Misreading monsoon

As usual in May, there is a welter of forecasts and opinions about the monsoon, the great majority of which are short on understanding and shorter on elementary science. The media – newspapers, television news channels, their websites – are to blame for spreading half-baked forecasts and wild prognoses. Not one of the numerous newspapers and TV channels, whatever the language they employ, bother to provide their reporters a basic grounding in the climatological system that gives us our monsoon.

In the first place, the India Meteorological Department (IMD) issues an operational forecast for the south-west monsoon season (June to September) rainfall for the country as a whole in two stages. The first stage forecast is issued in April and the second stage forecast is issued in June. These forecasts are prepared using state-of-the-art Statistical Ensemble Forecasting system (SEFS) and using the dynamical coupled Ocean-Atmosphere global Climate Forecasting System (CFS) model developed under Monsoon Mission of the Ministry of Earth Sciences.

On 15 April 2019 the IMD issued its first stage forecast. Based on our own in-field observations from the west coast, from the patterns of maximum termperature bands and variations in the lower and central peninsular region, from the sea surface temperatures in the Arabian Sea particular its southerly reaches and ditto for the Bay of Bengal, and from the wind patterns that can be experienced at various places in the peninsula and on the west coast, we find the IMD first stage forecast to be reliable.

It is the chronically ignorant media – which over the last few years has displayed a tendency to prefer some so-called private sector weather forecasters instead of what the Ministry of Earth Sciences provides – found irresponsibly claiming that the monsoon of 2019 will be ‘deficient’ and will also begin ‘late’. Neither of these terms is sensible in any way, and we take no satisfaction in noting that only a media that is insensible to planetary and mesoscale events like climate, will employ such insensible terms in reporting that is meant to educate and benefit the public.

IMD’s April forecast used the following five predictors: 1. the Sea Surface Temperature (SST) Gradient between North Atlantic and North Pacific (in December and January), 2. the Equatorial South Indian Ocean SST (in February), 3. the East Asia Mean Sea Level Pressure (in February and March), 4. North-west Europe Land Surface Air Temperature (in January), and 5. Equatorial Pacific Warm Water Volume (in February and March).

There are two forecasts the IMD makes. One is based on the Monsoon Mission CFS Model, which considers global atmospheric and oceanic initial conditions up to March 2019 and use 47 ensemble members (or kinds of data). The forecast based on the CFS model suggests that the monsoon rainfall during the 2019 monsoon season (June to September) averaged over the country as a whole is likely to be 94% ± 5% of the Long Period Average (LPA).

The second is the forecast based on the operational Statistical Ensemble Forecasting system (SEFS). This shows that quantitatively, the monsoon seasonal rainfall is likely to be 96% of the Long Period Average (LPA) with a model error of ± 5%. The SEFS comprises five category probability forecasts for the June to September rainfall over the country as a whole:

Overall therefore the IMD forecast is for the 2019 monsoon rainfall to be near normal. The IMD has already pointed out (which can be seen from the probabilities of the categories given in the table) that there is only a small chance for the monsoon rainfall to be above normal or excess. In view of the weather events and the climatological changes that we are seeing from day to day in May, ascribing a ‘lateness’ to the monsoon is absurd. Monsoon conditions already exist in and over the Indian land mass and in and over the great watery zones extending southwards from latitude 8 degrees North – and that is why we will find rain-bearing clouds crossing the south-western coastline in the first week of June 2019.

(Reposted from India Climate Portal.)

Advertisements

Three views of monsoon 2018

The India Meteorological Department (IMD) on 16 April had issued its first long range forecast for the 2018 South-West Monsoon season, which the IMD has historically taken to be 1 June to 30 September. The IMD had said that the “monsoon seasonal rainfall is likely to be 97% of the Long Period Average with a model error of ± 5%”. The IMD had also said that its forecast “suggests maximum probability for normal monsoon rainfall (96%-104% of the long period average) and low probability for deficient rainfall during the season”.

In early June, the IMD will issue its second long range forecast for the 2018 monsoon. Until then, I have studied three of the more reliable (in my view) international multi-model ensemble forecasts for the monsoon. What are ensemble forecasts? Each consists of several separate forecasts (some ensembles use 50) forecasts made by the same computer model – these are run on super-computers such as the High Performance Computer System of the Ministry of Earth Sciences (one is at the Indian Institute of Tropical Meteorology in Pune with 4.0 petaflops capacity and the other at the National Centre for Medium Range Weather Forecasting in Noida with 2.8 petaflops capacity).

The monsoon 2018 forecast for three-month blocks of the Multi-Model Ensemble (MME), USA National Centers for Environmental Prediction

The separate forecasts that make up one ensemble are all activated from the same starting time. The starting conditions for each differ from each other to account as far as possible for the staggering number of climatological, atmospheric, terrestrial and oceanographic variables that affect and influence our monsoon. The differences between these ensemble members tend to grow as the forecast travels two, three, four and more months ahead of the present.

I have considered the ensemble forecasts for the 2018 monsoon of the European Centre for Medium-Range Weather Forecast (ECMWF), the NOAA Climate Prediction Center and the Multi-Model Ensemble (MME) of the USA National Centers for Environmental Prediction. In this order, I find that the ECMWF forecast is somewhat pessimistic, the NOAA CPC is largely neutral and the MME is optimistic. The forecasting periods are in blocks of three months.

I have considered the ensemble forecasts for the 2018 monsoon of the Multi-Model Ensemble (MME) of the USA National Centers for Environmental Prediction, the NOAA Climate Prediction Center and the European Centre for Medium-Range Weather Forecast (ECMWF). In this order, I find that the MME is optimistic, the NOAA CPC is largely neutral and the ECMWF forecast is somewhat pessimistic.

Here are the details:

(1) The MME forecast, precipitation anomalies relative to the period 1993-2016, based on initial conditions calculated at the beginning of May 2018.
June July August (JJA) – west coast and Konkan, coastal Andhra Pradesh and Odisha, West Bengal, part of the North-East, the entire upper, middle and lower Gangetic region (Uttarakhand, Himachal, Uttar Pradesh, Bihar, Jharkhand), Madhya Pradesh, Chhattisgarh, Maharashtra and Telengana to have up to +1 mm/day. Rest of India other than Gujarat (-0.5 mm/day) normal.
July August September (JAS) – Gujarat to have up to -1 mm/day, Rajasthan up to -0.5 mm/day, Sikkim, Brahmaputra valley and Arunachal Pradesh up to -0.5 mm/day. Himachal Pradesh and Uttarakhand up to +0.5 mm/day, Madhya Pradesh, Odisha, Chhattisgarh, Jharkhand and West Bengal up to +0.5 mm/day. Andhra Pradesh and Tamil Nadu up to -0.5 mm/day.
August September October (ASO) – Gujarat up to -0.5 mm/day. Tamil Nadu up to -1 mm/day. Kerala and adjacent Karnataka up to -0.5 mm/day. Madhya Pradesh, Chhattisgarh, Odisha, West Bengal up to +1 mm/day. Tripura, Mizoram, Manipur, Nagaland, Uttarakhand, Himachal Pradesh up to +0.5 mm/day
September October November (SON) – Tamil Nadu, Kerala and adjacent Karnataka up to -1 mm/day. Maharashtra, Madhya Pradesh, Chhattisgarh, Odisha, West Bengal, Himachal Pradesh and Uttarakhand up to +0.5 mm/day.

(2) The NOAA CPC forecast, seasonal precipitation anomalies using initial conditions of 30 April 2018 to 9 May 2018.
May June July (MJJ) – for most of India a normal reading (+0.5 to -0.5 mm/day fluctuation) and for the west coastal, Konkan, Kerala, south Tamil Nadu and coastal Andhra Pradesh areas variation of up to +1.5 mm/day.
June July August (JJA) – for most of India a normal reading (+0.5 to -0.5 mm/day fluctuation).
July August September (JAS) – normal for most of India. Some areas in the central Deccan plateau, on the west coast and east coast variation of up to -1 mm/day.

(3) The ECMWF forecast, mean precipitation anomaly based on climate period data of 1993-2016 and initial conditions as on 1 May 2018.
June July August (JJA) – all of the southern peninsula and part of the Deccan region (Kerala, Tamil Nadu, Karnataka, Andhra Pradesh, part of Telengana) up to -100 mm for the period. West Bengal, Bihar, Jharkhand, Uttar Pradesh up to +100 mm for the period.
July August September (JAS) – all of the southern peninsula and the Deccan region – Kerala, Tamil Nadu, Karnataka, Andhra Pradesh, part of Telengana and Maharashtra up to -100 mm for the period.
August September October (ASO) – Maharashtra, Telegana, Andhra Pradesh and Karnataka up to -100 mm for the period.
September October November (SON) – Central and western India, eastern states and entire Gangetic region up to -50 mm for the period.

Monsoon 2015 could be 93% with ifs and buts

The India Meteorological Department has just released it’s long-awaited forecast for the 2015 Indian monsoon. In terms of the quantity of rainfall over the duration of the monsoon season (June to September) the IMD has said it will be 93% of the ‘Long Period Average’. This average is based on the years 1951-2000.

What this means is the ‘national’ average rainfall over the monsoon season for India is considered to be 89 centimetres, or 890 millimetres. So, based on the conditions calculated till today, the ‘national’ average rainfall for the June to September monsoon season is likely to be 830 millimetres.

There are caveats and conditions. The first is that the 93% forecast is to be applied to the long period average for each of the 36 meteorological sub-divisions, and a ‘national average’ does not in fact have much meaning without considerable localisation. The second is that the forecasting methodology itself comes with a plus-minus caution. There is “a model error of ± 5%” is the IMD’s caution.

This first forecast and the model that the forecast percentage has emerged from are thanks to the efforts of the Earth System Science Organization (ESSO), under the Ministry of Earth Sciences (MoES), and the India Meteorological Department (IMD), which is the principal government agency in all matters relating to meteorology. This is what the IMD calls a first-stage forecast.

As with all complex models, this one comes with several considerations. The ESSO, through the Indian Institute of Tropical Meteorology (IITM, which is in Pune), also runs what it calls an ‘Experimental Coupled Dynamical Model Forecasting System’. According to this, the monsoon rainfall during the 2015 monsoon season (June to September) averaged over India “is likely to be 91% ±5% of long period model average”. (The IMD forecast is available here, and in Hindi here.)

This is a lower figure than the 93% headline issued by the IMD. This too should be read with care as there are five “category probability forecasts” that are calculated – deficient, below normal, normal, above normal and excess. Each is accompanied by a forecast probability and a climatological probability (see the table). The maximum forecast probability of 35% is for a below normal monsoon, while the maximum climatological probability is for a normal monsoon.

As before, time will tell and the IMD will issue its second long range forecast in June 2015. Our advice to the Ministry of Earth Sciences and to the IMD is to issue its second long range forecast a month from now, in May, and also to confirm these forecasts two months hence in June, when monsoon 2015 will hopefully be active all over the peninsula. [This is also posted on India Climate Portal.]

The year of 1.275 billion Indians

Four years ago the Census of India 2011 completed its count – the world’s largest human counting exercise, which required an army of enumerators to carry out. The final tally – which was fixed in late 2012 although the provisional estimates for India and the states came within a few months of the census having counted the residents of the last remote hamlet – was 1,210,569,573 of us (833,463,448 rural and 377,106,125 urban).

That tally was valid for a particular day only, which was the last day of March 2011. Although the 1.2 billion statistic was used (and continues to be, popularly) to describe India, Bharat, our people and their lives and work, it is of course an incorrect number. The population has been rising from the first day of April 2011 as the daily count of births minus deaths (to give the net addition) adds to that 1.2 billion. Now, the net daily addition is about 40,000 a day!

But remember that those who were counted as net daily additions in 2001 – when the previous Census was done – are today only 14 years old, so the total population tally is not adults, it is infants, toddlers, children, early teenagers, late teenagers, youth, adults, mature adults, senior adults, elderly folk and really old folk.

It would have happened in February 2013, the crossing of the billion and a quarter mark (1.25 billion). What will our tally be in March 2015, two months from now? It depends how complex you want your calculations to be. The simplest method would be to use 1.210 billion as the 2011 tally and multiply that by the yearly population growth rate based on the increase in our population between 2001 and 2011. The result would be an estimate which does not recognise different population growth rates for rural and urban regions, and indeed these differ in every state. Naturally, for states with large populations (such as West Bengal, Bihar, Madhya Pradesh or Karnataka) the growth rate you apply can change your estimate by a significant number.

The population estimates for India by the Population Division of the UN’s Department of Economic and Social Affairs. The four estimates use low, medium, constant and high fertility rates as their basis. Source: World Population Prospects: The 2012 Revision (http://esa.un.org/unpd/wpp/index.htm)

It was, I had estimated last year, that the 400 million mark was crossed for urban India: “By the end of 2014 June, a group of cities will cross important population thresholds. This upward procession of population numbers – for districts, cities and states – is scarcely observed by administration or by citizens, but continues apace.”

To find the startling numbers in this graphic, I have used the individual growth rates for rural and urban populations in all states and union territories, based on the 2001-2011 period. This helps, but for the demographic purist it isn’t enough (and I haven’t attempted it here) because it does not factor in the changes in fertility rates – and since 2012 there have been reports that the fertility rates have dropped faster than expected for the period after 2011 – and, more minutely, changes in the age at which women marry, the spacing of children and so on; these are the formulae of the careful demographer.

To do so is necessary. It is even vital because so much of public and social provisioning depends on understanding population trends well and preparing for what the next five, ten and twenty years will bring. There is the matter of food and the public distribution system (see ‘India’s 681 million hungry rural citizens’ and the question: “What do and what can rural residents spend on food and the essentials of living in India?”).

There is the question of how much towns and cities can expand, and two years ago, according o the Census 2011 data, India in 2015 would have 63 cities with populations of over a million (ten more than we had in 2011), which led to the question of whether India is ruled for its cities.

But the big picture remains. In 2015 there are a number of population markers that will be reached. The population of rural Bihar will cross 99 million, the population of Rajasthan will approach 73 million, the population of urban Punjab will approach 11.2 million, the combined population of the seven North-Eastern states (excluding Assam) will reach 14.8 million, and the population of rural Gujarat will approach 35.2 million.

Urban populations have grown more rapidly and the consequences of such growth includes cities that are more difficult to live in (see “When the 65 million who live in India’s slums are counted”). But the trend tells us that in 2015 the population of urban West Bengal will cross 30 million, the population of urban Maharashtra will approach 55 million, the population of urban Karnataka will cross 26 million, and the population of urban Tamil Nadu will be more than its rural population. And finally, at some time during 2015, the number of Indians will touch 1,275 million (or 127 crore).

An epidemic of misreading rain

Who can you turn to? It’s easier to list those whom you shouldn’t turn to, the top rankers being the country’s press and television wallahs, followed at a not respectable distance by academic commentators, then come the government blokes and bureaucrats (some of whom do know the difference between isobars and salad bars, I’ll give them that). Lurking behind this cacophonous mob are the boffins of the IMD and its associated scientific chapters, a number of whom have got their sums right, but who aren’t given the space and encouragement to tell the great Bharatiya public what said public is yearning to hear simply because regulations forbid, just like it was in 1982, 1957, whenever.

As I may have mentioned before, this is Not A Good Thing. It has taken about a decade of mission mode tutoring (how the UPA bureaucrats loved that phrase, mission mode) to get the media wallahs to see the difference between weather and climate. A few may even have learned to read a wet bulb thermometer and puzzle their way through precipitation charts.

But overall, the profusion of android apps that profess to show cool graphics of clouds with lightning bolts erupting topside so that our humble ‘kisans’ know when it’s going to rain (i.e., by looking down at their screens instead of up at the sky) has not helped the Bharatiya public make more sense of less rain. We have squadrons of Insats and Kalpanas buzzing around the globe beaming pictures from the infra-red to the infra dig back home, every 60 or 90 minutes, busy enough to crash a flickr photo server, but the knowledge that said public can sift from it is sparse, rather like the rainfall over Barmer, Bikaner and Ajmer.

And so it goes, with the waiting for rain replacing with an equal banality waiting for Godot but with a far larger cast of characters, most of them insensible to the greater climatic drama being played out, 30,000 feet overhead, and at the poles, in the vast turquoise swells of the eastern Pacific where a malignant El Nino is brooding, in the Himalayan valleys where crystal zephyrs have been shoved aside by an airborne mat of PM2.5, or to the desiccation that creeps outwards from our towns and cities (7,935 of them, India’s triumphant ‘growth story’) that have enclosed sweeping hectares with cement, asphalt, and the hot foetid belches of factories and air-conditioners. GDP, they have been told, is the great liberator.

And that is why we have in place of the quiet concern of our forefathers in their dhotis, an electronic jumble of shrill alarm. “Weak monsoon intensifies drought like conditions in India” was one such headline, the text beneath finding the most ludicrous connections: “… threat of food inflation and weak rural demand in the first year of the Narendra Modi government”. Naturally, the cheerleaders of a demand-centric world cannot do otherwise.

And likewise with “Weak rains deliver India’s new Modi government its first economic challenge” that desultorily spies impending delays in the “sowing of main crops such as paddy, corn and sugarcane” and which notes mournfully that “about half of all farms lack irrigation systems” and, even worse, that “reservoir levels are only a fourth of last year’s levels”, this last despite the best efforts, ham-handed though they are, by the Central Water Commission to show India (for Bharat knows) that the reservoir levels in the 85 major reservoirs are low, but not much lower at this point in 2014 than they were in 2013. The GDP bullies dislike contrary numbers, and would go cross-eyed were someone to mischievously mention the existence of 4,845 large dams in India (the blue-ribboned 85 included) whose many water levels we don’t in fact know at all.

And similar vapidity from another quarter, which like its peers cloaks ineptitude with what it takes to be appropriate jargon, “The cumulative rainfall across the country has so far been 45 per cent below the Long Period Average (LPA) for 1951-2000” and brandishes even more frightful credentials with “a further breakdown of rain data recorded in different meteorological subdivisions shows that normal rainfall has been recorded in only seven of the 36 regions”. But which sere farmer and her wise daughters consider in their universe such things as meteorological subdivisions, when their world is what Balraj Sahni and Nirupa Roy in 1953 showed us so lambently, is no more than ‘do bigha zamin’?

But still the misreading gathers pace, as vexed fixations upon an existence merely economic chase away plain common-sense. For rains may come or rains may go, but in tractors – for so we are instructed by the agents of hardened merchants – we trust. To wit: “… tractor sales have typically expanded at a double-digit pace in the years when rains have disappointed… In the 11 years between fiscal 2003 and fiscal 2013, rains fell short by 5% or more on six occasions… In four of those six years, tractor sales grew at a double-digit pace”. Let us then leave behind our cares and go rollicking over the dusty, still dustier now, plains of the Deccan in tractors tooting red.

But a shadow of monsoon yet for Bharat, and at June’s end. It is past time that the prattling ceased and the learning began.

North India 2014, much dust, more heat, late rain

The sweltering of North India, aggravated by manic urbanisation, just as manic use of personal automobiles, the steady thinning of tree canopies, and small businesses forced to buy diesel generators – in the tens of thousands, each emitting hot fumes that further trap already heated layers of sooty air – is an annual pre-monsoon epic that no-one has the energy to write.

This panel of four maps shows us where India baked during the last week of 2014 May (and now, Delhi has experienced a record its residents did not want). The high temperature belts (top left map), 40-45 Celsius, covered most of central and north-western India (Maharashtra, Madhya Pradesh, Gujarat, Rajasthan, Gangetic Uttar Pradesh, Haryana, Punjab and Delhi). Minimum temperatures (top right), 20-15 Celsius, are seen in two pockets – south interior Karnataka and in the North-East over Manipur and Mizoram.

These temperature maps may be read with the rainfall for the same period, 2014 May 25-31, to correlate particularly the temperature anomalies (how much higher or lower the normal maxima and minima have departed) with where it has rained. The rainfall map (lower left) shows rain having fallen over south Karnataka, but also north West Bengal and eastern Bihar, coastal Odisha and southern Haryana. These appear to relate to a group of anomalies (lower right): the first being interior Tamil Nadu, north-eastern Karnataka and adjacent Andhra Pradesh; the second being eastern Uttar Pradesh and adjacent Bihar. [You can get the four maps in this zip file.]

Read these from top left – 21, 22 and 23 June. Lower row – 24, 25 and 26 June. The green shading is the rain-bearing cloud cover. After 20 June the peninsula will have rain in most meteorological zone but North and north-west India will still await the monsoon system.

What these don’t show, but which the longer range forecast Indian Institute of Tropical Meteorology (Pune, Maharashtra) has on record, is that monsoon 2014 will not touch northern India until the fourth week of June. Rain-bearing cloud and wind systems will cover, in this forecast, peninsular India by around the 16th or 17th of June, but it will be another week before they deliver some relief to the cemented and asphalted surfaces of the National Capital Territory and its parched surroundings.

These very helpful maps are used by the Pune-based met researchers as part of their monsoon monitoring and forecasting partnership with several international climatoloigcal research institutes, chief amongst them the National Oceanic and Atmospheric Administration (NOAA) of the USA through its Climate Prediction Centre.

The Tropmet – as the Pune group is usually called – has bequeathed to us a definition (perhaps for this season only, subject to revision when climate change asserts itself) of monsoon rain that is in part scientific and in part geographic, which I think is a good sign (the Indian Meteorological Department will disagree, but we know better). The faster Tropmet decides to communicate in language appreciated, and understood, by Bharatiya citizens, the more said citizens will find an interest in correcting the misconceptions of scientists.

Tropmet says: “Rainfall within the summer monsoon season is mainly punctuated by the northward propagating monsoon intraseasonal oscillations (MISOs) with timescales of 30-60 days that manifests as spells of heavy rainfall and periods of quiescent rainfall, instead of a continuous deluge.” In the Konkan, we like our continuous deluge and the old-timers would have sixteen names for different sorts of deluge (and an equally rich chest of monsoon nouns for other sorts of rain).

Weighing the monsoon winds for El Niño

Ignoring the torpor of the summer heat, the India Meteorological Department has dusted off the statutory paragraphs that give us in the sub-continent a first indication of what monsoon for the year may be like. The result this year, both scintillating pages, has been made that much more gripping by the inclusion of El Niño. The IMD’s treatment of the normal variables whose interplay determine the nature of any monsoon is perfunctory – which is surprising as the regional and international earth observation networks spare no detail and tend to inundate us with data and analysis.

Clouds and wind, land and farm. The equation that all rural districts make at this time of the year, but which is becoming more difficult with every year that climate change strengthens its grip.

But the IMD, especially for the south-west monsoon, has always preferred to be spartan. Perhaps there is some philosophical dictum that us non-meteorologists have yet to grasp, and if so then the only criticism we may be permitted, if the IMD had its way, its to ask for more such teaching. But the IMD does not have its way, and we publics whose monies support its work must continue to demand from the recalcitrant department better, much better, application and communication of its work.

The official release, ‘Long Range Forecast For the 2014 Southwest Monsoon Season Rainfall’, is delivered to us by the IMD, Ministry of Earth Sciences (MoES). There is the usual paragraph claiming a profundity of observation and of the IMD being a standard-bearer of superior method. “Operational models are critically reviewed regularly and further improved through inhouse research activities,” says the IMD. But what we still have, in a tradition that is probably three generations old, is the two stage forecast (one in April, the second in June). [Here is the release in Hindi.]

I think this proves how out of step the IMD – and the Ministry of Earth Sciences (grand title isn’t it?) – is with what citizens of India experience in their villages, towns, fields and hills. For, the south-west monsoon no longer arrives in the first week of June, and it no longer begins to depart by mid-September. Climate change began to alter that comfortable rhythm years ago, but the IMD’s forecasting grindstone is the same, never mind how many new earth observation satellites India pelts into orbit.

With all these provisos, stated and implicit, what has the IMD told us?

First, that the “experimental ensemble forecast based on IMD seasonal forecast model (SFM) indicates that the rainfall during the 2014 monsoon season (June to September) averaged over the country as a whole is likely to be 88% ± 5% of long period average (LPA)”. This means that in places it could be as low as 83% of the average, and no more than 92% of the average. Combine this with the assessments about the 2014 El Niño and we can see why, far from being satisfied that the IMD is considering both the monsoon and El Niño, we ought to monitor independently both and force the IMD to become more responsive.

Second, that “the experimental forecast based on the coupled dynamical model forecasting system suggest that the monsoon rainfall during the 2014 monsoon season (June to September) averaged over the country as a whole is likely to be 96% ± 5% of long period model average (LPMA)”. This is a more hopeful set, but also shows that the IMD, by telling us of two different scenarios from two models, is hedging its forecast, which is not what its job is.

Third, the IMD has said “the experimental five category probability forecasts for the 2014 monsoon season rainfall over the country as a whole using the experimental dynamical prediction system are 33% (deficient), 20% (below normal), 24% (normal), 6% (above normal) and 17% (excess)”. This means, using this ‘probability’, that a normal monsoon for 2014 has only a 1-in-4 chance whereas a deficient monsoon (that is, total rain less than 90% of the long period average) has a 1-in-3 chance.

This is a prognosis that stands between serious and grim, for a 10% drift towards the lower side of an expected average, for any of our 36 agro-meteorological regions, can spell ruin for farmers and severe hardship for water consumers. How have central and state governments prepared for such a forecast? We have no information, most likely because there has been no preparation (there are contingency plans for the chronically drought-prone districts, but these are normally triggered when there is an official declaration by  the state government that there are conditions of drought in parts of the state). Elections or no elections, El Niño cares not, and it is up to the state governments to make preparations for a monsoon 2014 whose delivery of water already looks uncertain.

Monsoon mysteries, mundane mathematics

Rainfall averages for the months of January to August 2013, nicely arranged in one handy panel.

June, July and August 2013 are behind us and India’s understanding of the south-west monsoon – and also of climate change during monsoon and outside the monsoon months – has scarcely improved.

Other than the announcements from the Indian Meteorological Department, which are unintelligible to anyone with even the smallest interest in monsoon and climate, there is little in India’s media or from official sources that helps turn an embarrassing scarcity of understanding into knowledge.

The IMD, I was told about two years ago by a senior crop scientist, is queried every day by officious underlings from New Delhi wanting to know weather conditions for wherever their bosses are to visit, and this means the ministers of the cabinet, sundry senior and junior ministers, all variety of politicians and all their scheming cronies and hangers-on. That is why the IMD has so many special forecasts for New Delhi on its website. For the Indian Meteorological Department, India outside Delhi seems to be an annoyance.

There are science institutes, such as the Indian Institute of Tropical Meteorology, which is, as far as I can make out, the seat of the scientific inquiry into monsoon and climate. The communication about its work however is either dismal or absurdly comic. Scientists long ago realised that to be of use to society, they had to make their questions and answers understood by those whom they wish to serve.

In such a spirit, India’s monsoon and climate scientists and researchers ought to have begun their careers with a session before their grandmothers – could they explain their work to their grandmothers? If not, they ought not to have been accepted in their present jobs (and their grandmothers should be inducted instead, for the wealth of their traditional knowledge).

And this is why we have gobbledygook such as this – from the IMD, about the 2013 September forecast:

“The 5-parameter PCR model was also used to prepare probability forecasts for the pre-defined 3 (tercile) categories of rainfall during the second half of the monsoon season. These are below normal (<90% of LPA), near normal (90-110% of LPA), above normal (>110% of LPA). The tercile categories have equal climatological probabilities (33.33%  each).  The  forecasted  probabilities  for  the  September  2013  rainfall  over  the country as a whole in percentage for the above tercile categories are  31%,  53%, and 16% respectively.”

Can the Director General of Meteorology – who is the head of the IMD – explain this to his washerman, his gardener and the chap who sells him vegetables? If not, he ought to make way for someone who can, because India understands the DG of M not at all, and we’re not paying his salary and departmental budget from our taxes to be misspoken to about the rain.

Just so that you know, this verbally challenged gentleman is assisted in his duties by five Additional Directors General and by 20 Deputy Directors General. And when I examined where this chatty group is deployed, the Delhi obsession (or fear thereof) became immediately clear, for of the five Additionals, four are in Delhi (the fifth is in Pune), and of the 20 Deputies, ten are in Delhi (the others are over the capital’s horizon, toiling somewhere in the torpor of the sub-continent.

The IMD runs six regional meteorological centres – each ruled by a humourless Deputy Director General, all morbid conversationalists – and these are to be found at Mumbai, Chennai, New Delhi, Kolkata, Nagpur and Guwahati. Then the IMD has what it importantly calls ‘operational units’: meteorological centres at state capitals, forecasting offices, agro-meteorological advisory service centres, flood meteorological offices, area cyclone warning centres and cyclone warning centres. All important centres, all equally incoherent about their work, all equally clueless about how to speak simply with those who need the forecasts the most.

But the disease of obscurantist climatic communication is to be found also in the Indian Institute of Tropical Meteorology, to wit: “However, the extended range predictability arises from Monsoon Intraseasonal Oscillations (MISOs) which are the quasi-periodic northward propagating large-scale convective cloud bands that manifest as the active/break spells, as described earlier.”

What will it take to get these scientists and researchers to understand that monsoon, rain, weather and climate are of great interest to most adults amongst the country’s 1.25 billion-strong population, and that by mumbling only to each other in equations and formulae they continue to do us a mis-service, and in fact shirk their main duty?

World grains harvest revised downwards for 2012-13

The IGC Grains and Oilseeds Index (GOI) & sub-Indices, charted from 2009 January (daily, with January 2000=100). Chart: IGC

The world 2012-13 total grains harvest (wheat and coarse grains) forecast has been revised lower by the International Grains Council (IGC) to 1,810 million tons (mt) this month and is now expected to fall year-on-year. The IGC’s forecast for the US maize crop has been cut by 50 mt, to 300 mt and the soyabean harvest has been reduced by 8.3 mt to 79 mt. Wheat output has also been revised lower in both Kazakhstan and Russia.

Maize (corn) and soyabean prices have soared to new highs on deteriorating output prospects in the US following the worst drought since 1956. Unfavourable weather conditions have also led to a scaling back of grains output and exportable surpluses in the Black Sea region, supporting gains in wheat and barley values. Global carryover grains stocks are expected to fall by 29 mt by the end of 2012-13, led by a 15 mt drop in wheat and 14 mt fall in maize; maize stocks are forecast at a six-year low.

Agrimoney has said that the downgrade takes the council’s estimate for the US corn harvest well below the US Department of Agriculture’s own forecast, of 329.5m tonnes. The USDA, whose estimates are followed particularly closely by traders, also still foresees a small rise in world corn inventories. However, analysts have already started the countdown to the next USDA Wasde report, on August 10, when it will revise estimates for crops worldwide.

The IGC Grains and Oilseeds Index (GOI) reached an all-time high on 20 July, and, despite some recent easing, is still up 14% month-on-month (m/m). World soyabean production is expected to recover sharply in 2012-13, rising by some 9% year-on-year, although the forecast hinges on a strong rebound in output from South America where planting begins in the fourth quarter of 2012.

Rice output in 2012-13 is forecast up 1%, compared to 3% growth in the previous year, due to less rapid expansion in Asia, but consumption growth is also likely to be lower. In contrast to the steep weather-driven gains in grains and oilseeds markets, rice prices declined marginally, mainly on supply-side pressures in Thailand.

Grains till 2016-17: the IGC speaks

The International Grains Council (IGC) has prepared a summary of projections for grains and cereals. The IGC Secretariat has said of its work that “the figures represent the Secretariat’s view of the general development of the global grains economy in the period to 2016-17, taking into account a number of broad assumptions”.

These include assumed trends in population growth, prices, developments in agriculture and trade policy, as well as prospects for the global economy. “The latter have become increasingly uncertain over the past year”, the IGC Secretariat has said. The IGC has added the proviso that these estimates and the forecast derived from them are subject to risk, and this analysis assumes that current economic problems do not worsen. Here are the sections:

Total Grains
* World grains production in 2016-17 is projected to reach 1.98bn tons, a 158m. increase (+9%) compared with 2011-12; wheat output is forecast to rise by 30m. (4%) and maize by 94m. (11%).
* Despite heightened economic uncertainty, the analysis assumes any slowdown in global economic growth will be temporary and increasing prosperity will boost grains consumption, particularly for feed and industrial uses. Feed use is expected to rise at a slightly faster pace than in recent years, while increases in industrial use will slow from the very rapid rates in the past decade. Diversifying diets, particularly in favour of livestock products, will slow the rise in direct use of grains for human food. Total grains consumption is projected at 1.98bn. tons in 2016-17 (1.83bn. in 2011-12), including 659m. (630m.) for human food, 846m. (769m.) for feed and 343m. (302m.) for industrial uses.

World grains stocks are forecast to show little change in the medium term and are set to remain relatively tight, especially for maize. At the end of 2016-17, world grain carryover stocks are projected at 354m. tons (compared with 360m. at the end of 2011-12), including 118m. (123m.) of maize, 196m. (202m.) of wheat and 26m. (23m.) of barley.
* World grains trade is projected to increase by about 2% per year, to 273m. tons in 2016-17, with wheat and maize rising to new records. Increasing demand for wheat-based foods will lift wheat import needs in Africa and Asia. Imports of maize for feed will rise, especially in Pacific Asia, with China seen as a more regular buyer.

Wheat
* Increases in world wheat production in the five years ending 2016 are expected to be broadly matched by use, and global stocks are expected to be maintained at close to recent levels.
* Planting decisions will be influenced by likely attractive prices for alternative crops, especially maize and oilseeds. Nevertheless, some rise in global wheat area is anticipated, led by gains in the CIS. After a relatively sharp increase of 1.6% in 2012-13, including a recovery in North America, global areas are projected to expand by around 0.4% annually. Taking into account slightly increased average yields over the period, world wheat production is projected to reach a record 714m. tons in 2016-17, representing an increase of 30m. compared with the estimate for 2011.
* World wheat consumption is projected to grow by 1.1% annually, close to the long-term average, reaching 716m. tons in 2016-17, up by 39m. compared with 2011-12. A continued increase in human food use accounts for half the rise, driven by expanding demand in developing countries. At 0.8% per year, the average annual increase is only slightly slower than the longer-term trend of 1.0%. Increases in world feed use mainly reflect a tight S&D outlook for maize and expectations that the cost of wheat will be more attractive than maize at times. Gains in industrial use are expected to accelerate, particularly for biofuels, although overall amounts will remain small relative to total consumption.
* World wheat carryover stocks are projected to stay relatively ample in the next five years, receding only slightly, to 196m. tons. Those in the eight major exporters are projected to show an initial rise, but then fall back to about the same level as currently.
* World wheat trade to 2016-17 is forecast to increase by around 2% per year, reaching a fresh record of 138m. tons. Increases in milling wheat trade will be sustained by rising demand in developing countries in Asia and Africa, while feed wheat may show some further gains if import costs are competitive with maize.

[The IGC forecast document is available here (pdf)]
[The spreadsheet (xls) with the major grains’ data is available here]

Rice
* Only a modest expansion in the global paddy (rice) area is forecast in the five years to 2016-17, with the average year-on-year increase projected at just 0.3% (compared to an average of 0.7% in the prior five-year period). To some extent, this reflects an expected contraction in China’s sowings, amid a continued shift to diets that are richer in protein. Taking into account slightly reduced average yield gains, global rice production (milled basis) is projected to increase by 23m. tons, to 482m. by 2016-17, an annual average growth rate of 1%.
* Global rice consumption is projected to reach 482m. tons by 2016-17, up by 25m. from 2011-12. At 1.1%, average growth, while broadly in line with the global population trend, will be lower than in previous years. This is due to a forecast contraction in China, as well as more moderate growth in other parts of Asia. Elsewhere, sub-Saharan Africa is expected to be one of the fastest growing regional markets, the result of a rising population and a shift away from traditional, locally-grown cereals.
* The world rice carryover is projected to rise only slightly over the next five years, to 103m. tons. In the five major exporters, stocks are expected to initially increase – centred on inventory accumulation in India and Thailand – before edging slightly lower. Their share of the world total will average around one-third through to 2016-17.
* Global rice trade is projected to expand by nearly 3% annually, to 37.2m. tons by 2017, broadly in line with maize but comfortably exceeding the year-to-year rise in wheat. Growth will be underpinned by larger shipments to Far East Asia, especially the Philippines, and sub-Saharan Africa. The latter sub-region will remain heavily dependent on imports to meet domestic requirements; their share of total consumption is forecast to average 45%.

Maize (Corn)
* The supply and demand for maize (corn) is projected to remain tight, with world inventories projected to drop to historically low levels.
* With firm global demand and generally tight availabilities expected to support world prices, maize plantings are projected to remain high across the forecast period. Increases in area and improvements in yields, especially in the US, Latin America and China, result in large consecutive crops. World maize production is forecast to increase to 949m. tons in 2016-17, some 94m. higher than the estimate for 2011.
* Global maize consumption is projected to rise to 949m. tons in 2016-17, up by 86m. from 2011-12. Growth in use is forecast to decelerate, mainly due to slowing industrial demand. With use for ethanol in the US levelling out, industrial consumption is projected to rise by 2% annually, compared to 12% in the last five years. Despite high prices, rising meat demand in developing countries will lift feed maize consumption by around 2% per year. Population growth, rising per capita incomes and changing dietary preferences are expected to boost meat consumption in parts of Asia, Latin America and Africa.
* World closing stocks are expected to tighten but, with supply and demand seen broadly in balance towards the end of the forecast period, the projected 2016-17 carryover of 118m. tons would only be 5m. below that at the end of 2011-12. US ending stocks are forecast to increase from recent lows, but China’s will decline.