Non-point sources of agricultural nutrients, phosphorus in particular, have become an important consideration for surface water quality deterioration in Manitoba water bodies. However, there is a shortage of information with respect to the relationship between measures of soil P and runoff P under field conditions in Manitoba. The objectives of this study are: (i) to quantify the relationship between measures of soil P (STP) and runoff P resulting from rainfall simulation trials and, (ii) to identify factors that influence this relationship.  Factors of interest include site location (i.e. soil type), crop residue management practices (with or without residue), variations in runoff P concentrations over time (early versus late in the runoff event), and soil P surplus. Field runoff simulation trials were conducted at five locations under agricultural annual crops (Deerwood, two sites in the Brandon area, and two sites near Oak Bluff), and three locations with perennial grass vegetation (Sanford, Rosser and Beaudry Provincial Park). Runoff simulations where conducted during the fall of 2009 (Sept. – Oct.) and summer - fall 2010 (July – Oct.) shortly after crop harvest.

Regression analysis showed that measures of dissolved P in runoff were linearly related with measurements of soil P content, with higher concentrations and loads of runoff P associated with greater soil P content. Various STP methods (Olsen P, Mehlich III P, Modified Kelowna P, CaCl₂ Extractable P, Water Extractable P) and degree of phosphorus saturation (DPS) methods were evaluated in terms of their ability to predict runoff dissolved P losses from agricultural land. Generally, STP methods were more accurate for predicting soil P losses than the more complex DPS methods, with Olsen P, Mehlich III P and Modified Kelowna P having the greatest correlation coefficients. Relationships between soil P and runoff dissolved P measures are further improved by accounting for the influence of factors listed above.