Seasonal Dynamics of Sterility Mosaic of Pigeonpea and its Prediction using Statistical Models for Banaskantha Region of Gujarat, India

Abstract

Sterility Mosaic Disease (SMD) is a major biotic stress limiting achievable yield levels of pigeonpea. Studies on field incidence of SMD carried out for four consecutive kharif seasons (2012-15) indicated the commencement of its infestation during second week of August with peak incidence during third week of October to November. Mean incidence of SMD was higher in 2013 (4.5%), 2014 (4.3%), 2012 (3.8%) with the least in 2015 (0.6%). Correlation analyses of SMD incidence with weather parameters of current, one and two weeks prior indicated significant and negative influence of evening humidity at current week; significant and positive influence of sunshine of current to two lagged weeks on mean SMD. Whereas, for maximum SMD, significant negative correlation is found with minimum temperature and evening humidity at current, one week and two week lags; significant positive correlation with sunshine at current to two week lags. Besides multiple regression model, advanced statistical models namely autoregressive integrated moving average model with exogenous variable (ARIMAX), support vector regression (SVR) model and artificial neural network (ANN) have been applied for predicting the mean and maximum SMD. A comparative performance of different models carried out in terms of root mean square error (RMSE) and mean square error (MSE) indicated that both MSE and RMSE of SVR model was less in comparison to regression, ARIMAX and ANN models for forecasting the incidence of sterility mosaic disease of pigeonpea.