The science behind predicting heat waves

Forecast accuracy for severe weather events has increased by about 50% in the past decade. Heatwaves can now be predicted 24 hours earlier, with 88% accuracy

The India Meteorological Department (IMD) sent out several heat wave alerts for various regions in the country in the past fortnight.

Heat wave forecasting by the agency has become significantly more accurate in the past few years. (PTI)

Even places like Mumbai where temperatures are typically cushioned by high moisture content because of its coastal environment weren’t spared. That said, Mumbai reportedly has had fewer heat stroke cases compared to this time in previous years. The reported fall in cases is being attributed to better implementation of heat wave action plans.

For local governments to be more prepared for extreme temperatures this summer, they must receive accurate, informative and timely forecasts from the IMD, which is the agency responsible for meteorological observations, weather forecasting and seismology in India.

Fortunately, heat wave forecasting by the agency has become significantly more accurate in the past few years.

What is the science behind heat wave predictions?

There is no universal method or index for declaring a heat wave. The nature of a heat wave depends on the geography and the time of the year, so definitions too vary. India’s National Disaster Management Authority (NDMA) defines it as a period of abnormally high temperatures, more than the normal maximum temperature, which adversely affects people living in these regions as they cause physiological stress, sometimes resulting in death.

The IMD follows certain criteria that need to be fulfilled before a scenario can be labelled as a heat wave, and these are based on the maximum temperature and the deviation from normal temperature at that particular weather station.

Heat waves typically occur in pre-monsoon months. Between eight to 10 such days are expected during this season but the past decade has seen more frequent and longer-lasting heat wave periods. Fortunately, meteorologists now have a better understanding of the dynamics of heat wave events. For example, we now know that these are not local phenomena.

“Though local conditions such as lack of rains and lack of soil moisture can support heat waves, they are actually caused by large-scale atmospheric circulation processes,” said MN Rajeevan, a meteorologist who was formerly secretary to the earth sciences ministry.

“Knowing the basic dynamics of heat waves means that we can predict them three to four days in advance with good confidence, and we can also predict how long they will last,” said Rajeevan who also served with the IMD for over 20 years.

How IMD’s heat wave forecast works

IMD relies on its large network of surface observatories located across the country to collect data on temperature, relative humidity, pressure, wind speed and direction. They use historical daily maximum temperature data to calculate the normal maximum temperature, and based on this, they define the criteria for heat waves in a particular region.

IMD scientists then feed initial conditions and observational data into numerical weather prediction models to generate the forecast. Different models are used for different types of forecasts — short-range, extended range and seasonal. “The model does not say that a heat wave will come,” clarified Rajeevan: “It can only predict the parameters conducive for a heat wave.”

Heat wave data accuracy

The forecast accuracy for all types of severe weather events has increased by about 50% just in the past decade, which is likely due to more data, better data processing methods and more research into understanding the physical mechanisms behind the phenomena.

Heat waves, specifically, can now be predicted 24 hours earlier, with 88% accuracy. In a paper published in the journal Earth and Space Science last year, Rajeevan and his co-author P. Rohini, evaluated the efficiency of various models in predicting heat waves in India.

Their studies showed that the European Centre for Medium-Range Weather Forecasts (ECMWF) model does particularly well at predicting heat waves in India. The researchers deduced that this could be because the mathematical equations of this model are capable of predicting large-scale atmospheric forcing, a climatic event that has been associated with heat wave events over India.

The primary parameter used by IMD to monitor heat waves in the country is daytime or maximum temperature. This makes sense because it is during the day that most humans are expected to be awake and outdoors, making it the time we are most vulnerable to extreme weather. But Rajeevan pointed out that heat can affect our bodies at night too.

Another aspect to keep in mind is that it is not enough to consider temperature alone to measure a heatwave. Even temperatures of 35 degrees can be detrimental to us if the humidity is 70% - a scenario playing out in many coastal regions of the country.

Rajeevan strongly suggested that India begin using indices such as the heat index used in the US that combines air temperature and relative humidity to provide ‘apparent temperature’ or what some weather apps depict as ‘Feels like’.

But even this may need to be customised for the Indian physiology. “Our bodies respond differently so IMD is trying to define what our threshold is,” said Rajeevan. This is tricky to do because there are wide differences - for example, age and lifestyle - even within our own population.

The future of heat wave forecasting

How much better can heat wave predictions get? One study published in Scientific Reports demonstrated that machine learning models can complement numerical models to predict maximum temperatures up to 10 days ahead. Beyond this may not be possible. “There is such a thing as a predictability limit,” reminded Rajeevan. “It is nature, after all.”

Nandita Jayaraj is the co-author of Lab Hopping: A Journey to Find India's Women in Science, which explores the gender gap in Indian science.