Assessing the impacts of climate change in Iberian mountains using the NextGEMS km-scale global climate simulations

Mountain areas play a pivotal role in the hydrological cycle of vast regions of the world, largely due to local processes such as orographic precipitation and the presence of seasonal or permanent snow cover. In the context of climate change, some of these processes are expected to be disrupted causing significant impacts to local ecosystems and nearby populations. This is particularly relevant for regions like the Iberian Peninsula, where the development of a persistent winter-spring snowpack confined to the various medium sized mountain ranges is key to offsetting water deficits during the dry summer season. Knowing the future climate of these mountains is therefore vital both for water resource management and for economic interests.

However, these mountain ranges are often characterized by medium heights and a small horizontal extent, making them very difficult to represent in most conventional coarse resolution global climate models and demanding thus the use of regional to local dynamical and statistical downscaling methods. Considering this, the new km-scale global climate simulations developed in the context of the European H2020 NextGEMS project and other similar initiatives may open up unprecedented opportunities to readily study future impacts of climate changes on these regions. These models allow the representation of local and regional processes while retaining the benefits of homogeneous global simulations. 

The present study firstly evaluates the capacity of historical km-scale simulations (1990-2019) to represent the climate of the main mountainous areas of the Iberian Peninsula, with a particular emphasis on variables impacting seasonal snow cover which are compared with different historical data sources, including local observations, reanalyses and satellite observations. We show a fairly acceptable agreement between the model climatology and regional reanalysis products specially for the annual number of days with snow cover. Regarding snow depth, however, the model shows a small positive bias in all regions except Sierra Nevada, where it has a negative bias. Following the assessment of potential model biases, the differences between the historical climatology and a 2020-2049 projection under scenario SSP3-7.0 are analysed. Among others, we show that in the projection significant decreasing trends are present in most snow cover metrics for all the considered mountain regions, even though they are more extreme at Sierra Nevada, where a significant reduction of total winter precipitation is also present.