Theme 2-3: Forage Production and Utilization--Poster Sessions
Description
Available empirical and mechanistic models to estimate total dry matter intake (DMI) and pasture dry matter intake (PDMI) of grazing lactating dairy cows have mainly been developed under intensive grazing conditions. The objective was to evaluate the adequacy of such existent models for use under semi-extensive grazing conditions, characterised by semi-natural grassland and less intensive resource use. Feed intake of lactating cows was measured on three commercial organic dairy farms in South Germany during one or two 6-d-periods/farm in 2019. Each period, DMI was determined in 10 or 20 cows per farm from their daily faecal output measured using titanium dioxide as marker and the apparent total tract digestibility of ingested organic matter derived from faecal crude protein concentration. PDMI was then calculated by subtracting weighed DMI of supplement feeds from total DMI. Further, individual milk yield and body weight were recorded, and samples of milk, pasture forage, supplement feed, and faeces were taken. For further analysis, means of observed values per farm and period (n = 7) were used, resulting in a diverse dataset of grazing systems differing in supplement feeding, daily pasture allocation, cow breed, and PDMI. Two semi-mechanistic PDMI models and six empirical and two semi-mechanistic models to predict total DMI were evaluated by their Mean Squared Error of Prediction, Relative Prediction Error (RPE) and Concordance Correlation Coefficient. One PDMI model produced acceptable (RPE < 20% of mean observed PDMI), and four total DMI models yielded satisfactory (RPE < 10%) prediction accuracy, however yet at a moderate precision (greatest Pearson correlation coefficient = 0.73). To further specify the grazing conditions under which the models reach both, satisfactory precision and accuracy, and due to the low number of observations, data of the present study will be complemented with data gathered on more farms in Southwest Germany in 2020.
Citation
Perdana-Decker, S.; Velasco, E.; Werner, J.; and Dickhoefer, U., "Evaluation of Herbage Intake Estimation Methods for Dairy Cattle Grazing on Semi-Extensive Pastures" (2022). IGC Proceedings (1993-2023). 16.
https://uknowledge.uky.edu/igc/24/2-3/16
Included in
Evaluation of Herbage Intake Estimation Methods for Dairy Cattle Grazing on Semi-Extensive Pastures
Available empirical and mechanistic models to estimate total dry matter intake (DMI) and pasture dry matter intake (PDMI) of grazing lactating dairy cows have mainly been developed under intensive grazing conditions. The objective was to evaluate the adequacy of such existent models for use under semi-extensive grazing conditions, characterised by semi-natural grassland and less intensive resource use. Feed intake of lactating cows was measured on three commercial organic dairy farms in South Germany during one or two 6-d-periods/farm in 2019. Each period, DMI was determined in 10 or 20 cows per farm from their daily faecal output measured using titanium dioxide as marker and the apparent total tract digestibility of ingested organic matter derived from faecal crude protein concentration. PDMI was then calculated by subtracting weighed DMI of supplement feeds from total DMI. Further, individual milk yield and body weight were recorded, and samples of milk, pasture forage, supplement feed, and faeces were taken. For further analysis, means of observed values per farm and period (n = 7) were used, resulting in a diverse dataset of grazing systems differing in supplement feeding, daily pasture allocation, cow breed, and PDMI. Two semi-mechanistic PDMI models and six empirical and two semi-mechanistic models to predict total DMI were evaluated by their Mean Squared Error of Prediction, Relative Prediction Error (RPE) and Concordance Correlation Coefficient. One PDMI model produced acceptable (RPE < 20% of mean observed PDMI), and four total DMI models yielded satisfactory (RPE < 10%) prediction accuracy, however yet at a moderate precision (greatest Pearson correlation coefficient = 0.73). To further specify the grazing conditions under which the models reach both, satisfactory precision and accuracy, and due to the low number of observations, data of the present study will be complemented with data gathered on more farms in Southwest Germany in 2020.
