Optimizing accuracy of sampling protocols to measure nutrient content of solid manure

Abstract

Precise applications of manure to cropland can help optimize productivity and minimize environmental nutrient losses. Applying manure precisely is a challenge because the nutrient content of manures is inherently variable and the accuracy of sampling protocols are unknown. This study aimed to quantify the accuracy of sampling protocols for static solid manure piles considering both the number and depth of grab samples entering a composite sample. Over 35 grab samples were collected from each of ten static piles of dairy manure in California's Central Valley. Grab samples were individually analyzed for dry matter (DM), ash, total nitrogen, potassium, and phosphorous concentrations. Resampling simulations quantified the precision and bias of sampling protocols varying in both grab sample number and depth. Results showed that number of grab samples required for measurements to meet an accuracy standard of ±10% of the true value varied significantly by pile makeup. Over 25 grab samples were often required for multi-source manure piles, where an average of six grab samples were required from single source piles. The DM concentration of manure piles decreased at depths greater than 0.4 m, and sampling simulations showed that measurements were biased unless 70-80% of grab samples were collected from the pile interior. Both the number and location of grab samples necessary to create a representative composite require resource investments by farmers, and should be considered to manage nutrient applications cropland.

Keywords: Manure management; Manure sampling; Nutrient management; Uncertainty analysis.