Back in February, a Japanese scientist made a forecast which should have had India worried right away. Toshio Yamagata, head of the Application Laboratory of Japan Agency for Marine-Earth Science and Technology, warned that the Indian sub-continent will experience a weaker monsoon this year. Around the same time, World Weather Inc. in the US cautioned that India should brace up for a relatively dry spell in August and September. Similar words of caution started coming in from other forecasters as well.
But the Indian Meteorological Department (IMD) brushed aside these warnings saying it was too early for a reliable forecast. When it did issue its first forecast in April, it said India would get 98 per cent rainfall this year. It was less than the ideal 100 per cent but was still robust. As time went by and rain continued to play truant, IMD scaled down its forecast to 96 per cent. There was no cause to panic still. Close to July end, the actual rainfall has been 78 per cent. That’s a worrisome deficit of 22 per cent — in 2009, when the last drought happened in India, the deficit was 23 per cent.
In the last 130 years, IMD has never forecast a drought, neither in 1987, 2002, 2004 or 2009. Each time, it predicted normal or near-normal rainfall. In the last 12 years, not once has IMD got the forecast right. The closest it came was in 2005 when the prediction was 98 per cent and the actual rainfall was 98.8 per cent. The worst was in 2002, when the forecast was 101 per cent rainfall, while the country actually received 79.4 per cent. A senior IMD official, who does not wish to be named, says that in 1979 when India witnessed a drought, the meteorological department’s forecast was that rainfall would be a poor 78 per cent. “But it was decided that this information will not be made public,” he says. “I don’t know why that was decided.”
|
IMD’s first forecast vis-à-vis actual rainfall |
1st forecast
(% of LPA)
rainfall
(% of LPA)
forecast and
actual rainfall
Source: PHD Research Bureau, compiled from CMIE
Jatin Singh, founder and CEO of Noida-based Skymet, India’s oldest private weather forecasting firm started in 2003, minces no words: “It’s a political problem. IMD’s hands are tied. It can never forecast a weak monsoon. That will create panic in the market.” Senior officials of several companies with stakes in the rural market also indicate that this is a factor. “Can you imagine the panic it will cause among farmers,” asks the head of a company which has sizeable business in the rural market. Prices of farm commodities would shoot up, the stock markets would go into a tailspin — it would be politically disastrous. IMD Director General L S Rathore does not respond to the question on whether declaring a normal or near-normal monsoon year after year was a political compulsion.
* * *
‘Political compulsion’ is only one part of the picture. The meteorological department does have a genuine problem when it comes to long range forecast. IMD’s work is divided between Pune and Delhi. The team in Delhi is focused on “nowcast” (forecasting up to six hours), short-range forecast (up to three days) and medium-range forecast (up to seven days). These forecasts are for a smaller area, more specific and hence more difficult. Yet, IMD scores better on this front. One reason for this is the dynamic model it uses for forecast. It simulates three-dimensional atmospheric conditions, and, based on certain prognostic equations, forecasts the state of the atmosphere in the future.
The long-range forecasts are done in Pune. And this is where the problem arises. Instead of the dynamic model, IMD uses the statistical model here. Instead of simulations of current weather conditions for predictions, one goes by historical experience in this model. Like, what will be the impact on rainfall be if there is an El Nino effect? “For example, 60 per cent of all droughts in India are related to El Nino (shift in ocean temperatures and atmospheric conditions in the tropical Pacific that disrupts weather around the world). This is historically established,” says Singh of Skymet. But the statistical model is a linear model. It does not prepare IMD for any uniqueness which might, and often does, creep in.
To address this problem, the Indian Institute of Tropical Meteorology (IITM) in Pune is working on a Rs 400-crore National Monsoon Mission. The project’s main aim is to forecast extreme events such as drought and floods, active and dry spells of monsoon and, more importantly, get information on how rainfall will be distributed all over the country. It has started using a dynamic model which has been developed by the US-based National Centers for Environmental Prediction. “The dynamic model involves using supercomputers and land and sea surface temperature data to predict the monsoon course,” says an IITM functionary. “By using coupled (or hybrid) models (both statistical and dynamic), we can collect [even more] more information.” Statistical models will still continue to be used but the idea is to graduate to dynamic models. Countries like US and UK have been using dynamic models for forecasting, and their met offices have been more accurate in their predictions than IMD.
The Monsoon Mission is being aided by a recently-acquired 2.5 petaflop supercomputer. IITM officers are not yet ready to talk about it. “Until now, IITM scientists have struggled to forecast extreme events — like floods or drought — in the medium- or long-term,” says Ajit Tyagi, a former director-general of IMD.
* * *
Every year, companies that have stakes in the rural markets refer to monsoon data. It helps them plan their go-to-market strategies. How useful do they find the IMD data? “Though we rely on long-term forecast to plan and budget our business, we do keep our ears to the ground and also employ the services of private forecasters who give us more localised forecasts,” says Rajiv Sinha, the joint managing director of DCM Shriram Consolidated which has interests in fertilizers, sugar, cement, seeds and rural retail. While IMD has said there is no need to panic, Sinha fears business will suffer 10-12 per cent due to the deficit in the monsoon rains “which seems to be happening this time”. Sources at Zuari, India’s largest private importer of fertilizers which has been tracking IMD’s monsoon forecasts closely, say they are hopeful that the meteorological department’s forecast will finally hold true and the rains will pick up by July-end.
Singh, meanwhile, is bracing himself for the worst scenario. If the rain deficit remains at 20 per cent by the end of July, August rainfall will have to be 30 per cent above normal for it to be a normal monsoon and this, he says, “is not going to happen.”
