VISHWAS PATIL, A RICE FARMER in Khed, Maharashtra, trusted the Indian Meteorological Department’s (IMD’s) prediction of adequate rains on time in June this year. The kharif or autumn harvest season for rice and other similar crops begins when the rains start, and there’s always a heavy demand for itinerant farmhands. Patil thought he would steal a march over his neighbours, and, based on the IMD forecast, hired extra farmhands for a week. The prediction was off by three days, sowing was delayed, and Patil ended up paying the workers for nothing.
Farmers need accurate long-term weather forecasts to ensure that planting and harvesting are not adversely affected. But agriculture is not the only segment of the economy that relies heavily on weather data. Commodities traders depend on them too. In FY11, the total turnover of commodity exchanges (MCX and the National Commodity Exchange) was Rs 112 lakh crore, up 50% from the previous year. The Indian market prices for agricultural commodities such as black pepper and cumin seeds already play a role in determining their global prices. The demand and supply of wheat in India also moves international markets. Since a major part of commodities trading involves futures, long-term predictions help traders get a good idea of the success or failure of specific crops, and they buy or sell accordingly.
Insurance companies also need similar data. So far, private insurers such as Bajaj Allianz and Tata AIG have been reluctant to offer crop insurance, since there were too many uncertainties. Between 1999 and 2010, for kharif crops alone the Agricultural Insurance Company of India collected a total premium of Rs 750 crore but ended up paying Rs 1,920 crore. With accurate weather data at their disposal, insurance companies can choose to cover farmers or seed companies against unforeseen weather conditions. For instance, ICICI Lombard, IFFCO-Tokio General Insurance, and Bajaj Allianz have begun insuring crops in Dindigul, a tomato growing area in Tamil Nadu, using weather data from the Tamil Nadu Agricultural University (TNAU), to estimate claims in advance and ascertain whether farmers are making false claims of bad weather.
THE PROBLEM FOR THESE SECTORS is that weather forecasts are provided by the IMD, which “has successfully predicted the monsoon only nine times in the last 23 years”, says a spokesperson of the Bharatiya Kisan Sangh, a farmers’ lobby. The IMD collects data from its satellite system, and processes it to provide short-, medium-, and long-term weather forecasts. The problem is, however, that the IMD’s equipment and ability to process raw data is increasingly suspect.
The government is aware of the IMD’s failings, and has recently set up a Rs 400 crore monsoon mission to improve weather reporting. The Indian Institute of Tropical Meteorology (IITM), set up as a unit of the IMD initially, but now an independent weather research department, has been given the mandate to determine a mathematical prediction model that will form the basis of future weather reports. B. N. Goswami, director of IITM, says: “Tropical climates are always difficult to predict and the recent global warming has only increased the complexities.”
While the government systems remain unreliable, companies are turning to the private sector for accurate data. Analysts conservatively estimate that the market size is around Rs 100 crore for customised weather data. That’s attractive enough to lure international biggies such as U.S. firm 3TIER and Britain-based GL Garrad Hassan, which have set up offices in Bangalore and Mumbai, respectively. There are also several domestic weather forecasting companies that have set up shop. Skymet Weather Services, Weather Risk, National Collateral Management Services (NCMSL), and even news agency Reuters and phonemaker Nokia, process the raw data they get from the IMD and create user-friendly, customised packages.
The market for these services is growing almost daily. State governments are among the biggest fish in Skymet’s net; other clients include private power distributors, such as Sanjeev Mehra, managing director of Noida-based Tata Power Trading, and Sanjeev Chauhan, general manager at Surat-based Torrent Power, who buy power at exchanges to make up for overnight shortages. Weather information is crucial for power companies and state electricity boards; if they know in advance that the temperature will go up, they can avoid a last-minute rush to buy power, which could cost up to four times more than usual. Skymet has also started an advisory for wind power companies.
Skymet is one of the early private players in this space. Started in 2003 by Jatin Singh, the company sources raw data from the IMD, the Indian Navy, and other similar sources. Its team of trained meteorologists processes the data and creates computer-based weather models, using servers from companies like Amazon. “We took a shortcut instead of spending money up front, and it has worked for us,” says Singh. Business is brisk; Singh says Skymet has been growing at 35% to 40% annually for the last six years.
Skymet’s shortcut is that it did not invest in equipment in the early stages. The traditional method of weather forecast—spend a few crores of rupees on instruments; automatic weather stations, lightning radars, or balloons to collect atmospheric data—can cost between Rs 5 lakh and Rs 1 crore. An S-band radar with a 500 kilometre radius costs about Rs 15 crore, while an X-band one with a smaller radius costs Rs 3 crore. The data is then analysed for short-and long-term predictions by computers with high processing power. Long-term predictions, which involve studying parameters such as wave heights and volcanic eruptions, use supercomputers that cost millions of dollars.
Despite not owning even processing power, Skymet says it has predicted rainfall better than the IMD; for 2010 and 2011, average accuracy in the rainy months of June to September was 70% to 80% compared to the IMD’s 50% to 60%.
AMONG SKYMET'S PROMINENT CLIENTS ARE state governments, trying to get focussed weather solutions to cater to their large rural populations. In Tamil Nadu, the IMD had only 16 observing stations and put out broad zonal reports for the southern or northern part of the state. It did not supply region-specific data, which farmers needed. Says Ajit Tyagi, a retired director general of the IMD: “The IMD did not focus on value-added services for companies or farmers.”
Some three years ago, the state determined that as farmers shift towards cash and short-term crops, they need more specific information to manage farming. Therefore, TNAU installed 224 weather stations to help farmers. These automatic stations, which are powered by solar panels, collect data on 10 weather parameters, including temperature, humidity, soil moisture, and leaf wetness. The data is transmitted hourly through mobile GPRS services. It is then centrally collated on the university server, analysed, and uploaded for public access. The university trained 750 agricultural officers to disseminate the information to farmers.
M. Rajavel, who is in charge of TNAU’s weather systems, says: “The state government is planning to expand the scope of the project as there is clear evidence that the flow of data has increased crop yield and reduced wastage.” TNAU is now getting enquiries from Hindustan Unilever, which wants to plan its sales strategy in rural areas based on estimated agricultural incomes.
Meanwhile, in Maharashtra, state agriculture minister Radhakrishna Vikhe Patil, fed up with the IMD’s predictions, has announced that the state will soon float a tender worth Rs 100 crore to set up 2,000 weather monitoring stations in partnership with the private sector. “The IMD’s predictions can’t be relied upon for commercial purposes,” he says.
Vikhe Patil says Maharashtra needs weather systems in a big way. Unlike neighbouring Gujarat, where 50% of farmland is irrigated, Patil says 82% of land in Maharashtra is not irrigated and depends on rainfall. To move away from this dependence and bring more area under farming, Vikhe Patil is advocating a new Rs 10,000 crore five-year programme to farm dry land in eastern Maharashtra. This will mean investment in micro-irrigation and climate control systems. Vikhe Patil, himself a farmer and agricultural university graduate, says: “The new project is an experiment where every parameter will be controlled and monitored, including the weather.”
Vikhe Patil wants to set up automatic weather stations through a public private partnership. The state government will pay Rs 40 crore with the rest coming from the partner. The goal is to access region-specific weather information, which the government can disseminate to farmers, while the private partner sells data to commercial customers.
As the new weather reporting business grows, the IMD is also beefing up its systems, slowly installing more weather reading stations. Goswami says that the IMD’s short-term predictions have been increasingly accurate after the adoption of a new, high-resolution computer-based predicting model two years ago. He admits, however, that long-term prediction will remain tricky for some time as the model needs to be brought up to date.