The effects of landscape patterns on ecosystem services of urban agglomeration in semi-arid area under scenario modeling

• The effects of landscape patterns on ESs were explored from current and future views. • The optimal sampling grid size was determined for landscape patterns in the YUA. • Spatial scale effects of landscape patterns on ESs were discussed. • The spatial heterogeneity of the effects of landscape patterns on ESs was exposed. Optimizing landscape patterns to promote ecosystem services (ESs) can alleviate human-land conflicts and contribute to achieving the United Nations Sustainable Development Goals (SDGs). However, delving deeper into the effects of landscape patterns on ESs under future scenarios remains challenging. This study combined various models, including the PLUS, InVEST, GeoDetector and MGWR models, to explore the effects of landscape patterns on ESs from current and future perspectives based on multi-scenario land use simulation and the evaluation of landscape patterns and ESs in the Ningxia Along the Yellow River Urban Agglomeration (YUA). The results show that (1) At the optimal scale of 2.7 km × 2.7 km, landscape types have become increasingly diverse and scattered, with habitat quality (HQ) and carbon sequestration (CS) declining and water yield (WY) increasing from 2005 to 2020. The change trends in landscape diversity, fragmentation under the cropland-ecological protection (CEP) and natural development (ND) scenarios during 2020–2035, are consistent with HQ trends from 2005 to 2020, while on the contrary inconsistent with WY. Notably, these trends under the CEP scenario change more slowly than those under the ND scenario. (2) Landscape indicators exert the strongest effects on HQ, followed by CS and WY. Landscape diversity emerges as the primary driver of ESs at the landscape level, while area indicators are predominant at the class level. (3) The spatial scale of landscape patterns’ influence on ESs indicates prioritization at both county and urban agglomeration scales for developing better landscape management measures. (4) The YUA should be divided into six types of regions for landscape management, and it’s necessary to formulate differentiated landscape management measures for each region. (5) This study reveals the spatial heterogeneity of the effects of landscape patterns on ESs, providing valuable partitioning management strategies for landscape planning to further optimize its patterns and thereby promote regional ESs.