Modern technology has helped improve the art of weather forecasting in India, but accuracy in predicting extreme weather events such as droughts remains elusive, adding an element of uncertainty to the country’s economic planning.
Getting monsoon forecasts right is essential for more than 230 million Indian farmers who rely on the seasonal rains from June to September to water more than half the country’s farmlands.
Yet, over the past seven years, Indian weather experts got their forecast on monsoon rains wrong thrice. They predicted an average monsoon year in 2009 but the season ended with a rainfall deficit of 22%, making for the worst monsoon in nearly four decades.
In 2007, in contrast to the forecast of below average rainfall, the 2007 monsoon turned out to be above average.
This year, against a forecast of below average rains, the seasonal showers have been delayed over most parts of north India by more than two weeks, raising fears of another year of drought. Already, farm minister Radha Mohan Singh has warned of a drought-like situation in parts of western India.Read: Drought fears loom over north India
In 2012, the government conceded in Parliament that the accuracy of long-range monsoon forecasts in the previous four years had only been about 50%. A year before that, a Deutsche Bank report said that since 1994, India’s weather office had only managed to forecast the June-September monsoon outcome correctly five times, discounting an error band of +/-5%, while on seven occasions the extent of error touched double digit.
While an improved forecasting model, rolled out last year, is helping the Indian Meteorological Department issue four to five monsoon forecasts instead of two, and provide region-specific information as well, the reliability of the forecast is suspect.
This is crucial for about 600 million Indians who are dependent on farming because there is a direct correlation between ample rains and their disposable incomes. This also makes monsoon forecasting a political hot potato. Farmers are the single largest factor influencing politics in India.
Back in 1886, when the India Meteorological Department made its first monsoon forecast, the pace of melting of snow on the Himalayas was used to predict the timing and quantity of rains.
Now, Indian officials use modern radars and complicated statistical models aided by super computers and satellite data to predict weather patterns. Statistical forecasts as well as more experimental dynamical forecasts methods are used.
In statistical forecasts, officials rely on inputs from up to six measured parameters that affect the monsoon, such as evolving conditions in the Pacific Ocean and snow cover conditions over Eurasia, officials say.
The problem is that they don’t represent the evolving physics of the situation, meaning they can quite easily go wrong.
Dynamical forecasts are those from coupled-ocean atmosphere models, rather like weather forecasting models, run for the season ahead and started from our current best observations of the atmosphere and ocean.Read: Moonsoon delayed; govt readies contingency plan
India is particularly vulnerable to predictions of drought, officials say.
The Indian weather office now forecasts monsoon rains based on its long-term rainfall data for the past 130 years. There are two annual forecasts, one in April and the other in June, after the monsoon rains cover half the country.