Mumbai’s skyscrapers could be affecting rainfall, forecast model is inaccurate, says IIT-B study
A six-member team from IIT-Bombay carried out extensive statistical modelling using data from automatic weather stationsmumbai Updated: May 15, 2017 09:33 IST
The city’s rising skyline could be affecting rainfall patterns, making the current monsoon prediction model less effective, says a study by the Indian Institute of Technology - Bombay (IIT-B). The study, a first-of-its-kind analysis on rainfall patterns based on data from Automatic Weather Stations (AWS), said there was a need to create a new model.
At present, the India Meteorological Department (IMD) – the official forecaster – uses the Ensemble Statistical Forecasting System, which uses 100-year old rainfall data and an average of six meteorological values such as sea surface temperatures in the Pacific and North Atlantic sea-level pressure to forecast the monsoon.
The study by a six-member team from IIT-B carried out extensive statistical modelling on data from AWS, which was set up by the municipal corporation after the July 2005 deluge, when the city received almost 944 mm rain on a single day – almost 40% of its annual rain. Results showed that non-uniformity in daily rainfall was gradually increasing. Also, rainfall at night was lesser in terms of volume and intensity compared to daytime rainfall. This rainfall variation over Mumbai, said researchers, is so high that no statistical model will be able to capture. The study assumes significance because a more accurate forecast system will help policy makers mitigate and be better prepared to handle extreme rainfall events.
“Mumbai is a very heterogeneous urban region. Heterogeneous building layers create heterogeneity in atmospheric instability as a result of which there is no systematic rainfall pattern,” said professor Subimal Ghosh, department of civil engineering, IIT-B. “Therefore, despite having a 15-minute prediction system in Mumbai, statistical forecasting model is not working for lack of a systematic rainfall pattern.”
Researchers said Mumbai needs a mechanistic model – such as Weather Research and Forecasting (WRF) model – which is a physics-based model, to understand the city’s precipitation pattern.
“Rainfall in Mumbai is very complicated, its meteorologically very difficult to understand. In addition to urbanisation, the city’s monsoon is influenced by the coast. Situated on the windward side of the Western Ghats also brings in heavy rainfall over Mumbai,” said professor Subhankar Karmakar, Centre for Environmental Science and Engineering, IIT-B, and co-author of the study. “Therefore, understanding rainfall only through statistical modelling is not enough; it doesn’t work for Mumbai.”
Before performing a statistical analysis, researchers expected good results because of the availability of rainfall data via 60 automatic weather stations across the city. These stations were set up post July 2005 deluge because meteorologists had then said they can’t forecast rainfall accurately for Mumbai as there was no proper rainfall data. AWS captures meteorological parameters such as humidity, precipitation and wind speed.
“But our results did not conclude anything meaningful because there is no spatio-temporal pattern of rainfall in Mumbai,” said Karmakar. “We can’t say rainfall was high or low in a particular month; if there is any spatial correlation in rainfall between two close-by stations, and even wind speed is not much of an influencing factor. Therefore, Mumbai rainfall has a very heterogeneous pattern.”
Even though the team analysed AWS data from 2006 to 2015, sub-hourly data from 26 AWS between 2013 and 2014 used for further analysis due to their consistency and continuity. The study has called for continuous validation of AWS data by the civic body through partnerships with institutions that can be run on the physics-based mechanistic model along with different meteorological data and an updated land-use-land-change map.