Some abstracts do not have video files because ASAS was denied recording rights.

Validation of the GreenFeed system against model predicted methane emissions

Saturday, July 23, 2016: 10:00 AM
155 E (Salt Palace Convention Center)
Pekka Huhtanen , Swedish University of Agricultural Sciences (SLU), Umea, Sweden
Mohammad Ramin , Swedish University of Agricultural Sciences (SLU), UmeƄ, Sweden
A. N. Hristov , The Pennsylvania State University, University Park, PA
Abstract Text:

The GreenFeed (GF) system (C-Lock Inc., Rapid City, SD, USA) was introduced to estimate methane (CH4) emission by measuring gas concentrations and flux when cattle visit a GF. The objective of the present study was to validate CH4 measured with the GF system with model predicted CH4 emissions. Evaluation was based on 55 treatment means from dairy and beef cattle studies, in which CH4 emission was measured by GF. Methane emission was predicted with the models of Yan et al. (2000; models Y1 and Y2), Ellis et al. (2007; E), Jentsch et al. (2007; J) and Ramin and Huhtanen (2013; R1 and R2). If the parameter values required in the models were not reported, tabulated values were used. A concentration of 18.5 MJ GE/kg DM was assumed for models based on GE intake. The evaluation was based on root mean squared prediction error (RMSPE) expressed as a proportion of observed mean. RMSPE was divided into errors resulting from mean bias, slope bias and random error across regression line. Observed mean (SD) CH4 emission DMI and dietary concentrations of CP and NDF were 386 (103) g/d, 18.4 (5.6) kg/d, 172 (34) g/kg DM and 401 (81) g/kg DM, respectively. Mean CH4 emission estimated by GF was close to values predicted by models Y, J and R that were developed from respiration chamber data (Table 1). However, observed CH4 values were much higher than predicted by model E that was developed from data determined by different techniques. RMSPE ranged from 6 to 9% of observed mean except model E with most of the error due to random variation. RMSPE was smaller when the effect of feeding level was taken into account compared to a model based only on intake (Y2 and J vs. Y1 and E). RMSPE was further reduced when the effects of diet digestibility and composition were taken into account in addition to intake (R1 and R2 vs. others). It is concluded that CH4 emissions estimated by GF were consistent with values predicted by models derived from large datasets from respiration chamber studies.

Keywords: methane, GreenFeed, model

Table 1. Predicted CH4 emission (Observed 386 g/d)