HS2.4.4

The predicted increase in drought occurrence and intensity will pose serious threats to global future water and food security. This was hinted by several historically unprecedented droughts over the last two decades, taking place in Europe, Australia, Amazonia or the USA. It has been hypothesised that the strength of these events responded to self-reinforcement processes related to land–atmospheric feedbacks: as rainfall deficits dry out soil and vegetation, the evaporation of land water is reduced, then the local air becomes too dry to yield rainfall, which further enhances drought conditions. Despite the 'local' nature of these feedbacks, their consequences can be remote, as downwind regions may rely on evaporated water transported by winds from drought-affected locations. Following this rationale, droughts may not only self-reinforce locally, due to land atmospheric feedbacks, but self-propagate in the downwind direction, always conditioned on atmospheric circulation. This propagation is not only meteorological but relies on soil moisture drought, and may lead to a downwind cascading of impacts on water resources. However, a global capacity to observe these processes is lacking, and thus our knowledge of how droughts start and evolve, and how this may change as climate changes, remains limited. Furthermore, climate and forecast models are still immature when it comes to representing the influences of land on rainfall.

Here, the largest global drought events are studied to unravel the role of land–atmosphere feedbacks during the spatiotemporal propagation of these events. We based our study on satellite and reanalysis records of soil moisture, evaporation, air humidity, winds and precipitation, in combination with a Lagrangian framework that can map water vapor trajectories and explore multi-dimensional feedbacks. We estimate the reduction in precipitation in the direction of drought propagation that is caused by the upwind soil moisture drought, and isolate this effect from the influence of potential evaporation and circulation changes. By doing so, the downwind lack of precipitation caused by upwind soil drought via water vapor deficits, and hence the impact of drought self-propagation, is determined. We show that droughts occurring in dryland regions are particularly prone to self-propagate, as evaporation there tends to respond strongly to enhanced soil stress and precipitation is frequently convective. This kind of knowledge may be used to improve climate and forecast models and can be exploited to develop geo-engineering mitigation strategies to help prevent drought events from aggravating during their early stages.

How to cite: Miralles, D. G., Schumacher, D. L., Keune, J., and Dirmeyer, P. A.: Drought spatiotemporal propagation via land feedbacks, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1505, https://doi.org/10.5194/egusphere-egu21-1505, 2021.

Recent droughts have shown that national water systems are facing increasing challenges over the last few years. As such, the Netherlands has seen increasing needs to adapt their water management in order to improve their preparedness for current and future drought events. Ideally, the necessary information needed for operational water management decisions should be readily available ahead in time and/or computed in a flexible and efficient way to ensure the various management actions. In this study we show that in addition to the physically based hydrological models, the upcoming and promising trend of incorporating machine learning (ML) in hydrology can increase the information produced to support national and operational water management.

To investigate the potential of ML for this case, we assessed 5 different ML methods to predict the following hydrological variables relevant for water management at a national scale: timeseries of discharge, groundwater levels, surface water levels and surface water temperatures. We developed a unified workflow for all the methods and variables of interest. As inputs, we only used a limited set of hydro-meteorological variables and general water management policies that are readily available on a daily basis and that can be used when the ML methods are used in seasonal forecasting mode.

We show that all methods have a good performance, with a normalized RMSE ranging between 0.0 and 0.4, and Random Forest outperforming other methods. This performance remains stable for low flows, where we observe that complex ML methods outperform simpler algorithms. The addition of water management in the ML routine increases overall performance, although limited. Finally, we observe that locations further upstream show a better performance due to the limited water management influence and close proximity to input observations.

Our study shows that ML has potential in predicting different hydrological variables at various locations at a national scale with only a simple input data set of 5 meteorological and hydrological variables. We additionally were able to capture and incorporate water management information in our analysis, creating a base for future experiments where a combination of seasonal forecasting and scenario analysis might reveal ML-informed mitigation strategies. As such, our approach may improve the preparedness of the national water system of the Netherlands for future drought events.

How to cite: Hauswirth, S. M., Bierkens, M., Beijk, V., and Wanders, N.: The potential of data driven approaches for quantifying hydrological extremes, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2624, https://doi.org/10.5194/egusphere-egu21-2624, 2021.

In the context of climate change, it is important to understand whether drought conditions over the growing season of agricultural crops have changed over the past decades. Common drought metrics used for such assessments compare hydrometeorological anomalies using a static time window. However, the growing season varies among crops as well as in space; driven by climatic differences, and time; driven by e.g. changes in climate or crop-genotypes. Focusing on Southwestern Germany, we aim to investigate how the ranking of drought years varies between crops as well as among static and spatiotemporally varying growing season scenarios. First, we derived annual information on the timing of different phenological phases of two crops, winter wheat and maize, resp. early and late covering, from observations available from the German Weather Services. We then interpolated the timing of these phenological phases to 1 km resolution grids covering all agricultural areas in the study region, using static and spatiotemporally varying interpolation scenarios. Following, we extracted climatological timeseries for all agricultural grid cells and used those to simulate the climatic water balance as well as soil moisture for each grid cell with the hydrological model TRAIN. Finally, we derived for each year different drought metrics, i.e. anomalies in precipitation, temperature, climatic water balance and minimum soil moisture, and correlated those with crop yield anomalies. Results revealed distinct differences in the start and end of the growing season among considered crops. Further, the timing of different phenological phases varied by over a month in both space and time. During the most prominent drought years (2003, 2015, 2018), the growing season of both crops was particularly dry, independent on whether a fixed or variable growing season was considered. On the other hand, there were also some crop specific drought years, e.g., 1991 for maize or 2008 for winter wheat. The difference in hydrometeorological anomalies derived for static and variable growing seasons mainly relates to differences in temperature, but also affected the ranking of some drought years according to other hydrometeorological variables. More apparent were differences between drought metrics, e.g. between the climatic water balance and minimum soil moisture. From these metrics, especially minimum soil moisture correlated well with maize yields, whereas correlations with winter wheat were generally weak for all metrics. To conclude, crop specific agricultural drought assessments could benefit from a crop-relevant growing season specific definition of drought.

How to cite: Tijdeman, E. and Menzel, L.: Crop specific assessment of droughts in the growing season considering the spatiotemporal variability in phenology, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14740, https://doi.org/10.5194/egusphere-egu21-14740, 2021.

In 2018, large areas of central and northern Europe were affected by an extreme drought. The water deficit propagated through the hydrologic cycle causing precipitation, soil moisture and, towards the end of 2018, streamflow and groundwater deficits. In Germany many socio-economic sectors were severely affected by the drought, e.g. the forestry sector has still not recovered. Main drivers for drought propagation are precipitation deficits. However, the natural variability of dry and wet precipitation patterns over time and space make characterization of droughts and predictions of impacts still challenging.

This study investigates German meteorological drought characteristics within general wet and dry spells since 1901 using station based daily precipitation data. Daily, monthly and seasonal aggregated indices such as the Standardized Precipitation Index (SPI) were used to characterize duration, severity and spatial extent of the 2018 drought. These characteristics were then compared with events of extreme droughts since 1901. Even though the meteorological drought of 2018 was extreme considering only precipitation data, we found comparable extremes in the past, for instance 1949 or 1964. However, based on what we observe in the SPI-12, clusters of extreme dry years in the 20th century were often followed by clusters of above average wet years, probably leading to a reduction of impacts in the following years. Since 2003, however, dry patterns predominate. Even though annual precipitation amounts are predicted to increase slightly in the study region this analysis shows the importance of analyzing sub annual as well as multi-year characteristics of precipitation patterns.

Including both wet and dry conditions when characterizing the severity of current drought events may improve our understanding of extreme meteorological drought events causing severe and long lasting impacts.

How to cite: Erfurt, M., Glaser, R., and Stahl, K.: Characterizing Germany’s 2018 drought in the context of wet and dry spells since 1901, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1037, https://doi.org/10.5194/egusphere-egu21-1037, 2021.

The 2018-2019 drought in northwestern Europe caused severe damage to a wide range of sectors, and has made clear that even in temperate-climate countries adaptations are needed to cope with increasing future drought frequencies. A crucial component of drought strategies is to monitor the status of groundwater resources. However, providing up-to-date assessments of regional groundwater drought development remains challenging due to the limited quality of available data. We set up a time series modelling-based method for data preparation to map the spatiotemporal development of the 2018-2019 groundwater drought in the southeastern Netherlands, based on a large amount of monitoring data. The data preparation method was evaluated for its usefulness and reliability for groundwater drought studies and prediction. The analysis showed that the 2018-2019 meteorological drought caused extreme groundwater drought throughout the southeastern Netherlands, breaking 30-year records almost everywhere. Drought onset and duration were strongly variable in space. Groundwater drought development appeared to be governed dominantly by the spatial distribution of rainfall and the geological-topographic setting. The time series modelling-based data preparation method was found a useful tool for the given situation to enable a detailed, consistent record of groundwater drought development. The time series simulations were generally found to be reliable; however, the use of time series simulations rather than direct measurement series may bias drought estimations especially at a local scale, and underestimate spatial variability. Finally, time series modelling was also found a promising tool for regional-scale drought nowcasting and prediction. Further development of time-series based validation and simulation methods, combined with accessible and consistent monitoring data, will be valuable to enable better groundwater drought monitoring in the future. 

How to cite: Brakkee, E., van Huijgevoort, M., and Bartholomeus, R.: Mapping the spatiotemporal development of groundwater drought from data: the 2018-2019 drought in the Netherlands, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-473, https://doi.org/10.5194/egusphere-egu21-473, 2021.

In the environment of the changing climate in Central Europe, the seasonality and magnitude of low flow events and hydrological droughts are projected to change in the near future. Ongoing increases in the air temperature, rates of evaporation and decreasing snow cover will significantly affect the summer deficit volumes even in the rivers of humid montane and highland areas in mid-latitudes. However, what if the significant changes have already been happening during the last decades? Therefore, this research is focused on analysis of the variability and seasonality of low flow events and hydrological drought events in fifteen near-natural catchments along the Czech–German and Czech–Polish national borders. To quantify the low flow regime changes of the study regions in the last 52 years (1968–2019), we applied tools from the R package lfstat. The 30-year moving averages of seasonality ratio (SR) and the seasonality index (SI) were derived to address the degree of change in each catchment. Moreover, the 7-day and 30-day mean summer minimum discharges were computed, as well as the streamflow deficit volumes for every episode of hydrological drought. The results showed a continual increase in the proportion of summer low flow and drought events during the study period along with a significant shift in the average date of low flow occurrence towards the beginning of the year. The most marked shifts in low flow seasonality were found mainly in catchments with the average altitude 800–1000 m a. s. l. Conversely, the low flow regime in catchments above 1000 m a. s. l. and also in the catchments below 800 m a. s. l. remained nearly stable throughout the 1968–2019 period. Moreover, the analysis of 7- and 30-day mean summer minimum discharges indicated a much-diversified pattern in the behavior of long-term trends than it was expected.

How to cite: Vlach, V., Ledvinka, O., and Matouskova, M.: Changing low flow seasonality in Central European headwaters, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-71, https://doi.org/10.5194/egusphere-egu21-71, 2021.

Several approaches to identify hydrological drought exist, which result in differences in drought frequency, timing, duration, and deficit volume (drought characteristics) using the same hydrometeorological data as input. This has created confusion within the hydro-meteorological community, as well as in operational water management services on the difference in drought characteristics obtained with the different approaches. The aim of this study, therefore, is to provide a comprehensive overview of the differences of hydrological drought, i.e. streamflow drought, using different identification approaches for the pan-European river network (>10,000 river grid cells). Time series of daily streamflow data were obtained from the LISFLOOD hydrological model forced with gridded meteorological observations from 1990 to 2018. Streamflow droughts were detected using the daily and monthly Variable Threshold methods (VTD and VTM), daily and monthly Fixed Threshold methods (FTD and FTM), and the Standardized Streamflow Index with 1-month accumulation period (SSI-1). For the threshold methods the Q80 (flow that is equaled or exceeded 80 percent of the time) is applied, whereas for the SSI a threshold of about -1 is used. We applied a centered 30-day moving average (30DMA) smoothing technique to the daily flow data to reduce the number of minor droughts. This is the first study that compares all these drought identification approaches in such a systematic way at this large scale. Our results (pan-European maps, tables) clearly show that characteristics of streamflow droughts derived with different approaches deviate, partly associated with different climate regions across Europe. The daily threshold methods (VTD and FTD) identify twice as much drought events than the monthly threshold methods (VTM and FTM) due to the daily resolution and minor droughts, even with smoothing. Average duration of FT droughts is longer than VT droughts. In addition, FT droughts have higher drought deficit volumes than VT droughts (~ 30-60%, dependent on climate region), whereas using monthly data (VTM and FTM) result in higher deficits (~10-60%) than daily data (VTD and FTD). In northern and central European regions (Köppen- Geiger Dfb, Dfc and ET climates), the variable threshold methods (VTD and VTM) generally detect drought earlier (March-July) than the fixed thresholds (FTD and FTM) (July-October). In the western European regions and the Mediterranean differences in timing among identification approaches are not so clear. The characteristics of SSI-1 drought, in general, are close to what is being identified with the VTM approach. Differences in drought characteristics highlight the importance of whether end-users should take seasonality into account or not (VT and SSI-1 versus FT) and consider temporal variability (daily versus monthly). Certainly, there is no unique hydrological drought definition that fits all purposes; hence we suggest that users should clearly agree among themselves upon a sharp definition on which type of streamflow drought is required to be identified for a specific application.

How to cite: Van Lanen, H. A. J. and Sutanto, S. J.: Implication of multi drought definitions to identify streamflow drought across Europe, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4063, https://doi.org/10.5194/egusphere-egu21-4063, 2021.

Drought is considered one of the costliest climate extremes that have wide impacts on humans and ecosystems. Understanding different drought stages, for example, onset, propagation, and its recovery, especially for tropical (the vulnerable region in Earth’s climate system) catchments are crucial for ecosystem sustainability and food security. Utilizing gauge-based quality-controlled daily streamflow records from 98 catchments of rain-fed Peninsular River Basins (PRB) in India, here we investigate different phases of hydrological droughts in a multi-stage framework. While several studies so far have investigated the propagation of hydrological droughts at a monthly resolution, a credible understanding of drought dynamics requires analyzing low-flow series at a higher spatial and temporal resolution, ensuring the issuance of timely alerts related to regional water scarcity.  Owing to high seasonality in the daily streamflow records, a variable threshold approach is adopted to delineate streamflow-based drought events. To assess the temporal evolution of droughts, the events are categorized into various inter-related phases, i.e., growth, persistence, and recovery stage over the study period 1965 – 2018. For most of the gauges, the mean timing of drought onset mostly lies between August and September revealing failure of monsoon as the primary causal factor for drought development in peninsular catchments. Furthering this, we identify four distinct hydrological drought regimes, which includes, Regime 1: persistent droughts with longer duration and moderate deficit volume with average termination during mid-monsoon (in September). These gauges are mostly situated in Central India and typically show a longer recovery time coincided with shorter return times (i.e., the time between two consecutive drought events), making it one of the most vulnerable regions in PRB; Regime 2: droughts with a shorter duration, least deficit volume with average termination in October, the post-monsoon period. These gauges are located in the western part of the country; Regime 3: droughts with the highest variability in drought deficit volume with the largest subsurface contribution from groundwater recharge. These sites are primarily located in eastern India and do not show any specific trend in the termination period; Regime 4: droughts with least regularity in drought termination with the average termination month clustered around November. These gauges are mostly concentrated in the southwestern part of the country. Our findings add value to the systematic understanding of hydrological drought propagations in rain-fed catchments, which serves as a basis for exploring future changes in droughts under concurrent shifts in rainfall and temperature extremes in a warming climate. 

How to cite: Ganguli, P., Singh, B., and Raut, A.: Spatiotemporal Clustering of Hydrological Droughts in Peninsular India, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-87, https://doi.org/10.5194/egusphere-egu21-87, 2021.

Assessment of flood and drought risks, and changes to these risks under climate change, is a critical issue worldwide. Statistical methods are commonly used in data-rich regions to estimate the magnitudes of river floods of specified return period at ungauged sites. However, data availability can be a major constraint on reliable estimation of flood and drought magnitudes, particularly in the Global South. Statistical flood and drought magnitude estimation methods rely on the availability of sufficiently long data records from sites that are representative of the hydrological region of interest. In the Philippines, although over 1000 locations have been identified where flow records have been collected at some time, very few records exist of over 20 years duration and only a limited number of sites are currently being gauged. We collated data from three archival sources: (1) Division of Irrigation, Surface Water Supply (SWS) (1908-22; 257 sites in total); (2) Japan International Cooperation Agency (JICA) (1955-91; 90 sites); and, (3) Bureau of Research and Standards (BRS) (1957-2018; 181 sites). From these data sets, 176 contained sufficiently long and high quality records to be analysed. Series of annual maximum floods were fit using L-moments with Weibull, Log-Pearson Type III and Generalised Logistic Distributions, the best-fit of these being used to estimate 2-, 10- and 100-year flood events, Q₂, Q₁₀ and Q₁₀₀. Predictive equations were developed using catchment area, several measures of annual and extreme precipitation, catchment geometry and land-use. Analysis took place nationally, and also for groups of hydrologically similar regions, based on similar flood growth curve shapes, across the Philippines. Overall, the best fit equations use a combination of two predictor variables, catchment area and the median annual maximum daily rainfall. The national equations have R² of 0.55-0.65, being higher for shorter return periods, and regional groupings R² are 0.60-0.77 for Q₁₀. These coefficients of determination, R², are lower than in some comprehensive studies worldwide reflecting in part the short individual flow records. Standard errors of residuals for the equations are between 0.19 and 0.51 (log₁₀ units), which lead to significant uncertainty in flood estimation for water resource and flood risk management purposes. Improving the predictions requires further analysis of hydrograph shape across the different climate types, defined by seasonal rainfall distributions, in the Philippines and between catchments of different size. The results here represent the most comprehensive study to date of flood magnitudes in the Philippines and are being incorporated into guidance for river managers alongside new assessments of river channel change across the country. The analysis illustrates the potential, and the limitations, for combining information from multiple data sources and short individual records to generate reliable estimates of flow extremes.

How to cite: Hoey, T., Tolentino, P., Guardian, E., Williams, R., Boothroyd, R., David, C. P., and Paringit, E.: Flood estimation for ungauged catchments in the Philippines using multiple archival data records, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4905, https://doi.org/10.5194/egusphere-egu21-4905, 2021.

How to cite: Lang, M., Renard, B., and Le Coz, J.: Hydrological consistency between the upstream and downstream estimates of Q1000 flood on the upper Rhine River, using historical series in Basel (1808-2017) and Maxau (1815-2018), EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8697, https://doi.org/10.5194/egusphere-egu21-8697, 2021.

Climate change has led to increased droughts and floods over mainland Australia, resulting in water scarcity, excessive surplus and socioeconomic losses. Therefore, it is of great significance to comprehensively evaluate droughts and floods from the meteorological and hydrological perspective. Firstly, we determine the Standard Precipitation and Evapotranspiration Index (SPEI) by correlation analysis to represent the meteorological conditions. To characterize the hydrological conditions, we calculate the hydrological drought indices including Standard Runoff Index (SRI), Soil Moisture Deficit Index (SMDI), and Total Storage Deficit Index (TSDI), using the runoff and soil moisture data from the Global Land Data Assimilation System (GLDAS) and the Terrestrial Water Storage Change (TWSC) data from Gravity Recovery And Climate Experiment (GRACE) respectively. Results show that the most severe hydrological drought over mainland Australia during the study period occurred from May 2006 to Jan. 2009 with the drought severity of -58.28 (cm months) and the most severe flood from Jun. 2010 to Jan. 2013 is with the severity of 151.36 (cm months). The comprehensive analysis of both meteorological and hydrological drought indices shows that both meteorological and hydrological drought indices can effectively detect the droughts and floods over mainland Australia. Moreover, the meteorological drought and flood are of higher frequency, while hydrological drought and flood have a relatively longer duration. Based on the cross-correlation analysis, we find that the SPEI can firstly reflect the droughts or floods over mainland Australia, and then the SRI, SMDI and TSDI reflect with the time lags of one, three and six months respectively. Furthermore, we calculate the frequency of drought and flood at the basin scale and find that SPEI and SMDI are equally sensitive to drought and flood, while TSDI is more sensitive to flood than drought. This study reveals the relationship between meteorological and hydrological conditions in mainland Australia in the last two decades and highlights its intensifying extreme climate conditions under the circumstances of the increasing temperature and complex changing precipitation.

How to cite: Wang, W., Shen, Y., Wang, F., and Li, W.: Assessment of Meteorological and Hydrological Droughts and Floods over mainland Australia based on Drought Indices, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8147, https://doi.org/10.5194/egusphere-egu21-8147, 2021.

The drought and floods are a natural phenomenon of ecosystems. Many studies found that the frequency and intensity of individual events of floods and drought are increasing in recent decades due to climate change. However, it is still unclear whether the frequency of combined flood-drought events is increasing in the same year or not under the climate change scenario. To identify drought and flood characteristics, we used the Standardized Weighted Average of Precipitation (SWAP), and copula bivariate distribution concept to estimate the joint probabilities of combined flood-drought events of the same year. We utilized gridded rainfall data from the India Meteorological Department at 0.25 degree for the present study. We estimated drought, flood and combined flood-drought events for the base period (1901-1930) and the current period (1991-2018). The analysis demonstrates that about 51.97% of the total grid points show an increasing monthly SWAP values trend in the summer monsoon season. However, in winter, only 15.55% of the total grid points show an increase in the trend of monthly SWAP values. The univariate flood and drought analysis revealed that 83.98%, 83.98% and 81.90% of total grids show a significant percentage change of drought at 5, 10 and 25-year return periods, respectively when the current period is compared with the base period. Still, only, 27.88%, 16.32% and 13.82% of the total grids show a significant change in the flood 5-year, 10-year and 25-year return periods, respectively. We also found that combined flood-drought events' frequency increases in 39.21%, 36.49% and 20.71% of total grid points corresponding to 5, 10 and 25-year return period values, respectively. This study concluded that less intensity drought, flood, and combined flood-drought events are increasing in more grid points. The study outcomes will help the decision-makers to make efficient decision to overcome the impacts of the hydroclimatic extremes.

How to cite: Chetan Kumar, S., Gupta, V., and Jain, M. K.: Impact of Climate Change on Combined Flood and Drought events in India, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9718, https://doi.org/10.5194/egusphere-egu21-9718, 2021.

Estimating design floods at location with no measurements or short records is a major challenge for operational hydrology. The aims of this study are to (i) develop a regional flood frequency model that consists of a regression model for the index flood and the parameters in the growth curve; (ii) assess and attribute the uncertainty to the components of the regional flood frequency model, (iii) develop flexible approaches for combining a regional flood frequency model with local data and provide recommendations for how to combine local and regional data. Annual maximum flood data from 529 gauging stations were used for the model development. We re-parametrized the Generalized Extreme Value (GEV) distribution into an index flood component and growth curve component, and we used the median flood as the index flood. The model was estimated using a MCMC algorithm within a Bayesian framework. The Bayesian approach was also used to combine local and regional data. Two approaches were used (i) combining local and regional data to estimate the index flood (ii) combining local and regional data to estimate both the index flood and the growth curve. Simulation experiments were carried out to assess the performance of these approaches. We see that in particular for data records shorter than 10 years, we can benefit from combining the local and the regional model by both approaches. We also constructed a prior for use in local analysis that complied with the distribution of the regional model for three key quantiles.  For the index flood, the regression model was successfully estimated and evaluated using a three-step cross validation approach. The most important variables for predicting the index flood were mean annual runoff, river length and lake percentage. The attribution of uncertainty showed that most of the uncertainty was found in the index flood component.

How to cite: Engeland, K., Reitan, T., Stenius, S. M., and Glad, P.: Design flood estimation at locations with no data or short records in a Bayesian framework, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13317, https://doi.org/10.5194/egusphere-egu21-13317, 2021.

Progressing towards a sustainable society implies the availability of reliable boundary conditions for various hydrodynamic flood models, including an extensive consideration of uncertainties. With an ever growing availability of data and models, the uncertainty sources are constantly increasing. Hence, an elaborate uncertainty analysis strategy has become a must. One way to deal with part of this uncertainty is by applying an ensemble approach, using different hydrological models in combination with various climate scenarios. However, impact modellers may find the growing number and the increasing length of input series for hydraulic models more challenging, since computing time, reliability of the analysis and project deadlines can cause a conflicting situation. In this context, there is a need for approaches that offer a compromise between computing the vast amount of long input series and adequately addressing the uncertainties within a reasonable time span. We present an approach which reduces the computation time, but simultaneously recognises the importance of robust results and the consideration of the different sources of uncertainty. By a stratification of the probability domain for extreme events (discharges, water levels,…) a set of hydrodynamic boundary conditions is generated. Each of these synthetic events gets a probability of occurrence, which changes according to either the considered confidence level or the considered ensemble member. In addition to the stratification approach, a tool for selecting synthetic events for design is developed. This tool allows end-users to create a subset of synthetic events which can be used as design events for a specific area and are representative for the full set of events. The approach is demonstrated for the River Dender catchment in Flanders using 40 years of hydro-meteorological data, an ensemble of 3 hydrological models and a detailed hydraulic model.

How to cite: Leyssen, G., van Uytven, E., Blanckaert, J., Adams, R., Franken, T., Nossent, J., and Pereira, F.: How to embrace big data and uncertainties within reasonable time constraints? A detailed flood study in Flanders, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16238, https://doi.org/10.5194/egusphere-egu21-16238, 2021.

The African continent is severely impacted by floods, with an increasing vulnerability to these events in the most recent decades. Our improved preparation against and response to this hazard would benefit from an enhanced understanding of the physical processes at play. A database recently compiled on Africa allows to conduct studies at the continental scale: the African Database of Hydrometric Indices (ADHI: https://doi.org/10.5194/essd-2020-281). Daily river discharge data have been extracted for 399 African rivers to analyze the seasonality of observed annual maximum discharge. In addition, extreme precipitation from CHIRPS and ERA5, and soil moisture from ERA5-Land between 1981 and 2018 have been also considered as potential flood drivers. The database includes a total of 11,302 flood events, covering most African regions. The analysis is based on directional statistics to compare the annual maximum river discharge with annual maximum rainfall and soil moisture. Results show that the annual peak flow in most areas is more strongly associated with the annual peak of soil moisture than of extreme precipitation. In addition, the interannual variability of flood magnitudes is better explained by the variability of annual maximum soil moisture or the precipitation summed over 5 days prior to an event, than by changes in the annual maximum daily precipitation. These results have important implications for the design of efficient flood forecasting systems or the investigation of the long-term evolution of these hydrological hazards.

How to cite: Tramblay, Y., Villarini, G., El Khalki, E. M., Gründemann, G., and Hughes, D.: On the influence of extreme rainfall and antecedent soil moisture on flood hazards in Africa, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2908, https://doi.org/10.5194/egusphere-egu21-2908, 2021.

Information on flood seasonality is required in many practical applications of hydrology and water resources management. However, an understanding of flood seasonality and how flood frequencies may have changed over time has not been established for the Congo Basin. The main objective of this study is therefore to identify flood seasonality and change in frequency the Congo Basin (CB). The analysis based on six major drainage areas of the CB, where we used a Peaks Over Threshold (POT) flood time-series with three peaks per year. The relative frequency method is applied to identify flood seasons, and then a cluster analysis is performed to classify flood into type based on monthly frequency. The directional statistics method is used to determine the mean day of flood and the flood variability measure. To identify flood frequency changes, the analysis of variance was applied to test the difference between two flood frequency time series blocks before and after the change point year. Results show that four gauging stations exhibit a unimodal flood seasonality distribution while two gauging stations have bimodal flood seasonality distribution, and two significant flood rich months are observed in all studied gauging stations.  The cluster analysis results in four spatially flood types with distinct seasonality characteristics. Mean flood dates show that the time interval between adjacent flood events in the south and south-east is shorter compared to time interval between flood events in the north and north-west. It is observed that, in almost all gauging stations, there is strong flood seasonality, and the geographical location of a watershed is indicative of its flood pattern. Most of significant decreasing frequencies are found in the southern part of the Congo Basin. There are no significant changes in flood frequency identified in the northern and eastern part of the Basin. However, flood frequencies have been increasing in the centre and western part of the Basin. This study suggest that, exploring flood generating factors and the drivers of change can provide insights for understanding the influence of these factors on floods as climate models projected changes in extreme precipitation and aridity in the future.  

How to cite: Gode, B. B., Jeff, N., Raphael, T., Lauwrence, H., Mark A, T., Vincent, L. M., and Paul, P. B.: Flood Seasonality in the Congo River Basin, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-64, https://doi.org/10.5194/egusphere-egu21-64, 2021.

How should design flood magnitudes be estimated under climate change? Apart from assuming stationarity, the two main approaches are hydrologic simulation and informed-parameter, which is generally based on either trend or climate covariates. Here, we compare these approaches across a large set of hydro-climatologically diverse basins located throughout the contiguous United States, splitting the historic record into a calibration and validation time period. We evaluate performance when the approaches are forced with observed climate as well as simulated climate from general circulation models. We also investigate how the strengths of the climate informed and hydrologic simulation approaches can be combined to improve projections; here, we use the outputs of hydrologic simulation as covariates in the climate informed approach. The results provide a quantitative perspective on key long-term flood projection issues and provide a route forward to improving projections given the identified strengths and weaknesses of each approach.

How to cite: Schlef, K., François, B., and Brown, C.: Comparing and Advancing Approaches to Long-Term Flood Projection, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3351, https://doi.org/10.5194/egusphere-egu21-3351, 2021.

Within risk modelling, event ‘footprints’ are used to demonstrate how an extreme event impacts different locations at a similar time. Currently, estimates of future impacts from extreme events are derived by applying climate change allowances to at-site flood frequency estimates based on observations from the current period. These modified flow frequency estimates are then used to calculate flood risk and associated losses using a variety of means.

The present work brings together these two strands to develop spatially resolved projections of changes in river flow and, together with new analyses of the spatial coherence, to generate a wider collection of plausible events to improve risk modelling of the rarest events. This wide collection of extreme flood events provides the foundational input for an event-based assessment of risk.

The research extends proven methods to generate extreme, widespread flood events directly based on outputs from a 1km grid-based hydrological model driven by UKCP18 datasets. These modelled events provide coherent and highly credible descriptions of changes in flow based on spatially coherent climate change information. In addition to the small number of widespread extreme events generated directly from the gridded hydrological model, copula-based methods have been extended and applied on a regional and even national scale at a 1km resolution over the GB river network. These extensions to the Heffernan-Tawn model and Empirical Copula models are being used to generate a collection of plausible extreme based on the climate of 1980-2010 and on climate projections for 2050-2080. The collection of events is then used to compare the characteristics and variability of widespread events across different climate ensemble members and compare between present and future estimates.

How to cite: Griffin, A., Stewart, L., Kay, A., Bell, V., Sayers, P., and Carr, S.: Simulating spatially coherent widespread flood events for risk modelling in the UK, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9671, https://doi.org/10.5194/egusphere-egu21-9671, 2021.

The rapid increase in heavy precipitation flooding events highlights the need for efficient flood forecasting techniques to facilitate flood hydrological research and effective flood management by civic bodies. The current study aims to develop a near-real-time flood forecasting framework by integrating a 3-way coupled hydrodynamic flood model framework with numerical weather modelling based rainfall forecasts. The proposed framework has been demonstrated over Mumbai city in India, which is subjected to flooding every year during the monsoon months. A fine-resolution atmospheric simulation with the Weather Research and Forecasting (WRF) model has been performed for rainfall forecasts, which serve as an input to the flood model. To access the impact of urbanization on rainfall extremes, three scenarios are considered in the WRF simulations, i.e., WRF model: (1) without Urban canopy model (WRF-NoUCM), (2) coupled with a single-layer Urban canopy model (WRF-SUCM), and (3) coupled with a multi-layer Urban canopy model (WRF-MUCM). Further, a three-way coupled flood model has been developed where the MIKE 11 model (streamflow) with the drainage network (stormwater drains) and the MIKE 21 model (overland flow) have been considered for flood inundation and subsequently hazard mapping. In addition, the tidal elevation is provided along the coastline in the model setup. The flood maps developed by three WRF forecasted rainfall scenarios have been compared with that of the maps developed with observed rainfall. The extent to which the scenarios have been able to imitate the pattern and extent of flooding generated by observed rainfall has been investigated to decide the best scenario to be adapted in the comprehensive flood forecasting network. This state-of-art flood forecasting approach may be implemented in other flood-prone coastal regions as a major non-structural flood management strategy to reduce flood risk and vulnerabilities for the people dwelling in those regions.

How to cite: Ghosh, M., Paul, S., Karmakar, S., and Ghosh, S.: Near-real-time flood forecasting for an urban coastal catchment: An approach in combination of numerical weather and 3-way coupled hydrodynamic flood modelling, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12834, https://doi.org/10.5194/egusphere-egu21-12834, 2021.

This study used the North American Multi-Model Ensemble (NMME) system to understand the role of near surface temperature in the prediction skill for US climate extremes. In this study, the forecasting skill was measured by anomaly correlation coefficient (ACC) between the observed and forecasted precipitation (PREC) or 2-meter air temperature (T2m) over the contiguous United States (CONUS) during 1982–2012. The strength of the PREC-T2m coupling was measured by ACC between observed PREC and T2m or forecasted PREC and T2m over the CONUS. This study also assessed the NMME forecasting skill for the summers of 2004 (spatial anomaly correlation between PREC and T2m: 0.05), 2011 (-0.65), and 2012 (-0.60) when the PREC-T2m coupling is weaker or stronger than the 1982–2012 climatology (ACC:-0.34). This study found that most of the NMME models show stronger (negative) PREC-T2m coupling than the observed coupling, indicating that they fail to reproduce interannual variability of the observed PREC-T2m coupling. Some NMME models with skillful prediction for T2m show the skillful prediction of the precipitation anomalies and US droughts in 2011 and 2012 via strong PREC-T2m coupling despite the fact that the forecasting skill is year-dependent and model-dependent. Lastly, we explored how the forecasting skill for SSTs over north Pacific and Atlantic Oceans affects the forecasting skill for T2m and PREC over the US. The findings of this study suggest a need for the selective use of the current NMME seasonal forecasts for US droughts and pluvials.

How to cite: Kam, J., Kim, S., and Roundy, J.: NMME-based Assessment of Prediction Skills of US Summertime Droughts and Pluvials: Role of Near-surface Temperature Prediction Skill, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1957, https://doi.org/10.5194/egusphere-egu21-1957, 2021.

Spatially extensive multi-year hydrological droughts threaten water resources availability and incur significant environmental and socio-economic consequences. Given the impacts of climate change, the UK is expected to remain vulnerable to future multi-year droughts. Existing approaches to quantify hydrological impacts of climate change are often scenario-driven and may miss out plausible outcomes with significant impacts. Event-based storyline approaches aim to quantify “storylines” of how a singular event with significant impacts could hypothetically have unfolded in alternative ways from plausible changes to its causal factors under present and future climate. This study uses the 2010-2012 UK drought, the most recent period of severe hydrological drought, as a basis, to create counterfactual storylines based on changes to 1) precondition severity, 2) temporal drought sequence and 3) climate change. Model simulations are performed using the GR4J hydrological model and drought characteristics for each counterfactual storyline is calculated using the Standardized Streamflow Index at multiple accumulation periods.

The storylines show that maximum intensity, mean deficit and duration of the 2010-2012 drought were highly conditioned by its meteorological preconditions. Recovery time from progressively drier preconditions reflect both spatial variation in drought characteristics and the influence of physical catchment characteristics, particularly hydrogeology, in the propagation of multi-year droughts. Plausible storylines of an additional dry year with dry winter conditions repeated before the observed drought or replacing the observed dramatic drought termination confirm the vulnerability of UK catchments to a “three dry winter” scenario. Application of the UKCP18 projections at four global warming levels explore the impacts of the drought in a warmer world. Drought conditions of the storylines could have matched and exceeded that experienced in past severe droughts, especially for southern catchments. The construction of storylines based on observed events can complement existing methods to stress test UK catchments against plausible unrealized droughts.

How to cite: Chan, W., Shepherd, T., Smith, K., Darch, G., and Arnell, N.: Storylines of UK drought based on the 2010-2012 event, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1544, https://doi.org/10.5194/egusphere-egu21-1544, 2021.

In this contribution future changes of surface water availability over the Austrian domain is investigated. We use an ensemble of downscaled and bias-corrected regional climate model simulations of the EURO-CORDEX initiative under moderate mitigation (RCP4.5) and Paris agreement (RCP2.6) emission scenarios. The climatic water balance and its components (rainfall, snow melt, glacier melt and potential evapotranspiration) are used as indicators for surface water availability and we focus on different altitudinal classes (lowland, mountainous and high alpine) to depict a variety of processes in complex terrain. Apart from analysing the mean changes of these quantities we also pursue a hazard risk approach by estimating changes in return periods of drought events of a given magnitude as observed in the reference period. The results show in general wetter conditions over the course of the 21^st century over Austria. Considering seasonal differences, winter and spring will be getting wetter due to an increase in precipitation along with a higher rainfall/snowfall fraction as a consequence of rising temperatures. In summer only little changes in the ensemble median of the climatic water balance are visible, hence uncertainties are large due to a considerable ensemble spread. However, by analysing changes in return periods of drought events, a robust signal of increasing risk of moderate and extreme drought events during summer is apparent. It emerges from an increase in interannual variability of the climatic water balance, which likely stems from intensified land-atmosphere coupling under climate change sustaining and intensifying spring preconditions towards even wetter or dryer summers.

How to cite: Haslinger, K., Laaha, G., Schöner, W., Konrad, A., Olefs, M., Koch, R., and Abermann, J.: Elevation-dependent drying signals under future climate change – a case study for Austria, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6533, https://doi.org/10.5194/egusphere-egu21-6533, 2021.

Droughts are anticipated to intensify or become more frequent in many parts of the world due to climate change. However, the issue of drought definition, namely the diversity of drought definition, makes it difficult to compare drought projections and hampers overviewing future changes in drought. This issue is widely known and underscored in recent reports of the Intergovernmental Panel on Climate Change, but the relative importance of the issue and its spatial distribution have never been quantitatively evaluated compared to other sources of uncertainty.

Here, using a multi-scenario and multi-model dataset with combinations of three climate change scenarios, four global climate models and seven global water models, we evaluated changes in the frequency of three categories of drought (meteorological, agricultural, and hydrological droughts) by a consistent standardized approach with four different temporal scales of accumulation periods to show how differences among the drought definitions could result in critical uncertainties. For simplicity, this study focuses on one drought index per drought category. Firstly we investigated the disagreement in the sign of changes between definitions, and then we decomposed the overall uncertainty to estimate the relative importance of each source of uncertainty. By a multifactorial ANOVA, uncertainty was decomposed into four main factors, namely drought definitions, climate change scenarios, global climate models and global water impact models, and their interactions.

Our results highlight specific regions where the sign of change disagrees between drought definitions. Importantly, changes in drought frequency in such regions tended to be statistically insignificant with low ensemble member agreement. Drought definition attributed to18% of the main factor uncertainty at the global scale, and the definition was the dominant uncertainty source over 11% of the global land area. The contribution of difference in the drought category showed a higher contribution to overall uncertainty than the difference in scales. The contribution of scenario uncertainty was the least among the main factors in general, though it is a dominant factor in the far-future in a couple of hotspot regions such as the Mediterranean region. Overall, model uncertainties were the primary source of uncertainty, and the definition issue was less important over large areas. However, definition uncertainty was the primal uncertainty source with significant changes in particular regions, such as parts of high-latitude areas in the northern hemisphere. One needs to pay attention to these regions in overviewing future drought change. Nonetheless, what this study quantified is the relative importance of uncertainty stemming from drought definition that should be avoidable or reducible if one treats drought specifically. Our results indicate that we can reduce uncertainty in drought projections to some extent and get a clearer picture by clarifying hydrological processes or sectors of interest.

How to cite: Satoh, Y., Shiogama, H., Hanasaki, N., Pokhrel, Y., Boulange, J., Burek, P., Gosling, S., Grillakis, M., Koutroulis, A., Schmied, H., Thiery, W., and Yokohata, T.: Decomposing the uncertainties in global drought projection, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6780, https://doi.org/10.5194/egusphere-egu21-6780, 2021.

How to cite: Reyniers, N., Addor, N., Darch, G., He, Y., Zha, Q., and Osborn, T.: Understanding changes in meteorological drought in regional UK Climate Projections (UKCP18), EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8713, https://doi.org/10.5194/egusphere-egu21-8713, 2021.