Evaluating high-impact precipitation forecasts from the Met Office Global Hazard Map (GHM) using a global impact database

Abstract

The Met Office Global Hazard Map (GHM) summarizes the risk of high-impact weather across the globe over the coming week using global ensemble forecast data. In addition to gridded daily probability forecasts, a symbol and polygon-based summary layer gives an at-a-glance view of likely high-impact weather for the week ahead. To evaluate the performance of the GHM, two complementary approaches were used. The first is an objective precipitation verification approach comparing the daily gridded precipitation forecasts with global precipitation observations. The second, and the main focus of this paper is a new, semi-automated evaluation approach that assesses the ability of the multi-model ensemble precipitation summary layer to highlight events that cause community impacts, as recorded in an impact database. The verification against observed precipitation confirms there is good skill in the precipitation forecasts and that the multi-model ensemble provides the best guidance to take forward into the summary layer. The verification against impacts indicates there is a good spatial relationship between the GHM precipitation forecasts and heavy rainfall impacts across the globe. Hit rates for all impact severities range from 40 to 60% for days 1–3 and taper off to lower hit rates at the longer lead times (10–20% for days 6–7). High-impact events are captured marginally less well than the low-, medium- and disastrous-impact event categories and this paper illustrates a number of approaches that could positively alter the profile of the hit-rate curve.