Climate change alters the rainfall patterns and causes insufficient rainfalls and irregular rainfall periods in Turkey. The reduction in rainfalls decreases the water resources in Turkey, and drought seems to be an inescapable end. This paper proposes a log-logistic accelerated failure time model with frailty to determine influence of some covariates. For this aim, we use the rainfall data from 70 meteorological stations of Turkey during 2012-2013. The survival models are compared by means of this data as a risk assessment tool in time to wet season and rainfall probability estimations. The data present that the log-logistic accelerated model is more suitable than the other survival models. The log-logistic accelerated model results show that duration time of rainfall depends on the temperature and the ratio of forest area.