Improving early warning of drought-driven food insecurity in southern Africa using operational hydrological monitoring and forecasting products

摘要

The region of southern Africa?(SA) has a fragile food economy and is vulnerable to frequent droughts. Interventions to mitigate food insecurityimpacts require early warning of droughts – preferably as early as possible before the harvest season (typically starting in April) and lean season (typically starting in November). Hydrologic monitoring and forecasting systems provide a unique opportunity to support early warning efforts, since they can provide regular updates on available root-zone soil moisture?(RZSM), a critical variable for crop yield, and provide forecasts of RZSM by combining the estimates of antecedent soil moisture conditions with climate forecasts. For?SA, this study documents the predictive capabilities of RZSM products from the recently developed NASA Hydrological Forecasting and Analysis System?(NHyFAS). Results show that the NHyFAS products would have identified the regional severe drought event – which peaked during December–February of?2015–2016 – at least as early as 1?November?2015. Next, it is shown that during?1982–2016, February RZSM (Feb-RZSM) forecasts (monitoring product) available in early November (early March) have a correlation of?0.49?(0.79) with the detrended regional crop yield. It is also found that when the February RZSM forecast (monitoring product) available in early November (early March) is indicated to be in the lowest tercile, the detrended regional crop yield is below normal about two-thirds of the time (always), at least over the sample years considered. Additionally, it is shown that the February RZSM forecast (monitoring product) can provide “out-of-sample” crop yield forecasts with comparable (substantially better with 40 % reduction in mean error) skill to December–February ENSO. These results indicate that the NHyFAS products can effectively support food insecurity early warning in the SA?region. Finally, since a framework similar to NHyFAS can be used to provide RZSM monitoring and forecasting products over other regions of the globe, this case study also demonstrates potential for supporting food insecurity early warning globally.