Head of climate prediction at IMD, D Sivananda Pai explains the delay in monsoon and 'data gaps' in forecasts, says IMD is working towards making taluka-level forecasts, and calls upon citizens too to play a more proactive role as extreme weather events increase.
As head of the Pune office of the India Meteorological Department, D Sivananda Pai has been making long-range forecasts for the Indian monsoon season for several years now. He has overseen a vast improvement in the organisation's forecasting capabilities, thanks to the massive investment in both human and technical resources. As another erratic monsoon season increases anxieties about the amount and quality of rainfall across the country, Pai explains the uncertainties in the science behind forecasting, the improvements that have happened over the years, and the impact of climate change on rainfall activity
AMITABH SINHA: There was a large rainfall deficit in June, and there are fears that we might be staring at another year of drought. How is the monsoon likely to play out for the rest of the season?
As you know, this year, the onset of monsoon over Kerala was delayed by eight days. Though the monsoon arrived over the South Andaman Sea on time, its further progress was delayed. While it was getting energised and progressing landwards, Cyclone Vayu developed and further delayed the monsoon. As a result, there was a 10-15-day delay in monsoon over central India. Vayu was expected to cross Gujarat coast and reach Rajasthan and cause rainfall, but after it entered the cold region it travelled parallel to the west coast. The net result was that in the first 15 days (after the onset), there was a large rainfall deficit. Though monsoon picked up thereafter, it could never wipe out the deficiency, and in June rainfall ended up being 33% below normal.
Most sub-divisions fell under deficient category except north interior
Karnataka, Andaman and Nicobar, West Rajasthan and Jammu and Kashmir. But now, (in the last week) there has been some improvement in rainfall activity. A few days ago, for the first time in this season, the actual rainfall received was more than the normal for the day.
Rainfall activity is now concentrated over some parts of northwest, central and northeast regions of India, while rainfall over southern peninsula has seen some reduction after June. The all-India rainfall deficiency has now reduced to 27%. With the effect of a weak El Nino decaying further, we expect rainfall activity to improve in the remaining days of this month. As predicted, the July rainfall will be 95% of the Long Period Average (LPA) followed by 99% of the LPA for August and normal rainfall for September, taking the season's rainfall to 96% of the LPA.
ANURADHA MASCARENHAS: How difficult is it to predict rainfall given that the monsoon has become so erratic?
Rainfall has always been erratic, globally. What has happened in recent years is that the extreme rainfall events have increased. As a result, the gap between an extreme rainfall event and a dry spell has also increased. Interestingly, the Indian monsoon rainfall does not show any significant deviation from its normal trends. But there is a slight decrease in rainfall in recent years, possibly because of the multi-decadal variation.
Regarding the difficulty in predicting rainfall, it varies with the kind of forecast being made. There are short-range, medium-range, extended-range and long-range forecasts that the India Meteorological Department (IMD) does. The short-range forecasts (up to five days) are the most accurate. Thanks to better data availability, greater computing power, improvements in infrastructure, and greater efficiencies in modelling techniques, there has been a vast improvement in the forecasts. Now, it is possible to predict the formation of a low pressure system about 300-400 km away a week in advance.
As we increase the time-range that we are looking at for our forecasts, uncertainties start to creep in. It is not because of lack of infrastructure or expertise. It is in the very nature of weather forecasting. It is not an exact science. It is probabilistic and complex. We need to feed very minute data into the models, and this data is not always available for all areas. There are data gaps.
We are relatively well off collecting data from land, but 70% of the Earth is ocean, and collecting data from oceans is very difficult. We have instruments in the sea, also onboard ships, and even aircraft supply data with regard to wind for example. But there are still gaps. We try to fill these data gaps through approximations and extrapolations. That introduces uncertainties.
ANURADHA MASCARENHAS: But weather data is widely shared between countries, isn't it?
Yes. There are elaborate international arrangements for sharing of data. All countries which collect data, and are part of this system managed by the World Meteorological Organisation, feed the data into a network. All this data is available to all of us, in real time.
We feed this data into our models which try to mimic nature's behaviour. But nature does not always behave in a certain and deterministic manner. There is some amount of randomness there as well. So there are uncertainties at different stages of forecast, during data collection, during computations, in the models, and in the interpretation of results. A forecast is made while accounting for all these uncertainties, and that is why the forecast talks about probabilities.
AMITABH SINHA: While you say that there has been no major deviation from the normal trend, there seems to be a rise in extreme rainfall events in India. Why is that?
The overall amount of rainfall has remained more or less the same. Of course, within this overall scenario, there are variations. There are more heavy rainfall events now, and the gaps between a wet period and a dry spell have increased.
The distribution of rainfall has also become more extreme, so either there are heavy rains or no rains at all. So, there are some trends of change. These trends show that central India, the western coast and northeast India are generally receiving more rainfall than earlier. If one considers the seasonal average for the country, there is no change.
AJAY JADHAV: Has there been any variation in the pre-monsoon showers as well? This year, for instance, many regions did not receive any rainfall before the season.
This year there was less rain during the pre-monsoon season. But it is important to know that rainfall during this season is mostly limited to the southern peninsula and the northeast regions. Rainfall in the northeast region was normal, but for the southern states, the rainfall was less. However, Jammu and Kashmir, Punjab, and Haryana got good rainfall during these months, benefiting cultivation in these parts. Haryana and Punjab are presently at comfortable positions as far as rainfall is concerned.
MANOJ MORE: The frequency of El Nino (an abnormal warming of the Pacific Ocean) is also increasing. So are we going to see more and more bad monsoon years?
2019 is the fifth El Nino year since 2000. El Nino generally starts developing during pre-monsoon (March to May) season, intensifies thereafter and peaks in the next winter (December-February), and then weakens during the March to May period… This can go on for even two years.
For its impact on monsoon, we check whether El Nino is present during the monsoon season or not. After November last year, El Nino gained in strength, then peaked during December-January and thereafter it has been weakening. As of now, the Nino 3.4 region (the region in the Pacific Ocean where an increased warming of waters seems to have the maximum impact on Indian monsoon) is 0.30 Celsius warmer than normal and Nino 4 region is 0.20 warmer. A significant impact of El Nino on Indian monsoon is observed mainly when these temperature anomalies are more than 0.50 Celsius.
AMITABH SINHA: So, is this an El Nino year or not?
Yes, it is an El Nino year. It will be counted as a weak El Nino, which will have relatively less impact. If one considers the present Sea Surface Conditions (SST) in the Indian Ocean, then it is conducive for Indian monsoon and forecasts indicate a positive Indian Ocean Dipole (IOD is alternative warming and cooling of waters in the Indian Ocean). Collectively, these two (weak El Nino and positive IOD) should ensure that the monsoon is not impacted adversely in any significant manner. That is our assessment. So in comparison to June, July and August rainfall should be better.
SUSHANT KULKARNI: How has computational technology changed our abilities to make accurate forecasts?
You need very fast supercomputers for weather forecasting. That is because of the large amount of parameters and the extremely complex calculations. Normal computers would take days or weeks to do these calculations which would make these useless because we need to issue forecasts in real time.
In the last few years, we have added substantially to our computing power. In fact, the high-powered computer that we have in Pune now is the third most powerful computer being used in weather forecasting anywhere in the world. This has improved our capabilities tremendously. Until five years ago, for example, to get an output for short range forecasting, we needed to run the weather model for two to three days. Now, with advancement in technology, we can get the output within 1 hour to 1.5 hours.
ANURADHA MASCARENHAS: You spoke about data gaps. How are climate studies managed in such a scenario?
Well, the data gaps exist in some parts of the world. Over the Indian region, there is no problem of data. We have a fairly good network of weather stations and instruments everywhere. Of course, we need to strengthen this network even further if we are to provide more granular local forecasts. And as I said earlier, wherever data gaps exist, we make do with approximations and extrapolations.
GEETA NAIR: How difficult is it to provide more local weather information?
As of now we are providing district-level forecasts for every district in the country. This is pretty good information. But as per the directions of the Prime Minister's Office and the NITI Aayog, we are also working on providing information at the taluka level, particularly for agriculture. This will take some time. We need to increase our monitoring network, we need to increase our computing power even further, and then we need to create abilities to prepare useful and relevant products to be disseminated at the local level. We are moving towards that.
AMITABH SINHA: What are the climate change trends that are becoming evident over India in terms of precipitation or temperature patterns?
According to the observed data for the last several years, there is no doubt that there has been a significant increase in the intensity and frequency of heavy rainfall events. And this is likely to continue. On the western coast, it is not the intensity but the frequency of rainfall events which seems to be on the rise. A similar trend has been observed in the Northeast region as well. In central India, however, both intensity and frequency of heavy rainfall events are increasing.
Similarly, heat wave incidents are also on the rise. There is a heat wave corridor, starting from Rajasthan, Gujarat, central India, Odisha, Andhra Pradesh… In most of these areas, there is an increase in heat-wave incidents. The incidents of cold wave, however, are showing a decreasing trend.
AMITABH SINHA: The northeast region usually gets good rainfall, but last year there was a large deficit. And this year too, till now at least, rainfall in the region has not been very good.
Last year's shortfall could be explained as part of annual variations. It's just been one month into the season this year. So, no, there is no special trend there.
But we have noticed an increase in the low-pressure systems during the monsoon season. These systems generally move in the westerly direction, from the Bay of Bengal towards Rajasthan along the monsoon trough. Whenever there is a depression in central India, the Northeast gets a bit less rain, mainly because it sucks in the moisture from surrounding areas. This increase in low pressure systems could also be the reason, at least partially, why more heavy rainfall events are being witnessed in central India.
GEETA NAIR: Was the recent deluge in Mumbai a result of bad urban planning or a deficiency in our ability to forecast heavy rainfall?
See, because of climate change, heavy rainfall events are increasing. That is very clear in India as well. There is little that can be done about that. But we need to learn how to manage these things. Even common people like us have to be sensitive to these things. For example, we have to make sure that we do not do anything that would choke drainage. I think this requires a cultural change from our side as well. The government agencies, no doubt, have a more proactive role to play but citizens must also pitch in. In the Mumbai event, heavy rainfall was predicted. But the urban management systems have to improve.
SUSHANT KULKARNI: How sensitive are other government agencies to the information provided by the IMD?
I think there is much more awareness now compared to any other time about the impact of weather and climate. More and more people, and also government agencies, now realise that the work they do, or even their normal lives, can be weather-sensitive, that these can be impacted by weather events. The cricket World Cup is currently going on. Weather information plays a vital role in team strategies these days. It was not this important even a few years ago. So, yes, there is much greater awareness about climate change and the impact it can have.
GEETA NAIR: The IMD produces huge amounts of weather data. Have you considered building a start-up-like ecosystem to use the data more effectively?
We get lots of requests for our data. Now we are putting a lot of data - mainly rainfall and temperature data - online, making it freely available to everyone. Earlier, we used to charge `10,000 to
`15,000 for this; over 1,000 USD from foreign agencies. We can't, of course, put all our data online, because it has commercial value as well… There are many people who can make useful weather products from this data and disseminate it to their clients. We welcome any such proposal, and we support them.
ANJALI MARAR: Is the Arabian Sea seeing an increased cyclonic activity? Is the western coast becoming dangerous?
There is nothing in our data to suggest that. And, in any case, most of the systems formed in the Arabian Sea generally tend to move towards the West, towards Oman and other places. These rarely come towards the Indian coastline.