Publication Date:
2023-09-06
Description:
Riverine low flow affects ecological functions and restricts the water availability for energy production and irrigation. Climate change is expected to alter the intensity, duration, and seasonality of these events. However, extreme value analysis is challenging because few to no statistically independent extreme low flow events occur per season. Here, we employ the large hydrological model ensemble WaSiM-LE with 50 realizations for the period 1980–2099 under RCP8.5. We explore four head catchments in Bavaria, two having a pluvial regime and two being also influenced by snow dynamics. We apply the Generalized Extreme Value distribution to assess changes in extreme low flow conditions reflected by the mean 7-day low flow (LM7Q) and the event duration. In future summer half-years, drier LM7Q and longer event duration is projected in all catchments. The seasonality experiences a shift from mid-summer into autumn due to increasing temperature and evaporation, while decreasing precipitation and soil moisture. Despite increasing winter rainfall, the winter-LM7Q stagnates until mid-century and decreases afterwards. However, the triggers of low flow change from lack of liquid precipitation to lag effects of low summer soil moisture, which also leads to an advance in seasonality to the beginning of the winter half-year. This study demonstrates a first step towards an extreme value statistical handling of low flow events applying a large ensemble. The ensemble addresses the uncertainty due to internal climate variability affecting extreme low flow events, where the large sample size results in robust estimations of long return periods.
Language:
English
Type:
info:eu-repo/semantics/conferenceObject
Permalink