Lake Qinghai, the largest closed-basin lake on the Qinghai–Tibet Plateau, plays a crucial role in maintaining regional ecological stability through its hydrological functions. In recent decades, the lake has exhibited a continuous rise in water level and lake area expansion, sparking growing interest in the mechanisms driving these changes and their future evolution. This study integrates the Soil and Water Assessment Tool (SWAT), simulations under future Shared Socioeconomic Pathways (SSPs) and statistical analysis methods, to assess runoff dynamics and lake level responses in the Lake Qinghai Basin over the next 30 years. The model was developed using a combination of meteorological, hydrological, topographic, land use, soil, and socio-economic datasets, and was calibrated with the sequential uncertainty fitting Ver-2 (SUFI-2) algorithm within the SWAT calibration and uncertainty procedure (SWAT–CUP) platform. Sensitivity and uncertainty analyses confirmed robust model performance, with monthly R² values of 0.78 and 0.79. Correlation analysis revealed that runoff variability is more closely associated with precipitation than temperature in the basin. Under SSP 1-2.6, SSP 3-7.0, and SSP 5-8.5 scenarios, projected annual precipitation increases by 14.4%, 18.9%, and 11.1%, respectively, accompanied by temperature rises varying with emissions scenario. Model simulations indicate a significant increase in runoff in the Buha River Basin, peaking around 2047. These findings provide scientific insight into the hydrological response of plateau lakes to future climate change and offer a valuable reference for regional water resource management and ecological conservation strategies.

Water conservation is among the most important ecosystem service functions, and its dynamic changes can reflect the health of regional ecosystems. As one of the ten major tributaries of the Yellow River, the Yi–Luo River in Henan Province plays a significant role in maintaining regional ecological health through the ecosystem water conservation functions of its basin. To explore the characteristics of these functions, this study employed ecosystem transfer matrix and water balance methods to analyze the changes in ecosystem types in the basin and the dynamic changes in the corresponding water conservation functions from 2010 to 2020. The results showed that from 2010 to 2020, the dominant ecosystem types were farmland, forestland, and shrubland, and the main transformations between different ecosystem types were from farmland to urban areas and wetland, from grassland to shrubland, and from shrubland to forestland. The water conservation functions of the different ecosystem types varied significantly, with the average water conservation capacities ranked as follows: forestland > wetland > shrubland > grassland > farmland > others > urban areas. Due to the decreased water conservation depths in the forestland and shrubland ecosystems, the overall water conservation depth decreased over the study period. In addition, the water conservation capacities of the farmland, forestland, and shrubland ecosystems were the highest, reaching 12.94 × 10⁸, 12.15 × 10⁸ and 9.73 × 10⁸ tons, respectively, accounting for 34.47%, 32.35%, and 25.91% of the total water conservation capacity, respectively. Overall, water conservation in the study area was high in the south and low in the north, and there was an overall decreasing trend over time. Moreover, the regions with critical and important water conservation functions accounted for the largest proportion (47.2%). The critical and important regions were mainly located in the higher altitudes of Luanchuan County and Lushi county, while the moderate and average regions were primarily located in the plains, and weak regions were distributed mainly in the urban area of Luoyang city. Through analysis of the spatiotemporal changes and degree of importance of water conservation functions in the Yi–Luo River Basin, the key regions for ecological protection and construction were identified. The study results can be used to help enhance ecological security in the Yellow River Basin of Henan Province and play a significant role in promoting the regional economy as well as the health and sustainable development of the ecological environment.

Qinghai Lake, the largest endorheic saline lake in China, has undergone a pronounced hydrological regime shift from a multi-decadal decline to a rapid post-2004 recovery, reflecting strong hydroclimatic non-stationarity in the northeastern Tibetan Plateau (TP). This paper supplements the current water level and lake area status of Qinghai Lake to provide basic background for future prediction. Reliable forecasting of such climate sensitive lake systems remains difficult because conventional statistical models often fail to capture non-linear fluctuations, whereas standalone deep learning models may overlook long-term deterministic evolution. To address this challenge, we developed a serial decomposition GeoAI framework that integrates autoregressive integrated moving average (ARIMA), one-dimensional convolutional neural networks (1D-CNNs), and long short-term memory (LSTM) networks for non-stationary water level forecasting. Using annual water level observations from 1960 to 2025, the ARIMA component was first used to extract the low-frequency deterministic trend, after which the CNN-LSTM module reconstructed the nonlinear residual variability. The model was trained on the 1960–2012 period and validated over 2013–2025, which represents the most dynamic expansion stage of Qinghai Lake. The hybrid framework outperformed the benchmark models, achieving a Root Mean Square Error (RMSE) of 0.2033 m, Mean Absolute Error (MAE) of 0.1727 m, and Mean Squared Error (MSE) of 0.0413 m² during validation. The decomposition strategy effectively reduced phase lag and amplitude attenuation, improving both predictive accuracy and process interpretability. Multi-step forecasting for 2026–2056 suggests that Qinghai Lake will continue to rise, reaching approximately 3204.08 m by 2056, although the growth rate is projected to slow as negative hydrological feedback strengthen. By explicitly separating deterministic climate scale signals from nonlinear short-term variability, the proposed framework provides a robust and transferable geoinformation based tool for forecasting water level dynamics and supporting adaptive management in climate sensitive, data scarce lake basins.