Vinson Kurian

Thiruvananthapuram , Aug. 9

THE calamitous `super drencher' that flooded Mumbai on July 26 has prompted meteorologists to go into a huddle and reflect on how Nature pulled a fast one to expose their level of preparedness in anticipating a contingency and issuing alerts.

A brainstorming session by the National Centre for Medium Range Weather Forecasting (NCMRWF), the medium-range forecaster, concluded that it had two options while dealing with such challenging tasks.

A presentation by Dr Akhilesh Gupta, Director, NCMRWF, said one option is to upgrade the observational system by deploying advanced Doppler weather radars, supported by a dense (10-km apart) network of observatories linking vulnerable zones for purposes of short duration prediction (made available few hours in advance).

The other option is the use of very high resolution (5-10 km) numerical models for exploring the feasibility of making predictions at least 24 hours in advance.

Giving a background, Dr Gupta said Mumbai (Santacruz) recorded an unprecedented 94.4 cm of rain on that eventful day. It now turns out that a nearby weather station at Vihar Lake recorded as high a precipitation as 104.5 cm. But, interestingly, Colaba, another suburb, recorded a paltry 7.3 cm.

The unusually high rainfall was caused possibly by a tiny offshore vortex of a size of 30 km embedded within the prevailing offshore trough.

The vortex was the result of the strong monsoon flow. It resembled a `cloudburst' event (generally observed in a hilly region) but had a relatively longer period of duration (about six to eight hours).

Tall clouds rising to 15 km in height were observed during the event. Normally, monsoon systems are not more than six to eight km in height.

Another feature was the severe thunderstorm activity and squally winds through the duration of the heavy downpour. Monsoon rains are not known to display such characteristics.

The India Meteorological Department (IMD) and NCMRWF had predicted heavy rainfall (of up to16 cm) over the Konkan coast (including Mumbai) that day. But the Mumbai station of IMD had in its local forecast for the day predicted heavy to very heavy rains over the city.

Beyond capability: "Both failed to predict an exceptionally heavy amount of 94.4 cm that fell over an area of 20 X 20 km. The simple and plain fact is that such intense rain activity over a short duration cannot be predicted several days in advance. The latest that one can hope to do that is a day ahead," Dr Gupta said.

Technology for prediction of such an event beyond one day does not exist in any part of the world. But the prediction of environmental conditions favourable for the build-up of such an event is possible with some accuracy. The region of heavy rainfall over a larger domain can also be predicted with reasonable accuracy.

Interestingly, a `re-run' (as distinct from real-time) of the high-resolution model of the UK Met Office predicted up to 80 cm of rain in Mumbai on July 26/27. But the global model of the agency could predict only up to 30 cm of rain a day in advance.

Ongoing efforts aimed at improving weather prediction modelling include collaboration with leading numerical weather prediction centres such as the UK Met Office, the European Centre for Medium Range Weather Forecasts, the National Centre for Environmental Prediction (NCEP) and the National Centre for Atmospheric Research (NCAR) of the US.

The two-year targeted Indo-US programme on land-falling cyclones in the Bay of Bengal is also on course. A similar programme worth a mention is the 10-year THORPEX project for the World Meteorological Organisation (WMO) for accelerated improvement in prediction of high-impact weather up to two weeks.

The computing infrastructure at the centre is being upgraded to enable running of high-resolution global/meso-scale models. The centre is also thinking of effecting a paradigm shift from computer-specific to service-specific models.

Significant enhancement in data storage and handling capability is was another option being explored. The centre is also working out a plan to bring research and operational groups on a fast communication network.

NCMRWF is an active partner in the NCAR's project for the development of the Weather Research and Forecasting (WRF) model. The WRF model for tropical cyclone (of about 1-km resolution) is being procured from the US as part of this collaboration. This will be tested and fine-tuned for the Indian region before being put into operation.

The centre is also making efforts to have close collaboration with NCEP on short/ medium/seasonal scale prediction of the monsoon, Dr Gupta said.

© Copyright 2000 - 2009 The Hindu Business Line