You are here

Heralding a new age of better weather forecasts

Scientists from IISc have come up with clever algorithms to improve the prediction of temperature in remote areas, which in turn can improve the accuracy of large scale weather events.

In order for meteorologists to forecast large-scale weather events like monsoons, typhoons, floods, droughts and desertification, they need to study hydrometeorology, or the study of the water cycle – how water moves from the earth to the atmosphere and back again. While today’s level of understanding on the subject is pretty good, there are still gaping holes in our knowledge that prevent us from accurately predicting large-scale disasters. One of the major flaws in today’s system is insufficient data.

Detailed data is available for certain regions, while more remote areas typically have lesser data. Since the information is incomplete, the precision of our forecasts also suffer. R. Bharath, V. V. Srinivas and B. Basu from the Department of Civil Engineering at the Indian Institute of Science have devised an ingenious method to improve this state of affairs.

Temperature is a key variable in hydrometeorology because it is necessary to estimate evaporation and transpiration rates. These play a huge part in the cycle of water into the atmosphere, and its return, making it integral to our understanding of diverse phenomena, like the probable rate of inflow into a dam, to determining how soil in a particular region holds its moisture. Known as spatial-temporal modelling, understanding how temperatures vary geographically over time provides us with key insights, which prove to be exceedingly beneficial in the fields of agriculture and disaster management.

The study undertaken by the IISc scientists uses cleverly designed algorithms, and predictors of temperatures to delineate and define special regions called 'homogeneous temperature regions'. They have used the available measurements of large scale atmospheric variables (LSAVs). Because of their calculations, the weather conditions of remote areas, where measurements of LSAVs are infrequent and arbitrary, can be extrapolated from available data with reasonable accuracy.

While this has been done using different approaches before, no other method has enabled scientists to achieve this level of accuracy, because most previously determined temperature regions depended too closely on conditions susceptible to ‘sampling variability’.

Defining the homogeneous temperature regions has been a huge undertaking, but it is only the beginning. “Research is underway to extend the proposed approach to delineate homogeneous hydrometeorological regions by simultaneously considering a larger set of variables such as precipitation, temperature, windspeed and humidity.” says Dr. V. V. Srinivas. “Such regions are expected to be useful to arrive at estimates of those variables for ungauged catchments of river basins, and those estimates could be subsequently utilized to estimate streamflow/flood and other hydrological processes in the ungauged catchments.”

What does their research mean in the immediate future? Well there are a host of opportunities for meteorologist to put the new information to use. Dr V. V. Srinivas says, “They can consider the newly delineated homogeneous temperature regions to study regional trends in mean, maximum, minimum and/or diurnal temperature. Further, information concerning spatio-temporal patterns of temperature and their variability is necessary to model various surface processes at global and local scales in disciplines like hydrology, anthropology, agriculture, forestry, environmental engineering and climatology. Temperature influences biological events like diseases, phenological events (e.g. the timing of natural events such as flowering, breeding) and agronomy.”

About the authors

R. Bharath, V. V. Srinivas and Bidroha Basu are at the Department of Civil Engineering, Indian Institute of Science, Bangalore.

Contact: V V Srinivas vvs@civil.iisc.ernet.in

The paper appeared in the INTERNATIONAL JOURNAL OF CLIMATOLOGY recently.

DOI: 10.1002/joc.4335