Status:Ongoing analysis Material from: Linus, Esti, Ivan
1. Impact
Early on 26 August tropical cyclone Harvey made landfall on the TExas coast between Corpus Cristi and Rockport. Later the cyclone became stationary for 3 days (check) and a huge amount of rainfall fell over the region (locally more than 1000 mm) and caused massive floodings in e.g Houston.
2. Description of the event
The cyclone formed on 18 August from a disturbance east of the West Indies. The cyclone later made landfall on Yucatan on 22 August. After entering the Gulf of Mexico the cyclone re-intensified. The cyclone made its second landfall on the Texas coast on 26 August.
The plot below shows the precipitation totals over Texas (to be added).
3. Predictability
3.1 Data assimilation
The plots below shows observation statistics for temperature from dropsondes along the path of Harvey within 1 degree (left) and 5 degrees (right).
The plots below shows the observation statistics for dropsondes for temperature (left) and wind vector (right) for 24 August 00z, when the Global Hawk mission dropped a lot of sondes.
The plots below show the same as above but for 24 August 12z.
3.2 HRES
The series of plots below shows 3-day accumulated precipitation (26 August 00z to 29 August 00z) and MSLP valid 27 August 12z from HRES forecasts.
The plot below shows the evolution of the central pressure in Best Track (black) and HRES forecasts with different initial times. The forecasts before 25 August missed the rapid intensification the day before landfall. Also note how the forecast from 24 August 12z kept the intensify for a long time, probably because the cyclone stalled closer to the coast as seen above.
3.3 ENS
The plots below show the tropical cyclone product for TC Harvey.
The figure below includes 3-day EFI and the precipitation from NEXRAD.
The plot below shows the ensemble evolution for 72-hour precipitation for Houston 26-29 August. The ensemble distribution is shown in blue box-and-whisker and HRES in red dot. The model climate is shown in red box-and-whisker positioned on 26 August.
3.4 Monthly forecasts
3.5 Comparison with other centres
4. Experience from general performance/other cases
5. Good and bad aspects of the forecasts for the event
- Early capturing of the risk for moving into Gulf of Mexico
- Picked up the risk the risk for heavy rainfall (especially HRES) from 21-22 August. Model generated more than 1000 mm/72h locally
- Did not predict the rapid intensification before landfall.