Predicting dairy cattle heat stress using machine learning techniques.

The objectives of the study were to use a heat stress scoring system to evaluate the severity of heat stress on dairy cows using different heat abatement techniques. The scoring system ranged from 1 to 4, where 1 = no heat stress; 2 = mild heat stress; 3 = severe heat stress; and 4 = moribund. The accuracy of the scoring system was then predicted using 3 machine learning techniques: logistic regression, Gaussian naïve Bayes, and random forest. To predict the accuracy of the scoring system, these techniques used factors including temperature-humidity index, respiration rate, lying time, lying bouts, total steps, drooling, open-mouth breathing, panting, location in shade or sprinklers, somatic cell score, reticulorumen temperature, hygiene body condition score, milk yield, and milk fat and protein percent. Three different treatments, namely, portable shade structure, portable polyvinyl chloride pipe sprinkler system, or control with no heat abatement, were considered, where each treatment was replicated 3 times with 3 second-trimester lactating cows. Results indicate that random forest outperformed the other 2 methods, with respect to both accuracy and precision, in predicting the sprinkler group's score. Both logistic regression and random forest were consistent in predicting scores for control, shade, and combined groups. The mean probability of predicting non-heat-stressed cows was highest for cows in the sprinkler group. Finally, the logistic regression method worked best for predicting heat-stressed cows in control, shade, and combined. The insights gained from these results could aid dairy producers to detect heat stress before it becomes severe, which could decrease the negative effects of heat stress, such as milk loss.

Graduate Student Literature Review: Heat abatement strategies used to reduce negative effects of heat stress in dairy cows.

The southeastern United States experiences an extended hot season with a high environmental temperature and relative humidity. With increasing global temperatures, managing dairy cattle in regions with tropical, subtropical, and Mediterranean climates is becoming an increasing challenge. Heat-stressed cows will decrease feed intake, decrease productivity, and increase respiration rate in an attempt to maintain internal body temperature. Temperature-humidity index (THI) is a unitless value that has been used to measure the magnitude of heat stress on dairy cows. Many researchers have studied the THI threshold at which dairy cattle begin to experience heat stress. When housing cows in a confinement setting, a pasture-based setting, or a combination of the two, the appropriate heat abatement should be implemented to allow cows to perform to their potential and to improve overall animal welfare. This review summarizes heat abatement strategies that have been studied to reduce the negative effects of heat stress.

Invited review: Physiological and behavioral effects of heat stress in dairy cows.

Animal welfare can be negatively affected when dairy cattle experience heat stress. Managing heat stress has become more of a challenge than ever before, due to the increasing number of production animals with increased milk yield, and therefore greater metabolic activity. Environmental temperatures have increased by 1.0°C since the 1800s and are expected to continue to increase by another 1.5°C between 2030 and 2052. Heat stress affects production, reproduction, nutrition, health, and welfare. Means exist to monitor and evaluate heat stress in dairy cattle, as well as different ways to abate heat, all with varying levels of effectiveness. This paper is a summary and compilation of information on dairy cattle heat stress over the years.

Symposium review: The most important factors affecting adoption of precision dairy monitoring technologies.

Precision dairy monitoring involves the use of technologies to measure physiological, behavioral, and production indicators on individual animals to detect events of interest. Estrus, disease, and calving detection are common applications, although estrus detection is the most tested and used. Many precision dairy monitoring technologies (PDMT) are commercially available and are being used in research and on farms. As a result, a common question from researchers and producers alike is, "what PDMT should I buy?" The answer to this question is inherently complicated because it depends on many factors, some of which researchers have yet to explore. The objective of this paper is to examine the less quantitively researchable aspects of PDMT adoption and use on-farm. This will be done through 3 lists of 5, determined from published theory and my own experience. First, the 5 main factors that influence adoption of an innovation: (1) relative advantage, (2) compatibility, (3) complexity, (4) trialability, and (5) observability. Each of these factors is at play to a different extent in the 5 adopter categories: (1) innovators, (2) early adopters, (3) early majority, (4) late majority, and (5) laggards. From my experience and research, the top PDMT are those that improve (1) farm efficiency; (2) farm economics; (3) decision-making; (4) animal welfare; and (5) producer happiness. Implementing PDMT on a farm is an enormous and potentially expensive decision. As this part of the industry continues to progress, the potential for different PDMT is endless. Sound research and producer feedback are imperative to ensuring that PDMT continue to improve and become more widely adopted.

Comparing dairy farm milk yield and components, somatic cell score, and reproductive performance among United States regions using summer to winter ratios.

Heat stress abatement is a challenge for dairy producers in the United States, especially in the southern states. Thus, managing heat stress is critical to maintain dairy cow performance in the summer. The ability to employ a metric to measure heat stress and evaluate abatement strategies may benefit dairy producers by providing meaningful feedback on the effectiveness of current and future management strategies with the goal of improving heat stress management. Therefore, this study aimed to explore the use of the summer to winter performance ratio metric to quantify and compare farm performance variables among regions of the United States. Monthly performance data recorded by the Dairy Herd Improvement Association from 2007 to 2016, for all US Dairy Herd Improvement Association herds processing records through Dairy Records Management Systems (Raleigh, NC), were obtained. Season dates were based on the astronomical definition of the Northern Hemisphere with summer as June 21 to September 21 and winter as December 21 to March 19. States were grouped into regions based on climate zone classification. Performance records included a total of 16,573 herds [Northeast (n = 7,955), Midwest (n = 6,555), Northern Plains (n = 305), Southeast (n = 1,370), and Southern Plains (n = 388) regions]. Herd test day performance variables energy-corrected milk, somatic cell score, milk fat and protein percentage, conception rate, heat detection rate, and pregnancy rate in summer and winter were used to calculate summer to winter ratios for each region. The MIXED procedure of SAS 9.4 (SAS Institute Inc., Cary, NC) was used to compare test day performance variables. The effects of year, mean days in milk, mean 150-d milk, mean herd size, and number of milkings per day were included as covariates in the models. Dairy cattle performance in all climate regions was negatively affected by summer heat stress, but some regions greater than others. A difference was also observed among regions when comparing summer to winter ratios for each performance parameter. This indicates that summer performance varies by climate region identified by the summer to winter ratio and demonstrates usefulness of the metric to monitor degree of heat stress based on performance.

Automated estrous detection using multiple commercial precision dairy monitoring technologies in synchronized dairy cows.

Estrus in dairy cattle varies in duration and intensity, highlighting the need for accurate and continuous monitoring to determine optimal breeding time. The objective of this study was to evaluate precision dairy monitoring technologies (PDMT) for detecting estrus. Estrus was synchronized in lactating Holstein cows (n = 109) using a modified G7G-Ovsynch protocol (last GnRH injection withheld to permit expression of estrus) beginning at 45 to 85 d in milk. Resumption of ovarian cyclicity at enrollment was verified by transrectal ultrasonography for presence of a corpus luteum. Cows were observed visually during 30 min (4 times per day) for behavioral estrus on d -1 to 2 (d 0 = day of estrus). Periods peri-estrus were defined by the temporal blood plasma progesterone patterns on d -5, -4, -3, -2, -1, 0, 2, 4, 6, and 8. Estrous detection by PDMT, an estrous behavior scoring system, and by visual observation of standing estrus were compared with the reference (gold) standard. Only 56% of cows that ovulated were observed standing by visual observation. Sensitivity and specificity for estrous detection were not different among all PDMT. Devices in this study measuring activity in steps, neck movement, high activity of head movement, or a proprietary motion index increased on the day of estrus 69 to 170% from the baseline before estrus. The change in rumination time on the day of estrus decreased for both neck and ear-based technologies (-2 to -16%). Temperature of the reticulorumen, vagina, and ear skin were not different on the day of estrus than day peri-estrus. Daily lying times decreased on average to 24.6% on the day of estrus for IceQube (IceRobotics Ltd., Edinburgh, Scotland). In contrast, lying time increased 15.5 and 33.1% for AfiAct Pedometer Plus (Afimilk, Kibbutz Afikim, Israel) and Track a Cow (ENGS Systems Innovative Dairy Solutions, Rosh Pina, Israel), respectively. All PDMT tested were capable of detecting estrus at least as effectively as visual observation. Four of the 6 PDMT that reported estrous alerts correctly detected 15 to 35% more cows than visual observation 4 times per day. Use of temporal progesterone patterns correctly identified more cows than visual observation alone. Dairy producers considering PDMT should focus on (1) the reference (gold) standard used to test efficacy of a device's alerts and (2) the device that will have the fewest false readings in their operations.

Mosquitoes (Diptera: Culicidae) Collected From Residential Yards and Dog Kennels in Florida Using Two Aspirators, a Sweep Net, or a CDC Trap.

Mosquito surveillance typically uses Centers for Disease Control and Prevention (CDC) mosquito light traps baited with CO2. From January 2013 to March 2015, we sampled seven field sites using three active mosquito-trapping techniques (two different aspirators and a sweep net) and the stationary CO2-baited CDC mosquito light trap to determine mosquito capture efficacy for each technique. Sampling occurred in four suburban backyards and three dog kennel facilities near Gainesville, FL, USA; species collection and relative abundance were measured. A total of 32 species and 70,090 individual mosquitoes were collected, including a new record for Alachua County, Florida, Aedes hendersoni (Cockerell). The dominant (>5% of total capture) mosquito species collected during the study included Aedes atlanticus (Dyar and Knab), Aedes infirmatus (Dyar and Knab), Anopheles crucians Wiedemann, Culiseta melanura (Coquillett), Culex erraticus (Dyar and Knab), Culex nigripalpus Theobald, and Uranotaenia sapphirina (Osten Sacken). The CDC trap captured the most species (29), followed by large aspirator (28), small aspirator (26), and the sweep net (23). All dominant species were captured with each sampling technique. Excluding Wyeomyia mitchellii (Theobald), all subdominant species (1-5% of total capture) were collected with each sampling technique. Future sampling should consider the utility (e.g., large numbers are readily collected) and limitations (e.g., personnel requirements) of aspirator collections when designing field-based mosquito sampling projects, especially those in residential areas or those focused upon species captured.

Machine-learning-based calving prediction from activity, lying, and ruminating behaviors in dairy cattle.

The objective of this study was to use automated activity, lying, and rumination monitors to characterize prepartum behavior and predict calving in dairy cattle. Data were collected from 20 primiparous and 33 multiparous Holstein dairy cattle from September 2011 to May 2013 at the University of Kentucky Coldstream Dairy. The HR Tag (SCR Engineers Ltd., Netanya, Israel) automatically collected neck activity and rumination data in 2-h increments. The IceQube (IceRobotics Ltd., South Queensferry, United Kingdom) automatically collected number of steps, lying time, standing time, number of transitions from standing to lying (lying bouts), and total motion, summed in 15-min increments. IceQube data were summed in 2-h increments to match HR Tag data. All behavioral data were collected for 14 d before the predicted calving date. Retrospective data analysis was performed using mixed linear models to examine behavioral changes by day in the 14 d before calving. Bihourly behavioral differences from baseline values over the 14 d before calving were also evaluated using mixed linear models. Changes in daily rumination time, total motion, lying time, and lying bouts occurred in the 14 d before calving. In the bihourly analysis, extreme values for all behaviors occurred in the final 24 h, indicating that the monitored behaviors may be useful in calving prediction. To determine whether technologies were useful at predicting calving, random forest, linear discriminant analysis, and neural network machine-learning techniques were constructed and implemented using R version 3.1.0 (R Foundation for Statistical Computing, Vienna, Austria). These methods were used on variables from each technology and all combined variables from both technologies. A neural network analysis that combined variables from both technologies at the daily level yielded 100.0% sensitivity and 86.8% specificity. A neural network analysis that combined variables from both technologies in bihourly increments was used to identify 2-h periods in the 8 h before calving with 82.8% sensitivity and 80.4% specificity. Changes in behavior and machine-learning alerts indicate that commercially marketed behavioral monitors may have calving prediction potential.

Influence of breed, milk yield, and temperature-humidity index on dairy cow lying time, neck activity, reticulorumen temperature, and rumination behavior.

The objective of this study was to compare weekly mean lying time (LT), neck activity (NA), reticulorumen temperature (RT), and rumination time (RU) among 3 breed groups, milk yield (MY), and temperature-humidity index (THI). Cows (n = 36; 12 Holstein, 12 crossbred, and 12 Jersey) were blocked by parity group (primiparous or multiparous), days in milk, and MY. Lying time, NA, RT, RU, and MY were recorded and averaged by day and then by week for each cow. For study inclusion, each cow was required to have 10 wk of LT, NA, RT, and RU data. Maximum THI were recorded and averaged daily. Mean (±SE) days in milk, LT, MY, RT, RU, NA, and maximum THI were 159.0 ± 6.0 d, 11.1 ± 0.1 h/d, 28.7 ± 0.5 kg/d, 38.8 ± 0.0°C, 6.4 ± 0.1 h/d, 323.8 ± 3.8 activity units, and 56.5 ± 0.6, respectively. The MIXED Procedure of SAS (SAS Institute Inc., Cary, NC) was used to evaluate fixed effects of breed, MY, parity, THI, and their interactions on LT, NA, RT, and RU with cow nested within breed as subject. All main effects remained in each model regardless of significance level. Stepwise backward elimination was used to remove nonsignificant interactions. The interactions of breed × parity group and maximum THI × parity group were associated with RT. Increasing THI coincided with increasing RT. Least squares means LT for multiparous cows was significantly greater than LT for primiparous cows (11.4 ± 0.3 and 10.5 ± 0.5 h/d, respectively). Least squares means NA for primiparous cows was greater than for multiparous cows of all breeds (372.1 ± 10.9 and 303.4 ± 7.8, respectively). The CORR Procedure of SAS was used to evaluate relationships among RT, RU, LT, NA, and MY. Rumination time was positively correlated with MY (r = 0.30) and negatively correlated with LT (r = -0.14). Reticulorumen temperature was negatively correlated with MY (r = -0.11). Rumination time was positively correlated with NA (r = 0.18) and negatively correlated with LT (r = -0.14). Lying time and NA were negatively correlated (r = -0.43). Neck activity was positively correlated with MY (r = 0.14). Lying time was negatively correlated with MY (r = -0.25). Milk yield was associated with RU, which may be related to cows with greater MY also having a greater feed intake. Lying time increased and NA decreased with increasing parity, which may be effects of social hierarchy, where primiparous cows are more susceptible to being pushed away from the feed bunk and freestalls. Milk yield was positively associated with RU. Greater milk production requires greater feed intake, which may result in longer RU than for low-yielding cows. Lying time decreased as milk yield increased. The behavioral and physiological differences observed in this study provide new insight into the effects that breed, parity, MY, and THI have on cows.

Short communication: Measuring feed volume and weight by machine vision.

Individual dairy cow feed intake is closely related to the health and productive output of each cow, with healthy cows generally eating more feed than unhealthy cows. Incorporating the use of an automated system to monitor feed consumption for each cow may be beneficial for dairy farm management. This study examined the use of an inexpensive 3-dimensional video camera to measure feed volume, from which we derived feed weight. Proof-of-concept testing was conducted to determine the effectiveness and capability of the machine vision feed-scanning system and its possible use in feed intake monitoring. Such systems are ideal because they do not impede the workflow of the farm or interrupt feeding behavior. This is an improvement over existing systems that are labor and cost intensive. Our conducted experiments involve measuring feed volume at known weights, up to 22.68 kg, with the resulting volume and weight values analyzed by means of linear and quadratic least squares t-test regression analysis. The effects of feed positioning in the bin and near-range sensor limitations were also examined. The results showed that an estimation of feed weight from 3-dimensional scan of volume measurements could be made to within 0.5 kg of the physically measured feed weight using a digital scale. Future efforts will focus on extending this work to active bunks with multiple cows eating throughout the day and testing total mixed rations of varied composition.

Behavioral and physiological changes around estrus events identified using multiple automated monitoring technologies.

This study included 2 objectives. The first objective was to describe estrus-related changes in parameters automatically recorded by the CowManager SensOor (Agis Automatisering, Harmelen, the Netherlands), DVM bolus (DVM Systems LLC, Greeley, CO), HR Tag (SCR Engineers Ltd., Netanya, Israel), IceQube (IceRobotics Ltd., Edinburgh, UK), and Track a Cow (Animart Inc., Beaver Dam, WI). This objective was accomplished using 35 cows in 3 groups between January and June 2013 at the University of Kentucky Coldstream Dairy. We used a modified Ovsynch with G7G protocol to partially synchronize ovulation, ending after the last PGF2α injection (d 0) to allow estrus expression. Visual observation for standing estrus was conducted for four 30-min periods at 0330, 1000, 1430, and 2200h on d 2, 3, 4, and 5. Eighteen of the 35 cows stood to be mounted at least once during the observation period. These cows were used to compare differences between the 6h before and after the first standing event (estrus) and the 2wk preceding that period (nonestrus) for all technology parameters. Differences between estrus and nonestrus were observed for CowManager SensOor minutes feeding per hour, minutes of high ear activity per hour, and minutes ruminating per hour; twice daily DVM bolus reticulorumen temperature; HR Tag neck activity per 2h and minutes ruminating per 2h; IceQube lying bouts per hour, minutes lying per hour, and number of steps per hour; and Track a Cow leg activity per hour and minutes lying per hour. No difference between estrus and nonestrus was observed for CowManager SensOor ear surface temperature per hour. The second objective of this study was to explore the estrus detection potential of machine-learning techniques using automatically collected data. Three machine-learning techniques (random forest, linear discriminant analysis, and neural network) were applied to automatically collected parameter data from the 18 cows observed in standing estrus. Machine learning accuracy for all technologies ranged from 91.0 to 100.0%. When we compared visual observation with progesterone profiles of all 32 cows, we found 65.6% accuracy. Based on these results, machine-learning techniques have potential to be applied to automatically collected technology data for estrus detection.

Stall cleanliness and stall temperature of two different freestall bases.

The objective of this study was to describe the differences in freestall cleanliness and stall temperature between a barn with Dual Chamber Cow Waterbeds (DCCW; Advanced Comfort Technology, Reedsburg, WI) and a barn with rubber-filled mattresses at the University of Kentucky Coldstream Dairy Research Farm from January 18, 2012, to May 3, 2013. Stall cleanliness was measured twice weekly (n=134) by the same 2 observers using a 0.91 m×0.91 m wire grid containing 128 equally sized rectangles (10.16 cm×5.08 cm). This grid was centered at the rear portion of the stall; a rectangle that was visibly wet or had any amount of feces present was defined as a dirty rectangle. Weekly stall temperature (n=66) was measured by the same observer during a.m. milkings in the same predetermined stalls. Feces and wet sawdust were removed from the stalls before stall temperatures were acquired. Temperatures were obtained using a handheld thermometer at 30.48 cm above the stall base as determined via dual laser measurements. Stall temperature was measured on the front, middle, and back of the stall first with clean sawdust and then with the sawdust removed from the stall and wiped clean with a towel. Daily temperature-humidity index (THI) was calculated using Kentucky climate data calculated through the University of Kentucky College of Agriculture via a data logger, located 5.63 km from the Coldstream Dairy Farm. Stall cleanliness was not different between the DCCW barn (26.09±0.89 rectangles) and the rubber-filled mattress barn (23.70±0.89 rectangles). Mean THI throughout the study was 64.39±0.82. Stall temperature was different among THI categories. Temperature-humidity index categories 1 (coldest), 2, 3, and 4 (warmest) had THI ranges of 22.94 to 50.77, 50.77 to 64.88, 64.88 to 78.75, and 78.75 to 101.59, respectively. Stall temperatures (°C; least squares means±SE) were 2.26±0.30, 8.86±0.30, 15.52±0.30, and 20.95±0.30 for THI categories 1 to 4, respectively. Stalls with rubber-filled mattresses had a lower temperature (°C) than DCCW with least squares means±SE of 10.52±0.21°C and 13.29±0.21°C, respectively. The DCCW were probably significantly warmer because water holds heat well. The DCCW may have more of a heat-insulating effect compared with rubber-filled mattresses.

Functional coupling of oxygen binding and vasoactivity in S-nitrosohemoglobin.

S-Nitrosohemoglobin (SNO-Hb) is a vasodilator whose activity is allosterically modulated by oxygen ("thermodyamic linkage"). Blood vessel contractions are favored in the oxygenated structure, and vasorelaxant activity is "linked" to deoxygenation, as illustrated herein. We further show that transnitrosation reactions between SNO-Hb and ambient thiols transduce the NO-related bioactivity, whereas NO itself is inactive. One remaining problem is that the amounts of SNO-Hb present in vivo are so large as to be incompatible with life were all the S-nitrosothiols transformed into bioactive equivalents during each arterial-venous cycle. Experiments were therefore undertaken to address how SNO-Hb conserves its NO-related activity. Our studies show that 1) increased O(2) affinity of SNO-Hb (which otherwise retains allosteric responsivity) restricts the hypoxia-induced allosteric transition that exchanges NO groups with ambient thiols for vasorelaxation; 2) some NO groups released from Cys(beta93) upon transition to T structure are autocaptured by the hemes, even in the presence of glutathione; and 3) an O(2)-dependent equilibrium between SNO-Hb and iron nitrosylhemoglobin acts to conserve NO. Thus, by sequestering a significant fraction of NO liberated upon transition to T structure, Hb can conserve NO groups that would otherwise be released in an untimely or deleterious manner.

Isolation, culture, and characterization of rat lung microvascular endothelial cells.

Highly pure primary cultures of rat lung microvascular endothelial cells were obtained from peripheral lung tissue using a combination of selective culture strategies. The cells had a characteristic morphology consistent with an endothelial origin and were positive for a number of endothelial cell markers, including uptake of fluorescent acetylated lactate dehydrogenase, binding of the lectin Bandeiraea simplicifolia I, and positive immunofluorescence staining with two endothelial cell monoclonal antibodies. The cells behaved as microvascular endothelial cells using an in vitro angiogenesis assay. This isolation method provides a simple method for culturing the pulmonary microvasculature of the rat and these studies support the idea that endothelial cells from different vessels exhibit phenotypic heterogeneity. This method should prove useful for studying specialized endothelial cell function and differentiation in vitro.