Tobias Vetter

Evaluation of sources of uncertainty in projected hydrological changes under climate change in 12 large-scale river basins

This paper aims to evaluate sources of uncertainty in projected hydrological changes under climate change in twelve large-scale river basins worldwide, considering the mean flow and the two runoff quantiles Q10 (high flow), and Q90 (low flow). First, changes in annual low flow, annual high flow and mean annual runoff were evaluated using simulation results from a multi-hydrological-model (nine hydrological models, HMs) and a multi-scenario approach (four Representative Concentration Pathways, RCPs, five CMIP5 General Circulation Models, GCMs). Then, three major sources of uncertainty (from GCMs, RCPs and HMs) were analyzed using the ANOVA method, which allows for decomposing variances and indicating the main sources of uncertainty along the GCM-RCP-HM model chain. Robust changes in at least one runoff quantile or the mean flow, meaning a high or moderate agreement of GCMs and HMs, were found for five river basins: the Lena, Tagus, Rhine, Ganges, and Mackenzie. The analysis of uncertainties showed that in general the largest share of uncertainty is related to GCMs, followed by RCPs, and the smallest to HMs. The hydrological models are the lowest contributors of uncertainty for Q10 and mean flow, but their share is more significant for Q90.

Modelling climate and land-use change impacts with SWIM: lessons learnt from multiple applications

Abstract The Soil and Water Integrated Model (SWIM) is a continuous-time semi-distributed ecohydrological model, integrating hydrological processes, vegetation, nutrients and erosion. It was developed for impact assessment at the river basin scale. SWIM is coupled to GIS and has modest data requirements. During the last decade SWIM was extensively tested in mesoscale and large catchments for hydrological processes (discharge, groundwater), nutrients, extreme events (floods and low flows), crop yield and erosion. Several modules were developed further (wetlands and snow dynamics) or introduced (glaciers, reservoirs). After validation, SWIM can be applied for impact assessment. Four exemplary studies are presented here, and several questions important to the impact modelling community are discussed. For which processes and areas can the model be used? Where are the limits in model application? How to apply the model in data-poor situations or in ungauged basins? How to use the model in basins subject to strong anthropogenic pressure? Editor D. Koutsoyiannis; Associate editor C. Perrin

Climatological drivers of changes in flood hazard in Germany

Since several destructive floods have occurred in Germany in the last decades, it is of considerable interest and relevance (e.g., when undertaking flood defense design) to take a closer look at the climatic factors driving the changes in flood hazard in Germany. Even if there also exist non-climatic factors controlling the flood hazard, the present paper demonstrates that climate change is one main driver responsible for the increasing number of floods. Increasing trends in temperature have been found to be ubiquitous in Germany, with impact on air humidity and changes in (intense) precipitation. Growing trends in flood prone circulation pattern and heavy precipitation are significant in many regions of Germany over a multi-decade interval and this can be translated into the rise of flood hazard and flood risk.

Detection of change in drought frequency in the Elbe basin: comparison of three methods

Abstract A comparison study is presented of three methods for evaluating trends in drought frequency: the standardized precipitation index (SPI), the Palmer drought severity index (PDSI), and a new method for estimation of dry spells (DS), which is based on average daily temperature and precipitation, and takes into account the length of a spell. The methods were applied to climate data from 450 stations in the Elbe River basin for the period 1951–2003, as well as data from several stations with longer observed time series. Statistical methods were used to calculate trend lines and evaluate the significance of detected trends. The dry spells estimated with the new method show significant trends in the whole lowland part of the Elbe basin during the last 53 years, and at the 10% level almost everywhere in the German part of the basin excluding mountains and the area around the river mouth. The SPI and PDSI methods also revealed significant trends, but for smaller areas in the lowland. The new DS method provides a useful supplement to other drought indices for the detection of trends in drought frequency. Furthermore, the DS method was able to detect statistically significant trends in areas where the other two methods failed to find significant trends, e.g. in the loess region in the southwest of the German part of the basin, where small insignificant changes in climate can lead to significant changes in water fluxes. This is important, because the loess region is the area within the basin having the highest crop yields. Therefore, additional research has to be done to investigate possible impacts of detected trends on water resources availability, and possible future trends in drought frequency under climate change.

Propagation of forcing and model uncertainties on to hydrological drought characteristics in a multi-model century-long experiment in large river basins

Recent climate change impact studies studies have presented conflicting results regarding the largest source of uncertainty in essential hydrological variables, especially streamflow and derived characteristics that describe the evolution of drought events. Part of the problem arises from the lack of a consistent framework to address compatible initial conditions for the impact models and a set of standardized historical and future forcings. The ISI-MIP2 project provides a good opportunity to advance our understanding of the propagation of forcing and model uncertainties on to century-long time series of drought characteristics using an ensemble of hydrological model (HM) projections across a broad range of climate scenarios and regions. To achieve this goal, we used six regional preconditioned hydrological models set up in seven large river basins: Upper-Amazon, Blue-Nile, Ganges, Upper-Niger, Upper-Mississippi, Rhine, and Upper-Yellow. These models were forced with bias-corrected outputs from five CMIP5 general circulation models (GCMs) under two extreme representative concentration pathway scenarios (i.e., RCP2.6 and RCP8.5) for the period 1971-2099. The simulated streamflow was transformed into a monthly runoff index (RI) to analyze the attributions of the GCM and HM uncertainties on to drought magnitudes and durations over time. The results indicated that GCM uncertainty mostly dominated over HM uncertainty for the projections of runoff drought characteristics, irrespective of the selected RCP and region. In general, the overall uncertainty increased with time. The uncertainty in the drought characteristics increased as the radiative forcing of the RCP increased, but the propagation of the GCM uncertainty on to a drought characteristic depended largely upon the hydro-climatic regime. Although our study emphasizes the need for multi-model ensembles for the assessment of future drought projections, the agreement between the GCM forcings was still too weak to draw conclusive recommendations.

Adjustment of global precipitation data for enhanced hydrologic modeling of tropical Andean watersheds

Global gridded precipitation is an essential driving input for hydrologic models to simulate runoff dynamics in large river basins. However, the data often fail to adequately represent precipitation variability in mountainous regions due to orographic effects and sparse and highly uncertain gauge data. Water balance simulations in tropical montane regions covered by cloud forests are especially challenging because of the additional water input from cloud water interception. The ISI-MIP2 hydrologic model ensemble encountered these problems for Andean sub-basins of the Upper Amazon Basin, where all models significantly underestimated observed runoff. In this paper, we propose simple yet plausible ways to adjust global precipitation data provided by WFDEI, the WATCH Forcing Data methodology applied to ERA-Interim reanalysis, for tropical montane watersheds. The modifications were based on plausible reasoning and freely available tropics-wide data: (i) a high-resolution climatology of the Tropical Rainfall Measuring Mission (TRMM) and (ii) the percentage of tropical montane cloud forest cover. Using the modified precipitation data, runoff predictions significantly improved for all hydrologic models considered. The precipitation adjustment methods presented here have the potential to enhance other global precipitation products for hydrologic model applications in the Upper Amazon Basin as well as in other tropical montane watersheds.

Direct aerosol effects during periods of solar dimming and brightening hidden in the regression residuals: Evidence from Potsdam measurements

A recent empirical study of Stanhill et al. [2014], which was based on the Angstrom-Prescott relationship between global radiation and sunshine duration, was evaluated. The parameters of this relationship seemed to be rather stable across the dimming and brightening periods. Thus, the authors concluded that the variation in global radiation is more influenced by changes in cloud cover and sunshine duration than by the direct aerosol effects. In our study, done for the Potsdam station (one of six globally distributed stations, the source of one of the longest observational records and closely located to former hotspots of aerosol emission), we tested and rejected the hypothesis that the dimming of global radiation directly caused by aerosols is negligible. The residuals of the Angstrom-Prescott regression reveal a statistically significant positive temporal trend and a temporal level segmentation. The latter was consistent with the temporal emission patterns around Potsdam. The trend in the residuals only disappeared when the model intercept varied according to the temporal level segmentation. The magnitude of the direct aerosol effect on the level changes in global radiation derived from the modified Angstrom-Prescott relationship was in the range indicated in previous studies. Thus from here, a specific request cannot be made for a revision of current climate models state-of-the-art representation of both the cooling effect directly caused by aerosols and the temperature sensitivity to the increase of greenhouse gases.