Hydrol. Earth Syst. Sci., 22, 437–461, 2018 https://doi.org/10.5194/hess-22-437-2018 © Author(s) 2018. This work is distributed under the Creative Commons Attribution 3.0 License. Can riparian vegetation shade mitigate the expected rise in stream temperatures due to climate change during heat waves in a human-impacted pre-alpine river? Heidelinde Trimmel1, Philipp Weihs1, David Leidinger1, Herbert Formayer1, Gerda Kalny2, and Andreas Melcher3 1Institute of Meteorology, University of Natural Resources and Life Science (BOKU), Vienna, Austria 2Institute of Soil Bioengineering and Landscape Construction (IBLB), University of Natural Resources and Life Science (BOKU), Vienna, Austria 3Institute of Hydrobiology and Aquatic Ecosystem Management (IHG), University of Natural Resources and Life Science (BOKU), Vienna, Austria Correspondence: Heidelinde Trimmel ([email protected]) Received: 13 May 2016 – Discussion started: 23 May 2016 Revised: 22 November 2017 – Accepted: 26 November 2017 – Published: 18 January 2018 Abstract. Global warming has already affected European 1 Introduction rivers and their aquatic biota, and climate models predict an increase of temperature in central Europe over all sea- Stream temperature is an important factor influencing the sons. We simulated the influence of expected changes in heat physical, chemical and biological properties of rivers and wave intensity during the 21st century on water temperatures thus the habitat use of aquatic organisms (Davies-Colley and of a heavily impacted pre-alpine Austrian river and anal- Quinn, 1998; Heino et al., 2009; Magnuson et al., 1979). ysed future mitigating effects of riparian vegetation shade Heino et al. (2009) suggest that freshwater biodiversity is on radiant and turbulent energy fluxes using the determinis- highly vulnerable to climate change with extinction rates ex- tic Heat Source model. Modelled stream water temperature ceeding those of terrestrial taxa. Stream temperature is highly increased less than 1.5 ◦C within the first half of the cen- correlated with the assemblages of fish and benthic inverte- tury. Until 2100, a more significant increase of around 3 ◦C brates along the river course (Dossi et al., 2015; Melcher et in minimum, maximum and mean stream temperatures was al., 2015). The duration and magnitude of the maximum sum- predicted for a 20-year return period heat event. The result mer stream temperatures in particular are limiting factors for showed clearly that in a highly altered river system riparian the occurrence of many fish species. High temperatures may vegetation was not able to fully mitigate the predicted tem- produce high physiological demands and stress while also perature rise caused by climate change but would be able to reducing the oxygen saturation in the water column. The in- reduce water temperature by 1 to 2 ◦C. The removal of ripar- creased metabolic requirements together with the decreased ian vegetation amplified stream temperature increases. Max- oxygen availability can prove to be a limiting factor or even imum stream temperatures could increase by more than 4 ◦C be lethal in combination; the average optimum temperature even in annual heat events. Such a dramatic water tempera- for cold water species is below 16 ◦C (Matulla et al., 2007; ture shift of some degrees, especially in summer, would indi- Pletterbauer et al., 2015). cate a total shift of aquatic biodiversity. The results demon- Continuous warming of water temperatures induces strate that effective river restoration and mitigation require changes from cold water to warm water fish species assem- re-establishing riparian vegetation and emphasize the impor- blages and slow altitudinal shifts of species, if the habitat is tance of land–water interfaces and their ecological function- suitable and no migration barriers exist. River continuum dis- ing in aquatic environments. ruption and river dimension reduce the fish zone extent sig- nificantly (Matulla et al., 2007; Bloisa et al., 2013). Extreme events where lethal thresholds of stream temperature are ex- Published by Copernicus Publications on behalf of the European Geosciences Union. 438 H. Trimmel et al.: Riparian forests and heat wave stream temperature ceeded can cause a disruption of animal communities or even controlled by air temperature, vapour pressure, wind speed extinction of (cold water) species (Melcher et al., 2013; Plet- and net radiation, play an important role (Caissie et al., 2007; terbauer et al., 2015). The largest uncertainties in forecasts Garner et al., 2014; Hannah et al., 2008; Johnson, 2004). of total suitable habitat are climate uncertainty (Wenger et One of the most influential factors regulating stream tem- al., 2013). All 230 stations of the Austrian hydrographic cen- perature is riparian vegetation (Caissie, 2006; Groom et al., tral office, with different elevations, distances from source 2011; Johnson, 2004; Moore et al., 2005; Rutherford et al., and catchment areas recorded increases in stream tempera- 1997). The streamside vegetation buffer width (Clark et al., ture of 1.5 ◦C during summer (June–August) and 0.7 ◦C dur- 1999), vegetation density and average tree height all have a ing winter (December–February) between 1980 and 2011 strong influence on stream temperature (Sridhar et al., 2004). (0.48 ◦C decade−1) (BMLFUW, 2011). This change is not Vegetation affects the sky view of the river and thereby short- likely to be due to natural climatic cycles but is part of a wave (Holzapfel et al., 2013) and longwave radiation flux, long-term trend caused by anthropogenic changes in the at- evaporation and convection heat flux, which are highly cor- mosphere (APCC, 2014). related to the openness of the sky. The reduction of shortwave Air temperatures have been rising and are expected to con- radiation can contribute significantly to reducing the heating tinue to rise globally within the next century (IPCC, 2013). In of rivers during warmer summers (Sinokrot and Stefan, 1993; eastern Austria, mean air temperature has risen by 2 ◦C since Parker and Krenkel, 1969; Rutherford et al., 1997). 1880, which is more than double the 0.85 ◦C rise recorded There are different approaches to predicting stream tem- globally (Auer et al., 2014). A further temperature increase perature. Water temperature can be predicted using statistical within the 21st century is very likely (APCC, 2014). If emis- functions (stochastic models) and its correlation (regression sion scenario A1B is assumed, mean air temperature in- models) to known variables (e.g. air temperature, water tem- creases of 3.5 ◦C over the level of the reference period 1961– perature of the previous days or streamflow). Use of air tem- 1990 by the end of the 21st century are expected in Austria perature as a surrogate for future water temperature can lead (APCC, 2014; Gobiet et al., 2014). to errors when linear (Erickson and Stefan, 2000; Webb and Temperature extremes have changed markedly and ex- Nobilis, 1997) or non-linear (Mohseni et al., 1998) regres- treme high temperature events, i.e. heat waves, are very likely sion models are applied (Arismendi et al., 2014). Stochas- to increase in the 21st century (APCC, 2014). Soil temper- tic models used to determine the long-term annual compo- ature is also expected to increase due to climate change and nent of temperatures and their short-term residuals separately will influence stream temperatures via substrate heat conduc- yield good results (Caissie et al., 2001). Including a discharge tion and groundwater flux (Kurylyk et al., 2015). For exam- term in the regression model can improve the model’s perfor- ple, in Austria, near-surface groundwater body temperature mance during heat wave and drought (low flow) conditions, is expected to rise by 0.5 to 1 ◦C on average by 2050 (BML- when water temperatures are most sensitive to air tempera- FUW, 2011). Austria lies between two zones of opposing ture (van Vliet et al., 2011). Energy balance models resolving precipitation trends (IPCC, 2013). Northern Europe shows an all energy fluxes affecting a river system are the best suited to increasing trend, while the Mediterranean has a decreasing predict stream temperature (Caissie et al., 2007) but demand trend (Böhm, 2006). In southeastern Austria, a precipitation the most input data. Only these models are able to simulate decrease of about 10–15 % has been recorded over the last energy flux changes caused by increased or decreased river 150 years (APCC, 2014; Böhm, 2012). Low flow discharge shade. rates of rivers are likely to decrease by 10 to 15 % by 2021– Though the influence of vegetation on water temperature is 2050 compared to 1976–2007 during all seasons (Nachtnebel evident, its ability to mitigate climate change is not yet suffi- et al., 2014; Mader et al., 1996; APCC, 2014). ciently understood. Latent and sensible heat fluxes as well as For the study region during summer heat waves, nei- longwave radiation balance are non-linearly dependent on air ther changes in groundwater nor snowmelt contributions are temperature. It is not obvious whether the same level of shade expected (APCC, 2014). Heavy and extreme precipitation will always lead to the same rate of heat reduction. Shading shows no clear increasing signal on average, but it is likely caused by tall but less dense trees may allow exchange of air, to increase from October to March (APCC, 2014). No clear while lower riparian vegetation may cause the same level of trend of increasing wind speed (Matulla et al., 2008; Benis- shade but would reduce air movement. Vegetation can reduce ton, 2007) or increase in sunshine hours (Ahrens et al., 2014) warming but may also reduce nightly cooling by altering the has been detected but changes in the climate system may also energy fluxes on a local scale, which can only be modelled include changes in those parameters (APCC, 2014). using deterministic methods. Stream temperature is controlled by advection of heat, dis- The conclusion may be drawn that many studies have persion and the net energy fluxes acting on the surface and already addressed the influence of riparian vegetation on river bed.