlation and fitted line of regression between LAI from litterfall data (litter-LAI) and LAI from light transmittance
measurements (light-LAI) in a tropical deciduous forest at Chamela, Jalisco, Mexico.
equation explained 95% of the variation; however, light-LAI was greater than litter-LAI by a
constant of 0.87 ±0.12 m2 m~ 2 (the slope was
1.03 ±0.05). The greater LAI predicted by the
light technique may reflect LAI of evergreen species, and perhaps leaf retention of a few deciduous species beyond the end of the litterfall collection. In both cases, the light-LAI technique
should provide a better estimate of stand LAI because it is based on the actual amount of foliage
in the canopy.
3.5. Seasonal variations of light-LAI
Seasonal LAI calculated using the Beer-Lambert equation and the monthly QJQ0 recording
from the permanent plots are shown in Fig. 7.
J.M. Maass et al. / Forest Ecology and Management 74(1995) 171-180
179
Seasonal Pattern of Leaf Area Index
1
o
6—
0
o
5 ~
fj
/
/
2-
/
?
n -
\
\p
0
^Ts ( )
°/^
/T* ^. > \ O 0
k»
«tvr jk
/
*
/
/*\
1
""
±" 3 ~
c
>v
f
O
o
O
*+
+
A M J J A S O N D J F M A M J J A S O N D J F M A M J
1990
1991
1992
* = Upper plot, + = Middle plot, o = Lower plot,
LJ Months with s 50mm of rainfall
Fig. 7. Seasonal LAI (determined by the light transmittance method) of a tropical deciduous forest at Chamela, Jalisco, Mexico.
Sampling was conducted from May 1990 to May 1991, on three permanent 80 m x 30 m plots at Watershed I. The line quantifies
the mean of the three plots (solid line shows the monthly sampling period, dashed line shows the 4 month sampling period).
The lower plot had consistently higher light-LAI
than middle and upper plots. This pattern is the
result of differences in site characteristics since
deeper soils at lower slope positions result in
higher soil moisture conditions and therefore
longer leaf retention.
Light-LAI increases at the beginning of the
rainy season, reaches a brief plateau during the
August-October period, and then slowly decreases toward the end of the dry season. The average yearly maximum light-LAI for Ws-I was 4.7
m2 mr 2 which was observed in September. In the
1990 and 1991 dry seasons, average annual minimum light-LAI was 1.0± 0.1 m2 m~ 2 which occurred toward the end of the dry season (AprilMay). However, in the 1992 dry season, winter
storms promoted leaf maintenance for a longer
period, raising LAIs to 2.67 + 0.2 m2 m~ 2 in all
plots.
Murphy and Lugo (1986b) report seasonal
LAIs for tropical deciduous forest at Guanica,
Puerto Rico (average annual precipitation of 860
mm). At their site, maximum LAI was 4.3 m2
m~ 2 during a wet period (July 1981), and minimum was 2.1 during the dry period (March
1982). The yearly rainfall variability in the vicinity of Guanica, explains the highly variable
year-to-year LAI values. They also mention that
only 26% of the species in the Guanica site are
completely deciduous, which explains why minimum LAI never became less than 2 m2 m~ 2 .
Ogawa et al. (1961) reports a higher LAI for a
dry monsoon forest at Ping Kong, Thailand (6.6
m2 m~ 2 ). Greater rainfall in Ping Kong (1300
mm) than in Chamela, may explain this higher
LAI value.
Usually, a single LAI value is used to characterize ecosystem canopy development. However, LAI can vary considerably within the year,
particularly in deciduous ecosystems. Since
monthly LAI values are rarely assessed, it has
been suggested that LAI should be supplemented
with an indication of leaf duration to obtain a
more precise parameter of canopy development
(Medina and Klinge, 1983). Our results show
that the ceptometer can be used to quantify seasonal LAI dynamics in dry tropical forests. At
Chamela, leaves usually remain on trees during
a 7-8 month period (June to January), but LAIs
are higher than 4 m2 m~ 2 only during 2.5-3.5
months of the year. Leaves also change their
functional properties during the growing season
affecting processes such as photosynthesis, herbivory, gas exchange or rainfall interception.
Further studies are needed to characterize seasonal changes in the quality of the foliage.
180
J.M. Maass etal. /Forest Ecology and Management 74 (1995) 171-180
Acknowledgments
Laboratory and field work help from Alberto
Hernandez, Rocio Esteban, Patricia Centeno,
Julieta Benitez and Ana Isabel Dominguez is
gratefully acknowledged. We also thank Peter
Vitousek, J. Douglas Deans and another anonymous referee for critically reviewing the manuscript; to Alfredo Perez-Jimenez for unpublished
information about site characteristics; and to Estacion de Biologia Chamela (Institute de Biologla-UNAM) for logistic support. Funding for this
study came from DGAPA-UNAM, CONACYT
and from the USDA-FS Coweeta Hydrologic
Laboratory.
References
Bolstad, P.V. and Gower, S.T., 1990. Estimation of leaf area
index in fourteen southern Wisconsin forest stands using
a portable radiometer. Tree Physiol., 7: 115-124.
Bullock, S.H., 1988. Rasgos del ambiente fisico y biologico
de Chamela, Jalisco, Mexico. Folia Entomol. Mex., 77: 517
Bullock, S.H. and Solis-Magallanes, A., 1990. Phenology of
canopy trees of a tropical deciduous forest in Mexico.
Biotropica, 22(1): 22-35.
Burton, A.J., Pregitzer, K.S. and Reed, D.D., 1991. Leaf area
and foliar biomass relationship in northern hardwood
forests located along an 800 km acid deposition gradient.
For. Sci., 37(4): 1041-1059.
Chason, J.W., Baldocchi, D.D. and Huston, M.A., 1991. A
comparison of direct and indirect methods for estimating
forest canopy leaf area. Agric. For. Meteorol., 57: 107128.
Decagon Devices, 1989. Sunfleck Ceptometer User's Manual. Decagon Devices, Pullman, WA, 27 pp.
Galicia, S.L., 1992. Influencia de la variabilidad de la forma
de la pendiente en las propiedades fisicas del suelo y su
capacidad de retention de agua, en una cuenca tropical
estacional. Bachelor Thesis, Facultad de Filosofia y Letras, Universidad Nacional Autonoma de Mexico, 116 pp.
Garcia-Oliva, F., Ezcurra, E. and Galicia, L., 1991. Pattern
of rainfall distribution in the central Pacific coast of Mexico. Geogr. Ann., 73A(3-4): 179-186.
Jarvis, P.G. and Leverenz, J.W., 1983. Productivity of temperate, deciduous and evergreen forests. In: O.L. Land,
P.S. Nobel, C.B. Osmond and H. Ziegler (Editors), Ecosystem Processes: Mineral Cycling, Productivity and
Man's Influence. Vol 12D. Physiological Plant Ecology:
New Series. Springer, New York, pp. 233-280.
Kozlowski, T.T., Kramer, P.J. and Pallardy, S.G., 1991. The
Physiological Ecology of Woody Plants. Academic Press,
New York, 657 pp.
Lott, E.J., 1985. Listados floristicos de Mexico III. La Estacion de Biologia Chamela, Jalisco, Herbario Nacional, In-
stitute de Biologia, UNAM, Mexico, 47 pp.
Lugo, A., Gonzalez-Liboy, J.A., Cintron, B. and Dugger, K.,
1978. Structure, productivity, and transpiration of a subtropical dry forest in Puerto Rico. Biotropica, 10(4): 278291.
Maass, J.M., Jordan, C. and Sarukhan, J., 1988. Soil erosion
and nutrient losses in seasonal tropical agroecosystems
under various management techniques. J. Appl. Ecol., 25:
595-607.
Martinez-Yrizar, A. and Sarukhan, J., 1990. Litterfall patterns in a tropical deciduous forest in Mexico over a fiveyear period. J. Trop. Ecol., 6: 433-444.
Martinez-Yrizar, A., Sarukhan, J., Perez-Jimenez, A., Rincon, E., Maass, J.M., Solis-Magallanes, A. and Cervantes,
L., 1992. Above-ground phytomass of a tropical deciduous forest on the coast of Jalisco, Mexico. J. Trop. Ecol.,
8: 87-96.
Medina, E. and Klinge, H., 1983. Productivity of tropical
forests and tropical woodlands. In: O.L. Lange, P.S. Nobel, C.B. Osmond and H. Ziegler (Editors), Physiological
Plant Ecology IV. Encyclopedia of Plant Physiology, New
Series, Vol. 12D. Springer, New York, pp. 281-303.
Monsi, M. and Saeki, T., 1953. Uber den Lichtfaktor in den
Pflanzengesellschaften und seine Bedeutung fur die
Stoffproduktion. Jpn. J. Bot., 14: 22-52.
Murphy, P.G. and Lugo, A.E., 1986a. Ecology of tropical dry
forest. Annu. Rev. Ecol. Syst., 17: 67-88.
Murphy, P.G. and Lugo, A.E., 1986b. Structure and biomass
of a subtropical dry forest in Puerto Rico. Biotropica,
18(2): 89-96.
Ogawa, H., Yoda, K., Ogino, K. and Kira, T., 1961. Comparative ecological studies on three main types of forest vegetation in Thailand. II Plant biomass. Nature Life Southeast Asia, 4: 49-81.
Parton, W.J., McKeown, B., Kirchner, V. and Ojima, D.,
1992. Century Users Manual. Natural Resource Ecology
Laboratory, Colorado State University.
Pierce, L.L. and Running, S.W., 1988. Rapid estimation of
coniferous forest leaf area index using a portable integrating radiometer. Ecology, 69(6): 1762-1767.
Rzedowski, J., 1978. Vegetation de Mexico, Limusa, Mexico, 431 pp.
Sarukhdn, J. and Maass, J.M., 1990. Bases ecologicas para un
manejo sostenido de los ecosistemas: el sistema de cuencas hidrograficas. In: E. Leff (Editor), Medio ambiente y
desarrollo en Mexico, Porrua-UNAM, Mexico, pp. 81114.
Swift, L.W., Swank, W.T., Mankin, J.B., Luxmoore, R.J. and
Goldstein, R.A., 1975. Simulation of evapotranspiration
and drainage from mature and clear-cut deciduous forest
and young pine plantation. Water Resourc. Res., 11(5):
667-673.
Vose, J.M. and Swank, W.T., 1990. Assessing seasonal leaf
area dynamics and vertical leaf area distribution in eastern white pine (Pinus strobus L.) with a portable light
meter. Tree Physiol., 7: 125-134.