Reducing Snow-Climate Uncertainty in Earth System modelling (ReSCUES)

点击翻译按钮获取中文摘要

Snow is a material with remarkable physical properties that profoundly alters the characteristics of the Earth's surface where it lies. Because snow has a high albedo (the fraction of solar radiation that it reflects rather than absorbs) and a high latent heat of fusion (the energy required to melt it), it delays the warming of the atmosphere and the ground in spring each year. Satellite measurements of Northern Hemisphere snow cover have now been available for 50 years, and a strong decreasing trend correlated with warming has been observed in spring over that period. Less snow accumulates in a warmer climate and melts sooner, increasing the absorption of solar radiation and reinforcing the warming (a strong positive feedback). Snow conducts heat poorly because it contains trapped air and so insulates the ground from cold temperatures in winter; this controls soil freezing and provides protection for short plants, small animals and soil microbes living in snowy regions, with important and complex impacts on the global carbon cycle. For all of these reasons, it is important that climate models should be able to predict snow cover accurately. Unfortunately, the latest climate models still differ greatly in their simulations of how snow cover varies from year to year in the current climate and how it will change in the future. There are many potential sources for this uncertainty, including errors in snowfall and temperature patterns predicted by models, multiple processes that control the rate of snowmelt but may be poorly represented in models, and uncertainty in setting optimal values for model parameters. It has proven very difficult to disentangle these sources of uncertainty and to determine how they can be reduced. In this project, we will use a new modelling system in which a single meteorological variable, model process or parameter value can be varied at a time, allowing highly controlled experiments to precisely determine how they influence simulations. Direct measurements of snow properties at research sites and satellite measurements of snow cover and albedo across the Northern Hemisphere will be used to identify the best simulations. Because snow melts both as the weather warms in spring and as the climate warms, improving the ability of models to simulate the current seasonal cycle and past trends can be expected to improve projections of future conditions, provided that the improvements are obtained for sound physical reasons. Improved predictions and better understanding of the sensitivity of snow to climate change will contribute to reviews of climate science by the Intergovernmental Panel on Climate Change which are essential resources for policymakers. Another important feature of snow is that it stores precipitation that falls in the mountains over winter and releases it in warmer times of year when human demand for water is higher. Many parts of the world are provided with water and threatened by floods from melting snow in upstream mountain regions. Even if the total amount of precipitation does not change in a warming climate, a shift to more falling as rain rather than snow will lead to river flows peaking earlier in the year, requiring major changes in the management of water resources. Global climate models, which cannot resolve processes occurring on scales smaller than a few hundred kilometres, are not adequate tools for informing water management decisions, but national weather services are now beginning to run forecasts for limited areas and short periods with kilometre-scale resolutions. We will use high-resolution meteorological data and the same modelling methods that we applied on the hemispheric scale to make and test predictions for snowmelt in well-instrumented areas of the French and Swiss Alps. Methods developed will be incorporated in a "downscaling toolkit" which will be made available to researchers and water managers by the International Network for Alpine Research Catchment Hydrology.