Drivers of carbon fluxes in Alpine tundra: a comparison of three empirical model approaches

This pre-print paper illustrate experimental work and the data analysis conducted in Gran Paradiso National Park, Italy, by researchers of the Institute of Geoscience and Earth Resources of The National Research Council of Italy in the framework of the H2020 project ECOPOTENTIAL (www.ecopotential-project.eu), regarding the CO₂ fluxes on high- altitude Alpine grasslands at Nivolet Plain (about 2700 m a.s.l.). Measurements of Net Ecosystem Exchange (NEE) and Ecosystem Respiration (ER) and other environmental variables have been conducted in three sampling sites belonging to the same catchment during 2017, 2018, 2019 summer seasons, using a portable accumulation chamber. In the paper,  we propose a comparison of three empirical modelling approaches using systematic statistical analysis, to identify the environmental variables controlling CO₂ fluxes.  Large year-to-year variations in the  gross primary production (GPP) and ecosystem respiration (ER) dependences on solar irradiance and temperature were observed,. We thus implemented a multi regression model in which additional variables were introduced as perturbations of the standard exponential and rectangular hyperbolic functions for ER and GPP, respectively.  A comparison of this model with other common modelling strategies, showed the benefits of this approach, resulting in large explained variances (83% to 94%). The optimum ensemble of variables explaining the inter- and intra-annual flux variability included solar irradiance, soil moisture and day of the year for GPP, and air temperature, soil moisture, air pressure and day of the year for the ER, in agreement with other studies. The modelling approach discussed provides a basis for selecting drivers of carbon fluxes and understanding their role in high-altitude Alpine ecosystems, also allowing for future short-range assessments of local trends