Will It Rain Today? Will It Not? Why Accurate Local Weather Forecasts Are Elusive To India
While weather forecasting accuracy is moving towards better accuracy and precision in India, challenges remain.
How many times did you get stuck in the rains this year?
How many times did you wish you knew when it would rain and when it would stop? On paper, it’s possible. But our forecasters never seem to get it right. Why?
Weather forecasting is moving towards better accuracy and precision in India, but factors ranging from network density to the country's tropical climate and topography, continue to pose challenges, GP Sharma, president at Skymet Weather Services, said.
Edited excerpts of the conversation:
Just considering how much rains affect us on a day to day basis, how accurate is localised Indian weather forecast?
For a larger area, the forecast for a majority of the period will turn out to be true barring some short term intermittent or very temporary drizzle.
Sometimes you have some low clouds which can give a sprinkle for a few minutes and then stop. Very short scale forecast of rain or drizzle is rather impossible to forecast. The possibility or the probability can always be mentioned, but a forecast is either yes or no; it's there or not there. In those terms it's a challenge.
In forecasts, we generally talk about a larger scale of weather phenomena. Localised forecast depends on local factors. For instance, prominent cities like Delhi NCR have become heat islands (urbanized areas that experience higher temperatures than outlying areas). Water bodies, forest covers also add to the localised conditions. Be it winters, summers or the monsoon season, localised factors come into play and it's very difficult to take stock of that.
Localised weather forecasts in India don't appear to be as accurate as in some other parts of the world. Why so?
Firstly, I don't agree straightaway that they are more accurate. We are comparing two different things. Weather phenomenon have a wide spectrum and what we are talking about is two different ends of the spectrum. Mid-latitude like in London and tropical weather conditions like in Delhi are incomparable. What we have is more challenging. Meteorology conditions are such that the tropics always throw more challenges for the forecaster than the mid-latitudes.
It's very easy to track the system in the mid-latitudes. The pressure variations are so large there, that it is easy to track the system; you can infer that something's going to happen. The pressure changes are minimal in the tropics, you won't even come to know that rain is there in the offing.
Also, the land and sea distribution is more uneven in the tropics, leading to variability.
Did the European countries forecast the heat wave in advance? Limitations always remain everywhere.
What is the margin of error when we speak about localised forecasting?
The platform of observations which you have, the density of network which you have, the human resources which you have at your disposal together make your forecasting successful.
Some regions are better placed as they have a very dense network of observations. Whether you run the numerical weather models or you run any other forecasting model, the density of network matters most. We are not where some other countries are in terms of density of network.
Technology has improved. We have added many things over a period of the last 15-20 years, with a visible improvement in forecasting standards. But our density of the platforms and network is still lacking.
Despite the prediction of a normal monsoon the past two years, we have not been able to accurately forecast what regional distribution would look like. Why?
The four homogeneous region are not actually homogeneous. A wider network area forecast is broken down to the regional focus. We have four homogeneous regions, but these areas are not homogeneous. There is a heterogeneity of the terrain. Forecasting becomes more difficult depending upon the variability of the terrain. It becomes a challenge.
I think we have improved even in the regional forecasts.
The central parts of the country are the focus because they are drought prone. These parts are forecasted to do well this monsoon and they are. Northwest India is forecasted to do better this monsoon after it failed last year and the year before.
Decoding monsoon always remains a big challenge for any scientist and for any agency. World over, agencies are studying this phenomenon and issue forecasts and successes and failures are a part of that. This year's monsoon forecast is on track.
With climate change, do you think it becomes harder to accurately forecast weather?
Officially, the monsoon is a four-month season from July 1 to Sept. 30. When you talk about climate change and global warming, this is the stark reality and there is no denial. Climate change and global warming gives more intense spells for shorter durations. Be it heat wave or cold wave or floods, frequency of those incidents increases. For instance, extreme rainfall will occur in a short duration of time or cloudbursts will happen more.
As far as the monsoon is concerned, intra-seasonal variation may not be there but seasonal variation will happen.
Coming to the Southwest Monsoons, almost two months down and two more to go. While the season began with deficient rainfall, we are currently 11% in excess of the normal as per IMD. What's the outlook for the next two months?
Our forecast is of a normal monsoon at 98% of LPA with a margin of error of plus/minus 5%. Skymet has definitely expressed some apprehension that the second half of the monsoon may not be as good as the first half, especially in the month of September.