Food insecurity is a global concern resulting from various complex processes and a diverse range of drivers. Due to its complexity, it is one of the most challenging drought impacts to predict. In this study, we introduce a novel machine learning model designed to forecast food crises in the Horn of Africa up to 12 months in advance. We trained an “XGBoost” model using more than 20 different input datasets to capture key food security drivers such as drought, economic shocks, conflicts and livelihood vulnerability. The model shows a promising ability to predict food security dynamics several months in advance (R2>0.6, three months in advance). Notably, it accurately predicted 20% of crisis onsets in pastoral regions (n = 84) and 40% of crisis onsets in agro-pastoral regions (n = 23) with a 3-month lead time. We compared these results to the established FEWS NET early warning system, and found a similar performance over these regions. However, our model is clearly less skilled in predicting food security for crop-farming regions than FEWS NET. This study underscores the importance of integrating machine learning into operational early-warning systems like FEWS NET and expanding these techniques to the continental or global-scale.

In this study, we present a machine‐learning model capable of predicting food insecurity in the Horn of Africa, which is one of the most vulnerable regions worldwide. The region has frequently been affected by severe droughts and food crises over the last several decades, which will likely increase in future. Therefore, exploring novel methods of increasing early warning capabilities is of vital importance to reducing food‐insecurity risk. We present a XGBoost machine‐learning model to predict food‐security crises up to 12 months in advance. We used >20 data sets and the FEWS IPC current‐situation estimates to train the machine‐learning model. Food‐security dynamics were captured effectively by the model up to 3 months in advance (R2 > 0.6). Specifically, we predicted 20% of crisis onsets in pastoral regions (n = 96) and 20%–50% of crisis onsets in agro‐pastoral regions (n = 22) with a 3‐month lead time. We also compared our 8‐month model predictions to the 8‐month food‐security outlooks produced by FEWS NET. Over a relatively short test period (2019–2022), results suggest the performance of our predictions is similar to FEWS NET for agro‐pastoral and pastoral regions. However, our model is clearly less skilled in predicting food security for crop‐farming regions than FEWS NET. With the well‐established FEWS NET outlooks as a basis, this study highlights the potential for integrating machine‐learning methods into operational systems like FEWS NET.

The Horn of Africa faces an ongoing multi-year drought due to five consecutive failed rainy seasons, a novel climatic event with unpreceded impacts. Beyond the starvation of millions of livestock, close to 23 million individuals in the region are currently facing high food insecurity in Kenya, Somalia and Ethiopia alone. The severity of these impacts calls for the urgent upscaling and optimisation of early action for droughts. However, drought research focuses mainly on meteorological and hydrological forecasting, while early action triggered by forecasts is seldom addressed. This study investigates the potential for early action for droughts by using seasonal forecasts from the European Centre for Medium-Range Weather Forecasts (ECMWF) SEAS5 system for the March-April-May (MAM) and October-November-December (OND) rainy seasons. We show that these seasonal rainfall forecasts reflect major on-the-ground impacts, which we identify from drought surveillance data from 21 counties in Kenya. Subsequently, we show that the SEAS5 drought forecasts with short lead times have substantial potential economic value (PEV) when used to trigger action before the OND season across the region (PEVmax = 0.43). Increasing lead time to one or two months ahead of the season decreases PEV, but the benefits persist (PEVmax = 0.2). Outside of Kenya, MAM forecasts have limited value. The existence of opportunities for early action during the OND season in Kenya and Somalia is demonstrated by high PEV values, with some regions recording PEVmax values close to 0.8. To illustrate the practical value of this research, we point to a dilemma that a pastoralist in the Kenyan drylands faces when deciding whether to adopt early livestock destocking. This study underscores the importance to determine the value of early actions for forecast users with different action characteristics, and to disseminate this value alongside the standard forecasts themselves. This allows users to trigger effective actions before drought impacts develop.

Drought currently threatens the food security of 26 million people in the Horn of Africa, and without urgent action this could have a devastating impact during the first half of 2022. Local agro-pastoralists living in the Horn of Africa drylands are among the most vulnerable people to these droughts. Drought forecasts are crucial for them to prepare, as these allow for life-saving forecast-based actions. However, there is a mismatch between forecasts and the impacts felt on the ground leading to local agro-pastoralists not using these forecasts. Therefore, we will develop and evaluate impact-based forecasting techniques for agro-pastoralists in the Horn of Africa using the TAMSAT-ALERT early-warning system. Over Kenya, these ensemble forecasts indicate a good relationship to pasture availability (NDVI/VCI) and maize yield, and can reliably anticipate drought conditions (<20th percentile soil moisture) up to 2 months before the end of the season. We build further on these results by testing its capabilities to forecast more specific drought variables relevant for agro-pastoralists. This includes a transition to drought (i.e. the drought onset), drought duration, the occurrence of dry spells in the rainy season, and possibly wet periods during the dry season. Moreover, we will assess how these variables relate to drought impacts, using impact data on livestock losses and IPC Acute Food Insecurity Classifications from FEWS-NET. This will pave the way for an estimation of the capabilities of TAMSAT-ALERT to serve as an impact-based forecasting system in African drylands. Furthermore, we will improve the model by integrating ensemble sub-seasonal European Centre for Medium-Range Weather Forecasts (ECMWF) precipitation forecasts, and will evaluate how much these forecasts increase the forecast accuracy of TAMSAT-ALERT. Using the potential economic value theory to trigger anticipatory action, we will reveal on when effective anticipatory actions can best be triggered, such as distribution of emergency fodder or migration of pastoralists. These research steps will provide decision-makers with tailored drought forecasts and decision-support, which will be crucial to increase drought resilience in African drylands.

Research shows that climate change will increase the intensity and frequency of extreme summer precipitation events as well as heatwaves, over the coming decades (IPCC, 2014; Russo et al., 2015). Moreover, the impact of heat waves will likely increase in cities due to the urban heat island (UHI) effect (Li & Bou-Zeid, 2013). Green infrastructure (e.g. parks, green roofs, etc.) is generally seen as an effective adaptation measure to address these challenges. The city of Amsterdam has started a project (RESILIO, https://resilio.amsterdam/en/smart-blue-green-roofs) to investigate a new innovation in this field: smart blue-green roofs. These roofs have the advantage over green roofs in that they have an extra water retention layer underneath the green layer, which can be used to buffer peak rainfall or as a capillary irrigation system for the plant layer in hot and dry summer days. The smart valve on the roof can be opened when extreme precipitation is predicted to capture extreme rainfall, but it is yet unknown if this forecast-based drainage provides added value to optimize the operation of the valve. Therefore, this study evaluates the performance of European Centre for Medium-Range Weather Forecasts (ECMWF) ensemble precipitation forecasts to trigger drainage from blue-green roofs. A conceptual hydrological model of a blue-green roof in Amsterdam is set up to simulate its operation for the last 5 years. Three drainage strategies can be triggered according to different probabilities of precipitation (30th, 60th and 90th percentile) based on ECMWF data. Each strategy is evaluated on how it leads to (1) minimize the overflow during peak rainfall into the city drainage system, and (2) to maintain high water levels during hot summer days to boost evaporative cooling. Preliminary results show that some early drainage strategies result in capturing 50-100% of rainfall (>10mm/hr), while enough water is available on most hot summer days (T>25℃) to ensure atmospheric cooling through plant transpiration. This implies that relatively low-resolution (18km) precipitation forecasts from ECMWF are valuable for anticipatory water management on a very local scale. These results also show the high potential of blue-green roofs for urban climate adaptation, and the need for anticipatory management of these nature-based solutions. The next research steps will include a city-scale roof suitability analysis that will reveal the value of this solution when implemented at most flat roofs in the city of Amsterdam. IPCC. (2014). Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Li, D., & Bou-Zeid, E. (2013). Synergistic interactions between urban heat islands and heat waves: The impact in cities is larger than the sum of its parts. Journal of Applied Meteorology and Climatology. https://doi.org/10.1175/JAMC-D-13-02.1 Russo, S., Sillmann, J., & Fischer, E. M. (2015). Top ten European heatwaves since 1950 and their occurrence in the coming decades. Environmental Research Letters. https://doi.org/10.1088/1748-9326/10/12/124003

As a result of high-density urbanization and climate change, both the frequency and intensity of extreme urban rainfall are increasing. Drainage systems are not designed to cope with this increase, and as a result, floods are becoming more common in cities, particularly in the rapidly growing cities of China. To better cope with more frequent and severe urban flooding and to improve the water quality of stormwater runoff, the Chinese government launched the national Sponge City Construction (SCC) program in 2014. The current SCC design standards and guidelines are based on static values (e.g., return periods, rainfall intensities, and volume capture ratio (VCR)). They do not fully acknowledge the large differences in climate conditions across the country and assume that the hydraulic conditions will not change over time. This stationary approach stems from the traditional engineering approach designed for grey infrastructure (following a "one size fits all" approach). The purpose of this study was to develop a methodology to assess the VCR baseline (before construction in the pre-development stage) and changes in VCR (difference between the VCR of the pre- and post-development stage). The VCR of the post-development stage is one of the required indicators of the Assessment Standard for Sponge Cities Effects to evaluate SCC projects. In this study, the VCR was derived from remote-sensing-based land use land cover (LULC) change analysis, applying an unsupervised classification algorithm on different Landsat images from 1985 to 2015. A visualization method (based upon Sankey chart, which depicts the flows and their proportions of components) and a novel and practical partitioning method for built-up regions were developed to visualize and quantify the states and change flows of LULC. On the basis of these findings, we proposed a new indicator, referred to as VCRa-L, in order to assess the changes in urban hydrology after SCC construction. This study employed the city of Nanjing as a case study and analyzed detailed information on how LULC changes over time of built-up areas. The surface area of the urban and built-up areas of Nanjing quadrupled from 11% in 1985 to 44% in 2015. In the same period, neither the entire city nor its subregions reached the VCR target of 80%. The proposed new methodology aims to support national, regional, and city governments to identify and prioritize where to invest and implement SCC measures more effectively in cities across China.

Lakes and reservoirs are crucial elements of the hydrological and biochemical cycle and are a valuable resource for hydropower, domestic and industrial water use, and irrigation. Although their monitoring is crucial in times of increased pressure on water resources by both climate change and human interventions, publically available datasets of lake and reservoir levels and volumes are scarce. Within this study, a time series of variation in lake and reservoir volume between 1984 and 2015 were analysed for 137 lakes over all continents by combining the JRC Global Surface Water (GSW) dataset and the satellite altimetry database DAHITI. The GSW dataset is a highly accurate surface water dataset at 30 m resolution compromising the whole L1T Landsat 5, 7 and 8 archive, which allowed for detailed lake area calculations globally over a very long time period using Google Earth Engine. Therefore, the estimates in water volume fluctuations using the GSW dataset are expected to improve compared to current techniques as they are not constrained by complex and computationally intensive classification procedures. Lake areas and water levels were combined in a regression to derive the hypsometry relationship (dh ∕ dA) for all lakes. Nearly all lakes showed a linear regression, and 42 % of the lakes showed a strong linear relationship with a R2 > 0.8, an average R2 of 0.91 and a standard deviation of 0.05. For these lakes and for lakes with a nearly constant lake area (coefficient of variation < 0.008), volume variations were calculated. Lakes with a poor linear relationship were not considered. Reasons for low R2 values were found to be (1) a nearly constant lake area, (2) winter ice coverage and (3) a predominant lack of data within the GSW dataset for those lakes. Lake volume estimates were validated for 18 lakes in the US, Spain, Australia and Africa using in situ volume time series, and gave an excellent Pearson correlation coefficient of on average 0.97 with a standard deviation of 0.041, and a normalized RMSE of 7.42 %. These results show a high potential for measuring lake volume dynamics using a pre-classified GSW dataset, which easily allows the method to be scaled up to an extensive global volumetric dataset. This dataset will not only provide a historical lake and reservoir volume variation record, but will also help to improve our understanding of the behaviour of lakes and reservoirs and their representation in (large-scale) hydrological models.

The October 2015 storm in Alexandria led to flooding of historical proportions. Although this October storm was a rare event (≈1:50 year), it could have been predicted at least 1 week in advance. If an appropriate warning system had been in place, measures could have been taken to alleviate its consequences. Therefore, the use of rainfall forecasting in storm water modelling can be one of the first useful and cheap mitigation measures. This forecasting system may further allow employing Anticipatory Water Management. The latter would imply that a large extra retention reservoir would be made available by pumping the water out from lakes of the city (Lake Maryot and Airport Lake) before it rains. Further research will be required to assess the technical and economic aspects of this option. The high variability and uncertainty of rainfall call for a robust and flexible strategy for Alexandria, which considers a portfolio of measures able to absorb the negative consequences of extreme events (designed for exceedance). This strategy should enable the responsible authorities to map out the future of the drainage system across the city over the next two or three decades and to identify the tipping point when upgrading of the existing drainage and irrigation system will be required. It will also provide the required evidence base on which a long-term wastewater and sewerage plan for Alexandria City should be implemented.