Visual and statistical inference of hourly and sub-hourly extreme rainfall trends Central Anatolia, Turkey case

Abstract

The analysis of long-term rainfall data in a changing climate is important because it has many sectoral applications such as agriculture, infrastructure, and water resources management. Statistical analyses of the annual maximum rainfall data were performed using the Mann–Kendall (MK) trend test to evaluate the annual maximum trend characteristics of the rainfall time series, the innovative trend analysis (ITA) method to detect categorial trends, and the ITA indicator to digitize the ITA results. Storm durations of 5, 10, 15, 30 min and 1, 3, 6, 24 h annual maximum rainfall series at 13 central stations in Central Anatolia, Turkey were used. According to the MK test results, there were no significant upward or downward monotonic trend at four stations, whereas the remaining nine stations showed a significant upward or downward monotonic trend. Significant negative and positive trends were identified for the sub-hourly and hourly rainfall, whereas significant positive trends were detected for hourly storm durations. Significant trend results were mostly consistent with the general ITA results. The sub-hourly storm duration data were more consistent in terms of significant trends. Conversely, when evaluated according to low, medium, and high data values in the rainfall series (categories), the high data values showed different trends. Although no trend was detected with the MK test, the ITA results showed an upward or downward trend for 25 rainfall series. 29 of 30 significant MK test results were consistent with the ITA indicator results, compared with 24 of 30 results of the visually inspected ITA results.