Non stationary duration-frequency modelling of hydro-climatic extremes

Rainfall Intensity-Duration-Frequency (IDF) curves are frequently used for the design and management of water resources infrastructure. IDF information is generally represented on plots of the intensity as a function of the duration, and for a family of curves representing a number of return periods. The use of IDF curves started gaining popularity worldwide since the 1930's (Bernard 1932; Hershfield 1961). It is now commonly accepted that climate change is affecting rainfall and hydrologic variable characteristics (Ouarda and El Adlouni 2011; Ouarda et al. 2014). Recently a few nonstationary IDF models have been proposed in the literature (Youssef and Ouarda 2015). In this work we propose a general methodology to integrate climate variability and change in IDF curves, by conditioning the parameters of the curves on the covariates of time and climate indices. Extended periods of extreme temperature can have damaging impacts on the ecosystems, agriculture, infrastructure, water resources availability, and public health. Mean temperatures have been increasing worldwide leading to a significant increase in the frequency, magnitude and duration of severe heat waves with considerable consequences. It is also important to understand the impacts of global-scale modes of climate variability on temperature extremes. In the present work we propose nonstationary TemperatureDuration-Frequency (TDF) curves relating temperature intensities corresponding to different durations to a number of return periods. The proposed curves account for climate external forcings through the introduction of covariates that account for temporal trend and teleconnections. These TDF curves can find applications in such fields as agriculture, public health, global security, water resources management, and energy generation. For instance, nonstationary TDF curves would allow integrating information concerning the evolution of temperature extremes in the design and operation (cooling management) of nuclear power plants.