Multiscenario simulation of LUCC and landscape ecological risk in the Tarim River basin based on the intPLUS model

Land use and cover change (LUCC) is a primary driving force that shapes the spatiotemporal patterns of landscape ecological risk (LER), particularly in ecologically fragile arid regions. We investigated the Tarim River Basin, the largest endorheic basin and a critical ecological security zone in arid northwestern China, with the aim to address the lack of quantitative evidence on how future policy pathways will reshape the LER. LUCC and LER changes were simulated under four 2035 scenarios: business-as-usual (BAU), prioritizing urban development (PUD), prioritizing ecological protection (PEP), and balanced urban–rural development (BUE). By coupling historical LUCC data (2010−2020) with the intPLUS model, we developed a spatially explicit framework integrating LUCC simulation with landscape ecological risk assessment. (1) After calibration against the observed 2020 pattern, the intPLUS model reproduced the 2020 land use at the 1-km scale (kappa = 0.941; OA = 0.972; FOM = 0.184), supporting its use for subsequent multiscenario forecasting. (2) From 2010 to 2020, cropland and developed land expanded by 10,694 and 1536 km 2 , respectively, whereas grassland decreased by 16,440 km 2 , increasing the mean LER from 0.031 to 0.034, with high-risk areas expanding by 28,283 km 2 . (3) By 2035, compared with the BAU scenario, the PEP, PUD, and BUE scenarios reduced high-risk areas by 1.16%, 1.24%, and 1.27%, respectively. Higher-risk areas were most policy-sensitive. Notably, under the selected scenario settings, the balanced urban–rural development scenario (BUE) compressed the high-risk core slightly more than the other intervention scenarios, whereas the prioritizing ecological protection (PEP) scenario produced a smaller reduction than the prioritizing urban development (PUD) scenario; these scenario-conditional comparisons indicate that the ranking of outcomes is not simply ordered by protection intensity. (4) LISA revealed a robust desert hotspot ring versus mountain cold-spot core structure ( p < 0.001), with policy interventions providing gradient buffering only in transition zones.