Relationships between paddock-level pasture mass and cow data derived from GPS and behavior-monitoring on-animal sensors in rotationally grazed dairy systems
ABSTRACT: Data from behavior-monitoring and location (global positioning system) devices fitted to dairy cows may improve our understanding of how animal behavior and movement are associated with feed availability and quality. We hypothesized that data from behavior-monitoring and location sensors m...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-02-01
|
| Series: | Journal of Dairy Science |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S0022030224013535 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832591046723239936 |
|---|---|
| author | S.J. Hendriks M. Qasim J.P. Edwards |
| author_facet | S.J. Hendriks M. Qasim J.P. Edwards |
| author_sort | S.J. Hendriks |
| collection | DOAJ |
| description | ABSTRACT: Data from behavior-monitoring and location (global positioning system) devices fitted to dairy cows may improve our understanding of how animal behavior and movement are associated with feed availability and quality. We hypothesized that data from behavior-monitoring and location sensors may be associated with feed availability in a paddock within a rotationally grazed dairy system. To investigate this, 100 cows were randomly assigned to one of 4 groups (n = 25 cows per group) and allocated to different target pasture allocations to meet either 80%, 100%, or 120% of their estimated ME requirements across 2 experimental periods (n = 20 d per experimental period), during late spring (experimental period 1; November 7 to November 26, 2021) and late summer (experimental period 2; February 27 to March 18, 2022). During both periods, all 4 groups were allocated 100% of ME requirements for 5 d (baseline). Then, 2 groups were under-allocated (80%), whereas the other 2 groups were over-allocated (120%) for 5 d. Subsequently, all groups returned to baseline for 5 d (100%), followed by a switch, where the under-allocated groups were over-allocated, and vice versa, for the final 5 d. Each cow was fitted with 5 devices that were commercially available in New Zealand and measured rumination, eating, grazing, and lying time, and activity. One device also determined cows' location, and these data were used to derive 3 additional behaviors: distance traveled, mean distance to herd mates, and proximity. These data were used as independent variables to build linear models with pasture mass as the response, which was estimated during the experiment using a calibrated rising plate meter. Distance traveled, standardized by pasture area allocated, was an important variable for explaining the variance in paddock-level pasture mass, alongside mean distance to herd mates and rumination time. Using location variables alone (distance traveled and distance to herd mates), adjusted R2 values were 0.38 and 0.40 for both pre- and postgrazing pasture mass (kg DM/cow), respectively. Further, including both location and behavior, model fit improved due to greater variation in pasture mass explained by these independent variables. The best linear model (adjusted R2 = 0.58) was for postgrazing pasture mass (kg DM/cow) with standardized distance traveled, distance to herd mates, and rumination time included as the independent variables. Model fit varied depending on location and behavior variables included, and devices lacking behavior data related to ingestion and mastication of feed (e.g., rumination, grazing, or eating data) were generally poorer performers. Our results demonstrate the additional value that location-based data can provide. Irrespective of this, predictive potential may be limited due to a moderate amount of variation in pasture mass explained by data from behavior-monitoring and location sensors using linear modeling approaches. Therefore, this method may not be suitably accurate to make near real-time grazing management decisions, but the results are promising as a concept. |
| format | Article |
| id | doaj-art-97f1fd271e9e45dea7c34b3431006336 |
| institution | Kabale University |
| issn | 0022-0302 |
| language | English |
| publishDate | 2025-02-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Journal of Dairy Science |
| spelling | doaj-art-97f1fd271e9e45dea7c34b34310063362025-01-23T05:25:21ZengElsevierJournal of Dairy Science0022-03022025-02-01108216441658Relationships between paddock-level pasture mass and cow data derived from GPS and behavior-monitoring on-animal sensors in rotationally grazed dairy systemsS.J. Hendriks0M. Qasim1J.P. Edwards2DairyNZ Ltd., Lincoln 7647, New Zealand; Corresponding authorDairyNZ Ltd., Hamilton 3240, New ZealandDairyNZ Ltd., Lincoln 7647, New ZealandABSTRACT: Data from behavior-monitoring and location (global positioning system) devices fitted to dairy cows may improve our understanding of how animal behavior and movement are associated with feed availability and quality. We hypothesized that data from behavior-monitoring and location sensors may be associated with feed availability in a paddock within a rotationally grazed dairy system. To investigate this, 100 cows were randomly assigned to one of 4 groups (n = 25 cows per group) and allocated to different target pasture allocations to meet either 80%, 100%, or 120% of their estimated ME requirements across 2 experimental periods (n = 20 d per experimental period), during late spring (experimental period 1; November 7 to November 26, 2021) and late summer (experimental period 2; February 27 to March 18, 2022). During both periods, all 4 groups were allocated 100% of ME requirements for 5 d (baseline). Then, 2 groups were under-allocated (80%), whereas the other 2 groups were over-allocated (120%) for 5 d. Subsequently, all groups returned to baseline for 5 d (100%), followed by a switch, where the under-allocated groups were over-allocated, and vice versa, for the final 5 d. Each cow was fitted with 5 devices that were commercially available in New Zealand and measured rumination, eating, grazing, and lying time, and activity. One device also determined cows' location, and these data were used to derive 3 additional behaviors: distance traveled, mean distance to herd mates, and proximity. These data were used as independent variables to build linear models with pasture mass as the response, which was estimated during the experiment using a calibrated rising plate meter. Distance traveled, standardized by pasture area allocated, was an important variable for explaining the variance in paddock-level pasture mass, alongside mean distance to herd mates and rumination time. Using location variables alone (distance traveled and distance to herd mates), adjusted R2 values were 0.38 and 0.40 for both pre- and postgrazing pasture mass (kg DM/cow), respectively. Further, including both location and behavior, model fit improved due to greater variation in pasture mass explained by these independent variables. The best linear model (adjusted R2 = 0.58) was for postgrazing pasture mass (kg DM/cow) with standardized distance traveled, distance to herd mates, and rumination time included as the independent variables. Model fit varied depending on location and behavior variables included, and devices lacking behavior data related to ingestion and mastication of feed (e.g., rumination, grazing, or eating data) were generally poorer performers. Our results demonstrate the additional value that location-based data can provide. Irrespective of this, predictive potential may be limited due to a moderate amount of variation in pasture mass explained by data from behavior-monitoring and location sensors using linear modeling approaches. Therefore, this method may not be suitably accurate to make near real-time grazing management decisions, but the results are promising as a concept.http://www.sciencedirect.com/science/article/pii/S0022030224013535GPS locationpasture massanimal behaviorwearablesprecision livestock farming |
| spellingShingle | S.J. Hendriks M. Qasim J.P. Edwards Relationships between paddock-level pasture mass and cow data derived from GPS and behavior-monitoring on-animal sensors in rotationally grazed dairy systems Journal of Dairy Science GPS location pasture mass animal behavior wearables precision livestock farming |
| title | Relationships between paddock-level pasture mass and cow data derived from GPS and behavior-monitoring on-animal sensors in rotationally grazed dairy systems |
| title_full | Relationships between paddock-level pasture mass and cow data derived from GPS and behavior-monitoring on-animal sensors in rotationally grazed dairy systems |
| title_fullStr | Relationships between paddock-level pasture mass and cow data derived from GPS and behavior-monitoring on-animal sensors in rotationally grazed dairy systems |
| title_full_unstemmed | Relationships between paddock-level pasture mass and cow data derived from GPS and behavior-monitoring on-animal sensors in rotationally grazed dairy systems |
| title_short | Relationships between paddock-level pasture mass and cow data derived from GPS and behavior-monitoring on-animal sensors in rotationally grazed dairy systems |
| title_sort | relationships between paddock level pasture mass and cow data derived from gps and behavior monitoring on animal sensors in rotationally grazed dairy systems |
| topic | GPS location pasture mass animal behavior wearables precision livestock farming |
| url | http://www.sciencedirect.com/science/article/pii/S0022030224013535 |
| work_keys_str_mv | AT sjhendriks relationshipsbetweenpaddocklevelpasturemassandcowdataderivedfromgpsandbehaviormonitoringonanimalsensorsinrotationallygrazeddairysystems AT mqasim relationshipsbetweenpaddocklevelpasturemassandcowdataderivedfromgpsandbehaviormonitoringonanimalsensorsinrotationallygrazeddairysystems AT jpedwards relationshipsbetweenpaddocklevelpasturemassandcowdataderivedfromgpsandbehaviormonitoringonanimalsensorsinrotationallygrazeddairysystems |