Applicability of a modified weather research and forecasting (WRF) model to heavy rainfall information

Abstract The Japan Meteorological Agency (JMA) issues “Forecast Information on Linear Rainbands” (FILR; translated by the authors) approximately half a day before the expected formation of linear rainbands. A linear rainband is identified according to a set of operational criteria. A modified Weather Research and Forecasting (WRF) model was applied to sixteen heavy rainfall events associated with erroneous FILR decisions. Ten of these events had no FILR issued but subsequently experienced linear rainbands (missed issuance). Model results were compared with radar/rain-gauge analyzed precipitation data, with a primary focus on maximum rainfall, one of the criteria. The model reasonably simulated the occurrence of heavy rainfall in eight of the ten events and the rainfall intensity in six of those eight events. Conversely, six of the sixteen events had FILRs issued but ultimately experienced no linear rainbands (false alarm). In five of these events, the model reasonably reproduced the observed rainfall that did not satisfy the linear-rainband criteria. These results suggest that, despite several remaining challenges, the modified WRF model has the potential to complement JMA FILR decisions and contribute to reducing both missed issuances and false alarms.