Relationship of soil phosphorus fractions to phosphorus soil tests and fertilizer response

Soil tests for available P may not be accurate because they do not measure the appropriate P fraction in soil. A sequential extraction technique (modified Hedley method) was used to determine if soil test P methods were accurately assessing available pools and if predictions of fertilizer response could be improved by the inclusion of other soil P fractions. A total of 145 soils were analyzed from field P fertilizer experiments conducted across Alberta from 1991 to 1993. Inorganic P (Pi) removed by extraction with an anion-exchange resin (resin P) was highly correlated with the Olsen and Kelowna-type soil test P methods and had a similar relationship with P fertilizer response. No appreciable improvement in the fit of available P with P fertilizer response was achieved by including any of the less available P fractions in the regression of P fertilizer response with available P. Little P was extractable in alkaline solutions (bicarbonate and NaOH), particularly in soils from the Brown and Dark Brown soil zones. Alkaline fractions were the most closely related to resin P, but the relationship depended on soil zone. Inorganic P extractable in dilute HCI was most strongly correlated with soil pH, reflecting accumulation in calcareous soils, while P(i) extractable in concentrated acids (HCI and H^SO^) was most strongly correlated with clay concentration. A positive but weak relationship was observed between these fractions and resin P. Complete fractionation of soil P confirmed that soil test P methods were assessing exchangeable, plant-available P.