Net ecosystem exchange comparative analysis of the relative influence of recorded variables in well monitored ecosystems

Weekly averaged datasets from fourteen AmeriFlux ecosystem monitoring sites spread across the Americas, processed to FLUXNET2015 standard, are statistically evaluated characterize their seasonal net exchange (NEE) trends. The cover wetland, cropland, woodland, grassland and tundra ecosystems. Up twenty measured variables variously correlated with NEE. A comparison of Pearson Spearman correlation coefficients reveals that behaving parametrically respect NEE for woodland (two out three sites) locations, but non-parametrically cropland sites. Multi-linear regression (MLR) analysis also distinguishes those MLR predicted versus calculated follow Y ≈ X relationships wetland sites, whereas other ecosystems results Y≠X Moreover, coefficient values optimum solutions each reveal quite distinct relative influences on values. These imply at can be relatively easily set recorded variables. On hand, types cannot variables, implying factors substantially influence