Driven by many factors, the water supply services (streamflow and groundwater) of many rivers in the dryland area of China have declined significantly. This aggravates the inherent severe water shortages and results in increased severity in the water use conflicts that are threatening sustainable development in the region. Innovative strategies towards more water-efficient land management are vital for enhancing water quantity to ensure water supply security. A key step in the successful development and implementation of such measures is to understand the response of hydrological processes and related services to changes in land management and climate. To this end, it was decided to investigate these processes and responses in the upper reaches of the Jing River (Jinghe), an important meso-scale watershed in the middle reaches of the Yellow River on the Loess Plateau (NW China). It has been shown that vegetation restoration efforts (planting trees and grass) are effective in controlling soil erosion on the Loess Plateau. Shifts in land cover/use lead to modifications of soil physical properties. Yet, it remains unclear if the hydraulic properties have also been improved by vegetation restoration. A better understanding of how vegetation restoration alters soil structure and related soil hydraulic properties, such as water conductivity and soil water storage capacity, is necessary. Three adjacent sites, with comparable soil texture, soil type, and topography but contrasting land cover (Black locust forest, grassland, and cropland), were investigated in a small catchment in the upstream Jinghe watershed (near Jingchuan, Gansu province). Seasonal variations of soil hydraulic properties in topsoil and subsoil were examined. Results revealed that the type of land use had a significant impact on field-saturated, near-saturated hydraulic conductivity, and soil water characteristics. Specifically, conversion from cropland to grass or forests promotes infiltration capacity as a result of increased saturated hydraulic conductivity, air capacity, and macroporosity. Moreover, conversion from cropland to forest tends to promote the formation of mesopores that increase soil water storage capacity. Tillage in cropland temporarily created well-structured topsoil, but also compacted subsoil, as indicated by low subsoil saturated hydraulic conductivity, air capacity, and plant available water capacity. An impact of land cover conversion on unsaturated hydraulic conductivities was not identified, indicating that changes in land cover do not affect functional meso- and microporosity. Changes in soil hydraulic properties and associated hydrological processes and services due to soil conservation efforts need to be considered, should soil conservation measures be implemented in water-limited regions for sustaining adequate water supply. To differentiate between the impacts of land management and climate change on streamflow, the variation of annual streamflow, precipitation, potential evapotranspiration, and climatic water balance in a small catchment of the upstream Jinghe watershed (near Pingliang, Gansu province) was examined during the period of 1955 – 2004. During this time the relative contributions of changes in land management and climate to the reduction of streamflow were estimated. A statistically significant decreasing trend of -1.14 mm y-1 in annual streamflow was detected. Furthermore, an abrupt streamflow reduction due to afforestation and construction of terraces and check-dams was identified around 1980. Remarkably, 74% of the total reduction in mean annual streamflow can be attributed to the soil conservation measures. Among various conservation measures, streamflow could be considerably reduced by afforestation and terracing (including damland creation), due to their low contribution to water yield. In contrast, slope farmland and grassland can maintain a certain level of water supply services due to higher runoff coefficients. According to a meta-analysis of the published studies on the Loess Plateau, the impact of changes in land management on annual streamflow appears to diminish with increasing catchment size while the impact of climate change appears uniform across space. This means that there is a dependency between the catchment size and the response of hydrological processes to environmental change. At least at the local scale, it appears that well-considered land management may help to ensure the water supply services. Due to limited surface water availability, groundwater is an essential water source for supporting ecosystem and socio-economic development in the dryland region. However, the groundwater process is susceptible and vulnerable to changes in climate and landscape (i.e., land cover and form) that in turn can result in profound adverse consequences on water supply services in water-limited regions. In addition, an improved understanding of the response of groundwater related processes to natural and artificial disturbances is likely to ensure more secure and more sustainable governance and management of such regions, as well as better options for adapting to climate change. Yet, this topic has seldom been researched, especially in areas that have already experienced large-scale alteration in landscape and are located in dryland regions, such as the Loess Plateau. Therefore, an investigation of the baseflow variation along the landscape change was conducted. The average annual baseflow has significantly decreased at catchment scale during the period of 1962 – 2002 without any obvious significant change in climate. At decadal scale, the reduction accounts for approximately 9% in the 1970s, 48% in the 1980s, and 92% in the 1990s, while the baseflow index declines averaging 5%, 16% and 67%, respectively. All of the monthly baseflow levels dropped at varying rates except in January, among which July was the most severe in terms of both magnitude (-4.17) and slope (-0.09 mm y-1). In perspective of landscape change, landform change (terrace and check-dam) tends to reduce baseflow by reallocation of surface fluxes and retention for crop growth causing limited deep drainage in other areas. Land cover change (i.e., afforestation) reduced the baseflow to a larger extent by enhanced evapotranspiration and thus hampered deep drainage as suggested by the soil moisture measurement underneath. The study indicates that knowledge about baseflow formation on catchment scale needs further improvement. Integrated soil conservation and water management for optimizing landscape structure and function in order to balance soil (erosion) and water (supply) related hydrological ecosystem services is vital. The governing processes to the changes of water-supply-services-related hydrological process (e.g., streamflow) are assumed to be different across space. To this end, the factors controlling streamflow were investigated on both a small and large scale. Streamflow in small catchments was found to be mainly controlled by precipitation and land cover type. On a larger scale, evaporative demand was found to be another additional major driving force. Hydrological modeling is a frequently used tool for the assessment of impacts of land use and climate change on water balance and water fluxes. However, application of the Soil and Water Assessment Tool (SWAT) model in the upstream Jinghe watershed was unsuccessful due to difficulties in calibration. The inability of the SWAT model to take the influence of terraces on steep slopes into consideration and the method how to calculate lateral flow were the main reasons for unsatisfactory calibration, at least for the current version of SWAT used in this study. Alternatively, Budyko’s frameworks were applied to predict the annual and long-term streamflow. However, the effect of changes in land management (e.g., afforestation) on streamflow could not be assessed due to a lack of vegetation factors. Therefore, an empirical analysis tool was derived based on an existing relationship for estimation. This method was found to be the most effective in reproducing the annual and long-term streamflow. The incorporation of temporal changes in land cover and form in the approach enables the estimation of the possible impact of soil conservation measures (e.g., afforestation or terracing). The importance of adaptive land management strategies for mitigating water shortage and securing the water supply services on the Loess Plateau was highlighted. A cross-sectoral view of the multiple services offered by managed ecosystems at different spatial scales under changing environments needs to be integrated to improve adaptive land management policy. In a water limited environment, such as the Loess Plateau, multiple ecosystem services including hydrological services need to be balanced with minimum trade-offs. This can only be achieved when management is based on a holistic understanding of the interdependencies among various ecosystem services and how they might change under alternative land management.

Driven by many factors, the water supply services (streamflow and groundwater) of many rivers in the dryland area of China have declined significantly. This aggravates the inherent severe water shortages and results in increased severity in the water use conflicts that are threatening sustainable development in the region. Innovative strategies towards more water-efficient land management are vital for enhancing water quantity to ensure water supply security. A key step in the successful development and implementation of such measures is to understand the response of hydrological processes and related services to changes in land management and climate. To this end, it was decided to investigate these processes and responses in the upper reaches of the Jing River (Jinghe), an important meso-scale watershed in the middle reaches of the Yellow River on the Loess Plateau (NW China). It has been shown that vegetation restoration efforts (planting trees and grass) are effective in controlling soil erosion on the Loess Plateau. Shifts in land cover/use lead to modifications of soil physical properties. Yet, it remains unclear if the hydraulic properties have also been improved by vegetation restoration. A better understanding of how vegetation restoration alters soil structure and related soil hydraulic properties, such as water conductivity and soil water storage capacity, is necessary. Three adjacent sites, with comparable soil texture, soil type, and topography but contrasting land cover (Black locust forest, grassland, and cropland), were investigated in a small catchment in the upstream Jinghe watershed (near Jingchuan, Gansu province). Seasonal variations of soil hydraulic properties in topsoil and subsoil were examined. Results revealed that the type of land use had a significant impact on field-saturated, near-saturated hydraulic conductivity, and soil water characteristics. Specifically, conversion from cropland to grass or forests promotes infiltration capacity as a result of increased saturated hydraulic conductivity, air capacity, and macroporosity. Moreover, conversion from cropland to forest tends to promote the formation of mesopores that increase soil water storage capacity. Tillage in cropland temporarily created well-structured topsoil, but also compacted subsoil, as indicated by low subsoil saturated hydraulic conductivity, air capacity, and plant available water capacity. An impact of land cover conversion on unsaturated hydraulic conductivities was not identified, indicating that changes in land cover do not affect functional meso- and microporosity. Changes in soil hydraulic properties and associated hydrological processes and services due to soil conservation efforts need to be considered, should soil conservation measures be implemented in water-limited regions for sustaining adequate water supply. To differentiate between the impacts of land management and climate change on streamflow, the variation of annual streamflow, precipitation, potential evapotranspiration, and climatic water balance in a small catchment of the upstream Jinghe watershed (near Pingliang, Gansu province) was examined during the period of 1955 – 2004. During this time the relative contributions of changes in land management and climate to the reduction of streamflow were estimated. A statistically significant decreasing trend of -1.14 mm y-1 in annual streamflow was detected. Furthermore, an abrupt streamflow reduction due to afforestation and construction of terraces and check-dams was identified around 1980. Remarkably, 74% of the total reduction in mean annual streamflow can be attributed to the soil conservation measures. Among various conservation measures, streamflow could be considerably reduced by afforestation and terracing (including damland creation), due to their low contribution to water yield. In contrast, slope farmland and grassland can maintain a certain level of water supply services due to higher runoff coefficients. According to a meta-analysis of the published studies on the Loess Plateau, the impact of changes in land management on annual streamflow appears to diminish with increasing catchment size while the impact of climate change appears uniform across space. This means that there is a dependency between the catchment size and the response of hydrological processes to environmental change. At least at the local scale, it appears that well-considered land management may help to ensure the water supply services. Due to limited surface water availability, groundwater is an essential water source for supporting ecosystem and socio-economic development in the dryland region. However, the groundwater process is susceptible and vulnerable to changes in climate and landscape (i.e., land cover and form) that in turn can result in profound adverse consequences on water supply services in water-limited regions. In addition, an improved understanding of the response of groundwater related processes to natural and artificial disturbances is likely to ensure more secure and more sustainable governance and management of such regions, as well as better options for adapting to climate change. Yet, this topic has seldom been researched, especially in areas that have already experienced large-scale alteration in landscape and are located in dryland regions, such as the Loess Plateau. Therefore, an investigation of the baseflow variation along the landscape change was conducted. The average annual baseflow has significantly decreased at catchment scale during the period of 1962 – 2002 without any obvious significant change in climate. At decadal scale, the reduction accounts for approximately 9% in the 1970s, 48% in the 1980s, and 92% in the 1990s, while the baseflow index declines averaging 5%, 16% and 67%, respectively. All of the monthly baseflow levels dropped at varying rates except in January, among which July was the most severe in terms of both magnitude (-4.17) and slope (-0.09 mm y-1). In perspective of landscape change, landform change (terrace and check-dam) tends to reduce baseflow by reallocation of surface fluxes and retention for crop growth causing limited deep drainage in other areas. Land cover change (i.e., afforestation) reduced the baseflow to a larger extent by enhanced evapotranspiration and thus hampered deep drainage as suggested by the soil moisture measurement underneath. The study indicates that knowledge about baseflow formation on catchment scale needs further improvement. Integrated soil conservation and water management for optimizing landscape structure and function in order to balance soil (erosion) and water (supply) related hydrological ecosystem services is vital. The governing processes to the changes of water-supply-services-related hydrological process (e.g., streamflow) are assumed to be different across space. To this end, the factors controlling streamflow were investigated on both a small and large scale. Streamflow in small catchments was found to be mainly controlled by precipitation and land cover type. On a larger scale, evaporative demand was found to be another additional major driving force. Hydrological modeling is a frequently used tool for the assessment of impacts of land use and climate change on water balance and water fluxes. However, application of the Soil and Water Assessment Tool (SWAT) model in the upstream Jinghe watershed was unsuccessful due to difficulties in calibration. The inability of the SWAT model to take the influence of terraces on steep slopes into consideration and the method how to calculate lateral flow were the main reasons for unsatisfactory calibration, at least for the current version of SWAT used in this study. Alternatively, Budyko’s frameworks were applied to predict the annual and long-term streamflow. However, the effect of changes in land management (e.g., afforestation) on streamflow could not be assessed due to a lack of vegetation factors. Therefore, an empirical analysis tool was derived based on an existing relationship for estimation. This method was found to be the most effective in reproducing the annual and long-term streamflow. The incorporation of temporal changes in land cover and form in the approach enables the estimation of the possible impact of soil conservation measures (e.g., afforestation or terracing). The importance of adaptive land management strategies for mitigating water shortage and securing the water supply services on the Loess Plateau was highlighted. A cross-sectoral view of the multiple services offered by managed ecosystems at different spatial scales under changing environments needs to be integrated to improve adaptive land management policy. In a water limited environment, such as the Loess Plateau, multiple ecosystem services including hydrological services need to be balanced with minimum trade-offs. This can only be achieved when management is based on a holistic understanding of the interdependencies among various ecosystem services and how they might change under alternative land management.