Status:Ongoing analysis Material from: Linus, Ervin, Mohamed


1. Impact

On 14 May the tropical cyclone Mocha made landfall in Myanmar and also impacted Bangladesh. The cyclone was during one stage one of the strongest ever in the Bay of Bengal. Significant evacuations were made ahead of the cyclone.

2. Description of the event

See Wikipedia article.

3. Predictability

3.1 Data assimilation

Between the forecasts from 13 May 00UTC and 13 May 12UTC, the intensity in the forecasts (both o-suite and e-suite) decreased significantly, while the cyclone continued to intensify in reality. Impacting the available observations, large differences was found to one specific buoy, which TC Mocha passed close to. During the passage the pressure observations suddenly raised instead to decreasing. An experiment was run without this buoy, and in that case the intensity was improved in the analysis.

These situations with instrument issues during passages of extremes (which could be the reason for the instrument faults) are tricky as the large EDA spread puts more weight to observations. In order to make the system less sensitive to these situations, one option are to use the variational quality also in the first minimisation of 4D-Var. Experiments for this case shows that it mitigated the problem to some degree.

3.2 HRES

3.3 ENS

The plots below show the tropical cyclone track for the operational ECMWF forecasts from 14 May  00UTC (first plot) to 8 May 00UTC (last plot). The symbols shows the position on 14 May 12UTC. HRES (red), ENS CF (blue), ENS PF (grey) and BestTrack (black).

Same as above but for 48r1 e-suite.

Central pressure (top), maximum wind speed (middle) and propagation speed (bottom) for HRES (red), ENS CF (blue), ENS PF (grey) and BestTrack (black) from o-suite.

Same as above but for e-suite.

3.4 Monthly forecasts

The plots below show the weekly tropical storm probability 8-16 May, with different initial dates.

The plots below show the weekly tropical storm activity anomaly 8-16 May, with different initial dates.

3.5 Comparison with other centres

4. Experience from general performance/other cases

5. Good and bad aspects of the forecasts for the event

  • Very good predictability of the genesis
  • Northward (left-hand side) track errors in the late medium-range and also short-range
  • Slow propagation speed error in the medium-range (could be related to the track error)
  • Issues with one buoy caused problems in the data assimilation.

6. Additional material