NL160_Cover_image_v4_for_digital_banner

ECMWF works with universities to support response to tropical cyclone Idai

Linus Magnusson, Ervin Zsoter, Christel Prudhomme, Calum Baugh, Shaun Harrigan (all ECMWF), Andrea Ficchì, Rebecca Emerton, Hannah Cloke, Liz Stephens, Linda Speight (all University of Reading, UK)

 

The tropical cyclone season of 2019 in the southern Indian Ocean was one of the most active on record, with 15 tropical storms. On 15 March, tropical cyclone Idai made landfall in Mozambique, causing around a thousand fatalities. This made Idai the deadliest cyclone in the southern Indian Ocean for more than 100 years. ECMWF’s forecasts predicted the landfall location and extreme precipitation and winds with high confidence about 5 days ahead of landfall. At the same time range, flood forecasts based on ECMWF’s precipitation forecasts indicated a moderate risk of severe flooding, rising to a very high risk after landfall. ECMWF worked with partners at the Universities of Reading and Bristol and the UK Government to support the humanitarian response to the disaster.

Meteorology

On 4 March, a tropical depression made landfall on the coast of Mozambique and propagated northwest. The depression brought anomalous rainfall and led to deadly floods across central Mozambique and southern Malawi from 5 March. On 8 March, the depression turned eastward, and on 9 March it moved back over the ocean. On 10 March, the depression intensified and became a tropical cyclone. A few days later, it started to propagate southwest and intensified further. The cyclone made landfall near Beira on 15 March and later moved further inland bringing heavy rainfall also to Zimbabwe. When Idai reached the coast of Mozambique, it brought winds of up to 170 km/h and a significant storm surge of about 4.5 m around Beira. According to NASA-GPM data, rainfall totals between 3 and 17 March reached 400–600 mm over much of the Sofala and Manica provinces of Mozambique, between Beira and Chimoio. These extreme rainfall amounts caused widespread flooding around most rivers in the region, including the Pungwe, Save, Buzi, Revue and Shire Rivers. According to the United Nations, an estimated 1.85 million people were affected by Cyclone Idai in Mozambique and required humanitarian assistance.

%3Cstrong%3E%20ECMWF%E2%80%99s%20track,%20position%20and%20intensity%20forecasts%20for%20Idai.%20%3C/strong%3E%20The%20plots%20show%20forecasts%20starting%20at%2000%C2%A0UTC%20on%207%C2%A0March%C2%A02019%20(left)%20and%2000%C2%A0UTC%20on%2010%C2%A0March%C2%A02019%20(right).%20The%20squares%20indicate%20position%20and%20intensity%20forecasts%20for%2012%C2%A0UTC%20on%2014%C2%A0March.%20The%20black%20line%20shows%20the%20observed%20track%20(%E2%80%98Best%20Track%E2%80%99%20estimate),%20while%20the%20observed%20position%20and%20intensity%20at%2012%C2%A0UTC%20on%2014%C2%A0March%20is%20indicated%20by%20the%20hourglass%20symbol.%20
ECMWF’s track, position and intensity forecasts for Idai. The plots show forecasts starting at 00 UTC on 7 March 2019 (left) and 00 UTC on 10 March 2019 (right). The squares indicate position and intensity forecasts for 12 UTC on 14 March. The black line shows the observed track (‘Best Track’ estimate), while the observed position and intensity at 12 UTC on 14 March is indicated by the hourglass symbol.

ECMWF’s forecasts

ECMWF’s ensemble forecasts (ENS) from 4 March signalled an increased probability of a tropical cyclone between mainland Africa and Madagascar for the following week. This coincides with the first appearance of the depression. In the forecast from 7 March, many ensemble members captured the depression’s reappearance over sea and subsequent intensification. There was, however, large uncertainty in the track forecast, and the predicted turning point was too far to the west, leading to a predicted landfall location too far north.

%3Cstrong%3E%20Extreme%20Forecast%20Index%20(EFI)%20and%20Shift%20of%20Tails%20(SOT)%20for%20precipitation.%20%3C/strong%3E%20There%20was%20a%20clear%20signal%20in%20the%20EFI%20(shading)%20and%20SOT%20(contours)%20for%20total%20precipitation%20on%2015%C2%A0March%20in%20the%20area%20affected%20by%20Cyclone%20Idai%20in%20the%20forecast%20from%2000%C2%A0UTC%20on%2010%C2%A0March.
Extreme Forecast Index (EFI) and Shift of Tails (SOT) for precipitation. There was a clear signal in the EFI (shading) and SOT (contours) for total precipitation on 15 March in the area affected by Cyclone Idai in the forecast from 00 UTC on 10 March.

On 10 March, when the depression intensified, the forecast became much more confident on the landfall location. From this point on, the ensemble forecasts predicted extreme rainfall in southern/central Mozambique and also the severe winds over Beira with high confidence.

GloFAS flood forecasts

Forecasts from the Global Flood Awareness System (GloFAS, www.globalfloods.eu), which is part of the EU-funded Copernicus Emergency Management Service (CEMS), use rainfall from ENS as input. From 10 March, as forecasts of the landfall location of Cyclone Idai became more certain, GloFAS forecasts showed moderate probabilities (> 40%) of severe flooding around Beira and for the most affected rivers in Mozambique (Pungwe, Buzi, Revue and Save). These probabilities did not increase but fluctuated in the following days, in line with the precipitation forecasts, until landfall on 15 March. On 12 March, severe flooding for the Pungwe River at Beira was predicted with a probability of more than 50%, while lower probabilities of severe flooding were indicated for the other main rivers. Only once Idai had made landfall and large rainfall amounts were observed or predicted by short-range forecasts (from 16 March), probabilities of severe flooding became very high (> 90%) across the whole region. At that point water levels were still increasing and flood peaks were consistently predicted to happen 3 to 7 days ahead.

%3Cstrong%3E%20GloFAS%20flood%20risk%20forecasts.%20%3C/strong%3E%20The%20shading%20of%20the%20rivers%20in%20the%20maps%20shows%20the%20predicted%20probability%20(in%C2%A0%25)%20of%20the%20streamflow%20exceeding%20the%20severe%20flood%20alert%20threshold%20(20-year%20return%20period)%20over%20the%20next%2030%C2%A0days%20in%20forecasts%20starting%20on%2010%C2%A0March%20(top)%20and%2016%C2%A0March%20(bottom).%20The%20triangles%20denote%20so-called%20reporting%20points,%20where%20the%20predicted%20evolution%20of%20the%20ensemble%20discharge,%20with%20the%20probability%20of%20exceeding%20different%20thresholds,%20is%20available%20on%20the%20GloFAS%20website.%20Upward%20pointing%20triangles%20indicate%20rising%20flow,%20downward%20pointing%20triangles%20indicate%20receding%20flow.%20The%20inset%20charts%20show%20the%20discharge%20evolution%20for%20the%20reporting%20point%20location%20indicated%20in%20the%20map%20(the%20Pungwe%20River%20at%20Beira,%2019.65%C2%B0S%20and%2034.65%C2%B0E).
GloFAS flood risk forecasts. The shading of the rivers in the maps shows the predicted probability (in %) of the streamflow exceeding the severe flood alert threshold (20-year return period) over the next 30 days in forecasts starting on 10 March (top) and 16 March (bottom). The triangles denote so-called reporting points, where the predicted evolution of the ensemble discharge, with the probability of exceeding different thresholds, is available on the GloFAS website. Upward pointing triangles indicate rising flow, downward pointing triangles indicate receding flow. The inset charts show the discharge evolution for the reporting point location indicated in the map (the Pungwe River at Beira, 19.65°S and 34.65°E).

Response

In the immediate aftermath of the landfall of Cyclone Idai, the CEMS-Floods team at ECMWF worked with the University of Reading, the University of Bristol, and the UK Government (Department for International Development, DFID) to provide humanitarian agencies involved in the response to Cyclone Idai with scientific information on flood hazard and population exposure.

Forecasts from GloFAS were used in combination with satellite imagery from Copernicus and flood extent maps from the University of Bristol to identify where, when and for how long flooding may occur, as well as where people may be impacted. Every two days, a flood emergency briefing was sent to DFID and shared with partners, including the United Nations Office for the Coordination of Humanitarian Affairs (UN OCHA), to help them plan their humanitarian actions.

UN OCHA and DFID subsequently asked ECMWF and the Universities of Reading and Bristol to provide flood briefings for Cyclone Kenneth from 24 April 2019, ahead of its landfall in northern Mozambique on the evening of 25 April. DFID told the team that the briefings had been “well received” and that “UN humanitarian response actors stated that the reports produced were tremendously helpful”.