Despite the prevalence of various forms of irrigation across the country, most agricultural land in India is rain-fed; dependent on the monsoon. While monsoons have emerged as the key drivers of the economy, monsoon forecasts in India are yet to attain the desired standards.
The two monsoons — south-west and north-east — enable the farming of various food-grains and cereals. Kharif crops such as rice, jowar, bajra, groundnut, coarse cereals, ragi and castor seed are dependent on the south-west (SW) monsoon which provides 70 per cent of the country’s rainfall. The average rainfall in the SW monsoon is 119 cm but there is great variation in regional distribution.
Average rainfall
The south-west monsoon covers the areas on the coastal regions of western India, the Western ghats as well as sub-Himalayan areas in the north-east. While the hills in Meghalaya receive the maximum rainfall during this season, at over 250 cm annually, the northern parts of Kashmir and western Rajasthan receive less than 40 cm. Interior parts of central India also receive rainfall during this season.
The SW monsoon starts in the June and continues till September, after which the north-east (NE) monsoon begins (October to December). The NE monsoon favours rabi crops such as pulses — tur, urad, gram — as also rapeseed, mustard, linseed, sunflower, wheat and barley. The NE monsoon covers the southern parts of the country, mainly Tamil Nadu, Pondicherry, coastal Andhra Pradesh and Rayalaseema.
Tamil Nadu receives 40 per cent of its rain in this monsoon. Around 60 per cent of the NE monsoon is in the coastal regions of Southern India and 40-50 per cent is in other districts. Parts of Kerala, Karnataka and Lakshadweep also receive nearly 20 per cent of rainfall in this monsoon, though the primary rainfall for these regions is during the SW monsoon. The average rainfall ranges 50-70 cm in the NE season, particularly for Tamil Nadu.
Monsoon forecasting
Given the significance of monsoon rains for Indian agriculture, forecasting becomes critical to give farmers enough warning to prepare themselves if the rainfall is going to be inadequate.
IMD is the principal agency, headquartered in Delhi (and under the Government of India), for meteorological observations, weather forecasting, rainfall monitoring and seismology to help in optimum operations of weather-sensitive activities such as agriculture, irrigation, shipping, aviation and offshore oil exploration. IMD releases monsoon forecasts on the south-west and north-east monsoons. The south-west monsoon forecast is usually released between March and April to prepare farmers for the sowing season of kharif crops, and also to alert other users and departments well in advance.
The south-west monsoon operational forecasts are issued in two stages. The first stage forecast is issued in April and the second in June. The second forecast would depict a clearer picture of El Nino and its influence on the Indian monsoon. For the year 2017 IMD has premised the first stage of forecast — that India will receive normal rainfall based on the conditions in Pacific and Indian oceans. That is, quantitatively the rainfall is expected to be 96 per cent of the Long Period Average (LPA) with 38 per cent probability of variation of +/- 5 per cent error.
Improving accuracy
The forecasts are prepared based on the Statistical Ensemble Forecasting System (SEFS) and Climate Forecasting System (CFS), developed under monsoon mission. It uses five predictors namely the Sea Surface Temperature (SST) gradient between north Atlantic and North Pacific, equatorial South India Ocean SST, East Asia mean sea level pressure, Northwest Europe land surface air temperature and equatorial Pacific warm water volume.
However, the accuracy of the IMD’s prediction has gone wrong many times in the past year. For instance IMD had forecast normal rainfall (96 per cent of LPA) for 2014 but the actual rainfall was 88 per cent of the LPA in what was actually a drought year for the country. Similarly, 2015 was also marked as a drought year due to deficient rainfall (86 per cent of the LPA) when IMD predicted below normal rainfall (93 per cent of the LPA) for that year.
These forecasts are improved on a regular basis and the IMD has been able to bring down its average absolute error to 5 per cent (of LPA) for the period 2003-2015 from 7.9 per cent in 1990-2012. The forecast error during some years, that is, the difference between what is forecast and the actual rainfall, was over 10 per cent. Despite the technological advancement in weather forecasting, 100 per cent accuracy is not easy based on statistical models.