Future shifts in extreme flow regimes in Alpine regions

Extreme low and high flows can have negative economic, social, and ecological effects and are expected to become more severe in many regions due to climate change. Besides low and high flows, the whole flow regime, i.e. annual hydrograph comprised of monthly mean flows, is subject to changes. Knowledge on future changes in flow regimes is important since regimes contain information on both extremes and conditions prior to the dry and wet season. Changes in individual lowand high-flow characteristics as well as flow regimes under mean conditions have been thoroughly studied. In contrast, little 5 is known about changes in extreme flow regimes. We here propose two methods for the estimation of extreme flow regimes and apply them to simulated discharge time series for future climate conditions in Switzerland. The first method relies on frequency analysis performed on annual flow duration curves. The second approach performs frequency analysis on the discharge sums of a large set of stochastically generated annual hydrographs. Both approaches were found to produce similar 100-year regime estimates when applied to a data set of 19 hydrological regions in Switzerland. Our results show that changes 10 in both extreme lowand high-flow regimes for rainfall-dominated regions are distinct from those in melt-dominated regions. In rainfall-dominated regions, the minimum discharge of low-flow regimes decreases by up to 50%, whilst the reduction is of 25% for high-flow regimes. In contrast, the maximum discharge of lowand high-flow regimes increases by up to 50%. In melt-dominated regions, the changes point into the other direction than those in rainfall-dominated regions. The minimum and maximum discharge of extreme regimes increase by up to 100% and decrease by less than 50%, respectively. Our findings pro15 vide guidance in water resources planning and management and the extreme regime estimates are a valuable basis for climate impact studies.

estimating extreme high-and low-flow regimes with a defined return period for current and future climate conditions. We propose two possible approaches for the estimation of such extreme regimes. The first approach is based on flow duration curves (FDCs). FDCs describe the whole distribution of discharge and are particularly suited for planning purposes (Vogel and Fennessey, 1994;Claps and Fiorentino, 1997). It has been shown that frequency analysis performed on annual FDCs allows for the estimation of extreme FDCs with pre-defined return periods (Castellarin et al., 2004;Iacobellis, 2008). While such estimates 5 contain information about the frequency of occurrence and the distribution of flow, they lack information on the seasonality of flow (Vogel and Fennessey, 1994). FDC estimates derived for a certain return period T therefore need to be recombined with a specific seasonality, e.g. the long-term one. This first estimation approach treats distribution and seasonality separately.
To overcome this problem, an alternative approach based on stochastically generated time series is proposed. Stochastically generated time series have been used in a number of water resources studies, including hydrologic design and drought planning 10 (Koutsoyiannis, 2000). Stochastic approaches generate large sets of realizations of possible discharge time series, thus sampling hydrologic variability beyond the historical record (Herman et al., 2016;Tsoukalas et al., 2018), potentially including extreme events and regimes. In hydrology, stochastic models have been developed so as to reproduce key statistical features of observed data, including the distribution and the temporal dependence (Sharma et al., 1997;Salas and Lee, 2010;Tsoukalas et al., 2018).
Many different approaches have been proposed for the stochastic simulation of streamflow time series. Often, indirect ap- 15 proaches, which combine the stochastic simulation of rainfall with hydrological models, have been used for the generation of stochastic discharge time series (Pender et al., 2015). These approaches are affected by uncertainties due to hydrological model selection and calibration, which can be avoided by using direct synthetic streamflow generation approaches (Herman et al., 2016). Direct approaches stochastically simulate discharge. The simplest type of models to describe daily streamflow are autoregressive moving average (ARMA) models (Pender et al., 2015;Tsoukalas et al., 2018). However, this type of models 20 only captures short-range dependence (Koutsoyiannis, 2000). Models also capturing long-range dependence include fractional Gaussian noise models (Mandelbrot, 1965), fast fractional Gaussian noise models (Mandelbrot, 1971), broken line models (Mejia et al., 1972), and fractional autoregressive integrated moving average models (Hosking, 1984). An alternative to these time-domain models are frequency domain models (Shumway and Stoffer, 2017). These latter use phase randomization to simulate surrogate data with the same Fourier spectra as the raw data (Theiler et al., 1992;Radziejewski et al., 2000). Despite 25 their favorable characteristics, such methods based on the Fourier transform have been rarely applied in hydrology (Fleming et al., 2002). We apply the approach of phase randomization to simulate stochastic discharge time series using the approach proposed by Brunner et al. (2019a). As opposed to classical phase randomization approaches, this approach does not rely on the empirical distribution but uses the flexible, four-parameter kappa distribution (Hosking, 1994), which allows for the generation of a wide range of realizations of high and low discharge values. Among these simulated series, extreme regimes 30 can be identified. After having identified a suitable approach for the estimation of extreme regimes, we apply this approach to discharge time series representing future climate conditions. A comparison to current estimates allows us to identify future changes in extreme high-and low-flow regimes.

Study area
The analyses were performed on a set of 19 hydrological regions in Switzerland ( Figure 1) with areas between 600 and 5000 km 2 , mean elevations between 550 and 2300 m a.s.l., and mean annual precipitation sums between 1000 and 1800 mm. The flow regimes north of the Alps (Plateau and Jura) are dominated by rainfall and characterized by high discharge in winter and 5 spring but low discharge in summer. In contrast, the regimes in the Alps are dominated by snow and ice melt and characterized by high discharge in summer. For illustration purposes, we chose four regions. Two of them (Jura and Thur) have a rainfalldominated regime and the other two (Valais and Engadin) a melt-dominated regime under the current climate.

Analysis framework
The analysis performed to detect changes in future extreme regimes consisted of three main steps ( Figure 2). First, different The two estimation techniques applied use frequency analysis on different quantities. The first method applies frequency analysis to the individual percentiles of the FDC. The second method uses stochastically simulated discharge time series to identify annual hydrographs with a certain non-exceedance probability. We refer to these methods as FDC and stochastic, respectively. The two methods are compared to a benchmark method (univariate), which performs univariate frequency analysis on the monthly discharge values and neglects the dependence between individual months. We here focus on the estimation of 5 high-and low-flow regimes with a return period of T = 100 years since this return period is commonly used for planning purposes. The methods outlined in this study, however, can be generalized to other return periods. In the following paragraphs, we describe the data sets (Figure 2 A, Sect. 2.3), the stochastic discharge generation procedure (Figure 2 B Figure 2. Illustration of study framework: 1) Comparison of different estimation techniques univariate, FDC, and stochastic, 2) estimation of current and future mean and extreme regimes using simulated discharge time series, and 3) comparison of current and future regime estimates.
The manuscript A) introduces the simulated data used, B) outlines the stochastic discharge generator, and C) describes the estimation approaches.

Hydrological simulations 10
We used discharge time series simulated with the hydrological model PREVAH (Viviroli et al., 2009b) as input for the analysis.
To represent current conditions, the model was driven with observed meteorological data for the period 1981-2010. To represent future conditions, it was driven with meteorological data obtained by regional climate model simulations for the period 2071-2100 (see below). PREVAH is a conceptual process-based model. It consists of several sub-models representing different parts of the hydrological cycle: interception storage, soil water storage and depletion by evapotranspiration, groundwater, snow accumulation and snow-and glacier melt, runoff and baseflow generation, plus discharge concentration and flow routing (Viviroli et al., 2009b). A gridded version of the model at a spatial resolution of 200 m was set up for Switzerland (Speich et al., 2015). For the calibration of the model parameters, meteorological and discharge time series from 140 mesoscale catchments 5 covering different runoff regimes were used. The model calibration was conducted over the period 1993-1997. Validation on discharge was performed with the period 1983-2005. More details on the calibration and validation procedures can be found in Köplin et al. (2010). The parameters for each model grid cell were derived by regionalizing the parameters obtained for the 140 catchments with ordinary kriging (Viviroli et al., 2009a;Köplin et al., 2010). The hydrological model has been calibrated using observed meteorological data but will subsequently be fed with meteorological data simulated by a set of GCM-RCM 10 combinations. It is assumed that the parameterset derived in the calibration procedure will still produce reliable results since Krysanova et al. (2018) have confirmed in a review that a good performance of hydrological models in the historical period increases confidence in projected impacts under climate change. Future glacier extents were simulated with two glacier evolution models. We used the global glacier evolution model (GloGEM; Huss and Hock, 2015) for short glaciers (glacier length < 1 km) and GloGEMflow (Zekollari et al., 2019) for long glaciers (length > 1 km). GloGEM simulates glacier changes with PREVAH is driven by time series of precipitation, temperature, relative humidity, shortwave radiation, and wind speed.
The meteorological forcing for current simulations was observed time series provided by the Federal Office of Meteorology and Climatology MeteoSwiss (2018) while the transient meteorological forcing for future simulations was derived from the CH2018 climate scenarios (National Centre for Climate Services, 2018). The meteorological data were interpolated to a 2 × 2 25 km grid using detrended inverse distance weighting where the detrending was based on a regression between climate variables and elevation (Viviroli et al., 2009b). The climate scenarios are based on the results from the EURO-CORDEX initiative (Jacob et al., 2014;Kotlarski et al., 2014), which are the most sophisticated and high-resolution coordinated climate simulations over  (Table A1 in Appendix), which provide temperature, precipitation, relative humidity, shortwave radiation, and wind speed, for the locations of various meteorological stations. The selection of scenarios included the three RCPs 2.6, 4.5, and 8.5 for which 8, 13, and 18 GCM-RCM combinations were available respectively. 10 out of the 39 GCM-RCM 35 combinations were available at a high resolution of 12.5 km and the remaining combinations at a resolution of roughly 50 km.
Using combinations at both resolutions allows for a larger ensemble, however, it means that those GCM-RCM combinations, which are available for both resolutions obtain more weight. During a model run, PREVAH reads the meteorological grids and further downscales the data to the computational grid of 200 × 200 m using bilinear interpolation. For temperature, a lapse rate of -0.65 • C/100m was additionally used for topographic corrections.

Stochastic simulation of discharge time series
The discharge simulated with the hydrological model for the current (1981-2010) and future (2071-2100) 30-years periods only represent small sets of possible annual hydrograph realizations. Among these realizations, certain hydrographs including extreme hydrographs such as a 100-year hydrograph were possibly not observed. We used a stochastic discharge simulation procedure to increase the number of possible annual hydrograph realizations. These realizations represent the discharge 10 statistics and temporal correlation structure of the available data, and extend the existing sample to yet unobserved annual hydrographs. To simulate such hydrographs, we used the method of phase randomization (Theiler et al., 1992;Schreiber and Schmitz, 2000). We combined this empirical procedure with the flexible four-parameter kappa distribution (Hosking, 1994) to allow for the extrapolation to yet unobserved values. This phase randomization approach preserves the autocorrelation structure of the raw series by conserving its power spectrum (Theiler et al., 1992). The procedure consists of three main steps 15 (Radziejewski et al., 2000). In a first step, the discharge series (here, the simulated discharge for past and future conditions) is converted from the time domain to the spectral domain by the Fourier transform (Morrison, 1994). The Fourier transform of a given time series where t is the time step, ω are the phases , and i = √ −1 is the imaginary unit. In this spectral domain, the data are represented 20 by the phase angle and by the amplitudes of the power spectrum as represented by the periodogram. The phase angle of the power spectrum is uniformly distributed over the range −π to π. In a second step, the phases in the phase spectrum are randomized while the power spectrum is preserved. In a third step, the inverse Fourier transform is applied to transform the data from the spectral domain back to the temporal domain. A step by step description of the stochastic simulation procedure and more background information on the Fourier transform are provided in Brunner et al. (2019a) and references therein. 25 An application of the simulation procedure to four example catchments in Switzerland has shown that both seasonal statistics and temporal correlation structures of discharge can be well reproduced (Brunner et al., 2019a). We therefore used this method to stochastically simulate 1500 years of discharge for each of the 19 regions in our data set. Stochastic series representing current conditions were generated by using the hydrological model simulations for 1981-2010 obtained by the 39 GCM-RCM combinations as input. Stochastic series representing future conditions were generated based on each of the hydrological model 30 simulations generated with the 39 GCM-RCM combinations for different scenarios.

Estimation of T -year hydrographs
We employed two methods for estimating 100-year low-and high-flow regimes: FDC and stochastic. The extreme regime estimates were compared to the stochastically generated hydrographs to check for plausibility. Furthermore, they were compared to a lower-bound (for low-flow regimes) or upper-bound (for high-flow regimes) benchmark regime derived by combining 100- year monthly discharge estimates obtained from univariate frequency analysis. This frequency analysis was performed on the 5 values of each month independently and the monthly values were fitted with a Generalized Extreme Value (GEV) distribution.
This distribution was not rejected according to the Anderson-Darling goodness-of-fit test computed using the procedure proposed by Chen and Balakrishnan (1995) (α = 0.05). The disadvantage of the univariate procedure is that the autocorrelation in the data, which is mainly visible for lags of 1 and 2 months, is neglected, which overestimates the extremeness of the 100-year low-flow regime and therefore produces unrealistic estimates. The univariate approach will therefore only be considered as a 10 benchmark for model comparison and not find consideration in the comparison of current and future extreme regime estimates.

FDC
A first extreme regime estimate was derived by performing the frequency analysis on annual FDCs. According to Vogel and Fennessey (1994), an annual FDC with an assigned return period can be obtained from the p th quantile function. To do so, we fitted a GEV distribution to the quantiles corresponding to each percentile. The GEV was not rejected based on the term (mean) regime of the daily input time series and varied for current and future conditions. The estimated extreme discharge regimes were aggregated to a monthly resolution to make them comparable to the univariate estimates.

Stochastic
The second method for the estimation of extreme regimes performs the frequency analysis directly on a large set of stochastically simulated annual hydrographs (here 1500 years). The frequency analysis was performed on the annual sums of the 25 stochastically generated hydrographs. We identified the hydrograph corresponding to the empirical 100-year annual discharge sum as the 100-year regime. The application of this procedure is only possible for long time series as given by the stochastic series, since a 100-year annual sum is not necessarily observed in a short record of, say, 30 years. Like the FDC estimates, the regimes derived from the stochastic approach were aggregated to a monthly resolution.

Comparison of current and future regime estimates
The two methods and the benchmark approach for the estimation of 100-year low-and high-flow regime estimates were applied to discharge time series representing current and future climate conditions. First, 100-year regimes were estimated for current conditions . For generating a control regime, we used the discharge simulated with the observed meteorological data. For representing uncertainty due to different GCM-RCM combinations for different scenarios, we derived one reference 5 regime for each discharge time series simulated by the 39 climate GCM-RCM combinations for different scenarios. This analysis provided us with a range of current regime estimates due to climate model uncertainty. The regime estimates derived from the 39 GCM-RCM combinations were used to derive a multi-model mean, which served as a reference for determining changes between current and future conditions. In a second step, 100-year estimates were derived for future conditions using 3 Results

15
The two estimation techniques and the benchmark approach provide distinct estimates for the 100-year low-flow regimes ( Figure 4). The univariate technique leads to the most extreme regimes whilst the FDC and stochastic methods lead to similar estimates. The univariate estimate should only be seen as a lower benchmark and not as an estimate for a "true" 100-year regime since the univariate approach neglects the dependence between monthly estimates. In contrast, the FDC and stochastic approaches produce more plausible estimates, i.e., estimates at the lower bound of the observed values. The summer low-flow 20 regimes estimated by the FDC technique are comparable to the regimes of the year 2003, which included a very dry summer (Beniston, 2004;Rebetez et al., 2006;Schär et al., 2004;Zappa and Kan, 2007).   The stochastic approach generally leads to more conservative estimates than the FDC approach in melt-dominated regions.
We attribute this to the fact that the stochastic approach performs frequency analysis on annual sums while the FDC approach performs frequency analysis on the percentiles of the FDC. The plausibility of the 100-year estimates derived by using the FDC and stochastic approaches is shown by a comparison with stochastically generated annual hydrographs ( Figure A1 in the Appendix for the low-flow estimates). The derived estimates, in fact, are embedded in the lower spectrum of the stochastically generated annual hydrographs. This is hardly the case for the univariate estimates, which lead to "unrealistically low" 100-year hydrographs partly outside of the range of the stochastically generated hydrographs. Similarly, the 100-year high-flow regime estimates derived by the FDC and stochastic methods are 5 embedded in the higher spectrum of the stochastically generated hydrographs while the univariate estimate is "unrealistically high". Since the univariate approach yields unrealistic estimates, it is not considered for further analysis.
3.2 Current and future low-flow regime estimates

Control simulation low and high flow
Reference simulations low and high flow Both mean and extreme regimes are subject to uncertainty when derived from simulated discharge. The uncertainty comes from the hydrological model and from the spread between the climate simulations. Figure 6 shows mean and extreme low-and high-flow regime estimates derived for the observed climatology for the four illustration catchments. It also shows the range of regimes obtained by using different GCM-RCM combinations and scenarios. This range of regimes generally encompasses 5 the regime derived from meteorological observations, which suggests that the climate model output realistically reproduces the observed climate. An exception is the Engadin, where the low-flow regimes derived from the GCM-RCM combinations overestimate summer low flows. This overestimation might be related to the univariate bias correction applied, which might not perfectly reflect the interplay between temperature and precipitation and therefore the timing of snowmelt processes (Meyer et al., 2019). The spread in the current regimes is larger for extreme than for mean conditions for the rainfall-dominated catchments Thur and Jura. In addition, the spread is larger for the high-than the low-flow extreme regimes except for the region Engadin. This range should be kept in mind when analyzing future regime estimates.  For the rainfall-dominated regions, changes in mean and extreme regimes are most visible for RCP 8.5. Here, the different 5 realizations lead to regimes with more pronounced summer low flows. In addition, there is a reduction in spring discharge under RCP 2.6 for both mean and extreme conditions when looking at the regimes derived from the FDC approach. In the case of melt-dominated regions, most GCM-RCM combinations lead to clear shifts towards regimes with earlier and reduced summer flows. These shifts are more pronounced for RCP 8.5 than RCPs 4.5 and 2.6. Note that the spread of future regimes is smaller for RCP 2.6 than RCPs 4.5 and 8.5 due to the smaller number of chains in the ensemble.

10
Differences between current and future mean and extreme low-flow regimes are summarized in Figure 8. The detected changes for RCP 2.6 and RCP 8.5 are similar (results for RCP 4.5 are not displayed but lie in between those of RCPs 2.6 and 8.5). Changes are projected for the minimum and maximum discharge of mean and extreme low-flow regimes and for their timing, but less for the mean of these regimes. The changes of the mean flow can reach up to 30%, while the maximum and minimum flows can change up to 100%. 5 Changes in melt-and rainfall-dominated regions are clearly different. Both the FDC and stochastic approach suggest changes in extreme low-flow regimes. In rainfall-dominated regions, an increase is expected for the discharge maximum independent of the estimation approach chosen. In contrast, a decrease is expected in the discharge minimum according to the stochastic approach while not clear changes are expected using the FDC approach. For melt-dominated regions, the change pattern is different. There, a decrease in maximum discharge is expected. An increase in minimum discharge is expected for mean

Current and future high-flow regime estimates
High-flow regime estimates are also expected to change (Figure 9) with no consistent change pattern visible at the first glance.
Changes in high-flow extreme regimes are slightly more pronounced for RCP 8.5 than for RCP 2.6 ( Figure 10; RCP 4.5 not shown because it is expected to provide results somewhere in betwen RCP 2.6 and 8.5). They are similar for the estimation techniques used (FDC/stochastic). As for the low-flow regimes, only moderate and mostly positive changes of less than 30% are expected in the mean discharge of extreme high-flow regimes. The changes in the maximum and minimum discharge of the high-flow regimes are much stronger, i.e up to 100%. In rainfall-dominated regions, changes in maximum discharge are mostly positive while they can be negative for melt-dominated regions. In these melt-dominated regions, an increase is expected in the minimum discharge of high-flow extreme regimes especially when using the FDC approach. In rainfall-dominated regions, changes in minimum discharge are mostly negative, especially for RCP 8.5. Changes in timing are different for the FDC and 5 stochastic approach and there is no consistent pattern across catchments. Minimum and maximum discharge can occur earlier or later in the year than under current conditions.

Estimation methods
The low-flow regime estimates derived with the univariate method are implausible because the method neglects the interdependence between flows of adjacent months. In contrast, both other methods, FDC and stochastic, lead to similar results. The differences between the two methods mainly lie in how the seasonality is derived. In the case of the FDC approach, mean 5 seasonality is used. In the case of the stochastic approach, a rather "random" seasonality is used since the regime is chosen according to the annual discharge sum. The use of one potential realization of seasonality in the stochastic approach compared to the use of a mean seasonality in the FDC approach has the disadvantage that it is less representative but the advantage that it is consistent with the corresponding annual discharge sum. The direction of changes derived from the two estimates are similar except for changes of minimum discharge in the low flow regime and minimum discharge in the high flow regimes.
Both types of estimates seem to be plausible in the light of the stochastically generated hydrographs, which represent a large set of possible realizations among which extreme hydrographs can be found. While the estimates derived by the two methods do not differ much, both methods have their advantages and disadvantages. The FDC approach is relatively simple to implement, but decouples seasonality from the distribution of daily discharge values. In contrast, the stochastic approach jointly considers magnitude and seasonality, but requires the implementation of a stochastic discharge generator. The main advantage of such 15 a generator is that the individual hydrograph realizations can be used for specific impact studies, which allows for directly performing the frequency analysis on the quantity of interest. There are several possible solutions to the multivariate problem of estimating extreme regimes, and none of these two methods can therefore be said to be the better one.
The estimation of extremes, be it of regimes or individual flow characteristics, is associated with several sources of uncertainty. These comprise the choice of an extreme value distribution used to fit the data (i.e. percentiles of FDCs, annual sums, 20 daily discharge sums) and the estimation of its parameters (Merz and Thieken, 2005;Brunner et al., 2018a). When applied to time series representing future conditions simulated with a hydrological model, additional uncertainty sources are involved.
These include the assumptions underlying the applied future global climate scenarios, global climate model structures, initial conditions, downscaling methods, modelled future glacier extents, the uncertainties inherent in the hydrological model results, and the calibration of its parameters (Wilby et al., 2008;Addor et al., 2014;Clark et al., 2016). Despite these uncertainties, 25 the extreme regime estimates can be used to identify future changes and as such these estimates can be further used in climate impact studies. Potential fields of application include water scarcity assessments, where such regime estimates are combined with estimates on water demand (Brunner et al., 2018b), eco-hydrological studies (Wood et al., 2008), or analyses on the future potential of hydropower production (Schaefli et al., 2019).

Changes in future regime estimates 30
Changes in all types of regimes (mean/extreme low flow/extreme high flow), were found to be distinct for melt-dominated and rainfall-dominated regions. This refers not only to the entire regime, but also to individual regime characteristics such as minimum, maximum, and mean flow as well as the timing of the minimum flow. The direction of change was different in rainfall-and melt-dominated regions for all regime types. An increase of up to 50% in the maximum discharge of mean and extreme low-and extreme high-flow regimes was found for rainfall-dominated regions. In contrast, a decrease of the minimum discharge by up to 100% is projected to occur for these catchments and all types of regimes. The opposite is true for meltdominated regimes, where the minimum discharge increases while the maximum and mean discharge decrease. The changes in extreme regimes can be explained by a reduction or an earlier contribution of snow-and glacier melt (Hanzer et al., 2018), and 5 by an increase in winter precipitation (Jenicek et al., 2018), which coincide with the high-flow season in rainfall-dominated regions but with the low-flow season in melt-dominated regions. For mean regimes, changes in melt-dominated regimes were found in previous studies (Barnett et al., 2005;Jenicek et al., 2018;Fatichi et al., 2014;Hanzer et al., 2018). Fatichi et al. were found by Jenicek et al. (2018) for several catchments in Switzerland and by Barnett et al. (2005) on a global scale. We 15 have shown here that these previous findings also apply to extreme regimes. The regime shifts detected have implications for various sectors. Regime shifts and more severe low flows were found to lead to more severe water scarcity situations, where water supply is insufficient to meet water demand (Brunner et al., 2019b). In the hydropower sector, future regime shifts are anticipated to lead to a reduction in production (Finger et al., 2012;Schaefli et al., 2019).

20
Extreme regime estimates were derived by frequency analysis performed on 1) annual flow duration curves (FDCs) and 2) on the discharge sums of stochastically generated annual hydrographs. Both were found to provide realistic, similar results. A range of future extreme regime estimates was obtained both for extreme and for mean conditions. In rainfall-dominated regions, the range of these future low-and high-flow estimates comprised the current estimate. On the contrary, in melt dominated regions, future high-flow and especially low-flow regimes were distinct from the current estimate. Changes in mean discharges 25 were moderate for all types of regimes and catchments, and did not exceed 30%. Projected changes in the minimum discharge of mean and extreme high-and low-flow regimes were positive in melt-dominated regions due to increases in winter precipitation and amount to up to a 100%. In contrast, mostly positive changes, of up to 50% in maximum discharge were found in rainfalldominated regions for all types of regimes. These positive changes in maximum discharge are linked to increases in winter precipitation, which coincide with the high-flow season. High-and low-flow regime estimates derived using the approaches 30 proposed in this study are important for climate impact studies addressing e.g. the future hydropower production potential or the occurrence of water shortage situations. The estimates also provide guidance for hydraulic design, emergency planning and drought and water management.