Prediction of Indian summer monsoon rainfall: a weighted multi-model ensemble to enhance probabilistic forecast skills

Abstract

India gets the maximum amount of rainfall during the months of June to September (JJAS) which is known as the summer monsoon season. The erratic nature of Indian summer monsoon rainfall (ISMR), in terms of both rainfall amount and distribution, is highly responsible for the interannual variability in agricultural production as well as occurrence of floods and droughts. Accurate seasonal predictions of ISMR are required for appropriate hydrological planning and disaster management systems. Studies have revealed that probabilistic prediction, based on the products of General Circulation Models (GCMs), can be generated in a parametric as well as non-parametric manner. The present paper discusses the enhancement of probabilistic prediction by improving the potential predictable signal obtained from these GCMs. A Singular-Value-Decomposition based multiple linear regression method (SVD-MLR) has been applied to improve the signal and a simple average of all GCMs (EM) has been used as the benchmark to examine the skill of the SVD-MLR method. The potential of the proposed method has been assessed through Brier Skill Score (BSS) and Rank Probability Skill Score (RPSS). A rigorous analysis has finally revealed that SVD-MLR method has better skill than EM in predicting the typical nature of observed monsoon rainfall in extreme years.

Forecast Customization System (FOCUS): A Multimodel Ensemble-Based Seasonal Climate Forecasting Tool for the Homogeneous Climate Zones of Myanmar

“…In this study, CFSv2 performs well over the shan region and dry zones, but GFDL predictability skills are low. Further investigation on MME schemes over the study region indicated that the AM-MME scheme is not able to enhance the overall skill of the forecast mainly because an ensemble member with higher skill gets the same weight as a member with lower skill [16]. However, the WA-MME method performs better as weights were calculated and assigned to each ensemble member.…”

“…e tool can be accessed from the following link: http://203.159. 16.146/ ForecastWeb/Login.aspx. Web data retrieval package, "wget," is used at the backend to automatically download required global forecast dataset from the respective websites.…”

“…Global climate models (GCMs) are useful tools for predicting seasonal climate; however, there is large uncertainty in its prediction, mainly because of the assumptions made in the initial atmospheric state [14]. To simulate and capture these uncertainties in the predictions, GCMs are processed with different initial conditions to generate multiple forecast members called the ensembles [15,16]. Multimodel ensemble (MME) is a process, where ensemble members of one GCM are statistically assembled with another GCM or a set of GCMs [17][18][19][20][21].…”

Rescaled range analysis and conditional probability-based probe into the intrinsic pattern of rainfall over North Mountainous India

“…Boxplot represents the degree of spread for a data set by dividing it by quartiles. In boxplots, spread for a data set is divided with quartiles (Acharya et al 2013). The boxplot consists of a 'box' which lies between the first quartile Q1 (25th percentile) and the third quartile Q3 (75th percentile).…”

Accounting for skill in trend, variability, and autocorrelation facilitates better multi-model projections: Application to the AMOC and temperature time series

“…A common forecasting problem is one of probabilistic multi-model forecasts of a stochastic dynamical system [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18]. Sometimes, when a collection of complex dynamical models is used to provide multi-model forecasts, these forecasts are weighted according to model performance compared to observations [1,5,10,[19][20][21][22][23]. The Bayesian approach to this problem assumes that associated with k dynamical models are k competing statistical models Mi for vector of observations y.…”