HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Full Text (PDF) All Versions of this Article: jas.2008-0960v1 86/10/2738    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Kebreab, E. Articles by Wirth, T. PubMed PubMed Citation Articles by Kebreab, E. Articles by Wirth, T.

Model for estimating enteric methane emissions from US dairy and feedlot cattle

E. Kebreab^*, K. A. Johnson, S. L. Archibeque, D. Pape and T. Wirth¹

^* National Centre for Livestock and Environment, Department of Animal Science, University of Manitoba, Winnipeg, Manitoba R3T 2N2, Canada , Washington State University, Department of Animal Science, Pullman, WA 99164 USA , Colorado State University, Department of Animal Science, Ft Collins, CO 80523 USA , ICF International, Washington DC, USA and ¹ Environmental Protection Agency, Washington DC, USA

kebreabe{at}cc.umanitoba.ca

Abstract

Methane production from enteric fermentation in cattle is one of the major sources of anthropogenic greenhouse gas emission in the United States and worldwide. National estimates of methane emissions rely on mathematical models such as the one recommended by the Intergovernmental Panel for Climate Change (IPCC). Models used for prediction of methane emissions from cattle range from empirical to mechanistic with varying input requirements. Two empirical (IPCC, 2006; Moe and Tyrrell, 1979) and 2 mechanistic models (Dijkstra et al., 1992 (COWPOLL); Baldwin, 1995 (MOLLY)) were evaluated for their prediction ability using individual cattle measurements. Model selection was based on mean square prediction error (MSPE), concordance correlation coefficient and residuals vs. predicted values analyses. In dairy cattle, COWPOLL had the lowest root MSPE and greatest accuracy and precision of predicting methane emissions (correlation coefficient estimate = 0.75). The model simulated differences in diet more accurately than the other models and the residuals vs. predicted value analysis showed no mean bias (P = 0.71). In feedlot cattle, MOLLY had the lowest root MSPE with almost all errors from random sources (correlation coefficient estimate = 0.69). The IPCC model also had good agreement with observed values and no significant mean (P = 0.74) or linear bias (P = 0.11) was detected when residuals were plotted against predicted values. A fixed methane conversion factor (Ym) might be an easier alternative to diet dependent variable Ym. Based on results, the 2 mechanistic models were used to simulate methane emissions from representative US diets and compared with the IPCC model. The average Ym in dairy cows was 5.63% of GE (range 3.78 to 7.43%) compared to 6.5% ± 1% recommended by IPCC. In feedlot cattle the average Ym was 3.88% (range 3.36 to 4.56%) compared to 3% ± 1% recommended by IPCC. Based on our simulations, using IPCC values can result in an overestimate of about 12.5% and underestimate of emissions by about 9.8% for dairy and feedlot cattle, respectively. In addition to providing improved estimates of emissions based on diets, mechanistic models can be used to assess mitigation options such as changing source of carbohydrate or addition of fat to reduce methane, which is not possible with empirical models. We recommend national inventories use diet specific Ym values predicted by mechanistic models to estimate methane emissions from cattle.

Key Words: cattle • greenhouse gas • methane • modeling