No abstract

<p>Drought events are projected to become more frequent and intense globally as a result of climate change. This weather hazard is particularly severe in the Horn of Africa, were it often produces damages to livestock, crop losses and food security emergencies. In order to reduce the severe impacts of this phenomenon and to promptly respond to humanitarian needs, it is essential to integrate the advances of the new generation satellite products to the current early warning systems. A study by West et al. (2019, 10.1016/j.rse.2019.111291) pointed-out how current drought monitoring systems require better spatial, temporal and spectral resolution to understand the complex nature of such events.</p> <p>For what concerns vegetation drought monitoring, current real-time systems (e.g. FEWS NET, GLDAS, NHyFAS) for emergency services  do not take advantage of the benefits of  geostationary satellite data and rather rely on Polar Operational Environmental Satellite data (POES) (Fensholt et al., 2011, 10.1016/j.jag.2011.05.009). Given their temporal resolution and considering that POES products present important data gaps problems due to the presence of clouds, they usually generate 10 or 16 days non-cloud contaminated vegetation composites. Consequently, this might affect timely response to natural hazards such droughts or floods. If on the one hand this data have been not fully accessible due to the requirements of considerable computational resources, the introduction of more powerful computers and distributed computing can narrow the gap.</p> <p>Precipitation deficits are a good indicator of meteorological drought. However, precise rainfall estimates are difficult to obtain given their large spatial variability. In Africa, convective storms can generate localized precipitation phenomena which can be difficult to measure. Additionally, weather stations data is limited and stations number have been decreasing. When using satellite products, studies on the African continent have shown how precipitation remote sensing products can present wet or dry biases (McNally et al., 2017, 10.1038/sdata.2017.12), which can in turn affect the outcomes of precipitation-derived meteorological indices (e.g. Standard Precipitation Index, SPI). Overall, the existence of different rain gauge, satellite or reanalysis products makes it non-trivial the identification of an optimal precipitation series (Le Coz & van de Giesen, 2020, 10.1175/JHM-D-18-0256.1) that can describe extreme events.</p> <p>The aim of this study is to show the benefits that the last generation of satellite products can offer to drought monitoring, in particular for those areas that lack reliable and dense in situ precipitation data. For this purpose, we will explore the relationships between meteorological and agricultural droughts for a subset of countries in the Greater Horn of Africa (Ethiopia, Somalia, Kenya), studying the relationships between vegetation health indexes (NDVI, VCI) and precipitation anomalies (e.g. SPI). The indexes related to the vegetation health have been calculated at daily time scale using the Meteosat SEVIRI radiometer, while the precipitation anomalies have been estimated using several precipitation products (e.g. rain gauge, satellite, reanalysis) at different short-term scales (30, 60, 90 and 180 days). A comprehensive intercomparison of the different precipitation products in the study area and their importance for the detection and forecast of agricultural droughts will be discussed.</p>