Predicting Impacts of Cyclones in South-East Africa (PICSEA)

Institute

  • University of Reading

Background

On average, 14 tropical cyclones per year form in the southern Indian Ocean, most in the months between November and April. Of these, about two or three per year make landfall in south-east Africa, most often in Mozambique and Madagascar. In these countries, tropical cyclones are associated with approximately one third of all extreme daily precipitation events, defined as days with rainfall accumulations greater than 50 mm (2 inches). Tropical cyclone landfalls in Mozambique in 2012 caused severe flooding, resulting in US$65 million in damage and 150 deaths. Two cyclone landfalls in Madagascar in early 2018 resulted in 23 deaths and displaced 21 000 people. The Seychelles archipelago is also affected by tropical cyclones, including category 5 (the most severe) Fantala in 2016.

Despite the vulnerability of the south-east African populations to tropical cyclones and related hazards, little is known about the ability of contemporary weather and climate prediction systems to forecast cyclone tracks, intensities, and wind and rain impacts. Further, there may be particular tropical atmospheric circulation patterns that provide 'windows of opportunity' for more accurate cyclone forecasts. For instance, El Niño conditions (warm equatorial Pacific Ocean temperatures) may provide the backdrop for more accurate predictions of tropical cyclones and their impacts.

Aims

The PICSEA project addresses these shortcomings by providing the most comprehensive assessment of forecast systems to date for tropical cyclones and their effects on south-east Africa. This assessment is needed desperately to give advice to national meteorological agencies, humanitarian organisations and the growing forecast-based finance community on how best to interpret forecasts of tropical cyclones in the southern Indian Ocean.

Specifically, PICSEA will determine which forecast systems, lead times and background tropical circulations lead to relatively more or less accurate tropical cyclone predictions. When should disaster management agencies trust a forecast for a landfalling tropical cyclone, and when should they not?

Approach

Initially, PICSEA will assess the accuracy of predictions of tropical cyclone tracks, intensities and associated hazards (primarily wind and rain) from three weather forecasting centres: the UK Met Office, the European Centre for Medium-range Weather Forecasts and the US National Centers for Environmental Prediction.

We have 10–30 years of forecast data for each centre, including multiple realisations of each forecast. We will determine to what extent, and how far in advance, contemporary prediction systems can forecast the extreme winds and rainfall associated with tropical cyclones in south-east Africa.

Next, PICSEA will determine whether there are particular background tropical conditions, such as El Niño or La Niña, that lead to more or less accurate forecasts. We will do this by evaluating forecast accuracy conditioned on the type of background conditions. Is skill for cyclone-related hazards greater during La Niña or El Niño? Are there particular circumstances under which forecasters and disaster management agencies should trust these forecasts more, or less?

Finally, PICSEA will work together with partner organisations — national meteorological organisations in Mozambique, Madagascar and the Seychelles, as well as the Red Cross/Red Crescent Climate Center, a key provider of scientific advice to humanitarian organisations and the forecast-based finance community — to develop guidance for interpreting tropical cyclone forecasts. We will work with forecasters and disaster management agencies to improve their understanding of when they can, and cannot, trust forecast information on cyclone impacts. PICSEA will also provide training in the use of this guidance, as well as background training on tropical meteorology, for forecasters in south-east African meteorological agencies.