1 Arbel et al. (2001) JDS 84: 600 Preliminary conditions changed? - - PDF document

▶
1
SMART_READER_LITE
LIVE PREVIEW

1 Arbel et al. (2001) JDS 84: 600 Preliminary conditions changed? - - PDF document

Sep 13, 2023 294 likes •387 views

Lecture outline Advanced Herd Management KU-Life 23-10-07 1. Presentation of the problem 2. Choice of method (Monte Carlo Sim ulation) 3. Application of the method Economic consequences of postponed first insemination of 4. Results cows

slide-1
SLIDE 1

1

A A R H U S U N I V E R S I T E T Det Jordbrugsvidenskabelige Fakultet

Economic consequences of postponed first insemination of cows in dairy herds

  • An application of Monte Carlo simulation

Søren Østergaard, Aarhus Universitet, Faculty of Agricultural Sciences, Department of Animal Health, Welfare and Nutrition

Advanced Herd Management KU-Life 23-10-07

Lecture outline

1. Presentation of the problem 2. Choice of method (Monte Carlo Sim ulation) 3. Application of the method 4. Results 5. Evaluation of the applied method

  • 1. Presentation of the technical

problem

Example: The application of SimHerd for the calculations of consequences due to prolonged lactation Educational example: Application of the Monte Carlo simulation

Changing the first drawn conclusions from the test data

Prolonged lactation

  • Fewer calves
  • Longer late lactation

(equals lower yield) + Fewer diseases + Fewer days being dry

Longer calving intervals Later insemination start - planned!

Economical loss – Danish kroner per empty day

4,6 - 7,6 Olds et al. (1979) 3,0 - 22 Pedersen (1981) 21 - 26 Bailie (1982) 2,1 - 7,1 Dijkhuizen et al. (1985) 10,2 Weaver & Goorger (1987) 0 - 3,9 Schm idt (1989) 0,3 - 12 Strandberg & Oltenacu (1990)

Typical reproduction management

Empty days are expensive

Many exam inations

0 - 25 Danish kroner per day

Reproduction objective

90 - 100 empty days in average

Early insemination start (40 days)!

slide-2
SLIDE 2

2

Preliminary conditions changed?

More persistent lactation curves

High yield and low body condition score at drying off

Lower meat/ milk price proportion Higher focus on diseases

Arbel et al. (2001) – JDS 84: 600

  • 1. calf

Older

  • Ins. start

90 & 150 60 & 120 Empty days 128 & 189 110 & 160 Milk 1-10 months +0,9% +2,0% Milk 1-5 following months +4,1% +0,9% Economic effect +5,2% +3,4%

1008 high-yielding cows from 19 herds

Financial analysis employed by Arbel et al.

Income

Milk (T> C) Culled cows (Test design!; NS) Calves born (NS)

Expenses

Feed consumption (Standard numbers for DIM; T> C) Purchased pregnant heifers (one per culled cow; NS) Work expenditure (Fixed number per annual cow; NS)

Period

1½ lactation; Same number of days for treatment and control

Financial analysis employed by Arbel et al.

Problem s

In practice: Will the number of culled cows, calves born and heifers in calf remain similar? Is the higher yield associated with the feed consumption? If the insemination start is postponed, how will the actual distribution of calving intervals be?

Will it differ between herds with different reproduction efficiency?

Does the chosen study period reflect the herd result?

… the problems – as to principles

Cow level vs. herd level

The state of a cow depend on:

The former state Transition probability to the present state (stochastic)

Herd production strategy (between herd variation) Condition of the rest of the herd (mechanistic system)

Presentation of the technical problem – summary:

The consequences of the postponed insemination start calculated at the herd level based on new estimates at the cow level

slide-3
SLIDE 3

3

  • 2. Justification of method

Monte Carlo simulation is chosen because:

A new recommendation based on experimental test results not including essential correlations at herd level Farm studies demand many farms performing the test design over a longer period (very expensive) SimHerd is chosen because it already exists

’Possible’ alternatives

Partial budgeting (Stochastic) Dynamic programming

The history of SimHerds

Developped by Jan Tind Sørensen (DJF) in 1989-91 Scientific tool Advisory tool interplaying with HerdView in 1991-92 SimHerd II including a disease module in 1998 SimHerd III including an even more detailed disease module in 2003 SimHerd IV including more details on mastitis Different special editions regularly updated

Applications of the SimHerd, examples:

  • Influence of the sickness level (1988)
  • Mortality and disease occurrences (1991)
  • SimHerd I + strategies of feeding, reproduction and renewal

(1992)

  • Length of empty period (1993)
  • Strategies of renewal and reproduction (1995)
  • BVD (1995)
  • Application of the TMR-1 (1996)
  • Grazing intensity (1997)
  • Organic vs. conventional production (1998)
  • Time for the first insemination (1998)
  • Strategies of combating the staphylococcus mastitis (1999)
  • SimHerd II + interplaying with the feed, health and production

(2000)

  • Application of a pen bull (2002)
  • AMS and renewal (2002)
  • Prolonged lactation (2003)
  • SimHerd III + Strategies of controlling milk fever (2004)
  • Strategies of controlling the PTB (2004)
  • Information value of progesterone measurements (2004)
  • Early treatment of mastitis and ketose (2004)
  • Economical values of breeding attributes (2004)

How does the model function?

Simulating the technical and econom ical consequences of the production strategies in a dairy cattle herd ’technical’ – e.g. m ilk yield and disease incidence ’econom ical’ – e.g. DB per annual yield of a cow ’production strategy’ – e.g. increased heat detection

The design of SimHerd

Dynam ic

Simulation of the development in herds over time Time step of one week

Mechanical

Simulation of the yearly results of the herd via ’parallel’ simulation of the single animals

Stochastic

Variation between the cows’ production capacity and the events on cow level simulated with random numbers and relevant probability distribution

Input og output

Input: State variables

List of all the cows and heifers in the herd with parameters for their current state at simulation start

Parameter values for the model

Biological parameters Parameters describing the production system (e.g. barn capacity Parameters describing the production strategies (e.g. feedstuff plan, reproduction and culling strategy)

Output: Technical annual results

Typically 10 years from 10 to 1000 repetitions

slide-4
SLIDE 4

4

The simulation steps

  • A cow in each state

Lactation stage Reproduction state (Heat, gestation, abortion, calving) Disease, death, voluntary and involuntary culling Feed intake Milk yield and growth

  • A heifer in each stage

Age, reproduction, culling, feeding

  • At herd level in each state

Replacement (max. number of cows, cows on culling list, heifer in calf, strategy of buying and selling heifers)

  • At the end of a sim ulation year, the results are saved
  • n a file

Why is SimHerd useful – generally speaking?

The choice of production strategies is of great econom ical im portance for the farmer Calculation of the consequences of the production strategies is complicated if done with exactitude

Interaction between the employed strategies Effect on cow level × number of cows

≠ Effect on herd level

… continued

”All things being equal” obtained via the model

In practice the production strategies will alter within the given test period

Sim Herd can serve as a educational tool demonstrating correlations at herd level SimHerd can point out the missing knowledge and calculate the consequences

  • 3. Describing the procedure

I. Form ulation of strategies/ scenarios II. Parameter estimation

  • If necessary based on published test results
  • III. Programm ing (if required)

IV. Validation

  • Internal (testing if the programme is doing as it is

supposed to)

  • External (Face validation; sensitivity analysis)

V. Sim ulation of experiment VI. Analysis and interpretation of results

Pressure for an increased degree of detail

  • ”But there is a correlation”
  • ”We should include it for safety reason”

BUT

  • Increased complexity makes the validation more

difficult!

’face validation’ is typically the only possibility as an external validation mothod

  • Influence on the present form of SimHerd

Validation is turning problematic

Optimise the complexity

We can predict scenarios with a combination of SimHerd and external calculations Additional state variables are often very expensive SimHerd still being used might indicate a respect for complexity

slide-5
SLIDE 5

5

  • 4. Results

Scenarios with prolonged lactation

  • Simulated using SimHerd
  • Ins. start, days

1 Short 2 Long, 1st 3 Long, all 1st calf Older 70 35 140 35 140 105

Scenarios with prolonged lactation

Unchanged risk of diseases, reproduction, feeding etc.

1st calf Older Milk 1-10 months +1,0 % +1,8 % Milk 1-5 months following +0,9 % ! +0,7 %

Two herds

Reproduction management Average Good Insemination chance % 45 60 Conception chance % 50 50

  • Ins. period – high yielding

189 175

  • Ins. period – low yielding

147 133

A typical Danish herd

120 cows Own breed Yielding level approx. 8500 Feeding includes 3 total m ixed rations Dry period of 7 weeks

  • Min. yield of 12 kg

18 - 20 % involuntary culling Calf m ortality from 8% to 12% Typical disease occurrences

Fewer culled cows

30 32 34 36 38 40 42 44 Tidlig alle Udskudt 1. Udskudt alle

  • Ins. start
  • Udskift. %

Middel God

Repro

slide-6
SLIDE 6

6

Fewer calvings

100 110 120 130 140 Tidlig alle Udskudt 1. Udskudt alle

  • Ins. start

Kælvninger Middel God

Repro

Fewer treatments for mastitis

30 32 34 36 38 40 Tidlig alle Udskudt 1. Udskudt alle

  • Ins. start

Mastitis, laktationer Middel God

Repro

Milk fever - almost unaltered

12 13 14 15 16 17 18 Tidlig alle Udskudt 1. Udskudt alle

  • Ins. start

Mælkefeber, laktationer Middel God

Repro

Lower net return per cow-year

Repro

12600 12800 13000 13200 13400 13600 Tidlig alle Udskudt 1. Udskudt alle

  • Ins. start

Kr pr. årsko Middel God

Less effect on net return per kilo milk

Repro

1,55 1,56 1,57 1,58 Tidlig alle Udskudt 1. Udskudt alle

  • Ins. start

Kr pr. kg EKM Middel God

Is the loss compensated by less work?

Postponed 1st Postponed all Fewer cases Av. Good Av Good calvings 8 10 23 23 mastitis 1½ 2 5 6 inseminations 22 27 68 71 Loss in 1.000 kroner:

  • max. 120 cows

16 20 69 49

  • max. 1 mio kg EKM

1 5 9 5 Extra cows!! 1 1 4 3

slide-7
SLIDE 7

7

Operational management questions

Higher conception rate with postponed insem ination start? Disturbances caused by increased rutting (AMS, bigger herds)? Can the “right” cows be chosen early enough? Is yield and body condition at drying off a problem

Persistency of the lactation curve (AMS; 3 milkings) · Feed system (TMR-1)

Application of a large part of roughage

Conclusion – postponed insemination start

The cow and the herd react in different ways

· Less yield per cow per year regardless of the higher daily yield · Cases of diseases per cow is only a little bit affected

Typically less net return

· Strategy: “only 1st calf” better than “all” · Only a small difference under the milk quota

Assum ption for success:

· Persistent lactation curves · Good reproduction / Enough heifers in calf

  • 5. Evaluation of the methods

suitability towards the problem

Generally suitably

Interaction between the animals in the herd Interaction between the production strategies (herd differences) Includes the barn capacity controlling the extend of the production Is the choice of scenario of the most relevant?

No direct optimization

…