India steps up its medium-range prediction

by Bob Henson

Medium-range weather forecasts—those spanning 3 to 10 days—might seem beside the point in the face of India’s legendary monsoon. But the nation’s summer rains arrive in pulses that are far from uniform, as was evident in the quirky monsoon of 2002 (see illusration and sidebar). Predicting short-term variations in the monsoon and other elements of Indian climate can make a big difference to agriculture and human culture.

That case has been proven through a decade of work at India’s National Centre for Medium Range Weather Forecasting. The modelers of NCMRWF are now expanding their facilities and their international ties. In July the NCMRWF hosted a two-day mesoscale modeling workshop in New Delhi that focused on the unique meteorology of the Himalayas, where the world’s tallest mountains tower above lowlands. The 90 participants included Al Kellie and Robert Gall (NCAR) and Karyn Sawyer (UOP Joint Office for Science Support). Also last summer, the center formally accepted a 24-processor Cray SV1 supercomputer acquired in 2001. It’s the first machine of its scope dedicated to the vagaries of Indian weather.

Founded in 1988, NCMRWF has earned a high profile in India by providing what might be the world’s most elaborate system for agricultural weather forecasting. The center employs NOAA’s Eta model, a global spectral model, and the Penn State/NCAR Mesoscale Model (MM5). Each day NCMRWF uses these tools to prepare a 120-hour forecast of key weather variables (with an effective lead time of four days) for each of 82 agrometeorological forecast offices across the country. Mass media outlets relay the outlooks from these offices to local farmers.

Among the success stories:

  • Wind and cloud forecasts help save an estimated US$200 per acre in annual costs of irrigation, pesticides, and fertilizer for onion, maize, and sorghum crops across South India.
  • Wind outlooks have allowed farmers in western India to reduce their insecticide and pesticide applications by 20%.
  • After timely warning of a major frost, farmers applied light irrigation to tomatoes and potatoes in North India, saving 30% of the crops that otherwise would have been lost. “Farmers who follow our advisories earn about twice as much profit as those who do not,” says L.S. Rathore, one of the lead meteorologists at NCMRWF.

A member of UCAR’s International Affiliates Program, NCMWRF recently signed a memorandum of understanding with NCAR and is building other model-oriented connections with the U.S. National Centers for Environmental Prediction, Florida State University, the University of Maryland, and NASA’s Goddard Space Flight Center. NCMWRF posts regular outlooks for parts of Kenya, Qatar, and Sri Lanka on its Web site at the request of those nations, and it wants to leverage its computing expertise to help developing countries in the region.

The door is open for other parties who’d like to work with the center, says NCMWRF head S.V. Singh. “We would

be interested in collaborating,” he says, on global and mesoscale models, data assimilation (including satellites and other remote sensors), thunderstorm simulation, ensemble and coupled ocean-atmosphere modeling, tropical cyclone forecasting, and visualization techniques.

The 2002 monsoon: not soon enough

Despite belated rains in August, the 2002 Indian monsoon marked the end of 14 generous years. Nationwide, the rainfall from this year’s monsoon came in at just 81% of average. It was the first monsoon since 1987 in which two criteria for an “all-India drought year” were met: national rainfall was at least 10% below average, and at least 20% of the country was drought stricken.

According to Someshwar Das, a lead scientist at NCMRWF, “The monsoon was quite unusual this year—which is quite usual, in the sense that no two monsoons are the same.” One quirky aspect was the accentuation of India’s usual west-to-east gradient in precipitation. The normally drier west had especially paltry rains this year, while parts of the east came in above average (see graphic).

The monsoon nearly sputtered to a halt in midsummer, with the driest July ever observed across India, before the rains returned in August. The stops and starts lent extra importance to the medium-range outlooks issued by NCMRWF. Virtually all the long-range forecasts issued last spring by the world’s top modeling centers had called for near-normal rainfall.

“It is possible,” says Das, “that some unforeseen changes in the planetary-scale atmosphere occurred during May that affected this year’s monsoon.” El Niño does tend to diminish the Indian monsoon, but there’s enough variability hidden in that relationship to keep it from being the most useful forecasting tool. To examine these and other possible cofactors, scientists from across the country met for a two-day brainstorming session in November at the Indian Institute of Science in Bangalore. Among the elements they scrutinized were an anticyclone parked over northern India from late June into August, a low-level inversion over the Arabian Sea, and extensive cyclone activity across the tropical Pacific.

“Results from modeling efforts and sensitivity experiments are still incomplete,” says Das. After these are discussed at a follow-up brainstorming session early in 2003, “we hope that we shall have some more insight about the episode.”


Also in this issue...

Putting travel risk on ice

More light, less heat

NSF center boosts space weather modeling

Science Bit: Acid-catalyzed process helps multiply particulates

Observationalists meet to map out HIAPER sensors

Training the next generation of emergency managers

Unidata’s new director: Mohan Ramamurthy

President’s Corner: Weather, Climate, and the Evolving U.S. Climate Change Science Program

Web Watch

UCAR Community Calendar

Governance Update