Verification of WRF model forecasts and their use for agriculture decision support in Bihar, India

. Weather forecasting with high spatial resolution has become increasingly relevant for decision support in agriculture and water management. Present work is carried out for verification of India Meteorological Department-Weather Research and Forecasting Model's (IMD-WRF) rainfall forecast with three day lead time over Nalanda, Supaul and East Chamaparan districts in Bihar, India. The model's skills up to a lead time of three days are evaluated with panchayat level daily in situ observations for Monsoon 2020 and 2021. Results confirmed show good agreement of forecast and observation throughout the domain and particularly over Supaul district, where about 70% of rain and no-rain days are correctly predicted for all panchayat. Also, False Alarm Ratio is < 0.3 in 90 percent of the total 812 panchayat and Hanssen and Kuiper's score is also found > 0.25 in almost all places. This evaluation supports the use of the WRF model forecast in agriculture up to three days in advance. However, the quantitative verification suggests that model output is more reliable for moderate rainfall.