<p>Spatial variability of ecosystem processes constitutes significant uncertainty source in greenhouse gas flux measurements and estimations. The major disadvantage of the chamber-based flux measurements is the poor spatial representativeness, but eddy-covariance measurements also have an uncertainty due to the unequal and not constant footprint area. One way to overcome these difficulties is the spatial sampling improving the field-scale data coverage.</p><p>The aim of this study was to describe the spatial variability of grassland soil CO<sub>2</sub> efflux under varying environmental conditions. For this reason, we conducted spatial measurements on a range of variables including soil respiration, above-ground biomass, greenness index of the vegetation, soil water content and soil temperature during a seven-year study in a dry grassland site in Hungary. Altitude and soil organic carbon (SOC) content of the measuring positions were also used as background factors. Measurements were repeated 19 times at 78 positions during the study, in the main phenological stages of the grassland vegetation: spring growth, summer drought, autumn regrowth. The sampling scheme was based on 80&#215;60 m grid of 10 m resolution. SOC content was highly variable among the positions due to the exposure differences and their environmental constrains. We analyzed the effect of the drivers on soil respiration grouping the measuring positions by the SOC content of the soil.</p>