Research work in the Climate Prediction and Analysis Division
1) Understanding monsoon variability
Understanding and quantifying the Indian monsoon phenomenon, in terms of intra-seasonal
as well as inter-seasonal variability is a key focus of the work of this group. This includes
research into the pathways for modulation of Indian Summer Monsoon Rainfall (ISMR), on daily
to decadal time scales, its spatial variation in these time scales as well as its correlation
with global factors.
2) Identification of effective predictors
Identification of new tele-connection patterns and mechanisms of monsoon variability and
development of new predictors for long range forecasting.
3) Development of more accurate statistical models
Many new statistical techniques like Parametric and power regression models, Ensemble Multiple
Regression, Artificial Neural Network, Canonical Correlation Analysis, Discriminant Analysis,
and Pursuit Projection Regression have been used to develop newer and more accurate statistical
models for Long Range Prediction of Indian Weather- especially South-west and North-east monsoon.
4) Assessment of the skills of various atmospheric and coupled models.
Various centres of the world run different climate models for providing long-range forecast of
Monsoon rainfall. IMD also runs the Seasonal Forecast Model (SFM) model operationally.
In order to supplement the information from the statistical forecast model, the skill of various
Dynamical and Statistical models have to be evaluated according to their ability to simulate the
Monsoon characteristics over the Indian subcontinent. The models are quantitatively verified using
the latest methodologies and according to WMO Verification Guidelines.
5) Development of high resolution daily gridded temperature and rainfall data sets for the Indian region
The daily gridded temperature field was created using the modified version of the Shepard’s angular
distance weighting algorithm for interpolating the Indian observing station temperature data into 10 X 10 degree grids.
The inverse distance weighted interpolation (IDW) scheme proposed by Shepard (1968) is also being used to
convert station rainfall data into a gridded format. The daily gridded rainfall data, at 1.0x1.0 (fixed network
of 2140 rain gauge stations, 1951 onwards), 1.0x1.0 (fixed network of 1384 rain gauge stations, 1901 onwards)
and 0.25x0.25 (variable from a network of 6955 rain gauge stations, 1901 onwards) degree resolution are created
and regularly updated using the quality controlled daily rainfall records. The 0.25x0.25 degree resolution daily
gridded rainfall data is being created everyday on real time basis (using real time rainfall data reported out of
a network of 2774 stations), and Actual-Normal-Anomaly maps are plotted everyday using this data.
6) Drought / Flood climatology over India
A Standardized Precipitation Index (SPI) developed by McKee et al. (1993), are computed using southwest
monsoon season rainfall data of Indian region. The drought / flood climatology and forecast for the Indian
region, based on SPI is not biased by aridity or wetness, and hence, is a better drought / flood index for
drought / flood monitoring over smaller spatial scales. Studies have shown that the SPI is also suitable
for examining break and active events in the monsoon rainfall.
The list of research papers published by IMD on Climate Prediction and Analysis Division is given here.