AN EMPIRICAL METHOD OF ESTIMATING SOIL TEMPERATURE ON CROPPED LAND ON CANADIAN PRAIRIES
Agronomists often require quick, easy methods of estimating soil temperatures under cereal production, either to fill in missing experimental measurements or to help explain apparent discrepancies in results. Methods available in the literature allow such estimates to be made from meteorological measurements and soil physical characteristics, but these methods are often mathematically complex. In the present paper a simple empirical regression and correlation approach was used to relate soil temperatures under cereal and fallow cropping systems to air temperature, and also to soil temperature at corresponding depths under grass plots at Swift Current, Saskatchewan. Relationships for the top 22.5 cm of soil were developed for the growing season and also for the whole year. Relationships between soil and air temperature were good near the soil surface, but deteriorated with depth even though highly significant r2 values were obtained. The best relationships were obtained between soil temperatures under the cereal system and temperatures under grass (r2 > 0.8 for growing season and > 0.9 for whole year). The relationships between mean daily temperatures under cereals (y) and those under grass at corresponding depths (x) were generally represented by y = x. The best Swift Current relationships for the growing season were used successfully [Formula: see text] to predict data for different years at Swift Current and Scott, Saskatchewan and at Lethbridge, Alberta. The error in prediction at the 10-cm depth was, on the average, 1–3 °C and at the 20-cm depth, 0–4 °C. The relationship developed will be more accurate in drier regions such as the southern prairies.