How Lazy Are Pet Cats Really? Using Machine Learning and Accelerometry to Get a Glimpse into the Behaviour of Privately Owned Cats in Different Households.

Surprisingly little is known about how the home environment influences the behaviour of pet cats. This study aimed to determine how factors in the home environment (e.g., with or without outdoor access, urban vs. rural, presence of a child) and the season influences the daily behaviour of cats. Using accelerometer data and a validated machine learning model, behaviours including being active, eating, grooming, littering, lying, scratching, sitting, and standing were quantified for 28 pet cats. Generalized estimating equation models were used to determine the effects of different environmental conditions. Increasing cat age was negatively correlated with time spent active (p < 0.05). Cats with outdoor access (n = 18) were less active in winter than in summer (p < 0.05), but no differences were observed between seasons for indoor-only (n = 10) cats. Cats living in rural areas (n = 7) spent more time eating than cats in urban areas (n = 21; p < 0.05). Cats living in single-cat households (n = 12) spent more time lying but less time sitting than cats living in multi-cat households (n = 16; p < 0.05). Cats in households with at least one child (n = 20) spent more time standing in winter (p < 0.05), and more time lying but less time sitting in summer compared to cats in households with no children (n = 8; p < 0.05). This study clearly shows that the home environment has a major impact on cat behaviour.

The Use of Triaxial Accelerometers and Machine Learning Algorithms for Behavioural Identification in Domestic Cats (Felis catus): A Validation Study.

Animal behaviour can be an indicator of health and welfare. Monitoring behaviour through visual observation is labour-intensive and there is a risk of missing infrequent behaviours. Twelve healthy domestic shorthair cats were fitted with triaxial accelerometers mounted on a collar and harness. Over seven days, accelerometer and video footage were collected simultaneously. Identifier variables (n = 32) were calculated from the accelerometer data and summarized into 1 s epochs. Twenty-four behaviours were annotated from the video recordings and aligned with the summarised accelerometer data. Models were created using random forest (RF) and supervised self-organizing map (SOM) machine learning techniques for each mounting location. Multiple modelling rounds were run to select and merge behaviours based on performance values. All models were then tested on a validation accelerometer dataset from the same twelve cats to identify behaviours. The frequency of behaviours was calculated and compared using Dirichlet regression. Despite the SOM models having higher Kappa (>95%) and overall accuracy (>95%) compared with the RF models (64-76% and 70-86%, respectively), the RF models predicted behaviours more consistently between mounting locations. These results indicate that triaxial accelerometers can identify cat specific behaviours.

The Behaviour of Sheep around a Natural Waterway and Impact on Water Quality during Winter in New Zealand.

Access of livestock, such as cattle, to waterways has been shown to be a cause of poor water quality due to pugging damage and excretion entering the water. In New Zealand, regulations require that cattle, deer, and pigs are excluded from accessing waterways, but there are no such requirements for sheep. The current study utilised 24 h video cameras, global positioning system units, and triaxial accelerometers to observe the interaction of Romney ewes (n = 40) with a natural waterway. Ewes were either restricted (week 1) or given access to a reticulated water trough (week 2). Proximity data showed that ewes spent more time within 3 m of the waterway when the trough was unrestricted than when restricted (14.1 ± 5.7 and 10.8 ± 5.1 min/ewe/day, respectively; p < 0.05). Ewes travelled shorter distances on the steeper areas of paddock than flatter areas. Similarly, ewes showed a spatial preference for the flat and low sloped areas of the paddock. Concentrations of suspended sediment and total phosphorus were higher during access to a reticulated water trough which coincided with the week with more rainy days. Phosphorus and E. coli concentrations in the stream water samples were the above recommended Australian and New Zealand Environment and Conservation Council water quality guidelines, especially after rainy days, but did not appear to be directly related to sheep activity. Overall, the results suggest that during winter, ewes interacted very little with the waterway and were thus unlikely to influence the levels of nutrient and pathogens in the waterway.

Variations in the 24 h temporal patterns and time budgets of grazing, rumination, and idling behaviors in grazing dairy cows in a New Zealand system.

This study investigated the variations in the temporal distributions and the lengths of times utilized for grazing, ruminating, and idling behaviors by grazing dairy cows over 24 h. Spring-calved lactating dairy cows (N = 54) from three breeds, Holstein-Friesian (HFR), Jersey (JE), and KiwiCross (KC) in different lactations (1st, 2nd, 3rd) and with different breeding worth index values (103 < BW > 151) were selected. The cows were managed through a rotational grazing scheme and milked once a day at 0500 hours. The cows grazed mainly pasture and consumed additional feeds (maize silage and turnips) in the summer and autumn seasons. AfiCollar was used to record grazing and rumination behaviors (min/h) in the individual cows throughout the lactation period (~270 d). The time neither utilized for grazing nor rumination was counted as idling behavior (min/h). A repeat measure design with PROC MIXED was performed in SAS considering the effects of breed, lactation, individual cow, the hour of the day, season, day within the season, and supplementary feed within the season to evaluate the difference in grazing, rumination, and idling behaviors. Hour of the day, season, day within season, and supplementary feed had significant effects on grazing, rumination, and idling behaviors. Regardless of the season and supplementary feed, cows spent most of the daytime grazing and most of the nighttime ruminating. Grazing activity remained consistently high throughout the day with two peaks around dawn and dusk and a short peak around midnight. Rumination activity remained high from the late evening until early morning. Grazing and ruminating patterns were similar between different breeds and lactations, however, JE cows grazed slightly longer than HFR and KC, and first-lactation cows grazed slightly longer than those in higher lactations. The onset and cessation of grazing activity by the cows were adjusted according to varying day lengths by season. Cows finished grazing earlier when they consumed additional supplements or silage along with pasture. Cows from different breed groups and lactations spent most of their 24 h grazing followed by ruminating and idling. Season and supplementary feed potentially affected the variations in behavior time budgets. These findings should support improving measures for grazing management to address pasture allocation and additional feed demands, and animal welfare in varying environmental and/or managemental conditions.

This study explored how grazing dairy cows pattern their essential such as including grazing, rumination, and idling, and how they distribute their time for those behaviors over 24 h. We used a group of spring-calved grazing dairy cows affiliated with different breeds, milking ages, and genetic merits and recorded their grazing and rumination behaviors for the whole milking period. An automated device, AfiCollar was used to continuously record minutes within an hour (min/h) utilized for grazing, rumination, and idling. The cows were mainly offered grass with some additional supplementary feeds on various days in summer and autumn and milked once a day at 0500 hours. Regardless of the breed, milking age, season, and supplementary feeds, grazing cows spent most of the daytime (from dawn to dusk) grazing and most of the nighttime ruminating, with a short grazing period around midnight. Cows adopted their grazing patterns according to varying day lengths during different seasons and finished grazing earlier when received supplementary feeds. Grazing cows allocated most of their time over 24 h for grazing followed by ruminating and idling. These findings could have implications to improve the measures for pasture management efficiency and additional feed demand, and animal welfare in varying environmental and/or managemental conditions.

Associations of Grazing and Rumination Behaviours with Performance Parameters in Spring-Calving Dairy Cows in a Pasture-Based Grazing System.

This study investigated the relationship of the length of time spent grazing and ruminating with the performance parameters of spring-calved grazing dairy cows (n = 162) over the lactation period for three lactation seasons (n = 54 per season). The cows were Holstein Friesian (HFR), Jersey (JE), and a crossbreed of Holstein Friesian/Jersey (KiwiCross), with 18 cows from each breed. The cows were either in their 1st, 2nd, 3rd, or 4th lactation year, and had different breeding worth (BW) index values (103 < BW > 151). The cows were managed through a rotational grazing scheme with once-a-day milking in the morning at 05:00 h. The cows were mainly fed on grazed pastures consisting of perennial ryegrass (Lolium perenne), red clover (Trifolium pretense), and white clover (Trifolium repens), and received additional feeds on various days in the summer and autumn seasons. This study used an automated AfiCollar device to continuously record the grazing time and rumination time (min/h) of the individual cows throughout the lactation period (~270 days) for three consecutive years (Year-1, Year-2, and Year-3). The milk yield, milk fat, milk protein, milk solids, liveweight, and body condition score data of the individual animals for the study years were provided by the farm. PROC CORR was used in SAS to determine the correlation coefficients (r) between the behaviour and production parameters. A general linear model fitted with breed × lactation year, individual cows, seasons, feed within the season, grazing time, rumination time, as well as their interactions, was assessed to test the differences in milk yield, milk fat, milk protein, milk solids, liveweight, and body condition score. The type I sum of squares values were used to quantify the magnitude of variance explained by each of the study factors and their interactions in the study variables. Grazing time exhibited positive associations with MY (r = 0.34), MF (r = 0.43), MP (r = 0.22), MS (r = 0.39), LW (r = -0.47), and BCS (r = -0.24) throughout the study years. Rumination time was associated with MY (r = 0.64), MF (r = 0.57), MP (r = 0.52), and MS (r = 0.57) in all study years, while there were no effects of rumination time on LW (r = 0.26) and BCS (r = -0.26). Grazing time explained up to 0.32%, 0.49%, 0.17%, 0.31%, 0.2%, and 0.02%, and rumination time explained up to 0.39%, 6.73%, 4.63%, 6.53%, 0.44%, and 0.17% of the variance in MY, MF, MP, MS, LW, and BCS, respectively.

Review and update of a Nutrient Transfer model used for estimating nitrous oxide emissions from complex grazed landscapes, and implications for nationwide accounting.

In New Zealand, nitrous oxide emissions from grazed hill pastures are estimated using different emission factors for urine and dung deposited on different slope classes. Allocation of urine and dung to each slope class needs to consider the distribution of slope classes within a landscape and animal behavior. The Nutrient Transfer (NT) model has recently been incorporated into the New Zealand Agricultural GHG Inventory Model to account for the allocation of excretal nitrogen (N) to each slope class. In this study, the predictive ability of the transfer function within the NT model was explored using urine deposition datasets collected with urine sensor and GPS tracker technology. Data were collected from three paddocks that had areas in low (<12°), medium (12-24°), and high slopes (>24°). The NT model showed a good overall predictive ability for two of the three datasets. However, if the urine emission factors (% of urine N emitted as N2 O-N) were to be further disaggregated to assess emissions from all three slope classes or slope gradients, more precise data would be required to accurately represent the range of landscapes found on farms. We have identified the need for more geospatial data on urine deposition and animal location for farms that are topographically out of the range used to develop the model. These new datasets would provide livestock urine deposition on a more continuous basis across slopes (as opposed to broad ranges), a unique opportunity to improve the performance of the NT model.

The Nutrient Transfer model allocates urine from grazing livestock to different slope classes. The predictive ability of the model was explored using urine sensor and tracker data of grazing livestock. The model showed a good overall predictive ability for two of the three datasets explored. There is a need for more geospatial urine deposition and animal location data on complex land.

Factors Affecting Grazing and Rumination Behaviours of Dairy Cows in a Pasture-Based System in New Zealand.

This study investigated the variation in daily time spent grazing and rumination in spring-calved grazing dairy cows (n = 162) of three breeds, Holstein-Friesian (HFR), Jersey (JE), and KiwiCross (KC) with different breeding worth index, and in different years of lactation (1st, 2nd, 3rd, 4th). The cows were managed through a rotational grazing system and milked once a day at 05:00 a.m. The cows grazed mainly pasture and received supplementary feeds depending on the season. Automated AfiCollar device continuously monitored and recorded grazing time and rumination time of the individual cows throughout the lactation period for three study years (Year-1, Year-2, Year-3) with 54 cows per year. A general linear mixed model fitted with breed × lactation year with days in milk (DIM), breeding worth (BW) index value, individual cow, season, and feed, and their interactions was performed in SAS. Variance partitioning was used to quantify the effect size of study factors and their interactions. Individual cows, DIM, and BW (except Year-3) had effects on grazing and rumination times throughout the study years. Grazing time and rumination time were different for different seasons due to varying supplementary feeds. Grazing time varied among breeds in Year-2 and Year-3, and among lactation years only in Year-1. Although rumination time differed among breeds in Year-3, it remained the same within different lactation years. Grazing time and rumination time had a negative relationship with each other, and their regression lines varied for different seasons. The total variance explained by the model in grazing time was 36-39%, mainly contributed by the individual cow (12-20%), season (5-12%), supplementary feed (2-6%), breed (1-5%), and lactation year (1-6%). The total variance explained in rumination was 40-41%, mainly contributed by the individual cow (16-24%), season (2-17%), supplementary feed (1-2%), breed (2-8%), and lactation year (~1%). These findings could contribute to improving the measures for feed resource management during different seasons over the lactation period for a mixed herd comprising JE, HFR and KC breeds in different years of lactation.

Validation of an Accelerometer Sensor-Based Collar for Monitoring Grazing and Rumination Behaviours in Grazing Dairy Cows.

This study evaluated the accuracy of a sensor-based device (AfiCollar) to automatically monitor and record grazing and rumination behaviours of grazing dairy cows on a real-time basis. Multiparous spring-calved dairy cows (n = 48) wearing the AfiCollar were selected for the visual observation of their grazing and rumination behaviours. The total observation period was 36 days, divided into four recording periods performed at different times of the year, using 12 cows in each period. Each recording period consisted of nine daily observation sessions (three days a week for three consecutive weeks). A continuous behaviour monitoring protocol was followed to visually observe four cows at a time for each daily observation session, from 9:00 a.m. to 5:00 p.m. Overall, 144 observations were collected and the data were presented as behaviour activity per daily observation session. The behaviours visually observed were also recorded through an automated AfiCollar device on a real-time basis over the observation period. Automatic recordings and visual observations were compared with each other using Pearson's correlation coefficient (r), Concordance correlation coefficient (CCC), and linear regression. Compared to visual observation (VO), AfiCollar (AC) showed slightly higher (10%) grazing time and lower (4%) rumination time. AC results and VO results had strong associations with each other for grazing time (r = 0.91, CCC = 0.71) and rumination time (r = 0.89, CCC = 0.80). Regression analysis showed a significant linear relationship between AC and VO for grazing time (R2 = 0.83, p < 0.05) and rumination time (R2 = 0.78, p < 0.05). The relative prediction error (RPE) values for grazing time and rumination time were 0.17 and 0.40, respectively. Overall, the results indicated that AfiCollar is a reliable device to accurately monitor and record grazing and rumination behaviours of grazing dairy cows, although, some minor improvements can be made in algorithm calibrations to further improve its accuracy.