Status: Finalised Material from: Linus, Ivan


 


1. Impact

Devastating floods affected south-western prefectures of Japan this weekend. At least 18 people had died and 14 were missing in the prefectures of Kumamoto and Kagoshima, with images from the hardest-hit areas showing houses completely destroyed by the strength of the flood waters. More than 400 mm were recorded in some places with more than 100 mm in just less than an hour. Large areas along the Kuma River were flooded with vehicles and buildings submerged and bridges destroyed. Mudslides destroyed houses and killed people. Excessive rain that affected Japan is related to northward propagating East Asian monsoon and quasi-stationary Meiyu front associated with this baroclinic zone separating warm and humid tropical air and cooler and drier polar air. MCSs and large rainfall totals are not uncommon along Meiyu front.

https://www.bbc.co.uk/news/world-asia-53304726

2. Description of the event

The plots below show analyses of MSLP and short forecasts of 6-hour precipitation from 1 July to 4 July (all 00UTC).

The plots below show analyses of T850 (shade) and z500 (contour) from 1 July to 4 July (all 00UTC).


The plot below show the accumulation since 7 June (thick) and daily precipitation (thin)  for station 47827 (inside the box in the plots above) for the past month. Observations (black), Short HRES (red) and climatology (grey).

3. Predictability

  

3.1 Data assimilation

 

3.2 HRES

The plots below show 24-hour precipitation on 3 July from observations (first plot) and HRES forecasts with 24-hours apart. One can note the the forecast from 1 July (third forecast plot) placed the front too far south.


3.3 ENS

The plots below show EFI for total precipitation on 3 July. The extreme rainfall was placed too far south in the forecast from 1 July, as in the HRES above).

The plot below shows the evolution of forecasts for 24-hour precipitation on 3 July in the box outlined above for HRES (red dot), ENS (blue box-and-whisker), model climate (red box-and-whisker) and average from observations inside the box (green). The drop in precipitation in forecasts from around 1 July is also visible here. (Model climate not added yet.)



3.4 Monthly forecasts

The plots below show weekly ensemble mean precipitation for 29 June to 6 July.

3.5 Comparison with other centres


4. Experience from general performance/other cases

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

  • The forecasts placed the front too far south in forecasts from around 1 July.


6. Additional material