Why IMD’s rain math doesn’t add up

Each blue bar represents the actual rainfall recorded by a district as a percentage of its normal. There are 614 district recordings in this chart.. The red dotted line is the 100% mark, and many of the bars end below, or way below, this mark. This is the district-level view of cumulative rainfall over eight rain weeks using the IMD’s own data.

Over eight weeks of recorded monsoon rain, the district-level data available with the India Meteorological Department (IMD) portrays a picture that is very different from its ‘national’ and ‘regional’ advice about the strength and consistency of rainfall.

In its first weekly briefing on the monsoon of August 2014, IMD said: “For the country as a whole, cumulative rainfall during this year’s monsoon (01 June to 30 July 2014) has so far upto 30 July been 23% below the Long Period Average.” Out of 36 meteorological sub-divisions, said the IMD, the rainfall has been normal over 15 and deficient over 21 sub-divisions.

The four regional readings that make IMD’s data look less worrisome than it actually is.

However, here is a far more realistic reading of the monsoon season so far, from the IMD’s own data. For the 614 individual readings from districts that have regular rainfall readings, we have the following: 86 districts have registered scanty rainfall (-99% to -60%); 281 districts have registered deficient rainfall (-59% to -20%); 200 districts have registered normal rainfall (-19% to +19%); and 47 districts have registered excess rainfall (+20% and more).

What this means, and the chart I have provided to illustrate the 614 individual values leaves us in no doubt, is that 367 out of 614 districts have had meagre rain for eight weeks. This also means that over eight weeks where there should have been rainfall that – as the IMD predicted in early June – would be around 95% of the ‘long period average’, instead three out of five districts have had less than 80% of their usual quota.

An India economical with monsoon truths

Monsoon measures for six weeks. A few more districts reporting the revised normal, but the deficient-2 category still has too many districts, and so does excess-2. And why so many ‘no data’ (many from the north-east)?

When a politician and a bureaucrat get together to supply punditry on the monsoon, the outcome is directionless confusion. There is no reason for our shared knowledge on monsoon 2014 to be reduced to a few boilerplate paragraphs and a couple of amateurish maps and charts, not with the equipment and scientific personnel the Republic of India has invested in so that we read the rain better. But Jitendra Singh, the Minister of State who is in charge of Science, Technology and Earth Sciences, and Laxman Singh Rathore, the Director General of the India Meteorological Department, have not progressed beyond the era of cyclostyled obfuscation.

The Press Information Bureau reported Singh as saying that there has been “significant increase in the monsoon during the last one week beginning from 13th July, and the seven days between last Sunday and this Sunday have recorded 11 percent increase in the monsoon country-wide”. Following suit, Rathore said: “The monsoon deficit has come down by 12 per cent and the overall deficit stands at around 31 per cent. This will bring in much needed relief to the farmers and solve the water issues.”

Coming from senior administrators, such fuzzy distraction cannot be tolerated. The Ministry of Earth Sciences, the India Meteorology Department and the Ministry of Information and Broadcasting must cease (desist, stop, halt – do it now) the use of a ‘national’ rainfall average to describe the progress of monsoon 2014. This is a measure that has no meaning whatsoever for cultivators in any of our agro-ecological zones, and has no meaning for any individual taluka or tehsil in the 36 meteorological sub-divisions. What we need to see urgently adopted is a realistic overview that numerically and graphically explains the situation, at as granular a level as possible.

When that does not happen, news media and information sources struggle to make sense of monsoon and climate and their reporting becomes dangerously misleading – consider “Late monsoon starts Indian farmer’s ‘journey to hell’ “; “Why the monsoon numbers hide reality” (this report is an attempt to point out the problem); “Monsoon deficit has come down to 31 per cent, no need to be ‘alarmist’: Met office”; “Satisfactory rainfall may wash away monsoon deficit”.

Using a revised categorisation of rainfall sufficiency levels (my method and the reasoning for it use is available here) we find that for the fifth and sixth weeks of monsoon, there has been a small improvement which does not lower the high likelihood of drought conditions becoming prevalent in districts and scarcity of water for our settlements – Messers Singh and Rathore please note (or visit the Indian Climate Portal Monsoon 2014 page which is an active repository of reportage, views, commentary and original data analysis on our monsoon).

The fifth monsoon week is 03 to 09 July 2014 and the sixth monsoon week is 10 to 16 July 2014. There has been a small addition to the revised normal rainfall category (-5% to +5%), rising from 18 districts recording normal rainfall in the 4th week to 22 in the 5th and 28 in the 6th. There has also been an improvement in the number of districts recording deficit-2 levels of rainfall (-21% and more) with 437 in the 4th week, 411 in the 5th week and 385 in the 6th week. For the remainder of July the likelihood of more rainfall in the districts that have recorded normal or excess-1 (+6% to +20%) is small, according to the available forecasts, and this means that monsoon 2014 will begin August with far fewer districts registering normal rainfall than they should at this stage.

The NOAA map of the land and sea percentiles. Note the warm water south of India and to the east of the Philippines.

With many sowing cycles beginning belatedly between now and the end of July, the Ministry of Earth Sciences, the India Meteorology Department, the Ministry of Agriculture and the Ministry of Water Resources are advised to work together (why aren’t they doing so already – or at least mandating ICAR institutes to supply them with analysis which they must absorb jointly?) to understand the impacts of regional, tropical and global climate trends that affect the Indian summer monsoon.

There is good reason to. According to NOAA, for 2014 June land and ocean surface temperatures jumped 0.72 Celsius above the 20th century average. These new records were pushed upwards by a broad warming of the ocean surface, and not only by an Equatorial Pacific whose waters are approaching the warmth usually seen during an El Nino. NOAA has said there was “extreme warming of almost every major world ocean zone” which warmed local air masses and had a far-reaching impact on global climate, “likely delaying the Indian monsoon”.

Monsoon 2014 and a third dry week

05 to 11 June is the first week. 12 to 18 June is the second week. 19 to 25 June is the third week. The bars represent the weeks and are divided by IMD’s rainfall categories, with the length of each category in a bar showing the proportion of that category’s number of districts. The colours used here match those used in IMD’s weekly rainfall map (below) which displays the category-wise rainfall in the 36 meteorological sub-divisions (but not by district).

IMD’s weekly rainfall chart, 19 to 25 June

We now have rain data for three complete weeks from the India Meteorological Department (IMD) and for all the districts that have reported the progress of the monsoon.

The overall picture remains grim. In the third week of the monsoon the number of districts that reported normal rains in that week (-19% to +19% of the average) is only 74. No rain (-100%) was reported by 71 districts Scanty rain (-99% to -60%) was reported by 221 districts, deficient rain (-59% to -20%) was reported by 125 districts, excess rain (+20% and more) was reported by 129 districts, and there was no data from 21 districts.

The Department of Agriculture and Cooperation, of the Ministry of Agriculture, has already issued its guidance to states on the contingency plans to be followed for a delayed monsoon. That is why it is important to make available the district-level normals and rainfall departures – the meteorological sub-divisions are too broad for such analysis and are irrelevant to any contingency plans and remedial work.

By end-June, when the IMD updates its outlook for the rest of monsoon 2014, we expect more detailed assessments of the districts to be publicly available – the agromet (agricultural meteorology section) already provides this to the states, with state agriculture departments given the responsibility of ensuring that the advice – which is especially important for farmers to plan the sowing of crop staples – reaches every panchayat.

The new measure of monsoon

Districts reporting monsoon data, over two weeks, colour-coded under a revised categorisation (explained in the text) for weekly rainfall. The left bar in each pair is the second week, the right bar is the first. Most districts are coloured light red, signifying rainfall much below the weekly normal. Peach is for the lesser deficient category. Green is normal. The two blue hues – lighter and darker – are for the two excess categories. It is immediately apparent that 485 out of 618 reporting districts (78%) have experienced less rainfall than they should have at this stage of the monsoon.

The changes that we find in the patterns, trends, intensity and quantity of India’s monsoon now require an overhaul in the way we assess what is satisfactory or not for environmental and human needs. India’s summer monsoon is already late, and where it is late but active it is weak. The indications from the central earth science agencies (including the India Meteorological Department), from the Indian Institute of Tropical Meteorology, from the National Centre for Medium Range Weather Forecasting are that it will be the end of June before the summer monsoon system settles over central India and the western Gangetic plains. Even so, it will be a relief from the searing temperatures but will not assure sowing conditions for farmers and cultivators, nor will it add to the stores of water in major and minor reservoirs.

In this commentary written for India Climate Portal, I have explained why IMD’s hoary top level categorisation of rainfall weekly quantities in the subdivisions must be replaced, both for what they describe and for how frequently they are described. These currently are: ‘normal’ in a subdivision is rainfall that is up to +19% above a given period’s average and down to -19% from that same average; likewise excess is +20% and more, deficient is -20% to -59% and scanty is -60% to -99%. The ‘normals’ are calculated based on the mean weekly rainfall for the period 1951-2000 with monitoring done in 641 districts distributed amongst the 36 meteorological subdivisions.

By categorising rainfall ‘normals’ and departures from  ‘normal’ to become more administratively impelling – these proposed corrections also simplify the interpretations possible for rainfall above and below ‘normals’ – greater awareness and preparedness of administrations, key agencies and citizens to the deficiencies of monsoon can be fostered. For the district tables below therefore, I have re-cast the categories as follows (all based on the long-term average provided by IMD): Normal in a district is +5% to -5%; Deficient 1 is -6% to -20%; Deficient 2 is -21% and more; Excess 1 is +6% to +20%; Excess 2 is +21% and more.

Whereas, for the same second rainfall week the IMD categories were ‘No Rain’ in 80 districts, ‘Scanty’ in 241 districts and ‘Deficient’ in 130 districts, under the proposed revision they will simply be ‘Deficient 2’ with 449 districts – thereby showing dramatically how widespread the conditions of the late and weak monsoon 2014 are – and ‘Deficient 1’ with 36 districts. Please read the rest at India Climate Portal.