The Hindu Business Line : Whither forecast?

Climatologists at regional levels of the India Meteorological Department took keen interest on the pathway of category-5 cyclonic storm Gonu, speeding up to 260 km per hour and killing over 60 people in Oman alone, maybe due to perplexity about the arrival of the South-West monsoon on the southeastern coast — although the two were not inter-related.

S-W monsoon continues to be maverick but the biggies at the IMD seem to be conceited with the grey theoretical aspects, with inadequate emphasis on the ‘green’ reality. The unusually prolonged hot-and-humid regime might have been difficult to predict but there was no semblance of concern at the IMD apex, which thought its duty was over with the “long range forecast for the 2007 South-West monsoon (June to September)”. Monsoon apparently is around southern Bengal but didn’t strike even on June 10 and is of a weaker pattern. There was no nor’wester — the prelude to monsoon in fairly full bloom during the period in mid-April.

At the outset, let’s note that monsoon forecasting in the tropics is not easy as stochastic models are loaded with constraints and assumptions. Which is why IMD projections are often off-the-reality as seen in 2005 and 2006. Farmers, it must be remembered, are concerned about the distribution of rainfall than the total rainfall during the rainy season. The prediction — 95 per cent normal monsoon “with a model error of +/- 5 per cent” is of no help to the farming community. They want to know how many rainy days there would be during the monsoon. In the past — prior to the 1980s — IMD used to publish data on both rainfall and rainy days. Why can’t the IMD scientists who collaborate with counterparts at the Indian Institute of Science, Bangalore, predict the number of rainy days?

Let’s get back to Gonu which, as thought, hit Oman in a major way. Dr Prabir Patra, research scientist, Yokohama-based Frontier Research Centre for Global Change, said in an e-mail: “I do not really know what causes this year’s anomalous sea-surface condition, but obviously warmer SSTs are predicted under various global warming scenarios. As you will note from three years of data, showing large inter-annual variability, we cannot assure global warming as the cause for this condition and climate change due to global warming is believed to occur in more than 10-20 years’ time periods. The other factor we may speculate is the role of aerosols — some studies have suggested more aerosols in the atmosphere results in lesser generation and development of tropical cyclones, which is very similar to results on role of aerosols on Indian summer monsoon rainfall.”

Many models, many differences

He focused on the sea-surface temperatures (SST) — see www.ssmi.com/sst/sst_data_daily.htm) — for three consecutive years on July 4 — 2005, 2006 and 2007. In fact, the aerosol factor remained ignored by the IMD. The 16-factor model, developed in 1988 by a group of scientists led by Dr Vasant Gowarikar, then secretary, department of science and technology, too disregarded this anthropogenic rider. In no time, the model drew flak from atmospheric physicists outside the IMD and DST, despite the claim that predictions were proved right.

Weather forecasting began in 1886 in British India by H.F. Blanford, a meteorological reporter. But the man who set the pace was Sir Gilbert Walker, as director-general of observatories in 1909. Sir Gilbert developed a multiple regression equation with four variables — the Himalayan snow accumulation in May-end, South American pressure during the spring, Mauritius pressure in May and rainfall in Zanzibar (now part of Tanzania) in April-May. It was alright at a time when human factors like deforestation, increased exploitation of fossil fuel and new settlements didn’t influence the natural surroundings at macro- and global levels, but not now.

The Gowarikar model was welcomed as the traditional approach was losing its relevance. But being a tropical phenomenon, clubbing of discrete variables was not enough and this had to be revised. The number of parameters was reduced to eight and then to five for the April forecast and the ten-variable regression equation was replaced by a six-parameter statistical forecasting model for June. The efficacy of this is to be verified and verification does not mean total rainfall but, more importantly, the dissipation thereof. The IMD too concedes that correlations in its statistical models flicker.

Actually, coefficients, derived empirically from time-series and/or cross-sectional data, are often too transient to be applicable even in moderately altered climatic situation, the IMD’s abiding faith in the “concept of ensemble forecasts” notwithstanding. This is essentially clubbing of a set of predictors such as North Atlantic Sea Surface Temperature (December + January), East Asia Mean Sea Level Pressure (February + March) and Equatorial Pacific Warm Water Volume (February + March) instead of a single model and may belie its confidence on monsoon prediction. Yet let’s hope it clicks.

The aerosol factor

Dr Patra, based on a collaborative research study (with other institutes), along with Swadhin K Behara, Jay R Herman, Shamil Maksyutov and Hajime Maketovo, in a paper — ‘Indian summer monsoon rainfall: Interplay of coupled dynamics, radiation balance and cloud microphysics’ (published in Atmospheric Chemistry and Physics) — showed that aerosols, transported from the northern Africa and Arabian Gulf region, contributed significantly to rainfall deficit during the summer of 2002.

Strangely enough, the IMD forecast model doesn’t consider the aerosol factor which is a reflex of anthropogenic factor as one of the major determinants of changed monsoonal pattern. In a recently published paper, ‘Exploring the sensitivity of basin-scale air-sea CO2 fluxes’, in the Journal of Geophysical Research), Patra et al looked into “the possible effects of aerosol-supplied nutrients on the inter-annual variabilities in surface ocean chlorophyll” . Their hypothesis is that the aerosol component “has to play a significant role in terms of micronutrients (for example, iron) supply for maintaining the high biological activity in the northwestern Arabian Sea during Jun-Jul-Aug period.” This leads to a matter of concern: rise in carbon dioxide in the atmosphere. Thus aerosols can work both ways — positive or negative. Which is why the aerosol factor cannot be ignored.

IMD has been collaborating with the Indian Institute of Tropical Meteorology, Pune; Indian Institute of Science, Bangalore; Space Applications Centre, Ahmedabad; National Institute of Oceanography and Centre for Mathematical Modelling and Computer Simulation, Bangalore, apart from exchange of opinion with the National Centres for Environmental Prediction (NCEP), USA; International Research Institute for Climate and Society (IRI), USA and so on.

All this is very impressive but the relevance to ground reality is the main point. Meteorological prediction in the tropics is very cumbrous. Even the fourth assessment report by The Intergovernmental Panel on Climate Change was criticised by atmospheric scientists. Frank Wentz of Remote Sensing Systems in Santa Rosa, California, in a paper in Science found models flawed as increase in humidity in warmer climates was underestimated.

Rabindranath Tagore wrote and tuned most of his songs on the rainy reason and his pen might have been attuned to dancing peacocks. But what was ecstatic for the great poet was nightmarish for the meteorologist entrusted with weather forecasting.

More Stories on : Climate & Weather | Insight | Cultivation

Article E-Mail :: Comment :: Syndication :: Printer Friendly Page

Stories in this Section
Whither forecast?