Study of water temperature changes of watersheds with different geological conditions due to climate change

The possible future impact of climate change on water temperature of streams flowing in different geologies in snowy cold regions were quantified using a physics-based model. The study area is the Kanayama Dam watershed (470 km2) located in the uppermost reaches of the Sorachi River in Hokkaido, the northernmost prefecture in Japan. It includes four subwatersheds, with areas of 16.1 to 22.1 km2, covered with forests. Two of these watersheds are composed of Pleistocene pyroclastic flow deposits, whereas the other two watersheds are composed of Mesozoic metamorphic rocks. A model was developed that incorporates an atmospheric and land surface process model that considers snow processes, a runoff model, and a water temperature estimation model. Different parameters were set for the runoff model according to the geological characteristics, and the parameters were used to calculate changes in water temperature in response to changes in the meteorological data in historical and future climate simulations. The simulation effectively reproduced observed variation in the discharge and water temperature in each watershed. Both the observed and simulation-reproduced data demonstrated that streams with newer Pleistocene pyroclastic flows, which have a greater proportion of groundwater runoff, have less variability in water temperatures than streams with older Mesozoic rocks. Finally, the characteristics of water temperature of the four watersheds in the late 21st century were calculated and analyzed using the simulation-reproduced data with downscaled future meteorological data, in accordance with the RCP 8.5 emission scenarios adopted by the IPCC AR5 report. Future water temperatures in summer were predicted to remain lower in the distribution areas of newer pyroclastic flows compared to those of older formations. Our findings would be useful to assess the impact of climate change on the stream ecosystems in snowy cold regions, and develop suitable adaptive measures for them.