The gradient method is widely used in conjunction with Fick’s Law to estimate emissions of soil CO2. This requires accurate estimation of the soil gas diffusion coefficient (DP) and the CO2 concentration gradient, typically from measured water content and CO2 concentration. Shallow application of water via precipitation or irrigation causes a temporary reduction in the near-surface soil gas diffusion coefficient, which compromises gradient-based flux estimates when undetected by a water content sensor. Our objectives were to analyze the effects of soil surface wetting and temperature on CO2 concentration and efflux using laboratory and field measurements and to compare four widely used models for estimating DP. A laboratory test was conducted to determine the effects of water application on the soil CO2 concentration and the CO2 efflux calculated with the gradient method. The DP parameter was very sensitive to the soil water content and was most reliably calculated using a power function model in a Millville silt loam field soil, while both power function and SAPHIR models were similarly reliable in a Kidman fine sandy loam laboratory test. A single CO2 sensor at a depth of 5 cm with water content monitored at 2.5 cm provided reasonable estimates of the soil CO2 efflux validated with an automated chamber. We found that under most conditions, the CO2 concentration gradient in the soil profile is a reasonable estimator of CO2 flux when measurements of the soil water content and known porosity values are used to estimate the gas diffusion coefficient. However, shallow wetting events require improved monitoring of spatial and temporal changes near the surface or appropriate modeling of hydrodynamics there.