Teknik Dergi, cilt.33, sa.2, ss.1-31, 2022 (SCI-Expanded)
This study examines the potential impacts of climate change on extreme precipitation.
Rainfall analysis with a stationary and nonstationary approach for observed and future
conditions is performed for (1950-2015 period) observed data of 5, 10, 15, 30 minutes and 1,
2, 3, 6 hours and projections (2015-2098 period) of 10, 15 minutes and 1, 6 hour for Ankara
province, Turkey. Daily projections are disaggregated to finer scales, 5 minutes storm
durations, then five minutes time series aggregated to the storm durations that are subject of
interest. Nonstationary Generalized Extreme Value (GEV) models and stationary GEV
models for observed and future data are obtained. Nonstationary model results are in general
exhibited smaller return level values with respect to stationary model results of each storm
duration for observed data-driven model results. Considering the projected data-driven model
results; on average nonstationary models produce mostly lower return levels for mid and
longer return periods for all storm durations and return periods except one-hour storm
duration. Depending on the models and Representative Concentration Pathways (RCP), there
are different results for the future extreme rainfall input; yet all results indicate a decreasing
extreme trend.