India receives 75 to 80 percent precipitation out of total 4000 km3 annual precipitation during rainy season under the influence of south-west monsoon (Kumar et al., 2005). In the arid and semi-arid regions, the yields of farmers are often badly affected due to extreme climate uncertainty. Due to its erratic nature and characteristic spatiotemporal variation, rainfall becomes the predominant key risk factor that has a direct or indirect effect on agriculture. Hence, the design and management of hydraulic structures, irrigation water supply, soil conservation planning, flood control systems, and optimal crop planning are based on rainfall depths that can be expected with a certain probability rather than the long-term average of rainfall data. There are various unpredictable sources of uncertainty concerning the physical processes that occur during hydrological events (Hosking and Wallis, 1997). The stochastic model (hydrological frequency analysis) can, however, be used as a tool to estimate how frequently a specified event will occur on average in a region based on the available data (Bhakar et al., 2006). In this method the magnitudes of events for design return periods are determined beyond the recorded range. Due to the high spatial and temporal variability of rainfall, irrigation water supply, flood control systems, and hydraulic structures are designed and managed based on rainfall depths that can be expected for a given probability.
Rainfall analysis using probability distribution models has been studied by a number of researchers. Kumar (2000) and Singh (2001) found that the Log Normal (LN) distribution is the best-fit probability distribution for annual maximum daily rainfall in India. Amin et al. (2016) found that the Log-Pearson Type-III (LP-III) distribution was the best-fit distribution to estimate annual maximum rainfall in the northern regions of Pakistan. Eslamian et al. (2007) suggested that the Generalized Extreme Value (GEV) and LP-III distributions provided the best fit to estimate maximum monthly rainfall as an extreme event in Iran. Lee (2005) and Ogunlela (2001) found that the LP-III distribution best fitted the rainfall distributions of Taiwan's Chia-Nan plain and Nigeria's Chia-Nan plain, respectively. For one to five consecutive days of maximum rainfall in Accra, Ghana, the LN-II distribution was shown to be the best-fit probability distribution (Kwaku et al., 2007). Olofintoye et al., (2009) identified that 50% of stations follow LP-III distributions and 40% follow Pearson Type-III distributions for peak daily rainfall in Nigeria. Sen et al. (1999) observed that the Gamma probability distribution provided the best fit to monthly maxima rainfall in arid regions of Libya. The US Water Resources Council (USWRC) recommended the LP-III distribution in 1967, and it was considered to be the best way of flood frequency analysis in the United States (Arora & Singh, 1989). Zalinaet al. (2002) concluded from their study on the annual maximum rainfall series in Malaysia that the GEV distribution is the best fit for analysing the annual maximum rainfall series. Hanson and Vogel (2008) reported that Pearson Type-III distribution fitted the best to the daily rainfall in the United States. According to Bhakar et al. (2008) the Gumbel distribution was the best fit for monthly maximum rainfall in India. Sharma and Singh (2010) evaluated that the LN and Gamma distributions were the best fit probability distributions for the annual and seasonal time scales, while the GEV distribution was found to be the best fit probability distribution for the weekly time scale in Pantnagar (India). The current study seeks the best-fit models for determining the frequency of extreme rainfall events as well as the maximum monthly and annual rainfall present over return periods of 3, 4, 5, 10, 15, 20, 25, 30 and 35 years in the Junagadh (Gujarat-India) region. These can be used to develop plans and policies for better management of water resources and agricultural issues. The findings would be useful for agriculturists, hydrologists, designers of hydraulic structures, irrigation engineers, environmental managers, and planners of water resources to develop better plans and policies.