Contents lists available at ScienceDirect

Catena

journal homepage: www.elsevier.com/locate/catena

Exploring the role of land restoration in the spatial patterns of deep soilwater at watershed scales

Yali Zhaoa,c,d, Yunqiang Wanga,b,c,⁎, Li Wangc,d, Xiaoyan Zhanga, Yunlong Yua,c, Zhao Jina,Henry Lina,e, Yiping Chena,c, Weijian Zhoua,b,c, Zhisheng Ana,b,c

a State Key Laboratory of Loess and Quaternary Geology, Institute of Earth Environment, Chinese Academy of Sciences, Xi'an, Shaanxi 710061, Chinab Interdisciplinary Research Center of Earth Science Frontier, Beijing Normal University, Beijing 100875, ChinacGraduate University of Chinese Academy of Sciences, Beijing 100049, Chinad State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau, Institute of Soil and Water Conservation, Chinese Academy of Sciences & Ministry ofWater Resources, Yangling 712100, Chinae Department of Ecosystem Science and Management, Penn State Univ., University Park, PA, USA

A R T I C L E I N F O

Keywords:Deep soilLand useSoil water managementGrain for Green projectGully Land Consolidation projectThe Loess Plateau

A B S T R A C T

Soil water is a key variable for re-vegetation and environmental restoration in water-limited terrestrial eco-systems such as the Chinese Loess Plateau. Large land restoration projects (e.g., the “Grain for Green” launchedin 1999 and the “Gully Land Consolidation” launched in 2011) had substantial impacts on the storage, dis-tribution, and spatial patterns of soil water, and these factors remain poorly understood across watershed scales.We measured the amount of water stored in soil layers from the surface down to 5m depth and characterized thevertical distribution of gravimetric soil water content (SWC) among four land uses (cropland, shrubland, for-estland, and orchard), two slope aspects (shady vs. sunny), and two landforms (slope vs. gully) in three wa-tersheds on the Chinese Loess Plateau. All three of the watersheds were affected by Grain for Green project, twowere affected by the Gully Land Consolidation project (named NG and GT-T watersheds) and one was unaffectedby restoration efforts (named GT-U watershed). In the three watersheds, the slope and gully SWCs varied from2.4 to 24.2% and from 4.8 to 46.6%, respectively, during the sampling period in October 2015 (end of the rainyseason). The amount and vertical distribution of slope SWC differed significantly among the land uses andbetween shady and sunny slopes in the three watersheds (p < 0.05). The mean gully SWC (20.4%) was sig-nificantly higher than the mean slope SWC (8.7%) for each of the three watersheds (p < 0.01). Gullies filled bythe Gully Land Consolidation project had a large capacity to store soil water by increasing the infiltration ofprecipitation, which accounted for 14.7% and 11.3% of the total annual rainfall in the NG and GT-T watersheds,respectively. Filled gullies can serve as large reservoirs of soil water to relieve the problem of water shortage andcan also increase the amount of land available for cultivation to ease deficits in food production. A combinationof the Grain for Green project on slopes for soil conservation and the Gully Land Consolidation project in gulliesfor storing more soil water and increasing farmland area is an effective land restoration strategy on the ChineseLoess Plateau and is helpful for managing water cycles in regions around the world with deep soils.

1. Introduction

The Chinese Loess Plateau covers an area of ~620,000 km2. It is atypical fragile terrestrial ecosystem; ~1.6 billion tonnes of sedimentshave been eroded into the Yellow River in past decades (Chen et al.,2007b; Zhang and Liu, 2005). The unique geomorphology createsspecific conditions that prompt of soil deposition, but ecological servicefunctions are weak making governmental strategies for long-term

sustainable development on the Chinese Loess Plateau difficult to for-mulate (Dang et al., 2013; Y.Q. Wang et al., 2015b). The Grain forGreen project and the Gully Land Consolidation project are two well-known initiatives implemented by the Chinese government to restorevegetation and create farmland for cultivation, respectively (Li et al.,2016; Liu et al., 2013; Uchida et al., 2005). The scales of both projectswere unprecedented, with budgets of ~337 billion RMB ($51.8 billion)and ~30 billion RMB ($4.6 billion), respectively (Jin, 2013; J.Y. Wang

https://doi.org/10.1016/j.catena.2018.09.004Received 7 April 2018; Received in revised form 21 July 2018; Accepted 2 September 2018

⁎ Corresponding author at: State Key Laboratory of Loess and Quaternary Geology, Institute of Earth Environment, Chinese Academy of Sciences, Xi'an, Shaanxi710061, China.

E-mail address: wangyunq04@163.com (Y. Wang).

Catena 172 (2019) 387–396

Available online 11 September 20180341-8162/ © 2018 Elsevier B.V. All rights reserved.

T

et al., 2013).The Chinese Loess Plateau is a predominantly water-limited area.

Soil water content (SWC) is the main factor limiting re-vegetation andenvironmental restoration efforts in this region. The SWC also plays akey role in both Grain for Green project and Gully Land Consolidationproject by controlling plant growth and land engineering, respectively.Therefore, determining the impacts of these projects on storage, verticaldistribution, and spatial patterns of soil water are essential. This isespecially important for water stored in deep soils which can be ex-tracted and used by perennial plants (widely planted in Grain for Greenproject) and can be sequestered and stored by land-filled gullies (resultsfrom Gully Land Consolidation project).

The landscape on the Chinese Loess Plateau has changed tre-mendously since the implementation of Grain for Green (Miao et al.,2012; Piao et al., 2005; B.Q. Zhang et al., 2013). Vegetation coveragenearly doubled between 1999 and 2013 (Chen et al., 2015). Wang et al.(2015a) reported that a large afforestation project was the largestcontributor to a reduction in soil erosion since the 1990s; this wasdetermined using an attribution approach to analyze 60 years of runoffand sediment load observations along the course of the Yellow Riverover the Chinese Loess Plateau. Grain for Green has achieved greatsuccess in environmental improvement and the control of soil erosion,but it has introduced a few new problems. For example, large-scaleconversion of farmland to forests and/or grasslands has led to a deficitof available farmland and to some extent, has threatened food securityof this region (Bennett, 2008; Chen et al., 2015; Liu et al., 2015). Theexcessive expansion of vegetation cover can also aggravate water re-source limitations which produce dried soil layers in the deep soilprofile. Such occurrences can adversely impact plant growth and lead tothe degradation or death of the restored vegetation (Li et al., 2008; Y.Q.Wang et al., 2010; Wang et al., 2009). Measures must be taken toachieve sustainable development.

Gully Land Consolidation was implemented in 2011 in Yan'an citywhich is located in the middle of the Chinese Loess Plateau. The goalwas to address the problem of the reduction in farmland which resultedfrom the wide implementation of Grain for Green. Gully LandConsolidation aims to create farmland by (1) consolidating gullies withsoil removed from the surrounding hills, (2) compacting the soil tomake it less susceptible to tunneling and subsidence and (3) leveling theland surface to create flat or gentle sloping fields (Liu et al., 2013). Thisproject plans to create 4.0million ha of farmland from 2011 to 2020 forplanting crops in what were originally gully channels. Liu et al. (2015)reported that Gully Land Consolidation can help to reduce soil transportby nearly 10% and create more terraces or land. However, the effect ofGully Land Consolidation project on SWC redistribution at the wa-tershed scale, and the amount of water that can be stored in new landsof the gully, remains unclear.

Considering that SWC plays a critical role in a series of eco-hydro-logical processes and may determine the transport of material and en-ergy, the absorption and utilization of water by plants, and is related tobiogeochemical processes (Bryan, 2000; Laio et al., 2001; Rodriguez-Iturbe, 2000; Vereecken et al., 2014; Western et al., 2002; Xia and Shao,2008), some researchers have investigated the heterogeneity of SWC inspace and time, and found that SWC can be affected by dozens of en-vironmental factors including land use/vegetation, topographic factors,and soil properties (Fu et al., 2003; Qiu et al., 2001; Qiu et al., 2003).However, the impacts of intensively-managed projects such as Grain forGreen and Gully Land Consolidation on the spatial patterns of deepSWCs at watershed scales are largely unknown, and this information isneeded to improve hydrological modeling and water resources man-agement.

Watersheds on the Chinese Loess Plateau can generally be dividedinto two parts: uplands which are usually hill slopes and top slopes,referred to as “slope SWC”, and gullies, referred to as “gully SWC”. Thesoil water regimes of slopes and gullies need to be studied to determinethe effect of Grain for Green and Gully Land Consolidation at a

watershed scale which is essential for a better understanding of theeffects of land restoration projects on hydrological processes.Evaluating the effect of Gully Land Consolidation on SWC redistributioncan further provide useful information to policy makers who are taskedwith making strategic decisions that will impact the long-term sus-tainability of the environment. Therefore, the objectives of this studywere: (1) to investigate the overall amount and vertical distribution ofSWC on slopes and in gullies within three watersheds, and determine ifthere are relationships between slope SWC and land use, and slope SWCand slope aspect, after 16 years of Grain for Green, and (2) to determinethe status of SWC in gullies in the three watersheds, and evaluate theeffect of Gully Land Consolidation on soil water storage at the wa-tershed scale.

2. Materials and methods

2.1. Description of the study area

We chose three representative watersheds to evaluate the effect ofGrain for Green and Gully Land Consolidation on soil water regimes. Allthree watersheds have been a part of the Grain for Green project since1999, and the vegetation coverage on the slopes has greatly increasedsince the initiation of Grain for Green. The gullies of the two water-sheds, Nangou (NG) and Gutun (GT-T), were filled with soil in 2013 aspart of Gully Land Consolidation. The third watershed, Gutun (GT-U),was used as a control for Gully Land Consolidation effects. The areas ofthe NG, GT-T and GT-U watersheds were 0.36, 0.45, and 0.43 km2,respectively (Fig. 1), and the areas of the gully land filled in NG and GT-T were 44,110 and 50,941m2, respectively.

All three watersheds are near the city of Yan'an in the middle of theChinese Loess Plateau (Fig. 1). This region has a typical continentalmonsoon climate with an annual mean precipitation of 541mm (Fig. 2)according to the Yan'an weather station data collected from 1956 to2015. A majority of the precipitation (70%) falls between June andSeptember, and a relatively high pan evaporation of 1000mm (Li et al.,2008). The mean annual temperature is approximately 9.8 °C (Fig. 2).The three watersheds are in a typical hilly-gully region and range inelevation from 966 to 1267m. The study area contains mainly loessialsoil with 4.1–7.1% clay, 63.3–70.5% silt, and 25.4–29.7% sand (TableS1). The soil has low fertility that is vulnerable to erosion. The pre-dominant types of land use are rain-fed farmland (in gullies), shrub-land, forest, and native grassland (on slopes). The most common typesof vegetation are locust (Robinia pseudoacacia L.), korshinsk pea shrub(Caragana korshinskii Kom.), sea buckthorn (Hippophae rhamnoides L.),and bunge needle grass (Stipabungeana Trin.). The trees (locust) andshrubs (sea buckthorn) on the slopes of the three watersheds wereplanted in 1999 when the Chinese government launched “Grain forGreen”, while grassland was naturally grown on the land after farm-lands were abandoned. Several (usually five years) years following theimplication of Grain for Green, the branches of locust trees werethinned once a year, the shrubs were kept without thinning, and whilegrazing is officially prohibited on the grasslands, it can occasionally befound as local farmers may secretly allow their sheep or cattle onto thegrassland. The current land use types of the three watersheds had beenestablished about 16 years prior to when the soil samples were col-lected. In contrast, the rain-fed agriculture in the gullies was establishedin 2013 when the Gully Land Consolidation project created the farm-land.

2.2. Soil sampling and laboratory analyses

We super imposed 80m×80m grids on Google maps of the NG,GT-T and GT-U watersheds (Fig. 1), ensuring that the sampling siteswere uniformly distributed in the watersheds and represented all landuses in each watershed. In each grid cell, we selected one representativesite where the dominant species appeared to collect soil samples on the

Y. Zhao et al. Catena 172 (2019) 387–396

388

slopes. Adjacent sampling sites were selected at intervals of ~40m torepresent the conditions of soil water throughout the three gullies. Weused a GPS receiver (5 m precision) to locate our pre-selected samplingsites and to record the longitude, latitude, and altitude of each site. Intotal, there were 49, 89 and 72 sampling points from the NG, GT-T andGT-U watersheds, respectively (Fig. 1).

Disturbed soil samples were collected to a depth of 500 cm at eachsite using a soil auger (5 cm in diameter) at 20-cm increments. A total of25 samples were collected from most sites, but for some sites, thenumber of soil samples throughout the profile were< 25 due to theoccurrence of shallow bedrock (below 300 cm, the cases in the GT-Uwatershed) or the occurrence of groundwater (as was found in the NGand GT-T watersheds). A total of 4873 soil samples were collected fromall sampling sites (N=210) with 1091, 2184 and 1598 samples fromthe NG, GT-T, and GT-U watersheds, respectively.

After collection, the soil samples were immediately sealed in air-tight aluminum cylinders and brought to laboratory. A subsample of thesoil was used to determine gravimetric SWC (g H2O/100 g dry soil,

expressed as % by weight in the rest of this paper) by weighing thesamples before and after oven-drying at 105 °C to a constant weight. Asecond subsample of soil was air-dried and passed through a 1-mmmesh for determining soil-particle composition by laser diffractionusing a Mastersizer 3000 (Malvern Instruments, Malvern, England). Allsoil samples for determining SWC were collected from October 1 to 29,2015 (end of rainy season), representing soil moisture status after therainy season.

Undisturbed soil cores were also collected to 500 cm depth (con-sidering the difficulty of collecting samplings in deep soil layers) in GT-T (N=7) and GT-U (N=3) watersheds using metal cylinders (dia-meter 5 cm, length 5 cm) for measuring soil bulk density (BD) and soilwater retention curves (SWRC). The BD data was used for calculatingsoil water storage (SWS). The SWRC data which was determined usingthe centrifugation method (Hitachi CR21G centrifuge; 20 °C) (Y.Q.Wang et al., 2015a; Y.Q. Wang et al., 2013) and this data was used forevaluating the feasibility of using SWC (not soil water potential) toquantify the soil water regime.

2.3. Statistical methods

Mean SWC in each watershed was calculated as:

(a) mean SWC at site i:

∑==

SWCN

SWC1i

l j

N

i j1

,

l

(1)

(b) mean SWC of layer j:

d

cb

a Fig. 1. Locations of (a) the Chinese Loess Plateau andthe study area, (b) the land filled gully in the Nangouwatershed (NG), (c) the land filled gully in the Gutunwatershed (GT-T), (d) the control gully in the Gutunwatershed (GT-U). The red dots represent samplingsites in the gullies, the green squares represent sam-pling sites on the slopes, and the blue lines representwater channels. (For interpretation of the references tocolor in this figure legend, the reader is referred to theweb version of this article.)

0

2

4

6

8

10

12

140

200

400

600

800

1000AP (mm) AT ( )

noitati

picerP

)m

m(

Tem

per

atu

re(

)

Fig. 2. Precipitation and air temperature of the study area from 1956 to 2015(AP, annual precipitation; AT, annual air temperature).

Y. Zhao et al. Catena 172 (2019) 387–396

389

∑==

SWCN

SWC1j

s i

N

i j1

,

s

(2)

where SWCi,j is the SWC of layer j at site i, Nl is the number of sampledsoil layers at each site (i.e., 25 in general, but restricted to 15 in somesites), and Ns is the number of sampling sites (i.e., 210).

The SWS in the gullies was calculated as:

= ⋅ ⋅ ⋅SWS SWC BD h 10i j i j i j, , , (3)

∑==

SWS SWSij

N

i j1

,

l

(4)

= ⋅SWV SWS AS i (5)

where SWSi,j is the SWS per unit area of layer j at site i, BDi,j is soil bulkdensity of layer j at site i, h is the soil-depth interval (in our study is20 cm), 10 is a unit conversion factor (mm cm−1), SWSi is the total SWSat site i in the gully, Nl is the number of sampled soil layers at each sitein the gully, SWV is the increase in SWS volume due to Gully LandConsolidation, SWSS is the difference of SWS in the 0–500 cm profile oftreated gully SWS minus the untreated gully in the 0–300 cm layers,and Ai is the segmented area of the land filled gully in the GT-T wa-tershed.

The highest SWC in the profile was used to replenish the vacant datafor calculating SWS for the sites where samples could not be collectedthroughout the entire 500 cm depth due to the occurrence of ground-water in the treated gullies in the NG and GT-T watersheds. Wetherefore used the highest SWC of the profile as the upper limit forcalculating SWV. Due to the occurrence of shallow bedrock in the GT-Uwatershed, soil samples were only collected to a maximum depth of300 cm. In addition, considering that several dams were established toensure stability of filled land in the gullies, we divided the filled landinto several segments to more precisely calculating SWV.

Primary statistical analyses, such as determinations of mean, stan-dard deviation (SD) and the coefficient of variation (CV) of measuredSWCs, were conducted using Microsoft Excel (version 2010). Leastsignificance difference (LSD) tests were used to identify significantdifferences in SWC among the land uses at each watershed, and t-testswere used to identify significant differences between slope SWC andgully SWC, and between sunny SWC and shady SWC for each watershedusing SPSS (version 16.0). The areas of the land filled gullies weremeasured using GIS software (ESRI® ArcMap™ 9.2).

3. Results

3.1. Soil physical properties in the 0–500 cm profile in the three watersheds

Soil texture was comparable between slopes and gullies in the threewatersheds with similar mean soil fraction contents (Fig. S1). Slopesites had a higher mean silt content (69.7%) and a lower proportion ofclay (26.0%) and sand (4.3%). The clay and sand contents exhibitedrelatively high variation (11% < CV < 26%) compared to silt(4% < CV < 9%) (Table S1), but the soil texture in the three water-sheds was similar. On the slopes, the sand contents decreased graduallyto a depth of 100 cm, while silt contents increased. In contrast, therewere only small vertical changes to soil texture which showed a slightwaving pattern in the gullies (Fig. S1). The CVs of the soil particlecomposition in the vertical profile were all< 10% (data not shown)which shows the weak variation in soil texture throughout the 500 cmsoil layers in the three watersheds.

Table 1 and Fig. 3 show that, along the 500 cm profile, there wereno significant differences for soil water retention parameters and re-lated hydraulic properties as fitted by using the van Genuchten equa-tion. This further indicated that it is feasible to use soil water content toevaluate the soil water quantity and its distribution instead of using soilwater potential.

3.2. Overall characteristics of soil water in the three watersheds

The SWCs at all sites ranged from 2.4 to 46.6%, with a mean of11.2% and a CV of 53% (Table 2). The overall vertical distribution ofSWC in the 0–500 cm soil profile for all sites in the three watershedswas generally stable near 10% but was higher in the 0–20 cm surfacelayer (~14%). The distribution of the CVs below 20 cm tended to de-crease with depth. The SWC varied less near the surface than in the20–500 cm soil layers.

Among the three watersheds, mean SWCs differed significantly be-tween the NG and GT-T, GT-U watersheds (p < 0.01), in the sequenceof the NG (12.3%) > GT-T (10.9%) > GT-U (10.6%) (Table 2). TheSWCs in the NG watershed ranged from 3.4 to 41.6% and fluctuatedslightly within 0–500 cm soil profile (Fig. 4a). The SWCs in the GT-Tand GT-U watersheds ranged from 2.4 to 32.7% and from 3.5 to 46.6%,respectively, and showed a similar vertical pattern that it first decreasedin the 0–20 cm surface layer and then remained stable at near 10% withincreasing soil depth.

The CVs of the SWCs in each soil layer ranged from 34 to 67%, 29 to59%, and 40 to 71% in the NG, GT-T and GT-U watersheds, respectively(Fig. 4), and the vertical distribution patterns of the CVs differed amongthe three watersheds, although the mean CVs indicated a similar var-iation (50% < mean CVs < 55%). The CVs for the NG watershed firstincreased to 120 cm and then gradually decreased, the CVs for the GT-Twatershed were highest at 40 cm and then decreased to 500 cm and theCVs for the GT-U watershed first increased in the 0–60 cm layer, de-creased to a depth of 400 cm and then increased below 400 cm.

3.3. Slope SWCs in the three watersheds

The slope SWCs of the three watersheds ranged from 2.4 to 24.2%,with a mean of 8.7%. The slope CVs ranged from 36 to 42%. In thevertical direction (Fig. 5a), slope SWCs tended to decrease in the0–100 cm layers and then gradually increased to a depth of 500 cm.Slope CVs were low but fluctuated in the top 120 cm of soil and thengenerally increased to 500 cm.

Slope SWC was significantly lower under forest and shrubland thanthat under grassland and orchard in all three watersheds (Table S2).The vertical distributions of slope SWCs under the different land uses inthe three watersheds generally first decreased to about 120 cm and thenincreased to 500 cm, although the amounts of decreases or increasesdiffered (Fig. 6). Slope SWCs under forest and shrubland decreasedrapidly (from 10% to about 6%), then increased slightly (except in theGT-U watershed) to about 8% and these land uses had lower levels ofsoil water compared to the other two land uses. Slope SWCs under thegrasslands and orchards first decreased (from 13 to about 8%) and thengradually increased to about 15%. The slope SWCs for the orchards and

Table 1Results of fitted van Genuchten equation parameters in 0–500 cm soil profilefrom the land filled gully in the Gutun watershed (GT-T).

Soil depth θr θs α n

50 cm 0.029 0.351 0.021 1.694100 cm 0.031 0.39 0.022 1.649150 cm 0.029 0.403 0.017 1.708200 cm 0.036 0.394 0.016 1.709250 cm 0.047 0.395 0.015 1.722300 cm 0.052 0.399 0.014 1.696350 cm 0.052 0.388 0.014 1.683400 cm 0.045 0.379 0.014 1.706450 cm 0.061 0.360 0.011 1.722500 cm 0.054 0.351 0.011 1.702

Note: θr is the residual soil water content (cm3 cm−3); θs is the saturated soilwater content (cm3 cm−3); n is a curve-shape parameter related to soil pore sizedistribution; α is the scaling parameter related to the inverse of the air entrypressure (cm−1).

Y. Zhao et al. Catena 172 (2019) 387–396

390

grasslands began to differ from those in the forest and shrublands belowa soil depth of 120 cm which had the lowest slope SWC. Slope SWC washighest under the orchard in the NG watershed.

The slope SWC was lower for the sunny than for the shady slopes inall watersheds but varied similarly between sunny and shady slopeswithin each watershed in the 0–500 cm soil profiles. The LSD test

indicated that the slope SWCs were significantly lower for the sunnythan for the shady slopes in the GT-T (7.3% and 8.8%, respectively) andthe GT-U watersheds (7.4% and 10.4%, respectively) (p < 0.01), butthe difference between the sunny and shady slopes was not significantin the NG watershed (8.5% and 8.9%, respectively).

Soil

wat

er c

onte

nt

(cm

3/c

m3)

0

0.1

0.2

0.3

0.4

0.5

0 5000 10000 15000

50 cm

0

0.1

0.2

0.3

0.4

0.5

0 5000 10000 15000

100 cm

0.0

0.1

0.2

0.3

0.4

0.5

0 5000 10000 15000

200 cm

0.0

0.1

0.2

0.3

0.4

0.5

0 5000 10000 15000

250 cm

0.0

0.1

0.2

0.3

0.4

0.5

0 5000 10000 15000

300 cm

0

0.1

0.2

0.3

0.4

0.5

0 5000 10000 15000

350 cm

0

0.1

0.2

0.3

0.4

0.5

0 5000 10000 15000

400 cm

0

0.1

0.2

0.3

0.4

0.5

0 5000 10000 15000Pressure Head (cm)

500 cm

0

0.1

0.2

0.3

0.4

0.5

0 5000 10000 15000Pressure Head (cm)

450 cm

0.0

0.1

0.2

0.3

0.4

0.5

0 5000 10000 15000

150 cm

Fig. 3. Soil water retention curve in 0–500 cm soil profile at the land filled gully in the Gutun watershed (GT-T) watershed. Red circles represent the measured dataand solid lines are the curves fitted by the van Genuchten model. (For interpretation of the references to color in this figure legend, the reader is referred to the webversion of this article.)

Y. Zhao et al. Catena 172 (2019) 387–396

391

3.4. Gully SWCs in the three watersheds

In the three watersheds, gully SWCs varied from 4.8 to 46.6%, witha mean of 20.7%, but the CVs varied little, ranging from 20 to 23%(Table 2). The mean gully SWC was significantly higher than the meanslope SWC (p < 0.01). Gully SWC tended to fluctuate slightly withincreasing soil layers (Fig. 5b). Gully CVs were high to a depth of140 cm and then decreased erratically to 500 cm depth.

Gully SWCs generally increased with depth in the NG and GT-Twatersheds (Fig. 7), but the gully CVs fluctuated sharply in some soillayers (e.g., at 220 and 320 cm in the NG watershed, at 220 and 480 cmin the GT-T watershed). In the GT-U watershed, soil samples were onlycollected to a maximum depth of 300 cm due to the shallow bedrock.Gully SWCs in the GT-U watershed tended to decrease within the0–300 cm profile, and the vertical distributions of gully SWCs werehighly variable (gully CVs varied from 10 to 38%). In contrast, the gullySWCs in the 0–140 cm layers were higher in the GT-U than the NG andGT-T watersheds.

The maximum sampling depths in the gullies were 500 cm in the NGand GT-T watersheds and 300 cm in the GT-U watershed, so it can beinferred that the mean depth to which the land was filled in the GullyLand Consolidation project was approximately 200 cm at the watershedscale, and this has been confirmed by local engineers and farmers. Wetherefore evaluated in the next section the impacts of Gully LandConsolidation on the soil water regimes in the gullies of the watersheds.

Table 2Statistical summary of soil-water content (SWC) in the land filled gully in the Nangou watershed (NG), the landfilled gully in the Gutun watershed (GT-T), and thecontrol gully in the Gutun watershed (GT-U). Slope SWC is the SWC of sites on slopes, while gully SWC is the SWC of sites in gullies in each of the watersheds.

Watershed Location N n Min (%) Max (%) Mean (%) SD CV (%)

Total All sites 210 4873 2.4 46.6 11.2 5.9 53Slope SWC 148 3822 2.4 24.2 8.7 3.2 36Gully SWC 62 1050 4.8 46.6 20.4 4.3 21

NG All sites 49 1091 3.4 41.6 12.3 6.8 55Slope SWC 31 775 3.4 18.9 8.7 3.1 36Gully SWC 18 316 4.8 41.6 21.3 4.9 23

GT-T All sites 89 2184 2.4 32.7 10.9 5.7 52Slope SWC 64 1597 2.4 23.3 8.3 3.3 39Gully SWC 25 587 7.5 32.7 19.1 3.8 20

GT-U All sites 72 1598 3.5 46.6 10.6 5.5 52Slope SWC 58 1450 3.5 24.2 9.4 3.9 42Gully SWC 14 148 10.9 46.6 22.5 4.9 22

Note: N is the number of sampling sites, n is the number of soil samples, SD is standard deviation, and CV is coefficient of variation.

Soil-water content %0 10 20 30 40

)mc(

htpe

dli

oS

0

100

200

300

400

50020 30 40 50 60 70 80

0 10 20 30 40

20 30 40 50 60 70 80

0 10 20 30 40

20 30 40 50 60 70 80

a b c

Coefficient of variation (%)Coefficient of variation (%)

Soil-water content % Soil-water content %

Coefficient of variation (%)

Fig. 4. Vertical distributions of soil-water content (SWC) inthe 0–500 cm soil profile in each of the three watersheds:means and standard deviations (red dots and bars) andcoefficients of variation (blue dots) in (a) the land filledgully in the Nangou watershed (NG), (b) the land filled gullyin the Gutun waterhsed (GT-T) and (c) the control gully inthe Gutun watershed (GT-U). Gray circles represent in-dividual sampling sites in each watershed. (For interpreta-tion of the references to color in this figure legend, thereader is referred to the web version of this article.)

Soil-water content (%)

0 10 20 30 40

)m

c(ht

ped

lio

S

0

100

200

300

400

500

Coefficient of variation (%)

0 10 20 30 40 50

a

0 10 20 30 40

0 10 20 30 40 50

Coefficient of variation (%)

Soil-water content (%)

b

Fig. 5. Vertical distributions of means and standard deviations of soil-watercontents (red dots and bars) and their coefficients of variation (blue dots) in the0–500 cm soil profiles in the three watersheds for slopes (a) and gullies (b). (Forinterpretation of the references to color in this figure legend, the reader is re-ferred to the web version of this article.)

Y. Zhao et al. Catena 172 (2019) 387–396

392

3.5. Effects of Gully Land Consolidation on soil water storage

In the GT-T and GT-U watersheds, BD gradually increased from soildepths of 0 to 30 cm, and then fluctuated from 1.40 to 1.60 g/cm3

below 30 cm depth. This suggests the shallow soil layers were easilyaffected by rainfall and human activities (Fig. 8). The BD of the GT-Twatershed was larger than that of the GT-U watershed which may haveresulted due to from machine compaction of the land that occurredduring the gully filling.

The mean SWS was 1062mm in the control gully in the GT-U wa-tershed, while the mean SWSs in the landfilled gullies in the NG andGT-T watersheds were 1702 and 1633mm (Fig. 9), respectively(Table 3). The comparable SWS between the NG and GT-T watersheds -the two land filled watersheds - suggests that the calculation of a meanland fill depth of 200 cm by the Gully Land Consolidation project isreasonable.

We calculated the soil water budgets as a result of the Gully LandConsolidation project according to Eq. (5). The estimated increasedyields of soil water by the Gully Land Consolidation project were2.8×104 and 2.7×104m3 in the land filled NG and GT-T gullies,respectively. The mean total amount of annual rainfall was 541mm(from 1956 to 2015), and the total areas of the two watersheds were0.36 and 0.45 km2, respectively. We used this information to estimatethat the increased yields of soil water by the Gully Land Consolidationproject accounted for 14.7 and 11.3% of the total annual rainfall in theNG and GT-T watersheds, respectively. Such a high proportion of re-distribution of water resources by the Gully Land Consolidation projectwould profoundly affect the spatial patterns of soil water resources atthe watershed scale.

Soil-water content %

0 5 10 15 20

htpe

dli

oS

mc

0

100

200

300

400

5000 5 10 15 20

Soil-water content %

0 5 10 15 20

Soil-water content %

Forest Grassland Shrubland Orchard

a b c

Fig. 6. Vertical patterns of means and standard deviations of soil-water content under the four land uses in each watershed: (a) land filled gully in the Nangouwatershed (NG), (b) land filled gully in the Gutun watershed (GT-T), and (c) control gully in the Gutun watershed (GT-U).

Soil water content (%)

10 15 20 25 30 35

)mc(

htpedlioS

0

100

200

300

400

500

Coefficient of variance (%)

0 5 10 15 20 25 30 35 40

a b

Fig. 7. Vertical distribution of (a) means and standard deviations of gully soil-water content and (b) their coefficients of variation in each of the three wa-tersheds: the land filled gully in the Nangou watershed (NG) (red), the landfilled gully in the Gutun watershed (GT-T) (blue), and the control gully in theGutun watershed (GT-U) (green), respectively. (For interpretation of the re-ferences to color in this figure legend, the reader is referred to the web versionof this article.)

1.2 1.3 1.4 1.5 1.6 1.7

)m

c(ht

pe

dli

oS

0

100

200

300

400

500

BD (g/cm3)

Fig. 8. Soil bulk density distribution to 500 cm in the land filled gully in theGutun watershed (GT-T) showing the means and standard errors (red dots andbars), and the control gully in the Gutun watershed (GT-U) showing means andstandard errors (blue dots and bars). (For interpretation of the references tocolor in this figure legend, the reader is referred to the web version of thisarticle.)

Y. Zhao et al. Catena 172 (2019) 387–396

393

4. Discussion

4.1. Factors affecting soil water content at the watershed scale

Soil water content is influenced by many biotic and abiotic factorsincluding climatic conditions (e.g. rainfall, evaporation, temperatureand wind speed) (Chen et al., 2008a; Wang et al., 2012), vegetationcharacteristics (e.g. land use pattern, plant species, coverage/densityand growth age) (Y.Q. Wang et al., 2011), soil properties (e.g. bulkdensity, saturated hydraulic conductivity and particle composition)(Namdar-Khojasteh et al., 2012), topographic elements (e.g. altitude,slope gradient, aspect and position) (Hu and Si, 2014), human activities(e.g. land use changes, vegetation management, grazing and landconsolidation) (Zhang et al., 2014) and their interactions at a widerange of spatial and temporal scales (Dekker et al., 2007; Hu and Si,2013; Van Ek et al., 2000; X.X. Wang et al., 2010). In our study, the soilmoisture status in the three watersheds was significantly influenced byland restoration projects (Grain for Green and Gully Land

Consolidation), which can be connected to the change of in land usepatterns on the slopes and the reconstructed topography characters ingullies before and after the two restoration projects.

The implication of Grain for Green dramatically altered the patternsof land uses in the watersheds as well as the soil water regimes. Thesignificant difference in SWC among forest/shrubland and grasslandmay be due to the higher capacity of water uptake by forest/shrublandassociated with the higher aboveground biomass, photosynthesis, rootdepth, and transpiration compared to those of grassland, where rootsare mostly distributed within the top 200 cm of soil. In contrast, thesimilar soil water status between forest and shrubland in a watershedmay have been caused by (1) a similar distribution pattern of roots,especially the rooting depth, which can both extend to 10m or more inour study area (Y.Q. Wang et al., 2013; Y.Q. Wang et al., 2015b), or by(2) very low slope SWCs, which can be defined as dried soil layers (Y.Q.Wang et al., 2011). In both forest and shrublands, such low SWCs werenot reflective of the difference of slope SWCs between forest andshrubland in the 0–500 cm profile. The SWC was unexpectedly higher

a b

Fig. 9. Filled land area in (a) the land filled gully in the Nangou watershed (NG), (b) the land filled gully in the Gutun watershed (GT-T). NAi (i= 1, 2, 3, 4) and TAi

(i= 1, 2, 3, 4, 5) are the segmented land area filled by check-dams in NG and GT-T, respectively.

Table 3Statistics of soil-water storage in the land filled gully in the Nangou watershed (NG), the land filled gully in the Gutun watershed (GT-T), and the control gully in theGutun watershed (GT-U).

Watershed LF (m2) SWS (mm) SWSS (mm) SWV (m3) TA (km2) P (mm) R (%)

NG NA1 5311 1516 454 2411NA2 11,633 1787 725 8433NA3 9528 1768 706 6726NA4 17,638 1737 675 11,905

44,110 28,665 0.36 541 14.7GT-T TA1 6432 1689 627 4032

TA2 8086 2000 938 7584TA3 10,556 1576 513 5415TA4 4608 1422 359 1654TA5 21,259 1479 417 8865

50,941 27,550 0.45 541 11.3GT-U 0 1062 0 0 0.43 541 0

Note: LF, landfilled area of each gully; TA, total area of each watershed; P, annual precipitation of each watershed; SWS, soil-water storage in each segment; SWSS,the difference of SWS in the 0–500 cm profiles of treated gully SWS minus the untreated gully; SWV, the increase in SWS volume due to the Gully Land Consolidation(GLC); R, ratio of increased gully yield by GLC to the total annual rainfall.

Y. Zhao et al. Catena 172 (2019) 387–396

394

in the orchard in the NG watershed than in the forest, shrubland andgrassland. This may have been because (1) the orchard trees wereplanted at the top of the slope leveled by local farmers to increase in-filtration of precipitation, or because (2) the orchard trees were plantedabout 10 years earlier so their capacity to transpire may not fully de-veloped (Jian et al., 2015; Y.Q. Wang et al., 2015c).

The mean gully SWC was affected by Gully Land Consolidation andwas expected to be significantly higher than the mean slope SWC(Fig. 5, p < 0.01) at the watershed scale because of channeling ofsurface runoff and subsurface flow to the gullies. The slope CVs of SWCwere generally higher than that of the gully CVs (Fig. 5 and Table 2),which is in accordance with the results of Hu et al. (2010), Hu and Si(2013) and Lin (2006) who reported that the variation in SWC underwet conditions were generally lower than that under dry conditionssince the factors controlling SWC under different water regimes weredifferent.

Slope aspect also plays an important role in determining the con-ditions of slope SWC (Huang et al., 2012; Liu et al., 2012; Qiu et al.,2001; Zhu et al., 2014). The difference of SWC between shady andsunny slopes was larger in the GT-T and GT-U watersheds than that inthe NG watershed. This result may be attributed to (1) the differences invegetation type and the density of vegetation coverage, and (2) thedifferent directions of the channels between the NG and GT-T/GT-Uwatersheds. Slope aspect is directly associated with the distribution ofsolar radiation, the microclimate, plant growth, temperature and mi-crobial activities, and thus soil water levels.

Our studies verified that SWC is affected by a series of factors suchas land use, slope position and aspect across a watershed scale. Moreimportantly, the human-induced land projects can aggravate, or evenreshape, the distribution patterns of soil water in the watershed, al-though the vertical soil hydraulic properties are similar (Fig. 3). Suchan effect should be taken into account when modeling hydrologicalprocesses as it has implications for the utilization of soil water resourcesin restored ecosystems.

4.2. Responses of soil water resources to the land restoration projects

The implementation of Grain for Green, which primarily aims toreduce soil erosion by transforming sloping cropland (previously re-claimed by local farmers for expansion of farmland) into grassland orforest, has greatly altered vegetation types, land use patterns, land-scapes and a series of eco-hydrological processes at local and regionalscales (Miao et al., 2012; Sun et al., 2015; B.Q. Zhang et al., 2013; Zhouet al., 2016). Changes in land use caused by Grain for Green canmarkedly affect the water cycle in a soil–plant–atmosphere system(Chen et al., 2007a; Fu et al., 2003; Li, 2000; Y.Q. Wang et al., 2015b;Yang et al., 2012) and thus alter the capacity of soil water storage (Cuoet al., 2013; Z.H. Zhang et al., 2013).

Our results suggest that Gully Land Consolidation increased SWSabout 640mm and 570mm in the NG and GT-T filled gullies (Table 3),respectively, and these filled gullies therefore represent a large re-servoir of soil water. The land created by Gully Land Consolidation isrelatively flat and composed of similar landscapes, so SWC is lessvariable within the gully soil profiles than within the slope soil profiles(Fig. 5b). The spatial variation of SWC at the watershed scale, however,increased as the gully SWC were included comparing to the casewithout gully SWC, which was consistent with the results reported byGao et al. (2016). The higher SWCs in the surface soil layers in the GT-Uwatershed compared to the NG and GT-T watersheds (Fig. 7) may bebecause the gully in the GT-U watershed functioned as a sink, but GullyLand Consolidation flattened and broadened the gullies in the NG andGT-U watersheds, dispersing water that could infiltrate which redis-tributed it more easily.

The land filled gullies can store water (about 2.8× 104 and2.7×104m3 in the land filled NG and GT-T gullies, respectively) forcrop cultivation on the new land, which can both solve the demand for

farmland by local farmers and relieve the pressure of water shortage.Compared with the GT-T gully, the higher SWS and SWV of the NGgully despite the smaller area of new land, can be explained by (1) thehigher clay content in the NG gully which may help it sequester moresoil water, and (2) the sand content in the GT-T gully was higher thanthat in the NG gully, which results in rapid drainage of soil water (Fig.S1 and Table S1).

New land in gullies, filled by Gully Land Consolidation to a depth ofapproximately 200 cm, could not only shorten the residence time ofrunoff generated on slopes, but also reduce the amount of runoff in thegully and decrease the flow velocity, allowing more water to infiltrateand be stored in the soils, which further results a decline of soil erosion.This explains why the SWS was nearly twice as high in the NG and GT-Twatersheds than in the GT-U watershed. The main benefit of Gully LandConsolidation is the increased ~200 cm soil in the gullies, plus in-creased area of gullies, helps store more water.

In summary, intensive land management projects from Grain forGreen and Gully Land Consolidation significantly affected the statusand patterns of both slope SWC and gully SWC (Figs. 7 and 9). Grain forGreen decreased slope SWC while Gully Land Consolidation sig-nificantly increased gully SWS. The two projects also influenced thevertical distribution of the SWCs both for slope SWC and gully SWC,and their variability (Figs. 5), which were highly correlated with rootpattern and other potential paths of water movement. Grain for Greenimplemented on slopes for conserving soil and Gully Land Consolida-tion conducted in gullies to increase the area of farmland and providefood is a combined strategy with unprecedented benefits for the Chi-nese Loess Plateau. The implementation of Grain for Green on slopes,however, can dramatically change the landscape and significantly affectSWC. The vegetation type and density must thus be carefully de-termined during the implementation of any procedure. Likewise, drai-nage in land filled gullies is a potential problem that may lead toponding and salinization, which is harmful to the growth of crops andthus needs to be considered.

More research is necessary to identify the benefits and dangers ofGully Land Consolidation because the long-term impacts on both socialand ecological development are difficult to foresee. From the viewpointof water management in this region, we can infer that Gully LandConsolidation has an important effect on soil water status by retaininglarge amounts of water in land filled gully soils. Gully LandConsolidation can thus be a useful strategy to provide farmland forcultivation, to increase the available water resources and to alleviatesoil desiccation, all of which would contribute to the sustainability ofthe ecology, environment and economic development in the hilly-gullyregion of the Chinese Loess Plateau (Liu et al., 2015). The effectivecombination of Grain for Green and Gully Land Consolidation should bea path to sustainable development that can both accelerate economicgrowth and promote ecological and environmental reconstruction onthe Chinese Loess Plateau.

5. Conclusions

We investigated the SWC to a depth of 500 cm in three watersheds,all were affected by Grain for Green on the Chinese Loess Plateau; twowatersheds, the NG and GT-T watersheds, contained land filled gulliesas part of Gully Land Consolidation, and one watershed, the GT-Uwatershed, contained a gully that was not filled. The slope SWC of thethree watersheds varied from 2.4 to 24.2%, and the gully SWC variedfrom 4.8 to 46.6%, with the gullies having mean SWCs that were 11.6%higher than on the slopes. The CVs were lower in the gullies than on theslopes. The vertical distribution of SWC on the slopes exhibited dif-ferent patterns among the four land uses. The SWC was higher on shadyslopes than on sunny slopes in all three watersheds. The mean SWS was1062mm in the GT-U watershed, while the mean SWSs in the land filledgullies in the NG and GT-T watersheds were 1702 and 1633mm, re-spectively. The increased yields of soil water as a result of Gully Land

Y. Zhao et al. Catena 172 (2019) 387–396

395

Consolidation were 2.8×104 and 2.7×104m3 in the NG and GT-Tgullies, respectively. Land filled gullies possessed a large capacity tostore water in the soils due to increased infiltration of precipitation; theincreased yield of soil water accounted for 14.7% and 11.3% of the totalannual rainfall in the NG and GT-T watersheds, respectively. Thecombination of Grain for Green on slopes and Gully Land Consolidationin gullies can both conserve soils and resolve water shortage issues thatare hindering development of this region. Our results also provideuseful information for developing important strategy decisions re-garding sustainable managements of soil and water resources in water-limited ecosystems.

Acknowledgements

This research was supported by the National Natural ScienceFoundation of China (Nos. 41571130083, 41471189, 41530854, and41722106), the “Ten Thousand Talent Program” for Young top-notchtalent, and the Shannxi Science and Technology Coordinator InnovativeEngineering Project (No. 2015KTZDNY01-04). Special thanks to Dr.William Blackhall for his help improving the manuscript.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.catena.2018.09.004.

References

Bennett, M.T., 2008. China's sloping land conversion program: institutional innovation orbusiness as usual? Ecol. Econ. 65, 699–711.

Bryan, R.B., 2000. Soil erodibility and processes of water erosion on hillslope.Geomorphology 32, 385–415.

Chen, L.D., Huang, Z.L., Gong, J., Fu, B.J., Huang, Y.L., 2007a. The effect of land cover/vegetation on soil water dynamic in the hilly area of the loess plateau, China. Catena70, 200–208.

Chen, L.D., Wei, W., Fu, B.J., Lü, Y.H., 2007b. Soil and water conservation on the LoessPlateau in China: review and perspective. Prog. Phys. Geogr. 31, 389–403.

Chen, H.S., Shao, M.A., Li, Y.Y., 2008a. The characteristics of soil water cycle and waterbalance on steep grassland under natural and simulated rainfall conditions in theLoess Plateau of China. J. Hydrol. 360, 242–251.

Chen, Y.P., et al., 2015. Balancing green and grain trade. Nat. Geosci. 8, 739–741.Cuo, L., Zhang, Y.X., Gao, Y.H., Hao, Z.C., Cairang, L.S., 2013. The impacts of climate

change and land cover/use transition on the hydrology in the upper Yellow RiverBasin, China. J. Hydrol. 502, 37–52.

Dang, X.H., Liu, G.B., Xue, S., Li, P., 2013. An ecological footprint and emergy basedassessment of an ecological restoration program in the Loess Hilly Region of China.Ecol. Eng. 61, 258–267.

Dekker, S.C., Rietkerk, M., Bierkens, M.F.P., 2007. Coupling microscale vegetation-soilwater and macroscale vegetation-precipitation feedbacks in semiarid ecosystems.Glob. Chang. Biol. 13, 671–678.

Fu, B.J., Wang, J., Chen, L.D., Qiu, Y., 2003. The effects of land use on soil moisturevariation in the Danangou catchment of the Loess Plateau, China. Catena 54,197–213.

Gao, X.D., Zhao, X.N., Wu, P.T., Brocca, L., Zhang, B.Q., 2016. Effects of large gullies oncatchment-scale soil moisture spatial behaviors: a case study on the Loess Plateau ofChina. Geoderma 261, 1–10.

Hu, W., Si, B.C., 2013. Soil water prediction based on its scale-specific control usingmultivariate empirical mode decomposition. Geoderma 193–194, 180–188.

Hu, W., Si, B.C., 2014. Revealing the relative influence of soil and topographic propertieson soil water content distribution at the watershed scale in two sites. J. Hydrol. 516,107–118.

Hu, W., Shao, M.A., Han, F.P., Reichardt, K., Tan, J., 2010. Watershed scale temporalstability of soil water content. Geoderma 158, 181–198.

Huang, Y.L., et al., 2012. Effect of land use and topography on spatial variability of soilmoisture in a gully catchment of the Loess Plateau, China. Ecohydrology 5, 826–833.

Jian, S.Q., Zhao, C.Y., Fang, S.M., Yu, K., 2015. Effects of different vegetation restorationon soil water storage and water balance in the Chinese Loess Plateau. Agric. For.Meteorol. 206, 85–96.

Jin, Z., 2013. The creation of farmland by gully filling on the Loess Plateau: a double-edged sword. Environ. Sci. Technol. 48, 883–884.

Laio, F., Porporato, A., Ridolfi, L., Rodriguez-Iturbe, I., 2001. Plants in water-controlledecosystems: active role in hydrologic processes and response to water stress: II.Probabilistic soil moisture dynamics. Adv. Water Resour. 24, 707–723.

Li, Y.S., 2000. Effects of forest on water circle on the Loess Plateau. J. Nat. Resour. 16,427–432.

Li, W., Wang, Q.J., Wei, S.P., Shao, M.A., Yi, L., 2008. Soil desiccation for Loess soils onnatural and regrown areas. For. Ecol. Manag. 255, 2467–2477.

Li, Y.H., Du, G.M., Liu, Y.S., 2016. Transforming the Loess Plateau of China. Front. Agr.Sci. Eng. 3, 181–185.

Lin, H., 2006. Temporal stability of soil moisture spatial pattern and subsurface pre-ferential flow pathways in the shale hills catchment. Vadose Zone J. 5, 317–340.

Liu, H.Y., et al., 2012. Topography-controlled soil water content and the coexistence offorest and steppe in northern China. Phys. Geogr. 33, 561–573.

Liu, Q., Wang, Y.Q., Zhang, J., Chen, Y.P., 2013. Filling gullies to create farmland on theLoess Plateau. Environ. Sci. Technol. 47, 7589–7590.

Liu, Y.S., Guo, Y.J., Li, Y.R., Li, Y.H., 2015. GIS-based effect assessment of soil erosionbefore and after gully land consolidation: a case study of Wangjiagou project region,Loess Plateau. Chin. Geogr. Sci. 25, 137–146.

Miao, C.Y., Yang, L., Chen, X.H., Gao, Y., 2012. The vegetation cover dynamics(1982–2006) in different erosion regions of the Yellow River Basin, China. LandDegrad. Dev. 23, 62–71.

Namdar-Khojasteh, D., Shorafa, M., Heidari, A., 2012. Estimating soil water content frompermittivity for different mineralogies and bulk densities. Soil Sci. Soc. Am. J. 76,1149–1158.

Piao, S.L., Fang, J.Y., Liu, H.Y., Zhu, B., 2005. NDVI-indicated decline in desertification inChina in the past two decades. Geophys. Res. Lett. 32, L06402.

Qiu, Y., Fu, B.J., Wang, J., Chen, L.D., 2001. Spatial variability of soil moisture contentand its relation to environmental indices in a semi-arid gully catchment of the LoessPlateau, China. J. Arid Environ. 49, 723–750.

Qiu, Y., Fu, B.J., Wang, J., Chen, L.D., 2003. Spatiotemporal prediction of soil moisturecontent using multiple-linear regression in a small catchment of the Loess Plateau,China. Catena 54, 173–195.

Rodriguez-Iturbe, I., 2000. Ecohydrology: a hydrologic perspective of climate-soil-vege-tation dynamics. Water Resour. Res. 36, 3–9.

Sun, W.Y., et al., 2015. Spatiotemporal vegetation cover variations associated with cli-mate change and ecological restoration in the Loess Plateau. Agric. For. Meteorol.209–210, 87–99.

Uchida, E., Xu, J., Rozelle, S., 2005. Grain for green: cost-effectiveness and sustainabilityof China's conservation set-aside program. Land Econ. 81, 247–264.

Van Ek, R., Witte, J.P.M., Runhaar, H., Klijn, F., 2000. Ecological effects of water man-agement in the Netherlands: the model DEMNAT. Ecol. Eng. 16, 127–141.

Vereecken, H., et al., 2014. On the spatio-temporal dynamics of soil moisture at the fieldscale. J. Hydrol. 516.

Wang, Z.Q., Liu, B.Y., Liu, G., Zhang, Y.X., 2009. Soil water depletion depth by plantedvegetation on the Loess Plateau. Sci. China Ser. D Earth Sci. 52, 835–842.

Wang, X.X., Shang, S.Y., Yang, W.H., Clary, C.R., Yang, D.W., 2010. Simulation of landuse-soil interactive effects on water and sediment yields at watershed scale. Ecol. Eng.36, 328–344.

Wang, Y.Q., Shao, M.A., Shao, H.B., 2010. A preliminary investigation of the dynamiccharacteristics of dried soil layers on the Loess Plateau of China. J. Hydrol. 381, 9–17.

Wang, Y.Q., Shao, M.A., Zhu, Y.J., Liu, Z.P., 2011. Impacts of land use and plant char-acteristics on dried soil layers in different climatic regions on the Loess Plateau ofChina. Agric. For. Meteorol. 151, 437–448.

Wang, Y.Q., Shao, M.A., Liu, Z.P., Warrington, D.N., 2012. Regional spatial pattern ofdeep soil water content and its influencing factors. Hydrol. Sci. J. 57, 265–281.

Wang, J.Y., Liu, Y.S., Liu, Z.G., 2013. Spatio-temporal patterns of cropland conversion inresponse to the “Grain for Green Project” in China's Loess Hilly region of YanchuanCounty. Remote Sens. 5, 5642–5661.

Wang, Y.Q., Shao, M.A., Liu, Z.P., 2013. Vertical distribution and influencing factors ofsoil water content within 21-m profile on the Chinese Loess Plateau. Geoderma 193,300–310.

Wang, S., et al., 2015. Reduced sediment transport in the Yellow River due to anthro-pogenic changes. Nat. Geosci. 9, 38–41.

Wang, Y.Q., Shao, M.A., Han, X.W., Liu, Z.P., 2015a. Spatial variability of soil parametersof the van Genuchten Model at a regional scale. CLEAN Soil Air Water 43, 271–278.

Wang, Y.Q., et al., 2015b. Choosing an optimal land-use pattern for restoring eco-en-vironments in a semiarid region of the Chinese Loess Plateau. Ecol. Eng. 74, 213–222.

Wang, Y.Q., Shao, M.N., Liu, Z.P., Zhang, C.C., 2015c. Characteristics of dried soil layersunder apple orchards of different ages and their applications in soil water manage-ments on the Loess Plateau of China. Pedosphere 25, 546–554.

Western, A.W., Grayson, R.B., Blöschl, G., 2002. Scaling of soil moisture: a hydrologicperspective. Annu. Rev. Earth Planet. Sci. 30, 149–180.

Xia, Y.Q., Shao, M.A., 2008. Soil water carrying capacity for vegetation: a hydrologic andbiogeochemical process model solution. Ecol. Model. 214, 112–124.

Yang, L., Wei, W., Chen, L.D., Mo, B.R., 2012. Response of deep soil moisture to land useand afforestation in the semi-arid Loess Plateau, China. J. Hydrol. 475, 111–122.

Zhang, X.C., Liu, W.Z., 2005. Simulating potential response of hydrology, soil erosion,and crop productivity to climate change in Changwu tableland region on the LoessPlateau of China. Agric. For. Meteorol. 131, 127–142.

Zhang, B.Q., Wu, P.T., Zhao, X.N., Wang, Y.B., Gao, X.D., 2013a. Changes in vegetationcondition in areas with different gradients (1980–2010) on the Loess Plateau, China.Environ. Earth Sci. 68, 2427–2438.

Zhang, Z.H., et al., 2013b. Changes in some soil properties induced by re-conversion ofcropland into grassland in the semiarid steppe zone of Inner Mongolia, China. PlantSoil 373, 89–106.

Zhang, F., et al., 2014. Assessment of effects of two runoff control engineering practiceson soil water and plant growth for afforestation in a semi-arid area after 10 years.Ecol. Eng. 64, 430–442.

Zhou, J., et al., 2016. Effects of precipitation and restoration vegetation on soil erosion ina semi-arid environment in the Loess Plateau, China. Catena 137, 1–11.

Zhu, H.D., et al., 2014. Soil moisture response to environmental factors following pre-cipitation events in a small catchment. Catena 120, 73–80.

Y. Zhao et al. Catena 172 (2019) 387–396

396