Observed and simulated soil moisture variability over the Lower Mississippi
Delta Region
Georgy V. Mostovoy and Valentine G. Anantharaj
Geosystems Research Institute, Mississippi State University, MS 39762
Journal of Hydrometeorology
(In press)
Geosystems Research Institute, Mississippi State University
PO Box 9652, Mississippi State, MS, 39762, USA
E-mail: [email protected]
Abstract
In order to better understand error and spatial variability sources of soil moisture
simulated with land surface models, observed and simulated values of soil moisture (offline
simulations with the Noah land surface model with four soil layers and approximately 1x1 km²
horizontal resolution were used) were compared. This comparison between observed and
modeled daily values of soil moisture was performed over the Lower Mississippi Delta region
during summer/fall months spanning years 2004 to 2006. The Noah simulations covered
2.5º×2.5º latitude-longitude domain and were forced by the NLDAS atmospheric forcing fields.
Hourly soil moisture measurements and other data, including local meteorological and soil
physical properties data from twelve SCAN sites, were used. It was shown that both the observed
and simulated level of soil moisture depend critically on the specified/sampled soil texture. Soil
types with a relatively high observed clay content (more than 50% of weight) retain more water
due to the low water diffusivity in comparison with silty/sandy soils having 20% or less of clay,
provided that other conditions are the same. This fact is in agreement with previous studies and
implies a strong soil texture control (through related hydraulic parameters, e.g. Richter et al.
2004; Braun and Schädler 2005) on the accuracy of simulated soil moisture. Sensitivity tests
using the Noah model were performed to assess the impact of using the hydraulic parameters
related to the site-specific soil texture on soil moisture quality. Indeed, at some SCAN sites, the
errors (root mean square difference and bias) were reduced. Simulated soil moisture showed at
least 50% reduction when the site-specific soil texture was used in Noah simulations instead of
that derived from the STATSGO data. The most significant improvement of simulated soil
moisture was observed within the top 0-10 cm layer where an original positive bias (an excessive
wetness) was almost eliminated. At the same time, excessive dryness (negative soil moisture
2
bias), which was dominant within second and third model layers was also reduced. These
improvements are expected to be valid at sites/regions with low (less than 0.3) vegetation
fraction.
3
1. Introduction
Soil moisture (SM) is a key variable of the land-atmosphere system, which along with
other environmental variables controls the rate of evaporation from the surface and the
partitioning of the moisture and energy fluxes across land-atmosphere interface. Hence, SM
represents an important input for various environmental and hydrometeorological models. For
that reason, accurate specification and prediction of the SM fields at different spatial scales
(ranging from local to continental) is a practical need for hydrometeorological applications. The
spatial density of available in-situ soil moisture measurements is not adequate to derive reliable
area-averaged SM estimates in the range from 1 km to 10 km grid resolution used for
initialization of mesoscale atmospheric models at regional or continental scales. Therefore, Land
Surface Models (LSM) are commonly used to simulate initial soil moisture states (Crawford et
al. 2000; Chen and Dudhia 2001).
The first LSMs (also referred to as land surface parameterization schemes) were
originated from rather simple but effective parameterization schemes of the land surface and
vegetation processes (e.g. Deardorff 1978). They were developed to describe the lower boundary
condition for the atmosphere in weather prediction and climate models. Based on the complexity
of the physical processes, the various LSMs can be classified into single soil layer, force-restore
(generally having two soil layers, e.g. Bosilovich and Sun, 1995), or multilayer diffusion type
models, as suggested by Shao and Henderson-Sellers (1996).
In virtually all current multilayer LSMs of moderate complexity, the vertical flow of
water within the soil layer of a finite depth is described as a diffusion process (e.g. Capehart and
Carlson 1994; Vitebro and Beljaars 1995; Liang et al. 1996; Smirnova et al. 1997; Irannejad and
4
Shao 1998; Walko et al. 2000; Chen and Dudhia 2001). The vertical distribution of soil moisture
is obtained by a numerical integration of the one-dimensional diffusion-type equation, known as
Richard’s equation in soil (Hillel 2004) and hydrology (Smith 2002) sciences, with specified
water diffusivity (Dw) and hydraulic conductivity (Kw) coefficients and given boundary
conditions at upper and lower boundaries of the soil column. These coefficients control the rate
of the flow of water in the soils, considered as a vertically homogeneous porous media (typically
non-saturated with water), so that the coefficients Dw and Kw depend on the SM content and the
soil texture, which in turn is characterized by the distribution of soil particles sizes. Twelve
major soil classes represented by the USDA soil texture triangle (Hillel 2004) are typically used
in LSM schemes to describe a natural continuum of soil types observed across the globe. A
power law having few empirically-derived parameters is usually assumed to describe the
relationship between Dw and Kw coefficients and the SM (Smith 2002). This manner of analytical
representation of Dw and Kw with constant parameters related to a particular soil texture class
provides a universal and efficient approach for modeling the SM at any geographical location
where the soil type is known. This surrogate methodology is widely applied although the
parameterization of soil hydraulic properties based on a dominant soil texture observed within
the model gridbox suffers from obvious uncertainties. It is evident that one of the restrictions of
this approach is related to the use of a limited number of soil classes to describe a continuum of
soil hydraulic properties. It can be expected that soil texture variations within boundaries of a
particular texture class, observed at the local/field scale, may explain a substantial part of
simulated SM errors. On the other hand, Baker (1978), studying variability of soil hydraulic
properties, reasoned that soils with different textures could be grouped together based on
similarities of their hydraulic properties.
5
Ek and Cuenca (1994) studied the sensitivity of surface fluxes to sample variations of the
hydraulic parameters having a range of one standard deviation within a particular soil texture
class (sandy loam was considered). The largest impact of these variations on surface fluxes was
observed over the bare soil for dry and moderate soil moisture states. Cuenca et al. (1996)
investigated the sensitivity of the surface layer temperature and fluxes with a single-column
atmospheric model integrated for six hours to different shapes of functions describing soil
hydraulic properties. Predictions using functions suggested by Clapp and Hornberger (1978)
were compared with those simulated with hydraulic functions proposed by van Genuchten
(1980), and results of this comparison showed moderate sensitivity of atmospheric surface layer
variables to changes in hydraulic functions.
The research mentioned above and similar sensitivity and validation studies with LSMs
were focused on assessment of the atmospheric surface layer response (Pan and Mahrt 1987;
Sridhar et al. 2002; Marshall et al. 2003) using in-situ measurements. Point SM simulations
performed over a catchment in southeast Australia using the ECMWF land surface scheme
(Richter et al. 2004) revealed that SM bias of this scheme could be reduced substantially by
using point-specific soil hydraulic parameters, including wilting point, field capacity, and
saturated soil conductivity. Simulated SM values may differ from the observed for a number of
reasons. But one of these reasons may be due to the differences between hydraulic properties of
the actual soil texture at the point location and those assigned in the LSM to the corresponding
dominant texture class within the grid cell or element. The importance of soil hydraulic
properties adjustment for better accuracy of simulated SM were demonstrated by Shao and
Irannejad (1999) and Braun and Schädler (2005). These results imply that observed SM biases
6
are often modulated by the uncertainties in the selected soil hydraulic parameters (Richter et al.
2004).
Only a few studies have been devoted toward the direct validation of simulated SM
(Crawford et al. 2000; Robock et al. 2003; Richter et al. 2004). Little is known about the impact
of hydraulic properties on spatial and temporal features of simulated SM although it is well
established that soil porosity is an important factor, which controls the accuracy of the simulated
values (Tischler et al. 2007) and spatial variations of observed (Yoo et al. 1998) soil moisture.
Therefore, the main objective of this study is to examine the impact of the differences between
the locally-sampled and the spatially-aggregated soil texture prescribed from the Soil Survey
data on the accuracy of simulated soil moisture characteristics. It is also anticipated that this
study will also be useful in the interpretation of observed spatial and temporal variability of SM
fields; and thereby contributing toward a better understanding of the role of various factors
influencing the spatial distribution of SM at a regional scale.
Hourly SM observations and physical soil property measurements from twelve Soil
Climate Analysis Network (SCAN) sites located mainly over the Lower Mississippi Delta
Region (SCAN 2007), or simply the Mississippi Delta, spanning the period from January 2004 to
December 2006 were used in this study. The Mississippi Delta was chosen primarily because of
the higher spatial density of SCAN sites located in the region, across the states of both Arkansas
and Mississippi. Retrospective SM simulations were performed using the Noah LSM (Ek et al.
2003) for these sites to assess the impact of using local soil hydraulic properties related to a sitespecific soil texture on the quality of simulated SM. The Noah model, having a moderate level of
complexity, is one of the widely used LSMs by the hydrometeorological community.
7
The rest of this paper is organized as follows. Section two describes SM and other
relevant data available from in-situ measurements at the SCAN sites in our domain. The salient
aspects of SM estimation using the Noah model and validation of simulated SM against SCAN
observations are presented in Section 3. Analysis of observed SM dynamics during dry-down
periods is described in Section 4. Performance of the Noah model with different specifications of
soil hydraulic properties is discussed in Section 5. Section 6 provides summary of the results and
conclusions. It should be noted that a volumetric fraction of the water content expressed in
percents [(water vol.)×100%/(soil vol.)] will be used as measurement units for the SM in this
study.
2. Data description
Soil moisture measurements from twelve SCAN sites (SCAN 2007) located across the
Lower Mississippi Delta Region (including five in the state of Arkansas and seven in the state of
Mississippi) were used for analysis and comparison with Noah LSM simulations. The
geographical distribution of these sites within the study area is depicted in Fig. 1. The SM
volumetric fraction is retrieved from water dielectric constant measured by the Hydra Probe II
SM sensor (Stevens 2007) every hour at depths of 5 cm, 10 cm, 20 cm, 51 cm, and 102 cm.
These depth levels and the layer thicknesses are depicted in Fig. 2b. This current study covers the
3-year period spanning from January 2004 to December 2006.
In addition to soil moisture measurements, the USDA SCAN network also provides sitespecific data about physical parameters of the soils including texture, particle size distribution,
8
water retention, and others. These data represent mean values for soil layers with a different
thickness in the range from about 10 cm to 30 cm, which depends on the location and depth.
3. Soil moisture simulations with Noah LSM
a. Soil moisture prediction
In the Noah model, the SM volumetric content (θ) is predicted by numerical integration
of the following diffusion-type equation (Chen and Dudhia 2001):

 
   K w
 Dw


S ,
t  z 
 z   z
(1)
where S is a SM sink due to water uptake by plant roots; and z is a vertical coordinate oriented
downward from the surface. Equation (1) describes a vertical flow of a fluid with constant
density through soil pores within a rigid, isotropic, and homogeneous soil column (Brutsaert
2006) with Po and E specified at the upper boundary of this column. Po and E represent fluxes of
precipitation (P) minus runoff (R) and water vapor (total evaporation) at the surface,
respectively. They are typically expressed in kg/(m²s) units, or in equivalent units of mm/s. The
water density in (1) is assumed to be 1 g/cm³.
 Dw
According to Eq. (1), both the diffusion flux

and the gravity flux Kw act to redistribute the water within the soil column. It is
z
assumed in (1) that the force of gravity is directed along the z-axis, so in case of rough terrain
9
with a moderate slope, the term involving Kw should be multiplied by cos (δ), where δ is the
terrain slope angle (Capehart and Carlson 1994).
A standard configuration of the Noah model with four soil layers (Chen and Dudhia
2001), with progressively increasing thickness from the top/surface layer, was used in the current
study. Figure 2 shows the location and the thickness of these layers. The three top layers in the
model may contain plant roots (only in these layers S ≠ 0), and the lowest 1-m layer serves as a
water reservoir with a free gravitational drainage condition at its bottom. Soil water diffusion
flux (  Dw

) is assumed to be zero across the lower boundary of the lowest layer. It should be
z
noted that before a numerical discretization of Eq. (1), it is integrated vertically within each
layer, leading to the equation for layer-averaged tendencies of SM content. Further in the text,
layer-averaged values of SM will be denoted by an over bar symbol. To evaluate a vertical
derivative of θ, a linear z-dependence for  values is usually adopted (Mahrt and Pan 1984).
The total evaporation rate E in the Noah model is described as a sum of (a) direct
evaporation from the top soil layer, (b) evaporation of the water intercepted by the canopy, and
(c) evapotranspiration from the canopy. The first two components are linearly proportional to
the potential evaporation (Ep) and (1-f), where f is the green vegetation fraction, and the
evapotranspiration is proportional to fEp. The Ep term, representing an atmospheric evaporative
demand, is evaluated from the surface energy balance using the approach suggested by Penman
(e.g. Mahrt and Ek 1984). Only a fraction β = (1   w ) / ( f   w ) of the rate (1- f)Ep is
available for the direct evaporation from the ground surface in the model, where θw and θf are
soil moisture content at wilting point and field capacity, respectively. Their values depend on the
soil texture class. The β coefficient, aka the soil moisture availability (e.g. Smirnova et al. 1997),
10
provides a simple soil control on the evaporation rate associated with the atmospheric demand
and was initially proposed by Deardorff (1978). The evapotranspiration may vary linearly
between (1- f)Ep (when the soil moisture content ( 1 ) within the top model layer is equal to θf )
and zero if the surface layer soil moisture content is θw . In other words, evaporation from the
soil stops if 1 = θw.
A parameterization for canopy transpiration represented by S was described by Chen and
Dudhia (2001). This parameterization scheme is a rather complex one, and it accounts for the
interaction between the SM and the vegetation described by several parameters such as the
minimal stomatal resistance (rmin), the Leaf Area Index (LAI), and others. Specifically,
evaporation of water intercepted by the canopy is given by fαEp, and canopy evapotranspiration
is expressed as f(1-α)Kv Ep. The coefficient α is represented by a power function of the canopy
water content normalized by the maximum (saturated) capacity (e.g. Chen and Dudhia 2001;
Smirnova et al. 1997). The coefficient Kv is determined by Kv = (ra + Δ)/( ra + rarcCh + Δ),
where ra is related to the resistance of the atmospheric surface layer; Ch is the surface exchange
coefficient; and Δ is a function of dq/dT (q is the air specific humidity at saturation and T is the
air temperature). The canopy resistance rc is defined as rc = rmin/(LAI F), the function F changes
within the range of 0-1 and takes into account influences of solar radiation, water vapor and soil
moisture deficits, and air temperature on rc. More details on F were considered by Chen and
Dudhia (2001).
Direct measurements of Dw and Kw are time and labor intensive; therefore, for practical
purposes of soil water flow modeling, the coefficients Dw and Kw are estimated from a capillary
bundle theory (Mualem 1976). This simplified approach requires an analytical specification of
11
the relationship between θ and soil water potential (ψ), also known as a water retention function
or curve (Smith 2002). In this case, both water diffusivity and hydraulic conductivity are
expressed as a combination of power law functions having three or more parameters describing
their shapes (e.g. Smith 2002).
For each major USDA soil texture class these hydraulic
parameters represent sample mean values with standard deviation of about 10%, obtained by a
least-square fitting procedure of water retention measurements. For example, Rawls et al. (1982)
published mean values and standard deviations of hydraulic parameters for 11 USDA texture
classes based on numerous soil water retention measurements of samples taken from 32 states.
Power law functions with empirically derived parameters reported by Clapp and Hornberger
(1978) are used in the Noah model to describe θ–dependency of the soil hydraulic properties ψ,
Kw, and Dw. Alternatively, these parameters describing the shape of ψ, Kw, and Dw functions can
be obtained by applying an inverse modeling approach. The inverse modeling method is highly
popular in the hydrological community and involves a solution for a minimization problem.
Generally, a root mean square difference between observed and modeled variables, such as
surface fluxes or the ground temperature Hogue et al. (2005), is subjected to minimization. A
recent application of such an approach to find an optimal set of the input hydraulic parameters
into the Noah model was described by Tischler et al. (2007). A marked improvement in error
reduction of the locally simulated SM was reported when the default parameters were substituted
by its optimal values (Tischler et al. 2007).
Finally, it should be mentioned that similar to the SM estimation, the soil temperature is
evaluated from the diffusion equation (Chen and Dudhia 2001) integrated in z over the same
layers as in Eq. (1). The surface skin temperature (Ts) is calculated from the surface energy
balance equation linearized in Ts. The soil temperature is held constant at the lower boundary
12
during integration. Both equations are numerically integrated in time, using an implicit scheme
with a time step equal to 15 min. Pan and Mart (1987) have shown that even for a two-layer
version of the Noah model, truncation errors associated with numerical differencing remained
rather small: not exceeding 10% for the surface evaporation for clay soils if compared with
higher vertical resolution simulations. Bosilovich and Sun (1998) also arrived at similar
conclusions that surface fluxes were relatively less sensitive to the number of soil layers in a
force-restore LSM.
b. Previous validation studies of Noah LSM
The Noah model and its components were rather extensively validated in the past using
various observations obtained during special field experiments, such as FIFE (e.g. Chen et al.
1996). Many of these, as well as some of the recent validation studies of the Noah model have
been focused on comparing the observed and simulated surface fluxes. Most of the recent
validations were based on measurements available from the Oklahoma Mesonet (Brock et al.
1995). They include studies published by Sridhar et al. (2002), Luo et al. (2003), Marshall et al.
(2003), Robock et al. (2003), and Chen et al. (2007).
Sridhar et al. (2002) reported a good performance for the uncoupled Noah model based
on surface fluxes validation over the state of Oklahoma during a one-year period. The Noah
model coupled with an operational forecasting model (Black 1994) was comprehensively
evaluated against the Oklahoma mesonet data by Marshall et al. (2003). They reported SM biases
with a typical range of ±5% for the top 0-10 cm layer during July 1998, but biases exceeding
10% were also observed over some counties (see their Fig. 22). Chen et al. (2007) examined SM
13
biases averaged over the available mesonet stations and revealed almost no bias ( < 1%) for the
top 0-10 cm layer and a substantial negative bias (-15%) for the 10-40 cm layer during the AprilJune 2002 period. They suggested explaining a slight wetness of the upper and excessive dryness
of the lower layer by an inadequate specification of the hydraulic conductivity in the Noah
model.
It is important to note that most of the above validation studies revealed a non-zero bias
of the simulated SM both at individual mesonet stations and for regionally averaged values as
well.
c. Noah/LIS set up
The Noah LSM, available within the state-of-the-art Land Information System (LIS)
developed at NASA Goddard Space Flight Center (Peters-Lidard et al.; 2004, Kumar et al.
2006), was configured at 0.01×0.01 latitude-longitude resolution (about 1×1 km²) over a
domain with an approximate latitude-longitude size of 2.5º×2.5º and covering the Lower
Mississippi Delta region located mainly in the state of Mississippi (see Fig. 1). The LIS provides
a common mechanism for the unified specifications of land topography, soil and vegetation
parameters, and facilitates the running of various LSMs either regionally or globally.
The Noah LSM (version 2.7.1) was used for soil moisture simulations with 4 standard
layers in the soil shown in Fig. 2a. The soil texture was represented by CONUS-SOIL (Miller
and White 1998) data based on USDA STATSGO database. The geographical distribution of
STATSGO soil classes within the Noah/LIS integration domain is shown in Fig. 1. Only five
texture classes (sandy loam, silt loam, silty clay loam, silty clay, and clay) are represented in the
14
model domain, as shown in Fig. 1. Clay soils (silty clay loam, silt clay, and clay) depicted in
blue-green colors are dominant over the Mississippi Delta with a small quantity of sandy soils
observed mainly along the Mississippi River (see Fig. 1). Sandy soils (sandy loam and silt loam)
prevail to the east and west of the Mississippi Delta, so there is a clear contrast of the soil texture
between the Delta and adjacent territories. The vegetation/land use description used in the Noah
model is based on the 13 land cover classification scheme developed at the University of
Maryland.
For retrospective simulations, the LIS framework supports various atmospheric data sets,
including GLDAS, GOES, NLDAS, ECMWF, each with different levels of spatial and temporal
resolutions. The Noah/LIS runs were performed using the North American Land Data
Assimilation System (NLDAS; Cosgrove et al. 2003) forcing for a 3-year period from the
beginning of January 2004 to the end of the year 2006. The input atmospheric forcing data
includes the following surface variables: air temperature and water vapor content, pressure,
components of the wind, downward fluxes of solar and longwave radiation, and rain- snowfall
rates. The NLDAS domain covers the Continental United States (CONUS) region and some
adjacent areas of Canada and Mexico. They are available online from the end of 1996 until the
present with 1/8th latitude-longitude resolution (approximately 15 km grid spacing).
The quality of NLDAS fields was validated against point observations and has proven
high. Luo and coauthors (2003) performed an evaluation study of the NLDAS forcing over the
southern Great Plains, spanning almost a two-year long period and showing that typical standard
deviations between NLDAS and those observed at surface stations variables are 2.3 ºC (for the
air temperature), 1.1 g/kg (for the water vapor specific humidity), 1.5 m/s (for the wind speed),
120 W/m² (for the downward solar radiation flux), and 0.65 mm/hr and 0.15 mm/hr (for the
15
hourly and daily precipitation rates, respectively). The hourly precipitation forcing in NLDAS is
based on daily precipitation reanalysis, which produced from gauge reports at NCEP Climate
Prediction Center. The NLDAS downward solar radiation flux at the surface is retrieved from
NOAA’s GOES satellite measurements. The other NLDAS forcing fields are interpolated both
temporally and horizontally from analysis fields produced by the NCEP Eta Data Assimilation
System at 40 km resolution every 3 hr.
The Noah/LIS model is integrated with 15 min time step. Green vegetation fraction fields
are obtained by space-time linear interpolation of multiyear monthly mean data having
0.15°×0.15° latitude-longitude resolution (Gutman and Ignatov 1998). Other static and
atmospheric forcing fields are also interpolated or aggregated from their native resolution to 1×1
km² cells using the same interpolation/aggregation capability of LIS. Depth-constant initial
conditions (30% for soil moisture and 10°C for soil temperature) are used within entire
integration domain shown in Fig. 1.
d. The dependence of simulated soil moisture on soil texture
The spatial pattern of the simulated soil moisture fields from the top 10 cm layer of the
Noah model is a representative example of the model state soon after a rainfall event (Figure 3a)
and that at the end of a drying period (Figure 3c). These figures reveal a close correspondence
between the spatial patterns of the soil texture depicted in Fig. 1 and those of the SM due to the
SM response to the hydraulic properties (e.g. Robock et al. 2003; Richter et al. 2004). Similar
examples of the spatial distribution of the SM over the part of the same region have previously
been shown by Mostovoy et al. (2007). A careful examination and comparison of Figures 1, 3a
16
and 3c reveal a close correlation between simulated SM and soil type. Areas of relatively low
SM content coincide well with corresponding areas of sandy loam soil depicted by the brown
color in Fig. 1. In fact, marked footprints of the sandy loam are clearly observed in the SM fields
along east/west boundaries of the model domain within a latitudinal zone bounded by 33ºN and
34ºN (see Fig. 3). Conversely, areas of relatively high SM correspond to those of clay soils,
which are dominant over the Delta and shown in Fig. 1 by blue and green colors. Mohr and
coauthors (2000) analyzed in-situ and remotely sensed observations across SW Oklahoma in
1997, and used their results to describe a similar control of the spatial distribution of soil texture
on the spatial distribution of SM fields.
This association between soil texture and SM is well established and supported by
previous empirical studies of SM spatial organization. Hollinger and Isard (1994) analyzed a
multi-year SM time series sampled biweekly during growing season over the state of Illinois at
15 grass-covered sites, and reported soil texture and structure to be the major factors controlling
water storage within 1 m top soil layer. These authors showed that fine-grained and wellstructured, silty-clayey soils with high porosity have twice as much water stored within the top 1
m layer than coarse-grained and poorly-structured sandy soils with relatively low porosity. Other
studies have shown that the spatial organization of soil moisture is controlled by the soil
porosity, which depends on the soil texture (e.g. Rodríguez-Iturbe et al. 1995; Yoo et al. 1998).
Using SM of the top 5 cm layer and field porosity data with 200 m resolution available from the
Washita’92 experiment, Yoo and coauthors (1998) demonstrated that spatial correlation
scales/lengths are about the same (around 2000 m, on average) for both SM and porosity fields.
Results of the above study indicated that soil texture/porosity is more influential in controlling
17
the spatial distribution of SM during drying out periods than rainfall and various landscape
factors, such as a terrain’s slope, orientation and vegetation patterns.
The simulated soil moisture fields of the top 10 cm layer in the Noah model have
distinctly elevated values within the Delta region where soils with relatively high clay content
prevail and, conversely, and lower SM values over the regions to the east and west from the
Delta where silt-loam-sandy soils are dominant. This qualitative agreement can be seen in Fig
3(a) and 3(c). In addition to these marked changes of the soil texture between the Mississippi
Delta and adjacent territories, there is a green vegetation contrast between the Delta (with the
relatively low values of f) and surrounding well-forested territories to the east and west having a
high level of f. Examples of the geographical distribution of the green vegetation fraction (f),
used in the Noah model simulation domain representing the multiyear mean monthly data of
vegetation fraction (Gutman and Ignatov 1998), are depicted in Fig. 3 (b, d) for August and
September.
Longitudinally averaged plots are used to better understand and interpret the
influences of soil texture and vegetation fraction on simulated soil moisture. The simulated
variables from the year 2006 are averaged within one-degree latitudinal band (from 33º N to 34º
N), except for the soil texture, which was aggregated within the same range, and then plotted in
Fig. 4 as a function of the longitude. Figure 4 shows good agreement between observed SM and
the aggregated texture classes along the longitudes. Both during August and October 2006 the
averaged SM values closely respond to soil texture transitions demonstrating elevated level of
SM over the Lower Mississippi Delta where clayey soil types are dominant (they are shown in
Fig. 4 as numbers representing the USDA soil texture classes ranging from 8 /Silty Clay Loam /
to 11 /Silty Clay/ and 12 /Clay/) as compared with regions of lower SM to the east and west
where sandy soil types (corresponding numbers ranging from 3 /Sandy Loam/ to 4 /Silt Loam/)
18
prevail. Note that the soil moisture responds to changes in soil texture in the same way at the end
of drying periods and after precipitation events (see corresponding SM lines in Fig. 4). Similar
features of SM responses were observed during other years (2004 and 2005) having marked
differences in precipitation amounts, as shown in Fig. 5. Soil moisture changes most sharply
along the eastern boundary of the Mississippi Delta where there is an almost discontinuous
decrease in the SM of about 10%. This analysis indicates the importance of soil texture in
maintaining spatial SM gradients both during relatively dry and wet conditions. Additionally,
frames (e) and (f) in Fig. 4 illustrate a total evapotranspiration response to different precipitation
inputs. The total evapotranspiration is lower over the Delta than over adjacent and more
vegetated/forested areas located to the east and west.
Finally, a formal stratification of simulated SM (averaged within one-degree latitudinal
range) according to the soil texture classes gives additional evidence of the soil texture control
on spatial SM patterns. Symbolic distributions (represented by box-plots showing the data range,
upper and lower quartiles, and the median) of the SM values within a particular soil texture class
are plotted in Fig. 6 for a three-year period spanning from 2004 to 2006, showing a rather
gradual increase in the median SM when the soil texture class changes from 3 (Sandy Loam) to
12 (Clay). Because the texture class 11 (Silty Clay) contains a relatively small number of sample
points (twelve), its distribution cannot be considered as statistically representative. The SM
differences across soil texture classes (as illustrated by Fig. 6) are related to the choice of
hydraulic properties (Richter et al. 2004). Overall, these plots support a general notion on the
importance of soil texture in maintaining a particular SM level within top 0-10 cm layer.
e. Comparison with SCAN soil moisture
19
In this section, validation results of SM simulated with the Noah/LIS model against
observations available at 12 SCAN sites over the MS Delta are presented. The Noah model
describes a homogeneous media having physical properties related to a specific soil texture class.
Depending on horizontal/vertical grid size and structure selection for the Noah model, this soil
texture class derived from STATSGO data is represented by one dominant soil type observed
within the model horizontal grid cell and the soil layer depth. In this study the 1-km grid size and
the 2-m soil depth are adopted.
The differences between the soil texture classes locally observed at SCAN sites for the
top 0-10 cm layer and those derived from 1-km STATSGO data are depicted in Fig. 7a. As
mentioned earlier, the soil texture specified in the Noah/LIS SM simulations is a constant with
the depth. Note that locally sampled soil texture data are not available for Earle, AR site. Four
sites among the eleven differ in their local soil types in the top 0-10 cm layer from the
STATSGO soil texture classes. More complete information about depth-variation of locally
observed soil texture types and clay/sand content within four Noah layers is shown in Fig. 7b.
Because the SCAN sampled clay/sand content data are represented by average values for layers
of about 10-cm to 15-cm in depth, they are aggregated vertically to match the thickness of Noah
layers. Silt loam (denoted by number 4) is the most prevalent soil texture, which is
observed/sampled at 7 sites for at least two top Noah layers (see Fig. 7b). At the same time,
relatively high variations of clay and sand ratios are observed among these sites (Fig. 7b),
although they are all categorized into the same texture class (silt loam). Note the high depthvariations of clay/sand content. Since a fixed set of hydraulic properties is assigned to every
texture class, these variations and vertical heterogeneity in sand/clay percentage can cause
20
additional deviations between simulated and observed SM. Considering the problem of clay/sand
variations within limits of texture classes, it might be a more accurate to relate hydraulic soil
properties not to the texture class, but rather directly to clay/sand content, which is also available
from the STATSGO data. Indeed, Cosby et al. (1984) reported a high correlation, exceeding 0.8
on average between texture-mean soil hydraulic parameters (saturated values of soil water
potential, Kw and others) and mean clay/sand content for these textures.
SM biases defined as Noah-simulated minus observed values for May-June, July-August,
and September-October periods of years 2005 and 2006 are depicted in Fig. 8 for 12 SCAN sites
within top three Noah layers. Table 1 illustrates the corresponding RMS differences between
simulated and observed SM for the same periods. Figure 9 shows that a rather high variability of
rainfall amount averaged over the selected sites and its frequency distribution is observed during
these two month periods. During July-August of 2005, for example, 11 significant precipitation
events with daily total exceeding 5 mm were observed and during September-October of the
same year only 3 such cases were registered (see Fig. 9). Except for Lonoke Farm, simulated SM
demonstrates a persistent negative bias (simulated SM is lower/dryer than observed value) within
the second and third Noah layers as shown in Fig. 8. SM bias patterns within the top 0-10 cm are
more variable among selected sites as compared with those of lower layers. In the top 0-10 cm
layer some sites such as Earle, Campus, DeWitt, Lonoke Farm, and N. Issaquena show a
consistent tendency for a positive SM bias. At Vance and Beasley Lake sites a persistent
negative SM bias prevails within the top layer and at Tunica and Perthshire Farm a persistent
positive SM bias (simulated values are too high) is dominant. A small positive SM bias, which
in general is less than 5 % in the top 0-10 cm is typical for Silver City and Good Timber Ck.
sites. SM statistics are not shown within third layer at Good Timber Ck. for 2006 because of a
21
poor quality of data. Overall, a rather good performance of the Noah model expressed in terms of
low values of SM bias and RSM difference within three Noah layers is observed at these two
sites. Note that Silver City has a soil texture gap between local (silt loam) and STATSGO (silty
clay loam) data within two top layers. A relatively small and consistent increase in the clay
content from about 13% (0-10 cm layer) to 28-30% in the 40-100 cm layer and almost no change
in sand content are observed at both of these sites (see Fig. 7). Although similar patterns of
clay/sand distribution are observed at Marianna, DeWitt, and Lonoke Fm., SM biases and RMS
differences are significantly higher at these sites as compared to those at Silver City and Good
Timber Ck. Therefore, there is no apparent link between levels of SM bias or RMS error shown
in Fig. 8 and Table 1 and the observed differences between local and STATSGO soil textures.
Only at Campus and DeWitt sites SM bias and RMS error are dramatically increased within the
third Noah layer (40-100 cm) where the differences between soil textures are locally observed
(see Fig. 8 and Table 1).
It should be noted that the above mentioned positive and negative SM biases, especially
within lower layers, are persistently manifested across various seasons/years having different
distribution and varying rate of precipitation. However, the top 0-10 cm layer shows a tendency
for increased/decreased SM bias and RMS difference during dryer/wetter periods. Comparing
Figs. 8 and 9 (middle frames), it can be concluded that during a relatively wet period of JulyAug. 2005 with precipitation amount of 212 mm, SM biases are significantly lower (except for
DeWitt, Campus, and Marianna sites) than those during July-Aug. 2006, which is a relatively dry
period having 104 mm of total precipitation. Although this tendency of reduced biases and RMS
errors (see Table 1) of SM is observed more clearly in the top layer, it is manifested in lower
22
layers as well. The same response of SM performance statistics to relatively dry/wet conditions
is also observed during Sept.-Oct. 2006 (see lower frames in Figs. 8 and 9).
One of the reasons for the excessive dryness within the second and third Noah layers may
be attributed to the high water diffusivity, which controls the speed of soil water propagation. It
will be shown later in section 5, that use of soil textures with low Dw substantially reduce the dry
drift of simulated SM within the lower layers of the Noah model.
Vegetation-precipitation interactions have proven to be an important factor controlling
surface runoff and SM distribution at seasonal time scales (e.g. Yildiz and Barros 2007). Fig. 10
shows variability of green vegetation fraction averaged for SCAN sites with the same vegetation
type during May-October period. According to the adopted UMD vegetation classification,
SCAN sites are categorized into either of two classes: wooded grassland (Good Timber Ck. and
Campus) or cropland (ten sites). Although the mean vegetation fraction for cropland sites
changes rather sharply during the growing season (from 0.42 in May to 0.78 in August and then
to 0.41 in October, as shown in Fig. 10) there is no an apparent impact of these changes in the
fractional vegetation cover (f) on SM performance statistics (biases and RMS errors), shown in
Fig. 8 and Table 1.
4. Observed soil moisture variability
The correlation between soil texture and simulated 0-10 cm SM patterns observed over
the Lower Mississippi Delta region (described in the section 3d) is fairly well confirmed by insitu SM measurements available over the same region. Fig. 11 shows examples of SM temporal
variability within 1 m top layer at five SCAN sites having different soil texture during years
23
2005 and 2006. For plotting purposes the observed SM, which is measured at five levels, is
linearly interpolated between these levels. Daily mean values of SM are depicted in Fig. 11, and
periods of missing measurements are shown by white bars as well. Local soil texture distribution
with depth sampled at SCAN sites and corresponding soil texture classes derived from
STATSGO data are also shown in Fig. 11.
Two upper frames in Fig. 11 illustrate the SM variability at two SCAN sites located NW
of the Delta at Campus, AR and Lonoke Farm, AR respectively, having silt loam soils. Three
lower frames depict SM time-depth plots for SCAN sites within the Mississippi Delta region;
they are Silver City, MS (silt loam, 4), Beasley Lake, MS (silty clay, 11), and Perthshire Farm,
MS (silty clay, 11). Locally sampled texture classes are indicated in parenthesis. Figure 7 clearly
depicts typical changes in SM dynamics caused by a transition in the soil texture from sandy to
clayey soils. Comparison between the upper frames representing sandy soils and the lower ones
relating to clayey soils supports the importance of soil texture in maintaining a specific level of
SM, particularly during the drying out stages of soils. Indeed, an increase in clay content up to
50% (as shown in two lower frames) leads to a very shallow layer having about 20 to 40 cm in
thickness affected by the surface evaporation and, as a consequence, relatively high SM.
Conversely, relatively low clay content (as illustrated by three upper frames in Fig. 11) produces
a deep (up to 1 m) soil evaporation layer and low SM.
The high correlation or correspondence between both observed and simulated SM and
soil texture suggests that accurate specification of soil texture classes, provided that they
correctly describe the soil hydraulic properties, is important as other factors such as vegetation
fraction and atmospheric forcing for improving the accuracy of the simulated SM and the overall
24
quality of land surface simulations. This inference is especially significant for the drying out
periods of soil matter.
It should be noted that a relatively small region covering approximately area of 2.5º×2.5º
in latitude-longitude, as compared with a typical size of large-scale weather systems
(atmospheric high, lows, and fronts), was considered in this study. Therefore, it would be
reasonable to accept almost horizontally homogeneous atmospheric conditions related to
amounts of total precipitation and evaporation (a component controlled by the atmospheric
demand) within this region. It is also obvious that in addition to the soil texture, horizontal
variations of other local factors, such as water table depth, soil cracks, and macro- and mini
terrain features favoring standing water and ponding, and others can affect locally observed SM
dynamics. These factors have not been considered in this study.
Because all SCAN sites are covered by the short grass with a relatively small locallyobserved f, a vegetation influence on SM is assumed not to be critical in this study. Use of
vegetation parameterizations has proven to be critical for simulations, especially long-term, of
surface fluxes although an impact of these parameterizations on the SM is not well established
and understood (Bosilovich and Sun 1998). Using the ECMWF LSM for multi-years runs over
Australia, Richter et al. (2004) demonstrated little sensitivity of modeled SM to variations of
vegetation parameters, such as f and LAI, in comparison to a relatively large response to changes
of soil hydraulic parameters. Results of our sensitivity tests with the Noah model showed that
activation of the short grass vegetation cover with f = 0.5 resulted in a soil water sink having
approximately constant rate, which mimics a water extraction by plant roots, and located within
three top model layers. The use of vegetation parameterization leads to additional monotonic
decrease of SM in time within second and third layers and the magnitude of this SM lowering is
25
proportional to f. In most test cases, however, use of the vegetation parameterization produced a
little impact on the top 0-10 cm SM in comparison with that in the lower layers. Hence, the bare
soil condition (f = 0) was chosen for SM sensitivity experiments described in the next section.
5. Results of sensitivity experiments
It can be expected from the results described in the previous two sections that SM
simulations with Noah model may be improved at SCAN sites if the STATSGO soil texture is
substituted with locally sampled values.
In order to evaluate the impact of using a local soil texture on modeled SM quality, two
sets of simulations were performed with the Noah model at SCAN sites where differences
between local and STATSGO soil textures were observed. They are both initialized from SM
and soil temperatures observed at SCAN sites. The SCAN data were aggregated vertically to
match to Noah layer thickness, as shown in Fig. 12 for the SM. In a first simulation, referred as a
control run, the soil texture was taken from the STATSGO data base and a second run was
carried out with the local soil texture. Both runs cover a set of two-month periods (May-June,
July-Aug. and Sept.-Oct.) spanning years 2004-2006 with different amount and rainfall
distribution. Since SCAN SM measurements cover only top 0-102 cm, they can only be
compared with SM within the top three Noah layers. Before this comparison, SCAN SM was
cubically interpolated to a vertical grid with 1 cm resolution between five observation levels, and
then the interpolated profile was used to estimate the layer-averaged SM. This procedure is
illustrated in Fig. 12, where the thickness of each layer corresponds to that of the Noah layers.
26
Figure 13 shows a comparison between simulated and observed SM for the top 3 Noah
layers at Perthshire Farm, MS site for three Sept.-Oct. periods for the years 2004, 2005 and 2006.
This particular example provides good evidence of an improvement in the quality SM
simulations in the top 0-10 cm due to the substitution of silty clay loam (the STATSGO texture)
with silty clay (the locally sampled texture). Both RMS difference and bias (estimated as
simulated minus observed value) are significantly reduced as compared with the control run,
shown in Table 2. This improvement is observed for drying out periods and persists during
different years and seasons (see Tables 2 and 3 for May-June and July-Aug.). Another typical
example of simulated SM improvement due to soil texture change (silt loam soil type was
substituted by sandy loam) is shown in Fig. 14 for Campus, AR SCAN site.
Although a certain reduction of the RMS difference and the bias also occurs within
second and third Noah layers, the model failed to reproduce the observed variability of the SM,
as illustrated by lower and middle frames in Figs. 13 and 14. In these layers, the reduced SM bias
may be directly attributed to the corresponding lowering of the water diffusivity (Dw) due to the
substitution of the STATSGO soil texture by observed. Recall that Dw of silty clay (the sample
texture) is smaller than Dw of silty clay loam. Relative changes in Dw and β coefficients due to a
set of typical soil texture transforms (shown in Tables 2 and 3 for five selected SCAN sites) are
depicted in Fig. 15. The coefficient Dw decreases most dramatically (up to twenty times and
more) when the soil texture changes from silt loam (4) to silty clay (11) and from silty clay loam
(8) to silty clay (11). Conversely, the change of soil texture from silt loam to sandy loam (3)
results in relatively small differences of Dw (it changes less then two times). The coefficient β
controls the rate of surface evaporation/transpiration and its changes among the texture classes
are generally smaller in comparison with those of Dw. Therefore Dw is a more important factor in
27
controlling soil water redistribution than the β coefficient, which depends both on SM content at
field capacity and at wilting point. Note that Dw is determined by saturated values of Kw, soil
water potential, and SM (or porosity).
Typical SM biases produced by the Noah model over the Mississippi Delta during spring,
summer, and fall periods have been described in the section 3e and in a previous study
(Anantharaj et al. 2007). Elimination of the original excessive wetness (a positive SM bias) for
the top model layer is an important improvement of the Noah model performance. It can be
expected that this improvement, if confirmed by similar experiments in other geographical
regions, has a positive impact on SM initialization and assimilation studies. Note that the
excessive wetness within top layers of different LSMs has been reported quite frequently in
literature (e.g. Irannejad and Shao 1998; Richter et al. 2004).
Eventhough the temporal variability of the soil moisture in the 0-10 cm top layer is
reasonably simulated, there is still a marked underestimation of the SM during and 10-14 days
after precipitation events (see Fig. 13). This deficiency, inherent in various LSM resulting in a
negative bias of the simulated SM (simulated soil matter is drier than observed), has been
described in earlier studies (e.g. Bosilovich and Sun 1998). Results of our additional sensitivity
tests have shown that this SM lowering could not simply be eliminated by increasing the
precipitation amount available for infiltration into the soil. Alternatively, this deficiency of
simulated SM can be reduced by choosing appropriate function shapes to describe soil hydraulic
properties, as suggested by Braun and Schädler (2005). Some sites, such as Tunica, MS, have a
high degree of vertical heterogeneity of the soil texture. At this SCAN site the soil texture
changes from top to bottom in a following order clay loam, silty clay loam, silty clay, and clay
(locally-sampled data aggregated to match four Noah layers). Other sites have less vertical
28
heterogeneity of the soil texture; but in general, clay content naturally increases with the soil
depth leading to a certain layering of the soil substrate. The quality of simulated SM time series
generally suffers with enhanced vertical heterogeneity of soil texture. However, for a given
profile of soil heterogeneity, the simulated SM may be improved by optimizing the LSM vertical
structure and accounting for the soil layering, as suggested by Irannejad and Shao (1998),
Richter et al. (2004), and Yang et al. (2005). Another emerging approach is to use advanced data
assimilation techniques, such a bias correction scheme to correct model errors using a three
dimensional Ensemble Kalman Filter (EnKF) method, to assimilate a set of in-situ observations
of sufficient spatial density (Luo and Houser 2008; Reichle and Koster 2004). Besides, remotely
sensed information offer the potential to develop new methodologies to properly characterize and
aggregate the soil texture information using artificial neural networks (Zhai et al. 2006) and
calibrate the LSM and optimize the soil hydraulic properties using pedotransfer functions
(Santanello et al. 2007).
Because the results described in this section were obtained for f = 0, additional
experiments demonstrating sensitivity of simulated SM to variations in vegetation fraction were
conducted. Figures 16 (Perthshire Farm) and 17 (Tunica) illustrate sensitivity of SM changes
associated with a transition between soil textures to variation in vegetation fraction during years
2005 and 2006 for Sept.-Oct. period. The maximum response in SM between Noah runs with
different textures is observed for f = 0. Conversely, a complete vegetation cover (f = 1) has a
tendency to produce a relatively small response in SM, especially in the top layer (see upper
frames if Figs. 16 and 17). Overall, Figs. 16 and 17 show that the improvement in SM statistics
related to the use of the local soil texture and discussed at this section is expected to be valid
only for sites/regions having the relatively small vegetation fraction (f ≤ 0.3).
29
The sensitivity of the Noah simulations to errors in atmospheric forcing has also been
investigated. However, only very small differences were observed in the simulated SM between
runs performed with the local atmospheric forcing and those executed with the NLDAS forcing
(results not shown). Typically, these differences are an order of magnitude smaller than that
associated with the substitution of the STATSGO soil texture by the local one. This inference is
in qualitative agreement with the results published by Luo and coauthors (2003) and Robock and
coauthors (2003), showing that the model’s physics adjustments were more significant for
controlling the SM level than reduction of the uncertainty in the atmospheric forcing. It should
be noted that typically the sign of the SM bias differs between the top and lower model layers. In
the present study, a positive SM or an excessive wetness was observed within the top Noah layer
and an excessive drying (a negative SM bias) within the lower layers, as shown in Tables 2 and
3. While the use of the depth-averaged SM for comparison may produce a better bulk agreement
with that of the observed, it will mask SM biases observed within specific layers. Therefore, the
use of layer-by-layer comparison represents more objective and rigorous validation of the Noah
model performance. Hence, it can be concluded that in the Lower Mississippi River Valley,
characterized by clayey soils, the soil texture (through related hydraulic properties) has a greater
influence on improving the accuracy of simulated soil moisture than vegetation and atmospheric
forcing. An adjustment of various hydraulic parameters represents a rather flexible and efficient
way to improve quality of SM predicted by LSMs. However, this is a complex and resource
intensive task, especially for distributed land surface and hydrological models. A complementary
approach is to use remotely sensed SM information, which when applied in conjunction with
data assimilation and parameter optimization techniques, offers a potential for producing
improved SM products at spatially relevant scales using land surface models.
30
6. Summary
Observed and simulated with the Noah model, values of soil moisture (offline
simulations were used with approximately 1×1 km² horizontal resolution) were compared on a
daily basis over the Lower Mississippi Delta region during summer/fall months spanning years
2004 to 2006. Hourly soil moisture measurements and other data including local meteorological
and soil physical properties data from twelve SCAN were used for these comparisons. For
comparison purposes, the SCAN soil moisture data available at 5 levels spanning from 5 cm to
102 cm were aggregated to match to a vertical size of three top Noah model layers having total
thickness of 100 cm. The Noah simulations covered 2.5º×2.5º latitude-longitude domain and
were forced by the NLDAS atmospheric forcing. It was shown that both observed and simulated
levels of soil moisture depend critically on specified/sampled soil texture. Soil types with high
content of clay matter (more than 50% of weight) contain more water due to low water
diffusivity leading to reduced rate of drying in comparison with silty/sandy soils having 20% or
less of clay, provided that other conditions are the same. This fact is in agreement with previous
studies (Mohr et al. 2000; Robock et al. 2003) and implies the importance of prescribing correct
soil texture in order to simulate and assimilate soil moisture accurately.
Therefore, sensitivity simulations were performed to assess the impact of using the sitespecific soil texture in-lieu of using aggregated soil texture data from the STATSGO database.
At some SCAN sites, errors (root mean square difference and bias) of simulated soil moisture
showed at least 50% reduction when site-specific soil texture was used in Noah simulations
instead of that derived from the STATSGO data. A major improvement of simulated soil
31
moisture was observed within the top 0-10 cm layer. The improvements are much more
substantial than those obtained by the substitution of NLDAS forcing by the local atmospheric
data. Elimination of the original excessive wetness within the top model layer (where typically a
positive SM bias is observed) and of the negative SM bias (excessive dryness of the soil matter)
within second and third Noah layers is an important improvement of the Noah model
performance. It can be expected that this improvement valid for relatively low vegetation
fraction (f ≤ 0.3), if confirmed by similar studies in other geographical regions, has a positive
impact on SM initialization and assimilation studies.
Although some reduction of the simulation error was also observed within lower model
layers (second and third), simulated soil moisture produced less temporal variability than
observed data. This particular deficiency of the Noah model can be attributed to several factors
including limitations of the model structure and configuration (assumed soil homogeneity related
to specific choice of hydraulic conductivity and water diffusivity, small number of model layers,
boundary conditions, etc.) and processes influencing the vertical distribution of soil water in
nature (vertical heterogeneity of the soil matter, cracks, hysteresis of wetting-drying cycles,
ponding, and variations of a water table depth from site to site) as well. Overall results of this
study suggest the need to focus on advancing new techniques to account the model biases due to
the uncertainties in the soil hydraulic properties used by land surface models.
Acknowledgements
This research was sponsored by the NASA Applied Sciences Program via NAS13-03032 and by
NOAA Office of Atmospheric Research via NA05OAR4601145 and NA06OAR4600181. We
32
appreciate the timely consultations provided by Garry Schaefer, NRSC-National Water and
Climate Center (Portland, OR) on understanding SCAN datasets. We acknowledge guidance and
consultations provided by Drs. Christa Peters-Lidard and Sujay Kumar, NASA GSFC,
Greenbelt, MD. We also thank Dr. Paul Houser for helpful discussions on land surface modeling
and the data assimilation. The authors greatly appreciate thorough comments provided by the
three anonymous reviewers.
33
REFERENCES
Anantharaj, V., G.V. Mostovoy, C.D. Peters-Lidard, P.R. Houser, and B. Li, 2007: An initial
assessment of soil moisture fields simulated by the Noah Land Surface Model at regional
and local scales., Eos Trans. AGU, 88(23), Jt. Assem. Suppl., Abstract H51D-02.
Baker, F.G., 1978: Variability of hydraulic conductivity within and between nine Wisconsin soil
series. Water Resour. Res., 14, 103-108.
Black, T.L., 1994: The new NMC mesoscale Eta Model: Description and forecast examples.
Wea. Forecasting, 9, 265-278.
Bosilovich, M.G., and W.-Y. Sun, 1995: Formulation and verification of a land surface
parameterization for atmospheric models. Bound.-Layer Meteor., 73, 321-341.
Bosilovich, M.G., and W.-Y. Sun, 1998: Monthly simulation of surface layer fluxes and soil
properties during FIFE. J. Atmos. Sci., 55, 1170-1184.
Braun, F.J., and G. Schädler, 2005: Comparison of soil hydraulic parameterizations for
mesoscale meteorological models. J. Appl. Meteor., 44, 1116-1132.
Brock, F.V., K.C. Crawford, R.L. Elliot, G.W. Cupers, S.J. Stadler, H.L. Johnson, and M.D.
Eilts, 1995: The Oklahoma Mesonet: A technical overview. J. Atmos. Oceanic. Technol.,
12, 5-19.
Brutsaert, W., 2006: Hydrology, An Introduction. Cambridge Univ. Press, New-York, NY, 605
pp.
Capehart, W.J., and T.N. Carlson, 1994: Estimating near-surface soil moisture availability using
a meteorologically driven soil-water profile model. J. Hydrology, 160, 1-20.
Chen, F., K. Mitchell, J. Schaake, Y. Xue, H-L Pan, V. Koren, Q. Y. Duan, M. Ek, and A. Betts,
1996: Modeling of land surface evaporation by four schemes and comparison with FIFE
observations. J. Geophys. Res., 101, 7251-7268.
Chen, F., and J. Dudhia, 2001: Coupling an advanced Land Surface-Hydrology Model with the
Penn State-NCAR MM5 Modeling System. Part I: Model implementation and sensitivity.
Mon. Wea. Rev., 129, 569-585.
Chen, F., and Coauthors, 2007: Description and evaluation of the characteristics of the NCAR
High-Resolution Land Data Assimilation System. J. Appl. Meteor., 46, 694-713.
34
Clapp, R.B., and G.M. Hornberger, 1978: Empirical equations for some soil hydraulic properties.
Water Resour. Res., 14, 601-604.
Cosby, B.J., G.M. Hornberger, R.B. Clapp, and T.R. Ginn, 1984: A statistical exploration of the
relationships of soil moisture characteristics to the physical properties of soils. Water
Resour. Res., 20, 682-690.
Cosgrove, B.A. and Coauthors, 2003: Real-time and retrospective forcing in the North American
Land Data Assimilation System (NLDAS) project. J. Geophys. Res., 108, 8842,
doi:10.1029/2002JD003118.
Crawford, T.D., D.J. Stensrud, T.N. Carlson, W.J. Capehart, 2000: Using a soil hydrology model
to obtain regionally averaged soil moisture values. J. Hydrometeor., 1, 353-363.
Cuenca, R.H., M. Ek, and L. Mart, 1996: Impact of soil water property parameterization on
atmospheric boundary layer simulation. J. Geophys. Res., 101, 7269-7277.
Deardorff, J.W., 1978: Efficient prediction of ground surface temperature and moisture with
inclusion of a layer of vegetation. J. Geophys. Res., 93, 1889-1903.
Ek, M.B., K.E. Mitchell, Y. Lin, E. Rogers, P. Grunmann, V. Koren, G. Gayno and J.D. Tarpley,
2003: Implementation of NOAH land surface model advances in the National Centers for
Environmental Prediction operational mesoscale Eta model. J. Geophys. Res., 108, 8851,
doi:10.1029/2002JD003296.
Ek, M., and R.H. Cuenca, 1994: Variation in soil parameters: Implications for modeling surface
fluxes and atmospheric boundary-layer development. Bound.-Layer Meteor., 70, 369383.
Gutman, G., and A. Ignatov, 1998: The derivation of green vegetation fraction from
NOAA/AVHRR data for use in numerical weather prediction models. Int. J. Remote
Sens., 19, 1533-1543.
Hillel D., 2004: Introduction to environmental soil physics. AP, Elsevier Science (USA), 494 pp.
Hogue, T.S., L. Bastidas, H. Gupta, S. Sorooshian, K. Mitchell, and W. Emmerich, 2005:
Evaluation and transferability of the Noah Land Surface Model in semiarid environments.
J. Hydrometeor., 6, 68-84.
Hollinger, S.E., and S.A. Isard, 1994: A soil moisture climatology of Illinois. J. Climate, 7, 822833.
35
Irannejad, P., and Y. Shao, 1998: Description and validation of the atmosphere-land-surface
interaction scheme (ALSIS) with HAPEX and Cabauw data. Global Planet. Change, 19,
87-114.
Kumar, S.V. and Coauthors, 2006: Land Information System – An interoperable framework for
high resolution land surface modeling. Environmental Modelling and Software, 21, 14021415.
Liang, X., E.F. Wood, and D.P. Lettenmaier, 1996: Surface soil moisture parameterization of the
VIC-2L model: Evaluation and modification. Global Planet. Change, 13, 195-206.
Luo, L., and Coauthors, 2003: Validation of the North American Land data Assimilation System
(NLDAS) retrospective forcing over the southern Great Plains. J. Geophys. Res., 108,
8843, doi:10.1029/2002JD003246.
Luo, Y. and P. R. Houser, 2008: Soil Moisture Data Assimilation with Noah Land Surface
Model Using the Land Information System (LIS). George Mason University, CREW
Technical Report. 14 pp.
Kabat, P., R.W.A. Hutjes, and R.A. Feddes, 1997: The scaling characteristics of soil parameters:
From plot scale heterogeneity to subgrid parameterization. J. Hydrology, 190, 363-396.
Mahrt, L., and M. Ek, 1984: The influence of atmospheric stability on potential evaporation. J.
Climate Appl. Meteor., 23, 222-234.
Mahrt, L., and H.-L. Pan, 1984: A two-layer model of soil hydrology. Bound.-Layer Meteor., 29,
1-20.
Marshall, C.H., K.C. Crawford, K.E. Mitchell, and D.J. Stensrud, 2003: The impact of the land
surface physics in the operational NCEP Eta Model on simulating the diurnal cycle:
Evaluation and testing using Oklahoma Mesonet Data. Wea. Forecasting, 18, 748-768.
Miller, D.A. and R.A. White, 1998: A conterminous United States multi-layer soil characteristics
data set for regional climate and hydrology modeling. Earth Interactions, 2, 1-26.
Mohr, K.I., J.S. Famiglietti, A. Boone, and P.J. Starks, 2000: Modeling soil moisture and surface
flux variability with an untuned Land Surface Scheme: A case study from the Southern
Great Plains 1997 Hydrology Experiment. J. Hydrometeor., 1, 154-169.
Mostovoy, G.V., V. Anantharaj, P.R. Houser, and C.D. Peters-Lidard, 2007: Use of
SCAN observations for validation of soil moisture spatial distribution simulated by the
Land-Surface Model over the Lower Mississippi Delta region. 11th Symposium Integrated
36
Observing and Assimilation Systems for Atmosphere, Oceans, and Land-Surface (IOASAOLS), Proc. 87th AMS Annual Meeting, San Antonio, TX, 5pp., 14-18 Jan., 2007 (CDROM).
Mualem, Y., 1976: A new model for predicting the hydraulic conductivity of unsaturated porous
media. Water Resour. Res., 12, 513-522.
Pan, H.-L., and L. Marhrt 1987: Interaction between soil hydrology and boundary-layer
development. Bound.-Layer Meteor. 38, 185-202.
Peters-Lidard, C.D., S.V., Kumar, Y. Tian, J.L., Eastman, and P.R. Houser, 2004: Global UrbanScale Land-Atmosphere Modeling with the Land Information System. In Proc. 84th AMS
Annual Meeting, Seattle, WA, 13 pp., 11-15 Jan., 2004.
Rawls, W.J., D.L. Brnakensiek, and K.E. Saxton, 1982: Estimation of soil water properties.
Trans. Am. Soc. Agric. Eng., 25, 1316-1320.
Reichle, R. H. and R. D. Koster, 2004: Bias reduction in short records of satellite soil moisture,
Geophysical Research Letters, 31 (19), L19501, doi:10.1029/2004GL020938, Oct 6,
2004
Richter, H., A.W. Western, and F.H. Chiew, 2004: The effect of soil and vegetation parameters
in the ECMWF Land Surface Scheme. J. Hydrometeor., 5, 1131-1146.
Robock, A., and Coauthors, 2003: Evaluation of the North American Land data Assimilation
System over the southern Great Plains during the warm season. J. Geophys. Res., 108,
8846, doi:10.1029/2002JD003245.
Rodríguez-Iturbe, I., G.K. Vogel, R. Rigon, D. Entekabi, F. Castelli, and A. Rinaldo, 1995: On
the spatial organization of soil moisture fields. Geophys. Res. Lettters, 22 (20), 27572760.
Santanello Jr., J.A., C. Peters-Lidard, M.E. Garcia, D.M. Macko, M.A. Tischler, M.S. Moran,
and D.P. Thoma, 2007: Using remotely sensed estimates of soil moisture to infer soil
texture and hydraulic properties across a semi-arid watershed. Rem. Sens. of Env., 110,
79-97.
Shao, Y., and A. Henderson-Sellers, 1996: Modeling soil moisture: A Project for
Intercomparison of Land Surface Parameterization Schemes Phase 2(b). J. Geophys. Res.,
101, 7227-7250.
37
Shao, Y., and P. Irannejad, 1999: On the choice of soil hydraulic models in land-surface
schemes. Bound.-Layer Meteor., 90, 83-115.
Smirnova, T.G., J.M. Brown, and S.G. Benjamin, 1997: Performance of different soil model
configurations in simulating ground surface temperature and surface fluxes. Mon. Wea.
Rev., 125, 1870-1884.
Smith, E.S. 2002: Infiltration theory for hydrologic applications. Am. Geophys. Union,
Washington, DC, 212 pp.
Soil
Climate
Analysis
network
(SCAN),
2007:
USDA
[Available
at
http://wcc.nrcs.usda.gov/scan.]
Sridhar, V., R.L. Elliot, F. Chen, and J.A. Brotzge, 2002: Validation of the NOAH-OSU land
surface model using surface
flux
measurements
in
Oklahoma.
107,
4418,
doi:10.1029/2001JD001306.
Stevens Water Monitoring System, Inc., 2007: The Hydra Probe Soil Sensor (Comprehensive
Stevens
Hydra
Probe
Users
Manual),
41pp.
[Available
online
at
http://www.stevenswater.com/infocenter.aspx.]
Tischler, M., M. Garcia, C. Peters-Lidard, M.S. Moran, S. Miller, D. Thoma, S. Kumar, and J.
Geiger, 2007: A GIS framework for surface-layer soil moisture estimation combining
satellite radar measurements and land surface modeling with soil physical property
estimation. Environmental Modeling and Software, 22, 891-898.
van Genuchten, M.T., 1980: A closed-form equation for predicting the hydraulic conductivity of
unsaturated soils. Soil Sci. Soc. Am. J., 44, 892-898.
Viterbo, P., and A.C.M. Beljaars, 1995: An improved land surface parameterization scheme in
the ECMWF model and its validation. J. Climate, 8, 2716-2748.
Walko, R.L., and Coautors, 2000: Coupled Atmosphere-Biophysics-Hydrology models for
environmental modeling. J. Appl. Meteor., 39, 931-944.
Yildiz, O., and A.P. Barros, 2007: Elucidating vegetation controls on the hydroclimatology of a
mid-latitude basin. J. Hydrology, 333, 431-448.
Yang, K., T. Koike, B. Ye, and L. Bastidas, 2005: Inverse analysis of the role of soil vertical
heterogeneity in controlling surface soil state and energy partition. ). J. Geophys. Res.,
101, D08101, doi:10.1029/2004JD005500.
38
Yoo, C., J.B. Valdes, and G.R. North, 1998: Evaluation of the impact of rainfall on soil moisture
variability. Adv. in Water Resour., 21, 375-384.
Zhai, Y, J. A. Thomasson, J.E. Boggess, and R. Sui, 2006: Soil texture classification with
artificial neural networks operating on remote sensing data. Comp. and Elec. in Agri., 56,
53-68.
39
Table 1. Root mean square difference between simulated and observed soil moisture at SCAN sites (month/year when data become
available are also shown). Upper rows stand for year 2005 and lower for 2006. Layers where locally-observed soil texture class
deviates from that of STATSGO data are shown in gray shading for Sept.-Oct.
Period
Layer
(cm)
Silver
City
02.2004
N.
Issaquena
Beasley
Lake
Vance
09/1999
09/1999
6.8
5.8
6.2
4.5
8.1
7.6
8.6
13.0
8.5
12.1
11.0
13.9
10.2
8.8
9.8
9.8
9.6
13.5
4.1
6.7
6.6
8.8
9.4
8.9
6.6
8.2
8.5
12.7
10.8
13.7
9.1
8.0
7.5
10.5
4.5
10.9
10.2004
0-10
MayJune
10-40
40-100
0-10
JulyAugust
10-40
40-100
0-10
Sept.Oct.
10-40
40-100
4.5
2.4
1.6
2.9
5.3
7.2
3.5
2.3
2.1
6.4
7.4
13.0
5.1
1.9
1.4
5.0
6.4
12.3
6.5
12.3
8.8
9.5
8.8
9.7
5.0
14.6
11.0
16.0
13.2
17.8
8.5
7.4
10.6
14.3
11.2
18.5
Perthshire
F
Tunica
Good
Timber
Marianna
Earle
Campus
DeWitt
Lonoke
Farm
09/1999
01/1999
06/2004
07/2004
02/2004
06/2004
09/1999
8.6
5.4
6.6
9.0
9.6
9.2
6.3
12.5
9.1
9.3
12.8
13.0
7.9
10.0
4.4
6.4
9.8
10.9
4.2
4.0
1.8
4.7
6.5
3.5
4.0
9.1
11.0
13.7
6.6
3.9
7.3
5.9
10.1
-
6.5
5.5
6.2
6.1
10.5
9.5
9.4
7.5
9.7
6.5
13.6
12.3
8.7
9.6
2.3
13.7
9.5
7.4
13.1
12.9
7.4
16.3
12.4
10.5
10.4
9.8
7.0
15.9
11.2
13.2
9.9
7.1
10.2
13.3
12.3
7.1
10.8
5.0
2.4
10.9
11.9
9.6
7.5
5.1
15.7
18.8
23.8
6.8
4.9
5.5
6.3
17.2
13.9
10.0
4.0
4.8
2.9
7.5
8.1
6.5
7.8
7.7
4.9
10.2
13.3
7.3
7.2
8.1
9.4
11.8
14.4
13.4
14.1
3.8
6.0
1.7
2.3
11.0
17.0
7.3
12.3
3.1
4.3
11.0
11.5
2.5
6.3
5.3
4.0
04/2002
7.8
10.6
8.3
8.0
8.6
7.7
6.3
12.1
10.1
13.3
11.5
14.5
7.6
9.0
10.1
12.2
10.1
15.5
Table 2. Bias error (BIAS) and Root Mean Square (RMS) difference between simulated and
observed soil moisture for three top Noah model layers. Upper row stands for control runs
statistics and lower for runs with local soil texture (corresponding change in soil texture from
local to STATSGO is shown by numbers below site name). Statistically significant improvement
for local texture runs as compared to control runs are shown in grey shading.
Site
Layer
Sept.-Oct. 2004
Sept.-Oct. 2005 Sept.-Oct. 2006
(cm)
0-10
Silver
City
8 -> 4
10-40
40-100
0-10
Tunica
10-40
4 -> 11
40-100
0-10
Beasley
Lake
8 -> 11
10-40
40-100
0-10
Perthshire
Farm
8 -> 11
10-40
40-100
0-10
Campus
4 -> 3
10-40
40-100
BIAS
RMS
BIAS
RMS
BIAS
RMS
2.4
1.6
0.1
-2.0
-5.2
-7.6
-4.7
-3.1
-0.5
-3.9
-1.7
-4.8
-7.5
-11.9
-6.8
-3.0
-7.2
-3.7
-0.1
-0.3
0.0
0.5
-5.7
-4.0
4.7
0.0
5.7
0.9
0.0
-4.3
3.9
3.0
2.8
4.4
6.6
9.4
7.4
7.0
4.9
6.7
3.4
6.2
8.1
13.6
7.1
3.9
7.4
3.8
7.6
7.6
6.0
5.3
5.9
4.2
5.9
2.6
7.2
4.5
0.8
4.4
2.6
2.6
-2.9
0.0
-7.5
-4.8
9.0
3.4
0.1
3.6
-5.6
-2.3
-10.9
-16.4
-8.1
-4.0
-7.2
-3.7
5.0
-1.2
-7.7
-3.2
-7.3
-3.6
1.5
-0.1
-4.9
-5.1
2.9
-1.6
3.1
3.6
5.2
2.0
9.6
6.2
10.9
4.8
5.4
6.5
6.4
3.4
11.9
17.4
8.4
4.1
7.6
3.9
7.5
4.5
8.0
3.5
7.6
3.8
4.6
3.9
6.7
6.8
3.0
1.8
1.3
0.9
-1.1
-3.1
-5.5
-7.8
8.5
0.3
-2.7
3.8
-5.9
0.0
-6.2
-12.3
-6.1
-2.1
-7.0
-3.4
6.8
0.2
-6.9
-2.6
-6.8
-3.1
-3.3
-4.7
-11.0
-11.3
3.2
0.4
3.2
2.4
2.1
3.9
5.7
8.1
8.8
2.2
4.5
5.1
6.4
1.7
6.9
13.8
7.5
3.9
7.3
3.6
8.1
2.6
7.6
2.9
7.2
3.3
5.4
6.1
12.1
12.5
3.5
1.1
41
Table 3. Same statistics as in Table 2, but for May-June and July-August and years 2005 and
2006.
Site
Layer
(cm)
0-10
Silver
City
1040
8 -> 4
40100
0-10
Tunica
4 -> 11
1040
40100
0-10
Beasley
Lake
1040
8 -> 11
40100
0-10
Perthshire
Farm
1040
8 -> 11
40100
0-10
Campus
4 -> 3
1040
40100
May-June
2005
July-Aug.
2005
May-June
2006
July-Aug.
2006
BIAS
RMS
BIAS
RMS
BIAS
RMS
BIAS
RMS
-1.1
-0.8
-2.0
-4.8
-5.6
-8.3
6.4
-0.6
-4.5
3.6
-7.7
-0.7
-10.1
-14.8
-5.7
-1.8
4.6
3.2
2.5
5.4
5.8
8.7
7.5
4.2
5.1
6.0
7.8
2.0
11.5
15.8
6.4
3.5
-5.2
-3.1
-3.9
-6.6
-6.0
-8.0
1.9
-1.5
-3.2
3.4
-6.2
-0.2
-11.2
-16.0
-5.8
-2.2
6.7
4.4
3.9
5.7
5.7
7.3
4.7
6.5
6.2
6.0
7.8
1.7
12.6
16.6
6.3
4.4
-2.9
-1.6
-1.4
-5.2
-5.3
-8.6
0.2
-7.5
-9.9
-2.3
-9.0
-2.3
-10.2
-14.1
-2.8
0.6
6.3
4.1
1.6
5.8
5.6
9.1
4.1
8.6
10.3
2.4
9.4
2.5
11.9
15.8
4.3
4.3
-2.6
0.1
-0.2
-4.8
-5.3
-9.1
10.8
2.4
-5.5
3.2
-7.8
-0.3
-15.5
-19.6
-5.3
-2.7
5.6
3.3
1.2
5.5
5.4
9.4
11.3
3.2
6.3
4.8
8.2
1.0
16.5
20.0
6.5
4.1
-7.2
-3.9
1.5
-3.2
-7.9
-4.0
-7.3
-4.0
9.0
0.3
-0.12
-3.6
7.5
4.0
5.8
6.3
8.2
4.0
7.5
4.1
9.3
3.2
1.7
4.1
-7.0
-3.8
-1.3
-7.2
-7.4
-2.7
-7.3
-3.2
13.3
5.1
8.9
6.0
7.2
3.8
6.0
7.8
7.8
3.0
7.6
3.4
14.5
6.4
10.1
8.0
-6.5
-3.6
1.8
-2.3
-7.6
-4.2
-6.7
-3.7
11.8
3.4
3.2
-0.2
6.8
3.7
6.4
7.9
7.9
4.2
7.0
3.8
12.1
5.9
4.7
4.1
-6.7
-3.8
1.4
1.4
-4.0
-2.4
-4.7
-3.0
7.0
2.1
8.0
4.7
6.9
3.9
2.7
2.5
4.1
2.5
4.9
3.1
11.0
4.9
9.3
5.5
-3.6
-6.9
3.9
7.1
-2.1
5.3
2.3
5.4
-4.0
-7.3
4.1
7.4
-0.4
-4.2
1.3
3.3
42
Figure captions
Figure 1. Geographical distribution of USDA soil classes and location of the SCAN sites used
for soil moisture analysis and validation within Noah/LIS integration domain. White
color stands for water bodies.
Figure 2. Soil layer distribution in the Noah model (a) and soil moisture measurements levels at
SCAN sites (b).
Figure 3. Examples of soil moisture (within top 0-10 cm layer) geographical distribution
simulated by the Noah model for August (a) and September (c) 2006. Right frames show
the spatial distribution of vegetation fraction for August (b) and September (d) used for
soil moisture simulations. Stars stand for SCAN sites.
Figure 4. Longitudinal variations of soil moisture simulated in the top 0-10 cm layer by the Noah
model, vegetation fraction, and NLDAS precipitation (a, b) and total evapotranspiration
(e, f) (all averaged within 1-degree latitude belt confined between 33º N and 34º N) and
dominant soil texture (c, d) within the same latitudinal zone across the Lower Mississippi
Delta during July-Aug. (a, c, e) and Sept.-Oct. (b, d, f) 2006. Two sets of soil moisture
values are shown: one (the lower line) corresponds to the end of drying period and the
other (the upper line) represents soil moisture distribution after subsequent rainfall
events. Rainfall and evapotranspiration amounts are converted to daily mean values for
the dates ranges corresponding to the soil moisture simulations.
Figure 5. Examples of soil moisture (within top 0-10 cm layer) and rainfall distribution are the
same as in Fig. 4, but for years 2004 (a, b) and 2005 (c, d).
Figure 6. Symbolic plots of soil moisture (values were averaged within 33º N – 34º N latitude
range) distributions stratified by soil texture type for years 2004 to 2006. July-August (a)
43
and September-October (b). Median, upper and lower quartiles, and data range are
shown. Numbers shown in right lower frame stand for total number of sample points
within specific texture class. Following abbreviations were used for texture classes:
Sandy Loam (SL), Silt Loam (SiL), Silty Clay Loam (SiCL), Silty Clay (SiC), and Clay
(C).
Figure 7. Comparison between local sample soil texture (top 0-10 cm layer) and that of
STATSGO data base (a). Depth-variation of local clay-sand content (b). Numbers below
symbols indicate layer of the Noah model. Symbols having no numbers stand for the
fourth (lower) Noah layer. Solid lines are boundaries between USDA soil texture classes.
Same abbreviations as in Fig. 6 are used to denote soil texture classes; Silt (Si), Loam
(L), Sandy Clay Loam (SCL), Clay Loam (CL), and Sandy Clay (SC).
Figure 8. Soil moisture biases in top three Noah layers at SCAN sites for May-June (upper
frames), July-August (middle), and September-October (lower) two-month periods of
years 2005 and 2006. Small open boxes within lower frames indicate site/layer where gap
between STATSGO and local soil texture is observed.
Figure 9. SCAN and NLDAS daily rainfall amount averaged over 12 SCAN sites (see Fig. 1 for
their location). Standard deviations from mean values are depicted by error bars for
SCAN data. Numbers shown in each frame are, from top to bottom, total number of rain
events (daily total >= 0.1 mm), number of significant events (daily total >= 5 mm), and
total rain amount (mm) within two-month period, respectively.
Figure 10. Variability of green vegetation fraction at SCAN sites during May-October period.
Two sites (Good Timber Ck. and Campus) belong to Wooded Grassland and other ten
sites to Cropland vegetation category. Standard deviations are shown by error bars. Data
44
are interpolated bilinearly to SCAN sites from multiyear vegetation fraction database
(Gutman and Ignatov 1998).
Figure 11. Time-depth plots of soil moisture (a) from Jan. 2005 to Dec. 2006 and local soil
texture (b) at five selected SCAN sites. Note close association between clay content
increase and changes (overall increase) of observed soil moisture.
Figure 12. Vertical interpolation example of SCAN soil moisture measurements (squares).
Cubically interpolated profile (thin line); thick vertical lines stand for layer-averaged
soil moisture; thin horizontal lines indicate boundaries between Noah layers.
Figure 13. Time series of observed and simulated with Noah model daily soil moisture at
Perthshire Farm, MS, SCAN site during Sept.-Oct. period and years 2004 to 2006.
Three upper rows correspond to first, second, and third model soil layers, respectively
(and their location is shown in Fig. 2). Observed daily rainfall amount is depicted in
the lower row by bars.
Figure 14. The same as in Fig. 10, but for Campus, AR, SCAN site.
Figure 15. Relative changes in water diffusivity Dw and coefficient β, associated with switches
between different soil textures, which are indicated by numbers. For example, if soil
texture is transformed from 8 (silty clay loam) to 11 (silty clay), relative change in Dw is
calculated as follows: [Dw (11)– Dw (8)] / Dw (11). Changes in β are estimated in the same
manner. Lines “8 −> 11” and “8 −> 4” from upper frame are duplicated in the lower
frame by closed circles and by thick dashed line, respectively.
Figure 16. Dependence of soil moisture (SM) difference between Noah runs with silty clay (11)
and silty clay loam (8) textures on variations of vegetation fraction represented by 0,
0.3, 0.5, 0.8, and 1. Perthshire Farm, Sept-Oct. 2005 (left frames) and 2006 (right
45
frames). Values of SM(11) minus SM(8) are depicted; 0-0.1m (upper frame), 0.1-0.4m
(middle), and 0.4-1m (lower) layer.
Figure 17.The same as in Fig. 16, but for Tunica and soil moisture (SM) difference between
Noah runs with silty clay (11) and silt loam (4) textures. Values of SM(11) minus SM(4)
are depicted.
46
longitude, W
Figure 1.
47
(a)
|
V
free drainage
(b)
|
V
bottom boundary layer
-800 ----------------------------------------
Figure 2.
48
(a)
(b)
(c)
(d)
Figure 3.
49
(a)
(b)
(c)
(d)
(e)
(f)
Figure 4.
50
(a)
(b)
(c)
(d)
Figure 5.
51
(a)
Figure 6.
52
(b)
(a)
(b)
Figure 7.
53
May-June
July – August
Figure 8.
54
September-October
Figure 8. Continued.
55
May-June
July-August
September-October
Figure 9.
56
Figure 10.
57
(a)
Figure 11.
58
(b)
Soil moisture content, volumetric fraction in %
Figure 11. Continued.
59
Figure 12.
60
Figure 13.
61
Figure 14.
62
Figure 15.
Figure 16.
64
Figure 17.
65