NCMRWF – shaping India’s weather forecast

NCMRWF – shaping India’s weather forecast


National Centre for Medium Range Weather Forecasting is constantly strengthening its resources to improve the weather forecasting capabilities of the country. In a tete-e-tete with Geospatial Media and Communications, Dr Swati Basu tells us more…

Dr Swati Basu
Dr Swati Basu
National Centre for Medium Range Weather Forecasting

Can you tell us about the mandate of National Centre for Medium Range Weather Forecasting (NCMRWF)?
NCMRWF’s mandate is to develop numerical weather prediction models for ultimate use by the operational agency of the country, that is, the Indian Meteorological Department (IMD). Through assimilation of various platform based observations, we try to improve the resolution of the models as well as improve the physical processes through parameterisation in the models, because it is essential that we are able to represent whatever physical processes are happening outside in the model as realistically as possible. So ultimately NCMRWF’s mandate is that we work on the advanced improved version of the operational model and after testing it, we hand it over to IMD for operational usage.

What’s the kind and scale of data provided by you?
The model generated output, let’s say, in terms of wind, temperature, rainfall, humidity – these are the basic parameters which are provided to IMD. The satellite data and other related things are an initial input to the model. We have a robust network of observatories all over the country as well as satellites and ship-based observations – information from all these networks are actually the current observations. Based on these current observations, we obtain the output which helps us determine the future trend. So, from basic parameters like rainfall, wind, temperature, pressure, we try to determine what’s going to be the future weather conditions. For example, suppose a cyclone is being formed today. If we want to know how the cyclone is going to behave in future, we need to look at the surface pressure, wind pattern, etc – these are the basic parameters and we provide these to IMD. Based on this information, they provide the warning, if necessary.

What kind of technology are you using to predict weather?
We are using 24 teraflop supercomputer. That’s why we are able to run high-resolution models. In order to understand how it works, we need to understand the meaning of resolution, let’s say, we have a 50 km resolution model. It means that if we divide the entire globe into small grids, these grids would be of 50:50 km. In other words, we can say that everything is homogenous. You can then imagine what it essentially means is that almost entire Delhi has the same weather. So the assumption that is going into this kind of model is unrealistic. Hence, we require a great deal of enhancement of computational resources for having a 10 km resolution model. Right now, with this new computer, we are able to run 22 km model. We have to go further down.

New technology involves training people as well. What’s being done in this regard?
NCMRWF scientists have been working on the high-performance computer since its inception (2009). Our scientists are quite competent because they have been dealing with high-power computer systems since a long time. Of course, from time-to-time, there is a need to train people. We do have some international collaborations with USA, UK, Korea, which involves exchange visits, research opportunities etc.

In what way is geospatial technology being used by you in your day-to-day operations?
From satellites, we get the total picture, that is, total moisture content etc – those are some basic things. Right from the surface till the top level of the atmosphere, let’s say, till 50 millibar or so, what we are interested in is knowing about the distribution of wind, humidity or water vapour at each point in horizontal as well as vertical plane. That’s why algorithms have to be developed, to know how these things are distributed in the vertical. For example, why is it that some cyclones get enhanced as they move while others just fizzle out? It is because of the combination of wind, pressure, humidity – how they are distributed in the vertical. That’s why their distribution is very important to us because it forms part of the initial profile. Once we get the initial profile, we need to know how it will change with time, only then we can predict better. So the initial distribution has to be as realistic as possible. This is where geospatial technology helps us understand both the vertical as well as the horizontal distribution – that’s the basic thing of what’s important to us. Vertical is very important – that’s actually very crucial because you cannot have observations at each point so somewhere something has to be derived. And whenever we are deriving something, there are bound to be assumptions. Everything depends on how realistic those assumptions are going to be. That’s why we keep on improving upon all these things. And that’s why some special observational campaigns are also taken up just to know the basic nature of this atmosphere.

Can you tell us how you communicate the weather information among various agencies and users?
There is a lot of data exchange between various departments of MoES, for example, between IMD and NCMRWF. We use high-performance computers in four units of Ministry of Earth Sciences (MoES) – IMD, IITM Pune, NCMRWF and INCOIS. All the high performance computers are connected so that we can optimally use the computational resources – whatever is available at any place. Then there is communication towards the user community. It’s not only important that we produce the forecast in time, it is also equally important that the information is disseminated to the end-users. IMD is doing a lot of work in this direction because our main users are farmers. IMD sends regular SMS and all these kind of services –I think its SMS services reach about 20 lakh farmers. We give the forecast to IMD and it takes it further to farmers. In fact, NCMRWF was created for farming community only, so these agro-advisory services was originally developed by us. After testing it properly, we handed it over to IMD who have now taken it further and it is now being used at the district level. Dissemination of information is mainly IMD’s job.

Can you tell us about your users? Which all organisations contact you for weather information?
We have Indian Air Force and Indian navy as our users. Another major user is Indian Army. It is important for army men to know weather conditions especially when they are scaling mountainous expedition. At that time, they require information about weather within a very small window, say, for next 24 hours. The information is critical for them to decide whether to go ahead with their expedition or not. Our users also include ISRO. It requires information about the basic parameters like wind, humidity, temperature, etc, at the time of the launch of satellite.

What kind of changes do you think we require in our system to be able to keep pace with today’s technological changes?
I think we are doing pretty well. You must realise that the model that we are running now is basically the state-of-art model which is being used in US also. So that way, our computational resources and the model are at par with other major global centers. But, of course, we have to keep on upgrading our systems. Because demands are increasing and we know exactly what is required – high-resolution models, more observations and assimilation of satellite as well as ocean-based observations. We have the experience of assimilating all kind of observations. So right combination of computational resources and constant improvement of resolution of models will be able to take us forward. At present, I can only say that as far as the kind of high-resolution models are concerned, we are running almost the similar thing that is available with the major global centers. We are constantly in touch with our foreign counterparts about the kind of models they have.