Comparing Q347 development and regionalization using recent data in two Swiss cantons

Abstract

In Switzerland, low flows are characterized by the discharge level corresponding to a 95% exceedance probability on a ten-year average flow duration curve, referred to as Q347. This threshold not only has significant implications for planning but also requires authorities to adjust the operation of relevant infrastructure to mitigate ecological impacts on watercourses. The value of Q347 can be determined from a flow duration curve if a discharge time series of at least ten years is available. However, for smaller catchments such time series are typically unavailable, necessitating the regionalization of Q347 values. In Switzerland, multiple linear regression has been established to estimate the area-specific discharge q347 for ungauged basins.

The primary objective of this study was to regionalize Q347 values for small, ungauged catchments (383 in the Canton of Solothurn and 9'034 in the Canton of Zürich) each with an area of less than 100 km². Daily discharge, precipitation, and temperature time series were collected for a 30-year study period from 1990 to 2020 from 56 gauged catchments in Solothurn, and similarly for a ten-year study period from 2013 to 2023 in Zürich, focusing on catchments smaller than 500 km² surrounding the target areas. A total of 30 “static” parameters delineating geometry, topography, geology, land use, and drainage along with nine “climatic” parameters describing temperatures, precipitation distributions, and potential evapotranspiration were defined and computed to characterize both gauged and ungauged catchments. The temporal variability of low flow events was then analysed for observed catchments in the two study areas. Over the past 30 years, the frequency of low flow events below the threshold has systematically increased, while the ten-year Q347 values for these catchments have decreased during the same period.

Three multiple linear regression methods were developed and implemented to be coupled with two adjustment techniques supplementing truncated discharge time series. Validation of the proposed models showed reduced errors and increased linear correlations between estimated and observed values compared to standard models. Notably, a spatially more homogeneous yet catchment-specific distribution of estimated values is observable. The proposed models yielded promising results, particularly when time series remain unadjusted, or adjustment is done using the Antecedent Precipitation Index (API) combined with the flow duration curve of a donor basin (Ridolfi et al., 2020). Using recent data for parameter selection and model fitting, especially in the Zürich study area (2013-2022), often resulted in lower Q347 values compared to standard models, reflecting an adaptation to current climatic trends.