Permafrost Favorability Index: Spatial Modeling in the French Alps Using a Rock Glacier Inventory

Abstract : In the present study we used the first rock glacier inventory for the entire French Alpsto model spatial permafrost distribution in the region. Climatic and topographic dataevaluated at the rock glacier locations were used as predictor variables in a GeneralizedLinearModel.Model performances are strong, suggesting that, in agreement with severalprevious studies, this methodology is able to model accurately rock glacier distribution. Amethodology to estimate model uncertainties is proposed, revealing that the subjectivityin the interpretation of rock glacier activity and contours may substantially bias the model.Themodel highlights a North-South trend in the regional pattern of permafrost distributionwhich is attributed to the climatic influences of the Atlantic and Mediterranean climates.Further analysis suggest that lower amounts of precipitation in the early winter and athinner snow cover, as typically found in the Mediterranean area, could contribute tothe existence of permafrost at higher temperatures compared to the Northern Alps. Acomparison with the Alpine Permafrost Index Map (APIM) shows no major differenceswith our model, highlighting the very good predictive power of the APIM despite itstendency to slightly overestimate permafrost extension with respect to our database. Theuse of rock glaciers as indicators of permafrost existence despite their time response toclimate change is discussed and an interpretation key is proposed in order to ensure theproper use of the model for research as well as for operational purposes.