PICO
PICO: Wed, 6 May, 08:30–12:30 | PICO spot 3
Global warming has significantly altered the spatiotemporal distribution of floods, leading to substantial variations in human adaptation patterns. Identifying the potential drivers of these changes and the underlying mechanisms of disaster adaptation is essential for formulating effective flood risk strategies. Based on observed streamflow records from 9,531 hydrological stations and data from 910 major flood events worldwide, this study reveals that most regions globally exhibit synchronized trends in drought and flood flows, with 28.14% showing a simultaneous increase and 33.36% showing a simultaneous decrease. To mitigate flood risk, residents in 53% of countries—most notably in the Middle East—demonstrate a tendency to migrate away from flood-prone areas. This retreat has significantly reduced flood-related mortality and forced displacement. Conversely, in regions with robust flood protection infrastructure, residents tend to maintain shorter migration distances. Further analysis of the drivers behind floodplain migration indicates that in developing nations, flood-induced mortality and displacement are the primary catalysts for relocation. In these contexts, the psychological memory of destruction or the urgent need for resources often compels residents to either flee or, paradoxically, migrate toward flood-prone zones. Under climate-driven pressures, the extent of flood inundation is a more significant determinant of migration patterns in regions such as Australia. Notably, in countries like the Philippines and Kenya, the mitigation of compound drought-flood extremes has encouraged further settlement in flood-prone areas, highlighting the complexity of multi-hazard interactions. This study systematically deciphers the mechanisms underlying flood adaptation strategies and attributes their primary drivers, providing a robust scientific framework for enhancing flood risk management and regional resilience.
How to cite: Wang, N.: Resident Adaptation Patterns Under the Influence of Global Flood Evolution, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6507, https://doi.org/10.5194/egusphere-egu26-6507, 2026.
Floodplains attract disproportionate concentrations of population and economic activity globally, yet the systemic flood risks emerging from this uneven development remain poorly characterized. Through a global analysis spanning 2000-2020, we quantify floodplain development patterns and associated flood losses across nations with varying income levels and flood protection capacities. Our results reveal that floodplains have experienced faster growth than non-floodplains in both population density and GDP density. These trends diverge sharply by income and protection levels: floodplain population density growth rates in low- and lower-middle-income countries outpaced those in high-income nations by factors of 2.33 and 7.58, respectively. Similarly, due to levee effect, regions with flood protection capacity of 100 years or more experienced GDP density growth that was 4.51 times higher than in regions with less than 10-year protection. The heightened sensitivity of flood losses to socio-economic growth stems from uneven floodplain development. This creates a divergent risk pattern: wealthier, well-protected regions accumulate greater economic assets at risk, whereas poorer, under-protected areas face the compounded burden of exposure to both population and GDP risks. Our findings highlight the urgent need for flood risk adaptation strategies that explicitly consider and address underlying floodplain socio-economic inequalities in exposure and protection.
How to cite: Lan, Z., Liu, W., Wang, T., and Sun, F.: Mapping global floodplain development disparities highlights drivers underlying intensifying flood losses, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5280, https://doi.org/10.5194/egusphere-egu26-5280, 2026.
The increasing frequency of extreme weather events is drawing attention to groundwater flooding, which is caused by rising groundwater levels and can result in significant damage to infrastructure, buildings, and the environment. Unlike fluvial or pluvial flooding, groundwater flooding is difficult to detect and not easily managed with traditional protective measures. Numerical models—particularly probabilistic approaches such as Bayesian inference—help to better quantify uncertainties in modeling and forecasting. Flood risk maps are essential for managing groundwater flooding; however, precise uncertainty analyses are often lacking. Citizen science and low-cost sensors can also contribute by bridging data gaps and encouraging public participation. This study presents a framework for assessing vulnerability to groundwater flooding that accounts for uncertainties and generates probabilistic maps. Using a case study from Garching in 2023, it demonstrates how modeling tools can be effectively utilized. Finally, the study suggests expanding monitoring tools and citizen engagement to strengthen risk communication, raise awareness, and better integrate groundwater flood protection measures.
How to cite: Chiogna, G. and Richieri, B.: Groundwater Flooding: Developing an approach to risk assessment and communication, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3383, https://doi.org/10.5194/egusphere-egu26-3383, 2026.
Extreme precipitation events have caused obvious damage to human environments and socioeconomic systems. However, the changes in extreme precipitation and their underlying causes remain unclear. This study analyzed daily precipitation data from 2,254 meteorological stations across China from 1981 to 2018, focusing on two key extreme precipitation indicators: Max 1-day precipitation amount (Rx1day) and Max 5-day precipitation amount (Rx5day). Trend analysis was conducted for 17 river basin divisions using the Mann-Kendall method. We also applied the field significance test, a statistical method to evaluate whether a spatial pattern of locally significant results, to determine whether observed trends at individual stations were statistically significant or due to random variation. The results showed that 59.3% and 58.6% of the stations exhibited increasing trends in Rx1day and Rx5day, respectively, with significant trends identified at 5.4% and 4.1% of the stations. The field significance test revealed a significant increasing in Rx1day across China at the 5% significance level. Among the 17 sub-basins, significant increases in extreme precipitation were observed in the Inland rivers of Xinjiang and Northern Tibet. The result was consistent with the warming and humidification trends in northwest China. We further analyzed the relationship between urbanization and extreme precipitation by using population density to distinguish rural and urban stations. We found that the spatial distribution of urban stations closely overlapped with stations experiencing increased extreme precipitation, while rural stations corresponded with those showing a decrease. With the progress of urbanization, variations in the trends observed at urban and rural stations have emerged. Nevertheless, urban stations exerted a more pronounced influence on the increasing trend of extreme precipitation.
How to cite: Wu, L.: Urbanization influence on changes of extreme precipitation in mainland China, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9087, https://doi.org/10.5194/egusphere-egu26-9087, 2026.
Global warming is altering snowmelt dynamics and flood generating mechanisms, yet their compound effects on cold-region floods remain unclear. Here, we investigate flood mechanism transitions and their drivers across 424 Northern Hemisphere snow-dominated catchments. Through comparative analysis, we pinpoint the specific impacts of these shifts on flood characteristics. Our results indicate that 48.3% of the catchments have undergone a snowmelt-to-rainfall transition in flood generating mechanisms. While this has not systematically altered long-term flood magnitude trends, it has significantly steepened the flood rising limb. Furthermore, although rising temperatures have advanced the timing of snowmelt and rain-on-snow floods, the shift toward rainfall dominance has largely offset this trend, leading to a stronger synchronization between flood timing and extreme precipitation. These findings offer critical insights for flood forecasting and water management in snow-dominated regions.
How to cite: Bai, X., Liu, W., Wang, H., Feng, Y., and Sun, F.: Changing flood-generating mechanisms and their impacts on flood characteristics in snow-dominated catchments, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5267, https://doi.org/10.5194/egusphere-egu26-5267, 2026.
Hydrological signatures (HS) have proven to be highly effective in calibrating physically-based hydrological models, enhancing their process consistency. However, their integration into parameter optimization for deep learning (DL)-based hydrological models has been limited. To address this gap, we propose a novel HS-informed framework that dynamically integrates hydrological signatures into DL parameterization through a multi-task learning approach. This study evaluates the impact of HS integration on model performance using a large-scale, global hydrological dataset. The HS-informed model achieved a significant performance improvement, with a median Nash-Sutcliffe Efficiency (NSE) of 0.739, compared to 0.666 for the baseline model across the test set. Notably, the most pronounced improvements in NSE were observed in hydrologically complex basins, including baseflow-dominated (+0.135), drought-prone (+0.148), and flood-prone basins (+0.159). Sensitivity analysis further revealed that the HS-informed model could leverage extended historical input data (over 120 days) to sustain robust performance (median NSE of 0.715) over a 30-day forecast period. Shapley Additive Explanations (SHAP) analysis highlighted two key mechanisms underlying these improvements: the enhanced recognition of long-term hydrological patterns through improved memory and a better representation of catchment heterogeneity by emphasizing non-climatic attributes. These findings demonstrate that integrating hydrological signatures offers a superior approach to traditional point-error-based calibration in AI-driven hydrological modeling.
How to cite: wang, Z., li, C., and cui, P.: A Novel Hydrological Signature-Informed Framework for Enhancing Extreme Streamflow Prediction Using Multi-Task Learning, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2556, https://doi.org/10.5194/egusphere-egu26-2556, 2026.
Intense rainfall events often trigger landslides and torrential flows, which are not only hazardous processes on their own, but can also generate cascading hazards through sudden and massive sediment delivery to river networks. Slope processes are therefore key drivers of geomorphic change in mountainous catchments, enhancing hillslope-channel connectivity and promoting rapid channel reorganisation. In light of the above, it is essential to characterise structural and functional connectivity (Heckmann et al., 2018), as well as geomorphic organisation before and after intense precipitation events, to better evaluate flood hazards and associated risks. Against this background, the January 2020 Gloria storm affected the Tordera River basin (Catalonia, NE Spain), where more than 480 mm of rainfall was recorded over 96 hours, with 24-hour accumulation over 200 mm, causing widespread sediment mobilisation and channel changes, as well as significant damage due to flooding and landslides.
In this study, we aim to evaluate: 1) how pre-event hillslope-channel connectivity influences the geomorphic response to extreme floods and the post-event geomorphic changes through an integrated analysis of the index of connectivity (IC); 2) the spatial distribution patterns of erosion and sedimentation through the geomorphic mapping of the active riverbed and sediment bars (both active and stable), before and after the Gloria storm, and the existing inventory of landslides caused by the event. High-resolution DTMs (Digital Terrain Models) were generated from airborne LiDAR surveys conducted in 2011, 2016, and 2023. The IC was derived from a pre-event DTM to characterise structural sediment connectivity following Cavalli et al. (2013), while erosion and sedimentation processes were quantified using the difference between DTMs (DTMs of Difference, DoDs) for pre-event (2016-2011) and post-event (2023-2016) periods. GIS-based geomorphic mapping of active channels and sediment bars before and after Gloria was used to assess event-scale channel reorganisation.
Preliminary results indicate a clear spatial correspondence between pre-event connectivity patterns and the magnitude of the geomorphic change observed during that extreme flood. Areas characterised by high pre-event erosion rates, identified from 2016-2011 DoDs, largely coincide with sectors where numerous landslides were triggered during the Gloria storm. High connectivity values also correspond to areas dominated by erosion and deposition in the 2016-2011 DoDs, highlighting the role of pre-event structural connectivity in conditioning sediment transfer pathways. Furthermore, active bars mapped after the event predominantly overlap with areas affected by pre-event erosion, whereas bars that remained stable during the storm are mainly associated with zones characterised by pre-event deposition. The active channel also experienced noticeable widening during the event, while the majority of vegetated bars that were stable before Gloria remained stable throughout the storm, reinforcing the link between pre-event geomorphic organisation and flood response. These findings highlight the importance of pre-event structural connectivity in controlling geomorphic response during extreme rainfall events, providing valuable insight for hazard assessment and river management.
Cavalli, M. et al. (2013). Geomorphometric assessment of spatial sediment connectivity in small Alpine catchments. Geomorphology, 188, 31-41.
Heckmann, T. et al. (2018). Indices of sediment connectivity: opportunities, challenges and limitations. Earth-Science Reviews, 187, 77-108.
How to cite: Jacobo-Quiñones, N., Guinau, M., Abancó, C., García-Sellés, D., López-Tarazón, J. A., Zapico, I., Tapia, M., González, M., and Pinyol, J.: Linking pre-event hillslope-channel connectivity to its geomorphic response during an extreme rainfall event: insights from the 2020 Gloria storm in the Tordera River basin (NE Spain), EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-7681, https://doi.org/10.5194/egusphere-egu26-7681, 2026.
This research investigates the influence of the El Niño–Southern Oscillation (ENSO) on extreme flood events in the United States and its potential connection to flood insurance claims from the National Flood Insurance Program (NFIP). Given the recently observed increase in the frequency of extreme weather events, this study aims to quantify the correlation between ENSO indicators and recorded economic losses at state and county levels across the USA. Emphasis is particularly placed on the state of California, which is highly sensitive to El Niño events.
The methodology is based on the integration of multiple datasets, including ENSO indices from NOAA, US-CAMELS streamflow data, COBE sea surface temperature (SST), digital elevation models (DEM), National Hydrography Dataset (NHD), OpenStreetMap (OSM), and US Census data. From these datasets, geospatial and physical features were extracted, such as hydrographic and road network density, mean elevation, distance to the coastline, county centroid coordinates, and population. These features were analyzed using statistical tools, including the Pearson correlation coefficient and Threshold Exceedance Analysis, applied across multiple percentile showing thresholds (90–99%).
In addition, a machine learning model was developed to predict flood insurance claims per 100,000 residents. The results indicate that correlations between ENSO indices and streamflow data are significantly stronger than those between ENSO indices and insurance claim records, highlighting the substantial influence of socioeconomic factors on the insurance claim filing process. California exhibits the highest positive correlation between the maximum annual ENSO index and insurance claims (r ≈ 0.35). The developed CatBoost model can be used to predict a high percentage (>60%) of their variability, using both static and dynamic features.
The study concludes that ENSO indices can contribute meaningfully to flood risk prediction frameworks. Future work will focus on extending the analysis to additional states or the entire USA and incorporating new explanatory features to further improve model performance.
How to cite: Tsolakidis, K.-C., Papoulakos, K., Tepetidis, N., Iliopoulou, T., Dimitriadis, P., Tsaknias, D., and Koutsoyiannis, D.: ENSO impacts on flood risk and insurance claims in the United States: a machine learning approach, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9856, https://doi.org/10.5194/egusphere-egu26-9856, 2026.
Kerala, on the windward side of the Western Ghats in southern India, receives about 3000 mm of annual rainfall under a tropical monsoon climate, driven by orographic south-west monsoon rainfall. The state has a high population density of about 859 persons per square kilometre and a limited geographical extent, with settlements concentrated in river valleys and downstream of reservoirs. These physiographic and socio-hydrological conditions make flood events critically important from both hydrological and societal perspectives. The catastrophic flood of 2018 further emphasized the need for an updated hydrological reassessment of existing dams and their spillway performance, and reservoir rule curves in Kerala.
Kerala has 53 large dams, of which 30 dams distributed across nine river basins are analysed in this study. The selected catchments are characterized by short hydrological response lengths and steep terrain, with longitudinal bed slopes ranging from 20 to 80 m km⁻¹. The Western Ghats rise sharply from near sea level to elevations of approximately 2500 m, promoting intense orographic rainfall, short travel times, and rapid runoff concentration. For the 30 dam catchments, the Time of Concentration (Tc) varies between 0.7 and 5 h, indicating fast-rising floods with minimal natural attenuation. Several catchments exhibit high hydrological response, with specific flood exceeding 13 m³ s⁻¹ km⁻². Most dams are located within 100 km of the Arabian Sea coastline and occur in serial or cascade arrangements along the same river valleys, a configuration that is hydrologically relevant for upstream–downstream flood interactions.
The study reassesses the Inflow Design Flood (IDF) and spillway adequacy of the selected dams. Of the 30 projects, 20 dams were completed before 1985, before the Bureau of Indian Standards (BIS) issued Indian Standard IS 11223:1985, which formally introduced IDF categories such as the Probable Maximum Flood (PMF), Standard Project Flood (SPF), and 100-year flood. In projects commissioned before 1985, spillway capacities were generally fixed using prevailing hydrological practices, limited storm data, and engineering judgment.
In the present reassessment, IDF estimation is carried out in accordance with BIS guidelines using a hydro-meteorological approach, and unit hydrograph parameters are derived from the Flood Estimation Report. Storm parameters are derived from the Probable Maximum Precipitation (PMP) Atlas for the West-Flowing Rivers of the Western Ghats, published by the India Meteorological Department (IMD) and the Central Water Commission (CWC), which compiles major historical storm events from 1905 to 2010. The revised design floods are compared with existing spillway capacities, and the analysis also examines relationships with Tc, gross storage, specific flood, and year of dam completion.
Results indicate that 26 out of the 30 dams show spillway inadequacy under the revised IDF. In several projects, design flood exceedance exceeds 200%, and in some cases, reaches more than 300%. Spillway inadequacy is more frequent in short-response catchments with lower Tc and higher specific flood values. This study offers a comparative hydrological perspective for steep tropical catchments in Kerala. It may support an informed, evidence-based reassessment of existing dams using updated datasets and contemporary analytical practices for prioritization of dam safety.
How to cite: Issac, I., Sen, S., and Goel, N. K.: Design Flood Revisions and Spillway Adequacy in Steep Tropical Catchments: A Multi-Dam Reassessment from Kerala, India , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-7809, https://doi.org/10.5194/egusphere-egu26-7809, 2026.
Groundwater abstraction can substantially reshape how climatic drought propagates into aquifer storage in semiarid Mediterranean settings. Here we propose an attribution framework to separate hydroclimatic and anthropogenic controls on standardized groundwater anomalies in two hydraulically connected aquifers of Spain’s Ebro Basin: the Plio-Quaternary of Alfamén and the Miocene of Campo de Cariñena.
We first reconstruct temporally continuous groundwater-level series for 1980–2025 using transfer-function noise (TFN) models in Pastas. Models are driven by daily precipitation and Penman–Monteith potential evapotranspiration, and include reconstructed monthly abstraction stresses. From these reconstructions we compute monthly Standardized Groundwater Indices (SGI) under current, pumped conditions and compare them to multiscale Standardized Precipitation–Evapotranspiration Index (SPEI) to quantify climate–groundwater coupling and identify spatial response types. We then isolate the effect of abstraction by building counterfactual “no-pumping” simulations through linear decomposition of calibrated TFN models and removal of pumping contributions, enabling within-piezometer comparisons against a reference-consistent baseline. Focusing on 2010–2025, we evaluate how abstractions alters anomalies beyond frequency using an SGI < −1 threshold, including month-level reclassification, event structure, peak timing, exceedance probabilities, and the instantaneous abstraction effect defined as ΔSGI = SGI_pumped − SGI_nopump.
Under pumping, climate–groundwater coupling strengthens monotonically with climatic accumulation, mean SGI–SPEI correlations increase from ~0.07–0.10 at 1-month SPEI to ~0.43 (Alfamén) and ~0.52 (Cariñena) at 48-month SPEI scales. Long-window coupling and response types show coherent spatial organization across intensively cultivated areas, particularly along the valley floor and lower piedmont. Persistent SGI declines under pumping concentrate in the central parts of both aquifers, broadly coinciding with irrigation hotspots, whereas piezometers near aquifer margins more often exhibit transient or non-significant declines. A key exception occurs in the shallow Plio-Quaternary of Alfamén near ephemeral streams, where episodic focused infiltration can temporarily offset local drawdown. Removing abstraction fundamentally shifts the apparent drought timescale. SGI without pumping shows no declining trends and aligns most strongly with annual climate variability (around SPEI12), with correlation peaks up to ~0.8 and network means near ~0.45 in both aquifers, indicating that observed downward SGI trends largely reflect externally imposed abstraction.
Counterfactual diagnostics reveal temporal reorganization. Pumping produces longer, more persistent anomalies episodes and seasonally biased onsets (late autumn/early winter, plus a June onset cluster in Cariñena), while peak timing of the events remains partly climate-governed. Exceedance probabilities of crossing SGI < −1 are higher at every monitoring point under pumping; the largest increases appear in central sectors, where sustained pumping and thicker saturated zones amplify cumulative stress on storage, but elevated likelihoods and ΔSGI also extend beyond the main abstraction hotspots into areas without raw drawdown signals. Over the monitoring network, we observe that pumping increases the likelihood and persistence of moderate groundwater anomalies, delays recovery, and lengthens the effective memory of the system, implying that SGI derived from observed heads in heavily exploited aquifers reflects a compound climate–management signal and should be complemented with counterfactual baselines and month-resolved persistence metrics for attribution and management.
How to cite: Cid-Escobar, D., Limones, N., Fernández, M., and De Stefano, L.: Beyond climatic-driven groundwater drought: extracting anthropogenic signatures fromstandardized groundwater indices, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11457, https://doi.org/10.5194/egusphere-egu26-11457, 2026.
Climate change and anthropogenic activities increasingly stress groundwater resources, even in generally water-rich areas like Germany, threatening socio-economic and ecological systems. Since the impacts of groundwater droughts often emerge slowly and implicitly, it remains unclear to what extent they are noticed and recognized by society.
We address this gap by linking hydrological observations of groundwater droughts in Germany with news-derived indicators of societal awareness at the national scale. We analyzed 30-year groundwater records from and 521 regions and 13,900 monitoring wells across aquifers of different depths, after quality control including outlier screening, level-shift detection, and linear interpolation of short gaps (≤1 month) to daily resolution. We then identified drought periods and quantified their duration and severity using the variable threshold method, and classified events by the strength of potential human influence. Drought events with strong human influence are defined as those for which the variability of the associated time series dominates more by long-term trend rather than by interannual variability, or the event itself is strongly affected by abrupt level shifts. Finally, drought periods with strong and weak human influence were linked to a multi-sector drought-impact dataset derived from German newspaper articles (2000–2024) to assess societal awareness of groundwater droughts nationwide.
We found at least one drought event in 89.4% of the time series. In regions, drought events with weak human influence lasted, on average, 127 days and had a mean severity (maximum deviation below the drought threshold) of 0.2 m. Societal awareness was generally highest during the early phases of groundwater droughts, prior to the maximum groundwater-level deviation. Strong human influence amplified drought conditions, increasing the number of events by 7.2% and their mean duration by 2.9% within each region, and also leading to much earlier societal awareness. However, awareness did not persist throughout the drought period: awareness strength declined much faster than the groundwater-level recovery rate, and no significant relationship was found between changes in awareness strength and groundwater levels in deep aquifers during drought periods. These findings suggest that "invisible" groundwater droughts, especially in deep aquifers, are not fully perceived by society and highlight the need for improved groundwater policy coordination at the national level.
How to cite: Wang, Z., Peña Guerrero, D., Sodoge, J., Ebeling, P., Zheng, Y., Siebert, C., Madruga de Brito, M., Merz, R., Stahl, K., and Tarasova, L.: Is society aware of “invisible” droughts? - a groundwater perspective , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9896, https://doi.org/10.5194/egusphere-egu26-9896, 2026.
Droughts are commonly classified into meteorological, agricultural, hydrological, and ecological types, yet how these categories interact dynamically and propagate across space and time at subseasonal scales remains poorly understood. Here we show that subseasonal droughts propagate as directional, cascading processes across the land-atmosphere system. We develop an event-based analytical framework using event coincidence analysis to identify subseasonal drought events as sustained extremes in precipitation-evapotranspiration balance, soil moisture, runoff, and vegetation condition across the contiguous United States from 1982 to 2025, using satellite observations and land data assimilation system simulations. We find robust lead-lag relationships and coherent propagation pathways in which meteorological droughts systematically precede agricultural, hydrological, and ecological droughts across space and time. Event coincidence analysis identifies statistically significant drought sources and sinks and their time-lagged directional dependencies, allowing directional propagation patterns to be traced across drought types and regions. We find consistent cross-type drought transitions in several climate-sensitive regions (for example, SPEI → soil moisture → NDVI in the Southern Plains), with meteorological droughts typically preceding agricultural and ecological impacts by several weeks and with variable amplification along the transition. Linking these propagation pathways to near-surface temperature, wind fields, and 850-hPa geopotential height shows that large-scale atmospheric circulation modulates timing and intensity of cross-type drought cascades. These findings show that subseasonal drought evolution is governed by directional temporal cascades and by coherent spatial propagation pathways across the land-atmosphere system, indicating non-local controls and distinct temporal signatures.
How to cite: Kumar, S. and Tian, D.: Cascading propagation of subseasonal droughts across the land-atmosphere system, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5997, https://doi.org/10.5194/egusphere-egu26-5997, 2026.
Soil moisture is a core element in shaping land–atmosphere interactions, playing a critical role in ecosystem functioning and sustaining water resources for human use. However, existing approaches, including numerical and AI-based methods, still suffer from notable limitations in soil moisture forecasting. In this study, we develop a novel AI-based soil moisture forecasting model (ASM), which is capable of providing low-resolution global forecasts and high-resolution regional forecasts of soil moisture at the subseasonal timescale. ASM consistently outperforms other representative state-of-the-art AI models across all forecast lead times. Compared with ECMWF, ASM is closer to the ground truth, and better preserve finer-scale spatial details. For regional predictions, ASM produces reliable high-resolution subseasonal soil moisture forecasts for two drought-prone regions selected as case studies: Southern Africa and Henan Province, China.
How to cite: Zhang, Q. and Huang, X.: Global–regional integrated subseasonal forecasts of soil moisture drought, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4172, https://doi.org/10.5194/egusphere-egu26-4172, 2026.
Flash droughts (FDs) are characterized by their rapid onset, but their societal and agricultural impacts depend critically on the duration of anomalous moisture stress. While the land–atmosphere processes governing FD initiation have been widely studied, the role of subsurface water storage in regulating persistence and recovery remains poorly constrained. Groundwater depth serves as the primary regulator of drought propagation. Rather than treating groundwater as a passive reservoir, this research investigates its active role in the initiation and subsequent evolution of FDs. We investigate how groundwater storage either dampens flash-drought intensification via upward moisture flux or catalyzes the evolution of these events into major hydrological crises. Our approach determines the precise influence of the water table on the intensification and multi-seasonal persistence of FD events. We utilize groundwater-level observations from the Central Ground Water Board of India, spanning 1996 – 2023, to construct seasonal groundwater depth fields (0.25° resolution) for pre-monsoon, monsoon, and post-monsoon conditions. FD events are identified using a gridded catalog derived from the Standardized Evaporative Stress Ratio (SESR). Our analysis will employ contingency-based statistical tests (χ²) and survival-type hazard analysis to quantify the probability of drought termination as a function of categorized water-table depths (shallow, intermediate, and deep). Spatial block-bootstrapping will be applied to account for regional spatial dependencies. We aim to identify critical groundwater depth thresholds beyond which the probability of flash-to-hydrological drought transition increases significantly. This work provides a new perspective on groundwater as a modulator of drought evolution in monsoon-dominated, groundwater-stressed environments.
How to cite: Vidushi, V. and Syed, T. H.: Groundwater Depth as a Control on Flash-Drought Dissipation Versus Hydrological-Drought Development in India, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6210, https://doi.org/10.5194/egusphere-egu26-6210, 2026.
Chile has endured a decade-long “mega-drought,” yet it remains unclear whether this represents a temporary climate anomaly or the onset of long-term aridification. While droughts are typically temporary events, persistent or recurrent droughts can indicate a transition toward aridification, that is, a gradual shift to drier conditions. We assessed how temporal changes in water supply and demand at multiple time scales affect vegetation productivity and land cover changes in continental Chile to diagnose the region's climate trajectory from drought to aridification. Since 2000, much of the region has seen a continuous decrease in water supply alongside a rise in atmospheric water demand. Further, in water-limited ecoregions, evapotranspiration, likely reflecting reduced transpiration or vegetation cover, has declined over time, with this trend intensifying over longer time scales. A long-term decline in water availability and shifting demand have led to declining vegetation productivity, especially in the Chilean Matorral and the Patagonia Steppe ecoregions. We discovered a link between these declines and drought indices related to soil moisture and actual evapotranspiration at time scales of up to 12 months. Further, our results indicate that the trends in drought indices account for up to 78% of shrubland and 40% of forest area changes across all ecoregions. The most important variable explaining cropland changes is the burned area. Our findings suggest that Chile is undergoing a transition from episodic drought to aridification, underscoring the need for adaptation strategies aligned with this emerging baseline.
How to cite: Zambrano, F., Vrieling, A., Meza, F., Duran-Llacer, I., Fernández, F., Venegas-González, A., Raab, N., and Craven, D.: From Drought to Aridification: Land-Cover Fingerprints of a Drying Chile, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20700, https://doi.org/10.5194/egusphere-egu26-20700, 2026.
Surface water availability (WA), defined as precipitation minus evapotranspiration, is affected by changes in vegetation structure. These biophysical impacts can alter the distribution of water availability, shifting both its average and extreme values, while the divergence is not yet quantified. Using long-term remote sensing observations, our analysis reveals that increases in leaf area index (LAI) lead to a widespread decline in average water availability, with a global reduction of -2.11 mm/month m2 m-2. Additionally, we show that in humid regions, extreme water availability—represented by the 15th and 85th percentiles of water availability from 2001 to 2020—exhibits stronger sensitivity to LAI variations than average water availability. Overall, the fraction of variance in low water availability explained by greening is minimal (-2.7%), followed by average water availability (6.8%), while high water availability exhibits the largest fraction (-23.6%).
How to cite: Li, Z., Liu, W., Wang, T., and Sun, F.: Biophysical impacts of Earth greening modulate average and extreme water availability, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8602, https://doi.org/10.5194/egusphere-egu26-8602, 2026.
Increasingly frequent and severe droughts pose substantial risks to agricultural water systems globally. Farmers can mitigate drought impacts through on-farm adaptation strategies, such as reducing drainage or increasing groundwater retention. However, the feedback between farmers’ adaptive behaviour and groundwater dynamics remains poorly understood. To address this gap, we developed an agent-based model to evaluate how individual farmers’ adaptation decisions influence local and regional groundwater systems. The model couples farmer decision-making, grounded in protection motivation theory, with the hydrological dynamics of the eastern Netherlands simulated using the WALRUS hydrological model. We ran scenarios based on different climate conditions and land use configurations to assess the effects of adaptation behaviour. Our findings show that farmers with adaptation measures experience substantially less drought damage associated with low groundwater levels during moderate droughts (65% reduction), but these measures are less effective during extreme droughts (13% reduction). Farmers who adopt these measures also experience slightly increased damage during wet periods, indicating a higher risk of waterlogging. Importantly, both benefits and drawbacks extend beyond the farm scale, affecting groundwater levels of both adapting and non-adapting farmers in the area. Ongoing work explores the spatial distribution of these effects in more detail to better understand the neighbourhood effects for both social and hydrological dynamics. The findings of our study can be used to support strategies that minimise trade-offs between groundwater extremes through both individual and collective adaptation.
How to cite: De Graaff, L., Mazzoleni, M., Wens, M. L. K., Brauer, C. C., and Van Loon, A. F.: Exploring neighbourhood effects of farm-level drought adaptation on groundwater extremes with a coupled agent-based and hydrological model , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6318, https://doi.org/10.5194/egusphere-egu26-6318, 2026.
In recent years, the Netherlands has experienced extreme climatic events, including droughts in 2018 and 2022 and record-breaking wet years, such as 2024. These events are prompting a paradigm shift among water managers and users from rapidly draining water to holding it when possible to mitigate dry years, while maintaining the capacity to deal with floods. This research aims to examine how water users' decisions to adapt to drought can influence the water system, and vice versa, while accounting for trade-offs with flood risk. We address this research question using an agent-based model (ABM) based on a small agricultural catchment (Hupsel, ~1,400 ha) in which dairy farming is the predominant land use. The ABM has two main components: 1) the hydrological system; and 2) the human decision-making system. The hydrological system focuses on shallow groundwater and surface water, represented by a MODFLOW model that includes drainage and surface water networks, a single-layer sandy aquifer, and different land use types via the unsaturated zone flow (UZF) package. In the human decision-making system, farmers can decide among different adaptation options to drought based on the protection motivation theory: 1) adopt groundwater irrigation; 2) retain water in ditches to enhance recharge; 3) remove drains or ditches to enhance recharge further; and 4) change crops to less water-demanding ones. Results focused on irrigation indicate that consecutive years of drought lead to higher irrigation adoption, which, in turn, depletes the aquifer and makes the water system more sensitive to dry spells. ABMs are a valuable tool to explore the feedback between humans and the water system in a spatially explicit way, moving beyond the usual representation of anthropogenic interventions as model boundary conditions.
How to cite: Henao Casas, J. D., De Graaff, L., Van Huijgevoort, M., Van Der Velde, Y., and Van Loon, A.: Exploring human-groundwater feedbacks in the Dutch agricultural context under climate extremes using an agent-based model, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6478, https://doi.org/10.5194/egusphere-egu26-6478, 2026.
