1. Introduction
Land surface models (LSMs) have been used widely in studying interactions within the soil, vegetation and atmosphere continuum, in addition to predicting water and energy fluxes. Improved understanding of land–atmosphere interactions potentially enhances the ability of weather and climate models to predict future conditions (Barlage et al. 2015; Chen and Dudhia 2001; Gao et al. 2015; Kumar et al. 2014; Sadeghi et al. 2019). Detailed land–atmosphere processes and vegetation characteristics are incorporated in state-of-the-art versions of LSMs for improved predictions of soil–atmosphere boundary fluxes (Chen and Dudhia 2001; Gayler et al. 2013; Niu et al. 2011; Oleson et al. 2013). In addition, high-performance computing facilities and improved spatiotemporal resolution of remote sensing and ground-based measurements have contributed to enhancement of these models. This has resulted in extensive applications of LSMs to characterize evapotranspiration (ET) and soil water dynamics at regional scales (Cai et al. 2014a,b; Gayler et al. 2013; Koster and Suarez 1992; Chen et al. 1996; Long et al. 2014).
Various studies in the past have highlighted the influence of soil moisture in land surface processes and considered it as a key variable and fundamental in governing the exchange of water and energy flux at the soil–atmosphere boundary (Delworth and Manabe 1988, 1993; Dirmeyer and Shukla 1994; Dirmeyer 1995; Gayler et al. 2014; Namias 1952; Poltoradnev et al. 2018; Xia et al. 2014; Zheng et al. 2015). Soil moisture content interacts with atmospheric processes and directly influences the partitioning of net energy at the land surface into sensible and latent heat flux, hence controlling water and energy exchanges between the land surface and atmosphere (Goodrich et al. 1994; Heathman et al. 2009; Qiu et al. 2013; Seneviratne et al. 2010, 2013; Xia et al. 2014). There have been continuous efforts toward improving and archiving in situ soil moisture observations to provide support for calibrating land surface models and validating satellite-based soil moisture products (Crow et al. 2005; Ford and Quiring 2014; Quiring et al. 2016; Xia et al. 2014, 2015). On the other hand, soil moisture predictions from land surface simulations such as, North American Land Data Assimilation System (NLDAS) or Global Land Data Assimilation System (GLDAS) are used as surface initial conditions for numerical weather prediction and for operational drought management as well as for seasonal hydrological prediction (Mitchell et al. 2004; Rodell et al. 2004; Xia et al. 2014). However, the predicted soil moisture outputs from these models have not been comprehensively evaluated at regional or global scale due to the general lack of soil moisture field observations (Spennemann et al. 2015; Xia et al. 2014).
The soil and vegetation parameters incorporated in such models are at global or regional scale and are not able to account for spatial heterogeneity. The soil matrix acts as a reservoir with unknown functions that regulate surface ET (Gochis et al. 2010; Scott et al. 1997). Surface fluxes are modulated by soil physical and hydraulic properties, which are affected by soil texture in LSMs (Gao et al. 2015; Kishné et al. 2017; Michael and Cuenca 1994). Meanwhile, soil structure, organic matter, bulk density, and preferential flow through macropores also influence water flow through soil, but are not captured simply by the texture class from the lookup tables that are typically used for soil parameterization in LSMs (Gutmann and Small 2005, 2007). Consequently, LSMs are commonly calibrated by adjusting parameters to match simulation results to field observation to improve model performance (Gutmann and Small 2010; Hogue et al. 2006; Kishné et al. 2017; Poltoradnev et al. 2018; Shellito et al. 2016; Yin et al. 2016). Thus, the LSM generated soil moisture becomes a model-dependent quantity and may cause inconsistency when used by different models (Koster et al. 2006, 2009).
Simulated subsurface water flow is dependent on the specified soil hydraulic properties. Therefore, the soil moisture and fluxes simulated by LSMs are affected by the soil parameterization. Soil hydraulic parameters used in LSMs are often simplified and may be limited to a small number of soil types based on textural classes from a lookup table. Soil moisture is a prime variable controlling the transfer of water and energy fluxes from land surfaces (Dirmeyer et al. 2006; Dong and Crow 2018; Xia et al. 2014; Koster et al. 2009). Several past studies have focused on evaluating the effects of soil parameterization on model performance in simulating land surface processes (Bi et al. 2016; Breuer et al. 2012; Chen et al. 2013; Gayler et al. 2013, 2014; Gochis et al. 2010; Garrigues et al. 2015).
Breuer et al. (2012) found that the use of higher-resolution spatial soil data in the land surface model produced significant variation in planetary boundary layer heights, which affected soil water content at field capacity and permanent wilting point, thus impacting the estimation of latent heat flux or ET. Similarly, in a field-based study by Gayler et al. (2013), the impact of detailed soil hydraulic properties and root growth on seasonal patterns of simulated ET and soil moisture were investigated. Their study suggested the need for more practical methods to estimate soil parameters for use in land surface models at catchment scales. Garrigues et al. (2015) found that the errors in the soil hydraulic parameters in the Interactions between the Soil, Biosphere, and Atmosphere (ISBA-A-gs) land surface model had the largest influence on ET when compared with uncertainties in the large-scale climate parameters. Bi et al. (2016) compared soil moisture predictions from GLDAS with in situ observations over the Tibetan plateau and found that errors in model parameters for soil properties caused biases in model prediction.
Soils are highly heterogeneous spatially, with distinct plant-dependent root systems (Gayler et al. 2014), soil textures (Yang et al. 2005), organic matter content (Chen et al. 2016), and stone fragment distributions (Cousin et al. 2003; Parajuli et al. 2017, 2019). In addition, soils may contain large openings, commonly known as macropores, which can have a dominating impact on subsurface flow and transport (Beven and Germann 1982, 2013). Characterizing heterogeneity and incorporating some of these concepts or features like macropores in land surface modeling at regional or global scales have proven to be difficult, due in part to a lack of large-scale experimental data. These complexities and variability of physical and hydraulic properties increases the uncertainty in calculated heat and water transfer rates within the soil (Koster and Suarez 1992; Li et al. 2013; Gayler et al. 2013; Ke et al. 2013). Nevertheless, accounting for more detailed subsurface properties and processes in LSMs is needed to improve simulation of soil moisture and water and energy fluxes between the land surface and atmosphere (Koster and Suarez 1992; Gayler et al. 2013; Li et al. 2013; Ke et al. 2013).
Soil moisture content simulated by LSMs is highly dependent on the process models and model parameterization. Differences in process model descriptions and parameterization methods can produce inconsistent results among different LSMs even when the models are driven by the same boundary forcing (Dirmeyer et al. 2006; Koster et al. 2009; Zheng et al. 2017). LSM simulation of soil moisture is recognized as being subject to errors from multiple sources (Song et al. 2014; Xue et al. 2013), but studies have suggested that development of methods for more detailed soil parameterization could improve the ability of LSMs to provide more accurate predictions (Gayler et al. 2013, 2014).
Current LSMs derive information from soil maps generated by the National Resources Conservation Service (NRCS) in the form of the Soil Survey Geographic (SSURGO) and the State Soil Geographic (STATSGO) databases over the United States. The SSURGO database is a digital general soil association map with the most detailed level of soil information produced with data collected and archived in 7.5-min topographic quadrangle units, while STATSGO is designed primarily for regional or multistate soil databases that are compiled by generalizing SSURGO soil survey maps. The Food and Agricultural Organization’s (FAO) soil databases are also used for global or regional simulations of LSMs. These were developed in the early 1990s at 1-km resolution for the contiguous United States (CONUS) and at 10-km resolution elsewhere (FAO 1991; Yang et al. 2011). These maps provide two soil layers: 0–30 and 30–100 cm. There has been notable improvement in soil mapping in recent years. For example, a new soil database known as “POLARIS” has remapped the SSURGO database using high-resolution geospatial environmental data and machine learning algorithms to obtain soil hydraulic parameters at various depths for CONUS at a 30-m spatial resolution (Chaney et al. 2016). The SoilGrids250m database is another noteworthy resource that provides global prediction for several soil properties such as organic carbon, bulk density, pH, soil texture fractions, and coarse fragments at seven standard depths (0, 5, 15, 30, 60, and 200 cm), with 250-m resolution (Hengl et al. 2017). These soil datasets offer options for including more detailed soil information in order to improve soil parameterization in LSMs at regional scales.
Challenges arise in trying to represent soil heterogeneity in natural settings including characteristics such as the coarse soil fraction, or stone content of a soil and its size distribution (particle size greater than 2 mm) throughout the soil profile (Cousin et al. 2003; Novák and Šurda 2010; Parajuli et al. 2017). Stone content in soil alters the physical properties affecting available water content and soil hydrodynamics, which further impacts root water uptake and ET (Cousin et al. 2003; Parajuli 2018). The magnitude of this effect depends upon the type and origin of stones, the volumetric fraction of stone content and the size and porosity of stones (Cousin et al. 2003; Parajuli et al. 2017; Naseri et al. 2019). Although the soil water holding capacity is largely affected by the presence of stones in the soil, which usually decrease the water available in the stony soil, the majority of hydrologic and land use studies ignore the presence of stones in the soil and likely overestimate ET. Other features that are typically ignored in LSMs include macropores in structured soils (Beven and Germann 2013). Macropores can allow for the rapid, preferential flow of water through soils during periods of intense rainfall or snowmelt. However, macropore sizes, distribution, and connectivity, are exceptionally difficult to characterize at large scales. Macropore flow is also not taken into account in Darcy’s law or the Richards equation (Richards 1931), and is therefore not within the scope of the current study.
The main objective of this study was to examine variously parameterized Noah land surface models with multiphysics option (Noah-MP) and to determine the impact of stone content on simulations of ET and moisture dynamics in stony soils. Noah-MP was selected for use in this study because of its application in a wide range of studies. Our approach was to first simulate ET using Noah-MP with default soil parameters from the lookup table with a single soil type for the entire soil profile. We then estimated ET using Noah-MP using five distinct soil layers, for which soil hydraulic parameters were again estimated for each layer from a lookup table. Then we adjusted the soil parameters based on textural information with and without accounting for stone content information obtained from field observations for each layer. The soil hydraulic parameter lookup table for Noah-MP is based on U.S. Department of Agriculture (USDA) soil textural classes. Detailed information about the experimental design of the numerical simulations is presented in the following sections.
2. Materials and methods
a. Site description and data
The data for this study were collected from the Low Sage Eddy Covariance (LS-EC) tower (43.14°N, 116.74°E) and the nearby Lower Sheep Weather Station (LSWS) located within the Reynolds Creek Experimental Watershed (RCEW) in southwestern Idaho. The site is located at an altitude of 1600 m above the sea level and is part of the Reynolds Creek Critical Zone Observatory. Climate in the RCEW varies with the montane elevation gradient, while the mean annual temperature and precipitation for the LSWS are 8.5°C and 345 mm, respectively (Fellows et al. 2017).
The LS-EC tower (AmeriFlux site U.S.-Rls) is located approximately 500 m from the LSWS and is mounted at a height of 2.09 m above the ground surface to measure water and carbon fluxes within the sagebrush ecosystem. Shortwave and longwave radiation, air temperature, and humidity were collected at the LS-EC tower every 30 min using a four-component net radiometer (Kipp and Zonen model CNR-1), and a temperature/humidity probe (Vaisala, Inc., model HMP155C). Soil heat flux was determined using six soil heat flux sensors (REBS model HFT3) installed 0.08 m deep within the soil, along with three sets of self-averaging thermocouples installed at 0.02 and 0.06 m. A three-dimensional sonic anemometer (Campbell Scientific, Inc., model CSAT3) and an open-path infrared gas analyzer (LI-COR, Inc., model LI-7500a) were sampled at 10 Hz as part of the eddy covariance system. Further details of the system and processing of the data are reported by Fellows et al. (2017).
Meteorological observations, including humidity, air temperature, wind speed, precipitation, and incoming shortwave radiation were obtained from the LSWS. These variables were used as atmospheric forcing data to drive a one-dimensional Noah-MP simulation. The data were collected every 15 min and processed at 30-min intervals. Precipitation was measured using a dual-gauge system especially designed for windy and snow-dominated conditions and processed hourly. Volumetric soil water content and soil temperature were recorded every hour at depths of 5, 15, 30, 60, and 90 cm using Stevens Water Monitoring Systems, Inc., model HydraProbe II soil moisture sensors.
During the process of soil moisture sensor installation, the excavated soil was analyzed to determine soil texture, bulk density, root distribution, stone content, etc. The site was highly heterogeneous in terms of soil distribution with significant stone content within the profile. The soil description for the site was obtained from Patton et al. (2017), presented in Table 1, along with soil information obtained from POLARIS, SSURGO, and FAO soil maps.
Soil textural descriptions of the low sage site soil profile at the Reynolds Creek Watershed based on field observation and information obtained from POLARIS, SSURGO, and FAO. Soil bulk density values are reported for the fine soil fraction without stone.
The soil textural analysis was performed for multiple horizons at increments of 10 cm. However, only five distinct soil layers were assumed for the improved Noah-MP simulation. Soils collected from the site had relatively high stone contents with nearly 40% stones by volume between 30 and 80 cm (Table 1; Fig. 1).
Distribution of stone content along the soil profile of the soil pit dug during installation of sensors at the Reynolds Creek Lower Sheep Weather Station.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Distribution of stone content along the soil profile of the soil pit dug during installation of sensors at the Reynolds Creek Lower Sheep Weather Station.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Distribution of stone content along the soil profile of the soil pit dug during installation of sensors at the Reynolds Creek Lower Sheep Weather Station.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
b. Noah-MP land surface model
Noah-MP evolved from the Noah LSM (Ek et al. 2003) to incorporate different schemes for runoff, leaf dynamics, stomatal resistance, and a soil moisture factor (Niu et al. 2011; Yang et al. 2011). In addition to the previously available Noah LSM, Noah-MP has added biophysical processes, such as accounting for an unconfined aquifer for groundwater storage and a dynamic water table, interactive vegetation canopy, multilayer snowpack, and a simple Topography Based Hydrological Model (TOPMODEL)-based runoff production function (Dickinson et al. 1998; Niu et al. 2005, 2007; Yang and Niu 2003; Cai et al. 2014b). While Noah-MP has mostly been used as a component of the Weather Research and Forecasting (WRF) Model, it is also available as a standalone one-dimensional offline version. In this study, the offline mode of Noah-MP, consistent with WRF, version 3.6, was used in simulating the land surface processes with atmospheric forcing.
The soil water flow component in the Noah-MP LSM is simulated using the diffusivity form of the Richards equation (Richards 1931), also called the “Richardson–Richards” equation by Raats and Knight (2018), who argued that this equation was first introduced by Richardson (1922):
where D is soil water hydraulic diffusivity (m2 s−1) and S represent the sink term. The upper boundary condition for the model is provided as the potential ET and precipitation, and the lower boundary is set to free drainage. The soil hydraulic diffusivity for one-dimensional vertical flow in soil is expressed in terms of the soil hydraulic conductivity K, matric potential h, and volumetric water content θ:
Note that, for simplicity, the absolute value of h is employed in Eq. (2) and the following equations.
The Clapp–Hornberger parameterization of water retention and hydraulic conductivity functions (Clapp and Hornberger 1978) are used with Eq. (1) to simulate the soil water flow (Brooks and Corey 1964; Chen and Dudhia 2001; Cosby et al. 1984):
where hs is the matric potential at air entry, often referred to as “bubbling pressure” (m); Ks is saturated soil hydraulic conductivity (m s−1); and b is a fitting parameter related to the soil pore-size distribution. The parameters in this study were determined from the Noah-MP soil parameter lookup table for the soil type based on soil textural class. The K(θ) and D(θ) functions are nonlinearly dependent on θ and vary by several orders of magnitude for small variation in θ as the soil gets drier.
To partition net energy within a grid cell, Noah-MP uses semitile subgrid scheme to account for land surface heterogeneity (Niu et al. 2011; Yang et al. 2011). In this scheme, the shortwave radiation transfer is computed for entire grid using a modified two-stream approximation assuming uniform distribution of vegetation over a grid cell. While, longwave radiation, ground heat, sensible heat, and latent heat are computed separately over two tiles with vegetation (Fveg) and bare fraction (1 − Fveg). Hence, the latent heat flux (LE) of a model grid cell is (Niu et al. 2011):
where LEg,b is the latent heat exchanged by the bare ground (i.e., 1 − Fveg) by emitting the part of net longwave radiation to the atmosphere, LEg,υ is the latent heat exchanged by the vegetative fractional area by emitting the part of ground-absorbed solar radiation as longwave radiation to the canopy, and LEυ is the latent heat exchanged by the vegetative fractional area by emitting the part of canopy-absorbed solar radiation as longwave radiation to the atmosphere.
c. Experimental design
In this study, four numerical experiments were designed to evaluate the performance of the Noah-MP LSM for simulating soil moisture and ET under different soil parameterization (summarized in Table 2). Each experiment was forced with the same set of atmospheric boundary conditions and vegetation parameters between 15 April and 28 September 2015. The first simulation (NMP I) was performed using default settings where a single soil type was considered for the entire soil profile, represented by a single set of soil parameters (θs, Ks, hs, and b) obtained from the Noah-MP soil parameter lookup table for general soil classification defined by the USDA soil textural class based on sand, silt and clay percentages obtained from field observation as presented in Table 1. In the second simulation (NMP II), five soil layers were defined; however, the soil parameters were obtained from the soil parameter lookup table as in NMP I. The third simulation (NMP III) used the same five soil layers with more descriptive soil parameters derived for each soil layer using pedotransfer functions based on textural information of sand, silt, and clay percentage and bulk density obtained from the field as presented in Table 1. The saturated water content θs was estimated by neural network prediction (Rosetta Lite, version 1.1; Schaap et al. 1998, 2001; Schaap and Bouten 1996; Šimůnek et al. 2016), requiring the same soil information on sand, silt, and clay percentage and bulk density. Additional parameters Ks, hs, and b in Eqs. (3) and (4) were estimated using the pedotransfer functions of Clapp and Hornberger (1978). In the fourth simulation (NMP IV), in addition to the soil parameterization in NMP III, we accounted for the stone content by adjusting the total water holding capacity of the soil (i.e., saturated water content) for each layer based on the stone porosity and volumetric stone content as shown in Table 1. The stone porosity was considered to be 15% assuming an average between highly porous (coarse sandstones) and negligibly porous (fine sandstones) based on previous studies (Parajuli et al. 2017, 2019). The effective water content at saturation for stony soil was calculated as a weighted average:
where θmix, θsoil, and θstone are volumetric water contents for stony soil, soil matrix, and stone fragments (m3 m−3) and υ is the volumetric stone content (m3 m−3). The saturated hydraulic conductivity Ks for the stony soil was assumed to be equal to that of the fine soil matrix, which is, however, dependent on position and volume fraction of the stone fragments in the soil profile (Hlaváčiková et al. 2016).
Summary of the four numerical experiments employed in this study.
Accurate measurement of the bulk stony soil moisture content is challenging because stone obstructions are generally avoided during sensor installation. The soil moisture sensors were installed in the pit wall avoiding visible stones that would bend rods or prevent insertion. Thus, the ability of soil moisture sensors to include stone content with the soil matrix is limited by their sensing volume and the size of the stone fragments as reported in previous studies (Coppola et al. 2013; Vaz et al. 2013). Hence, the measurements of soil volumetric water content should be adjusted for the effect of stone content in each soil layer. Figure 2 provides as an example of how a sensor might be installed in a stone-free part of the soil matrix, i.e., avoiding stones during installation and thus minimizing the effect of stones on the sensor measurement of soil moisture. In other words, water content sensors installed in stony soils are unlikely to yield soil moisture readings that are representative of the bulk stony soil. The sensor reading therefore would represent water content of the soil matrix θsoil, while the bulk stony soil water content would be generally lower as described in Eq. (6), which depends on stone porosity and stone volume fraction within the soil matrix. Based on this principle, measurements of soil water content at various depths were adjusted based on porosities of the soil matrix ηsoil and stone ηstone as well as the volumetric stone fraction υ information available for each soil layer. The adjustment factor kadj and the stone-adjusted soil water content θmix are written as
Schematic diagram showing the installation of a sensor in (a) soil without stone and (b) soil with stone to demonstrate that in both cases the sensor is typically reading the soil matrix water content except where small stones fit between/around sensor rods; i.e., sensor installers avoid stony areas.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Schematic diagram showing the installation of a sensor in (a) soil without stone and (b) soil with stone to demonstrate that in both cases the sensor is typically reading the soil matrix water content except where small stones fit between/around sensor rods; i.e., sensor installers avoid stony areas.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Schematic diagram showing the installation of a sensor in (a) soil without stone and (b) soil with stone to demonstrate that in both cases the sensor is typically reading the soil matrix water content except where small stones fit between/around sensor rods; i.e., sensor installers avoid stony areas.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
d. Evaluation statistics
In this study, the numerical model performance was evaluated based on the agreement between measurements and the values predicted by the Noah-MP model, indicated by the root-mean-square error (RMSE), average bias, coefficient of correlation r, and Nash–Sutcliffe efficiency (NSE) coefficient (Nash and Sutcliffe 1970), written as
where Oi and Mi are the observed and model predicted values of the same variable, respectively, and Oavg is the average of the observed values. The NSE ranges from negative infinity to 1 (perfect fit). In general, model prediction is considered acceptable when the NSE > 0.50 (Andersen et al. 2001; Shrestha et al. 2014).
Furthermore, the soil moisture anomaly was calculated, and abovementioned statics were computed by comparing simulated and stone-adjusted soil moisture anomalies. The soil moisture anomaly (SMA) was computed as
where θi is soil moisture content at each time step and θavg is long-term average of soil moisture calculated for the simulation period.
3. Results
a. Simulation of soil moisture
The sensor-measured and the stone-adjusted soil water content values for each layer are presented in Fig. 3. The stone-adjusted water content is substantially reduced relative to the measured value in layers 3 (30-cm depth) and 4 (60-cm depth), due to the high volumetric stone contents of 0.39 and 0.38 m3 m−3, respectively. The measured and stone-adjusted values of soil water content were compared with simulations from Noah-MP as described in Table 2 at 5-, 15-, 30-, 60-, and 90-cm depths (Fig. 3), where the simulation results are visibly closer to the stone-adjusted values than the sensor measurements. The soil moisture measurements show marked drydowns from the beginning of the simulation period (15 April) with a recharge to the 30-cm depth in mid-July. Considering a single soil type for the entire soil profile (NMP I), the Noah-MP simulation produced less drydown in the soil. Under this consideration, overestimation of soil moisture in the top two layers can be perceived from Fig. 3. A similar trend is observed in the second simulation (NMP II), where the soil profile is divided into five different layers based on a general soil textural class for each layer. The third soil moisture simulation (NMP III) revealed significant improvement in correlation with measurements when the soil hydraulic parameters were derived from the pedotransfer function based on actual observation of sand, silt, and clay fractions and bulk density.
Comparison of daily average sensor-measured, stone-adjusted, and modeled volumetric water contents from four Noah-MP simulation results (NMP I, NMP II, NMP III, and NMP IV). Observed daily precipitation from the simulation period was also included between 15 Apr and 28 Sep 2015.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Comparison of daily average sensor-measured, stone-adjusted, and modeled volumetric water contents from four Noah-MP simulation results (NMP I, NMP II, NMP III, and NMP IV). Observed daily precipitation from the simulation period was also included between 15 Apr and 28 Sep 2015.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Comparison of daily average sensor-measured, stone-adjusted, and modeled volumetric water contents from four Noah-MP simulation results (NMP I, NMP II, NMP III, and NMP IV). Observed daily precipitation from the simulation period was also included between 15 Apr and 28 Sep 2015.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
In general, Noah-MP simulation of soil moisture was improved with parameters based on actual field observation (NMP III, IV). Moreover, by considering the presence of stone content in the NMP IV simulation, soil moisture was predicted with higher accuracy when compared to stone-adjusted soil moisture observations. This can be observed especially for the layers 3 and 4 because of their greater overall impact on bulk soil moisture content due to higher stone content (~40%; Table 1). A notable disagreement in measured and simulated soil moisture was detected at the 30-cm depth for all four simulations. A low Ks value for the clay in the third layer could be a reason for restriction of water flow to the bottom of the layer. While the single computation node at the center of the layer poorly represents the measured soil water content at 30 cm. Thus, the impact of rain events was not well represented by the Noah-MP model’s soil water simulation at 30 cm.
Table 3 presents the RMSE, average bias, NSE, and r computed between the daily average of measured and stone-adjusted soil water content at the 5-, 15-, 30-, 60-, and 90-cm depths with those of the four Noah-MP simulations. Smaller RMSE values represent better agreement between measured and modeled values, whereas NSE values closer to 1 infer better model prediction. Table 3 shows that NMP III provided better result when compared with measured soil water content as the model did not account for the presence of stones. The lowest RMSE and highest NSE for NMP IV suggested that Noah-MP simulation of soil water content can be improved considerably with detailed information of soil texture and stone content. The simulated soil moisture from NMP I and NMP II are positively biased mostly for the first and second layers when compared with both sensor measured and stone-adjusted soil moisture. Whereas, a significant reduction in the average bias for soil moisture in all the layers can be observed for NMP IV (Table 3). Furthermore, the r suggests a significant improvement in soil moisture simulations using NMP IV, mostly in layers 3–5. To evaluate the temporal consistency, the statistics for soil moisture anomalies from observation (stone adjusted) and simulation were computed (Table 4). The results showed improvement in NMP IV, with reduced RMSE and least average bias in all layers. Thus, using soil parameters derived from actual physical properties of the soil including detailed layering in addition to information on stone content has the potential to improve LSM estimation of soil water dynamics.
RMSE, average bias, NSE, and correlation coefficient values of simulated results at 5-, 15-, 30-, 60-, and 90-cm depths for each of four simulations (NMP I, NMP II, NMP III, and NMP IV) described in Table 2. Simulated results are compared with the daily average of measured (left column) and stone-adjusted (right column with italics) soil water contents for each simulation and depth.
RMSE, average bias, and NSE values between the anomalies of simulated results at 5-, 15-, 30-, 60-, and 90-cm depths for each of four simulations (NMP I, NMP II, NMP III, and NMP IV) described in Table 2 and daily average of stone-adjusted soil water contents for each depth.
b. Evaluation of evapotranspiration estimates
The daily averaged values of eddy covariance measured latent heat flux were compared with the simulated values produced by the four Noah-MP simulations as presented in Fig. 4. The Noah-MP simulated latent heat flux values were overestimated for NMP I and NMP II simulations at the beginning of the simulation period but were able to match the observations during drier periods. The Noah-MP simulations with improved soil parameters (NMP III) and with stone content information (NMP IV) matched the eddy covariance estimates well, especially at the end of the period. The overestimation of ET can be observed following rain events in mid-July for NMP I and NMP II. For the same period, NMP III and NMP IV performed well.
Time series of daily average of measured latent heat flux from the LS-EC tower compared with the four Noah-MP simulations as detailed in Table 2.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Time series of daily average of measured latent heat flux from the LS-EC tower compared with the four Noah-MP simulations as detailed in Table 2.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Time series of daily average of measured latent heat flux from the LS-EC tower compared with the four Noah-MP simulations as detailed in Table 2.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
The error statistics between the measured and simulated latent heat flux under different experiments is presented with scatterplots in Fig. 5. The error statistics between the measured and simulated latent heat fluxes are shown in terms of r2, NSE, bias, and RMSE. The lines of perfect fit (1:1 line), 95% prediction interval and regression lines are also shown. The scatterplots in Fig. 5 exhibit improvement in correlation between the measured latent heat flux with those simulated by Noah-MP using improved soil parameterization. The highest r2 and NSE of 0.75 and 0.62, respectively, as well as lowest RMSE and bias of 13.4 W m−2 (0.5 mm day−1) and −0.07 W m−2, respectively (Fig. 5), suggests the improved soil parameterization with consideration of stone content in NMP IV more accurately simulated latent heat fluxes or ET.
Scatterplots between daily averaged latent heat flux from measured by the LS-EC tower and estimated by four Noah-MP simulations (Table 2). Regression lines and prediction intervals with 95% confidence are also shown. The coefficient of determination r2, Nash–Sutcliffe model efficiency coefficient (NSE), average bias, and root-mean-square error (RMSE) values computed between measured and simulated latent heat flux are reported.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Scatterplots between daily averaged latent heat flux from measured by the LS-EC tower and estimated by four Noah-MP simulations (Table 2). Regression lines and prediction intervals with 95% confidence are also shown. The coefficient of determination r2, Nash–Sutcliffe model efficiency coefficient (NSE), average bias, and root-mean-square error (RMSE) values computed between measured and simulated latent heat flux are reported.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Scatterplots between daily averaged latent heat flux from measured by the LS-EC tower and estimated by four Noah-MP simulations (Table 2). Regression lines and prediction intervals with 95% confidence are also shown. The coefficient of determination r2, Nash–Sutcliffe model efficiency coefficient (NSE), average bias, and root-mean-square error (RMSE) values computed between measured and simulated latent heat flux are reported.
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
The cumulative ET simulated under NMP IV (Noah-MP with stone content) over the simulation period (15 April–28 September 2015) produced the closest match with the measured cumulative ET of 302 mm (Fig. 6). The cumulative ET predicted by NMP I and NMP II through June were similar to measured cumulative ET. This suggests that effects of stone content on ET under wet conditions may be negligible when root water uptake from the soil matrix satisfies the demand (Parajuli et al. 2019). However, later in the summer, an overestimation of ET is clearly visible in the NMP I and NMP II simulations where stone content is neglected (Fig. 6). The NMP IV estimates of ET considering stone content yielded closer results when compared with the eddy covariance measurements (Fig. 6).
Cumulative daily evapotranspiration derived from latent heat flux measured by the LS-EC tower compared with the four Noah-MP simulations (Table 2) over the simulation period (15 Apr–28 Sep 2015).
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Cumulative daily evapotranspiration derived from latent heat flux measured by the LS-EC tower compared with the four Noah-MP simulations (Table 2) over the simulation period (15 Apr–28 Sep 2015).
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
Cumulative daily evapotranspiration derived from latent heat flux measured by the LS-EC tower compared with the four Noah-MP simulations (Table 2) over the simulation period (15 Apr–28 Sep 2015).
Citation: Journal of Hydrometeorology 21, 8; 10.1175/JHM-D-19-0075.1
4. Discussion
Soil water has substantial influence on surface energy balance as it influences the partitioning of available energy into soil, latent, and sensible heat fluxes. Increased soil water increases water available for soil evaporation and plant transpiration, thereby directly impacting the surface energy balance. Furthermore, increased soil water increases soil thermal conductivity inducing higher heat transport into the soil as soil heat flux. Hence, the components of the energy balance either influence or are impacted by soil water, highlighting their importance in simulation of soil water content that is critically tied to heat fluxes and ET. Compared to measured values, simulated soil water content considering soil stone information (NMP IV) yielded reduced water content values in layers where stone content was highest at 30- and 60-cm depths (Fig. 3). This discrepancy was addressed by assuming that sensor-based measurements did not account for stone influence simply because sensors are inserted to avoid stones. Although the NMP IV simulation provided better estimates of ET, soil water content estimates showed high RMSE and low NSE in most layers (Table 3). After adjusting soil water content measured values based on stone content information, the match between adjusted measurements and simulated water contents was substantially improved as indicated by the lowest RMSE values and highest NSE values at all layers for the NMP IV simulations (Table 3). The RMSE between sensor-measured soil moisture and that predicted by the NMP IV simulations at the 30-cm depth was 0.068 m3 m−3 (Table 3). The higher RMSE in this layer was likely due to the inefficiency of the Noah-MP model in representing the hydraulic behavior of the high clay content layer. Furthermore, the comparison of soil water content for the layer at 30 cm simulated by Noah-MP, whose computation node lies in the middle of the layer, is not spatially well aligned with the sensor located at the bottom of the third layer, which is close to the sandy-clay layer in the fourth layer. Thus, the mismatch between the vertical scale of soil moisture simulated by Noah-MP and the point observations may have contributed to the bias in soil moisture estimation at different layers. A more comprehensive investigation on the errors associated with sensors installed in stony soils is needed to better understand how to accurately monitor stony soil water content.
The latent heat flux simulated by the Noah-MP model depends on water availability in the root zone, thus the effect of modification in soil parameterization can be seen in the ET estimation (Fig. 5). As expected, the best result of ET estimation was obtained from the NMP IV simulation where the soil parameters were specific to each layer, each of which also accounted for stone content. Statistically, the RMSE values were reduced to 13 from 20 W m−2, and the NSE was improved from 0.19 to 0.62 (Fig. 5), this despite the inherent uncertainties associated with the ET simulation. In addition, although eddy covariance represents the most direct and accurate determinations of latent heat flux and ET at the local scale, there are limitations in its application as the vapor flux footprint varies with wind speed and direction. The flux footprint of eddy covariance measurements extends well beyond the point scale simulations using Noah-MP but for one-dimensional vertical fluxes these should be highly correlated. Furthermore, instrumentation error and site condition can affect the flux measurements. Ryu et al. (2008) reported the uncertainty in ET measured by eddy covariance over a California grassland to be nearly 9% at the 90% confidence level. The NMP IV simulations of ET yielded an NSE value of 0.62 while achieving the lowest RMSE (Fig. 5). Regardless of the difference between spatial scales of the eddy covariance footprint and the single column simulation of Noah-MP, the results suggest that the improvement in soil parameterization considering effects of stone content can notably advance estimates of ET using land surface models in stony soil ecosystems.
5. Conclusions
Stony soils are common in many natural ecosystems such as mountains and forest environments. The presence of stones in soil can reduce porosity and permeability and typically reduces the available water for plants because stones often have lower porosity than the soil matrix. Although soil water content measurements are made in stony soils, sensing volumes are typically not representative of the bulk stony soil water content due to the small measurement volumes of the sensors and their placement within the soil matrix, often away from stones during installation. With this assumption in mind, we applied an adjustment to the soil water content measurement based on the volumetric stone content and on the stone properties as a corrected representation of stony soil water content. The performance of Noah-MP in simulating soil water content and evapotranspiration was evaluated using various representations of soil and stone characteristics. These comparisons revealed varied influences of stone and soil matrix on simulated soil water content and evapotranspiration, with the best results coming as expected from the simulation which accounted for soil stone content. This result points to the potentially substantial role of stones in modulating soil water dynamics and evapotranspiration, while the magnitude of stone influence depends largely on the type and porosity of stones. Hence, in addition to the higher resolution representation of subsurface properties, it is equally important to account for the presence of stone content. Montane forest ecosystems typically have soils with higher stone content that exhibit a considerable impact on soil water dynamics, evapotranspiration and water balance processes. While this study was based on single location, future research focusing on extracting and characterizing stone content information and incorporating this information into large-scale land surface model applications will potentially result in more accurate simulation of soil water content and evapotranspiration.
Acknowledgments
This research was supported by the iUTAH project funded through an NSF EPSCoR Grant EPS 1208732 awarded to Utah State University, as part of the State of Utah Research Infrastructure Improvement Award. This research was also supported by the Utah Agricultural Experiment Station (Seed Grant, Project UTA01189), Utah State University, and approved as journal paper number 9188. The Reynolds Creek Critical Zone Observatory is supported by the National Science Foundation under Award NSF EAR 1331872. Any opinions, findings, and conclusions or recommendations expressed are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
REFERENCES
Andersen, J., J. C. Refsgaard, and K. H. Jensen, 2001: Distributed hydrological modelling of the Senegal River Basin—Model construction and validation. J. Hydrol., 247, 200–214, https://doi.org/10.1016/S0022-1694(01)00384-5.
Barlage, M., M. Tewari, F. Chen, G. Miguez-Macho, Z. L. Yang, and G. Y. Niu, 2015: The effect of groundwater interaction in North American regional climate simulations with WRF/Noah-MP. Climatic Change, 129, 485–498, https://doi.org/10.1007/s10584-014-1308-8.
Beven, K., and P. Germann, 1982: Macropores and water flow in soils. Water Resour. Res., 18, 1311–1325, https://doi.org/10.1029/WR018i005p01311.
Beven, K., and P. Germann, 2013: Macropores and water flow in soils revisited. Water Resour. Res., 49, 3071–3092, https://doi.org/10.1002/wrcr.20156.
Bi, H., J. Ma, W. Zheng, and J. Zeng, 2016: Comparison of soil moisture in GLDAS model simulations and in situ observations over the Tibetan Plateau. J. Geophys. Res. Atmos., 121, 2658–2678, https://doi.org/10.1002/2015JD024131.
Breuer, H., F. Ács, Á. Horváth, B. Laza, I. Matyasovszky, P. Németh, T. Weidinger, and K. Rajkai, 2012: A sensitivity study on the soil parameter-boundary layer height interrelationship. ISRN Meteor., 2012, 786 592, https://doi.org/10.5402/2012/786592.
Brooks, R. H., and A. T. Corey, 1964: Hydraulic properties of porous media and their relation to drainage design. Trans. ASAE, 7, 0026–0028, https://doi.org/10.13031/2013.40684.
Cai, X., Z. L. Yang, C. H. David, G. Y. Niu, and M. Rodell, 2014a: Hydrological evaluation of the Noah-MP land surface model for the Mississippi River Basin. J. Geophys. Res. Atmos., 119, 23–38, https://doi.org/10.1002/2013JD020792.
Cai, X., Z. L. Yang, Y. Xia, M. Huang, H. Wei, L. R. Leung, and M. B. Ek, 2014b: Assessment of simulated water balance from Noah, Noah-MP, CLM, and VIC over CONUS using the NLDAS test bed. J. Geophys. Res. Atmos., 119, 13 751–13 770, https://doi.org/10.1002/2014JD022113.
Chaney, N. W., E. F. Wood, A. B. McBratney, J. W. Hempel, T. W. Nauman, C. W. Brungard, and N. P. Odgers, 2016: POLARIS: A 30-meter probabilistic soil series map of the contiguous United States. Geoderma, 274, 54–67, https://doi.org/10.1016/j.geoderma.2016.03.025.
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, https://doi.org/10.1175/1520-0493(2001)129<0569:CAALSH>2.0.CO;2.
Chen, F., and Coauthors, 1996: Modeling of land surface evaporation by four schemes and comparison with FIFE observations. J. Geophys. Res. Atmos., 101, 7251–7268, https://doi.org/10.1029/95JD02165.
Chen, L., Y. Li, F. Chen, A. Barr, M. Barlage, and B. Wan, 2016: The incorporation of an organic soil layer in the Noah-MP land surface model and its evaluation over a boreal aspen forest. Atmos. Chem. Phys., 16, 8375–8387, https://doi.org/10.5194/acp-16-8375-2016.
Chen, Y., K. Yang, J. Qin, L. Zhao, W. Tang, and M. Han, 2013: Evaluation of AMSR-E retrievals and GLDAS simulations against observations of a soil moisture network on the central Tibetan Plateau. J. Geophys. Res. Atmos., 118, 4466–4475, https://doi.org/10.1002/JGRD.50301.
Clapp, R. B., and G. M. Hornberger, 1978: Empirical equations for some soil hydraulic properties. Water Resour. Res., 14, 601–604, https://doi.org/10.1029/WR014i004p00601.
Coppola, A., G. Dragonetti, A. Comegna, N. Lamaddalena, B. Caushi, M. A. Haikal, and A. Basile, 2013: Measuring and modeling water content in stony soils. Soil Tillage Res., 128, 9–22, https://doi.org/10.1016/j.still.2012.10.006.
Cosby, B. J., G. M. Hornberger, R. B. Clapp, and T. Ginn, 1984: A statistical exploration of the relationships of soil moisture characteristics to the physical properties of soils. Water Resour. Res., 20, 682–690, https://doi.org/10.1029/WR020i006p00682.
Cousin, I., B. Nicoullaud, and C. Coutadeur, 2003: Influence of rock fragments on the water retention and water percolation in a calcareous soil. Catena, 53, 97–114, https://doi.org/10.1016/S0341-8162(03)00037-7.
Crow, W. T., R. D. Koster, R. H. Reichle, and H. O. Sharif, 2005: Relevance of time-varying and time-invariant retrieval error sources on the utility of spaceborne soil moisture products. Geophys. Res. Lett., 32, L24405, https://doi.org/10.1029/2005GL024889.
Delworth, T. L., and S. Manabe, 1988: The influence of potential evaporation on the variabilities of simulated soil wetness and climate. J. Climate, 1, 523–547, https://doi.org/10.1175/1520-0442(1988)001<0523:TIOPEO>2.0.CO;2.
Delworth, T. L., and S. Manabe, 1993: Climate variability and land-surface processes. Adv. Water Resour., 16, 3–20, https://doi.org/10.1016/0309-1708(93)90026-C.
Dickinson, R. E., M. Shaikh, R. Bryant, and L. Graumlich, 1998: Interactive canopies for a climate model. J. Climate, 11, 2823–2836, https://doi.org/10.1175/1520-0442(1998)011<2823:ICFACM>2.0.CO;2.
Dirmeyer, P. A., 1995: Problems in initializing soil wetness. Bull. Amer. Meteor. Soc., 76, 2234–2240, https://doi.org/10.1175/1520-0477-76.11.2211.
Dirmeyer, P. A., and J. Shukla, 1994: Albedo as a modulator of climate response to tropical deforestation. J. Geophys. Res., 99, 20 863–20 877, https://doi.org/10.1029/94JD01311.
Dirmeyer, P. A., X. Gao, M. Zhao, Z. Guo, T. Oki, and N. Hanasaki, 2006: GSWP-2: Multimodel analysis and implications for our perception of the land surface. Bull. Amer. Meteor. Soc., 87, 1381–1398, https://doi.org/10.1175/BAMS-87-10-1381.
Dong, J., and W. T. Crow, 2018: Use of satellite soil moisture to diagnose climate model representations of European soil moisture-air temperature coupling strength. Geophys. Res. Lett., 45, 12 884–12 891, https://doi.org/10.1029/2018GL080547.
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, 369–383, https://doi.org/10.1007/BF00713776.
Ek, M., 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, https://doi.org/10.1029/2002JD003296.
FAO, 1991: The Digitized Soil Map of the World (release 1.0). Rep. 67/1, Food and Agriculture Organization.
Fellows, A. W., G. N. Flerchinger, M. S. Seyfried, and L. Kathleen, 2017: Data for partitioned carbon and energy fluxes within the Reynolds Creek Critical Zone Observatory. Boise State University, accessed 17 February 2018, https://doi.org/10.18122/B2TD7V.
Ford, T. W., and S. M. Quiring, 2014: Comparison and application of multiple methods for temporal interpolation of daily soil moisture. Int. J. Climatol., 34, 2604–2621, https://doi.org/10.1002/joc.3862.
Gao, Y., K. Li, F. Chen, Y. Jiang, and C. Lu, 2015: Assessing and improving Noah-MP land model simulations for the central Tibetan Plateau. J. Geophys. Res. Atmos., 120, 9258–9278, https://doi.org/10.1002/2015JD023404.
Garrigues, S., and Coauthors, 2015: Evaluation of land surface model simulations of evapotranspiration over a 12-year crop succession: Impact of soil hydraulic and vegetation properties. Hydrol. Earth Syst. Sci., 19, 3109–3131, https://doi.org/10.5194/hess-19-3109-2015.
Gayler, S., J. Ingwersen, E. Priesack, T. Wöhling, V. Wulfmeyer, and T. Streck, 2013: Assessing the relevance of subsurface processes for the simulation of evapotranspiration and soil moisture dynamics with CLM3.5: Comparison with field data and crop model simulations. Environ. Earth Sci., 69, 415–427, https://doi.org/10.1007/s12665-013-2309-z.
Gayler, S., and Coauthors, 2014: Incorporating dynamic root growth enhances the performance of Noah-MP at two contrasting winter wheat field sites. Water Resour. Res., 50, 1337–1356, https://doi.org/10.1002/2013WR014634.
Gochis, D. J., E. R. Vivoni, and C. J. Watts, 2010: The impact of soil depth on land surface energy and water fluxes in the North American monsoon region. J. Arid Environ., 74, 564–571, https://doi.org/10.1016/j.jaridenv.2009.11.002.
Goodrich, D. C., T. J. Schmugge, T. J. Jackson, C. L. Unkrich, T. O. Keefer, R. Parry, L. B. Bach, and S. A. Am, 1994: Runoff simulation sensitivity to remotely sensed initial soil water content. Water Resour. Res., 30, 1393–1405, https://doi.org/10.1029/93WR03083.
Gutmann, E. D., and E. E. Small, 2005: The effect of soil hydraulic properties vs. soil texture in land surface models. Geophys. Res. Lett., 32, L02402, https://doi.org/10.1029/2004GL021843.
Gutmann, E. D., and E. E. Small, 2007: A comparison of land surface model soil hydraulic properties estimated by inverse modeling and pedotransfer functions. Water Resour. Res., 43, W05418, https://doi.org/10.1029/2006WR005135.
Gutmann, E. D., and E. E. Small, 2010: A method for the determination of the hydraulic properties of soil from MODIS surface temperature for use in land-surface models. Water Resour. Res., 46, W06520, https://doi.org/10.1029/2009WR008203.
Heathman, G. C., M. Larose, M. H. Cosh, and R. Bindlish, 2009: Surface and profile soil moisture spatio-temporal analysis during an excessive rainfall period in the southern Great Plains, USA. Catena, 78, 159–169, https://doi.org/10.1016/j.catena.2009.04.002.
Hengl, T., and Coauthors, 2017: SoilGrids250m: Global gridded soil information based on machine learning. PLOS ONE, 12, e0169748, https://doi.org/10.1371/journal.pone.0169748.
Hlaváčiková, H., V. Novák, and J. Šimůnek, 2016: The effects of rock fragment shapes and positions on modeled hydraulic conductivities of stony soils. Geoderma, 281, 39–48, https://doi.org/10.1016/j.geoderma.2016.06.034.
Hogue, T. S., L. A. Bastidas, H. V. Gupta, and S. Sorooshian, 2006: Evaluating model performance and parameter behavior for varying levels of land surface model complexity. Water Resour. Res., 42, W08430, https://doi.org/10.1029/2005WR004440.
Ke, Y., L. R. Leung, M. Huang, and H. Li, 2013: Enhancing the representation of subgrid land surface characteristics in land surface models. Geosci. Model Dev., 6, 1609–1622, https://doi.org/10.5194/gmd-6-1609-2013.
Kishné, A. S., Y. T. Yimam, C. L. Morgan, and B. C. Dornblaser, 2017: Evaluation and improvement of the default soil hydraulic parameters for the Noah Land Surface Model. Geoderma, 285, 247–259, https://doi.org/10.1016/j.geoderma.2016.09.022.
Koster, R. D., and M. J. Suarez, 1992: A comparative analysis of two land surface heterogeneity representations. J. Climate, 5, 1379–1390, https://doi.org/10.1175/1520-0442(1992)005<1379:ACAOTL>2.0.CO;2.
Koster, R. D., and Coauthors, 2006: GLACE: The Global Land–Atmosphere Coupling Experiment. Part I: Overview. J. Hydrometeor., 7, 590–610, https://doi.org/10.1175/JHM510.1.
Koster, R. D., Z. Guo, R. Yang, P. A. Dirmeyer, K. Mitchell, and M. J. Puma, 2009: On the nature of soil moisture in land surface models. J. Climate, 22, 4322–4335, https://doi.org/10.1175/2009JCLI2832.1.
Kumar, A., F. Chen, M. Barlage, M. B. Ek, and D. Niyogi, 2014: Assessing impacts of integrating MODIS vegetation data in the Weather Research and Forecasting (WRF) Model coupled to two different canopy-resistance approaches. J. Appl. Meteor. Climatol., 53, 1362–1380, https://doi.org/10.1175/JAMC-D-13-0247.1.
Li, H., M. S. Wigmosta, H. Wu, M. Huang, Y. Ke, A. M. Coleman, and L. R. Leung, 2013: A physically based runoff routing model for land surface and Earth system models. J. Hydrometeor., 14, 808–828, https://doi.org/10.1175/JHM-D-12-015.1.
Long, D., L. Longuevergne, and B. R. Scanlon, 2014: Uncertainty in evapotranspiration from land surface modeling, remote sensing, and GRACE satellites. Water Resour. Res., 50, 1131–1151, https://doi.org/10.1002/2013WR014581.
Mitchell, K. E., and Coauthors, 2004: The multi-institution North American Land Data Assimilation System (NLDAS): Utilizing multiple GCIP products and partners in a continental distributed hydrological modeling system. J. Geophys. Res., 109, D07S90, https://doi.org/10.1029/2003JD003823.
Namias, J., 1952: The annual course of month-to-month persistence in climatic anomalies. Bull. Amer. Meteor. Soc., 33, 279–285, https://doi.org/10.1175/1520-0477-33.7.279.
Naseri, M., S. C. Iden, N. Richter, and W. Durner, 2019: Influence of stone content on soil hydraulic properties: Experimental investigation and test of existing model concepts. Vadose Zone J., 18, 1–10, https://doi.org/10.2136/vzj2018.08.0163.
Nash, J. E., and J. V. Sutcliffe, 1970: River flow forecasting through conceptual models Part I—A discussion of principles. J. Hydrol., 10, 282–290, https://doi.org/10.1016/0022-1694(70)90255-6.
Niu, G.-Y., Z.-L. Yang, R. E. Dickinson, and L. E. Gulden, 2005: A simple TOPMODEL-based runoff parameterization (SIMTOP) for use in global climate models. J. Geophys. Res., 110, D21106, https://doi.org/10.1029/2005JD006111.
Niu, G.-Y., Z.-L. Yang, R. E. Dickinson, L. E. Gulden, and H. Su, 2007: Development of a simple groundwater model for use in climate models and evaluation with gravity recovery and climate experiment data. J. Geophys. Res., 112, D07103, https://doi.org/10.1029/2006JD007522.
Niu, G.-Y., and Coauthors, 2011: The community Noah land surface model with multiparameterization options (Noah-MP): 1. Model description and evaluation with local-scale measurements. J. Geophys. Res., 116, D12109, https://doi.org/10.1029/2010JD015139.
Novák, V., and P. Šurda, 2010: The water retention of a granite rock fragments in High Tatras stony soils. J. Hydrol. Hydromech., 58, 181–187, https://doi.org/10.2478/v10098-010-0017-x.
Oleson, K., and Coauthors, 2013: Technical description of version 4.5 of the Community Land Model (CLM). NCAR Tech. Note NCAR/TN-503+STR, 420 pp., https://doi.org/10.5065/D6RR1W7M.
Parajuli, K., 2018: Advancing methods to quantify actual evapotranspiration in stony soil ecosystems. Ph.D. dissertation, Utah State University, 139 pp., https://digitalcommons.usu.edu/etd/7242.
Parajuli, K., M. Sadeghi, and S. B. Jones, 2017: A binary mixing model for characterizing stony-soil water retention. Agric. For. Meteor., 244–245, 1–8, https://doi.org/10.1016/j.agrformet.2017.05.013.
Parajuli, K., S. B. Jones, D. G. Tarboton, G. N. Flerchinger, L. E. Hipps, L. N. Allen, and M. S. Seyfried, 2019: Estimating actual evapotranspiration from stony-soils in montane ecosystems. Agric. For. Meteor., 265, 183–194, https://doi.org/10.1016/j.agrformet.2018.11.019.
Patton, N. R., K. A. Lohse, S. E. Godsey, M. S. Seyfried, and B. T. Crosby, 2017: Dataset for predicting soil thickness on soil mantled hillslopes. Boise State University, accessed 17 February 2018, https://doi.org/10.18122/B2PM69.
Poltoradnev, M., J. Ingwersen, K. Imukova, P. Högy, H. D. Wizemann, and T. Streck, 2018: How well does Noah-MP simulate the regional mean and spatial variability of topsoil water content in two agricultural landscapes in southwest Germany? J. Hydrometeor., 19, 555–573, https://doi.org/10.1175/JHM-D-17-0169.1.
Qiu, J., W. T. Crow, J. Dong, and G. S. Nearing, 2020: Model representation of the coupling between evapotranspiration and soil water content at different depths. Hydrol. Earth Syst. Sci., 24, 581–594, https://doi.org/10.5194/hess-24-581-2020.
Quiring, S. M., T. W. Ford, J. K. Wang, A. Khong, E. Harris, T. Lindgren, D. W. Goldberg, and Z. Li, 2016: The North American soil moisture database: Development and applications. Bull. Amer. Meteor. Soc., 97, 1441–1459, https://doi.org/10.1175/BAMS-D-13-00263.1.
Raats, P. A., and J. Knight, 2018: The contributions of Lewis Fry Richardson to drainage theory, soil physics, and the soil-plant-atmosphere continuum. Front. Environ. Sci., 6, 13, https://doi.org/10.3389/fenvs.2018.00013.
Richards, L. A., 1931: Capillary conduction of liquids through porous mediums. Physics, 1, 318–333, https://doi.org/10.1063/1.1745010.
Richardson, L. F., 1922: Weather Prediction by Numerical Process. Cambridge University Press, 237 pp.
Rodell, M., and Coauthors, 2004: The Global Land Data Assimilation System. Bull. Amer. Meteor. Soc., 85, 381–394, https://doi.org/10.1175/BAMS-85-3-381.
Ryu, S. R., J. Chen, A. Noormets, M. K. Bresee, and S. V. Ollinger, 2008: Comparisons between PnET-Day and eddy covariance based gross ecosystem production in two Northern Wisconsin forests. Agric. For. Meteor., 148, 247–256, https://doi.org/10.1016/j.agrformet.2007.08.005.
Sadeghi, M., M. Tuller, A. W. Warrick, E. Babaeian, K. Parajuli, M. R. Gohardoust, and S. B. Jones, 2019: An analytical model for estimation of land surface net water flux from near-surface soil moisture observations. J. Hydrol., 570, 26–37, https://doi.org/10.1016/j.jhydrol.2018.12.038.
Schaap, M. G., and W. Bouten, 1996: Modeling water retention curves of sandy soils using neural networks. Water Resour. Res., 32, 3033–3040, https://doi.org/10.1029/96WR02278.
Schaap, M. G., F. J. Leij, and M. Th. van Genuchten, 1998: Neural network analysis for hierarchical prediction of soil hydraulic properties. Soil. Sci. Soc. Amer. J., 62, 847–855, https://doi.org/10.2136/sssaj1998.03615995006200040001x.
Schaap, M. G., F. J. Leij, and M. Th. van Genuchten, 2001: Rosetta: A computer program for estimating soil hydraulic parameters with hierarchical pedotransfer functions. J. Hydrol., 251, 163–176, https://doi.org/10.1016/S0022-1694(01)00466-8.
Scott, R., D. Entekhabi, R. Koster, and M. Suarez, 1997: Timescales of land surface evapotranspiration response. J. Climate, 10, 559–566, https://doi.org/10.1175/1520-0442(1997)010<0559:TOLSER>2.0.CO;2.
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, 125–161, https://doi.org/10.1016/j.earscirev.2010.02.004.
Seneviratne, S. I., and Coauthors, 2013: Impact of soil moisture-climate feedbacks on CMIP5 projections: First results from the GLACE-CMIP5 experiment. Geophys. Res. Lett., 40, 5212–5217, https://doi.org/10.1002/grl.50956.
Shellito, P. J., E. E. Small, and M. H. Cosh, 2016: Calibration of Noah soil hydraulic property parameters using surface soil moisture from SMOS and basinwide in situ observations. J. Hydrometeor., 17, 2275–2292, https://doi.org/10.1175/JHM-D-15-0153.1.
Shrestha, S., M. Khatiwada, M. S. Babel, and K. Parajuli, 2014: Impact of climate change on river flow and hydropower production in Kulekhani hydropower project of Nepal. Environ. Processes, 1, 231–250, https://doi.org/10.1007/s40710-014-0020-z.
Šimůnek, J., M. T. Van Genuchten, and M. Šejna, 2016: Recent developments and applications of the HYDRUS computer software packages. Vadose Zone J., 15, 1–25, https://doi.org/10.2136/vzj2016.04.0033.
Song, M., Y. Ma, Y. Zhang, W.