Non-stationarity in US droughts and implications for water resources planning and management
Abstract
The concepts of return period and reliability are widely used in hydrology for quantifying the risk of extreme events. The conventional way of calculating return period and reliability requires the assumption of stationarity and independence of extreme events in successive years. These assumptions may not be true for droughts since a single drought event can last for more than one year. Further, droughts are known to be influenced by multi-year to multi-decadal oscillations (eg. El Nino Southern Oscillation (ENSO), Atlantic Multidecadal Oscillation (AMO), Pacific Decadal Oscillation (PDO)), which means that the underlying distribution can change with time. In this study, we develop a non-stationary frequency analysis for relating meteorological droughts in the continental US (CONUS) with physical covariates. We calculate the return period and reliability of meteorological droughts in different parts of CONUS by considering the correlation and the non-stationarity in drought events. We then compare the return period and reliability calculated assuming non-stationarity with that calculated assuming stationarity. The difference between the two estimates is used to quantify the extent of non-stationarity in droughts in different parts of CONUS. We also use the non-stationary frequency analysis model for attributing the causes of non-stationarity. Finally we will discuss the implications for water resources planning and management in the United States.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2017
- Bibcode:
- 2017AGUFM.H21F1540A
- Keywords:
-
- 1812 Drought;
- HYDROLOGY;
- 1817 Extreme events;
- HYDROLOGY;
- 1840 Hydrometeorology;
- HYDROLOGY;
- 1880 Water management;
- HYDROLOGY