PRESS RELEASE

New Delhi, 15 April 2004

 

INDIA METEOROLOGICAL DEPARTMENT

 

Long Range Forecast for 2004

South-West Monsoon Season (June-September) Rainfall

 

 

1.          Background

 

In 2003, IMD introduced several new models for Long Range Forecasts of South-west Monsoon season (June-September) rainfall over India and adopted a 2-stage forecasting strategy. The first forecast for South-west Monsoon season (June-September) rainfall for the country as a whole was issued on 16 April, 2003 based on the newly developed 8 Parameter Power Regression and Probabilistic models.

 

In the second stage, the forecast update for South-West Monsoon season (June-September) rainfall for the country as a whole was issued on 9 July, 2003 based on the newly developed 10 parameter Power Regression and Probabilistic models.  In addition, forecast for July rainfall over the country as a whole and forecasts for South-west Monsoon season (June-September) rainfall for 3 broad homogeneous regions of India were also issued.  

 

IMD's new operational models have been peer reviewed and the results were published in Current Science (Vol.86, No.3, 10 February 2004, pp 422-431), a leading international research journal published by the Indian Academy of Sciences, Bangalore. The performance of these new models was proved to be correct for the year 2003.

 

Based upon the newly developed 8 parameter Power Regression and Probabilistic models, which use data up to March, IMD has prepared the Long Range Forecasts for the 2004 South-west Monsoon season (June-September) rainfall over the country as a whole.

 

2. Details of the operational models

 

          IMD uses 8-parameter Power Regression and Probabilistic models for the long range forecasts issued in April.  Both the models use the same 8 regional and global parameters as the predictors, which need data up to March. The Power Regression model is a statistical model based on a non-linear regression technique. Data of 38 years (1958-1995) have been used for the model development and the model error is ± 5%.

 

 

 

 

The Probabilistic model is based on the Linear Discriminant Analysis (LDA) technique, which gives the probability of monsoon season rainfall for the country as a whole in 5 broad rainfall categories.  The Probabilistic model was developed using 40 years of data (1958-1997).

 

3.  Forecast for 2004 South-West Monsoon Rainfall

 

Based upon the newly-adopted 8-Parameter Power Regression Model, IMD's Long Range Forecast for the 2004 South-west Monsoon season (June-September) is that the rainfall for the country as a whole is likely to be 100% of the Long Period Average (LPA) with a model error of ± 5 %.

 

For the 2004 South-west Monsoon season (June-September) for the country as a whole, the new 8-Parameter Probabilistic model indicates  

 

16% Probability of below normal rainfall (90 to 97 % of LPA)

 

58% Probability of near normal rainfall (98 to 102 % of LPA)

 

18% Probability of above normal rainfall (103 to 110 % of LPA)

 

4% Probability of excess rainfall (more than 110 % of LPA)

 

There is only 4% probability for South-west Monsoon season (June-September) rainfall to be deficient (rainfall less than 90%).

 

IMD will issue a Long Range Forecast update in the last week of June 2004.

 

--------------------