Impacts of Land Surface Parameterizations on Simulations over the Arid and Semiarid Regions: The Case of the Loess Plateau in China

Sha Lu aSchool of Atmospheric Sciences, Nanjing University, Nanjing, China
bJoint International Research Laboratory of Atmospheric and Earth System Sciences, Nanjing, China

Search for other papers by Sha Lu in
Current site
Google Scholar
PubMed
Close
,
Weidong Guo aSchool of Atmospheric Sciences, Nanjing University, Nanjing, China
bJoint International Research Laboratory of Atmospheric and Earth System Sciences, Nanjing, China

Search for other papers by Weidong Guo in
Current site
Google Scholar
PubMed
Close
,
Jun Ge aSchool of Atmospheric Sciences, Nanjing University, Nanjing, China
bJoint International Research Laboratory of Atmospheric and Earth System Sciences, Nanjing, China

Search for other papers by Jun Ge in
Current site
Google Scholar
PubMed
Close
, and
Yu Zhang cPlateau Atmosphere and Environment Key Laboratory of Sichuan Province, School of Atmospheric Sciences, Chengdu University of Information Technology, Chengdu, China

Search for other papers by Yu Zhang in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

The arid and semiarid areas of the Loess Plateau are extremely sensitive to climate change. Land–atmosphere interactions of these regions play an important role in the regional climate. However, most present land surface models (LSMs) are not reasonable and accurate enough to describe the surface characteristics in these regions. In this study, we investigate the effects of three key land surface parameters including surface albedo, soil thermal conductivity, and additional damping on the Noah LSM in simulating the land surface characteristics. The observational data from June to September from 2007 to 2009 collected at the Semi-Arid Climate and Environment Observatory of Lanzhou University (SACOL) station in northwest China are used to validate the Noah LSM simulations. The results suggest that the retrieved values of surface albedo, soil thermal conductivity, and additional damping based on observations are in closer agreement with those of the MULT scheme for surface albedo, the J75_NOAH scheme for soil thermal conductivity, and the Y08 scheme for additional damping, respectively. Furthermore, the model performance is not obviously affected by surface albedo parameterization schemes, while the scheme of soil thermal conductivity is vital to simulations of latent heat flux and soil temperature and the scheme of additional damping is crucial for simulating net radiation flux, sensible heat flux, and surface soil temperature. A set of optimal parameterizations is proposed for the offline Noah LSM at the SACOL station when the MULT scheme for surface albedo, the J75_NOAH scheme for soil thermal conductivity, and the Y08 scheme for additional damping are combined simultaneously, especially in the case of sensible heat flux and surface soil temperature simulations.

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Weidong Guo, guowd@nju.edu.cn

Abstract

The arid and semiarid areas of the Loess Plateau are extremely sensitive to climate change. Land–atmosphere interactions of these regions play an important role in the regional climate. However, most present land surface models (LSMs) are not reasonable and accurate enough to describe the surface characteristics in these regions. In this study, we investigate the effects of three key land surface parameters including surface albedo, soil thermal conductivity, and additional damping on the Noah LSM in simulating the land surface characteristics. The observational data from June to September from 2007 to 2009 collected at the Semi-Arid Climate and Environment Observatory of Lanzhou University (SACOL) station in northwest China are used to validate the Noah LSM simulations. The results suggest that the retrieved values of surface albedo, soil thermal conductivity, and additional damping based on observations are in closer agreement with those of the MULT scheme for surface albedo, the J75_NOAH scheme for soil thermal conductivity, and the Y08 scheme for additional damping, respectively. Furthermore, the model performance is not obviously affected by surface albedo parameterization schemes, while the scheme of soil thermal conductivity is vital to simulations of latent heat flux and soil temperature and the scheme of additional damping is crucial for simulating net radiation flux, sensible heat flux, and surface soil temperature. A set of optimal parameterizations is proposed for the offline Noah LSM at the SACOL station when the MULT scheme for surface albedo, the J75_NOAH scheme for soil thermal conductivity, and the Y08 scheme for additional damping are combined simultaneously, especially in the case of sensible heat flux and surface soil temperature simulations.

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Weidong Guo, guowd@nju.edu.cn

1. Introduction

Arid and semiarid areas cover about 50% of the East Asia land surface and support 14.4% of the global population in 2000 (Safriel and Adeel 2005; Wang et al. 2012). China has the largest population in the world and is a country with the majority of arid and semiarid regions in East Asia and even the world (Piao et al. 2010; J. P. Huang et al. 2019). The arid and semiarid regions have increased rapidly in recent years in China, with an increase of 33% during 1994–2008 compared to 1948–62, which will enhance the risk of desertification in the near future (J. P. Huang et al. 2019).

The intensity of the regional climate response over the arid and semiarid areas has been amplified by the land–atmosphere interactions. Therefore, the land–atmosphere interactions over the arid and semiarid regions are currently identified as a hot spot or issue, and these interactions control the energy balance, water and heat transfer, and carbon cycle through turbulent flux at the atmospheric boundary layer, which play an important role in climate change (Wu et al. 2009; Berg et al. 2016; Huang et al. 2017a,b). The land surface parameter changes over the arid and semiarid regions can induce some positive feedbacks that reinforce and prolong the droughts. Climate variations can affect the land surface characteristic parameters in the arid and semiarid areas, such as the surface albedo, soil thermal conductivity, soil moisture and temperature, and surface aerodynamic and thermodynamic roughness. In turn, changes in land surface parameters alter atmospheric variations via fluxes and energy, water, carbon, and momentum, creating feedback that further affects the arid and semiarid climates (Dickinson et al. 1986; Pielke et al. 1998; Zhou et al. 2001; Maestre et al. 2013; Cheng et al. 2015; Berg et al. 2016; Xiao and Duan 2016; Duan et al. 2017; Huang et al. 2017a,b; Lu and Zuo 2018).

The land surface model (LSM) is a useful tool for studying the land–atmosphere interactions. The land surface parameters and their schemes are crucial for LSM simulation performance. Although the parameterization scheme is improved locally, the model simulation performance will be correspondingly effectively enhanced. The LSM with multiple parameterization options has great potential to facilitate physically based ensemble climate predictions, identification of the optimal combinations of schemes and explanation of model differences, and identification of critical processes controlling the coupling strength between the land surface and the atmosphere. Moreover, the multiple options of schemes for various processes enable us to explore multimodel ensemble simulations. However, model-projected future changes are still more uncertain over dry lands, partly because of the large internal climate variability in regional precipitation. For example, Ji et al. (2015) compared the observational data with Coupled Model Intercomparison Project phase 5 (CMIP5) simulations over the period 1948–2005 and found that the simulated mean precipitation on the global scale was too high over the arid and semiarid regions.

At present, most existing LSMs have large discrepancies in the simulation of land surface processes over the arid and semiarid areas. The characteristics of land surface parameters vary with different underlying surfaces, and some land surface parameterization (LSP) schemes are empirical and local. Therefore, the simulation performance of LSM can be greatly improved if the characteristics and parameterization methods of the land surface process in the arid and semiarid areas are analyzed or further optimized for the purpose of choosing the proper LSP schemes by carrying out many field observations, which is conducive to forecasting weather and climate changes. Some field experiments have recently been carried out over the arid and semiarid regions in China over the past 30 years containing the Heihe River Basin Field Experiment (HEIFE) (Hu 1994; Hu and Gao 1994; Wang and Mitsuta 1991, 1992), the Semi-Arid Climate and Environment Observatory of Lanzhou University (SACOL) (Huang et al. 2008; Guan et al. 2009; Wang et al. 2010), etc. These projects are the basic conditions for developing and verifying much more reasonable and accurate land surface parameters and LSP schemes for LSMs.

The Loess Plateau is the largest arid and semiarid zone in China, which is a highland region of northwest China, covering about 640 000 km2 (Liu 1985; An 2000; An et al. 2001; Guo et al. 2002; Fu et al. 2017; Ge et al. 2020). Some special land surface processes of the Loess Plateau not only influence regional atmospheric circulation and climate, but also affect the monsoon circulation in all of East Asia. It is an important source region of the dust aerosol featured by its unique underlying surface (Wang et al. 2010; Huang et al. 2013; J. P. Huang et al. 2019), and this region is also known as the transitional zone of climate and ecosystem change suffering the severe aridity trend in past decades (Zhang 1979, 1993; Fu and Wen 2002; Fu and Penning De Vries 2006; Huang et al. 2016a,b). With the characteristics of low nutrition content, low vegetation coverage, and low water conservation capacity in the soil, the arid and semiarid regions of the Loess Plateau in China are pretty sensitive to local, regional, and even global climate changes (Ma and Fu 2006, 2007; Fu and Ma 2008; Huang et al. 2016a,b; J. P. Huang et al. 2019). Because the ecological environment is extremely fragile, this region is also vulnerable to drought and degradation (Reed et al. 2012; Getirana et al. 2014, 2020). Fu et al. (2017) demonstrated that the vegetation over the Loess Plateau is fragile and highly dependent on water availability. As precipitation is pretty low and has an uneven distribution in both time and space, the soil water content is not easily conserved environmentally in the arid and semiarid regions of the Loess Plateau (Guan et al. 2009). It is very important and necessary to study the land surface processes over the Loess Plateau due to the uniqueness of the soil, geography, and climate regimes of this region. Cao et al. (2018, 2019) and Lv et al. (2019a,b) performed numerical experiments focused on the Loess Plateau to examine the impact of revegetation or afforestation on rainfall. However, there was little international long-term observation for the Loess Plateau, and little was learned of quantitatively and comprehensively assessing some more accurate or optimal LSP schemes in the LSM over the Loess Plateau.

In this study, we explore the impacts of three key land surface parameters and the optimal LSP combination scheme by a multivariable integrated evaluation (MVIE) method to quantitatively and comprehensively evaluate the performance of Noah LSM in simulating the land surface characteristics over the arid and semiarid regions of the Loess Plateau. This study is organized as follows. In section 2, we briefly introduce the observational experiment, data, Noah LSM, and assessment method. Comparisons of different LSP schemes for the land surface parameters are given in section 3. Noah LSMs configured with 48 combination LSP schemes are shown in section 4. In section 5, we summarize the main results and discussion.

2. Descriptions of observational site and model

a. Experiment station and data

The SACOL station (35°57′46″N, 104°08′13″E; elevation, 1965.8 m) is situated at the arid and semiarid areas of the Loess Plateau in China, located about 48 km away from the downtown area of Lanzhou city near the southern bank of the Yellow River in Gansu Province (Huang et al. 2008; Guan et al. 2009; Wang et al. 2010). The position of SACOL station is shown in Fig. 1. It is one of the reference sites of the international Coordinated Energy and Water Cycle Observations Project (CEOP). The surface is mostly covered by short grass with protophyta, which is usually less than 15 cm tall and covers less than 80% of the surface in summer and autumn (Wang et al. 2010). The land–atmosphere interactions are intense in this period. This study applies observations obtained by the SACOL station covering 4 months (June–September) from 2007 to 2010. The SACOL station is located on a nearly north–south mesa with a fetch length of about 120 m in the most common wind direction. The mesa has a limited width of approximately 200 m from the east to the west and is about 600 m in length from the north to the south. There is a large V-shaped valley to the west of the site and a small valley one to the east. The SACOL site was established in 2005 and started continuous observations in May 2006. The site with its surroundings suffered little or no human activities, so it can represent a primary regime of landform and vegetation in an arid and semiarid climate of the Loess Plateau (Huang et al. 2008; Xiao et al. 2012).

Fig. 1.
Fig. 1.

(a) Map of SACOL’s location and (b) location of the SACOL station.

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

Observational data from SACOL are widely used now. Many studies have demonstrated that SACOL observational data are of high quality and play a crucial role in investigating land–atmosphere interactions and even predicting climate change over the arid and semiarid regions of the Loess Plateau in China (Guan et al. 2009; Wang et al. 2010; Huang et al. 2013, 2016a, 2017a,b; J. P. Huang et al. 2019). Other specified information about the SACOL station and instruments can be found in Wang et al. 2010.

b. Noah LSM

The original Ohio State University (OSU) land surface model was developed in the 1980s (Mahrt and Pan 1984) and officially named the Noah LSM in 2000. We used version 3.4 of the Noah LSM in this study, which follows the force–restore principle, and the finite difference in space and Crank–Nicolson method in time are used to integrate the physical and chemical processes. In the initial version of the Noah LSM, the soil was divided into two layers, and the soil temperatures and soil water content were calculated by thermal diffusion and Richardson’s equations, respectively (Chen and Dudhia 2001). The latent heat flux was calculated via the Penman formula. Land surface static data are essential for Noah LSM, the vegetation types were taken from the U.S. Geological Survey (USGS) classification scheme, and the global soil dataset was obtained from the Food and Agriculture Organization database (http://www.fao.org). The calculation module incorporated the bioclimate, energy, canopy radiation transfer, momentum/sensible heat/latent heat flux exchange, soil and snow temperatures, hydrology, and dynamic vegetation. Within the Noah LSM, the effects of local factors such as the atmosphere, vegetation, and snow cover on the land surface processes are comprehensively considered. Noah LSM is able to simulate soil temperatures, soil water content, canopy water content, snow depth, energy flux, and upward longwave and shortwave radiation intensities (Sridhar et al. 2002). The model has been widely used in the weather/climate model by the National Centers for Environmental Prediction (NCEP) Global Forecast System (GFS), and it has been coupled with the Weather Research and Forecasting (WRF) Model and regional climate models (RegCMs). Through applications and validations, Noah LSM has been developed continuously. The main improvements include the soil layer being divided into four layers (0.1, 0.3, 0.6, and 1.0 m from top to bottom), modification to the canopy conductance formulation, evapotranspiration on bare land and vegetation phenology, and a new runoff model (Niu et al. 2011). Additionally, the parameterization scheme of thermodynamic roughness and coefficient of sensible heat exchange on the surface layer have been improved.

The simulation experiments cover 4 months (June–September) from 2007 to 2009 and 2010. The model integration starts on 23 May for every year and ends on 1 October of the same year. The first 9 days are taken as the spinup period, and the 4-month simulations are used for analysis. The input variables for the Noah LSM mainly include the atmospheric forcing, and the initial conditions for the soil temperature and moisture. The atmospheric forcing variables are hourly air temperature, specific humidity, u- and υ-component wind speed, air pressure, precipitation, atmospheric longwave radiation, solar shortwave radiation, solar shortwave radiation, multilayer soil temperature and soil moisture, soil heat flux, sensible heat, and latent heat flux. The soil layer is divided into four nodes that are 0.1, 0.3, 0.6, and 1.0 m from top to bottom. The soil temperature and moisture initialized for the model employ the observations at the corresponding layers, and the layers without observations are assumed to have the values through linear interpolation with the adjacent nodes.

c. Assessment metrics

To evaluate the model, three statistical metrics including the mean bias error (MBE), root-mean-square error (RMSE), and correlation coefficient (R) are applied. Additionally, an MVIE method is applied to evaluate the overall model performance in simulating multiple fields. The core concept of MVIE is to group various fields into a field and compare the constructed field with the observations. The MVIE method can be flexibly applied to full fields (including both the mean and the anomaly) or anomaly fields depending on the application. A detailed introduction and application of this method can be found in Lu et al. (2021), Lu and Zuo (2021), Xu et al. (2016, 2017), and F. Huang et al. (2019).

3. Comparisons of different LSP schemes

a. Surface albedo

1) Computational method

The surface albedo (α), which is the ratio of shortwave radiation reflected by Earth to the incoming shortwave radiation at the top of the atmosphere, may be considered a core parameter in regulating the climate system and its variability due to its substantial role in controlling the surface energy budget and temperature in the land surface process (Stephens 2015; Wielicki et al. 2005; Donohoe and Battisti 2011; Kwon et al. 2016; Kumar et al. 2019). The computational method used to determine the surface albedo is as follows:
α=RUdtRDdt,
where RU is the upward direct radiation to the land surface (W m−2), RD is the solar radiation downward to the land surface (W m−2), and t is time. To choose clear days, the datasets with a solar altitude angle greater than 10° [from 0600 to 1800 local standard time (LST) every day] are used in this study. The monthly mean values (June–September) of α from 2007 to 2009 were calculated via the radiation observations from the SACOL station are listed in Table 1. The monthly mean α from June to September mostly showed a decreasing trend from June to September, which indicated that surface albedo has a certain intermonthly variation due to the influences of surface and seasonal conditions.
Table 1

The monthly mean surface albedo (α), soil thermal conductivity (λs), and additional damping (kB−1) from the year of 2007 to 2009.

Table 1

2) Parameterization schemes

The Noah LSM mainly considers the seasonal variation of the surface albedo, which requires inputting the monthly mean values (snow-free) for simulation site and then interpolating them to Julian day of time step. It is noteworthy that the monthly mean surface albedo we inputted is assumed valid at 15th of every month, and Noah LSM physics will internally add snow cover effects to albedo. As seen from Table 1, the average values from June to September are 0.193, 0.189, 0.174, and 0.166, respectively (hereafter called the ALB_NOAH scheme). Although it makes α stable in the magnitude, it cannot accurately and appropriately describe the daily variations (figure not shown).

Dickinson et al. (1986) (hereafter called the D86 scheme) thought that the bare soil surface albedo is related to the surface soil water content (θ), which is used in the BATS (Biosphere–Atmosphere Transfer Scheme) and described as follows:
α=α0+0.110.4θ,
in which α0 is the albedo for saturated soil, and there α0 = 0.15.
Wang et al. (2005) (hereafter called the W05 scheme) analyzed the global 0.05° Moderate Resolution Imaging Spectroradiometer (MODIS) Bidirectional Reflectance Distribution Function (BRDF) and albedo data over 30 desert locations and indicated that bare soil albedo mostly varies with solar zenith angle (Schaaf et al. 2002; Wang et al. 2004):
α(θ)=αr{1+B1[g1(θ)g1(60°)]+B2[g2(θ)g2(60°)]},
where α is the surface albedo; θ is the solar zenith angle; αr is the albedo at a 60° solar zenith angle and depends on season and location. The parameters B1 and B2 are the average of the ratios of the volumetric and geometric parameters in the MODIS algorithm over αr for 30 pixels, respectively. In general, B1 = 0.346, B2 = 0.063. The functions g1 and g2 are from the MODIS algorithm:
g1(θ)=0.0075740.070987θ2+0.307588θ3,g2(θ)=1.2849090.166314θ2+0.04184θ3.

In this study, we develop two different parameterization schemes for the surface albedo by using the least squares fitting method based on SACOL observations. It is assumed that the α parameterization scheme is closely associated with the solar altitude angle (hθ) and the surface (5 cm) soil water content (θ5cm) for SACOL site during the period from June to September. Figure 2 displays the variations of α with hθ and θ5cm, respectively. It is obvious that α decreases as hθ increases, this relationship can be fitted by the function α=0.071e3.538hθ+0.172, and the square of the fitting correlation coefficient is 0.114 (Fig. 2a). Figure 2b shows that α can also decrease as θ5cm increases, the relationship can be fitted by the linear function α=0.247θ5cm+0.221, and the square of the fitting correlation coefficient is 0.272. The specific processes of the ADD [Eq. (A7)] and MULT [Eq. (A13)] parameterization schemes for α are described in the appendix. More importantly, the square of the fitting correlation coefficient is 0.330 in Fig. A1c, which is larger than that fitted by Fig. A1b, proving that the MULT scheme is slightly better than the ADD scheme.

Fig. 2.
Fig. 2.

(a) The change of the surface albedo (α) with the solar altitude angle (hθ) and (b) the variation of the surface α with the 5-cm soil water content (θ5cm) (note: the red solid lines are the function fitting lines of the α with the hθ and θ5cm, respectively).

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

Figure 3 displays the comparisons of the surface albedo between the observed values and the simulated values by five different parameterization schemes including D86, W05, ALB_NOAH, ADD, and MULT. Table 2 lists the statistical evaluations between the observed and simulated surface albedo calculated by different parameterization schemes. It is clear that the absolute values of MBE and RMSE of the MULT scheme are the smallest, and the R of the MULT scheme is the largest compared to the other four schemes. Therefore, the MULT scheme is the best parameterization scheme for improving the simulation performance.

Fig. 3.
Fig. 3.

Comparisons of the surface albedo (α) between the observed values and the simulated values by five parameterization schemes: (a) D86, (b) W05, (c) ALB_NOAH, and (d) ADD and MULT.

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

Table 2

The statistical evaluations [mean bias error (MBE), root-mean-square error (RMSE), correlation coefficient (R)] between the observed and the simulated surface albedo (α), and additional damping (kB−1) values by different parameterization schemes.

Table 2

b. Soil thermal conductivity

1) Computational method

Soil thermal conductivity (λs) can directly affect the vertical heat conduction and transport process below the surface, and the instantaneous change process of soil temperature with boundary conditions is determined by λs (Passerat de Silans et al. 1996). In general, it is defined as
λs=GTsoil/z,
in which z is the soil depth (m), G is the soil heat flux at the depth (W m−2), and Tsoil is the observational soil temperature (K). Therefore, we can obtain the 5-cm soil thermal conductivity (λs5) by applying Eqs. (6a) and (6b):
G5=λs5(Tsoilz)5,
(Tsoilz)5=T10T00.1,
where G5 is the 5-cm soil heat flux (W m−2), λs5 is the 5-cm soil thermal conductivity (W m−1 K−1), T10 is the 10-cm soil temperature (K), and T0 is the surface soil temperature (K) retrieved by the observed longwave radiation. Note that it is assumed that the surface soil thermal conductivity is approximately equal to the 5 cm soil thermal conductivity; thus, λs5 is replaced with λs in the next text of this study. Table 1 lists the monthly mean values (June–September) of the λs from 2007 to 2009.

2) Parameterization schemes

According to the Chinese Soil Database (http://www.soil.csdb.cn/), the soil texture of the SACOL station is clay loam, the content of organic matter in the topsoil (0–17 cm) is 12.8 g kg−1, and the percentages of particles composition within 2–0.02 and 0.02–0.002, and less than 0.002 mm are 55.09%, 27.70%, and 17.20%, respectively.

The λs is affected by soil properties such as soil composition, soil density, soil porosity, and soil water content. The parameterization scheme of λs in the Noah LSM derives from Johansen (1975) (hereafter called the J75_NOAH scheme), which can be simplified as Eqs. (7a)–(7e) under the condition of clay loam without permafrost and snow cover:
λs={Ke(λsatλdry)+λdrySr>1×105λdrySr1×105,
λsat=λss1nλswn,
λss=λsqqλso1q,
λdry=0.135γd+64.727000.947γd,
Ke={0.7lgSr+1.0Sr>0.05lgSr+1.0Sr>0.1,
where λsat is the soil saturated thermal conductivity (W m−1 k−1); λdry is the dry thermal conductivity of the soil (W m−1 k−1) with Kestern number (Ke), which takes an empirical function of dry density [γd = 2700(1 − n), n is soil porosity]; Sr is the saturation ratio index, which is defined as the ratio of the soil water content θ to soil porosity n ( Sr=θ/n); λss is the solid heat conductivity of the soil (W m−1 k−1) obtained by the heat conductivity of the quartz (λsq = 7.7 W m−1 k−1) and the heat conductivity of other minerals (λso = 2.0 W m−1 k−1); and λsw is the soil water conductivity (λsw = 0.57 W m−1 k−1).
Farouki (1981) (hereafter called the F81 scheme) modified the calculation method of the λss based on the J75_NOAH where “%sand” is the content of sand and “%clay” is the content of clay:
λs={Ke(λsatλdry)+λdrySr>1×105λdrySr1×105,
λsat=λss1nλswn,
λss=8.80(%sand)+2.92(%clay)(%sand)+(%clay),
λdry=0.135γd+64.727000.947γd,
Ke={0.7logSr+1.0Sr>0.05logSr+1.0Sr>0.1.
Luo et al. (2009) (hereafter called the LUO09 scheme) proposed that the parameterization scheme can be expressed by the Eqs. (9a)–(9e); Ke is the Kestern number. Table 3 shows the λdry, λsat, difference between λdry and λsat, and the mean values of Ke calculated by different parameterization schemes as follows:
λs={Ke(λsatλdry)+λdrySr>1×105λdrySr1×105,
λsat=λss1nλswn,
λss=λsqqλso1q,
λdry=χ10ηn,
Ke=κSr1+(1κ)Sr,
where χ, η, and κ are different parameters associated with the soil types (Table 4). Figure 4 presents the monthly mean values of the λs between the J75_NOAH, F81, and LUO09 parameterization scheme simulations and the observations. It can be seen that the λs simulated by the LUO09 scheme is the largest, the second is the F81 scheme, and the smallest is the J75_NOAH scheme. The retrieved values of the J75_NOAH scheme are in the closest agreement with the observations.
Fig. 4.
Fig. 4.

Comparisons for the monthly mean values of the λs between the J75_NOAH, F81, and LUO09 parameterization scheme simulations and the observations.

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

Table 3

The dry heat conductivity (λdry), the saturated heat conductivity (λsat), the differences between λdry and λsat, and the mean values of Ke number calculated by different parameterization schemes.

Table 3
Table 4

The different parameter values of χ, η, and κ for the different types of soil in the LUO09 parameterization scheme.

Table 4

c. Additional damping

1) Computational method

The additional damping (kB−1) is defined as the logarithm of the ratio of the surface aerodynamic roughness (z0m) to the surface thermodynamic roughness (z0h), which is one of the important ways to describe the sensible heat flux exchange between the land and the atmosphere. According to the Monin–Obukhov similarity theory and the wind profiles, it can be described as follows:
kB1=lnz0mz0h,
H=ρCp(θsθa)rah,
H=ρCpu*θ*,
rah=Pr0ku*[lnzz0m+lnz0mz0hψh(ξ)],
where H is the sensible heat flux (W m−2); ρ is the air density at the reference height (kg m−3); Cp is the specific heat capacity of air (J m−3 K−1); u* is the friction velocity (m s−1); θ* is the dimensionless temperature (K); θs is the near surface potential temperature (K); θa is the atmospheric potential temperature (K); k is the von Kármán constant (k = 1.4); rah is the aerodynamic resistance(s m−1); z is the height in the surface layer (m); Pr0 is a constant (under stable conditions; Pr0 = 1.0; under unstable conditions, Pr0 = 0.95); ξ is the atmospheric stability; and ψh(ξ) is the sensible heat flux gradient relation function of ξ.
In stable conditions:
ψh(ξ)=βhξ.
In unstable conditions:
ψh(ξ)=2ln(1+y22),
y=(1γhξ)1/4,
in which βh = 5.0 and γh = 16.0.
By Eqs. (23)–(29), kB−1 can be expressed as
kB1=1Pr0ρcp(θsθa)Hku*lnzz0m+ψh(ξ).

Table 1 shows the monthly mean values (June–September) of kB−1 from 2007 to 2009. It is obvious that the average value of kB−1 in every year has little difference, and the monthly average value of kB−1 has a decreasing trend from June to September. The average kB−1 of monthly mean values from June to September are 6.65, 6.24, 5.44, and 4.20, respectively, which are far greater than the default mean values of kB−1 calculated by Noah LSM (1.56, 1.43, 1.36, and 1.29, respectively).

2) Parameterization schemes

Sheppard (1958) (hereafter called the S58 scheme) first parameterized kB−1 with the roughness Reynolds number ( Re*) as follows:
kB1=ln(PrRe*),
where Pr is the Prandtl constant (Pr = 0.71).
Kustas et al. (1989) (hereafter called the K89 scheme) found that the simple relationship of zonal horizontal wind and temperature difference between the near surface and the atmosphere with conventional observation data can well describe the change of kB−1 on the condition that the sparse vegetation underlying surface in arid regions, as shown in Eq. (19). At the same time, the introduction of the temperature difference between the near surface and the atmosphere also solves the problem of daily variation of kB−1 to a certain degree:
kB1=0.17u(θsθa),
where u is the zonal horizontal wind speed (m s−1).
The parameterization scheme of kB−1 in the Noah LSM proposed by Zilitinkevich (1995) (hereafter called the Z95_NOAH scheme) is expressed as Eq. (20), which is in the form of the power function about Re*:
kB1=0.1Re*0.5.
Yang et al. (2008) (hereafter called the Y08 scheme) also developed a parameterization scheme of kB−1 expressed as Eq. (21):
kB1=70νuexp(βu0.5|θ0.25|)ν=1.328×105(PsP)(θθs)1.754,
where ν is the kinematic viscosity of air (m2 s−1); u* is the friction velocity (m s−1); β is a constant (β = 7.2); Ps is the standard atmospheric pressure (Pa); P is the air pressure (Pa); and θ is the atmospheric potential temperature (K).

Figure 5 shows the comparisons of the kB−1 between the observed values and the simulated values by the S58, K89, Z95_NOAH, and Y08 parameterization schemes. Figure 6 displays the average daily variation of kB−1 between the different parameterization schemes and the observations. It can be seen that the values simulated by the S58, Z95_NOAH, and Y08 schemes have a small range of variation and are relatively stable. The values of the Z95_NOAH scheme are approximately 0, which are far less than the measured value of kB−1. The values of kB−1 estimated from the K89 scheme show great uncertainty mainly due to the large variation of wind speed and temperature difference between the land and the atmosphere. The S58 and Z95_NOAH parameterization schemes have no obvious daily variation characteristics. The K89 and Y08 schemes show certain parabolic daily variations, especially the K89 scheme. Table 2 lists the statistical evaluations between the observed and simulated kB−1 calculated by different parameterization schemes. The simulations of the Y08 scheme perform better than other schemes, with the smallest MBE and RMSE but the largest R.

Fig. 5.
Fig. 5.

Comparisons of the additional damping (kB−1) between the observed values and the simulated values by different parameterization schemes: (a) S58, (b) K89, (c) Z95_NOAH, and (d) Y08.

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

Fig. 6.
Fig. 6.

The average daily variation of the additional damping (kB−1) between the different parameterization schemes and the observations.

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

4. Influence on simulations with different LSP scheme combinations

To investigate the applicability of the LSP schemes over the arid and semiarid regions in China, different parameterization combination schemes are coupled into the Noah LSM to simulate the land surface process of the SACOL station from June to September in 2010. The 48 experiments composed of different combinations of LSP schemes are shown in Table 5. It is noteworthy that the α parameterization schemes select ALB_NOAH, D86, W05, and MULT, respectively. The calculation values of the ADD and MULT schemes are similar for the α schemes, and the MULT scheme is slightly better than the ADD scheme; thus, only the MULT scheme is considered in the combination experiments. The λs parameterization schemes are J75_NOAH, F81, and LUO09. The kB−1 parameterization schemes chosen are Z95_NOAH, S58, K89, and Y08.

Table 5

The different parameterization scheme combinations of the 48 experiments.

Table 5

a. Net radiation flux

Figure 7 shows the statistical diagram of the MBE, RMSE, and R between the net radiation flux (Rn) simulated by 48 experiments and the observations. The simulated values of Rn for each parameterization combination scheme are in closer agreement with the observations. The maximum MBE absolute value of Rn between the simulation and the observation is about 24 W m−2, and the minimum MBE absolute value is about 2 W m−2. The maximum RMSE of Rn between the simulation and the observation is less than 32 W m−2, and the minimum is approximately 9 W m−2. The R of Rn between the simulation and the observation is more than 0.99 (figure not shown). Overall, Rn is overestimated when the kB−1 parameterization scheme is taken as Z95_NOAH, while Rn is underestimated when the S58, K89, or Y08 scheme is selected.

Fig. 7.
Fig. 7.

The statistical evaluations between the net radiation flux (Rn; W m−2) simulated by the 48 experiments and observations: (a) mean bias error (MBE) and (b) root-mean-square error (RMSE).

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

b. Sensible and latent heat fluxes

Figure 8 displays the statistical diagram of the MBE, RMSE, and R between the H simulated by 48 experiments and the observations. All combination schemes overestimate H, which is closely associated with the selection of the kB−1 parameterization. The maximum MBE absolute value of H between the simulation and the observation is greater than 38 W m−2, and the minimum MBE absolute value is less than 9 W m−2. The maximum RMSE of H between the simulation and the observation is greater than 66 W m−2, and the minimum is less than 35 W m−2. The maximum R of H between the simulation and the observation is greater than 0.94, and the minimum is less than 0.90. For the four different parameterization schemes of α, the MULT and D86 schemes have poor effects of H mostly because of the inaccurate soil water content. The MULT scheme is more affected by the soil water content with larger biases [Eq. (17)]. The W05 scheme only considers the influence of the solar altitude angle with the best performance [Eq. (3)]. The RMSEs of the other three parameterization schemes except for the Z95_NOAH scheme are mostly less than 40 W m−2, and the R values of the Z95_NOAH and Y08 schemes are significantly greater than those of the S58 and K89 schemes for the kB−1 results, while the Y08 scheme performs best. The LUO09 scheme is better for simulating λs, followed by the F81 scheme, and then the J75_NOAH scheme. In short, the schemes of kB−1 and λs have significant effects on H. The Y08 scheme for kB−1 and the LUO09 scheme for the λs simulation perform relatively better. Although the simulated H by the J75_NOAH scheme for λs is the closest to the measured value, kB−1 provided by each parameterization scheme is too small resulting in H being overestimated. However, the λs is relatively large, and more energy will be input underground so that H will be reduced correspondingly. The biases of the two factors superpose making the simulated H closer to the measured value. By and large, the selection of the kB−1 parameterization scheme has a great influence on the simulation of H, and the Y08 parameterization scheme performs better.

Fig. 8.
Fig. 8.

As in Fig. 7, but for the sensible heat flux (H; W m−2).

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

Figure 9 displays the statistical diagram of the MBE, RMSE, and R between the latent heat flux (LE) simulated by 48 experiments and the observations. It is obvious that LE simulations by the 48 experiments are all lower than the measured values. The maximum MBE absolute value of LE between the simulation and the observation is about 15 W m−2, the minimum MBE absolute value is about 8.5 W m−2. The maximum RMSE of LE between the simulation and the observation is less than 37 W m−2, and the minimum is about 32 W m−2. The maximum R of LE between the simulation and the observation is greater than 0.86, and the minimum is greater than 0.82. The changes of latent heat flux (LE) simulated by different parameterization schemes are small. The LE simulation results are better when the parameterization combination scheme is the Z95_NOAH scheme for kB−1 and the J75_NOAH scheme for λs. The α parameterization schemes have no significant difference in simulating LE.

Fig. 9.
Fig. 9.

As in Fig. 7, but for the latent heat flux (LE; W m−2).

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

c. Soil temperature

The statistical diagrams of the MBE, RMSE, and R between the different layer soil temperatures simulated by 48 experiments and the observations are given in Figs. 10 and 11. The surface soil temperature (Tg), 5-cm soil temperature (T5cm), 10-cm soil temperature (T10cm), and 20-cm soil temperature (T20cm) simulations are all the best when the parameterization scheme of kB−1 is the Z95_NOAH, followed by the Y08, S58, and K89. It is found that the different layer soil temperatures are underestimated when the kB−1 parameterization scheme is the default scheme of Z95_NOAH in Noah LSM. The Tg, T5cm, T10cm, and T20cm simulations are all overestimated on the condition that the kB−1 parameterization scheme is Y08, S58, or K89. This is mainly because kB−1 is underestimated, resulting in the simulation result of H being overestimated. The maximum MBE absolute value of Tg between the simulation and the observation is about 3 K, the minimum MBE absolute value is less than 0.5 K. The maximum RMSE of Tg between the simulation and the observation is about 3.8 K, the minimum is less than 1.5 K. The maximum R of Tg between the simulation and the observation is greater than 0.99, and the minimum is greater than 0.98 (figure not shown). The maximum MBE absolute value of T10cm between the simulation and the observation is about 2.6 K, the minimum MBE absolute value is also less than 0.5 K. The maximum RMSE of T10cm between the simulation and the observation is about 3.0 K, the minimum is less than 1.0 K. The maximum R of T10cm between the simulation and the observation is greater than 0.98, and the minimum is greater than 0.92 (figure not shown). By and large, the T5cm, T10cm, and T20cm simulations are similar with Tg (Fig. 11).

Fig. 10.
Fig. 10.

As in Fig. 7, but for the surface soil temperature (Tg; K).

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

Fig. 11.
Fig. 11.

As in Fig. 7, but for the soil temperatures at different layers (K). (a1),(b1) The 5-cm soil temperature (T5cm), (a2),(b2) the 10-cm soil temperature (T10cm), and (a3),(b3) the 20-cm soil temperature (T20cm).

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

Figure 12 presents the Taylor diagram between the different layer soil temperatures simulated by the 48 experiments and the observations simulated by the 48 experiments and the observations. It can be seen that the kB−1 schemes have a significant impact on the Rn simulation results, especially the Y08 scheme. There are no obvious differences for the different parameterization schemes of α in simulating the different layer soil temperatures, while the W05 scheme performs relatively better. Moreover, it can be concluded that when the kB−1 parameterization scheme is S58 and the λs parameterization scheme is LUO09, the simulation result of Tg is in closer agreement with the observation. Above all, the selection of the kB−1 parameterization scheme has the greatest impact on the Tg simulation. In particular, when the Y08 parameterization scheme of kB−1 combines with the J75_NOAH of λs, the simulated Tg is more ideal. Actually, the simulation effect is also better when the S58 scheme of kB−1 combines with the LUO09 scheme of λs, but which results from the superposition value of two parameter errors, not due to the approximation values of real parameters (Fig. 12). As shown in Fig. 12, each layer soil temperature simulation is the best when the λs parameterization scheme is the J75_NOAH, followed by the F81 scheme and the LUO09 scheme. This is mainly because the ground heat flux (G0) is the upper boundary condition depending on the soil thermodynamic process, but the kB−1 parameterization scheme has a great impact on G0 as mentioned above. At the same time, the simulation performance of the Z95_NOAH scheme is the best, followed by the Y08 scheme; thus, the T5cm simulation is apparently affected by the kB−1 parameterization scheme. The soil vertical heat exchange process is more obvious at the 20-cm soil depth, while the transfer process inside the soil is more determined by λs so that T20cm is easily affected by the λs parameterization scheme. In other words, the deeper the depth is, the smaller the difference brought by the parameterization scheme of kB−1, but the Z95_NOAH scheme performs much better. Accordingly, the selection of the λs parameterization scheme is increasingly significant to simulations of deeper soil temperatures, and J75_NOAH is the optimal scheme.

Fig. 12.
Fig. 12.

The Taylor diagram between the soil temperatures at different layers simulated by the 48 experiments and the observations. (a1)–(a3) The 5-cm soil temperature (T5cm), (b1)–(b3) the 10-cm soil temperature (T10cm), and (c1)–(c3) the 20-cm soil temperature (T20cm). Note: the various colors represent the simulation results with different parameterization schemes of the additional damping (kB−1), the surface albedo (α), and the soil heat conductivity (λs) in (a1)–(a3), (b1)–(b3), and (c1)–(c3), respectively.

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

d. Multivariable integrated evaluation

Figure 13 presents the MVIE diagram to quantitatively and comprehensively summarize and rank the performance in simulating mean Rn, H, LE, and soil temperatures at different layers (Tg, T5cm, T10cm, T20cm) from June to September in the year of 2010 of the 48 experiments. The simulation skill scores illustrate that the 46th experiment with the combined LSP scheme out of all 48 simulations best matches the observations, which is composed of the MULT scheme for α, the J75_NOAH scheme for λs, and the Y08 scheme for kB−1. The value of the skill score approaches 1 in a perfect simulation. The upper-left triangle and the lower-right triangle take into account both the normalized root-mean-square length (RMSL) and similarity coefficient (SC) respectively. The upper-left triangle emphasizes on the simulation of the amplitude of multivariable field more, while the lower right triangle is more sensitive to the pattern similarity. The more left-handed result in Fig. 13 represents the higher value of simulation skill score in the 48 experiments. As a result, the performance of the 46th experiment ranks at the top out of all simulations. This is the optimal parameterization scheme combination for simulating land surface characterization over the arid and semiarid areas of the Loess Plateau in China.

Fig. 13.
Fig. 13.

The comprehensive performance ranks of 48 experiments in Table 5.

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

5. Discussion and conclusions

As a region of unique landform and climate and thus a special zone of the arid and semiarid climate (such regions constitute about 30% of the global land), the northwestern Loess Plateau of China acts as a not negligible portion of land–atmosphere interactions. To improve the simulation performance of LSM over the arid and semiarid areas of the Loess Plateau in China, we compare the simulations of different LSP schemes in the Noah LSM based on the observation data at the SACOL site during June to September from 2007 to 2009. Furthermore, we analyze the 48 combination scheme simulations through the way of bidirectional coupling between the Noah LSM and the LSP schemes, and a set of optimal combined LSP schemes for typical land surfaces over the arid and semiarid of the Loess Plateau in northwest China is proposed. Main conclusions are as follows.

  1. Disturbance tests indicate that α, λs, and kB−1 are all sensitive to the LSM performance. The disturbance of α has a prominent effect on the simulation results of Rn, T5cm, T10cm, and T20cm. An increase in α is associated with a decrease in the simulation of Rn, T5cm, T10cm, and T20cm. The disturbance of λs has a remarkable impact on the T5cm, T10cm, and T20cm simulations, and a decrease in λs is consistent with the decrease in the simulation of T5cm, T10cm, and T20cm. The disturbance of kB−1 has a significant influence on the simulation performance of H and Tg, and an increase in kB−1 results in a decrease in H and an increase in Tg.

  2. The MULT scheme of α is the best by comparing the parameter values derived from different parameterization schemes and the parameter values derived from observation data. The J75_NOAH scheme performs best when comparing the monthly average value of λs calculated by observation data and three parameterizations. The kB−1 value retrieved by the Y08 scheme is the closest to the value retrieved from measurement.

  3. The 48 experiments show that the model simulation performance is not obviously affected by different α parameterization schemes, while the parameterization scheme of λs is important to the simulation results of LH, Rn, T5cm, T10cm, and T20cm. The parameterization scheme of kB−1 is crucial for simulating Rn, H, and Tg. A set of optimal parameterization schemes is obtained for the offline Noah LSM based on the SACOL observations over the arid and semiarid regions when the MULT scheme for α, the J75_NOAH scheme for λs, and the Y08 scheme for kB−1 are combined synchronously, especially for H and Tg.

This study analyzes the land surface characteristic parameters by combining statistical analysis with numerical simulation and selects a set of optimal parameterization schemes for the SACOL station. Due to limited field observations, this study is only validated by a single-point offline simulation at the SACOL station. However, we recommend this optimal combination parameterization scheme to validate the simulation performance on the Loess Plateau or at least the western part of the Loess Plateau because the SACOL station is typical and significant for representing the land–atmosphere interactions over the arid and semiarid regions of the Loess Plateau in northwest China. In addition, more field observations should also be carried out and collected to further develop and validate the land surface process parameterization schemes, and some optimization methods such as artificial neural networks and satellite remote sensing information should be employed for selecting appropriate parameters in LSMs.

Acknowledgments.

This study is jointly supported by the National Science Foundation of China (Grants 42021001, 42130602, 42005096). This work is also supported by the Jiangsu Collaborative Innovation Center for Climate Change. We are grateful for the excellent support provided by the people at SACOL, Lanzhou University.

APPENDIX

Development of the ADD and MULT Schemes for the Surface Albedo

If it is assumed that hθ and θ5cm are simply adding linear relationships on α respectively, the surface albedo can be parameterized by
α=f1(θ)+g1(hθ),
there is a linear relationship between f1(θ) and θ5cm, that is,
f1(θ)=a0θ+a1.
It can be regarded as dry soil when θ5cm is small (θ5cm < 0.1), and α is only affected by hθ. There is
α=a1+g1(hθ)=g2(hθ).
Thus, Eq. (A1) can be changed as
α=f2(θ)+g2(hθ).
The least squares method is used to fit the relation of Eq. (A3). We can obtain that
g2(hθ)=0.076e2.062hθ+0.172.
As shown in Fig. A1a, the square of the fitting correlation coefficient is 0.294, which is larger than the square of the fitting function in Fig. 2a, indicating that the interference effect of soil moisture content on surface albedo is weakened when θ5cm is small to some extent. Then, the least squares method is used to fit the relation of f2(θ) to obtain
f2(θ)=0.253θ+0.030.
Similarly, Fig. A1b displays that the square of the fitting correlation coefficient is 0.314, which is higher than the square of the fitting function in Fig. 2b, demonstrating that the fitting accuracy between the soil water content and surface albedo is improved after removing the influence of hθ. Consequently, the addition scheme [hereafter called the ADD scheme; Eq. (A7)] can be expressed as follows:
α=0.2020.253θ+0.076e2.062hθ.
Fig. A1.
Fig. A1.

(a) The relationship between the surface albedo (α) and the solar altitude angle (hθ) for the dry soil [the red line is the fitting curve of Eq. (A5)]. (b) The relationship between α and the θ5cm after removing the influence of hθ for the addition rule [the red line is the fitting line of Eq. (A6)]. (c) The relationship between α and θ5cm after removing the influence of hθ for the multiplication rule [the red line is the fitting line of Eq. (A12)].

Citation: Journal of Hydrometeorology 23, 6; 10.1175/JHM-D-21-0143.1

If it is assumed that hθ and θ5cm independently affect α, the surface albedo parameterization scheme can be described as
α=f3(θ)g3(hθ).
It can be regarded as dry soil when θ5cm is small (θ5cm < 0.1), and α is only affected by hθ. There is
α=cg3(hθ)=g4(hθ), and
f4(θ)=αg4(hθ).
According to the least squares method (Figs. A1a,c), further it can obtain that
g4(hθ)=0.076e2.062hθ+0.172,
f4(θ)=1.319θ+1.159.
Hence, the multiplication scheme [hereafter called the MULT scheme; Eq. (A13)] can be expressed as follows:
α=f4(θ)g4(hθ)=(1.319θ+1.159)(0.076e2.062hθ+0.172).

REFERENCES

  • An, Z. S., 2000: The history and variability of the East Asian paleomonsoon climate. Quat. Sci. Rev., 19, 171187, https://doi.org/10.1016/S0277-3791(99)00060-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • An, Z. S., J. E. Kutzbach, W. L. Prell, and S. C. Porter, 2001: Evolution of Asian monsoons and phased uplift of the Himalaya-Tibetan plateau since late Miocene times. Nature, 411, 6266, https://doi.org/10.1038/35075035.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Berg, A., and Coauthors, 2016: Land-atmosphere feedbacks amplify aridity increase over land under global warming. Nat. Climate Change, 6, 869874, https://doi.org/10.1038/nclimate3029.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cao, Q., D. Y. Yu, M. Georgescu, and J. G. Wu, 2018: Substantial impacts of landscape changes on summer climate with major regional difference: The case of China. Sci. Total Environ. 625, 416427, https://doi.org/10.1016/j.scitotenv.2017.12.290.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cao, Q., J. G. Wu, D. Y. Yu, and W. Wang, 2019: The biophysical effects of the vegetation restoration program on regional climate metrics in the Loess Plateau, China. Agric. For. Meteor., 268, 169180, https://doi.org/10.1016/j.agrformet.2019.01.022.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, F., and J. Dudhia, 2001: Coupling an advanced land surface-hydrology model with the Penn State–NCAR MM5 modeling system. Part II: Preliminary model validation. Mon. Wea. Rev., 129, 587604, https://doi.org/10.1175/1520-0493(2001)129<0587:CAALSH>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cheng, S. J., X. D. Guan, J. P. Huang, F. Ji, and R. X. Guo, 2015: Long-term trend and variability of soil moisture over East Asia. J. Geophys. Res. Atmos., 120, 86588670, https://doi.org/10.1002/2015JD023206.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dickinson, R. E., A. Henderson-Sellers, P. J. Kennedy, and M. F. Wilson, 1986: Biosphere–Atmosphere Transfer Scheme (BATS) for the Community Climate Model. NCAR Tech. Note NCAR/TN-275+STR, 72 pp., https://doi.org/10.5065/D6668B58.

    • Crossref
    • Export Citation
  • Donohoe, A., and D. S. Battisti, 2011: Atmospheric and surface contributions to planetary albedo. J. Climate, 24, 44024418, https://doi.org/10.1175/2011JCLI3946.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Duan, A. M., R. Z. Sun, and J. H. He, 2017: Impact of surface sensible heating over the Tibetan Plateau on the western Pacific subtropical high: A land-air-sea interaction perspective. Adv. Atmos. Sci., 34, 157168, https://doi.org/10.1007/s00376-016-6008-z.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Farouki, O. T., 1981: Thermal properties of soils. CRREL Monograph 81-1, U.S. Army Corps of Engineers, 136 pp., https://apps.dtic.mil/sti/pdfs/ADA111734.pdf.

    • Crossref
    • Export Citation
  • Fu, B., S. Wang, Y. Liu, J. B. Liu, W. Liang, and C. Y. Miao, 2017: Hydrogeomorphic ecosystem responses to natural and anthropogenic changes in the Loess Plateau of China. Annu. Rev. Earth Planet. Sci., 45, 223243, https://doi.org/10.1146/annurev-earth-063016-020552.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fu, C. B., and F. Penning De Vries, Eds., 2006: Initial science plan of the Monsoon Asia Integrated Regional Study. MAIRS Working Paper Series 1, IAP-CAS, 86 pp.

    • Crossref
    • Export Citation
  • Fu, C. B., and G. Wen, 2002: Some key issues of aridity trend in northern China (in Chinese). Climatic Environ. Res., 7, 2229.

  • Fu, C. B., and Z. G. Ma, 2008: Global change and regional aridfication (in Chinese with English abstract). Chin. J. Atmos. Sci., 32, 752760.

  • Ge, J., A. J. Pitman, W. D. Guo, B. L. Zan, and C. B. Fu, 2020: Impact of revegetation of the Loess Plateau of China on the regional growing season water balance. Hydrol. Earth Syst. Sci., 24, 515533, https://doi.org/10.5194/hess-24-515-2020.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Getirana, A. C. V., and Coauthors, 2014: Water balance in the Amazon basin from a land surface model ensemble. J. Hydrometeor., 15, 25862614, https://doi.org/10.1175/JHM-D-14-0068.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Getirana, A. C. V., and Coauthors, 2020: GRACE improves seasonal groundwater forecast initialization over the United States. J. Hydrometeor., 21, 5971, https://doi.org/10.1175/JHM-D-19-0096.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Guan, X. D., J. Huang, N. Guo, J. R. Bi, and G. Y. Wang, 2009: Variability of soil moisture and its relationship with surface albedo and soil thermal parameters over the Loess Plateau. Adv. Atmos. Sci., 26, 692700, https://doi.org/10.1007/s00376-009-8198-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Guo, Z. T., and Coauthors, 2002: Onset of Asian desertification by 22 Myr ago inferred from loess deposits in China. Nature, 416, 159163, https://doi.org/10.1038/416159a.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hu, Y., 1994: Research advance about the energy budget and transportation of water vapor in the HEIFE area (in Chinese). Adv. Earth Sci., 9, 3034, http://www.adearth.ac.cn/CN/Y1994/V9/I4/30.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hu, Y., and Y. Gao, 1994: Some new understandings of processes at the land surface in arid area from the HEIFE (in Chinese). Acta Meteor. Sin., 52, 285296.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, F., Z. Xu, and W. Guo, 2019: Evaluating vector winds in the Asian-Australian monsoon region simulated by 37 CMIP5 models. Climate Dyn., 53, 491507, https://doi.org/10.1007/s00382-018-4599-z.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, J. P., and Coauthors, 2008: An overview of the semi-arid climate and environment research observatory over the Loess Plateau. Adv. Atmos. Sci., 25, 906921, https://doi.org/10.1007/s00376-008-0906-7.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, J. P., M. X. Ji, Y. Z. Liu, L. Zhang, and D. Y. Gong, 2013: An overview of arid and semi-arid climate change (in Chinese). Adv. Climate Change Res., 9, 914.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, J. P., M. X. Ji, Y. K. Xie, S. S. Wang, Y. L. He, and J. T. Ran, 2016a: Global semi-arid climate change over last 60 years. Climate Dyn., 46, 11311150, https://doi.org/10.1007/s00382-015-2636-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, J. P., H. P. Yu, X. D. Guan, G. Y. Wang, and R. X. Guo, 2016b: Accelerated dryland expansion under climate change. Nat. Climate Change, 6, 166171, https://doi.org/10.1038/nclimate2837.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, J. P., and Coauthors, 2017a: Dryland climate change: Recent progress and challenges. Rev. Geophys., 55, 719778, https://doi.org/10.1002/2016RG000550.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, J. P., H. P. Yu, A. G. Dai, Y. Wei, and L. T. Kang, 2017b: Drylands face potential threat under 2°C global warming target. Nat. Climate Change, 7, 417422, https://doi.org/10.1038/nclimate3275.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, J. P., J. R. Ma, X. D. Guan, Y. Li, and Y. L. He, 2019: Progress in semi-arid climate change studies in China. Adv. Atmos. Sci., 36, 922937, https://doi.org/10.1007/s00376-018-8200-9.

    • Search Google Scholar
    • Export Citation
  • Ji, M., J. P. Huang, Y. Xie, and J. Liu, 2015: Comparison of dryland climate change in observations and CMIP5 simulations. Adv. Atmos. Sci., 32, 15651574, https://doi.org/10.1007/s00376-015-4267-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Johansen, O., 1975: Thermal conductivity of soils. Ph.D. thesis, Norwegian University of Science and Technology, 291 pp.

    • Crossref
    • Export Citation
  • Kumar, S. V., D. M. Mocko, S. Wang, C. D. Peters-Lidard, and J. Borak, 2019: Assimilation of remotely sensed leaf area index into the Noah-MP land surface model: Impacts on water and carbon fluxes and states over the Continental United States. J. Hydrometeor., 20, 13591377, https://doi.org/10.1175/JHM-D-18-0237.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kustas, W. P., B. J. Choudhury, M. S. Moran, R. J. Reginato, and H. L. Weaver, 1989: Determination of sensible heat flux over sparse canopy using thermal infrared data. Agric. For. Meteor., 44, 197216, https://doi.org/10.1016/0168-1923(89)90017-8.

    • Search Google Scholar
    • Export Citation
  • Kwon, Y., Z. L. Yang, L. Zhao, T. J. Hoar, A. M. Toure, and M. Rodell, 2016: Estimating snow water storage in North America using CLM4, DART, and snow radiance data assimilation. J. Hydrometeor., 17, 28532874, https://doi.org/10.1175/JHM-D-16-0028.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liu, T. S., 1985: Regional record of aeolian processes: The distribution of loess. Loess and the Environment, China Ocean Press, 14–15.

    • Crossref
    • Export Citation
  • Lu, S., and H. C. Zuo, 2018: Improvement and validation of the Common Land Model on cropland covered by plastic film in the arid region of China. J. Appl. Meteor. Climatol., 57, 20712089, https://doi.org/10.1175/JAMC-D-17-0185.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lu, S., and H. C. Zuo, 2021: Sensitivity of South Asian summer monsoon simulation to land surface schemes in Weather Research and Forecasting model. Int. J. Climatol., 41, 68056824, https://doi.org/10.1002/joc.7278.

    • Search Google Scholar
    • Export Citation
  • Lu, S., W. D. Guo, Y. K. Xue, F. Huang, and J. Ge, 2021: Simulation of summer climate over Central Asia shows high sensitivity to different land surface schemes in WRF. Climate Dyn., 57, 22492268, https://doi.org/10.1007/s00382-021-05876-9.

    • Search Google Scholar
    • Export Citation
  • Luo, S. Q., and Coauthors, 2009: Soil thermal conductivity parameterization establishment and application in numerical model of central Tibetan Plateau (in Chinese). Chin. J. Geophys., 52, 919928, https://doi.org/10.3969/j.issn.0001-5733.2009.04.008.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lv, M. X., Z. G. Ma, M. X. Li, and Z. Y. Zheng, 2019a: Quantitative analysis of terrestrial water storage changes under the Grain for Green Program in the Yellow River basin. J. Geophys. Res. Atmos., 124, 13361351, https://doi.org/10.1029/2018JD029113.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lv, M. X., Z. G. Ma, and S. M. Peng, 2019b: Responses of terrestrial water cycle components to afforestation within and around the Yellow River basin. Atmos. Ocean. Sci. Lett., 12, 116123, https://doi.org/10.1080/16742834.2019.1569456.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ma, Z. G., and C. B. Fu, 2006: Some evidence of drying trend over northern China from 1951 to 2004. Chin. Sci. Bull., 51, 29132925, https://doi.org/10.1007/s11434-006-2159-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ma, Z. G., and C. B. Fu, 2007: Global aridification in the second half of the 20th century and its relationship to large-scale climate background. Sci. China, 50D, 776788, https://doi.org/10.1007/s11430-007-0036-6.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maestre, F. T., and Coauthors, 2013: Changes in biocrust cover drive carbon cycle responses to climate change in drylands. Global Change Biol., 19, 38353847, https://doi.org/10.1111/gcb.12306.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mahrt, L., and H. Pan, 1984: A two-layer model of soil hydrology. Bound.-Layer Meteor., 29, 120, https://doi.org/10.1007/BF00119116.

  • 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.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Passerat de Silans, A. M. B., B. A. Monteny, and J. P. Lhomme, 1996: Apparent soil thermal diffusivity, a case study: HAPEX-Sahel experiment. Agric. For. Meteor., 81, 201216, https://doi.org/10.1016/0168-1923(95)02323-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Piao, S. L., and Coauthors, 2010: The impacts of climate change on water resources and agriculture in China. Nature, 467, 4351, https://doi.org/10.1038/nature09364.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pielke, R. A., R. Avissar, M. R. Raupach, H. A. Dolman, X. B. Zeng, and S. Denning, 1998: Interactions between the atmosphere and terrestrial ecosystems: Influence on weather and climate. Global Change Biol., 4, 461475, https://doi.org/10.1046/j.1365-2486.1998.t01-1-00176.x.

    • Search Google Scholar
    • Export Citation
  • Reed, S. C., K. K. Coe, J. P. Sparks, D. C. Housman, T. J. Zelikova, and J. Belnap, 2012: Changes to dryland rainfall result in rapid moss mortality and altered soil fertility. Nat. Climate Change, 2, 752755, https://doi.org/10.1038/nclimate1596.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Safriel, U., and Z. Adeel, 2005: Dryland systems. Ecosystems and Human Well-being. Current State and Trend, R. Hassan et al., Eds., Island Press, 623662.

    • Crossref
    • Export Citation
  • Schaaf, C. B., and Coauthors, 2002: First operational BRDF, albedo nadir reflectance products from MODIS. Remote Sens. Environ., 83, 135148, https://doi.org/10.1016/S0034-4257(02)00091-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sheppard, P. A., 1958: Transfer across the Earth’s surface and through the air above. Quart. J. Roy. Meteor. Soc., 84, 205224, https://doi.org/10.1002/qj.49708436102.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sridhar, V., R. L. Elliott, C. Fei, and J. A. Brotzge, 2002: Validation of the NOAH-OSU land surface model using surface flux measurements in Oklahoma. J. Geophys. Res., 107, 4418, https://doi.org/10.1029/2001JD001306.

    • Search Google Scholar
    • Export Citation
  • Stephens, G. L., 2005: Cloud feedbacks in the climate system: A critical review. J. Climate, 18, 237273, https://doi.org/10.1175/JCLI-3243.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, G., J. Huang, W. Guo, J. Zuo, J. Wang, J. Bi, Z. Huang, and J. Shi, 2010: Observation analysis of land-atmosphere interactions over the Loess Plateau of northwest China. J. Geophys. Res., 115, D00K17, https://doi.org/10.1029/2009JD013372.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, J. M., and Y. Mitsuta, 1991: Turbulence structure and transfer characteristics in the surface layer of the HEIFE Gobi area. J. Meteor. Soc. Japan, 69, 587593, https://doi.org/10.2151/jmsj1965.69.5_587.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, J. M., and Y. Mitsuta, 1992: Evaporation from the desert: Some preliminary results of HEIFE. Bound.-Layer Meteor., 59, 413418, https://doi.org/10.1007/BF02215461.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, L., P. D’Odorico, J. P. Evans, D. J. Eldridge, M. F. McCabe, K. K. Caylor, and E. G. King, 2012: Dryland ecohydrology and climate change: Critical issues and technical advances. Hydrol. Earth Syst. Sci., 16, 25852603, https://doi.org/10.5194/hess-16-2585-2012.

    • Search Google Scholar
    • Export Citation
  • Wang, Z., X. Zeng, M. Barlage, R. E. Dickinson, and C. B. Schaaf, 2004: Using MODIS BRDF and albedo data to evaluate global model land surface albedo. J. Hydrometeor., 5, 314, https://doi.org/10.1175/1525-7541(2004)005<0003:UMBAAD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, Z., M. Barlage, X. Zeng, R. E. Dickinson, and C. B. Schaaf, 2005: The solar zenith angle dependence of desert albedo. Geophys. Res. Lett., 32, L05403, https://doi.org/10.1029/2004GL021835.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wielicki, B. A., T. Wong, N. Loeb, P. Minnis, K. Priestley, and R. Kandel, 2005: Changes in Earth’s albedo measured by satellite. Science, 308, 825, https://doi.org/10.1126/science.1106484.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wu, G. X., Y. M. Liu, X. Zhu, W. Li, R. Ren, A. M. Duan, and X. Liang, 2009: Multi-scale forcing and the formation of subtropical desert and monsoon. Ann. Geophys., 27, 36313644, https://doi.org/10.5194/angeo-27-3631-2009.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xiao, X., H. C. Zuo, Q. D. Yang, S. J. Wang, L. J. Wang, J. W. Chen, B. L. Chen, and B. D. Zhang, 2012: On the factors influencing surface-layer energy closure and their seasonal variability over the semi-arid Loess Plateau of Northwest China. Hydrol. Earth Syst. Sci., 16, 893910, https://doi.org/10.5194/hess-16-893-2012.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xiao, Z. X., and A. M. Duan, 2016: Impacts of Tibetan Plateau snow cover on the interannual variability of the East Asian summer monsoon. J. Climate, 29, 84958514, https://doi.org/10.1175/JCLI-D-16-0029.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xu, Z., Z. Hou, Y. Han, and W. Guo, 2016: A diagram for evaluating multiple aspects of model performance in simulating vector fields. Geosci. Model Dev., 9, 43654380, https://doi.org/10.5194/gmd-9-4365-2016.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xu, Z., Y. Han, and C. Fu, 2017: Multivariable integrated evaluation of model performance with the vector field evaluation diagram. Geosci. Model Dev., 10, 38053820, https://doi.org/10.5194/gmd-10-3805-2017.

    • Search Google Scholar
    • Export Citation
  • Yang, K., and Coauthors, 2008: Turbulent flux transfer over bare-soil surfaces: characteristics and parameterization. J. Appl. Meteor. Climatol., 47, 276290, https://doi.org/10.1175/2007JAMC1547.1.

    • Search Google Scholar
    • Export Citation
  • Zhang, X. S., 1979: The Tibetan Plateau and the vegetation of China-geographical distribution characteristics of the Plateau in China related to the Plateau’s Role in atmospheric circulation (in Chinese). J. Xinjiang Agric. Univ., 1, 413, https://doi.org/CNKI:SUN:XJNY.0.1979-01-001.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, X. S., 1993: A vegetation-climate classification system for global change studies in China. Quat. Sci., 13, 157169.

  • Zhou, L., C. J. Tucker, R. K. Kaufmann, D. Slayback, N. V. Shabanov, and R. B. Myneni, 2001: Variations in northern vegetation activity inferred from satellite data of vegetation index during 1981 to 1999. J. Geophys. Res., 106, 20 06920 083, https://doi.org/10.1029/2000JD000115.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zilitinkevich, S., 1995: Non-local turbulent transport: Pollution dispersion aspects of coherent structure of convective flows. Trans. Ecol. Environ., 6, 5360, https://doi.org/10.2495/AIR950071.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save
  • An, Z. S., 2000: The history and variability of the East Asian paleomonsoon climate. Quat. Sci. Rev., 19, 171187, https://doi.org/10.1016/S0277-3791(99)00060-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • An, Z. S., J. E. Kutzbach, W. L. Prell, and S. C. Porter, 2001: Evolution of Asian monsoons and phased uplift of the Himalaya-Tibetan plateau since late Miocene times. Nature, 411, 6266, https://doi.org/10.1038/35075035.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Berg, A., and Coauthors, 2016: Land-atmosphere feedbacks amplify aridity increase over land under global warming. Nat. Climate Change, 6, 869874, https://doi.org/10.1038/nclimate3029.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cao, Q., D. Y. Yu, M. Georgescu, and J. G. Wu, 2018: Substantial impacts of landscape changes on summer climate with major regional difference: The case of China. Sci. Total Environ. 625, 416427, https://doi.org/10.1016/j.scitotenv.2017.12.290.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cao, Q., J. G. Wu, D. Y. Yu, and W. Wang, 2019: The biophysical effects of the vegetation restoration program on regional climate metrics in the Loess Plateau, China. Agric. For. Meteor., 268, 169180, https://doi.org/10.1016/j.agrformet.2019.01.022.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, F., and J. Dudhia, 2001: Coupling an advanced land surface-hydrology model with the Penn State–NCAR MM5 modeling system. Part II: Preliminary model validation. Mon. Wea. Rev., 129, 587604, https://doi.org/10.1175/1520-0493(2001)129<0587:CAALSH>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cheng, S. J., X. D. Guan, J. P. Huang, F. Ji, and R. X. Guo, 2015: Long-term trend and variability of soil moisture over East Asia. J. Geophys. Res. Atmos., 120, 86588670, https://doi.org/10.1002/2015JD023206.

    • Crossref
    • Search Google Scholar
    • Export Citation