Development of mathematical models to predict volume and nutrient composition of fresh manure from lactating Holstein cows

Abstract

Abstract. Organic compounds in dairy manure undergo a series of reactions producing pollutants such as ammonia and methane. Because various organic compounds have different reaction rates, the emissions could be accurately determined if amounts and concentrations of individual nutrients in manure are known. A set of empirical models were developed for predicting faecal and urinary water, carbon (C), nitrogen (N), acid detergent fibre and neutral detergent fibre output (kg/day) from lactating Holstein cows. Dietary nutrient contents, milk yield and composition, bodyweight, age and days in milk were used with or without dry matter intake (DMI) as potential predictor variables. Multi-collinearity, goodness of fit, model complexity, and random study and animal effects were taken into account during model development, which used 742 measured faecal or urinary nutrient output observations (kg/day). The models were evaluated with an independent dataset (n = 364). When DMI was used as a predictor variable, the models predicted faecal and urinary nutrient outputs successfully with root mean square prediction error as a percentage of average observed values (RMSPE%) ranging from 9.1% to 20.7%. All the predictions except urine output had RMSPE% ranging from 18.3% to 24.6% when DMI was not used. The nutrient output predictions were in reasonable agreement with observed values throughout the data range (systematic bias <14% of total bias). Fresh manure C : N ratio predictions were acceptable (RMSPE% = 14.3-15.2%) although the systematic bias were notable (17.1-20.7% of total bias). The models could be integrated successfully with process-based manure or soil models to assess nutrient transformation in dairy production systems.

Methane and nitrous oxide emissions from Canadian dairy farms and mitigation options: An updated review

“…Feeding increased levels of corn silage can also change nutrient digestibility and excretions, which would ultimately affect CH 4 emission potential of manure (Külling et al 2002;Nousiainen et al 2009;Appuhamy et al 2014). In Hassanat et al (2013) and Benchaar et al (2014), increasing corn silage at the expense of alfalfa silage or barley silage linearly increased apparent totaltract digestibility of organic matter (OMd) in Canadian cows, likely resulting in low C excretions in feces and thereby potentially low manure CH 4 emissions.…”

Prediction of drinking water intake by dairy cows

“…Models including multiple explanatory variables were then formed with all possible combination of variables having notable effects (P < 0.10), when fitted individually. However, highly correlated variables (|r| > 0.5) were not included together to minimize multicollinearity issues (Appuhamy et al, 2014a). For example, DMI and milk yield were regressed separately (r = 0.67, Table 3).…”

The Ruminant Farm Systems Animal Module: A Biophysical Description of Animal Management

“…Manure production and methane emissions of cows are divided into lactating and dry cows. For lactating cows, fecal water, total solids, urine excretion, manure excretion, volatile solids (separated into degradable and nondegradable volatile solids) are calculated according to [ 35 ] using dry matter intake as a predictor variable. The user has the option to select from three enteric methane emission calculations: (1) the US animal model described by Niu et al [ 36 ], (2) the Mills et al [ 37 ] Mitsherlich Model 3, or (3) the IPCC [ 38 ] Tier 2 model.…”

“…The Mills et al [ 37 ] and IPCC [ 38 ] models were selected to provide a comparison to other models commonly used in the dairy industry and environmental sciences. For instance, the Mills et al [ 35 ] Mitsherlich 3 equation is used in the Cornell Net Carbohydrate and Protein System [ 39 ] and Integrated Farm System Model [ 28 ]. The dietary gross energy is calculated for the IPCC [ 38 ] Tier 2 model according to [ 31 ] (Equation (S3)).…”