Land use/land cover (LULC) change has greatly altered ecosystem carbon storage capacity and may eventually profoundly impact global climate change. Characterizing the LULC change and its impact on wetland ecosystem carbon stock provides useful data and insights that can guide decision-making procedures aimed at achieving sustainable development objectives. The Yellow River Delta (YRD) represents the most intact coastal wetland and is considered to be the most recent wetland ecosystem in China. It exhibits significant carbon stock capacity and ecological value. Based on the LULC data of the YRD in 2002, 2007, 2012, 2017, and 2022, this paper quantitatively evaluates the spatiotemporal changes in LULC and carbon stock in the region and analyzes the response characteristics of carbon stock to LULC change. The results show significant reductions in cropland and tidal flat wetland areas from 2002 to 2022, resulting in a decrease of 1,428,735.77 t and an increase of 139,856.58 t in carbon stock, respectively. The built-up land area expanded considerably, and carbon stock was lost by 1,467,915.82 t. Spatially, the carbon stock exhibited a pattern of “low along the coast, high inland; low in the center, high around the periphery”. In addition, protecting cropland, reducing building, facilitating the conversion of reservoirs and ponds to forest, and transforming tidal flat wetlands into reservoirs and ponds can increase the region’s carbon storage capacity. These findings provide valuable insights for regional carbon management strategies and ecological protection policies, supporting the sustainable development goals of the Yellow River Delta.

Ecological security is of central significance for maintaining the integrity and stability of regional ecosystems. Nevertheless, current academic research systematically studying the complex driving mechanisms behind ecological security is deficient. Therefore, this study utilised ecological sources (ESs) and ecological resistance surfaces (ERSs) to calculate and obtain the Ecological Suitability Index (ESI), thereby reflecting the ecological security status. The drivers of ESI in different areas of the Yellow River Basin (YRB) from 1990 to 2020 were explored using the random forest (RF) model and GeoDetector. The results of the study show that: (1) according to the results of the RF model, LANDUSE, NDVI, and PRE are the most important driving factors of the ESI in the upper (UYR), middle (MYR), and lower (LYR) reaches of the YRB; (2) according to the results of GeoDetector factor detection, in the UYR and MYR, the ESI is primarily influenced by natural factors such as NDVI and PRE, showing particularly significant impacts in 2010 and 2020; (3) in the LYR, the ESI is mainly influenced by NDVI, but the degree of influence from socioeconomic factors has significantly strengthened. This study provides a decision-making direction for ecological protection and coordinated development in the YRB.

Understanding the interaction between ecological carrying capacity and high-quality development is essential for reconciling ecological protection with socioeconomic development in the Yellow River Basin. This study, spanning the period from 2000 to 2023, constructed an ecological carrying capacity index across four dimensions, including water resources, energy, food, and ecological regulation and carbon source-sink balance, and a high-quality development index across five dimensions, including ecology, water resources, economy, culture, and livelihood. TOPSIS, the coupling coordination degree model, exploratory spatial data analysis, the standard deviational ellipse, and multiscale geographically weighted regression were used to characterize spatiotemporal patterns and spatially varying associations. The results showed that: (1) Ecological carrying capacity declined slightly from 0.297 in 2000 to 0.292 in 2005 and then rose to 0.307 in 2023, whereas high-quality development increased from 0.199 in 2000 to 0.267 in 2015 and subsequently stabilized. (2) The coupling coordination degree increased from 0.476 in 2000 to 0.516 in 2023 and exhibited significant positive spatial autocorrelation (global Moran’s I = 0.327–0.440, p < 0.01). (3) MGWR outperformed GWR (R² = 0.641–0.785 vs. 0.482–0.624), indicating spatial nonstationarity. The ecosystem quality index showed the most stable positive association with coupling coordination degree, the association with patent grants strengthened markedly by 2023, the positive association with GDP per capita diminished, and population density remained negatively associated. This study proposes targeted, zone-specific regulatory recommendations, providing a decision-making basis for scientifically differentiated regulation in the Yellow River Basin.

Downward shortwave radiation (DSR) varies strongly across mountain slopes, and products designed for horizontal surfaces may not resolve this variability. We estimated slope-parallel DSR at eight Saihanba sites using Sentinel-2 imagery, topography, atmospheric inputs, and BRDF-informed surface albedo. Both branches used XGBoost, with the clear-sky branch further incorporating a parameterized radiation anchor, bounded residual learning, and support-aware output correction. On the 2024 test set (n = 1451), the full framework achieved an RMSE of 92.60 W/m², an NRMSE of 15.79% and an R² of 0.897. Pooled leave-one-site-out validation across all 6096 samples yielded an RMSE of 113.68 W/m². For horizontal-surface observations at 10 Heihe River Basin stations, NRMSE was 17.60%. Clear-sky estimates depended most on solar-incidence geometry, whereas top-of-atmosphere spectra dominated the cloudy-sky XGBoost model.