Phreatic evaporation can constitute a critical vertical linkage in the vertical interaction between surface water and groundwater systems. Accurate simulation and quantification of phreatic evaporation are of great significance for the assessment and sustainable management of shallow groundwater resources in the Huaibei Plain. This study aims to simulate and project the phreatic evaporation using the CMIP6 multi-model ensemble. The measurements were performed to extract the data from the Wudaogou Hydrological Experimental Station. The commonly used formulas were applied to evaluate the regional features of phreatic evaporation. The Ye Shuiting formula was further optimized after evaluation. An advanced iterative algorithm was employed to integrate the spatial distribution of the groundwater depth data over the plain. The historical variation and dynamics of phreatic evaporation were simulated, particularly focus on the lime concretion black soil and yellow fluvo-aquic soil areas. Future projections of precipitation and evaporation from five CMIP6 climate models were used as the primary climatic forcing factors. Multi-model ensemble approaches were also employed to integrate these climate projections using long short-term memory (LSTM). As such, the phreatic evaporation iterative algorithm was combined with the optimal Ye Shuiting formula. Future trends of phreatic evaporation were predicted under the SSP1-2.6, SSP2-4.5, and SSP5-8.5 scenarios. The results show that: 1) The Ye Shuiting formula shared the high applicability to both lime concretion black soil and yellow fluvo-aquic soil in the study area. The optimal Ye Shuiting formula outperformed the rest empirical ones. The annual average phreatic evaporation was ranked in the descending order of: yellow fluvo-aquic soil area (254.5 mm)> areal average of Huaibei Plain (179.3 mm)> lime concretion black soil area (108.5 mm). Phreatic evaporation in all subregions also exhibited an increasing trend during the historical period. 2) The LSTM multi-model ensemble demonstrated strong performance in reproducing the variations in precipitation and evaporation during the baseline period. Projections indicated that the future precipitation and evaporation exceeded the historical levels under the three emission scenarios, with the ranking order of SSP5-8.5, SSP1-2.6, SSP2-4.5. The phreatic evaporation iterative algorithm showed that the amplitude of groundwater depth fluctuations was smaller under the three future emission scenarios than that during the historical period. Specifically, the groundwater depths under the SSP1-2.6 and SSP2-4.5 scenarios were lower than those during the historical period, whereas those under SSP5-8.5 exhibited a significantly larger relative to the historical period. 3) The magnitudes of phreatic evaporation variations under the three future emission scenarios exceeded those in the historical period, indicating an overall increasing trend. The amplification amplitudes and change rates were ranked in descending order of the SSP1-2.6, SSP2-4.5, and SSP5-8.5. Notably, the change rate and uncertainty in winter were significantly higher than those in spring, summer, and autumn. The high accuracy of phreatic evaporation can greatly contribute to the data reference for the simulation and prediction of the phreatic evaporation and water cycle components in the Huaibei Plain.

By using simulation data of precipitation and temperature in the Huaihe River Basin from 1985 to 2014 from 22 global climate models (GCMs) of the coupled model intercomparison project in phase 6 (CMIP6), the simulation performance of GCMs was comprehensively assessed by means of a rank score method with seven indexes selected, and the sensitivity of the scores to each index was analyzed. The spatial simulation performance of the selected optimal GCMs was evaluated. The results reveal significant differences in the simulation performance of monthly average precipitation and temperature in the Huaihe River Basin among the models. Overall, the GCMs perform better in simulating temperatures, but there is a general overestimation of monthly average precipitation. The top five models with the best comprehensive rank score are EC-Earth3 (7.83), EC-Earth3-Veg (7.66), ACCESS-CM2 (7.62), TaiESM1 (7.27), and FGOALS-f3-L (7.20). The rank score results for precipitation are most sensitive to the standard deviation, the statistics of Mann-Kendall trend analysis (z), and the slope of Mann-Kendall (β), while those for temperature show high sensitivity to z and β. Different combinations of rank score indexes moderately influence the scores. The EC-Earth3 model, identified as the most optimal, accurately reproduces the spatial distribution of precipitation but performs slightly less effectively for the spatial distribution of temperature.

Using six multimodal integration methods, including arithmetic averaging, weighted averaging, multiple linear regression, BP neural network, long-short-term memory (LSTM) neural network, and random forest (RF), this study integrated five global climate models (GCMs) data in CMIP6, and based on historical precipitation and temperature data of the Yellow River Basin water conservation region, the simulation performance of different integration methods were evaluated. The multimodal integration method with the best performance was selected to predict future precipitation and temperature under three scenarios (SSP1-2.6, SSP2-4.5, and SSP5-8.5). The results show that the multimodal integration could well reproduce the variations of historical precipitation and temperature, and the LSTM neural network method has the best performance. In three scenarios, future average annual precipitation all increases, but the change of seasonal precipitation in the future varies. Under the SSP1-2.6 scenario, the annual precipitation peaks occur at the beginning of each period, while annual precipitation increases in the near term and decreases obviously in the long term under the SSP2-4.5 and SSP5-8.5 scenarios. Future temperature in three scenarios shows upward trends of different degrees, and the amplitude and rate of temperature increase from small to large are: SSP1-2.6, SSP2-4.5, SSP5-8.5. Future temperature increases greatest in autumn and least in winter. There are large uncertainties in the future precipitation and temperature predicted by multimodal integration methods, and the uncertainty in the medium to long term is greater than in the short term. The uncertainty in future precipitation projection is relatively greater than that of temperature, and the uncertainty in autumn and winter is significantly greater than that in spring and summer.

To explore the spatiotemporal evolution patterns of ecosystem services in the water conservation area of the Yellow River, the InVEST model was employed to calculate five ecosystem service factors from 2000 to 2020, including carbon storage, soil retention, habitat quality, water yield capacity, and water conservation. The spatiotemporal evolution characteristics of these factors and their trade-offs and synergies were analyzed. The results indicate that from 2000 to 2020, the water yield capacity, water conservation, and soil retention in the Yellow River water conservation area exhibited consistent fluctuations, while habitat quality and carbon storage showed an overall declining trend. There were varying degrees of synergistic correlations between the ecosystem service factors, with the strongest synergies observed between soil retention and water yield capacity, as well as between soil retention and carbon storage. Other factors exhibited moderate or weak synergistic relationships. The five ecosystem service factors were closely related to land use types, and there are significant differences in each factor under different land use types. Synergies dominated between the various ecosystem service factors, with the overall synergistic relationship showing a slight fluctuating upward trend over time. In terms of spatial distribution, the grasslands in the central and southern regions were dominated by high-high synergies, while low-low synergies were primarily found in the grasslands and forests of the northern region. High-low and low-high trade-offs were mainly scattered in cultivated lands and construction lands in the eastern and western regions.