Search

Accurate river streamflow forecasts are a vital tool in the fields of water security, flood preparation and agriculture, as well as in industry more generally. Traditional physics-based models used to produce streamflow forecasts have become increasingly sophisticated, with forecasts improving accordingly. However, the development of such models is often bound by two soft limits: empiricism – many physical relationships are represented empirical formulae; and data sparsity – long time series of observational data are often required for the calibration of these models. Artificial neural networks have previously been shown to be highly effective at simulating non-linear systems where knowledge of the underlying physical relationships is incomplete. However, they also suffer from issues related to data sparsity. Recently, hybrid forecasting systems, which combine the traditional physics-based approach with statistical forecasting techniques, have been investigated for use in hydrological applications. In this study, we test the efficacy of a type of neural network, the long short-term memory (LSTM), at predicting streamflow at 10 river gauge stations across various climatic regions of the western United States. The LSTM is trained on the catchment-mean meteorological and hydrological variables from the ERA5 and Global Flood Awareness System (GloFAS)–ERA5 reanalyses as well as historical streamflow observations. The performance of these hybrid forecasts is evaluated and compared with the performance of both raw and bias-corrected output from the Copernicus Emergency Management Service (CEMS) physics-based GloFAS. Two periods are considered, a testing phase (June 2019 to June 2020), during which the models were fed with ERA5 data to investigate how well they simulated streamflow at the 10 stations, and an operational phase (September 2020 to October 2021), during which the models were fed forecast variables from the European Centre for Medium-Range Weather Forecasts (ECMWF) Integrated Forecasting System (IFS), to investigate how well they could predict streamflow at lead times of up to 10 d. Implications and potential improvements to this work are discussed. In summary, this is the first time an LSTM has been used in a hybrid system to create a medium-range streamflow forecast, and in beating established physics-based models, shows promise for the future of neural networks in hydrological forecasting.

The European Centre for Medium-Range Weather Forecasts (ECMWF) mission is to deliver high-quality global medium-range numerical weather predictions and monitoring of the Earth system to its member states. The modelling and forecasting of natural hazards are an important part of this mission. Challenges in this area include the integration of innovative observations into the Earth system; realistic representations of water, energy and carbon cycles; coupling and initialisation of all Earth system components; adequate representation of uncertainties; supporting the development of user-specific products to enable optimal decision-making under uncertainties; and advances in software engineering. The new ECMWF strategy identified three pillars to sustain its future development (ECMWF, 2021a): science and technology (world-leading weather and Earth system science, cutting-edge technology and computational science), impact (high-quality products fit for purpose, efficient and easy access to products), and people (inspiring and hiring the best experts). Progress in all these areas will need enhanced collaboration with member states and partners across Europe and beyond.

Estimating how much water is flowing through rivers at the global scale is challenging due to a lack of observations in space and time. A way forward is to optimally combine the global network of earth system observations with advanced numerical weather prediction (NWP) models to generate consistent spatio-temporal maps of land, ocean, and atmospheric variables of interest, which is known as a reanalysis. While the current generation of NWP models output runoff at each grid cell, they currently do not produce river discharge at catchment scales directly and thus have limited utility in hydrological applications such as flood and drought monitoring and forecasting. This is overcome in the Global Flood Awareness System (GloFAS; http://www.globalfloods.eu/, last access: 28 June 2020) by coupling surface and sub-surface runoff from the Hydrology Tiled ECMWF Scheme for Surface Exchanges over Land (HTESSEL) land surface model used within ECMWF's latest global atmospheric reanalysis (ERA5) with the LISFLOOD hydrological and channel routing model. The aim of this paper is to describe and evaluate the GloFAS-ERA5 global river discharge reanalysis dataset launched on 5 November 2019 (version 2.1 release). The river discharge reanalysis is a global gridded dataset with a horizontal resolution of 0.1∘ at a daily time step. An innovative feature is that it is produced in an operational environment so is available to users from 1 January 1979 until near real time (2 to 5 d behind real time). The reanalysis was evaluated against a global network of 1801 daily river discharge observation stations. Results found that the GloFAS-ERA5 reanalysis was skilful against a mean flow benchmark in 86 % of catchments according to the modified Kling–Gupta efficiency skill score, although the strength of skill varied considerably with location. The global median Pearson correlation coefficient was 0.61 with an interquartile range of 0.44 to 0.74. The long-term and operational nature of the GloFAS-ERA5 reanalysis dataset provides a valuable dataset to the user community for applications ranging from monitoring global flood and drought conditions to the identification of hydroclimatic variability and change and as raw input for post-processing and machine learning methods that can add further value. The dataset is openly available from the Copernicus Climate Change Service Climate Data Store: https://cds.climate.copernicus.eu/cdsapp#!/dataset/cems-glofas-historical?tab=overview (last access: 28 June 2020) with the following DOI: https://doi.org/10.24381/cds.a4fdd6b9 (C3S, 2019).

New precipitation (P) datasets are released regularly, following innovations in weather forecasting models, satellite retrieval methods, and multi-source merging techniques. Using the conterminous US as a case study, we evaluated the performance of 26 gridded (sub-)daily P datasets to obtain insight into the merit of these innovations. The evaluation was performed at a daily timescale for the period 2008–2017 using the Kling–Gupta efficiency (KGE), a performance metric combining correlation, bias, and variability. As a reference, we used the high-resolution (4 km) Stage-IV gauge-radar P dataset. Among the three KGE components, the P datasets performed worst overall in terms of correlation (related to event identification). In terms of improving KGE scores for these datasets, improved P totals (affecting the bias score) and improved distribution of P intensity (affecting the variability score) are of secondary importance. Among the 11 gauge-corrected P datasets, the best overall performance was obtained by MSWEP V2.2, underscoring the importance of applying daily gauge corrections and accounting for gauge reporting times. Several uncorrected P datasets outperformed gauge-corrected ones. Among the 15 uncorrected P datasets, the best performance was obtained by the ERA5-HRES fourth-generation reanalysis, reflecting the significant advances in earth system modeling during the last decade. The (re)analyses generally performed better in winter than in summer, while the opposite was the case for the satellite-based datasets. IMERGHH V05 performed substantially better than TMPA-3B42RT V7, attributable to the many improvements implemented in the IMERG satellite P retrieval algorithm. IMERGHH V05 outperformed ERA5-HRES in regions dominated by convective storms, while the opposite was observed in regions of complex terrain. The ERA5-EDA ensemble average exhibited higher correlations than the ERA5-HRES deterministic run, highlighting the value of ensemble modeling. The WRF regional convection-permitting climate model showed considerably more accurate P totals over the mountainous west and performed best among the uncorrected datasets in terms of variability, suggesting there is merit in using high-resolution models to obtain climatological P statistics. Our findings provide some guidance to choose the most suitable P dataset for a particular application.

An early warning system for drought events can provide valuable information for decision makers dealing with water resources management and international aid. However, predicting such extreme events is still a big challenge. In this study, we compare two approaches for drought predictions based on forecasted precipitation derived from the Ensemble extended forecast model (ENS) of the ECMWF, and on forecasted monthly occurrence anomalies of weather regimes (MOAWRs), also derived from the ECMWF model. Results show that the MOAWRs approach outperforms the one based on forecasted precipitation in winter in the north-eastern parts of the European continent, where more than 65 % of droughts are detected 1 month in advance. The approach based on forecasted precipitation achieves better performance in predicting drought events in central and eastern Europe in both spring and summer, when the local atmospheric forcing could be the key driver of the precipitation. Sensitivity tests also reveal the challenges in predicting small-scale droughts and drought onsets at longer lead times. Finally, the results show that the ENS model of the ECMWF successfully represents most of the observed linkages between large-scale atmospheric patterns, depicted by the weather regimes and drought events over Europe.

This paper considers whether there is any added value in using seasonal climate forecasts instead of historical meteorological observations for forecasting streamflow on seasonal timescales over Europe. A Europe-wide analysis of the skill of the newly operational EFAS (European Flood Awareness System) seasonal streamflow forecasts (produced by forcing the Lisflood model with the ECMWF System 4 seasonal climate forecasts), benchmarked against the ensemble streamflow prediction (ESP) forecasting approach (produced by forcing the Lisflood model with historical meteorological observations), is undertaken. The results suggest that, on average, the System 4 seasonal climate forecasts improve the streamflow predictability over historical meteorological observations for the first month of lead time only (in terms of hindcast accuracy, sharpness and overall performance). However, the predictability varies in space and time and is greater in winter and autumn. Parts of Europe additionally exhibit a longer predictability, up to 7 months of lead time, for certain months within a season. In terms of hindcast reliability, the EFAS seasonal streamflow hindcasts are on average less skilful than the ESP for all lead times. The results also highlight the potential usefulness of the EFAS seasonal streamflow forecasts for decision-making (measured in terms of the hindcast discrimination for the lower and upper terciles of the simulated streamflow). Although the ESP is the most potentially useful forecasting approach in Europe, the EFAS seasonal streamflow forecasts appear more potentially useful than the ESP in some regions and for certain seasons, especially in winter for almost 40 % of Europe. Patterns in the EFAS seasonal streamflow hindcast skill are however not mirrored in the System 4 seasonal climate hindcasts, hinting at the need for a better understanding of the link between hydrological and meteorological variables on seasonal timescales, with the aim of improving climate-model-based seasonal streamflow forecasting.

We undertook a comprehensive evaluation of 22 gridded (quasi-)global (sub-)daily precipitation (P) datasets for the period 2000–2016. Thirteen non-gauge-corrected P datasets were evaluated using daily P gauge observations from 76 086 gauges worldwide. Another nine gauge-corrected datasets were evaluated using hydrological modeling, by calibrating the HBV conceptual model against streamflow records for each of 9053 small to medium-sized ( 2) catchments worldwide, and comparing the resulting performance. Marked differences in spatio-temporal patterns and accuracy were found among the datasets. Among the uncorrected P datasets, the satellite- and reanalysis-based MSWEP-ng V1.2 and V2.0 datasets generally showed the best temporal correlations with the gauge observations, followed by the reanalyses (ERA-Interim, JRA-55, and NCEP-CFSR) and the satellite- and reanalysis-based CHIRP V2.0 dataset, the estimates based primarily on passive microwave remote sensing of rainfall (CMORPH V1.0, GSMaP V5/6, and TMPA 3B42RT V7) or near-surface soil moisture (SM2RAIN-ASCAT), and finally, estimates based primarily on thermal infrared imagery (GridSat V1.0, PERSIANN, and PERSIANN-CCS). Two of the three reanalyses (ERA-Interim and JRA-55) unexpectedly obtained lower trend errors than the satellite datasets. Among the corrected P datasets, the ones directly incorporating daily gauge data (CPC Unified, and MSWEP V1.2 and V2.0) generally provided the best calibration scores, although the good performance of the fully gauge-based CPC Unified is unlikely to translate to sparsely or ungauged regions. Next best results were obtained with P estimates directly incorporating temporally coarser gauge data (CHIRPS V2.0, GPCP-1DD V1.2, TMPA 3B42 V7, and WFDEI-CRU), which in turn outperformed the one indirectly incorporating gauge data through another multi-source dataset (PERSIANN-CDR V1R1). Our results highlight large differences in estimation accuracy, and hence the importance of P dataset selection in both research and operational applications. The good performance of MSWEP emphasizes that careful data merging can exploit the complementary strengths of gauge-, satellite-, and reanalysis-based P estimates.

In light of strong encouragement for disaster managers to use climate services for flood preparation, we question whether seasonal rainfall forecasts should indeed be used as indicators of the likelihood of flooding. Here, we investigate the primary indicators of flooding at the seasonal timescale across sub-Saharan Africa. Given the sparsity of hydrological observations, we input bias-corrected reanalysis rainfall into the Global Flood Awareness System to identify seasonal indicators of floodiness. Results demonstrate that in some regions of western, central, and eastern Africa with typically wet climates, even a perfect tercile forecast of seasonal total rainfall would provide little to no indication of the seasonal likelihood of flooding. The number of extreme events within a season shows the highest correlations with floodiness consistently across regions. Otherwise, results vary across climate regimes: floodiness in arid regions in southern and eastern Africa shows the strongest correlations with seasonal average soil moisture and seasonal total rainfall. Floodiness in wetter climates of western and central Africa and Madagascar shows the strongest relationship with measures of the intensity of seasonal rainfall. Measures of rainfall patterns, such as the length of dry spells, are least related to seasonal floodiness across the continent. Ultimately, identifying the drivers of seasonal flooding can be used to improve forecast information for flood preparedness and to avoid misleading decision-makers.

Many fields, such as drought-risk assessment or reservoir management, can benefit from long-range streamflow forecasts. Climatology has long been used in long-range streamflow forecasting. Conditioning methods have been proposed to select or weight relevant historical time series from climatology. They are often based on general circulation model (GCM) outputs that are specific to the forecast date due to the initialisation of GCMs on current conditions. This study investigates the impact of conditioning methods on the performance of seasonal streamflow forecasts. Four conditioning statistics based on seasonal forecasts of cumulative precipitation and the standardised precipitation index were used to select relevant traces within historical streamflows and precipitation respectively. This resulted in eight conditioned streamflow forecast scenarios. These scenarios were compared to the climatology of historical streamflows, the ensemble streamflow prediction approach and the streamflow forecasts obtained from ECMWF System 4 precipitation forecasts. The impact of conditioning was assessed in terms of forecast sharpness (spread), reliability, overall performance and low-flow event detection. Results showed that conditioning past observations on seasonal precipitation indices generally improves forecast sharpness, but may reduce reliability, with respect to climatology. Conversely, conditioned ensembles were more reliable but less sharp than streamflow forecasts derived from System 4 precipitation. Forecast attributes from conditioned and unconditioned ensembles are illustrated for a case of drought-risk forecasting: the 2003 drought in France. In the case of low-flow forecasting, conditioning results in ensembles that can better assess weekly deficit volumes and durations over a wider range of lead times.

Too often, credible scientific early warning information of increased disaster risk does not result in humanitarian action. With financial resources tilted heavily towards response after a disaster, disaster managers have limited incentive and ability to process complex scientific data, including uncertainties. These incentives are beginning to change, with the advent of several new forecast-based financing systems that provide funding based on a forecast of an extreme event. Given the changing landscape, here we demonstrate a method to select and use appropriate forecasts for specific humanitarian disaster prevention actions, even in a data-scarce location. This action-based forecasting methodology takes into account the parameters of each action, such as action lifetime, when verifying a forecast. Forecasts are linked with action based on an understanding of (1) the magnitude of previous flooding events and (2) the willingness to act "in vain" for specific actions. This is applied in the context of the Uganda Red Cross Society forecast-based financing pilot project, with forecasts from the Global Flood Awareness System (GloFAS). Using this method, we define the "danger level" of flooding, and we select the probabilistic forecast triggers that are appropriate for specific actions. Results from this methodology can be applied globally across hazards and fed into a financing system that ensures that automatic, pre-funded early action will be triggered by forecasts.

Meteorological centres make sustained efforts to provide seasonal forecasts that are increasingly skilful, which has the potential to benefit streamflow forecasting. Seasonal streamflow forecasts can help to take anticipatory measures for a range of applications, such as water supply or hydropower reservoir operation and drought risk management. This study assesses the skill of seasonal precipitation and streamflow forecasts in France to provide insights into the way bias correcting precipitation forecasts can improve the skill of streamflow forecasts at extended lead times. We apply eight variants of bias correction approaches to the precipitation forecasts prior to generating the streamflow forecasts. The approaches are based on the linear scaling and the distribution mapping methods. A daily hydrological model is applied at the catchment scale to transform precipitation into streamflow. We then evaluate the skill of raw (without bias correction) and bias-corrected precipitation and streamflow ensemble forecasts in 16 catchments in France. The skill of the ensemble forecasts is assessed in reliability, sharpness, accuracy and overall performance. A reference prediction system, based on historical observed precipitation and catchment initial conditions at the time of forecast (i.e. ESP method) is used as benchmark in the computation of the skill. The results show that, in most catchments, raw seasonal precipitation and streamflow forecasts are often more skilful than the conventional ESP method in terms of sharpness. However, they are not significantly better in terms of reliability. Forecast skill is generally improved when applying bias correction. Two bias correction methods show the best performance for the studied catchments, each method being more successful in improving specific attributes of the forecasts: the simple linear scaling of monthly values contributes mainly to increasing forecast sharpness and accuracy, while the empirical distribution mapping of daily values is successful in improving forecast reliability.

Probabilistic hydro-meteorological forecasts have over the last decades been used more frequently to communicate forecast uncertainty. This uncertainty is twofold, as it constitutes both an added value and a challenge for the forecaster and the user of the forecasts. Many authors have demonstrated the added (economic) value of probabilistic over deterministic forecasts across the water sector (e.g. flood protection, hydroelectric power management and navigation). However, the richness of the information is also a source of challenges for operational uses, due partially to the difficulty in transforming the probability of occurrence of an event into a binary decision. This paper presents the results of a risk-based decision-making game on the topic of flood protection mitigation, called "How much are you prepared to pay for a forecast?". The game was played at several workshops in 2015, which were attended by operational forecasters and academics working in the field of hydro-meteorology. The aim of this game was to better understand the role of probabilistic forecasts in decision-making processes and their perceived value by decision-makers. Based on the participants' willingness-to-pay for a forecast, the results of the game show that the value (or the usefulness) of a forecast depends on several factors, including the way users perceive the quality of their forecasts and link it to the perception of their own performances as decision-makers.

Soil moisture memory is a key component of seasonal predictability. However, uncertainty in current memory estimates is not clear and it is not obvious to what extent these are dependent on model uncertainties. To address this question, we perform a global sensitivity analysis of memory to key hydraulic parameters, using an uncoupled version of the H-TESSEL land surface model. Results show significant dependency of estimates of memory and its uncertainty on these parameters, suggesting that operational seasonal forecasting models using deterministic hydraulic parameter values are likely to display a narrower range of memory than exists in reality. Explicitly incorporating hydraulic parameter uncertainty into models may then give improvements in forecast skill and reliability, as has been shown elsewhere in the literature. Our results also show significant differences with previous estimates of memory uncertainty, warning against placing too much confidence in a single quantification of uncertainty.

For many years, meteorological models have been run with perturbated initial conditions or parameters to produce ensemble forecasts that are used as a proxy of the uncertainty of the forecasts. However, the ensembles are usually both biased (the mean is systematically too high or too low, compared with the observed weather), and has dispersion errors (the ensemble variance indicates a too low or too high confidence in the forecast, compared with the observed weather). The ensembles are therefore commonly post-processed to correct for these shortcomings. Here we look at one of these techniques, referred to as Ensemble Model Output Statistics (EMOS) (Gneiting et al., 2005). Originally, the post-processing parameters were identified as a fixed set of parameters for a region. The application of our work is the European Flood Awareness System (http://www.efas.eu), where a distributed model is run with meteorological ensembles as input. We are therefore dealing with a considerably larger data set than previous analyses. We also want to regionalize the parameters themselves for other locations than the calibration gauges. The post-processing parameters are therefore estimated for each calibration station, but with a spatial penalty for deviations from neighbouring stations, depending on the expected semivariance between the calibration catchment and these stations. The estimated post-processed parameters can then be used for regionalization of the postprocessing parameters also for uncalibrated locations using top-kriging in the rtop-package (Skøien et al., 2006, 2014). We will show results from cross-validation of the methodology and although our interest is mainly in identifying exceedance probabilities for certain return levels, we will also show how the rtop package can be used for creating a set of post-processed ensembles through simulations.

The African Flood Forecasting System (AFFS) is a probabilistic flood forecast system for medium- to large-scale African river basins, with lead times of up to 15 days. The key components are the hydrological model LISFLOOD, the African GIS database, the meteorological ensemble predictions by the ECMWF (European Centre for Medium-Ranged Weather Forecasts) and critical hydrological thresholds. In this paper, the predictive capability is investigated in a hindcast mode, by reproducing hydrological predictions for the year 2003 when important floods were observed. Results were verified by ground measurements of 36 sub-catchments as well as by reports of various flood archives. Results showed that AFFS detected around 70 % of the reported flood events correctly. In particular, the system showed good performance in predicting riverine flood events of long duration (> 1 week) and large affected areas (> 10 000 km2) well in advance, whereas AFFS showed limitations for small-scale and short duration flood events. The case study for the flood event in March 2003 in the Sabi Basin (Zimbabwe) illustrated the good performance of AFFS in forecasting timing and severity of the floods, gave an example of the clear and concise output products, and showed that the system is capable of producing flood warnings even in ungauged river basins. Hence, from a technical perspective, AFFS shows a large potential as an operational pan-African flood forecasting system, although issues related to the practical implication will still need to be investigated.

Timely forecasts of the onset or possible evolution of droughts are an important contribution to mitigate their manifold negative effects. In this paper we therefore analyse and compare the performance of the first month of the probabilistic extended range forecast and of the seasonal forecast from the European Centre for Medium-range Weather Forecasts (ECMWF) in predicting droughts over the European continent. The Standardized Precipitation Index (SPI-1) is used to quantify the onset or likely evolution of ongoing droughts for the next month. It can be shown that on average the extended range forecast has greater skill than the seasonal forecast, whilst both outperform climatology. No significant spatial or temporal patterns can be observed, but the scores are improved when focussing on large-scale droughts. In a second step we then analyse several different methods to convert the probabilistic forecasts of SPI into a Boolean drought warning. It can be demonstrated that methodologies which convert low percentiles of the forecasted SPI cumulative distribution function into warnings are superior in comparison with alternatives such as the mean or the median of the ensemble. The paper demonstrates that up to 40 % of droughts are correctly forecasted one month in advance. Nevertheless, during false alarms or misses, we did not find significant differences in the distribution of the ensemble members that would allow for a quantitative assessment of the uncertainty.

The Canadian Forest Fire Weather Index (FWI) System is the mostly widely used fire danger rating system in the world. We have developed a global database of daily FWI System calculations, beginning in 1980, called the Global Fire WEather Database (GFWED) gridded to a spatial resolution of 0.5° latitude by 2/3° longitude. Input weather data were obtained from the NASA Modern Era Retrospective-Analysis for Research and Applications (MERRA), and two different estimates of daily precipitation from rain gauges over land. FWI System Drought Code calculations from the gridded data sets were compared to calculations from individual weather station data for a representative set of 48 stations in North, Central and South America, Europe, Russia, Southeast Asia and Australia. Agreement between gridded calculations and the station-based calculations tended to be most different at low latitudes for strictly MERRA-based calculations. Strong biases could be seen in either direction: MERRA DC over the Mato Grosso in Brazil reached unrealistically high values exceeding DC = 1500 during the dry season but was too low over Southeast Asia during the dry season. These biases are consistent with those previously identified in MERRA's precipitation, and they reinforce the need to consider alternative sources of precipitation data. GFWED can be used for analyzing historical relationships between fire weather and fire activity at continental and global scales, in identifying large-scale atmosphere–ocean controls on fire weather, and calibration of FWI-based fire prediction models.

Large-scale fires occur frequently across Indonesia, particularly in the southern region of Kalimantan and eastern Sumatra. They have considerable impacts on carbon emissions, haze production, biodiversity, health, and economic activities. In this study, we demonstrate that severe fire and haze events in Indonesia can generally be predicted months in advance using predictions of seasonal rainfall from the ECMWF System 4 coupled ocean–atmosphere model. Based on analyses of long, up-to-date series observations on burnt area, rainfall, and tree cover, we demonstrate that fire activity is negatively correlated with rainfall and is positively associated with deforestation in Indonesia. There is a contrast between the southern region of Kalimantan (high fire activity, high tree cover loss, and strong non-linear correlation between observed rainfall and fire) and the central region of Kalimantan (low fire activity, low tree cover loss, and weak, non-linear correlation between observed rainfall and fire). The ECMWF seasonal forecast provides skilled forecasts of burnt and fire-affected area with several months lead time explaining at least 70% of the variance between rainfall and burnt and fire-affected area. Results are strongly influenced by El Niño years which show a consistent positive bias. Overall, our findings point to a high potential for using a more physical-based method for predicting fires with several months lead time in the tropics rather than one based on indexes only. We argue that seasonal precipitation forecasts should be central to Indonesia's evolving fire management policy.

ERA-Interim/Land is a global land surface reanalysis data set covering the period 1979–2010. It describes the evolution of soil moisture, soil temperature and snowpack. ERA-Interim/Land is the result of a single 32-year simulation with the latest ECMWF (European Centre for Medium-Range Weather Forecasts) land surface model driven by meteorological forcing from the ERA-Interim atmospheric reanalysis and precipitation adjustments based on monthly GPCP v2.1 (Global Precipitation Climatology Project). The horizontal resolution is about 80 km and the time frequency is 3-hourly. ERA-Interim/Land includes a number of parameterization improvements in the land surface scheme with respect to the original ERA-Interim data set, which makes it more suitable for climate studies involving land water resources. The quality of ERA-Interim/Land is assessed by comparing with ground-based and remote sensing observations. In particular, estimates of soil moisture, snow depth, surface albedo, turbulent latent and sensible fluxes, and river discharges are verified against a large number of site measurements. ERA-Interim/Land provides a global integrated and coherent estimate of soil moisture and snow water equivalent, which can also be used for the initialization of numerical weather prediction and climate models.

Droughts are widespread natural hazards and in many regions their frequency seems to be increasing. A finer-resolution version (0.05° × 0.05°) of the continental-scale hydrological model PCRaster Global Water Balance (PCR-GLOBWB) was set up for the Limpopo River basin, one of the most water-stressed basins on the African continent. An irrigation module was included to account for large irrigated areas of the basin. The finer resolution model was used to analyse hydrological droughts in the Limpopo River basin in the period 1979–2010 with a view to identifying severe droughts that have occurred in the basin. Evaporation, soil moisture, groundwater storage and runoff estimates from the model were derived at a spatial resolution of 0.05° (approximately 5 km) on a daily timescale for the entire basin. PCR-GLOBWB was forced with daily precipitation and temperature obtained from the ERA-Interim global atmospheric reanalysis product from the European Centre for Medium-Range Weather Forecasts. Two agricultural drought indicators were computed: the Evapotranspiration Deficit Index (ETDI) and the Root Stress Anomaly Index (RSAI). Hydrological drought was characterised using the Standardized Runoff Index (SRI) and the Groundwater Resource Index (GRI), which make use of the streamflow and groundwater storage resulting from the model. Other more widely used meteorological drought indicators, such as the Standardized Precipitation Index (SPI) and the Standardized Precipitation Evaporation Index (SPEI), were also computed for different aggregation periods. Results show that a carefully set-up, process-based model that makes use of the best available input data can identify hydrological droughts even if the model is largely uncalibrated. The indicators considered are able to represent the most severe droughts in the basin and to some extent identify the spatial variability of droughts. Moreover, results show the importance of computing indicators that can be related to hydrological droughts, and how these add value to the identification of hydrological droughts and floods and the temporal evolution of events that would otherwise not have been apparent when considering only meteorological indicators. In some cases, meteorological indicators alone fail to capture the severity of the hydrological drought. Therefore, a combination of some of these indicators (e.g. SPEI-3, SRI-6 and SPI-12 computed together) is found to be a useful measure for identifying agricultural to long-term hydrological droughts in the Limpopo River basin. Additionally, it was possible to undertake a characterisation of the drought severity in the basin, indicated by its time of occurrence, duration and intensity.

Global seasonal forecasts of meteorological drought using the standardized precipitation index (SPI) are produced using two data sets as initial conditions: the Global Precipitation Climatology Centre (GPCC) and the European Centre for Medium-Range Weather Forecasts (ECMWF) ERA-Interim reanalysis (ERAI); and two seasonal forecasts of precipitation, the most recent ECMWF seasonal forecast system and climatologically based ensemble forecasts. The forecast evaluation focuses on the periods where precipitation deficits are likely to have higher drought impacts, and the results were summarized over different regions in the world. The verification of the forecasts with lead time indicated that generally for all regions the least reduction on skill was found for (i) long lead times using ERAI or GPCC for monitoring and (ii) short lead times using ECMWF or climatological seasonal forecasts. The memory effect of initial conditions was found to be 1 month of lead time for the SPI-3, 4 months for the SPI-6 and 6 (or more) months for the SPI-12. Results show that dynamical forecasts of precipitation provide added value with skills at least equal to and often above that of climatological forecasts. Furthermore, it is very difficult to improve on the use of climatological forecasts for long lead times. Our results also support recent questions of whether seasonal forecasting of global drought onset was essentially a stochastic forecasting problem. Results are presented regionally and globally, and our results point to several regions in the world where drought onset forecasting is feasible and skilful.

Near-real-time drought monitoring can provide decision-makers with valuable information for use in several areas, such as water resources management, or international aid. One of the main constrains of assessing the current drought situation is associated with the lack of reliable sources of observed precipitation on a global scale available in near-real time. Furthermore, monitoring systems also need a long record of past observations to provide mean climatological conditions. To address these problems, a novel probabilistic drought monitoring methodology based on ECMWF probabilistic forecasts is presented, where probabilistic monthly means of precipitation were derived from short-range forecasts and merged with the long-term climatology of the Global Precipitation Climatology Centre (GPCC) data set. From the merged data set, the standardised precipitation index (SPI) was estimated. This methodology was compared with the GPCC first guess precipitation product as well as SPI calculations using the ECMWF ERA-Interim reanalysis and Tropical Rainfall Measuring Mission (TRMM) precipitation data sets. ECMWF probabilistic forecasts for near-real-time monitoring are similar to GPCC and TRMM in terms of correlation and root mean square errors, with the added value of including an estimate of the uncertainty given by the ensemble spread. The real-time availability of this product and its stability (i.e. that it does not directly depend on local rain gauges or single satellite products) are also beneficial in the light of an operational implementation.

Systems for the early detection of floods over continental and global domains have a key role in providing a quick overview of areas at risk, raise the awareness and prompt higher detail analyses as the events approach. However, the reliability of these systems is prone to spatial inhomogeneity, depending on the quality of the underlying input data and local calibration. This work proposes a simple approach for flood early warning based on ensemble numerical predictions of surface runoff provided by weather forecasting centers. The system is based on a novel indicator, referred to as an extreme runoff index (ERI), which is calculated from the input data through a statistical analysis. It is designed for use in large or poorly gauged domains, as no local knowledge or in situ observations are needed for its setup. Daily runs over 32 months are evaluated against calibrated hydrological simulations for all of Europe. Results show skillful flood early warning capabilities up to a 10-day lead time. A dedicated analysis is performed to investigate the optimal timing of forecasts to maximize the detection of extreme events. A case study for the central European floods of June 2013 is presented and forecasts are compared to the output of a hydro-meteorological ensemble model.

Catastrophe risk models used by the insurance industry are likely subject to significant uncertainty, but due to their proprietary nature and strict licensing conditions they are not available for experimentation. In addition, even if such experiments were conducted, these would not be repeatable by other researchers because commercial confidentiality issues prevent the details of proprietary catastrophe model structures from being described in public domain documents. However, such experimentation is urgently required to improve decision making in both insurance and reinsurance markets. In this paper we therefore construct our own catastrophe risk model for flooding in Dublin, Ireland, in order to assess the impact of typical precipitation data uncertainty on loss predictions. As we consider only a city region rather than a whole territory and have access to detailed data and computing resources typically unavailable to industry modellers, our model is significantly more detailed than most commercial products. The model consists of four components, a stochastic rainfall module, a hydrological and hydraulic flood hazard module, a vulnerability module, and a financial loss module. Using these we undertake a series of simulations to test the impact of driving the stochastic event generator with four different rainfall data sets: ground gauge data, gauge-corrected rainfall radar, meteorological reanalysis data (European Centre for Medium-Range Weather Forecasts Reanalysis-Interim; ERA-Interim) and a satellite rainfall product (The Climate Prediction Center morphing method; CMORPH). Catastrophe models are unusual because they use the upper three components of the modelling chain to generate a large synthetic database of unobserved and severe loss-driving events for which estimated losses are calculated. We find the loss estimates to be more sensitive to uncertainties propagated from the driving precipitation data sets than to other uncertainties in the hazard and vulnerability modules, suggesting that the range of uncertainty within catastrophe model structures may be greater than commonly believed.

Drought monitoring is a key component to mitigate impacts of droughts. Lack of reliable and up-to-date precipitation data sets is a common challenge across the globe. This study investigates different data sets and drought indicators on their capability to improve drought monitoring in Africa. The study was performed for four river basins located in different climatic regions (the Oum er-Rbia in Morocco, the Blue Nile in eastern Africa, the Upper Niger in western Africa, and the Limpopo in southeastern Africa) as well as the Greater Horn of Africa. The five precipitation data sets compared are the ECMWF ERA-Interim reanalysis, the Tropical Rainfall Measuring Mission satellite monthly rainfall product 3B-43, the Global Precipitation Climatology Centre gridded precipitation data set, the Global Precipitation Climatology Project Global Monthly Merged Precipitation Analyses, and the Climate Prediction Center Merged Analysis of Precipitation. The set of drought indicators used includes the Standardized Precipitation Index, the Standardized Precipitation-Evaporation Index, and Soil Moisture Anomalies. A comparison of the annual cycle and monthly precipitation time series shows a good agreement in the timing of the rainy seasons. The main differences between the data sets are in the ability to represent the magnitude of the wet seasons and extremes. Moreover, for the areas affected by drought, all the drought indicators agree on the time of drought onset and recovery although there is disagreement on the extent of the affected area. In regions with limited rain gauge data the estimation of the different drought indicators is characterized by a higher uncertainty. Further comparison suggests that the main source of differences in the computation of the drought indicators is the uncertainty in the precipitation data sets rather than the estimation of the distribution parameters of the drought indicators.

Subsistence farming in southern Africa is vulnerable to extreme weather conditions. The yield of rain-fed agriculture depends largely on rainfall-related factors such as total seasonal rainfall, anomalous onsets and lengths of the rainy season and the frequency of occurrence of dry spells. Livestock, in turn, may be seriously impacted by climatic stress with, for example, exceptionally hot days, affecting condition, reproduction, vulnerability to pests and pathogens and, ultimately, morbidity and mortality. Climate change may affect the frequency and severity of extreme weather conditions, impacting on the success of subsistence farming. A potentially interesting adaptation measure comprises the timely forecasting and warning of such extreme events, combined with mitigation measures that allow farmers to prepare for the event occurring. This paper investigates how the frequency of extreme events may change in the future due to climate change over southern Africa and, in more detail, the Limpopo Basin using a set of climate change projections from several regional climate model downscalings based on an extreme climate scenario. Furthermore, the paper assesses the predictability of these indicators by seasonal meteorological forecasts of the European Centre for Medium-Range Weather Forecasts (ECMWF) seasonal forecasting system. The focus is on the frequency of dry spells as well as the frequency of heat stress conditions expressed in the temperature heat index. In areas where their frequency of occurrence increases in the future and predictability is found, seasonal forecasts will gain importance in the future, as they can more often lead to informed decision-making to implement mitigation measures. The multi-model climate projections suggest that the frequency of dry spells is not likely to increase substantially, whereas there is a clear and coherent signal among the models of an increase in the frequency of heat stress conditions by the end of the century. The skill analysis of the seasonal forecast system demonstrates that there is a potential to adapt to this change by utilizing the weather forecasts, given that both indicators can be skilfully predicted for the December–February season, at least 2 months ahead of the wet season. This is particularly the case for predicting above-normal and below-normal conditions. The frequency of heat stress conditions shows better predictability than the frequency of dry spells. Although results are promising for end users on the ground, forecasts alone are insufficient to ensure appropriate response. Sufficient support for appropriate measures must be in place, and forecasts must be communicated in a context-specific, accessible and understandable format.

The humanitarian crises caused by the recent droughts (2008–2009 and 2010–2011) in East Africa have illustrated that the ability to make accurate drought forecasts with sufficient lead time is essential. The use of dynamical model precipitation forecasts in combination with drought indices, such as the Standardized Precipitation Index (SPI), can potentially lead to a better description of drought duration, magnitude and spatial extent. This study evaluates the use of the European Centre for Medium-Range Weather Forecasts (ECMWF) products in forecasting droughts in East Africa. ECMWF seasonal precipitation shows significant skill for March–May and October–December rain seasons when evaluated against measurements from the available in situ stations from East Africa. The forecast for October–December rain season has higher skill than for the March–May season. ECMWF forecasts add value to the consensus forecasts produced during the Greater Horn of Africa Climate Outlook Forum (GHACOF), which is the present operational product for precipitation forecast over East Africa. Complementing the original ECMWF precipitation forecasts with SPI provides additional information on the spatial extent and intensity of the drought event.

Evaporation is a key process in the water cycle with implications ranging, inter alia, from water management to weather forecast and climate change assessments. The estimation of continental evaporation fluxes is complex and typically relies on continental-scale hydrological models or land-surface models. However, it appears that most global or continental-scale hydrological models underestimate evaporative fluxes in some regions of Africa, and as a result overestimate stream flow. Other studies suggest that land-surface models may overestimate evaporative fluxes. In this study, we computed actual evaporation for the African continent using a continental version of the global hydrological model PCR-GLOBWB, which is based on a water balance approach. Results are compared with other independently computed evaporation products: the evaporation results from the ECMWF reanalysis ERA-Interim and ERA-Land (both based on the energy balance approach), the MOD16 evaporation product, and the GLEAM product. Three other alternative versions of the PCR-GLOBWB hydrological model were also considered. This resulted in eight products of actual evaporation, which were compared in distinct regions of the African continent spanning different climatic regimes. Annual totals, spatial patterns and seasonality were studied and compared through visual inspection and statistical methods. The comparison shows that the representation of irrigation areas has an insignificant contribution to the actual evaporation at a continental scale with a 0.5° spatial resolution when averaged over the defined regions. The choice of meteorological forcing data has a larger effect on the evaporation results, especially in the case of the precipitation input as different precipitation input resulted in significantly different evaporation in some of the studied regions. ERA-Interim evaporation is generally the highest of the selected products followed by ERA-Land evaporation. In some regions, the satellite-based products (GLEAM and MOD16) show a different seasonal behaviour compared to the other products. The results from this study contribute to a better understanding of the suitability and the differences between products in each climatic region. Through an improved understanding of the causes of differences between these products and their uncertainty, this study provides information to improve the quality of evaporation products for the African continent and, consequently, leads to improved water resources assessments at regional scale.

Hydrological ensemble prediction systems (HEPS) have in recent years been increasingly used for the operational forecasting of floods by European hydrometeorological agencies. The most obvious advantage of HEPS is that more of the uncertainty in the modelling system can be assessed. In addition, ensemble prediction systems generally have better skill than deterministic systems both in the terms of the mean forecast performance and the potential forecasting of extreme events. Research efforts have so far mostly been devoted to the improvement of the physical and technical aspects of the model systems, such as increased resolution in time and space and better description of physical processes. Developments like these are certainly needed; however, in this paper we argue that there are other areas of HEPS that need urgent attention. This was also the result from a group exercise and a survey conducted to operational forecasters within the European Flood Awareness System (EFAS) to identify the top priorities of improvement regarding their own system. They turned out to span a range of areas, the most popular being to include verification of an assessment of past forecast performance, a multi-model approach for hydrological modelling, to increase the forecast skill on the medium range (>3 days) and more focus on education and training on the interpretation of forecasts. In light of limited resources, we suggest a simple model to classify the identified priorities in terms of their cost and complexity to decide in which order to tackle them. This model is then used to create an action plan of short-, medium- and long-term research priorities with the ultimate goal of an optimal improvement of EFAS in particular and to spur the development of operational HEPS in general.

The last decade has seen growing research in producing probabilistic hydro-meteorological forecasts and increasing their reliability. This followed the promise that, supplied with information about uncertainty, people would take better risk-based decisions. In recent years, therefore, research and operational developments have also started focusing attention on ways of communicating the probabilistic forecasts to decision-makers. Communicating probabilistic forecasts includes preparing tools and products for visualisation, but also requires understanding how decision-makers perceive and use uncertainty information in real time. At the EGU General Assembly 2012, we conducted a laboratory-style experiment in which several cases of flood forecasts and a choice of actions to take were presented as part of a game to participants, who acted as decision-makers. Answers were collected and analysed. In this paper, we present the results of this exercise and discuss if we indeed make better decisions on the basis of probabilistic forecasts.

Vast parts of Africa rely on the rainy season for livestock and agriculture. Droughts can have a severe impact in these areas, which often have a very low resilience and limited capabilities to mitigate drought impacts. This paper assesses the predictive capabilities of an integrated drought monitoring and seasonal forecasting system (up to 5 months lead time) based on the Standardized Precipitation Index (SPI). The system is constructed by extending near-real-time monthly precipitation fields (ECMWF ERA-Interim reanalysis and the Climate Anomaly Monitoring System–Outgoing Longwave Radiation Precipitation Index, CAMS-OPI) with monthly forecasted fields as provided by the ECMWF seasonal forecasting system. The forecasts were then evaluated over four basins in Africa: the Blue Nile, Limpopo, Upper Niger, and Upper Zambezi. There are significant differences in the quality of the precipitation between the datasets depending on the catchments, and a general statement regarding the best product is difficult to make. The generally low number of rain gauges and their decrease in the recent years limits the verification and monitoring of droughts in the different basins, reinforcing the need for a strong investment on climate monitoring. All the datasets show similar spatial and temporal patterns in southern and north-western Africa, while there is a low correlation in the equatorial area, which makes it difficult to define ground truth and choose an adequate product for monitoring. The seasonal forecasts have a higher reliability and skill in the Blue Nile, Limpopo and Upper Niger in comparison with the Zambezi. This skill and reliability depend strongly on the SPI timescale, and longer timescales have more skill. The ECMWF seasonal forecasts have predictive skill which is higher than using climatology for most regions. In regions where no reliable near-real-time data is available, the seasonal forecast can be used for monitoring (first month of forecast). Furthermore, poor-quality precipitation monitoring products can reduce the potential skill of SPI seasonal forecasts in 2 to 4 months lead time.

Anticipation and preparedness for large-scale flood events have a key role in mitigating their impact and optimizing the strategic planning of water resources. Although several developed countries have well-established systems for river monitoring and flood early warning, figures of populations affected every year by floods in developing countries are unsettling. This paper presents the Global Flood Awareness System (GloFAS), which has been set up to provide an overview on upcoming floods in large world river basins. GloFAS is based on distributed hydrological simulation of numerical ensemble weather predictions with global coverage. Streamflow forecasts are compared statistically to climatological simulations to detect probabilistic exceedance of warning thresholds. In this article, the system setup is described, together with an evaluation of its performance over a two-year test period and a qualitative analysis of a case study for the Pakistan flood, in summer 2010. It is shown that hazardous events in large river basins can be skilfully detected with a forecast horizon of up to 1 month. In addition, results suggest that an accurate simulation of initial model conditions and an improved parameterization of the hydrological model are key components to reproduce accurately the streamflow variability in the many different runoff regimes of the earth.

Global flood hazard maps can be used in the assessment of flood risk in a number of different applications, including (re)insurance and large scale flood preparedness. Such global hazard maps can be generated using large scale physically based models of rainfall-runoff and river routing, when used in conjunction with a number of post-processing methods. In this study, the European Centre for Medium Range Weather Forecasts (ECMWF) land surface model is coupled to ERA-Interim reanalysis meteorological forcing data, and resultant runoff is passed to a river routing algorithm which simulates floodplains and flood flow across the global land area. The global hazard map is based on a 30 yr (1979–2010) simulation period. A Gumbel distribution is fitted to the annual maxima flows to derive a number of flood return periods. The return periods are calculated initially for a 25 × 25 km grid, which is then reprojected onto a 1 × 1 km grid to derive maps of higher resolution and estimate flooded fractional area for the individual 25 × 25 km cells. Several global and regional maps of flood return periods ranging from 2 to 500 yr are presented. The results compare reasonably to a benchmark data set of global flood hazard. The developed methodology can be applied to other datasets on a global or regional scale.

The Normal Quantile Transform (NQT) has been used in many hydrological and meteorological applications in order to make the Cumulated Distribution Function (CDF) of the observed, simulated and forecast river discharge, water level or precipitation data Gaussian. It is also the heart of the meta-Gaussian model for assessing the total predictive uncertainty of the Hydrological Uncertainty Processor (HUP) developed by Krzysztofowicz. In the field of geo-statistics this transformation is better known as the Normal-Score Transform. In this paper some possible problems caused by small sample sizes when applying the NQT in flood forecasting systems will be discussed and a novel way to solve the problem will be outlined by combining extreme value analysis and non-parametric regression methods. The method will be illustrated by examples of hydrological stream-flow forecasts.

Flood forecasting increasingly relies on numerical weather prediction forecasts to achieve longer lead times. One of the key difficulties that is emerging in constructing a decision framework for these flood forecasts is what to dowhen consecutive forecasts are so different that they lead to different conclusions regarding the issuing of warnings or triggering other action. In this opinion paper we explore some of the issues surrounding such forecast inconsistency (also known as "Jumpiness", "Turning points", "Continuity" or number of "Swings"). In thsi opinion paper we define forecast inconsistency; discuss the reasons why forecasts might be inconsistent; how we should analyse inconsistency; and what we should do about it; how we should communicate it and whether it is a totally undesirable property. The property of consistency is increasingly emerging as a hot topic in many forecasting environments.

In this paper, we discuss a joint approach to calibration and uncertainty estimation for hydrologic systems that combines a top-down, data-based mechanistic (DBM) modelling methodology; and a bottom-up, reductionist modelling methodology. The combined approach is applied to the modelling of the River Hodder catchment in North-West England. The top-down DBM model provides a well identified, statistically sound yet physically meaningful description of the rainfall-flow data, revealing important characteristics of the catchment-scale response, such as the nature of the effective rainfall nonlinearity and the partitioning of the effective rainfall into different flow pathways. These characteristics are defined inductively from the data without prior assumptions about the model structure, other than it is within the generic class of nonlinear differential-delay equations. The bottom-up modelling is developed using the TOPMODEL, whose structure is assumed a priori and is evaluated by global sensitivity analysis (GSA) in order to specify the most sensitive and important parameters. The subsequent exercises in calibration and validation, performed with Generalized Likelihood Uncertainty Estimation (GLUE), are carried out in the light of the GSA and DBM analyses. This allows for the pre-calibration of the the priors used for GLUE, in order to eliminate dynamical features of the TOPMODEL that have little effect on the model output and would be rejected at the structure identification phase of the DBM modelling analysis. In this way, the elements of meaningful subjectivity in the GLUE approach, which allow the modeler to interact in the modelling process by constraining the model to have a specific form prior to calibration, are combined with other more objective, data-based benchmarks for the final uncertainty estimation. GSA plays a major role in building a bridge between the hypothetico-deductive (bottom-up) and inductive (top-down) approaches and helps to improve the calibration of mechanistic hydrological models, making their properties more transparent. It also helps to highlight possible mis-specification problems, if these are identified. The results of the exercise show that the two modelling methodologies have good synergy; combining well to produce a complete joint modelling approach that has the kinds of checks-and-balances required in practical data-based modelling of rainfall-flow systems. Such a combined approach also produces models that are suitable for different kinds of application. As such, the DBM model considered in the paper is developed specifically as a vehicle for flow and flood forecasting (although the generality of DBM modelling means that a simulation version of the model could be developed if required); while TOPMODEL, suitably calibrated (and perhaps modified) in the light of the DBM and GSA results, immediately provides a simulation model with a variety of potential applications, in areas such as catchment management and planning.

The paper presents a methodology for the estimation of uncertainty of inundation extent, which takes account of the uncertainty in the observed spatially distributed information and implements a fuzzy evaluation methodology. The Generalised Likelihood Uncertainty Estimation (GLUE) technique and the 2-D LISFLOOD-FP model were applied to derive the set of uncertain inundation realisations and resulting flood inundation maps. Conditioning of the inundation maps on fuzzified Synthetic Aperture Radar (SAR) images results in much more realistic inundation risk maps which can better depict the variable pattern of inundation extent than previously used methods. It has been shown that the evaluation methodology compares well to traditional approaches and can produce flood hazard maps that reflect the uncertainties in model evaluation.

The political pressure on the scientific community to provide medium to long term flood forecasts has increased in the light of recent flooding events in Europe. Such demands can be met by a system consisting of three different model components (weather forecast, rainfall-runoff forecast and flood inundation forecast) which are all liable to considerable uncertainty in the input, output and model parameters. Thus, an understanding of cascaded uncertainties is a necessary requirement to provide robust predictions. In this paper, 10-day ahead rainfall forecasts, consisting of one deterministic, one control and 50 ensemble forecasts, are fed into a rainfall-runoff model (LisFlood) for which parameter uncertainty is represented by six different parameter sets identified through a Generalised Likelihood Uncertainty Estimation (GLUE) analysis and functional hydrograph classification. The runoff of these 52 * 6 realisations form the input to a flood inundation model (LisFlood-FP) which acknowledges uncertainty by utilising ten different sets of roughness coefficients identified using the same GLUE methodology. Likelihood measures for each parameter set computed on historical data are used to give uncertain predictions of flow hydrographs as well as spatial inundation extent. This analysis demonstrates that a full uncertainty analysis of such an integrated system is limited mainly by computer power as well as by how well the rainfall predictions represent potential future conditions. However, these restrictions may be overcome or lessened in the future and this paper establishes a computationally feasible methodological approach to the uncertainty cascade problem.

Quantifying flood hazard is an essential component of resilience planning, emergency response, and mitigation, including insurance. Traditionally undertaken at catchment and national scales, recently, efforts have intensified to estimate flood risk globally to better allow consistent and equitable decision making. Global flood hazard models are now a practical reality, thanks to improvements in numerical algorithms, global datasets, computing power, and coupled modelling frameworks. Outputs of these models are vital for consistent quantification of global flood risk and in projecting the impacts of climate change. However, the urgency of these tasks means that outputs are being used as soon as they are made available and before such methods have been adequately tested. To address this, we compare multi-probability flood hazard maps for Africa from six global models and show wide variation in their flood hazard, economic loss and exposed population estimates, which has serious implications for model credibility. While there is around 30%–40% agreement in flood extent, our results show that even at continental scales, there are significant differences in hazard magnitude and spatial pattern between models, notably in deltas, arid/semi-arid zones and wetlands. This study is an important step towards a better understanding of modelling global flood hazard, which is urgently required for both current risk and climate change projections.