Analysis of large flood events: Based on flood data during 1985-2016 in Australia and India

This study uses past flood data in Australia and India in order to observe emerging patterns and trends to enhance prediction of flood hazards. The dataset comprises of 10 attributes and 348 flood instances (121 floods events in Australia and 227 floods events in India). Then two flood risk assessment metrics were used by previous research, is utilized to understand flood aspects such as: (i) severity class (associated with the frequency of flood) and (ii) magnitude (related to flood severity, flood duration and affected area of floods). The collected data from the Dartmouth Flood Observatory records is then compared and contrasted to the attributes derived and measured in raw data form. The 25th and 75th percentiles slightly show high values of flood severity class in Australian data 1-2 while in India 1-1.5. Similarly, values of magnitudes were also seen in Australian flood data reports as 5.2-6.6 while the flood Data of Indian was 5.4-6.5. Meanwhile, the highest mean values calculated were 25th-75th percentiles of low values of a number of fatalities (12.25-141.50) and a number of displaced people (0-151.5k) were shown in Indian floods whereas a number of fatalities (0-2) and a number of displaced people (0-525) were seen in Australian floods. This study has found a slight increase of trends during the 32-year period in the number of reported floods from Australia and India, which exceeds the severity class and magnitude thresholds. Henceforth, this study provides new evidence that supports the existing hypothesis that the earth is becoming more prone to natural hazards in recent times, possibly, due to climate change. Nevertheless, this study believes existing two flood risk assessment metrics (flood severity class and flood magnitude) can be improved by including annual flood increment rate for particular countries to the assessment criteria. The outcome of this analysis can be used for bias correction of flood prediction using theoretical or data driven models.