Author(s): Yogeesh KJ, Abhishek Singh, Ranjitha K and Raghuraman M

Abstract: The scope of any prediction model depends on its practical utility and economic importance of crop. The model developed for pest of rice, Brown Plant Hopper (BPH) has a role to evade the crop loss if we predict the pest before occurrence with significant accuracy. The mean to variance ratio of the BPH count for the study period was > 1 and it indicates that data was over dispersed and also The Kolmogorov - Smirnov test and Shapiro-Wilk’s test was found to be significant and indicating the non-normality of the data. Since the insect count data were discrete and dispersed fitted with general linear models like Poisson, Quasi-Poisson and Negative binomial distribution. The poison model provided good fit with minimum performance indicators like RMSE and MAPE compared to other models. The population dynamics of BPH in the study area influenced by maximum temperature (p=0.043) and relative humidity (p=0.05). Similarly the BPH was observed more in 32^nd and 44^th standard meteorological weeks of crop growth period.

Pages: 1370-1372 | Views: 158 | Downloads: 4

Download Full Article: Click Here