Linking landscape patterns to rainfall-runoff-sediment relationships: A case study in an agriculture, forest, and urbanization-dominated mountain watershed

• Effects of LP on soil erosion diminished when STM exceeded a threshold. • The impact of LP on RRSR was higher than that on the STM. • Enhance forest-cultivated land boundaries and reduce constructed land fragmentation could mitigate soil transfer. • The variations of soil erosion could be predicted by several major LP metrics. Human-induced alterations in Landscape Patterns (LP) significantly influence hydrological processes, yet their comprehensive impact on rainfall-runoff-sediment relationships (RRSR) remains unexplored. This study introduces the Sediment Transport Modulus per unit of erosive Rainfall (STMR) as a novel metric, alongside the Sediment Transport Modulus (STM) and Suspended Sediment Concentration (SSC), to characterize RRSR. Using hydrometeorological data (1963–2022) and land use maps (1990–2022), the Hushan River Watershed (HSRW) in eastern Poyang Lake, China, was selected to examine the effects of LP on RRSR through LP metrics calculation, correlation analysis, redundancy analysis, and multiple linear regression. The results reveal significant increases in STM, STMR, and SSC, accompanied by notable LP changes during the study period. Overall, LP had a stronger influence on RRSR than sediment yield alone. While the LP of cultivated and forest land showed limited effects on RRSR compared to STM, constructed land exhibited a more pronounced impact. Notably, LP influence diminished when sediment yields exceeded a threshold. Key contributors to RRSR variability included Shannon’s Diversity Index, the percentage of cultivated and forest land, and patch density of cultivated and constructed land. It can be concluded that LP effectively predicted RRSR variations. Strategic LP management, such as enhancing the boundaries between forest and cultivated land, reducing the fragmentation of construction land, and strengthening soil erosion control measures along riverbanks can reduce the transfer of soil during the runoff convergence process. These insights are crucial for optimizing landscapes to enhance ecological resilience.