Active and break spells of the Indian summer monsoon

In this paper, we suggest criteria for the identification of active and break events of the Indian summer monsoon on the basis of recently derived high resolution daily gridded rainfall dataset over India (1951–2007). Active and break events are defined as periods during the peak monsoon months of July and August, in which the normalized anomaly of the rainfall over a critical area, called the monsoon core zone exceeds 1 or is less than −1.0 respectively, provided the criterion is satisfied for at least three consecutive days. We elucidate the major features of these events. We consider very briefly the relationship of the intraseasonal fluctuations between these events and the interannual variation of the summer monsoon rainfall.     

Northeast monsoon rainfall variability over south peninsular India and its teleconnections

Rainfall over south peninsular India during the northeast (NE) monsoon season (Oct–Dec) shows significant interannual variation. In the present study, we relate the northeast monsoon rainfall (NEMR) over south peninsular India with the major oscillations like El Ni?o Southern Oscillation (ENSO), Indian Ocean Dipole (IOD), and Equatorial Indian Ocean Oscillation (EQUINOO) in the Indian and Pacific Oceans. For establishing the teleconnections, sea surface temperature, outgoing long wave radiation, and circulation data have been used. The present study reveals that the positive phase of ENSO, IOD, and EQUINOO favor the NEMR to be normal or above normal over southern peninsular India. The study reveals that the variability of NEMR over south peninsula can be well explained by its relationship with positive phase of ENSO, IOD, and EQUINOO.     

Role of intra-seasonal oscillations in modulating Indian summer monsoon rainfall

The day-to-day behavior of Indian summer monsoon rainfall (IMR) is associated with a hierarchy of quasi-periods, namely 3?C7, 10?C20 and the 30?C60?days. These two periods, the 10?C20?days and the 30?C60?days have been related with the active and break cycles of the monsoon rainfall over the Indian sub-continent. The seasonal strength of Indian summer monsoon rainfall may depend on the frequency and duration of spells of break and active periods associated with the fluctuations of the above intra-seasonal oscillations (ISOs). Thus the predictability of the seasonal (June through September) mean Indian monsoon depends on the extent to which the intra-seasonal oscillations could be predicted. The primary objective of this study is to bring out the dynamic circulation features during the pre-monsoon/monsoon season associated with the extreme phases of these oscillations The intense (weak) phase of the 10?C20 (30?C60) days oscillation is associated with anti-cyclonic circulation over the Indian Ocean, easterly flow over the equatorial Pacific Ocean resembling the normal or cold phase (La Nina) of El Nino Southern Oscillation (ENSO) phenomenon, and weakening of the north Pacific Sub-tropical High. On the other hand the weak phase of 10?C20?days mode and the intense phase of 30?C60?days mode shows remarkable opposite flow patterns. The circulation features during pre-monsoon months show that there is a tendency for the flow patterns observed in pre-monsoon months to persist during the monsoon months. Hence some indications of the behavior of these modes during the monsoon season could be foreshadowed from the spring season patterns. The relationship between the intensity of these modes and some of the long-range forecasting parameters used operationally by the India Meteorological Department has also been examined.     

New Models for Long Range Forecasts of Summer Monsoon Rainfall over North West and Peninsular India

Summary New models based on (a) Multivariate Principal Component Regression (PCR) (b) Neural Network (NN) and (c) Linear Discriminant Analysis (LDA) techniques were developed for long-range forecasts of summer monsoon (June–September) rainfall over two homogeneous regions of India, viz., North West India and Peninsular India. The PCR and NN models were developed with two different data sets. One set consisted 42 years (1958–1999) of data with 8 predictors and the other, 49 years (1951–1999) of data with 6 predictors. The predictors were subjected to the Principal Component Analysis (PCA) before model development. Two different neural networks were designed with 2 and 3 hidden neurons. To avoid the nonlinear instability, 20 ensemble runs were made while training the network and the ensemble mean results are discussed. The LDA model was developed with 42 years of data (1958–1999) for classifying three rainfall intervals with equal prior probability of 0.33. Both the PCR and NN models showed useful forecast skill for NW India and Peninsular India. Models with 8 predictors performed better than the models with only 6 predictors. The NN model with 3 hidden neurons performed better than model with 2 hidden neurons. For NW India, the NN model performed better than the PCR model. The RMSE of the NN model and PCR model with 8 predictors for NW India (Peninsular India) during the independent period 1984–99 was 12.5% (12.2%) and 12.6% (11.5%), respectively. Corresponding figures for the models with 6 predictors are 15.0% (13.0%) and 13.9% (11.4%) respectively. During the independent period, model errors were large in 1991, 1994, 1997 and 1999. However all the models showed deteriorating predictive skill after 1988, both for NW India and Peninsular India. The LDA model correctly classified 62% of grouped cases for NW India and Peninsular India. The LDA model showed better skill in classifying deficient rainfall (< − 8%) over NW India and excess rainfall (> 3%) over Peninsular India. Received October 2, 1999 Revised December 28, 1999     

New statistical models for long-range forecasting of southwest monsoon rainfall over India

The India Meteorological Department (IMD) has been issuing long-range forecasts (LRF) based on statistical methods for the southwest monsoon rainfall over India (ISMR) for more than 100 years. Many statistical and dynamical models including the operational models of IMD failed to predict the recent deficient monsoon years of 2002 and 2004. In this paper, we report the improved results of new experimental statistical models developed for LRF of southwest monsoon seasonal (June–September) rainfall. These models were developed to facilitate the IMD’s present two-stage operational forecast strategy. Models based on the ensemble multiple linear regression (EMR) and projection pursuit regression (PPR) techniques were developed to forecast the ISMR. These models used new methods of predictor selection and model development. After carrying out a detailed analysis of various global climate data sets; two predictor sets, each consisting of six predictors were selected. Our model performance was evaluated for the period from 1981 to 2004 by sliding the model training period with a window length of 23 years. The new models showed better performance in their hindcast, compared to the model based on climatology. The Heidke scores for the three category forecasts during the verification period by the first stage models based on EMR and PPR methods were 0.5 and 0.44, respectively, and those of June models were 0.63 and 0.38, respectively. Root mean square error of these models during the verification period (1981–2004) varied between 4.56 and 6.75% from long period average (LPA) as against 10.0% from the LPA of the model based on climatology alone. These models were able to provide correct forecasts of the recent two deficient monsoon rainfall events (2002 and 2004). The experimental forecasts for the 2005 southwest monsoon season based on these models were also found to be accurate.     

Long Range Forecasting Indian Summer Monsoon Rainfallby a Hybrid Principal Component Neural Network Model

Summary  The existing methods based on statistical techniques for long range forecasts of Indian monsoon rainfall have shown reasonably accurate performance, for last 11 years. Because of the limitation of such statistical techniques, new techniques may have to be tried to obtain better results. In this paper, we discuss the results of an artificial neural network model by combining two different neural networks, one explaining assumed deterministic dynamics within the time series of Indian monsoon rainfall (Model I) and other using eight regional and global predictors (Model II). The model I has been developed by using the data of past 50 years (1901–50) and the data for recent period (1951–97) has been used for verification. The model II has been developed by using the 30 year (1958–87) data and the verification of this model has been carried out using the independent data of 10 year period (1988–97). In model II, instead of using eight parameters directly as inputs, we have carried out Principal Component Analysis (PCA) of the eight parameters with 30 years of data, 1958–87, and the first five principal components are included as input parameters. By combining model I and model II, a hybrid principal component neural network model (Model III) has been developed by using 30 year (1958–87) data as training period and recent 10 year period (1988–97) as verification period. Performance of the hybrid model (Model III) has been found the best among all three models developed. Rootmean square error (RMSE) of this hybrid model during the independent period (1988–97) is 4.93% as against 6.83%of the operational forecasts of the India Meteorological Department (IMD) using the 16 parameter Power Regression model. As this hybrid model is showing good results, it is now used by the IMD for experimental long-range forecasts of summer monsoon rainfall over India as a whole. Received August 20, 1998/Revised April 20, 1999     

Empirical prediction of Indian summer monsoon rainfall with different lead periods based on global SST anomalies

Summary The main objective of this study was to develop empirical models with different seasonal lead time periods for the long range prediction of seasonal (June to September) Indian summer monsoon rainfall (ISMR). For this purpose, 13 predictors having significant and stable relationships with ISMR were derived by the correlation analysis of global grid point seasonal Sea-Surface Temperature (SST) anomalies and the tendency in the SST anomalies. The time lags of the seasonal SST anomalies were varied from 1 season to 4 years behind the reference monsoon season. The basic SST data set used was the monthly NOAA Extended Reconstructed Global SST (ERSST) data at 2° × 2° spatial grid for the period 1951–2003. The time lags of the 13 predictors derived from various areas of all three tropical ocean basins (Indian, Pacific and Atlantic Oceans) varied from 1 season to 3 years. Based on these inter-correlated predictors, 3 predictor sub sets A, B and C were formed with prediction lead time periods of 0, 1 and 2 seasons, respectively, from the beginning of the monsoon season. The selected principal components (PCs) of these predictor sets were used as the input parameters for the models A, B and C, respectively. The model development period was 1955–1984. The correct model size was derived using all-possible regressions procedure and Mallow’s “Cp” statistics. Various model statistics computed for the independent period (1985–2003) showed that model B had the best prediction skill among the three models. The root mean square error (RMSE) of model B during the independent test period (6.03% of Long Period Average (LPA)) was much less than that during the development period (7.49% of LPA). The performance of model B was reasonably good during both ENSO and non-ENSO years particularly when the magnitudes of actual ISMR were large. In general, the predicted ISMR during years following the El Ni?o (La Ni?a) years were above (below) LPA as were the actual ISMR. By including an NAO related predictor (WEPR) derived from the surface pressure anomalies over West Europe as an additional input parameter into model B, the skill of the predictions were found to be substantially improved (RMSE of 4.86% of LPA).