Singh’s articulation of the benefits of the study and the choice of data across 193 years highlights all that is wrong with current weather forecasting techniques and resulting policies. Much of contemporary weather forecasting here relies on narrow data sets, sometimes just “20-30 years”, according to Singh who proffers the example of Maharashtra to indicate just how this affects policy planning.
In July, Maharashtra declared near-drought conditions in the state because it received low rainfall. “The last four-five years (that has averaged around 900mm of rainfall) have just been an unusually good run for the state,” said Singh. “Look back from the 1900s and it’s frequently around 750mm.”
The 30-year spike
The model also appears to have busted a few myths surrounding the monsoon.
Meteorologists here refer to a 30-year cycle to explain rainfall patterns. According to this cycle, rainfall increases for a period of 30 years, aided by extremely high rainfall, even floods, in various parts of the country. Then, for the next 30 years it decreases, aided by near-drought conditions, even full-blown droughts. And then the cycle is repeated. Singh said his study disproves this.
A longer-term view shows “nearly 60 years of consistently low rainfall”, he said.
The study’s findings become especially significant when transposed on an irrigation map of India with many of the states and regions experiencing a decline in rainfall levels not having developed their irrigation networks to the extent they should have. According to data from the ministry of water resources (which is updated only till the 9th Plan), only 11 states have developed their irrigation facilities to between 60% and 90% of their potential, with most states having developed them only to between 40% and 60% of their potential.
“There’s also a limit to the extent of irrigation facilities that can be developed, because of restrictions such as sustainability of structures and subsequent water salinity issues,” said Srinivas Jain, researcher at the Indian Institute of Technology, Delhi, who specializes in water management studies. “Then, there are inter-state water disputes, all of which means that access to water is a major problem and rainwater harvesting is our only way out.”
Pattern recognition
Madhavan Rajeevan, senior scientist at the National Atmospheric Research Laboratory, an Indian Space Research Organisation body, who was part of a similar study conducted in 2006 by the India Meteorological Department, or IMD, that analysed rainfall data between 1901 and 2003, said some regions such as Kerala and Madhya Pradesh showed declining trends in their rainfall.
“It’s an interesting and useful study, in that it’s the first time pre-1990 data is used to study the climate then, but the new study doesn’t substantially change our understanding of rainfall trends,” Rajeevan said.
The new study and its results come at a time when Indian meteorologists have moved away from historical data-based models because of their limited success (perhaps on account of considering limited data sets).
Most scientists here have moved to so-called general circulation models (GCM) for forecasting rainfall. These models rely more on simulating the atmosphere and physical conditions on a given day and then extrapolate them to see how the atmosphere behaves.