*4.1. Necessity to Assessing WESV*

The significance of evaluating the value of ecosystem services is to better manage the ecological environment and natural resources to achieve sustainable development. The water issue is prominent today, so it is more practical to discuss the value of water resources ecosystem services. In China, the value of ecosystem services is always in the form of the upper limit of the ecological protection compensation standard. The water ecosystem services value is often calculated as part of a comprehensive ecosystem service value and is rarely discussed in isolation. Even in the basin ecological compensation, only the fluctuation of the direct use value caused by the change of water quantity is calculated, and how the ecosystem service value of water resources including groundwater changes is not fully explored. In 2019, the ecological protection and high-quality development of the Yellow River Basin was established as one of China's major national strategies. Ecological compensation is one of the key tasks of the strategy. The Loess Plateau, located in the middle reaches of the Yellow River, is the main source of sediment in the Yellow River. The region has scarce water resources, a large population, and an urgent need for development. Hence, from the perspective of the integrity of water resources, this study selects typical regions to assess WESV and analyzes the temporal and spatial variation characteristics of WESV, which can provide a basis for the ecological compensation development.

#### *4.2. The Spatiotemporal Distribution of WESV Response to Different Influencing Factors*

Previous studies have shown that the value of water resources is affected by multiple factors, such as water quantity, water quality, use of water resources, economic development, and educational level of residents [59,60].

With the deepening of relevant research, the water resources value has been extended to the value of ecosystem services that water resources can provide for human well-being. Correspondingly, WSVE is also affected by economic, social, natural and cultural factors. Many researches have explored the mechanism by which ESV is affected by various driving factors, including land use change, socioeconomic development indicators, and human acceptance willingness. Furthermore, the distribution characteristics of the sensitivity of ESV to various influencing factors at different temporal and spatial scales were obtained. WESV is an important part of ESV and is also affected by various factors. As mentioned above, the extent to which WESV is affected by external factors should be studied separately.

With the rapid economic development and the sharp expansion of cities, the differences between different regions in natural resources, economy, society and culture are gradually increasing. In this study, the central part of the Loess Plateau was selected as the research area, and four main influencing factors were selected. The GWR model was used to analyze the influence of per capita GDP, population density, the proportion of water areas, and water consumption on WESV in different regions.

Per Capita GDP is an important parameter to measure the development degree of a region. The rise in prices and the massive consumption of water resources brought about by economic development will lead to an increase in WESV. Meanwhile, WESV will greatly decreased because of the ecological degradation caused by development, that is, the reduction in water area. Therefore, the relationship between water conservation and economic development should be balanced. The effect of per capita GDP on ESV was confirmed by Song F. [61] in a study on the value of wetland ecosystem services. Dai X. concluded that per capita GDP was negatively correlated with ESV in Chengdu [62]. As can be seen from Figure 4, per capita GDP showed time instability. The study area has undergone a process of environmental damage and recovery during a 10-year development period, which is consistent with the process in China. During 2015–2020, the construction of ecological civilization has been raised to a new level. At last, a pattern in which the degree of economic development roughly matches WESV is formed. Many studies have found that per capita GDP has an unstable impact on ESV, by influencing the level of awareness, willingness to protect the environment, and investment in environmental protection [63–65].

It is generally believed that the increase in population density is accompanied by the expansion of human activity areas, which will encroach on other land-use types. The water area is also experiencing depletion of rivers and lakes due to the massive depletion of surface water while the water area was occupied. Under the dual effect of the two, WESV is bound to decrease, which also explains the negative correlation between population density and WESV. It was confirmed in Chen Y.'s study of the relationship between population density and agro-ecosystem services [66]. Jiang Z. also obtained a consistent conclusion in his study of the South Four Lakes that population density was negatively correlated with Esv [67]. In recent years, the population growth rate in the study area has slowed down as in the whole country. WESV is less sensitive to population density.

In the same region, WESV will increase with the increase in water area, which can also be explained by the calculation method adopted in this study. The equivalent factor method is to calculate the corresponding ESV according to the type of land use. Figure 6 showed that the proportion of water area has the greatest influence on WESV. During the initial stage of the study, the drier regions in the north were more sensitive. With the comprehensive management of the Mu Us Desert, the environment in the northern region has been gradually improved, and the scarcity of water resources and the vulnerability of the water environment have been alleviated. The center of influence then transferred to the west, where the pace of governance was relatively slow.

In 2010 and 2015, the impact of water consumption on WESV had a certain regularity in space, but the spatial instability that occurs cannot be ignored. Figure 7a,b can well demonstrate this phenomenon. From 2015 to 2020, the state strictly controlled the regional water consumption, and the strictest water resources management system was successfully implemented locally [68], which well explained the phenomenon that the spatial distribution pattern in 2020 was more regular.

There is spatial heterogeneity in the sensitivity of WESV to explanatory variables. Using the GWR model can more clearly show the spatial pattern of the influence degree and its evolution trend over time.

#### *4.3. Limitations*

Based on the water area obtained from remote sensing image statistics and the water consumption obtained from statistical data, this study proposed a method to quickly calculate WESV and analyze the temporal and spatial distribution of driving factors. However, the impact of water quality on WESV does not depend on water consumption or water area. Deterioration of water quality not only reduces the quality of services provided by the water resource but also incurs additional remediation costs [69]. This study did not include water quality as the basic data in the WESV assessment for three reasons. First, different types of WESVs have different sensitivity to water quality, and its mechanism of action and degree of influence are not clear, which will cause uncertainty in the calculation of WESV. Second, the equivalent factor method used in this paper is calculated from the ecological service value and the economic value of grain, and the output of grain is affected by water quality. That is to say, this study considered the impact of water quality on WESV to a certain extent. Third, the spatial and temporal scales of this study determined that water quality will not have an essential impact on WESV. In space, the county-level administrative region is the smallest research unit; in terms of time, a year is the minimum span of time. From this point of view, the water quality in the region is relatively stable, and this is also verified by the water quality data released by the environmental department. Therefore, the WESV calculated in this study is still representative and reliable without considering water quality.
