Field-Scale Spatial Variability of Soil Moisture and L-Band Brightness Temperature from Land Surface Modeling

Camille Garnaud Meteorological Research Division, Environment and Climate Change Canada, Dorval, Quebec, Canada

Search for other papers by Camille Garnaud in
Current site
Google Scholar
PubMed
Close
,
Stéphane Bélair Meteorological Research Division, Environment and Climate Change Canada, Dorval, Quebec, Canada

Search for other papers by Stéphane Bélair in
Current site
Google Scholar
PubMed
Close
,
Marco L. Carrera Meteorological Research Division, Environment and Climate Change Canada, Dorval, Quebec, Canada

Search for other papers by Marco L. Carrera in
Current site
Google Scholar
PubMed
Close
,
Heather McNairn Science and Technology Branch, Agriculture and Agri-Food Canada, Ottawa, Ontario, Canada

Search for other papers by Heather McNairn in
Current site
Google Scholar
PubMed
Close
, and
Anna Pacheco Science and Technology Branch, Agriculture and Agri-Food Canada, Ottawa, Ontario, Canada

Search for other papers by Anna Pacheco in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Although soil moisture is an essential variable within the Earth system and has been extensively investigated, there is still a limited understanding of its spatiotemporal distribution and variability. Thus, the objective of this study is to attempt to reproduce the spatial variability of soil moisture and brightness temperature as measured by point-based and airborne remote sensing measurements. To do so, Environment and Climate Change Canada’s Surface Prediction System (SPS) is run at very high resolution (100 m) over a region of Manitoba (Canada) where an extensive soil moisture experiment took place in the summer of 2012 [SMAP Validation Experiment 2012 (SMAPVEX12)]. Results show that realistic finescale soil texture improves the quality of SPS outputs. Soil moisture spatial average evolution in time is well simulated by SPS. Simulated spatial variability is underestimated when compared to point-based measurements, although results are improved when examined domainwide versus comparisons using grid points corresponding to measurement sites. SPS brightness temperature fields compare well with remote sensing data in terms of spatial variability. It is shown that during drier periods, factors other than soil texture become important with respect to soil moisture spatial variability. However, during periods with plenty of precipitation, soil texture seems essential in improving simulated soil moisture spatial variability at high resolutions. These results support the conclusion that SPS could provide very high–resolution soil moisture products for research and operational purposes if high-resolution soil texture and vegetation products are made available on a larger scale.

Denotes content that is immediately available upon publication as open access.

For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author e-mail: Camille Garnaud, camille.garnaud@canada.ca

Abstract

Although soil moisture is an essential variable within the Earth system and has been extensively investigated, there is still a limited understanding of its spatiotemporal distribution and variability. Thus, the objective of this study is to attempt to reproduce the spatial variability of soil moisture and brightness temperature as measured by point-based and airborne remote sensing measurements. To do so, Environment and Climate Change Canada’s Surface Prediction System (SPS) is run at very high resolution (100 m) over a region of Manitoba (Canada) where an extensive soil moisture experiment took place in the summer of 2012 [SMAP Validation Experiment 2012 (SMAPVEX12)]. Results show that realistic finescale soil texture improves the quality of SPS outputs. Soil moisture spatial average evolution in time is well simulated by SPS. Simulated spatial variability is underestimated when compared to point-based measurements, although results are improved when examined domainwide versus comparisons using grid points corresponding to measurement sites. SPS brightness temperature fields compare well with remote sensing data in terms of spatial variability. It is shown that during drier periods, factors other than soil texture become important with respect to soil moisture spatial variability. However, during periods with plenty of precipitation, soil texture seems essential in improving simulated soil moisture spatial variability at high resolutions. These results support the conclusion that SPS could provide very high–resolution soil moisture products for research and operational purposes if high-resolution soil texture and vegetation products are made available on a larger scale.

Denotes content that is immediately available upon publication as open access.

For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author e-mail: Camille Garnaud, camille.garnaud@canada.ca

1. Introduction

Soil moisture is an essential variable within the Earth system. It regulates the evapotranspiration rate from the ground and the biosphere, thus affecting the energy and water balance at the surface–atmosphere interface. Its effect on the partitioning of net radiation into sensible and latent heat has an impact on boundary layer development. Thus, in some regions, precipitation can be strongly constrained by evapotranspiration and thus soil moisture (Eltahir 1998; Betts and Viterbo 2005). Moreover, spatial heterogeneity in soil moisture, through its influence on differential boundary layer structures, can generate mesoscale atmospheric circulations (Seneviratne et al. 2010; Small 2001). On longer time scales, several studies have shown that land water storage induces a “memory” to the Earth system (Seneviratne et al. 2006; Lorenz et al. 2010), which is crucial to take into account in subseasonal and seasonal forecasts (Conil et al. 2007; Koster et al. 2010).

Owing to the impact of soil moisture on the atmosphere on different time scales, this variable is essential for delivering quality meteorological and environmental forecasts, including flood and drought forecasts. However, since soil moisture interacts with a range of processes (precipitation, evapotranspiration, land surface energy, and water fluxes, etc.) and is affected by numerous factors (orography on all scales, vegetation, and soil texture, etc.), it exhibits significant spatiotemporal variability across different scales. Ample efforts have been made in the scientific community to better understand such a complex variable (e.g., Teuling and Troch 2005; Brocca et al. 2007; Mittelbach and Seneviratne 2012; Li and Rodell 2013; Wang et al. 2015). Nonetheless, there is still a limited understanding of soil moisture spatiotemporal distribution and variability, particularly at finer scales. Thus, because of the lack of reliable large-scale, fine-resolution observations, land surface models (LSMs) with observation-based forcing are efficient tools to study soil moisture, even though they are greatly dependent on forcing data quality and discrepancies have been found between models (Seneviratne et al. 2010).

The demand for high-resolution soil moisture products is high. For agricultural purposes, for example, detailed soil moisture distribution is essential (Champagne et al. 2012; McNairn et al. 2012; Crow et al. 2012a). Within the hydrological context, local effects influencing runoff and infiltration are crucial to the quality of different water cycle variables, including river flow, as well as drought monitoring (Manns et al. 2014; Wood et al. 2015). Furthermore, atmospheric models are starting to be run at very high resolutions (100–250 m) for special events such as the Vancouver Winter Olympics, the Pan American Games, and for other research purposes (Thompson et al. 2007; Leroyer et al. 2011; Chen et al. 2011; Leroyer et al. 2014). In these cases, the atmospheric model requires very high–resolution initial conditions, but it is still too costly to run a land surface assimilation system at these resolutions. Thus, very high–resolution LSM products could prove to be very useful in the downscaling of lower-resolution assimilation products. Furthermore, high resolution can be used as support during field campaign setups and calibration–validation (cal–val) efforts within the context of spatial remote sensing missions. Last, depending on this study’s results, the creation of high-resolution observed soil texture maps on a larger scale could be justified. Fortunately, such maps are already available from Agriculture and Agri-Food Canada (AAFC) for all agricultural regions of Canada (mostly southern parts of Canada).

Environment and Climate Change Canada (ECCC; formerly known as Environment Canada) has developed the Surface Prediction System (SPS), and its newest land surface scheme, Soil, Vegetation, and Snow (SVS), has recently been tested in a stand-alone approach with respect to soil moisture and at different resolutions. Work is underway to implement in experimental mode a 250-m version of SPS to downscale numerical weather prediction outputs from a 2.5-km atmospheric model and to downscale land data assimilation data used to initialize local atmospheric model runs at high resolutions (100–250 m). This newest development further enhances the need for an evaluation of the LSM’s simulation of soil moisture.

In Alavi et al. (2016), the performance of SVS runs at a 10-km resolution is evaluated against the Interactions between Surface, Biosphere, and Atmosphere (ISBA) scheme, as well as a large set of in situ and brightness temperature (TB) data from the Soil Moisture and Ocean Salinity (SMOS) satellite over North America. SVS was found to yield time evolution more accurately and higher correlations with observations, as well as reduced errors. Furthermore, Alavi et al. (2016) point out that at this resolution (10 km) SVS is not affected significantly by different soil datasets, but is mostly affected by the different vegetation datasets.

At 100-m resolution, Garnaud et al. (2016) showed that SPS with SVS lacked spatial variability in soil moisture compared to a ground measurement network data (Brightwater Creek Network, Saskatchewan, Canada). The model was run with a uniform soil texture throughout the study domain, and in this case the lack of detailed soil texture information was thought to be a contributing factor in the low spatial variability of the SPS output.

The research presented here is the continuation of Garnaud et al. (2016) in that it attempts to enhance the quality of SPS with SVS, particularly with respect to soil moisture variability, but it also evaluates the model from two different points of view: ground-based and airborne observations. Indeed, airborne or satellite radiometer measurements of L-band TB, which are known to be particularly sensitive to soil moisture, are now commonly used to evaluate land surface models (de Rosnay et al. 2009; Albergel et al. 2012; Parrens et al. 2014). Thus, SPS is run once again at 100-m resolution but over a region of Manitoba (Canada) where an extensive soil moisture experiment took place in 2012 [Soil Moisture Active Passive (SMAP) Validation Experiment 2012 (SMAPVEX12); McNairn et al. 2015]. At this site, detailed soil texture data are available at high resolution, as well as regular ground observations and airborne radiometer measurements of L-band brightness temperature. Despite the fact that it may strike some people as an obvious result, there is not a great deal of objective evidence that highly distributed soil texture maps actually lead to improved representations of finescale soil moisture in land models. This is something that high-resolution land modelers commonly assume but is almost never actually demonstrated.

2. Models and methods

a. Surface Prediction System

SPS (also known as GEM-Surf; Leroyer et al. 2011; Bernier et al. 2012; Rochoux et al. 2016) is used operationally and for research purposes at resolutions ranging from 25 km to 100 m. SPS consists of four main components: 1) land schemes to simulate the temporal evolution of the soil and surface processes, 2) atmospheric forcing (i.e., surface pressure, near-surface air temperature and humidity, wind, solar and longwave radiation, and precipitation), 3) geophysical land surface characteristics, and 4) initial conditions for the surface models (e.g., soil moisture and surface temperature).

SPS is rather unique in the way spatial downscaling is performed. This is achieved in SPS by modification of the atmospheric forcing based on high-resolution orography (not a significant factor in this study) and by providing surface fields (orography, vegetation, and soils) as accurately as possible at the local scale. This approach has been shown to perform quite well in mountainous and urban areas (e.g., Bernier et al. 2011; Leroyer et al. 2011).

In SPS, the surface heterogeneity is represented at subgrid scales as an aggregation of different covers (i.e., land, urban, water, continental ice, and sea ice) as shown in Fig. 1. Each type of cover (or tile) features specific input parameters and parameterization schemes to simulate the related physical processes (Rochoux et al. 2016; Molod and Salmun 2002; Essery et al. 2003). In this study, the land surface modeling component of SPS is the SVS scheme, which is described in Alavi et al. (2016) and Husain et al. (2016). This scheme is currently developed and tested as a replacement to the ISBA scheme (Noilhan and Planton 1989) in the Meteorological Service of Canada’s (MSC) operational numerical environmental prediction applications.

Fig. 1.
Fig. 1.

Schematic of the modeling system. SPS is driven by hourly downscaled forecast fields from the 15-km version of GEM. Forcing fields are temperature T, humidity q, and winds (U, V) at the atmospheric model’s first level (~40 m above the surface) as well as surface pressure P0, radiation (rad), and rate of precipitation (Pr). In SPS, the surface heterogeneity is represented at subgrid scales as an aggregation of different covers (i.e., land, urban, water, continental ice, and sea ice). Adapted from Bernier et al. (2012).

Citation: Journal of Hydrometeorology 18, 3; 10.1175/JHM-D-16-0131.1

Compared to the Canadian implementation of ISBA (Bélair et al. 2003a,b; Carrera et al. 2010), SVS calculates the energy and water budgets for vegetation, two snow packs, and ground underneath vegetation and snow with a new tiling approach. The scheme also has improved parameterization of the vegetation thermal coefficient and includes a photosynthesis process in order to evaluate the surface stomatal resistance. There are new formulations for land surface albedo and emissivity and a new snowpack under vegetation. SVS also calculates root density functions depending on the vegetation type and multilayer water vertical transport in the soil, for each of its seven soil layers. In SVS, the soil moisture is represented in each of these layers based on vertical transport, evapotranspiration, and surface/lateral flows.

b. Atmospheric forcing

Atmospheric forcing data from ECCC’s Regional Deterministic Prediction System (RDPS), the 15-km version of the regional Global Environmental Multiscale (GEM) model (Mailhot et al. 2006), is used to drive the evolution of the land surface state simulated by SPS in external mode. Downwelling shortwave and longwave radiation and surface pressure are taken at the surface, while air temperature, specific humidity, and wind are taken from GEM’s lowest vertical level, which is roughly at 20 m for temperature and humidity and 40 m for wind. As detailed in Bernier et al. (2012), the downscaling procedures of forcing data are embedded within SPS, including interpolation to the high-resolution surface of air temperature, surface pressure, and specific humidity.

With respect to precipitation, the forcing data come from the Canadian Precipitation Analysis (CaPA; Mahfouf et al. 2007). CaPA combines a short-range 6-h RDPS precipitation forecast with available precipitation gauge observations using an optimum interpolation (OI) methodology.

c. Geophysical fields

Since SPS is run with a 100-m grid spacing, it is important to describe the surface geophysical state with the greatest accuracy possible. Several databases from different sources are used. For crops, the fractional cover of each type during the summer of 2012 was obtained from AAFC crop inventory (http://open.canada.ca/data/en/dataset/ba2645d5-4458-414d-b196-6303ac06c1c9; contains information licensed under the Open Government Licence of Canada).

The seasonal cycle of the roughness length z0 for each crop type was derived from crop height measurements taken throughout the 2012 growing season in the study domain (see section below). The seasonal cycle of the leaf area index (LAI) was obtained from the Canola Council of Canada for canola, from United Kingdom’s Home Grown Cereals Authority for wheat, and from Setinoyo et al. (2008) for soybean. Maximum rooting depth of each crop type, which remains constant throughout the simulation, was derived from Cutforth et al. (2013) and Dwyer et al. (1988). For other vegetation types, mainly deciduous broadleaf trees, evergreen broadleaf shrubs, and grass, fractional cover comes from the European Science Agency’s GlobCover (version 2.3, 2009) at 300-m resolution, and the LAI, rooting depth, and z0 are obtained from lookup tables.

The high-definition spatial distribution of vegetation, along with the orography and the geographical location of the study domain are shown in Fig. 2. The orography of the region is obtained from the Canadian Digital Elevation Data 1:50 000 (CDED50) geospatial dataset with a 20-m resolution over Canada. The soil composition comes from the AAFC soil survey data at a 316-m resolution obtained on the NSIDC website (http://nsidc.org/data/docs/daac/smap/smapvex12/sv12stm/). The soil data were then interpolated to a 100-m resolution to fit our grid using the nearest-neighbor method to avoid averaging. Figure 3 shows the soil content of sand and clay over the domain.

Fig. 2.
Fig. 2.

(a) Orography (m) of the study domain along with the location of the 102 measurement sites from SMAPVEX12 (black filled circles), (b) the geographical location of the study domain, and (c) its primary land-cover type at each grid point.

Citation: Journal of Hydrometeorology 18, 3; 10.1175/JHM-D-16-0131.1

Fig. 3.
Fig. 3.

Percentage of (a) sand and (b) clay in soil coming from the AAFC soil survey data.

Citation: Journal of Hydrometeorology 18, 3; 10.1175/JHM-D-16-0131.1

d. Experimental design and verification dataset

SPS was run in external mode from 1 March to 31 July 2012 centered near Elm Creek, Manitoba, Canada, with a 100-m grid spacing over a 13 km × 70 km domain and with a 10-min time step. Forcing data, as described above, are given to SPS with an hourly time step, and thus they are interpolated to the model time step. The analysis presented in this study focuses on the 43-day period, from 7 June to 19 July 2012.

The time period and domain correspond to that of SMAPVEX12, a joint Canada–U.S. intensive measurement campaign to support the development, enhancement, and testing of SMAP soil moisture retrieval algorithms (McNairn et al. 2015). During this field experiment and several times per week, NASA flew two aircraft, the Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) and another with a Passive/Active L-band Sensor (PALS). In this study, the PALS data are used as they provide radiometer products such as vertically and horizontally polarized brightness temperatures over the entire study domain at a 1500-m resolution.

Alongside the airborne data acquisitions, ground crews collected soil moisture data and vegetation measurements on 55 annual and perennially cropped agricultural fields and at four forest sites, as shown by the black dots in Fig. 2a. This figure, depicting the orography of the region in the background, shows the lower altitude in the southeastern corner where soils with high clay content dominate. Out of the 102 available soil moisture sample points, 97 were used in this study, as the other 5 are outside the study domain. The 97 sample points include 6 measurements originating from permanent in situ stations operated by AAFC [the Real-Time In Situ Soil Monitoring for Agriculture (RISMA) network] and 37 from temporary in situ stations installed by the U.S. Department of Agriculture (USDA) for the duration of the field experiment. The remaining soil moisture data represent the field average of measurements collected by crews on 54 SMAPVEX12 experimental fields, near coincident with PALS overflights.

Garnaud et al. (2016) showed that in a similar setup but over a domain in Saskatchewan, SPS lacked in spatial variability, possibly due to a uniform soil texture throughout the study domain. Thus, in order to enhance the quality of the model at such high resolution with respect to spatial variability, two simulations of SPS are performed in this study, with the second aiming an additional realistic increase in spatial variability.

SPS_1 uses a standard configuration with a homogeneous soil texture as given by the Soil Landscapes of Canada dataset (version 2.1) of Agriculture Canada, which has a 10-km resolution. This setup is identical to the one in the previous study by Garnaud et al. (2016). SPS_2 differs from SPS_1 only in that it uses soil texture coming from the high-resolution database created by the AAFC soil survey, as described above. It will yield information on the impact of soil texture on the simulated soil moisture.

The outputs of both simulations are then evaluated against in situ data, as well as airborne remote sensing data from PALS. Indeed, brightness temperatures are commonly used to evaluate soil moisture simulated by land surface models (de Rosnay et al. 2009; Albergel et al. 2012; Parrens et al. 2014), since at L band the microwave emission measured by a radiometer is mostly sensitive to soil moisture in the top 5 cm of soil (Entekhabi et al. 2010). To compare SPS-simulated results to PALS brightness temperature data at an incidence angle of 40°, the Community Microwave Emission Modeling (CMEM) platform, developed by the European Centre for Medium-Range Weather Forecasts (ECMWF; de Rosnay et al. 2009; Drusch et al. 2009; Sabater et al. 2011), is used as a forward model to generate L-band TBs from SPS-simulated land surface states (i.e., soil moisture content, soil temperature, snow mass, snow density, water bodies surface temperature, vegetation type and cover, soil texture, and topography).

CMEM is a modular code that includes a choice of several parameterizations or submodels. In this study, CMEM submodels were chosen to follow Carrera et al. (2015) and are shown in Table 1. To facilitate comparison, these 100-m TBs are further upscaled to 1-km resolution by taking a linear average within each 1 km × 1 km pixel, while PALS data are downscaled from 1.5- to 1-km resolution using the nearest-neighbor method. Since the airborne measurements were made early in the mornings, CMEM outputs are taken at 0700 local time for better comparison. Thus, CMEM allows us to move into TB space in order to evaluate SPS-simulated soil moisture from another point of view, that is, airborne radiometer measurements, which has the advantage of covering the entire domain (vs in situ data that are very localized). Indeed, no observations or measurement systems provide information on soil moisture at the resolution and coverage needed. In this particular study, airborne measurements are available and provide remote sensing (TB) measurements that have the right resolution and coverage. Brightness temperature is not soil moisture but it can be linked to soil moisture observations; thus, at this point in time, it is the best we can do to evaluate objectively the model soil moisture outputs.

Table 1.

CMEM configuration.

Table 1.

Simulated data are analyzed using various statistics throughout the results section. To strengthen the analysis, the bootstrapping method, which allows assigning measures of accuracy to sample estimates, is used. Confidence intervals (CI) of 95% are thus calculated using 1000 samples.

3. Results

The model sensitivity to both configurations is shown in this section. The study period covering from 7 June to 19 July 2012 is analyzed in its entirety, corresponding to the duration of SMAPVEX12. Moreover, three days were chosen throughout this period (12 June, 27 June, and 13 July) on which NASA flew the aircraft with PALS as examples to better visualize spatial variability.

a. Soil moisture

Figure 4 shows the impact of the two configurations on estimates of soil moisture in the top soil layer (0–5 cm) for 12 June, 27 June, and 13 July. In situ soil moisture is represented as circles, with the color of each circle indicating measured moisture. The transition from homogeneous (SPS_1) to heterogeneous soil texture (SPS_2) results in spatial variability in bare ground albedo as simulated by SVS. This leads to horizontal gradients in latent and sensible heat fluxes (not shown), as well as in soil moisture. Sandier textured soils located approximately west of 98°W, as seen in Fig. 3, drain more quickly relative to the heavy clays located in the southeast. The first day, 12 June, is different from the other two days since no rain fell on this particular day, yet heavy rain occurred on the previous day. For the other days, 27 June is during a period of dry-down while 13 July is in a period of low precipitation rates.

Fig. 4.
Fig. 4.

Soil moisture (m3 m−3) on three days as simulated by SPS with two different configurations. The color-filled circles are observations collected during SMAPVEX12, with colors corresponding to the same color scale as the simulated data.

Citation: Journal of Hydrometeorology 18, 3; 10.1175/JHM-D-16-0131.1

Spatial correlations between simulated data and observed data at the 97 measurement sites are shown in Table 2. The high-resolution soil texture in SPS_2 clearly improves the spatial correlation between simulated and observed data on these three days. The greatest improvement is when the ground is the wettest on 12 June; the correlation of SPS_1 with observations is very small (0.03) with the CI including the zero correlation, while SPS_2 shows a correlation of close to 0.70. This improvement is most obvious where the region of heavier clays transitions to lighter sandy and loam soils around 98°W (Fig. 3). Note that the darker blue patches in the northwestern part of the domain in both simulations in Fig. 4 correspond to forested areas. Forest cover tends to promote retention of near-surface soil moisture longer than crops because of greater shading and deeper roots.

Table 2.

Spatial correlations between simulated and observed data at the 97 measurement points on different days. The 95% CI calculated using the bootstrapping method is given in parentheses.

Table 2.

To better visualize the spatial variability of each SPS simulation compared to in situ observations, Fig. 5 shows the scatterplot of observed data versus model data at grid points that correspond to measurements sites. Note that the comparison is made between very localized measurement, even though 54 of the 97 sample sites are actually field averages, and a simulated 10 000 m2 average (i.e., the area covered by a 100 m × 100 m grid point). Nonetheless, these plots give a good representation of the simulated spatial variability.

Fig. 5.
Fig. 5.

Soil moisture (m3 m−3) scatterplot of simulated vs observed data for the two SPS simulations on three days during the study period.

Citation: Journal of Hydrometeorology 18, 3; 10.1175/JHM-D-16-0131.1

The low spatial variability associated with the output from SPS_1, as displayed in Fig. 4, and its low correlation to observed data (Table 4, described in greater detail below) are particularly noticeable in its corresponding scatterplots (Fig. 5). These figures show a clear improvement with the use of varying soil texture, but the scatterplots of SPS_2 still show clustering of points. This signifies that parameterization in SVS or other factors such as subgrid orography and vegetation are still affecting SPS outputs in terms of soil moisture spatial variability.

These three days were merely samples of the study period. Thus, in order to visualize the improvements made by the different simulation configurations with respect to spatial variability over the course of SMAPVEX12, Fig. 6 shows the time evolution of the spatial standard deviation of observed and simulated soil moisture. In Fig. 6a, this statistic is based on all grid points, while in Fig. 6b it is based on the 97 measurement locations only. The observations show a high spatial variability during crop growth in June followed by a decrease in variability during crop maturity in July. Although both simulations underestimate spatial variability, the trend seen in the observations is best replicated by SPS_2 in Figs. 6a and 6b, while SPS_1 simulates opposite trends.

Fig. 6.
Fig. 6.

Soil moisture spatial std dev (m3 m−3) evolution in time from 7 Jun to 19 Jul 2012 of two SPS simulations (blue and red with corresponding 95% CIs) and observations (black). (a) Simulated domainwide std dev compared with observations and (b) the spatial std dev of grid points corresponding to measurement sites only compared with observations.

Citation: Journal of Hydrometeorology 18, 3; 10.1175/JHM-D-16-0131.1

The shaded regions in both graphs show the CI around the simulated data. The CIs in Fig. 6a are barely visible since the simulated data include 91 000 grid points, meaning that the statistics on such a large sample are very robust. Interestingly, the CIs in Fig. 6b are largest during the driest period also corresponding to crop maturity in early July, particularly in SPS_1.

It is also interesting to note that, in general, spatial variability simulated by SPS is higher, and thus closer to that observed, when calculated with the entire domain than with 97 grid points corresponding to measurement sites. Indeed, SPS_2’s bias is about 33% domainwide and about 50% when looking at measurement sites only. Differences are expected between the two approaches since we compare an entire domain with a few grid points, and agreement between the simulations and observations depends heavily on how well the distribution of ground-data sites reflects the domainwide average. Since most of these grid points are close together and not necessarily representative of the entire domain and soil moisture is an autocorrelated field, it explains why domainwide data give better results than grid points corresponding to in situ measurements.

During the study period it rained frequently until 22 June, which explains the low spatial variability in SPS_1 during that period, since soil moisture was constantly and spatially evenly very high. From the end of June until the end of the simulation, rain was sparse and soil moisture variance increases because of other elements impacting soil moisture when it is relatively low, such as vegetation, slope, etc.

The temporal evolution of soil moisture of both simulations compared to measured soil moisture along with their CIs is shown in Fig. 7. In Fig. 7a, the domainwide spatial average is used for both SPS simulations, while in Fig. 7b grid points corresponding to the 97 measurement locations are averaged. Observed data are the spatial average of the 97 SMAPVEX12 measurement sites in both cases. For the two simulations, the mean absolute bias, mean bias, standard deviation of the error, and Pearson correlation coefficient with respect to measured soil moisture are shown in Table 3.

Fig. 7.
Fig. 7.

As in Fig. 6, but for spatial average.

Citation: Journal of Hydrometeorology 18, 3; 10.1175/JHM-D-16-0131.1

Table 3.

Temporal statistics between the two simulations’ spatial average (domainwide and measurement sites) and in situ observations for 0–5-cm soil moisture. The 95% CI calculated using the bootstrapping method is given in parentheses.

Table 3.

Figure 7 shows a general wavelike trend in average soil moisture with a peak between 12 and 24 June and a trough in early July. This trend is well simulated by both simulations, although SPS_1 tends to overestimate soil moisture with a near-constant bias. In periods of dry-down, such as during the end of June and beginning of July, both simulations tend to overestimate water loss from the top soil layer compared to observations. This could be the result of faulty parameterizations of shading, evapotranspiration, and infiltration in periods of water stress.

Once again, the CIs from the bootstrapping method are negligible when used on the domainwide data, but they remain relatively small when used on measurement sites only, acknowledging the robustness of these results. Similarly to previous results, both simulations compare better to observations when using domainwide data, with high-resolution soil texture noticeably improving the soil moisture spatial mean simulated by SPS_2 compared to SPS_1. This is reflected in the mean error of SPS_2 that is close to zero (Table 3).

Modeling and ground-observation biases are essential to keep in mind. As mentioned before, the ground-observation points are close together and not necessarily representative of the entire domain and soil moisture is spatially autocorrelated, leading to a possible sampling bias. The model itself tends to underestimate soil moisture variance at measurement points, as reflected in Fig. 6b. Thus, when looking at Fig. 6a, it should be noted that there is a cancelling of two unrelated biases: the sampling bias and the modeling bias. Figure 7 is not affected as much since clustered sampling of an autocorrelated field does not produce bias in field means as it does for field variances. This is reflected in the weaker difference between Figs. 6a and 6b versus Figs. 7a and 7b.

Interestingly, SPS_1 has the highest temporal correlation of both simulations (Table 3), 0.901 and 0.907 with domainwide and measurement-sites-only averages, respectively. This confirms that SPS_1 mostly has a bias problem (as seen in Fig. 7) and not a temporal correlation problem. This may suggest that the correlation coefficient alone does not adequately reflect the relationship between simulated and measured soil moisture. Both simulations have stronger correlations with observations when comparing with the measurement-sites-only average rather than when comparing domainwide average, suggesting that evolution of soil moisture in time is best described by data corresponding to measurement sites. However, the standard deviation of the error for the two simulations is also greater for these data points. Thus, the temporal evolution is best described by measurement-sites-only averages, but the averages themselves are best described by SPS when taken domainwide.

b. Brightness temperature

At L band, the microwave emission, that is, the TB, measured by PALS instruments (see section 2d) originates primarily from the top 5 cm of the soil and is sensitive to soil moisture in this surface volume. However, both surface roughness and vegetation contribute to microwave emissions. Thus, in order to better understand the extent of the impact of soil moisture on brightness temperature, TB anomalies were calculated with the hypothesis that vegetation state is near constant over the study period. The TB anomalies were thus calculated for each grid point using the mean value over the 11 days during which PALS data are available. Figure 8 shows these anomalies for horizontally polarized TB on 12 June 2012 for PALS; SPS_2 (as an example); and a third simulation, SPS_2*. The latter is the result of SPS_2 and CMEM when CMEM is fed with an arbitrary constant soil moisture of 0.15 m3 m−3 but the rest of SPS_2 remains the same. As seen in Fig. 8, although SPS_2 underestimates TB anomalies’ spatial variability by about half, most of the spatial variability is due to soil moisture, thus justifying the use of TBs to evaluate SPS-simulated soil moisture.

Fig. 8.
Fig. 8.

Horizontally polarized TB (K) anomalies on 12 Jun 2012 from (a) PALS, (b) SPS_2, and (c) SPS_2* when CMEM is fed with a constant 0.15 m3 m−3 soil moisture, with the respective spatial std dev in each subplot. The mean values to calculate the anomalies are taken from the 11 days when PALS data are available.

Citation: Journal of Hydrometeorology 18, 3; 10.1175/JHM-D-16-0131.1

Please note that only horizontally polarized TB data are shown in this section since vertically polarized TB data gave similar results. Thus, the data presented in Fig. 9 are 1-km horizontally polarized TBs: Fig. 9 (right) shows PALS data, while Fig. 9 (left, center) shows the TBs simulated by SPS and CMEM, for 3 days during the study period. Drier (wetter) soils are generally associated with higher (lower) TBs.

Fig. 9.
Fig. 9.

Horizontally polarized TB (K) from PALS and both SPS simulations as simulated by CMEM at a 1-km resolution on three days during the study period.

Citation: Journal of Hydrometeorology 18, 3; 10.1175/JHM-D-16-0131.1

As expected, PALS data show higher TBs and thus drier soils in the more elevated terrain in the northwest, where sandy and loamy soils dominate. The wetter soils to the east have significantly higher clay content. The transition between the two types of soils is better simulated by SPS_2 as a result of high-resolution soil texture. However, the impact of the different configurations on simulated TBs is not as strong as on simulated soil moisture. This is also denoted in Table 4, which lists the spatial variability of both simulations as well as observed data across the domain. SPS_1 even shows greater spatial variance on 13 July than SPS_2, such that it is closest to the observed variance.

Table 4.

Spatial variance (std dev) of simulated and observed PALS data on different days. The 95% CI calculated using the bootstrapping method is given in parentheses.

Table 4.

Figure 10 shows scatterplots of simulated versus observed TBs, along with the Pearson correlation coefficient in Table 5. The weak improvements from the different configurations are better visualized here. Indeed, from SPS_1 to SPS_2 the correlation coefficient increases significantly from 0.305 to 0.599 on 12 June. Nonetheless, please note that these scatterplots include 910 comparison points (i.e., 910 grid points of 1 km × 1 km horizontal resolution), compared to the 97 points in Fig. 5, and that SPS_1’s correlation to PALS (0.305) is an order of magnitude higher than its correlation to in situ data (0.031) on that day. The low TBs over forested areas are also better represented by SPS_2 during the three days, although in this case, the L-band emissivity is strongly altered by vegetation and is not necessarily representative of soil moisture.

Fig. 10.
Fig. 10.

Horizontally polarized TB (K) scatterplot of simulated vs observed data for both SPS simulations on three days.

Citation: Journal of Hydrometeorology 18, 3; 10.1175/JHM-D-16-0131.1

Table 5.

Spatial correlations between simulated and observed PALS data on different days. The 95% CI calculated using the bootstrapping method is given in parentheses.

Table 5.

However, these are single-day results, and thus the temporal evolution of the domainwide spatial average and variability were calculated for the entire 43-day campaign and are provided in Fig. 11. The average TB tends to be underestimated by SPS, although SPS_2 shows better results than SPS_1. With respect to spatial variability, SPS follows the general pattern of high variability during the first half followed by a period of lower variability when crops have reached maturity. SPS is not able, however, to perfectly follow the strong increase in mean TB (due to a decrease in soil moisture) accompanied by a strong decrease in spatial variability during the end of June. Still, both SPS simulations represent a TB spatial variability that is highly representative of data observed at a larger scale, except during the transition periods: a marked improvement compared to the underestimation of the soil moisture spatial variability compared to in situ observations, as previously shown. The reason is probably because TB observations sample equally all regions of the domain while in situ data are mostly clustered around the center of the domain as shown in Fig. 2a.

Fig. 11.
Fig. 11.

Domainwide TB (K) (a) average and (b) std dev evolution in time of both SPS simulations (blue and red with corresponding 95% CIs) and observations (black) when horizontally polarized.

Citation: Journal of Hydrometeorology 18, 3; 10.1175/JHM-D-16-0131.1

It is interesting to note that both observation datasets as well as simulated data show decreased variance in soil moisture (except SPS_1) and TBs during the dry period, that is, the second half of the study period. Since the opposite would have been expected, we do not have any rational explanations at this point. Also, SPS_1 better simulates spatial correlation (Table 5) and variance (Table 4, Fig. 9b) on 13 July while greatly underestimating the spatial mean (Fig. 9b), more than SPS_2. This tendency is also observed when compared to in situ data. This suggests that during drier periods, factors other than soil texture become important with respect to soil moisture, such as vegetation, micro-orography, organic matter, and rock fraction. However, during periods with plenty of precipitation, soil texture seems essential in improving soil moisture simulated data at high resolutions.

4. Discussion and conclusions

ECCC’s SPS was run at very high resolution (100 m) over southern Manitoba (Canada), where an extensive field experiment took place in the summer of 2012 (SMAPVEX12). The objective of this study was to reproduce the spatial variability of soil moisture and brightness temperature as measured by point-based ground measurements and airborne remote sensing data. Two SPS simulations were run, one with uniform soil texture and the other with observed high-resolution soil texture, to improve on the spatial variability of simulated soil moisture and brightness temperature. Both simulations were then evaluated with two datasets: ground-based in situ data as well as airborne L-band radiometer measurements of brightness temperature. Both datasets have their advantages and disadvantages. In situ data are precise and frequent measurements of soil moisture, but they are very localized, that is, they do not cover all the study domain and are not representative of large areas. Airborne TB measurements, on the other hand, cover all the study domain, are relatively frequent, and are very sensitive to soil moisture, but they are at 1.5-km resolution and imply transfer radiative modeling (CMEM).

When compared to point-based ground measurements (based on field average measurements and data recorded by in situ soil moisture stations), the inclusion of detailed soil texture improved SPS-simulated soil moisture in terms of spatial variability. However, although results are improved compared to those in Garnaud et al. (2016), spatial variability remains underestimated, especially during the first period of the study period corresponding to relatively high precipitation and crop growth. This could be due to the predefined growth cycle of crops in SPS. Since precipitation forcing is provided by a 15-km atmospheric model, data were obtained from three ECCC measuring stations surrounding the study domain (data not shown), and it was concluded that the precipitation spatial variability among the stations was too small to account for the model underestimating the spatial variability of soil moisture.

Interestingly, both spatial and temporal variability are closer to observations when the variability and average are calculated using domainwide results versus only grid points corresponding to field and in situ point measurements. It could be speculated that the increased variability in both time and space is a result of a greater number of grid points in the domainwide data. Indeed, since most of the in situ measurement points are close together and not necessarily representative of the entire domain and soil moisture is an autocorrelated field, it may explain why domainwide data give better results than grid points corresponding to in situ measurements. It implies that both in situ data and modeled data can be used in cal–val efforts in the context of a satellite mission such as NASA’s SMAP mission, which acts at coarser resolutions (3–36 km). Indeed, comparison between footprint-scale satellite data and in situ and modeled data would involve spatial averaging of the latter two.

In the context of L-band microwave emissions, with detailed soil texture in the simulation setup, SPS brightness temperature fields compare well with remote sensing data in terms of spatial variability, except during transition periods, and very well in terms of spatial average. Since TB is very sensitive to soil moisture, as demonstrated in this study, these results emphasize that SPS is able to represent soil moisture relatively well at 1-km resolution.

It is important to note that SPS_1 better simulates spatial correlation and variance during the end of the study period when compared to PALS data, while underestimating the spatial mean more than SPS_2. This tendency is also observed when compared to in situ data. This suggests that during drier periods, factors other than soil texture become important with respect to soil moisture spatial variability, such as vegetation, micro-orography, organic matter, and rock fraction. However, during periods with plenty of precipitation, soil texture seems essential in improving simulated soil moisture spatial variability at high resolutions.

As a result, this study shows that, if run on a larger domain, SPS could be used to provide very high–resolution soil moisture maps in the context of hydrology modeling and agricultural planning, granted that high-resolution soil texture and vegetation maps are made available. Furthermore, for research purposes at ECCC, SPS will now be used to downscale land data assimilation data to initialize the atmospheric model run at high resolutions (100–250 m).

During SMAP cal–val efforts, satellite data are compared to core validation sites around the world (http://smap.jpl.nasa.gov/science/validation/). It has already been noted that, although most sites have already validated the remote sensing data, SMAP data do not compare well with a few sites. In such cases, in light of results in this study, SPS could be used to better understand the discrepancies between core sites and satellite data. Furthermore, in addition to core validation sites, which have numerous measuring stations in each site, a sparse network also supplies validation data to the cal–val efforts. However, these are single point-based measurements, and they are to be compared to data from a satellite pixel (approximately 1300 km2). Crow et al. (2012b) have explored this problem and presented techniques using distributed land surface modeling and sparse ground observations for satellite soil moisture cal–val. Thus, when needed, SPS could be run over a domain surrounding the point measurement in order to evaluate whether the measurement is representative of the satellite pixel area. At this resolution and similar to Chaney et al. (2015), SPS could also be used for future ground network design in order to avoid biases linked to the positioning of measuring instruments, as it may be the case for SMAPVEX12.

Further improvements would include specification of vegetation based on C-band active measurements and perturbations associated with micro-orographical features.

Acknowledgments

The authors thank Maria Abrahamowicz and Dorothy Durnford for their technical assistance and Vanh Souvanlasy for his help preparing the surface fields. This work was funded by the Canadian Space Agency through the Government Related Initiative Program (GRIP).

REFERENCES

  • Alavi, N., S. Bélair, V. Fortin, S. Zhang, S. Z. Husain, M. L. Carrera, and M. Abrahamowicz, 2016: Warm season evaluation of soil moisture prediction in the Soil, Vegetation, and Snow (SVS) scheme. J. Hydrometeor., 17, 23152332, doi:10.1175/JHM-D-15-0189.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Albergel, C., G. Balsamo, P. de Rosnay, J. Munos-Sabater, and S. Boussetta, 2012: A bare ground evaporation revision in the ECMWF land-surface scheme: Evaluation of its impact using ground soil moisture and satellite microwave data. Hydrol. Earth Syst. Sci., 16, 36073620, doi:10.5194/hess-16-3607-2012.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bélair, S., R. Brown, J. Mailhot, B. Bilodeau, and L.-P. Crevier, 2003a: Operational implementation of the ISBA land surface scheme in the Canadian regional weather forecast model. Part II: Cold season results. J. Hydrometeor., 4, 371386, doi:10.1175/1525-7541(2003)4<371:OIOTIL>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bélair, S., L.-P. Crevier, J. Mailhot, B. Bilodeau, and Y. Delage, 2003b: Operational implementation of the ISBA land surface scheme in the Canadian regional weather forecast model. Part I: Warm season results. J. Hydrometeor., 4, 352370, doi:10.1175/1525-7541(2003)4<352:OIOTIL>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bernier, N., S. Bélair, B. Bilodeau, and L. Tong, 2011: Near-surface and land surface forecast system of the Vancouver 2010 Winter Olympic and Paralympic Games. J. Hydrometeor., 12, 508530, doi:10.1175/2011JHM1250.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bernier, N., S. Bélair, B. Bilodeau, and L. Tong, 2012: Assimilation and high resolution forecasts of surface and near surface conditions for the 2010 Vancouver Winter Olympic and Paralympic Games. Pure Appl. Geophys., 171, 243256, doi:10.1007/s00024-012-0542-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Betts, A., and P. Viterbo, 2005: Land-surface, boundary layer, and cloud-field coupling over the southwestern amazon in ERA-40. J. Geophys. Res., 110, D14108, doi:10.1029/2004JD005702.

    • Search Google Scholar
    • Export Citation
  • Brocca, L., R. Morbidelli, F. Melone, and T. Moramarco, 2007: Soil moisture spatial variability in experimental areas of central Italy. J. Hydrol., 333, 356373, doi:10.1016/j.jhydrol.2006.09.004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Carrera, M., S. Bélair, V. Fortin, B. Bilodeau, D. Charpentier, and I. Doré, 2010: Evaluation of snowpack simulations over the Canadian Rockies with an experimental hydrometeorological modeling system. J. Hydrometeor., 11, 11231140, doi:10.1175/2010JHM1274.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Carrera, M., S. Bélair, and B. Bilodeau, 2015: The Canadian Land Data Assimilation System (CALDAS): Description and synthetic evaluation study. J. Hydrometeor., 16, 12931314, doi:10.1175/JHM-D-14-0089.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Champagne, C., A. Berg, H. McNairn, G. B. Drewitt, and E. C. Huffman, 2012: Evaluation of soil moisture extremes for agricultural productivity in the Canadian Prairies. Agric. For. Meteor., 165, 111, doi:10.1016/j.agrformet.2012.06.003.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chaney, N., J. Roundy, J. Herrera-Estrada, and E. Wood, 2015: High-resolution modeling of the spatial heterogeneity of soil moisture: Applications in network design. Water Resour. Res., 51, 619638, doi:10.1002/2013WR014964.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, F., S. Miao, M. Tewari, J. Bao, and H. Kusaka, 2011: A numerical study of interactions between surface forcing and sea breeze circulations and their effects on stagnation in the greater Houston area. J. Geophys. Res., 116, D12105, doi:10.1029/2010JD015533.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Conil, S., H. Douville, and S. Tyteca, 2007: The relative influence of soil moisture and SST in climate predictability explored within ensembles of AMIP type experiments. Climate Dyn., 28, 125145, doi:10.1007/s00382-006-0172-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Crow, W., S. Kumar, and J. Bolten, 2012a: On the utility of land surface models for agricultural drought monitoring. Hydrol. Earth Syst. Sci., 16, 34513460, doi:10.5194/hess-16-3451-2012.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Crow, W., and Coauthors, 2012b: Upscaling sparse ground-based soil moisture observations for the validation of coarse-resolution satellite soil moisture products. Rev. Geophys., 50, RG2002, doi:10.1029/2011RG000372.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cutforth, H. W., and Coauthors, 2013: Comparing rooting characteristics and soil water withdrawal patterns of wheat with alternative oilseed and pulse crops grown in the semiarid Canadian Prairie. Can. J. Soil Sci., 93, 147160, doi:10.4141/cjss2012-081.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • de Rosnay, P., and Coauthors, 2009: The AMMA Land Surface Model Intercomparison Experiment coupled to the Community Microwave Emission Model: ALMIP-MEM. J. Geophys. Res., 114, D05108, doi:10.1029/2008JD010724.

    • Search Google Scholar
    • Export Citation
  • Drusch, M., T. Holmes, P. DeRosnay, and G. Balsamo, 2009: Comparing ERA-40 based L-band brightness temperatures with Skylab observations: A calibration/validation study using the Community Microwave Emission Model. J. Hydrometeor., 10, 213226, doi:10.1175/2008JHM964.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dwyer, L., D. Stewart, and D. Balchin, 1988: Rooting characteristics of corn, soybeans and barley as a function of available water and soil physical characteristics. Can. J. Soil Sci., 68, 121132, doi:10.4141/cjss88-011.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Eltahir, E., 1998: A soil moisture–rainfall feedback mechanism: 1. Theory and observations. Water Resour. Res., 34, 765776, doi:10.1029/97WR03499.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Entekhabi, D., and Coauthors, 2010: The Soil Moisture Active Passive (SMAP) mission. Proc. IEEE, 98, 704716, doi:10.1109/JPROC.2010.2043918.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Essery, R., M. Best, R. Betts, P. Cox, and C. Taylor, 2003: Explicit representation of sub-grid heterogeneity in a GCM land surface scheme. J. Hydrometeor., 4, 530543, doi:10.1175/1525-7541(2003)004<0530:EROSHI>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garnaud, C., S. Bélair, A. Berg, and T. Rowlandson, 2016: Hyperresolution land surface modeling in the context of SMAP Cal–Val. J. Hydrometeor., 17, 345352, doi:10.1175/JHM-D-15-0070.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Husain, S., N. Alavi, S. Bélair, and M. Carrera, 2016: The multibudget Soil, Vegetation, and Snow (SVS) scheme for land surface parameterization: Offline warm season evaluation. J. Hydrometeor., 17, 22932313, doi:10.1175/JHM-D-15-0228.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kirdyashev, K. P., A. Chukhlantsev, and A. Shutko, 1979: Microwave radiation of the earth’s surface in the presence of vegetation cover. Radiotech. Elektron., 24, 256264.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., and Coauthors, 2010: Contribution of land surface initialization to subseasonal forecast skill: First results from a multi-model experiment. Geophys. Res. Lett., 37, L02402, doi:10.1029/2009GL041677.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Leroyer, S., S. Bélair, J. Mailhot, and I. B. Strachan, 2011: Microscale numerical prediction over Montreal with the Canadian external urban modeling system. J. Appl. Meteor. Climatol., 50, 24102428, doi:10.1175/JAMC-D-11-013.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Leroyer, S., S. Bélair, S. Husain, and J. Mailhot, 2014: Subkilometer numerical weather prediction in an urban coastal area: A case study over the Vancouver metropolitan area. J. Appl. Meteor. Climatol., 53, 14331453, doi:10.1175/JAMC-D-13-0202.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Li, B., and M. Rodell, 2013: Spatial variability and its scale dependency of observed and modeled soil moisture over different climate regions. Hydrol. Earth Syst. Sci., 17, 11771188, doi:10.5194/hess-17-1177-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lorenz, R., E. Jaeger, and S. I. Seneviratne, 2010: Persistence of heat waves and its link to soil moisture memory. Geophys. Res. Lett., 37, L09703, doi:10.1029/2010GL042764.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mahfouf, J.-F., B. Brasnett, and S. Gagnon, 2007: A Canadian Precipitation Analysis (CaPA) project: Description and preliminary results. Atmos.–Ocean, 45, 117, doi:10.3137/ao.v450101.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mailhot, J., and Coauthors, 2006: The 15-km version of the Canadian regional forecast system. Atmos.–Ocean, 44, 133149, doi:10.3137/ao.440202.

  • Manns, H. R., A. Berg, and H. McNairn, 2014: Impact of soil surface characteristics on soil water content variability in agricultural fields. Hydrol. Processes, 28, 43404351, doi:10.1002/hyp.10216.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McNairn, H., A. Merzouki, A. Pacheco, and J. Fitzmaurice, 2012: Monitoring soil moisture to support risk reduction for the agriculture sector using radarsat-2. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 5, 824834, doi:10.1109/JSTARS.2012.2192416.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McNairn, H., and Coauthors, 2015: The Soil Moisture Active Passive Validation Experiment 2012 (SMAPVEX12): Prelaunch calibration and validation of the SMAP soil moisture algorithms. IEEE Trans. Geosci. Remote Sens., 53, 27842801, doi:10.1109/TGRS.2014.2364913.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mittelbach, H., and S. I. Seneviratne, 2012: A new perspective on the spatio-temporal variability of soil moisture: Temporal dynamics versus time-invariant contributions. Hydrol. Earth Syst. Sci., 16, 21692179, doi:10.5194/hess-16-2169-2012.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Molod, A., and H. Salmun, 2002: A global assessment of the mosaic approach to modeling land surface heterogeneity. J. Geophys. Res., 107, doi:10.1029/2001JD000588.

    • Search Google Scholar
    • Export Citation
  • Noilhan, J., and S. Planton, 1989: A simple parameterization of land surface processes for meteorological models. Mon. Wea. Rev., 117, 536549, doi:10.1175/1520-0493(1989)117<0536:ASPOLS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Parrens, M., J.-C. Calvet, P. de Rosnay, and B. Decharme, 2014: Benchmarking of L-band soil microwave emission models. Remote Sens. Environ., 140, 407419, doi:10.1016/j.rse.2013.09.017.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rochoux, M., S. Bélair, M. Abrahamowicz, and P. Pellerin, 2016: Subgrid-scale variability for thermodynamic variables in an offline land surface prediction system. J. Hydrometeor., 17, 171193, doi:10.1175/JHM-D-15-0016.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sabater, J. M., P. DeRosnay, and G. Balsamo, 2011: Sensitivity of L-band NWP forward modeling to soil roughness. Int. J. Remote Sens., 32, 56075620, doi:10.1080/01431161.2010.507260.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schmugge, T. J., and B. J. Choudhury, 1981: A comparison of radiative transfer models for predicting the microwave emission from soil. Radio Sci., 16, 927938, doi:10.1029/RS016i005p00927.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Seneviratne, S. I., and Coauthors, 2006: Soil moisture memory in AGCM simulations: Analysis of Global Land–Atmosphere Coupling Experiment (GLACE) data. J. Hydrometeor., 7, 10901112, doi:10.1175/JHM533.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Seneviratne, S. I., T. Corti, E. L. Davin, M. Hirschi, E. B. Jaeger, I. Lehner, B. Orlowsky, and A. J. Teuling, 2010: Investigating soil moisture–climate interactions in a changing climate: A review. Earth Sci. Rev., 99, 125161, doi:10.1016/j.earscirev.2010.02.004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Setinoyo, T., A. Weiss, J. Specht, K. Cassman, and A. Dobermann, 2008: Leaf area index simulation in soybean grown under near-optimal conditions. Field Crops Res., 108, 8292, doi:10.1016/j.fcr.2008.03.005.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Small, E., 2001: The influence of soil moisture anomalies on variability of the North American monsoon system. Geophys. Res. Lett., 28, 139142, doi:10.1029/2000GL011652.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Teuling, A., and P. Troch, 2005: Improved understanding of soil moisture variability dynamics. Geophys. Res. Lett., 32, L05404, doi:10.1029/2004GL021935.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Thompson, W., T. Holt, and J. Pullen, 2007: Investigation of a sea breeze front in an urban environment. Quart. J. Roy. Meteor. Soc., 133, 579594, doi:10.1002/qj.52.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, J. R., and T. Schmugge, 1980: An empirical model for the complex dielectric permittivity of soils as a function of water content. IEEE Trans. Geosci. Remote Sens., 18, 288295, doi:10.1109/TGRS.1980.350304.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, T., T. Franz, V. Zlotnik, J. You, and M. Shulski, 2015: Investigating soil controls on soil moisture spatial variability: Numerical simulations and field observations. J. Hydrol., 524, 576586, doi:10.1016/j.jhydrol.2015.03.019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wigneron, J. P., L. Laguerre, and Y. H. Kerr, 2001: A simple parameterization of the L-band microwave emission from rough agricultural soils. IEEE Trans. Geosci. Remote Sens., 39, 16971707, doi:10.1109/36.942548.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wood, E. F., S. Schubert, A. Wood, C. Peters-Lidard, K. Mo, A. Mariotti, and R. Pulwarty, 2015: Prospects for advancing drought understanding, monitoring, and prediction. J. Hydrometeor., 16, 16361657, doi:10.1175/JHM-D-14-0164.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save
  • Alavi, N., S. Bélair, V. Fortin, S. Zhang, S. Z. Husain, M. L. Carrera, and M. Abrahamowicz, 2016: Warm season evaluation of soil moisture prediction in the Soil, Vegetation, and Snow (SVS) scheme. J. Hydrometeor., 17, 23152332, doi:10.1175/JHM-D-15-0189.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Albergel, C., G. Balsamo, P. de Rosnay, J. Munos-Sabater, and S. Boussetta, 2012: A bare ground evaporation revision in the ECMWF land-surface scheme: Evaluation of its impact using ground soil moisture and satellite microwave data. Hydrol. Earth Syst. Sci., 16, 36073620, doi:10.5194/hess-16-3607-2012.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bélair, S., R. Brown, J. Mailhot, B. Bilodeau, and L.-P. Crevier, 2003a: Operational implementation of the ISBA land surface scheme in the Canadian regional weather forecast model. Part II: Cold season results. J. Hydrometeor., 4, 371386, doi:10.1175/1525-7541(2003)4<371:OIOTIL>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bélair, S., L.-P. Crevier, J. Mailhot, B. Bilodeau, and Y. Delage, 2003b: Operational implementation of the ISBA land surface scheme in the Canadian regional weather forecast model. Part I: Warm season results. J. Hydrometeor., 4, 352370, doi:10.1175/1525-7541(2003)4<352:OIOTIL>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bernier, N., S. Bélair, B. Bilodeau, and L. Tong, 2011: Near-surface and land surface forecast system of the Vancouver 2010 Winter Olympic and Paralympic Games. J. Hydrometeor., 12, 508530, doi:10.1175/2011JHM1250.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bernier, N., S. Bélair, B. Bilodeau, and L. Tong, 2012: Assimilation and high resolution forecasts of surface and near surface conditions for the 2010 Vancouver Winter Olympic and Paralympic Games. Pure Appl. Geophys., 171, 243256, doi:10.1007/s00024-012-0542-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Betts, A., and P. Viterbo, 2005: Land-surface, boundary layer, and cloud-field coupling over the southwestern amazon in ERA-40. J. Geophys. Res., 110, D14108, doi:10.1029/2004JD005702.

    • Search Google Scholar
    • Export Citation
  • Brocca, L., R. Morbidelli, F. Melone, and T. Moramarco, 2007: Soil moisture spatial variability in experimental areas of central Italy. J. Hydrol., 333, 356373, doi:10.1016/j.jhydrol.2006.09.004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Carrera, M., S. Bélair, V. Fortin, B. Bilodeau, D. Charpentier, and I. Doré, 2010: Evaluation of snowpack simulations over the Canadian Rockies with an experimental hydrometeorological modeling system. J. Hydrometeor., 11, 11231140, doi:10.1175/2010JHM1274.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Carrera, M., S. Bélair, and B. Bilodeau, 2015: The Canadian Land Data Assimilation System (CALDAS): Description and synthetic evaluation study. J. Hydrometeor., 16, 12931314, doi:10.1175/JHM-D-14-0089.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Champagne, C., A. Berg, H. McNairn, G. B. Drewitt, and E. C. Huffman, 2012: Evaluation of soil moisture extremes for agricultural productivity in the Canadian Prairies. Agric. For. Meteor., 165, 111, doi:10.1016/j.agrformet.2012.06.003.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chaney, N., J. Roundy, J. Herrera-Estrada, and E. Wood, 2015: High-resolution modeling of the spatial heterogeneity of soil moisture: Applications in network design. Water Resour. Res., 51, 619638, doi:10.1002/2013WR014964.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, F., S. Miao, M. Tewari, J. Bao, and H. Kusaka, 2011: A numerical study of interactions between surface forcing and sea breeze circulations and their effects on stagnation in the greater Houston area. J. Geophys. Res., 116, D12105, doi:10.1029/2010JD015533.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Conil, S., H. Douville, and S. Tyteca, 2007: The relative influence of soil moisture and SST in climate predictability explored within ensembles of AMIP type experiments. Climate Dyn., 28, 125145, doi:10.1007/s00382-006-0172-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Crow, W., S. Kumar, and J. Bolten, 2012a: On the utility of land surface models for agricultural drought monitoring. Hydrol. Earth Syst. Sci., 16, 34513460, doi:10.5194/hess-16-3451-2012.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Crow, W., and Coauthors, 2012b: Upscaling sparse ground-based soil moisture observations for the validation of coarse-resolution satellite soil moisture products. Rev. Geophys., 50, RG2002, doi:10.1029/2011RG000372.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cutforth, H. W., and Coauthors, 2013: Comparing rooting characteristics and soil water withdrawal patterns of wheat with alternative oilseed and pulse crops grown in the semiarid Canadian Prairie. Can. J. Soil Sci., 93, 147160, doi:10.4141/cjss2012-081.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • de Rosnay, P., and Coauthors, 2009: The AMMA Land Surface Model Intercomparison Experiment coupled to the Community Microwave Emission Model: ALMIP-MEM. J. Geophys. Res., 114, D05108, doi:10.1029/2008JD010724.

    • Search Google Scholar
    • Export Citation
  • Drusch, M., T. Holmes, P. DeRosnay, and G. Balsamo, 2009: Comparing ERA-40 based L-band brightness temperatures with Skylab observations: A calibration/validation study using the Community Microwave Emission Model. J. Hydrometeor., 10, 213226, doi:10.1175/2008JHM964.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dwyer, L., D. Stewart, and D. Balchin, 1988: Rooting characteristics of corn, soybeans and barley as a function of available water and soil physical characteristics. Can. J. Soil Sci., 68, 121132, doi:10.4141/cjss88-011.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Eltahir, E., 1998: A soil moisture–rainfall feedback mechanism: 1. Theory and observations. Water Resour. Res., 34, 765776, doi:10.1029/97WR03499.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Entekhabi, D., and Coauthors, 2010: The Soil Moisture Active Passive (SMAP) mission. Proc. IEEE, 98, 704716, doi:10.1109/JPROC.2010.2043918.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Essery, R., M. Best, R. Betts, P. Cox, and C. Taylor, 2003: Explicit representation of sub-grid heterogeneity in a GCM land surface scheme. J. Hydrometeor., 4, 530543, doi:10.1175/1525-7541(2003)004<0530:EROSHI>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garnaud, C., S. Bélair, A. Berg, and T. Rowlandson, 2016: Hyperresolution land surface modeling in the context of SMAP Cal–Val. J. Hydrometeor., 17, 345352, doi:10.1175/JHM-D-15-0070.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Husain, S., N. Alavi, S. Bélair, and M. Carrera, 2016: The multibudget Soil, Vegetation, and Snow (SVS) scheme for land surface parameterization: Offline warm season evaluation. J. Hydrometeor., 17, 22932313, doi:10.1175/JHM-D-15-0228.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kirdyashev, K. P., A. Chukhlantsev, and A. Shutko, 1979: Microwave radiation of the earth’s surface in the presence of vegetation cover. Radiotech. Elektron., 24, 256264.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., and Coauthors, 2010: Contribution of land surface initialization to subseasonal forecast skill: First results from a multi-model experiment. Geophys. Res. Lett., 37, L02402, doi:10.1029/2009GL041677.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Leroyer, S., S. Bélair, J. Mailhot, and I. B. Strachan, 2011: Microscale numerical prediction over Montreal with the Canadian external urban modeling system. J. Appl. Meteor. Climatol., 50, 24102428, doi:10.1175/JAMC-D-11-013.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Leroyer, S., S. Bélair, S. Husain, and J. Mailhot, 2014: Subkilometer numerical weather prediction in an urban coastal area: A case study over the Vancouver metropolitan area. J. Appl. Meteor. Climatol., 53, 14331453, doi:10.1175/JAMC-D-13-0202.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Li, B., and M. Rodell, 2013: Spatial variability and its scale dependency of observed and modeled soil moisture over different climate regions. Hydrol. Earth Syst. Sci., 17, 11771188, doi:10.5194/hess-17-1177-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lorenz, R., E. Jaeger, and S. I. Seneviratne, 2010: Persistence of heat waves and its link to soil moisture memory. Geophys. Res. Lett., 37, L09703, doi:10.1029/2010GL042764.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mahfouf, J.-F., B. Brasnett, and S. Gagnon, 2007: A Canadian Precipitation Analysis (CaPA) project: Description and preliminary results. Atmos.–Ocean, 45, 117, doi:10.3137/ao.v450101.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mailhot, J., and Coauthors, 2006: The 15-km version of the Canadian regional forecast system. Atmos.–Ocean, 44, 133149, doi:10.3137/ao.440202.

  • Manns, H. R., A. Berg, and H. McNairn, 2014: Impact of soil surface characteristics on soil water content variability in agricultural fields. Hydrol. Processes, 28, 43404351, doi:10.1002/hyp.10216.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McNairn, H., A. Merzouki, A. Pacheco, and J. Fitzmaurice, 2012: Monitoring soil moisture to support risk reduction for the agriculture sector using radarsat-2. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 5, 824834, doi:10.1109/JSTARS.2012.2192416.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McNairn, H., and Coauthors, 2015: The Soil Moisture Active Passive Validation Experiment 2012 (SMAPVEX12): Prelaunch calibration and validation of the SMAP soil moisture algorithms. IEEE Trans. Geosci. Remote Sens., 53, 27842801, doi:10.1109/TGRS.2014.2364913.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mittelbach, H., and S. I. Seneviratne, 2012: A new perspective on the spatio-temporal variability of soil moisture: Temporal dynamics versus time-invariant contributions. Hydrol. Earth Syst. Sci., 16, 21692179, doi:10.5194/hess-16-2169-2012.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Molod, A., and H. Salmun, 2002: A global assessment of the mosaic approach to modeling land surface heterogeneity. J. Geophys. Res., 107, doi:10.1029/2001JD000588.

    • Search Google Scholar
    • Export Citation
  • Noilhan, J., and S. Planton, 1989: A simple parameterization of land surface processes for meteorological models. Mon. Wea. Rev., 117, 536549, doi:10.1175/1520-0493(1989)117<0536:ASPOLS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Parrens, M., J.-C. Calvet, P. de Rosnay, and B. Decharme, 2014: Benchmarking of L-band soil microwave emission models. Remote Sens. Environ., 140, 407419, doi:10.1016/j.rse.2013.09.017.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rochoux, M., S. Bélair, M. Abrahamowicz, and P. Pellerin, 2016: Subgrid-scale variability for thermodynamic variables in an offline land surface prediction system. J. Hydrometeor., 17, 171193, doi:10.1175/JHM-D-15-0016.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sabater, J. M., P. DeRosnay, and G. Balsamo, 2011: Sensitivity of L-band NWP forward modeling to soil roughness. Int. J. Remote Sens., 32, 56075620, doi:10.1080/01431161.2010.507260.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schmugge, T. J., and B. J. Choudhury, 1981: A comparison of radiative transfer models for predicting the microwave emission from soil. Radio Sci., 16, 927938, doi:10.1029/RS016i005p00927.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Seneviratne, S. I., and Coauthors, 2006: Soil moisture memory in AGCM simulations: Analysis of Global Land–Atmosphere Coupling Experiment (GLACE) data. J. Hydrometeor., 7, 10901112, doi:10.1175/JHM533.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Seneviratne, S. I., T. Corti, E. L. Davin, M. Hirschi, E. B. Jaeger, I. Lehner, B. Orlowsky, and A. J. Teuling, 2010: Investigating soil moisture–climate interactions in a changing climate: A review. Earth Sci. Rev., 99, 125161, doi:10.1016/j.earscirev.2010.02.004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Setinoyo, T., A. Weiss, J. Specht, K. Cassman, and A. Dobermann, 2008: Leaf area index simulation in soybean grown under near-optimal conditions. Field Crops Res., 108, 8292, doi:10.1016/j.fcr.2008.03.005.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Small, E., 2001: The influence of soil moisture anomalies on variability of the North American monsoon system. Geophys. Res. Lett., 28, 139142, doi:10.1029/2000GL011652.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Teuling, A., and P. Troch, 2005: Improved understanding of soil moisture variability dynamics. Geophys. Res. Lett., 32, L05404, doi:10.1029/2004GL021935.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Thompson, W., T. Holt, and J. Pullen, 2007: Investigation of a sea breeze front in an urban environment. Quart. J. Roy. Meteor. Soc., 133, 579594, doi:10.1002/qj.52.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, J. R., and T. Schmugge, 1980: An empirical model for the complex dielectric permittivity of soils as a function of water content. IEEE Trans. Geosci. Remote Sens., 18, 288295, doi:10.1109/TGRS.1980.350304.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, T., T. Franz, V. Zlotnik, J. You, and M. Shulski, 2015: Investigating soil controls on soil moisture spatial variability: Numerical simulations and field observations. J. Hydrol., 524, 576586, doi:10.1016/j.jhydrol.2015.03.019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wigneron, J. P., L. Laguerre, and Y. H. Kerr, 2001: A simple parameterization of the L-band microwave emission from rough agricultural soils. IEEE Trans. Geosci. Remote Sens., 39, 16971707, doi:10.1109/36.942548.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wood, E. F., S. Schubert, A. Wood, C. Peters-Lidard, K. Mo, A. Mariotti, and R. Pulwarty, 2015: Prospects for advancing drought understanding, monitoring, and prediction. J. Hydrometeor., 16, 16361657, doi:10.1175/JHM-D-14-0164.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    Schematic of the modeling system. SPS is driven by hourly downscaled forecast fields from the 15-km version of GEM. Forcing fields are temperature T, humidity q, and winds (U, V) at the atmospheric model’s first level (~40 m above the surface) as well as surface pressure P0, radiation (rad), and rate of precipitation (Pr). In SPS, the surface heterogeneity is represented at subgrid scales as an aggregation of different covers (i.e., land, urban, water, continental ice, and sea ice). Adapted from Bernier et al. (2012).

  • Fig. 2.

    (a) Orography (m) of the study domain along with the location of the 102 measurement sites from SMAPVEX12 (black filled circles), (b) the geographical location of the study domain, and (c) its primary land-cover type at each grid point.

  • Fig. 3.

    Percentage of (a) sand and (b) clay in soil coming from the AAFC soil survey data.

  • Fig. 4.

    Soil moisture (m3 m−3) on three days as simulated by SPS with two different configurations. The color-filled circles are observations collected during SMAPVEX12, with colors corresponding to the same color scale as the simulated data.

  • Fig. 5.

    Soil moisture (m3 m−3) scatterplot of simulated vs observed data for the two SPS simulations on three days during the study period.

  • Fig. 6.

    Soil moisture spatial std dev (m3 m−3) evolution in time from 7 Jun to 19 Jul 2012 of two SPS simulations (blue and red with corresponding 95% CIs) and observations (black). (a) Simulated domainwide std dev compared with observations and (b) the spatial std dev of grid points corresponding to measurement sites only compared with observations.