If Indian Meteorological Department has to sustain its near 100 per cent success rate in predicting the onset of monsoon, the southwest monsoon has to touch down along the Kerala coast on June 6 or at the most on June 10. But is it just enough for the IMD to predict the onset of monsoon right? Should it also not say, with increasing accuracy, whether it would be a drought year or the year of deluge?
IMD’s indigenously developed statistical prediction technique, with a model error of + or – 4 days, has been bang on since 2005, the year it was introduced. But never has it been able to warn the people, with any certainty, about floods or drought. In the previous two years, the southwest monsoon has arrived on the dot, on exactly the date the IMD had predicted: on May 30 (2017) and May 29 (2018). In 2016 and 2014, it arrived a day after the forecast onset date of June 7 (2016) and June 5 (2014).
In 2013, the southwest monsoon descended slightly early on June 1 though it was predicted only on June 3. This was within the error bracket of + or – 4 days. In the eight years since 2005, till 2013, the onset of the monsoon had been slightly early or late than IMD’s prediction, never on the dot like in 2017 and 2018 but nonetheless it swung only within the error bracket of -4 days.
In 2006, the onset took place four days before the forecast date of May 30. And in 2007, it took place four days after the predicted date of May 24. In all other years, the maximum difference between actual onset and prediction was plus or minus two days.
It was in 2015 alone that the IMD prediction had gone a bit awry. The forecast was May 30 but the rains came only six days later on June 5. Fortunately, the wrong prediction was nothing ominous as was feared then. The monsoon was normal during that year.
This year, too, the IMD officials say that the monsoon will play true to script. "The monsoon clouds have advanced over Andaman Sea by May 27 and the cross equatorial monsoon flow has already appeared over Andaman Sea. These are clear signs that monsoon will strike on June 6 or 8," the official said.
The IMD's model, which is based on Principal Component Regression technique, uses the following six predictors: i) Minimum Temperature over North-west India, ii) Pre-monsoon rainfall peak over south Peninsula, iii) Outgoing Long wave Radiation (OLR) over south China Sea, (iv) Lower tropospheric zonal wind over southeast Indian ocean, (v) upper tropospheric zonal wind over the east equatorial Indian Ocean, and (vi) Outgoing Long wave Radiation (OLR) over south-west Pacific region.
Where IMD falls short
Predicting monsoon is the easy part, said Manoj, a weather expert working in CUSAT, Kochi. "The IMD has perfected the onset prediction. But it is the prediction of the active-break cycles that has turned to be problematic," Manoj said. The active-break cycle refers to the alternating periods of activity and lull during a southwest monsoon period, which stretches from June to the end of September.
The monsoon begins with a bang, and then after a time it goes dull. It picks up again and then, after a time, loses steam, takes a break. "We are looking at a drought, if there are three or more than three extended break periods of 14 days or more during the southwest monsoon period. (This had happened in 2013 and 2016). If the break periods are less, then we are looking at a deluge. (Last year, there was virtually no break period in between. The monsoon was almost fully active across its three-month existence in 2018)," Manoj said.
Even IMD officials admit that the country has to perfect its prediction technique when it comes to forecasting active-break cycles. "We are also fully aware that it is not just enough to predict the cycle. We have to do it fairly early, at least two weeks in advance," the IMD official said. "Now, at the most we can predict the cycle only five days before and this too will not be accurate," he added.
Weather scientists say this failure in foreseeing the active-break cycles has got to do with the convection pattern (the rise of hot air) in India. "Since India is very near to the equator, the sun falls nearly perpendicular on it. The convection therefore is quick and non-linear, making it hard to capture its effects in conventional thermodynamic models. We have to upgrade our technology to get this prediction more right," the IMD official said.
