Assessing future flood risks to cropland under climate change in the Second Songhua River Basin, China: A depth-duration framework for adaptive management

Study region The Second Songhua River Basin (SSRB), a representative mid-latitude agricultural basin in Jilin Province, northeastern China. Study focus To overcome the limitations of traditional single-indicator flood hazard assessments that overlook agricultural waterlogging stress, this study develops a process-oriented, depth-duration dual-factor risk assessment framework. We couple a bias-corrected CMIP6 multi-model ensemble with a calibrated distributed Rainfall-Runoff-Inundation (RRI) model. Future cropland inundation extent, depth, and duration are simulated under multiple emission scenarios (SSP245/SSP585), future horizons, and return periods (50/100-year). A dual-factor decision matrix is constructed to classify cropland into four risk quadrants to guide differentiated drainage planning. New hydrological insights for the region Future flood hazards show a shifting trend from localized ponding to contiguous deep waterlogging belts. Under the far-future SSP585 scenario, inundated cropland reaches 3,107.33 km 2 (uncertainty range: 2249.50–3830.13 km 2 ) for a 50-year flood and peaks at 3287.41 km 2 for a 100-year event (uncertainty range: 2364.96–4179.78 km 2 ). By deploying the proposed depth–duration framework, we successfully decoupled and interconnected the compound hazard components: within the 100-year footprint, a critical 931.34 km 2 (uncertainty range: 654.72–1174.25 km 2 ) of the affected cropland is identified as the highest-risk zone, suffering from concurrent severe submersion (>2.0 m) and prolonged waterlogging (>3 days). The resulting risk matrix effectively categorizes the basin into four distinct adaptive management zones, providing a scientific baseline for implementing differentiated farmland governance and supporting climate-adaptive risk management.