Nutrient management planning for sow operations using extant and new feed consumption models and Manure Analysis


Project Lead:  Dr. Ermias Kebreab, University of California

Executive Summary

Download Full Final Report here (pdf)

Successful nutrient management planning of sow operations depends on accurate estimation of nutrient output from the animal. The objective of the study was to evaluate an empirical model that predicts phosphorus (P) and nitrogen (N) output and used to calculate land base requirement for manure spreading by Manitoba Conservation. In addition, the effect of microbial phytase supplementation on land base requirement was investigated. The model was originally developed for Quebec, Canada and implemented in Manitoba, Canada. An experiment was conducted at the T.K. Cheung Centre, University of Manitoba, Winnipeg, MB. Eighteen sows were randomly allocated to receive a diet that meets NRC requirement for P (6.0 g total P/kg; NOPHYTASE) or reduced P (5.1 g P/kg) supplemented with microbial phytase at 500 FTU/kg (PHYTASE). Samples of feed, feces and urine were collected for five days after seven days of acclimatization. Urine was collected using urinary catheters. The results were a composite of three collection periods. Reducing P intake by nearly 2 g/d and supplementing phytase resulted in numerically higher P and N retentions in the sows. Furthermore, there was a tendency for the sows to excrete less P and N in manure. Measured total P in manure (feces + urine) was compared to model predicted values. Evaluation was based on mean square prediction error (MSPE) which was further decomposed into error of prediction due to variation from regression line, central tendency (or deviation from the mean) and random (unexplained error). The MSPE analysis showed that root MSPE as percentage of observed mean was 10% in NOPHYTASE and 30% in PHYTASE treatments. This was further confirmed in the decomposition of sources of error with only 26% and 50% coming from random variation in NOPHYTASE and PHYTASE diets, respectively. This means that the model predicted P excretions from the standard diet better than the phytase supplemented diet. The annual land base requirement for manure spreading based on the P content of manure from sows fed NOPHYTASE and PHYTASE diets were 0.24 and 0.23 ha/sow, respectively. However, the model predicted land requirements of 0.32 and 0.28 ha/sow, respectively. The overestimation by the model, which was directly linked to P output predictions, has a significant implication to producers in determining the amount of animals they are allowed to keep. Therefore, it is recommended that the model should be refined to reflect local conditions if it were to be used as nutrient management planning and monitoring tool. It is particularly important to revise the model if producers are using microbial phytase to reduce P intake and output. The project results can be used as a benchmark to the amount of P and N expected to be excreted from non-lactating dry sows in Manitoba.