Codes for building and comparing multi regression models of carbon dioxide fluxes in the Alpine grasslands at Nivolet Plain, Gran Paradiso National Park (IGG-CNR-CZO@NIVOLET)

Codes for building multi regression model of carbon dioxide (CO₂) fluxes measured using the flux chamber method in the high-altitude Alpine grassland located at Nivolet Plain, Gran Paradiso National Park, Italy (about 2700 m.a.s.l.). Data represent the average values obtained from four sampling sites, characterized by soils developed over carbonate rocks ('CA') (45.500212N-7.152213E), glacial deposits ('GL') (45.490167N-7.139916E), gneiss rocks ('GN') (45.490256N-7.149253E) and alluvial deposits ('AL') (45.492656 N-7.146092 E). Net Ecosystem Exchange (NEE), Ecosystem Respiration (ER) and Gross Primary Production (GPP) are reported in \(\mu\)molCO₂/m²/s. Flux data are complemented by measurements of soil temperature (Soil_Temp, Celsius degree), soil volumetric water content (VWC, %), air temperature (Air_Temp, Celsius degree), air moisture (Air_Moisture, %), solar irradiance (Radiance, W/m²) and atmospheric pressure (Pressure, hPa). Other shortcuts:day of the year, DOY (1-365); decimal time, Time_Dec=hour+minute/60 (0-24 in local time, UTC+2)

Files:

'dataset.csv' contains data.

'BuildERModel.m' and 'BuildGPPModel.m' are Matlab codes to select the best model according  to the Akaike Information Criterion. These codes provide a supervised method to build multi regression models for ER and GPP, respectively. Convergence of the nonlinear fit in 'BuildGPPModel.m' is not ensured and can be obtained by changing the initial conditions ('a1g'). Use intrinsic Matlab functions 'corrcoef' to compute the correlation between predictors before fitting.

'Comparisons.m' allows to compare couples of sites or years using the shuffling method ('shuffER.m' and 'shuffGPP.m') to assess the significance of parameter difference.

'shuffER.m' and 'shuffGPP.m' provide the probability of difference being induced by random coincidence according to the shuffling method.

See also: Magnani, M., Baneschi, I., Giamberini, M., Mosca, P., Raco, B., & Provenzale, A. (2020). Drivers of carbon fluxes in Alpine tundra: a comparison of three empirical model approaches. Science of The Total Environment, 732, 139139.