The ability to distinguish spatial variability either from deterministic trends or from experimental treatment effects contributes to the accuracy and interpretation of field experiments. The objective of this study was to characterize three components of soil and soil plus vegetation CO2 efllux: the positional trend, spatial correlation, and random variation. Soil CO2 efflux was measured in a wheat (Triticum aestivum L.) field at two different dates with differing soil water contents, and soil plus vegetative CO2 efflux was measured at three grassland sites. Carbon dioxide efflux was determined by the alkali absorption method using static chambers located at 3‐m intervals within a 18 m by 18 m square grid. Positional trends were identified using a multiple regression technique, and with a t‐test extended into two dimensions. Data were detrended using a generalized least squares (GLS) regression procedure. Semivariograms were used for analysis of spatial correlation and random variability. The assumption of spatial homogeneity (first‐order stationarity) was not founded for CO2 efflux for four of the five data sets. Positional trends accounted for 16 to 48% of the total variability in these cases. Spatial correlation was not detected, although ignoring positional trends may well have resulted in the opposite conclusion. Spatial structure was affected by the soil water content under wheat. Defining the spatial structure of soil respiration requires determinations under a range of environmental conditions. Advantages of the trend identification and quantification procedures utilized are direct application of common regression techniques, direct evaluation of first‐order stationarity following trend removal and correction for correlated error structure.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.