Subsoil compaction is an increasing problem in modern agriculture, but is not easily recognized in practice, also because of possible within-field spatial variations. This paper addresses the question of how within-field spatial variations in soil bulk density and other soil characteristics relate to within-field spatial variations in crop yield and potential CO2 and N2O emissions from soil. Four fields (5 to 20 ha each) were selected at the suggestion of crop farmers, and sampled using a random soil sampling design (100 samples per field). Undisturbed soil samples were taken at depth of 5–10, 30–35, and 50–55 cm and soil bulk density and potential CO2 and N2O emissions measured under controlled conditions. At each sampling point, also top soil (0–20 cm) samples were taken for determination of pH, texture, SOM, and (micro)nutrients, and soil penetration resistance measurements and visual assessments of soil structure were made. Wheat yields were recorded with harvesters equipped with GPS and yield recorders. Mean soil bulk density in the sub-soil (30–35 cm) ranged between fields from 1.36 ± 0.08 to 1.60 ± 0.11 g cm−3. Mean wheat yields ranged between fields and years from 7.6 ± 0.6 and 11.3 ± 2.4 Mg ha−1. Semi-variogram analyses showed that crop yields and soil properties were mostly spatially dependent; nugget-to-sill ratios were < 25% with ranges of 137 to 773 m. The ratio of CO2 emissions to N2O emissions was negatively related to soil bulk density, especially following N application. In conclusion, within-field spatial variations in subsoil bulk density were successfully related to spatial variations in crop yield and potential CO2 and N2O emissions. The ratio of CO2 emissions to N2O emissions had a much greater response to spatial variations in soil bulk density than wheat yield. Our study suggests that N2O emission factors may depend on (sub)soil bulk density.