IMD to adopt dynamic model to make short-term forecasts

Written by Sandip Das | New Delhi | Updated: May 14 2012, 07:15am hrs
In what could prove to be a revolution of sorts for key sectors of the economy, such as agriculture, tourism, transport, sports and travel, the India Meteorological Department (IMD) is planning a foray into short-term weather forecasting on the lines of the models used in the US and Europe.

The IMD, currently, uses the statistical model to make seasonal climate forecast. This has often proved to be an unreliable method for predicting the rainfall pattern.

The statistical models are not proving too accurate and we dont have enough skills for using the dynamic models, Shailesh Nayak, secretary, ministry of earth science, told FE. Within the next couple of years, we will be able to make short-term forecasts using the dynamic model, he noted.

Using the dynamic model, IMD would shortly make five days' weather forecast for each region of the country. Nayak said under the R400-crore Monsoon Mission, approved by the Cabinet recently, Pune-based Indian Institute of Tropical Meteorology is working on improving long-range and seasonal forecasts, while the National Centre for Medium Range Weather Forecasting in Noida is experimenting on improving medium-range (up to a fortnight) forecasts.

IMD had predicted deficient rainfall for 2009. But the actual rainfall turned out to be far below the IMD estimates. IMD had also failed to predict the droughts of 2002 and 2004.Nayak admits that while IMD's macro forecast for the entire country had largely been accurate during the last few years, we have been unable to predict variability in the rainfall pattern in specific regions. Dynamic models are used for predicting rainfall over a smaller scale, which is not possible in the statistical forecasting system where the forecast is generated using various information received from agencies in the US, Europe and Australia. The southwest monsoon (JuneSeptember) rainfall, crucial to the economy, is currently calculated as the area weighted average of the seasonal monsoon rainfall data of all 36 meteorological sub-divisions in the country.

IMD recently predicted that the southeastern monsoon would be mostly likely normal, with a rainfall of 99% of the long period average (LPA), which is the average annual rainfall of 89 cm during 1951-2000. IMD makes its predictions on the basis of five parameters, which include the North Atlantic sea surface temperature, the North West Europe land surface air temperature, and the East Asia mean sea level pressure.