Coupled model based operational extended range forecast of temperature over India during winter season of 2020-2021

Abstract

Agriculture, health, transportation, and hydrology are among the many sectors that rely significantly on intra-seasonal temperature fluctuations during the winter. The intra-seasonal variability of minimum (Tmin) and maximum temperature (Tmax) is observed across India during the winter season of November to February (NDJF). This study analyse the real-time extended range forecast (ERF) skill of Tmin and Tmax over India during the winter season, i.e. NDJF of 2020-2021, using the Climate Forecast System version 2 (CFSv2) coupled model, operational in India Meteorological Department. The statistical metrics, such as forecast accuracy, bias, probability of detection (POD), false alarm ratio (FAR), probability of false detection (POFD), critical success index (CSI), and the equitable threat score (ETS), are used to assess the temperature forecasting capability. A four-week quantitative comparison of observed and models predicted Tmin is carried out, utilising every Wednesday’s initial conditions as the starting point. Throughout central India and northwest India, the intra-seasonal variability of observed Tmax and Tmin revealed a considerable reduction and rise in temperature over the winter season, with Tmin fluctuation showing greater variability than Tmax variability. The trend and intra-seasonal variations in Tmax and Tmin over India were well reflected in the real-time extended range forecast during the season, up to 2 to 3 weeks. In a homogeneous region, northern and central India have higher skill levels than the southern peninsula and northeast India. At the district level in central and northwest India, categorical and quantitative forecasting ability was also found to be much more significant. As a result, for applications in various industries, the operational ERF of Tmax and Tmin with a lead time of 2 to 3 weeks may provide a reliable forecast.