This presentation provides an introduction to precision dairy monitoring. The wide range of opportunities for future dairy management are discussed. Then, the challenges of turning these dreams into reality are covered.

1. Precision dairy monitoring
opportunities and challenges
JEFFREY BEWLEY,
PHD, PAS

3. Precision Dairy Monitoring

4. Management by Exception
0
10
20
30
40
50
60
70
80
90
100
0
6
12
18
24
30
36
42
48
54
60
66
72
78
84
90
96
102
108
114
120
Activity
Time in Hours
Estrus

5. PRECISION DAIRY MONITORING
APPLICATIONS
• Estrus Detection
• Mastitis Detection
• Fresh Cow Disease Detection
• Lameness Detection
• Calving Detection
• Genetic Traits
• Management Monitoring

7. Fatness or
Thinness
MobilityHoof
Health
Mastitis
Respiration
Rumination/pH
Temperature
Milk
content
Heart rate
Animal
position/location
Chewing
activity
Lying/
standing
behavior
Methane
emissions
Feed
intake
What Does a Cow
Have to Say?

8. Milk Yield Rumination Time Lying Time Eating Time
43 pounds 500
minutes
18 hours
300
minutes
Somatic Cell Count
Body Condition
Score
Temperature Progesterone
682,000 cells/mL 3.0 104.0°F 10 ng/ml
© Copyright Showeet.com

9. IDEAL TECHNOLOGY
• Explains an underlying biological process
• Can be translated to a meaningful action
• Cost-effective
• Flexible, robust, reliable
• Simple and solution focused
• Readily available information

10. The Options Are Endless

11. Leg Devices

12. Ear Devices

13. Neck Devices

14. Sleep Monitoring System
• Sleep importance-immune function,
well- being, disease, facilities
decisions
• Develop and test a non-invasive
monitor using an accelerometer
• Measure head and neck
movement to classify
sleep/wake behaviors through
human observation
• 92 to 93% agreement with
human observations Klefot et al., 2016

15. Rumen Devices

16. Calving Devices

17. Parlor and Image Devices

18. RealTime Location Systems

19. Future is in
Image Analysis

20. Feed Intake: 3D Imaging(99%R2)
Shelley et al., 2013

21. Locomotion Example
Zhao et al., 2017

22. Video Behavior

23. Technology Validation

24. • Technologies :
• AfiAct Pedometer Plus
• Cowalert IceQube
• CowManager Sensoor
• Smartbow
• Track a))) Cow
Lying, Rumination, and
Feeding Validation
Borchers et al., 2016

25. Lying Behavior
Technology
Number of
cows
Correlation to visual
observations
(r)1
CowAlert IceQube 48 1.00**
Track a Cow 44 1.00**
AfiAct Pedometer Plus2 48 1.00**
Borchers et al., 2016

26. Feeding Behavior
Technology
Number
of
cows
Correlation to visual
observations
(r)1
CowManager SensOor 46 0.87**
Track a Cow 41 0.93**
Borchers et al., 2016

27. Rumination Time
Technology
Number
of
cows
Correlation to visual
observations
(r)1
CowManager SensOor 46 0.69**
Smartbow 46 0.96**
Borchers et al., 2016

28. Eartag based system-triangulation
Comparison with laser measure device
Root mean square error of (x,y)
1.22 m (SE 1.32 m)
Close estimate of the location of the cow
Smartbow Position
Monitoring
Feed$bunk$
Sec+on$B$
Milking$parlor$
Holding$pen$
6"
Sec+on$A$
X$"
Hay$
Hay$
X$axis"
Y$axis"
31.97 m
18.07m52.39m
45.68 m 77.27 m
34.23m
0
Sensor
No reference wall
No sensor
Wolfger et al., 2014

29. 109 lactating Holstein cows at the
University of Kentucky Coldstream Dairy
Modified G7G-Ovsynch used for
synchronization at 45-85 DIM
Estrus gold standard was verification of
luteal regression and ovulation using
temporal progesterone patterns and
ultrasonography
Visual observation 4X a day for 30min
each for 4 days
All cows equipped with 9 commercially
available precision dairy technologies
Multiple Technology Estrus
Detection Study
Mayo et al., 2015

31. University of Kentucky Research
Disease Status

32. Early Disease Detection
Detect diseases
earlier than with
visual observation
alone
Improve individual
cow treatment
results
Indicate a larger,
herd-level, problem
leading to improved
prevention strategies

33. Hyperketonemia: Increased lying
† P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001Tsai et al. 2016, unpublished

34. Hyperketonemia:Increasedruminationtime
† P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001Tsai et al. 2016, unpublished

35. Hyperketonemia: Higher fat:protein
† P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001Tsai et al. 2016, unpublished

36. Hypocalcemia: Increased lying
† P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001Tsai et al. 2016, unpublished

37. Metritis: Increased lying
† P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001Tsai et al. 2016, unpublished

38. How good are
we at finding
events of
interest?

40. HowManyCowsWithConditionDo WeFind?
Example: 100 metritis events
80 Metritis Events Identified by Technology
20 Metritis Events
Missed by Technology

41. HowManyAlertsCoincidewithanActualEvent?
Example: 100 metritis alerts
90 Alerts for Cows that Have Metritis
10 Alerts for Cows that
Do No Have Metritis

42. TimelineAlert/EventComparisons
Eckelkamp et al., 2017

43. FarmerHandlingofSensorAlerts
33%
50%
16%
Alert accepted and
animal checked
Alert accepted and
animal not checked
Alert rejected
Eckelkamp et al., 2017

44. Healthstatussummaryofcowsenrolledin
theintensivehealthcow checksfrom 1to14
DIM Variable Subclinical
ketosis
Hypocalcemia Clinical
metritis
No disease
n cows diagnosed, cow days 208 176 30 -
Mean ± SD rumination time, h/d 5.49 ± 1.93 5.20 ± 1.77 5.25 ± 2.21 6.08 ± 1.90
Mean ± SD reticulorumen temperature, °C 39.04 ± 1.16 38.99 ± 0.59 39.36 ± 0.56 39.05 ± 1.26
Mean ± SD lying time, h/d 10.37 ± 3.43 10.33 ± 3.94 11.41 ± 3.11 9.51 ± 3.13
Mean ± SD milk yield, kg/d 27.60 ± 8.20 25.09 ± 8.67 25.45 ± 10.65 28.92 ± 7.39
Stone et al. 2016, unpublished

45. Sensitivityandspecificityofruminationtime,activity,
reticulorumentemperature,lyingtime,andlyingboutsoneach
diseaseusingdifferentalertthresholdsfordiseasedetection
Disease Sensitivity (%) Specificity (%) Accuracy
Subclinical ketosis 98 6 94
55 91 57
Hypocalcemia 97 8 55
85 22 55
Clinical metritis 86 52 86
67 85 67
Stone et al. 2016, unpublished

46. Gartner Product Life Cycle

47. How much information do we
really need?

48. Are we focused too heavily
on disease detection?

49. Data Integration

50. DataSync: Interface with
anyone, no charge
• DHI Records
• Parlor Records
• Activity / Sensor systems
www.bovisync.com

51. No Magic Solutions:
The Cow is Still First
Cow People will Benefit
Most

53. CAT5CableisaRaccoonDelicacy

54. WARNING:
Lightning will strike the same
technology twice

55. Arewemeasuringthetargetswe
intendto?
Can’t forget
the need for
third party
validations

56. X ≠ X and Y ≠ Y
Tsai et al., 2016
Means for 135
cows
DIM 1-21

57. Disappearing Data
847 cow days (29%) out of
possible 2898
Tech
3
Tech
2
Tech
1 • 138 cows
• DIM 1 to 21
• 2898 cow days
• 7 technologies
Tsai et al., 2016

58. Other Cautions
• Huge within cow and within herd variation
• Many management factors and environmental
conditions affect these variables
• Some times tags randomly stop reading
• Tag placement is important
• Group/pen changes affect behaviors
• Some cows don’t read the book
• Not all changes are linear

59. TheIntangibleValueof Information

60. 4. What is the policy
for upgrading to new
versions of devices?
5. What are full costs
(hardware, devices,
maintenance, data
storage)?
6. What protocols are
available for handling
alerts?
6 Questions Producers Should Ask

62. 1. Data Integration

63. 2. Understanding Technology
Economics

65. 4.FarmerUserGroups

66. 5. Group or Herd Level Data
• Most useful for within group or within
herd changes
• May be useful for cohort comparisons
• Keep in mind natural variation and lag
• Be extremely cautious comparing across
herds
• Question conventional wisdom

67. Core
Temperature
Control
Center
Fan and
Sprinkler
Control
6.Cow ResponsiveEnvironments

68. 7. Calf Monitoring
Calf feeder
data
Wearables
Physiology?

69. 8. Genetic Evaluations
• May provide information previously
unavailable for genetic evaluations
• New or improved traits (i.e. feed intake,
lameness, BCS, heat tolerance, fertility)
• Improved data accuracy (i.e. yield, fat,
protein, SCC, health traits)
• Synergies with genomics

70. 9. Image
Analysis

71. 10.PowerofMilk

72. Exciting
Opportunities
Realistic
Expectations
Big Data
Moving Forward

73. Questions?
Jeffrey Bewley, PhD, PAS
jbewley@bovisync.com
jbewley@cowfocused.com