Abstract <p>The field observation data on soil respiration (SR), or the CO₂ emission from soil, over five years (2020–2024) in four regions of the European Russia (Chuvash Republic, Ryazan oblast, Tula oblast, and Kursk oblast) within the biome of broad-leaved forests and forest-steppes were analyzed. In each region, SR was measured in the same set of natural (forests and steppes) and agricultural (arable lands, hayfields, and pastures) ecosystems. SR was estimated with portable infrared gas analyzers and chamber technique in 10&#xa0;spatial replicates at the height of growing season. In the Kursk oblast, gas measurements were more intensive (once–twice per month) as part of year-round monitoring. The goal was to quantify the relative contributions of SR spatial and temporal variations at different scales of observations, which increases the reliability of SR field estimates for subsequent extrapolation to larger regions or for forecast. Nonparametric statistical method PERMANOVA was used. On the scale of a year or a biotope, the contribution of temporal variation (49.0–59.8%) to the total variance of SR significantly exceeds the contribution of spatial variation (5.4–9.0%). A decrease in the spatiotemporal scale of observations to long-term series and regional ecosystems reduces the contribution of temporal variation to 27% and increases the contribution of spatial variation to 23.2%. In general, temporal variation on the considered scales appears to be more important for the overall variation in SR, suggesting the methodological recommendation to increase the frequency in assessing the SR intra-annual dynamics rather than to increase the number of measurement points in local ecosystems.</p>