Generalized linear models for analyzing count data of rainfall occurrences

Abstract Having the adequate knowledge about behavior of climatic variables on occurrences rainfall is needed to country’s economists and agriculturists for saving people from devastating natural hazards like flash flood, drought, heavy rainfall, etc. Therefore, study has been taken initiative identify influence rainfall. The developed generalized linear models (GLMs) Poisson distribution weekly fortnightly count data daily summer monsoon seasons five regional stations Bangladesh. For these models, minimum maximum temperatures relative humidity are considered as explanatory variables. stations, model selection procedures AIC BIC indicate that GLMs satisfactorily explain in seasons. season if one unit increases, then probability rainy days will be increased by 12 percent Feni station, 6 Sylhet, Khulna Rajshahi 7 Dhaka station. Besides, increases temperature, 22 percent, 19 24 17 Feni, Khulna, respectively. Further, temperature indicates negative all period. play remarkable role changing regions country.