Soil heat transfer is complex, and conduction-alone models may not always perform well in estimating soil apparent thermal diffusivity. Soil apparent thermal diffusivity is related to soil temperature change propagation rates. Soil temperature data collected at the Tazhong station in China were used to examine the characteristics of soil apparent thermal diffusivity determined by three different algorithms and the sum of vertical gradient of soil apparent thermal diffusivity and apparent water flux density . The results showed that 1) soil apparent thermal diffusivity obtained with a conduction–convection algorithm had a better agreement with soil apparent thermal diffusivity obtained with a phase algorithm than with soil apparent thermal diffusivity obtained with an amplitude algorithm except for the case of = 0; 2) when > 0, , and when < 0, ; 3) for a given soil temperature phase shift, increased (decreased) with increasing logarithmic amplitude attenuation when the phase shift was larger (smaller) than the logarithmic amplitude attenuation, reached a maximum value when the phase shift equaled the logarithmic amplitude attenuation, and increased with increasing logarithmic amplitude attenuation; and 4) for a given logarithmic amplitude attenuation, decreased with increasing phase shift and increased (decreased) with increasing phase shift when the phase shift was larger (smaller) than times the logarithmic amplitude attenuation. These mathematical conclusions were also confirmed with field data.