The water-yield function in the Yellow River Basin (YRB) of China for maintaining the basin’s ecological water balance plays a crucial role. Understanding its spatiotemporal variation and the underlying drivers in the basin is crucial for the management, utilization, and development of water resources. Thus, we used the InVEST model to explore its spatiotemporal dynamics across multiple scales (“basin–county–pixel”). Then, we integrated socio-economic and natural factors to elucidate the driving forces and spatial heterogeneity of water-yield dynamics. Our findings indicated that water-yield trends increased in 71.76% of the YRB, and significant water-yield increases were detected in 13.9% of the basin over the past 40 years. A phase-wise comparison revealed a shift in water yield from a decreasing trend in the first two decades to a significant increasing trend in the last two decades. Hotspot analysis revealed that hotspots of increasing water-yield trends have shifted from the downstream section of the basin toward the southwest, while hotspots of decreasing water-yield trends first concentrated in the basin’s southern part and then disappeared. Both natural and socioeconomic factors have exerted positive and negative impacts on water-yield dynamics. Among them, the dynamics of water yield have been predominantly driven by natural variables.

Benefiting from information complementarity, the fusion of hyperspectral imagery (HSI) and light detection and ranging (LiDAR) data for land cover classification has attracted significant attention in the remote sensing community. However, due to modality imbalance and information redundancy, effectively extracting and integrating complementary knowledge from HSI and LiDAR data remains a major challenge. To address the aforementioned issues, a class semantic prototype guided fusion network (CSPGFNet) is proposed to realize efficient and accurate classification by task-relevant feature mining, fusion, and interaction. Specifically, feature extraction sub-networks with multi-scale and multi-type convolutional structures are designed for each modality to mitigate semantic imbalance caused by inherent dimensional discrepancies. Moreover, a feature fusion-interaction module based on cross-attention mechanism is designed to fuse and interact task-relevant spatial–spectral and elevation information from modalities with class semantic prototype (CSP) as the bridge. Furthermore, a composite loss that incorporates multi-factor constraints is optimized to ensure information complementarity, semantic consistency and task relevance of the whole network. Experimental evaluations on three benchmark datasets demonstrate the effectiveness of the proposed method.