This paper is presented by Paul R. Moorcroft, David Medvigy and Steven C. Wofsy.

Terrestrial ecosystem models are the principal tools to assess the impacts of climate change on vegetation. Models are typically forced with meteorology from general circulation models or reanalysis products that provide representative means for solar radiation, precipitation, and temperature, but often do not correctly capture their higher-order statistics (variances and covariances). Here we use an empirically-constrained ecosystem model for forests in the Northeastern United States to show that high frequency variances and covariances of meteorological drivers have powerful ecological consequences. Monthly averaged meteorological drivers artificially enhance photosynthesis, respiration and Net Ecosystem Productivity. Conifer productivity is differentially increased, causing long-term term shifts in forest composition. We conclude that the high frequency variances and covariances of meteorological drivers cannot be neglected in climate-ecosystem studies.