Real-time quantitative rainfall forecasts at hobli-level over Karnataka: evaluation for the winter monsoon 2010

by P. Goswami, V. Rakesh, G. K. Patra and V. S. Prakash

Advance and accurate forecasts of rainfall can aid many sectors, from agriculture to disaster mitigation. However, for effective application, such forecasts must be at relevant spatial scales; given the tremendous spatial variability of rainfall, only forecasts at high resolution can serve the user’s need.

Worldwide, such highlevel rainfall forecasting is in its infancy. It is recognized that scientific inputs through sustained evaluation, supplemented by region-specific customization, are necessary for improvement of the current level of skill in high-resolution forecasting.

To create a science-based platform for advance and accurate forecast at relevant scales through such scientific efforts and interactive evaluation, CSIR Centre for Mathematical Modelling and Computer Simulation (C-MMACS) and Karnataka State Natural Disaster Monitoring Centre (KSNDMC, Govt Karnataka) have embarked upon a programme aimed at generating rainfall forecasts at the hobli-level. The goal of the programme is to generate and disseminate forecasts customized to user needs and applications, such as irrigation schedule and crop insurance. The most commonly used current strategy for high-resolution forecasting is to use Limited Area Models (LAM), three dimensional models of the atmospheric dynamics. The strength of these models is their ability to resolve surface forcings such as orography, vegetation and convective dynamics at scales not yet feasible in General Circulation Models (GCM). However, it is recognized that off-the-shelf models can not be expected to provide useful skill. World over, dynamical meso-scale forecasting is an evolving technology, with many challenges. The challenges grow as the spatial scale of the forecast is reduced; indeed, advance forecasting at hobli-level has not yet been tried in India. We have used a state-of-the-art LAM, called Weather Research and Forecasting (WRF) Model, calibrated for meso-scale forecast over Karnataka. The calibration and optimization were carried out based on methodologies developed at C-MMACS as well as elsewhere. The forecasts are generated using the C-MMACS High-Performance Computing Platform, and issued twice daily (morning and afternoon) through an Industrial Partner of C-MMACS (M/S Frontier Pusher); these forecasts are disseminated to various users by KSNDMC. Here we present an evaluation of real-time 24-h quantitative precipitation forecasts at hobli-level over Karnataka for the northeast monsoon season (October- December) of 2010. A statistical evaluation of the forecasted rainfall at hobli-level is carried out by comparing with the observations from the telemetric rain-gauge network established by KSNDMC. In addition, data from a variety of sources, including the CSIR Climate Observation and Modelling Network (COMoN) is used to improve the forecast through assimilation of local data. Several measures of forecast quality are used to evaluate the forecast skill. The objective at present is to provide accurate forecasts of rainfall category rather than the precise amount. The results show that forecasts capture the spatio-temporal variability in observed rainfall over Karnataka well enough to be useful. At the same time, certain areas of systematic bias are identified for further calibration to improve forecast skill over these areas.

You are here: Home

Working Hours

Open Monday - Friday, 8:30 AM - 5:00 PM IST

  This email address is being protected from spambots. You need JavaScript enabled to view it.

  Fax: +91-80-2522-0392

  Call: +91-80-2505 1920, 2505 1921