An experimental approach is described that examines the influence of landscape terrain and land use on fluxes of important greenhouse gases (CH4, N2O and CO2) in soil. The landscape is gridded into 'field' units (cells), and each cell is characterized. For example, a 500 X 500 m rolling landscape, consisting of forest and croplands, is gridded into 400 field units (25 X 25 m cells). Cell gridding and classification of the slope, elevation and land use are partially automated using photogrammetric techniques. An analytical plotter locates 400 sample points on (1:5000 scale) aerial photographs and issues X, Y ground coordinates, then determines elevation (Z), and computes slope for each cell. Through photo interpretation, each cell is classified bylandform, land use, and vegetation. These data are superimposed on soil maps. Each cell is classified by the soil characteristics that conceptually regulate the microbial processes and the gas fluxes in question (at proximal level). Geostatistics (e.g. semivariograms) allows examination of the spatial relationships of the factor(s) that control the gas fluxes in question. Based on the study's objective, appropriate cells can be statistically selected (identified) by an experimental designed (e.g. factorial). The gas fluxes and other soil and environmental factors of potential explanatory importance (e.g. NH4+, NO3-, C, N content of soil, soil moisture and temperature) should be measured simultaneously in selected cells. The number of sample measurements required within each cell depends on the methods of measurements and spatial variability of the variable in question.