The 25 June 2009 issue (Vol 96, No. 12) of Current Science carries this review of the book “Monsoon Prediction” by Prof. R. R. Kelkar:

The history of monsoon prediction in India dates back to 1886. India Meteorological Department (IMD) has the distinction of being the first national weather service in the world to start the operational monsoon prediction work. It all started in 1886, when Blanford used the relationship between winter Himalayan snow cover and Indian monsoon for predicting the ensuing monsoon rainfall. Sir Gilbert Walker, in 1910s and 1920s, initiated high-quality research in understanding the monsoon variability and development of statistical models. Since then, meteorologists have been working relentlessly to understand and predict Indian monsoon variability and numerous research papers have been published on this important topic. Many reviews have also been published. Monsoon prediction is now a challenging research area with plenty of scope and opportunity. However, we did not have a good compilation of all the results pertaining to monsoon prediction research. The book under review fills this gap. The author, R. R. Kelkar, who served IMD for more than 35 years was also responsible for operational monsoon prediction in IMD. He has used his rich experience in shaping this useful book. The book also includes a critical analysis of our prediction capability and its future scope and opportunities.

The book contains six chapters, describing the major components of the monsoon, global and regional teleconnections, different methods of seasonal monsoon prediction, characteristics of intra-seasonal variations and prediction, projection of monsoon in the 21st century and finally, problems and prospects associated with monsoon prediction.

The first chapter describes different ways of defining a monsoon, monsoon regions, and major climatological features of the Indian monsoon. There are many ways of defining a monsoon. But for the Indians, it is the annual cycle of rains over the country. Monsoon theories included the simple land–sea breeze theory to a complex theory that involves the atmosphere, land and ocean. Indian monsoon variability is influenced by many components of the global climate system. They are known as teleconnection patterns, linking the Indian monsoon variability to major climate anomalies around the world such as Eurasian snow cover and snow depth, El Nino/Southern Oscillation (ENSO), sea surface temperature anomalies over the Indian and Atlantic oceans, and land surface temperature anomalies.

In the second chapter, various monsoon teleconnection patterns are discussed. Teleconnections with ENSO and snow cover are probably studied more and are discussed in detail. However, a detailed discussion on the role of Indian Ocean climate anomalies (Indian Ocean Dipole and EQUINOO) on Indian monsoon is missing. The relationship between sunspot activity and Indian monsoon is a research area with some renewed interest.

In India, statistical models are extensively used for monsoon prediction (long-range forecast). Most of these statistical models are based on the same principles used by Gilbert Walker in 1920s. IMD now uses an updated statistical model for issuing operational long range forecasts based on the ensemble technique. Chapter 3 discusses the details of these statistical models, including the parameters used in the model. The models include simple linear regression, discriminant models for probability distributions and the latest ensemble method based on projection pursuit regression.

Prediction of the quantum of rainfall during the whole season (seasonal prediction) is useful for planning purposes. However, for the farmers, prediction on intra-seasonal timescales (for example, 15–20 days) is more relevant for their day-to-day operations. Similarly, prediction of intra-seasonal activities such as onset and withdrawal, and the active–break cycles of the Indian monsoon is crucial for agricultural planning. However, little work has been done on this important area. Prediction of various monsoon intra-seasonal elements is discussed in Chapter 4. This is an emerging area of research both in India and abroad. This chapter also includes a brief discussion on the use of atmospheric and coupled general circulation models for seasonal monsoon prediction.

In the particular context of climate change due to global warming, it is important to know the future projections of the Indian monsoon. Chapter 5 discusses the projection of the Indian monsoon behaviour in the 21st century. The chapter starts with a discussion on the observed climate change over India using instrumental datasets, followed by future emission scenarios and climate projections. Even though there is no appreciable trend in the quantum of monsoon rainfall averaged over the country, there are significant trends in monsoon rainfall over different parts of the country. Also, there are temporal changes during the monsoon season. Most of the current climate models indicate that monsoon rainfall would increase in future climate. However, there are many uncertainties in the climate model projections. Most of the current climate models are unable to simulate the observed features of the Indian monsoon in their totality. It is important to understand and appreciate these uncertainties in the global climate models in interpreting and using the future climate-change projections.

In the final chapter, the problems and prospects of monsoon prediction are discussed. There are known limitations of the statistical models, like secular variations of correlation, choice of optimum number of predictors and training period, etc. However, in spite of these known constraints, statistical models continue to be popular for several reasons. The users, while interpreting forecasts from these statistical models, should understand and appreciate the limitations of these models. There are other issues like the definition of all-India summer monsoon rainfall (averaged over the whole country and whole season from 1 June to 30 September), changes in the global teleconnection patterns, especially with ENSO. Statistical models cannot be used for prediction on much smaller spatial and temporal scales. For this specific need, we have to use dynamical models. The future prospects of monsoon prediction are completely dependent on the skill of advanced dynamical models, which showed an improvement in the model skill over the recent years. Specifically, monsoon field experiments and Indian climate research programmes will improve our understanding of the monsoon variability and thus monsoon prediction.

This book is specifically designed on monsoon prediction as it discusses the current state-of-the-art of monsoon prediction, its problems and prospects. It provides a critical analysis on how good we are in monsoon prediction and what are its future prospects in the context of improvement in dynamical models and monsoon field experiments. The book contains a good compilation of useful and updated references at the end of each chapter. The book will be useful to meteorologists, especially operational forecasters and students in meteorology and anyone interested in knowing more about monsoon prediction.