Weather forecasts that err, even as they become more accurate

We have indeed come a long way past the era when the standard response to weather forecasts ('it may or may not rain') was one of ridicult,  but what we witness now is that weather belies our increased ability to precisely foretell it.

Two years ago,  several farmers fron Anantgaon in Maharashtra, lodged a complaint with the police against India Metereological Department (IMD);   another farmers' organization went a step further and threatened to close down the Pune station of IMD – and citing reasons.  IMD predicted in early June that there would be normal monsoon that year.  And on 12 June the authorities directed the drought-hit farmers of Marathwada to start their kharif crop planting.  It rained quite well till 20 June.  But thereafter rain retreated for several weeks.   By the time chief minister Fadnavis asked the farmers on 9 July that they need start planting only after 20^th,  they had planted more than three quarters of their crops.  And they all were destroyed.  The farmers alleged that such a big loss happened because IMD misled them.  Although that was true as far as they were concerned,   it can be argued based on fact that IMD's forecast,  based on the data and analysis available at the time,   was not wrong.

Despite handicaps like inaccuracies in data,  poor quality of equipment that analyse them and dearth of experts,  forecasts now ae less erroneous than earlier.  But the fact that weather changes happen faster than the speed of technology,  makes matters more complicated.  It is a rare examnple that this year southwest monsoon retreated all over the country and southeast monsoon hit the Tamil Nadu coast on the same day.   Not only is it rare for southwest monsoon to subside so late,  it is also unusual that from the start of the retreat it disappeared in so few number of days.   Even in actual showers, a rare phenomenon was experienced: in the first month of monsoon, i.e. June,  the showers this  year was less by a third.  But the following months,  it was torrential rain.

However much be the variations in monsoons year after year,  in the past there were certain consistent calculations and signs.  But as per indications now,  even those premises are not permanent.   A good example is the flood of last year.  If we take the total rainfall of 2018, we will not find anything unusual.  As per statistics since 1902,  there was more rain than last year's in 33 other years.  And even last year,  during the months of June,  July and September the rainfall was only at the normal scale.   But what turned the tables was the extremely heavy rainfall that came during the middle month of August,  especially during the last two weeks.  This year too, there came indications of the weather assuming  a new pattern of drought during months considered to be rainy,  and then heavy downpour in the intervening months.  Scientists see this as signs of global warming and global climate change. And this by itself,  is a matter that commands the urgent attention of the country.  Clubbed with the truth that a majority of the country's farmers still depend on agriculture,  is also the fact that 51% of the farming itself rely on rainfall.   When forecasts go wrong,  that affects farming and the lives of farmers.  Further,  it will also adversely impact the existence and safety of commercial and industrial enterprises of diverse segements of people.   Hence the urgent need of accurate weather forecasts.

Currently,  IMD and private agencies conduct weather forecasts between them with five different time spans – daily alerts,  alerts for three days,  ten days, a month and for a season.  For the last few years,  long-term forecasts have been going wrong.  This year's rainfall belied the forecasts of not only IMD,  but even of private agencies like Skymet.  When IMD predicted 96 per cent rain,  Skymet did 93 per cent;   and what we got was 110 per cent rainfall.  Even if we explain that they go wrong because of climate change,  there remains the imperative for greater accuracy.  In fact,  climate change itself is a phenomenon that is accurately forecast at global level.  The question is whether we do factor it into our forecasts.  Greater accuracy can be achieved,   at the same time when enhancing technology and human resources,  by including in study models the impact of climate change not affected by geographical boundaries also.    Further,  the customary concepts about monsoon need to be revisited in light of the new experience of heavy downpour in a span of a few days.  And everything right from day-to-day life to long term farming and disaster management will, have to be rejigged in line with that.