Reconstruction and future projection of daily flow rate of the glacial Careser stream in the southern Alps
Description
This dataset contains the reconstructed melting season time series of the daily flow rate of the Careser stream (Trento, Italy) at the Careser Baia gauging station, from May to October, for the period 1976 - 2019. The file named Flowrate_summer_DNNprediction.csv contains the measured flowrate, Q (m3 s-1), and the modeled flowrate by the Dense feed-forward Neural Network (DNN), Qpred NN(m3 s-1), used to fill the gaps in the flowrate time series. The remaining columns show the input information employed in the DNN model: the daily precipitation, P (mm), partitioned into the liquid, Pliq (mm) and solid Psol (mm) components, the mean, minimum and maximum air temperature averaged across the surface of the glacier (Tmean (°C), Tmin (°C) and Tmax (°C)), the ablation volume of the glacier for that specific year (Volume_lost (m3)); and time of the measurement (year, month and julian day). The two additional files (yearly_area_rcp45.csv and yearly_flowrate_rcp45.csv) contain the evolution of the glacier area (m2) and summer mean flow rate (m3 s-1) under the RCP 4.5 climatic scenario. The second column of these files reports the historical values of area and flow rate, respectively. The third column shows the 9-year moving average of this latter variable. The following 24 columns show future projections of area and flow rate respectively with the combinations of the eight EURO-CORDEX models considering three different models of summer ablation, corresponding to the linear regression of the measured ablation with the mean air temperature and the shift of ± the standard deviation, in order to evaluate the uncertainty of the projections. Details on the modeling approach and the discussion of the results are provided in the following paper: https://doi.org/10.1016/j.jhydrol.2023.130316