Evaluating Future Runoff, Water Stress and Fluvial Flooding in China under Climate Change

China faces complex changes in runoff variability, water stress, and fluvial flood hazards under climate change and socioeconomic development. Understanding how these processes may evolve under future scenarios is critical for long-term water resource planning and climate adaptation. However, existing studies often suffer from fragmented assessments. For example, runoff projections are typically limited to catchment-scale models or coarse-resolution global datasets, making it difficult to capture China’s diverse hydrological patterns at a national scale in high resolution. Water stress assessments frequently apply simplified assumptions or inconsistently combine climate and socioeconomic scenarios, leading to internally inconsistent projections. Moreover, flood risk studies in China rarely include nationwide evaluations of inundation exposure, and no studies to date have examined such risks under climate overshoot scenarios, despite their increasing plausibility under current emission trajectories. This study presents an integrated assessment of future runoff, water stress and fluvial flooding across China under a range of climate and socioeconomic scenarios. To achieve this, an integrated and novel multi-model framework was developed, combining a land surface model (JULES), machine learning algorithms, and a hydrodynamic flood model (CaMa-Flood). The analysis considers three climate-socioeconomic pathways: SSP245 (moderate emissions), SSP585 (high emissions), and SSP534-OS (an overshoot scenario involving temporarily exceeding 2 °C warming before returning below it through mitigation). This modelling framework provides a novel, multi-dimensional assessment of runoff, water stress and fluvial flooding in China by combining high-resolution, nationally consistent hydrological simulations of runoff, scenario-consistent water availability and demand projections, and spatially explicit flood inundation analyses. It represents the first application of the JULES model for hydrological simulation in China, integrates physically based and machine learning approaches within a scenario-consistent framework, and presents the first assessment of overshoot scenario impacts on fluvial flooding and inundation. The runoff projections reveal contrasting seasonal trends between northern and southern China under SSP245 and SSP585 scenarios. Southern regions of China are projected to experience wetter summers and drier winters, while northern regions show the opposite pattern. Notably, under the overshoot scenario SSP534-OS, summer runoff in the Yangtze Basin is expected to exceed that under SSP585 in the mid-century, though this trend reverses by the end of century. For the investigated scenarios, extreme high and low runoff events are projected to increase substantially. Over 56% of China, particularly in the south, is expected to face intensified extreme high runoff, while more than 40% may experience more frequent low runoff, especially in central and southern regions. Under SSP585, these trends become more severe over time, with basin-mean high runoff exceeding 140% of historical levels by the end of the century. Water stress projections suggest moderate national-level stress throughout the century. However, severe regional disparities emerge. The northwest and central provinces are expected to face worsening stress due to declining water availability and increasing demand, particularly under SSP585. In contrast, northeastern provinces such as Heilongjiang and Hebei may see improved conditions, benefiting from increased water supply and reduced demand. The main drivers of these shifts are regional changes in water availability and sectoral water use, with urban growth, industrial expansion, and agricultural demand playing key roles. Fluvial flood risk analysis under SSP534-OS and SSP585 indicates that the overshoot pathway may reduce flood risk in the long term but poses heightened mid-century risks. National flood magnitude increases more sharply under SSP585, reaching a 90% rise by end of the century. However, in 2041-2070, flood hazards under SSP534-OS exceed those under SSP585 across 49% of China, particularly in the Yangtze Basin, indicating the early intensification of overshoot-related risks. Flood exposure assessments show that key basins - including the Yangtze, Yellow, Huaihe, and Haihe River basins - are vulnerable to both cropland and population inundation. Under SSP534-OS, exposure is expected to peak earlier, but SSP585 is projected to lead to dramatically higher impacts by 2100. Population exposure under SSP585 is projected to nearly triple that of SSP534-OS, while cropland inundation ratios exceed 25% in some regions. These findings emphasise the dual need for early adaptation and long-term mitigation to manage flood risks. Overall, this study provides new insights into the spatial and temporal dynamics of hydrological change across China, highlighting critical regional vulnerabilities and offering a scientific basis to support future strategies for water resource management, climate adaptation, and risk mitigation.<p></p>

Author dc:creator

• Danyang Gao (21061607)

• hydrology
• climate change
• water-related risks
• hydrological modelling
• machine learning

Statement dc:rights

• All rights reserved
• Open Access after 2027-08-23

Identifier

10779/exe.31365904.v1

OAI identifier oai:identifier

oai:figshare.com:article/31365904