Uncertainties in ecosystem gas exchange measurements due to differences in infrared gas analyzer response

A key component for the application of eddy covariance technique, used worldwide to quantify ecosystem energy and material exchanges, are fast-response water and carbon dioxide gas analyzers. Here I present data from on-going comparisons that examine the performance of closed (LI7200, Li-Cor Inc.) and open path infrared gas analyzers (LI7500) as well as between several open path Li7500 analyzers used in conjunction with a common sonic anemometer for quantifying daily to yearly net ecosystem exchange of carbon dioxide (NEE) and evapotranspiration (ET) in semiarid landscapes. While differences between ET fluxes tend to be small, I observed significant and systematic differences between NEE values computed from data obtained with open and closed path analyzers, as well as those estimated between the same model of open path analyzer. These differences can contribute systematic error to daily and seasonal estimates of NEE and complicate our interpretation of source/sink behavior, both within one site and especially across sites. I also examine how such instrument-specific biases can be corrected such that meaningful cross-site comparisons can be made.