Met office failed to predict intensity of rainfall

As Jammu and Kashmir battle unprecedented floods, the question that comes to everybody's mind is: Could this have been predicted? A warning along with forecast of the possible magnitude of the rains and an alert state administration could have saved precious lives and misery.

When meteorologists can successfully forewarn about cyclones with more catastrophic consequences, why the same could not be done for flood, which is a reality for India for centuries?

Flood forecasting, scientists say, is a different ball game. While models are available, they have to be customised for each of the river basins by feeding them with several types of local data collected from space as well as ground surveys. It is a long-drawn, time-consuming job that involves coordination of multiple institutions under the Central and state governments.
The government's approach, so far, is to accept flood as a reality and prepare for relief and rehabilitation.

Flood alerts start with accurate rain forecasting. To be fair to the  agency, Indian Meteorological Department warned the Jammu and Kashmir administration in the first week of September about “heavy (6-12 cm) to very heavy (12-24 cm) rainfall.”

The state administration issued warnings from  mosques and police vehicles. Unfortunately, not many took the warnings seriously.

 “In any disaster situation, there are three factors – forecast, response of the administration and people’s response. People consider cyclones as hazard and are ready to evacuate but don’t consider rain as a hazard. In Phailin, 24 people died in the cyclone whereas as many as 80 lost their lives in the rain that followed,” Shailesh Nayak, secretary in the Ministry of Earth Sciences that governs IMD, told Deccan Herald.

The administration and warnings failed to properly communicate the gravity of the situation that people experienced later. In nine districts, precipitation was more than 200 per cent of the average normal rainfall. In most of the weather stations, it was between 15-20 cm and one station, Katra recorded 27 cm rainfall. The heavy downpour threw the administration out of gear.

Many scientists admit India’s weather forecasting system has to be improved both in quality and quantity. It is possible only with better instrumentation, climate models and computation.

IMD’s existing climate models currently run on a 9 km by 9 km scale and fed by a limited number of data sets from the ground and space. The Ministry of Earth Sciences has recently acquired super-computers and climate models from the USA and UK to make predictions in 1.5 km by 1.5 km grid within a few minutes. But it would take at least a couple of years to test and validate the models. More data collection platforms have to be installed to improve the prediction accuracy.   
 
The flood forecast system, too, is archaic. The Central Water Commission (CWC) began flood forecasting in 1958 with only one station. The number of stations was increased to 84 in 1980, 145 in 1985 and 175 in 2006. Since then there is no addition of new flood forecasting stations. A new trend of urban flooding, however, has emerged in the last decade. Srinagar also  is facing the same on a larger scale.

Flood forecasts

The CWC flood forecasting network covers most of the traditional flood prone inter-state river basins. Flood forecasts are issued for 175 stations, out of which 147 are for river stage forecast and 28 for inflow forecast for reservoirs and barrages. The CWC historically focuses on Ganga (87 stations) and  Brahmaputra (27) though there are a few forecasting stations for the Godavari (18), Krishna (9), west flowing rivers including the  Tapi (15), Mahanadi (4) Barak (5) and Pennar (1). The network covers only 17 states.  Jammu and Kashmir is not one of them.
While the CWC takes into account the real time water level rise in rivers, the run-off factor, storage capacity in the barrages and history to give flood alerts three days in advance, new scientific models on flood forecasting are more comprehensive.

Several such models like SWAT and HEC-RAS are globally available. The National Centre for Medium Range for Weather Forecasting, Noida, too developed a flood forecasting model many years ago. But all of these models are required to be customised for each river basin.

For that, several local data has to be collected and fed into the model and its prediction has to be validated before the models are put to operational use. This is a mammoth exercise requiring coordination between several Central and state government agencies.

“What is lacking is a vision or political willingness. We need a national mission on flood forecasting,” said a scientist who worked on a forecasting model. After being stuck in the files for years, the government cleared a plan to have a Himalayan meteorology network only after the Uttarakhand flood in 2013 and the first Doppler radar is expected to be installed in Srinagar by 2014-end.

But to place orders for other instruments would take another six to eight months as specifications are still being drawn up, says Nayak.

Several years ago, the National Disaster Management Authority brought out detailed guidelines on the management of floods in general as well as urban floods like the one that hit Srinagar. But the guidelines were never converted into an actionable programme, even though several of the Indian metropolis was hit by urban floods in the last 15 years.