Three connected programmes: the changing water cycle, water-related disasters and security, and new methods for hydrology.三个相互关联的研究方向:变化环境下的水循环、水灾害与水安全,以及水文水资源新技术。
Qilian Mountains祁连山
Water Cycle and Ecohydrology水循环和生态水文
The team investigates how global warming and human activities are altering fundamental water cycle processes. They seek to reconcile different data sources (like satellite, model (such as PML-V2 evapotranspiration and HBV-PML hydrological models), and ground observations) to better quantify these changes, asking critical questions such as why parts of the water cycle appear to be stagnating despite warming, and how human actions reshape water availability and ecosystem functions—like the roles of forests as “sponges” or “pumps”—at regional and global scales.团队研究全球变暖和人类活动如何改变水循环的基本过程,融合卫星、模型(如 PML-V2 蒸散发模型与 HBV-PML 水文模型)和地面观测以更好地量化这些变化,并回答关键问题:为何升温背景下部分水循环过程反而趋于停滞,人类活动又如何在区域和全球尺度上重塑水资源可用性和生态系统功能,例如森林究竟扮演“海绵”还是“水泵”的角色。
Rain falls, vegetation returns vapour, and runoff travels down the river network降水落地,植被蒸散发上升,径流沿河网下泄
Hengduan Mountains: 1 km terrain, with rivers routed from precipitation minus evapotranspiration (CRU TS 4.09; PML-V2.2). Rain and vapour are drawn to show the cycle, not measured amounts.横断山区:1 km 地形,河网由降水减蒸散发(CRU TS 4.09;PML-V2.2)汇流得到。雨滴与水汽仅示意水循环过程,并非实测量。
PML_v2PML_v2 蒸散发产品
Penman-Monteith-Leuning Evapotranspiration V2 (PML_V2) products include evapotranspiration (ET), its three components, and gross primary production (GPP) at 500m and 8-day resolution during 2000-2023.Penman-Monteith-Leuning 蒸散发第二版(PML_V2)产品提供 2000—2023 年 500 m、8 天分辨率的蒸散发及其三个组分和总初级生产力(GPP)。
PML_v2Related publications (4)相关论文(4)
Zhang, Y., Peña-Arancibia, J.L., McVicar, T.R., Chiew, F.H.S., Vaze, J., Liu, C., Lu, X., Zheng, H., Wang, Y., Liu, Y.Y., Miralles, D.G., Pan, M., 2016. Multi-decadal trends in global terrestrial evapotranspiration and its components. Sci. Rep. 6, 19124. doi.org/10.1038/srep19124.
Zhang, Y., Kong, D., Gan, R., Chiew, F.H.S., McVicar, T.R., Zhang, Q., & Yang, Y. (2019). Coupled estimation of 500m and 8-day resolution global evapotranspiration and gross primary production in 2002-2017. Remote Sensing Environ. 222, 165-182. doi.org/10.1016/j.rse.2018.12.031.
He, S., Zhang, Y., Ma, N., Tian, J., Kong, D., & Liu, C. (2022). A daily and 500 m coupled evapotranspiration and gross primary production product across China during 2000–2020, Earth Syst. Sci. Data., doi.org/10.5194/essd-2022-183.
Wang, X., Zhang, Y., Zhang, X., He, S., Kong, D., Tian, J., Wei, H., Wang, L., Quan, Y., Zheng, Y., & Wang, Y. (2025). Satellite data-driven estimation of daily and 500 m net ecosystem exchange over China during 2003–2020. Remote Sensing of Environment, 331, 115047. doi.org/10.1016/j.rse.2025.115047
HBV-PMLHBV-PML 模型
The HBV-PML model, provides critical data support for global water resource assessments. This innovative model integrates the remote sensing evapotranspiration model PML with the classical hydrological model HBV, significantly improving the accuracy of evapotranspiration simulation through remote sensing observations. It effectively overcomes the limitations of traditional models in representing vegetation dynamics.HBV-PML 模型将遥感蒸散发模型 PML 与经典水文模型 HBV 相结合,利用遥感观测显著提高蒸散发模拟精度,克服了传统模型难以刻画植被动态的局限,为全球水资源评估提供关键数据支撑。
HBV-PML
Water-Related Disasters and Security水灾害和水安全
This theme focuses on the mechanisms and impacts of extreme hydrological events, including droughts and floods. The group studies how droughts develop and propagate through different environmental layers, how flood patterns have evolved in China over decades, and how to enhance societal and ecological resilience to water-related extremes in a changing climate.这一方向聚焦干旱和洪水等极端水文事件的机理与影响,研究干旱如何在不同环境圈层中发展和传播,中国洪水格局在数十年间如何演变,以及如何在气候变化下增强社会和生态系统对水相关极端事件的韧性。
Storm暴雨
Flood洪水
Drought干旱
Storm, flood and drought on the middle Yangtze长江中游的暴雨、洪水与干旱
Three Gorges and the middle Yangtze on the 1 km terrain. A storm rains on the upper catchment; the flood that follows covers the land below a water level rising to 24 m above the nearest channel, from the height above the nearest drainage (HAND; Nobre et al., 2011); a drought then dries the smaller streams first. The gauge on the Yangtze near Shashi marks a flood warning (red) and a low-flow warning (amber). Rain, water levels and the drought illustrate the processes studied; they are not observations or forecasts.三峡与长江中游(1 km 地形)。暴雨降落在上游流域;随后的洪水依据相对最近河道高差(HAND;Nobre 等,2011)显示水位升至高出最近河道 24 米过程中被淹没的土地;之后干旱先使较小的支流断流。长江沙市附近的水位标尺标出洪水预警线(红)与低水位预警线(橙)。降雨、水位与干旱过程仅示意研究内容,并非观测或预报。
Fire Impacts and Cascading Hazards林火影响与级联灾害
This research examines how wildfires alter catchment water balance and intensify subsequent flood hazards. Studies in southeast Australia show that fire-induced vegetation and land-surface changes can modify evapotranspiration and streamflow, while amplifying flood peaks under certain hydroclimatic conditions. These findings highlight the cascading hydrological impacts of forest fires and their implications for flood risk.本研究关注林火如何改变流域水量平衡并加剧后续洪水灾害。在澳大利亚东南部的研究表明,林火引起的植被与陆面变化会改变蒸散发和径流,并在特定水文气候条件下放大洪峰。这些发现揭示了林火的级联水文影响及其对洪水风险的意义。
Fire Impacts and Cascading HazardsRelated publications (2)相关论文(2)
Xu, Z., Zhang, Y., Zhang, X., Ma, N., Tian, J., Kong, D., & Post, D. (2022). Bushfire‐Induced Water Balance Changes Detected by a Modified Paired Catchment Method. Water Resources Research, 58(11). doi.org/10.1029/2021wr031013
Xu, Z., Zhang, Y., Blöschl, G., & Piao, S. (2023). Mega Forest Fires Intensify Flood Magnitudes in Southeast Australia. Geophysical Research Letters, 50(12). doi.org/10.1029/2023gl103812
Drought Recovery and Hydrological Resilience干旱恢复与水文韧性
This research investigates how large river basins recover from severe droughts and why recovery varies across regions. By integrating GRACE satellite observations with global water availability estimates, we quantify terrestrial water storage recovery and identify pronounced disparities between humid and dryland basins. We further examine how post-drought water availability, recurrent droughts, and human water use influence recovery trajectories and long-term hydrological resilience.本研究关注严重干旱后大型流域水文系统的恢复过程及其区域差异。通过融合 GRACE 卫星观测与全球水资源可利用量估算,量化陆地水储量恢复过程,揭示湿润与干旱流域之间显著的恢复差异,并进一步分析干旱后水分可利用量、重复干旱和人类用水对水文恢复及长期韧性的影响。
Drought Recovery and Hydrological ResilienceRelated publications (1)相关论文(1)
Xu, Z., Zhang, Y., Blöschl, G., Wagener, T., AghaKouchak, A., Zhang, X., Wang, L., Tian, J., Ma, N., Li, X., Li, C., Huang, Q., Tang, Z., Yang, X., Xia, J., & Liu, C. (2026). Satellite evidence of unequal basin-scale recovery from severe droughts. National Science Review, nwag624. doi.org/10.1093/nsr/nwag624
Innovating Water Science水文水资源与新技术
The group leverages advanced technologies such as high-resolution remote sensing and artificial intelligence (AI) to improve the monitoring and prediction of hydrological processes. They develop fine-scale datasets for evapotranspiration and vegetation productivity, integrate AI with physical models to enable better forecasting in ungauged basins, and work toward real-time flood prediction systems that support water resources management under climate change.团队利用高分辨率遥感和人工智能等先进技术改进水文过程的监测与预测:研发蒸散发和植被生产力的精细尺度数据集,融合人工智能与物理模型以提升无资料流域的预报能力,并致力于发展实时洪水预报系统,为气候变化下的水资源管理提供支撑。
Satellite卫星遥感
AI model人工智能
Forecast预报
Satellite retrieval, AI model and forecast over the Loess Plateau黄土高原上空的卫星遥感、人工智能与预报
Loess Plateau and the Yellow River: mean evapotranspiration for 2001–2020 from our PML-V2.2 product on 0.1° cells, light to dark blue for 100 to 700 mm per year. The network’s inputs take the colour of the cells beneath them. The satellite pass, the neural network and the forecast on the rivers illustrate how remote sensing and AI are combined with physical models; they are not a model run.黄土高原与黄河:本组 PML-V2.2 产品 2001–2020 年平均蒸散发(0.1° 网格),由浅至深蓝表示 100–700 毫米/年。神经网络输入节点的颜色取自其下方网格的数值。卫星过境、神经网络与河网上的预报仅示意遥感、人工智能与物理模型的结合方式,并非模型运行结果。
AI for Hydrology Modeling人工智能水文建模
We introduce an AI-driven Transformer-based deep learning model to address data gaps in GRACE and GRACE Follow-On satellite data. By integrating meteorological and hydrological features like temperature, precipitation, and evapotranspiration, the model effectively fills the 11-month data gap, showing superior accuracy, particularly in arid regions. The transformer-based approach outperforms traditional methods, demonstrating the potential of AI in enhancing satellite data continuity and reliability for environmental monitoring.我们提出基于 Transformer 的人工智能深度学习模型,用于填补 GRACE 与 GRACE Follow-On 卫星观测之间的数据缺口。模型融合气温、降水和蒸散发等气象水文要素,有效填补约 11 个月的观测空缺,精度优于传统方法,在干旱区表现尤为突出,展示了人工智能在提升卫星数据连续性与可靠性方面的潜力。
AI for Hydrology ModelingRelated publications (1)相关论文(1)
Wang, L., & Zhang, Y. (2024). Filling GRACE data gap using an innovative transformer-based deep learning approach. Remote Sensing of Environment, 315, 114465. doi.org/10.1016/j.rse.2024.114465
CAS-Canglong中科苍龙
We present a novel AI-based approach for sub-seasonal to seasonal (S2S) global sea surface temperature (SST) prediction, using a 3D Transformer model called CAS-Canglong. By leveraging the self-attention mechanism inherent in Transformer architectures, the model efficiently handles the complexities of ocean-atmosphere interactions and improves prediction accuracy. Compared to traditional physics-based models, CAS-Canglong demonstrates superior skill in predicting SST patterns, enhancing computational efficiency and forecasting accuracy for up to three months in advance. This approach shows great potential for improving predictions of extreme weather events like droughts and floods, as well as for future climate studies.我们提出面向全球海表温度次季节至季节(S2S)预测的人工智能方法——三维 Transformer 模型“中科苍龙”(CAS-Canglong)。模型利用自注意力机制刻画复杂的海气相互作用,与传统物理模式相比,在最长三个月的提前期内具有更高的海温预测技巧和计算效率,可为干旱、洪水等极端事件预报和未来气候研究提供支撑。
CAS-CanglongRelated publications (1)相关论文(1)
Wang, L., Zhang, X., Leung, L. R., Chiew, F. H. S., AghaKouchak, A., Ying, K., & Zhang, Y. (2024). CAS-Canglong: A skillful 3D Transformer model for sub-seasonal to seasonal global sea surface temperature prediction (No. arXiv:2409.05369). arXiv. doi.org/10.48550/arXiv.2409.05369