Sustainability recording traits

Introduction to ICAR sustainability traits

The purpose of ICAR sustainability traits is to provide a harmonized approach to assess the sustainability of dairy herds. By providing a common definition of these traits, we encourage organizations that are involved in milk recording, breeding or any other way of data recording in dairy herds to develop tools to support farmers to increase the sustainability of their dairy herd.

The traits have been selected and defined by a group of ICAR related experts. As definition of sustainability itself the group has used the definition provided by the SAI platform (https://saiplatform.org/):

“Sustainable agriculture is the efficient, long-term production of safe, high-quality agricultural product, in a way that protects and improves the natural environment, the social and economic conditions of the farmers, their employees and local communities, and safeguards the health and welfare of all farmed species.”

It is regarded not to be ICAR’s role to standardize the make-up of Sustainability Indices. The weight of the various traits is a matter for the members or countries themselves to decide. Therefore, ICAR does not provide a sustainability index, but lets the user make a choice which traits to include in their own sustainability index. A selection of traits can be used to create an index that fits the data available and the specific circumstances in your organization or your country.

With this list of traits ICAR aims to identify the key traits in recording that effect sustainability, to provide definitions of these key traits and to harmonize measurement methods of these key traits.

ICAR sustainability traits are selected in such a way that they cover the most important aspects of the performance of the herd regarding sustainability. The traits have been defined in a way that they generally reflect data collected over a 365-day period in one herd. Data collected during a one-year period is more stable to influences on animal performance due to geography, seasonal calving, environmental impact related to weather conditions, herd size fluctuations etc.

Definitions of traits might differ according to how the used data is measured. Some traits are based on 365-day counts of the number of cows present in the dairy herd. Other traits are based on snapshot data (for example test-day average days in milk).

The list contains several categories of traits:

Feeding and production
Fertility
Health
Longevity and culling
Young stock

The list of sustainability traits can be found below in Table 1 as short list with just the name and category. Different colours are used to distinguish the different categories, these colours have no particular meaning. The list of sustainability traits can also be found in Appendix 1 as detailed list with the definitions of the traits. Appendix 2 of this Section contains prediction equations for feed intake, feed efficiency and methane for Dairy Cattle.

We recommend users of this list of traits to select one or more traits per category and to combine these traits into a sustainability index suitable to their national system. The weight per trait could be determined by each user. The sustainability index could be made available to members of your organization to support the sustainability of their herd or to proof sustainability or product quality to e.g. dairy processors.

Acknowledgements

This document is the work of the ICAR Sustainability Task Force. ICAR gratefully acknowledges their contribution. Members of the ICAR Sustainability Task Force were Tone Roalkvam (Chair, Tine, Norway), Martin Burke (ICAR, the Netherlands), Fabian Bernal (DeLaval, Sweden), Christa Egger-Danner (RinderZucht, Austria), Robert Fourdraine (Dairy Record Management Systems, USA), Birgit Grendl-Gredler (Wageningen University, the Netherlands), René van der Linde (ICAR, the Netherlands) and Débora Santschi (Lactanet, Canada).

List of ICAR sustainability recording traits

Table 1. List of ICAR sustainability recording traits.
Number	Trait	Category
1	Age at slaughter (beef cattle)	Feeding and Production
2	Average Days in Milk	Feeding and Production
3	Body weight	Feeding and Production
4	Daily gain	Feeding and Production
5	Dry Matter Intake	Feeding and Production
6	Energy Corrected Milk	Feeding and Production
7	Feed efficiency	Feeding and Production
8	Methane Emissions	Feeding and Production
9	MUN /Urea rates in milk	Feeding and Production
10	% Cows with functional BCS	Feeding and Production
11	Apparent Pregnancy Loss Rate	Fertility
12	Average Days Open	Fertility
13	Average Calving Interval	Fertility
14a	Non-Return Rate 56 days	Fertility
14b	1^st Service Conception Rate	Fertility
15	Pregnancy Rate	Fertility
16	% Cows culled due to reproductive problems	Fertility
17	% Cows with fertility disorders	Fertility
18	Average Somatic Cell Count	Health
19	Chronic infection rate	Health
20	Dry Cow Cure Rate	Health
21	Fresh Cow Infection Rate	Health
22	Selective Dry Cow Therapy Rate	Health
23	% Cows culled due to udder health	Health
24	% Cows culled due to lameness	Health
25	% Cows culled due to other disorders/diseases	Health
26	% Cows with FPR < 1 at first test day	Health
27	% Cows with FPR >1.3/1.5 at first test day	Health
28	% Cows with lameness	Health
29	% Cows with mastitis	Health
30	% Cows with subclinical metabolic issue	Health
31	Age at culling (dairy cattle)	Longevity
32	Average Daily Production of culled animals	Longevity
33	Average Lactation Number	Longevity
34	Average Lifetime Production of culled animals	Longevity
35	% Cows died ≤ 60 days in milk	Longevity
36	Age at first calving	Young stock
37	Young stock EBV ranking	Young stock
38	Young stock sire EBV ranking	Young stock
39	% Female young stock involuntary culled	Young stock
40	% Calves born dead	Young stock
41	% Female calves with diarrhea	Young stock
42	% Female calves with respiratory diseases	Young stock
43	% Mortality of female calves until 90 days	Young stock

Appendix 1 : List of definitions of traits to assess sustainability at herd level

The list of sustainability traits can be found below in Table 2 as detailed list with the definitions of the traits. Different colors are used to distinguish the different categories, these colors have no particular meaning.

Table 2. List of ICAR sustainability recording traits.
Number	Trait and category	Formula
1	Age at slaughter (beef cattle)
1	Feeding and Production	The average age at slaughter (AAS) is calculated as the slaughter date minus the date of birth of all animals that are slaughtered during the past 365 days. To be expressed in days or months (days/(365.25/12)). Date of slaughter and date of birth needs to be known.
2	Average Days in Milk
2	Feeding and Production	Days in milk is defined as date of test day minus date of calving. N = number of test days in the past 365 days. M = number of cows in the milking herd each test day. The annual average days in milk (DIM) is calculated in two steps. Step 1: calculate per test day the average DIM and the number of cows in the milking herd [excluding dry cows]. Step 2: take the total of all test days of number of cows * average DIM on each test day and divide this by the sum of all cows on all test days in the past 365 days.
3	Body weight
3	Feeding and Production	The average body weight (BW) is the sum of the BW per day of all cows in the milking herd on each day during the past 365 days divided by the sum of all cows in the milk herd on each day in that herd during the past 365 days.
4	Daily gain
4	Feeding and Production	Average daily gain is the average of the slaughter weight minus the start weight divided by the age at slaughter for young animals. Start weight is either measured or a certain standard (breed specific) value is taken. For lactating animals it is defined as the weight at slaughter minus the weight at calving divided by the date at slaughter minus the date of calving.
5	Dry Matter Intake
5	Feeding and Production	The average Dry Matter Intake (DMI) is the sum of the DMI in kg per day of all cows in the milking herd on each day during the past 365 days divided by the sum of all cows in the milking herd on each day in that herd during the past 365 days. If Dry Matter Intake cannot be measured, prediction equations can be used (see Appendix 2). To note: If information to calculate energy intake (or protein intake) is available, this should be considered (see Appendix 2).
6	Energy Corrected Milk
6	Feeding and Production	N = number of test days in the past 365 days. M = number of cows in the milking herd each test day. The average Energy Corrected Milk (ECM) is calculated in two steps. Step 1: calculate per test day the average ECM and the number of cows in the milking herd [excluding dry cows]. Step 2: take the total of all test days of number of cows * average ECM on each test day and divide this by the sum of all cows on all test days in the past 365 days. Formula’s to calculate energy corrected milk can be found in Section 2 of the ICAR guidelines.
7	Feed efficiency
7	Feeding and Production	The average feed efficiency is the sum of the energy corrected milk of m cows in the milking herd on each day during the past 365 days divided by the sum of the feed intake of m cows in the milking herd on each day during the past 365 days. Data should be adjusted for stage of lactation (to the same stage of lactation across herds, e.g. 150 days) of the cow. To note: If information on feeding ration is available, this should be included to increase the prediction (see Appendix 2).
8	Methane Emissions
8	Feeding and Production	Average methane emissions are the sum of the methane emissions of m cows in the milking herd on each day during the past 365 days divided by the numbers of cows (m) in the milking herd on each day during the past 365 days. If methane emission cannot be measured, prediction equations can be used (see Appendix 2).
9	MUN /Urea rates in milk
9	Feeding and Production	N = number of test days in the past 365 days. M = number of cows in the milking herd each test day. The average MUN rate is calculated in two steps. Step 1: calculate per test day the average MUN rate and the number of cows in the milking herd [excluding dry cows]. Step 2: take the total of all test days of number of cows * average MUN rate on each test day and divide this by the sum of all cows on all test days in the past 365 days.
10	% Cows with functional BCS
10	Feeding and Production	The percentage cows with functional BCS is calculated as the number of cows with BCS within a desirable range (between 2 and 4 on 1-5 scale) scored during the past 365 days divided by the number of cows with BCS scored during the past 365 days. Cows can be included multiple times.
11	Apparent Pregnancy Loss Rate
11	Fertility	Apparent pregnancy loss rate (PLR) is defined as the number of cows with two or more breedings and the second or later breeding at least 65 days after the previous breeding [where second breeding occurred in the last 365 days] divided by the number of cows that had their first breeding in the period of > 65 and ≤ 365+65 days ago. The assumption is that the cow became pregnant on first breeding and then lost it within that 65 day period and was rebred.
12	Average Days Open
12	Fertility	Days open (DO) is the date of last insemination minus calving date. It is optional to include cows that have not been bred and past voluntary waiting period (VWP). Exclude cows with DO NOT BREED code. It is assumed that the average DO is calculated each test day. N = number of test days in the past 365 days. M = number of cows that are included in the calculation each test day. The average DO is calculated in two steps. Step 1: calculate per test day the average DO and the number of cows in the calculation. Step 2: take the total of all test days of number of cows * average DO on each test day and divide this by the sum of all cows that were included in calculating DO on all test days in the past 365 days. The traits Average Calving Interval and Average Days Open are highly related to each other and should therefore be regarded as alternatives for each other.
13	Average Calving Interval
13	Fertility	Calving Interval (CI) is the date of calving for the current lactation minus the calving date of the previous lactation. It can only be calculated for cows with two consecutive calvings (lactations). The average CI is calculated as the sum of the CI of all animals that in the past 365 days have calved for at least the second time divided by the number of animals that in the past 365 days have calved for at least the second time. The traits Average Calving Interval and Average Days Open are highly related to each other and should therefore be regarded as alternatives for each other.
14a	Non-Return Rate 56 days
14a	Fertility	The non-return rate 56 days (NR56) is calculated as the number of animals with their first insemination (ins.) > 56 and ≤ 56+365 days ago and no re-insemination within 56 days after 1^st insemination divided by the number of animals with their first insemination > 56 and ≤ 56+365 days ago. The main difference between NR56 (14a) and FSCR (14b) is that NR56 is calculated based on re-insemination or not and FSCR is calculated based on pregnancy confirmation.
14b	1^st Service Conception Rate
14b	Fertility	The 1^st service conception rate (FSCR) is calculated as the number of animals with their first insemination (ins.) > 56 and ≤ 56+365 days ago and confirmed pregnancy after 56 days divided by the number of animals with their first insemination > 56 and ≤ 56+365 days ago.
15	Pregnancy Rate
15	Fertility	Pregnancy rate can be defined as the percentage of cows reported pregnant and reported open in period > 42 days and ≤ 63 days ago from calculation divided by the number of cows reported open > 42 and ≤ 63 days ago and beyond VWP (i.e. eligible to be bred). The annual pregnancy rate is calculated as the total number of cows confirmed pregnant divided by the number of cows eligible to be bred over the past 365 days. Usually the pregnancy rate is calculated once per week or per 3 weeks.
16	% Cows culled due to reproductive problems
16	Fertility	The percentage of cows culled due to reproductive problems is calculated as the number of cows culled in the past 365 days with main culling reason reproductive problems divided by the average number of cows with at least one calving (dry and producing) present in the past 365 days. In case of more than one culling reason, reproductive problems should be at least one of the reasons.
17	% Cows with fertility disorders
17	Fertility	The percentage of cows having fertility disorders (e.g. silent heat, cysts, metritis, rep ..) is calculated as the number of cows having fertility disorders in the past 365 days divided by the average number of cows with at least one calving (dry and producing) present in the past 365 days.
18	Average Somatic Cell Count
18	Health	N = number of test days in the past 365 days. M = number of cows in the milking herd each test day. The average somatic cell count (SCC) is calculated in two steps. Step 1: calculate per test day the average SCC and the number of cows with SCC available. Step 2: take the total of all test days of number of cows * average SCC on each test day and divide this by the sum of all cows on all test days in the past 365 days.
19	Chronic infection rate
19	Health	A chronic infection is defined as cows having in the same lactation two consecutive test days with SCC ≥ 200.000. N = number of test days in the past 365 days. M = number of cows with two consecutive test days in the same lactation. The chronic infection rate is calculated in two steps. Step 1: calculate per test day the number of cows with a chronic infection and the number of cows with two consecutive test days in the same lactation. Step 2: take the total of all test days of number of cows with chronic infection and divide this by the sum of all cows with two consecutive test days in the same lactation on all test days in the past 365 days.
20	Dry Cow Cure Rate
20	Health	A dry cow cured is defined as cows having a SCC ≥ 200.000 on the last test day of the previous lactation and a SCC < 200.000 on the first test day of the current lactation. N = number of test days in the past 365 days. M = number of cows having their first test day of a lactation on the particular test day and having a last test day in the previous lactation. The dry cow cure rate (CR) is calculated in two steps. Step 1: calculate per test day both the number of cured cows and the total number of cows with their first test day of a lactation on that test day and having a last test day in the previous lactation. Step 2: take the total of all test days of number of dry cows cured and divide this by the sum of all cows with two consecutive test days in the same lactation on all test days in the past 365 days.
21	Fresh Cow Infection Rate
21	Health	A fresh cow infected is defined as cows having a SCC ≥ 200.000 on the first test day of the current lactation. N = number of test days in the past 365 days. M = number of cows having their first test day of a lactation. The fresh cow infection rate is calculated in two steps. Step 1: calculate per test day both the number of fresh cows infected and the total number of fresh cows (with their first test day of a lactation on that test day). Step 2: take the total of all test days of number of fresh cows infected and divide this by the sum of all cows that have calved since the previous test day on all test days in the past 365 days.
22	Selective Dry Cow Therapy Rate
22	Health	The Selective Dry Cow Therapy Rate is calculated as the number of cows that were dried off without antibiotics during the past 365 days divided by the total number of cows that were dried off during the past 365 days.
23	% Cows culled due to udder health
23	Health	The percentage of cows culled due to udder health is calculated as the number of cows culled in the past 365 days with main culling reason mastitis divided by the average number of cows with at least one calving (dry and producing) present in the past 365 days. In case of more than one culling reason, mastitis should be at least one of the reasons.
24	% Cows culled due to lameness
24	Health	The percentage of cows culled due to lameness or other claw health reasons is calculated as the number of cows culled in the past 365 days with main culling reason lameness divided by the average number of cows with at least one calving (dry and producing) present in the past 365 days. In case of more than one culling reason, lameness should be at least one of the reasons.
25	% Cows culled due to other disorders/diseases
25	Health	The percentage of cows culled due to other disorders/diseases (e.g. Pneumonia, Scour, etc.) is calculated as the number of cows culled in the past 365 days with main culling reason other disorders/diseases divided by the average number of cows with at least one calving (dry and producing) present in the past 365 days. In case of more than one culling reason, due to other disorders/diseases should be at least one of the reasons.
26	% Cows with FPR < 1 at first test day
26	Health	This is calculated as the number of cows that have calved in the past 365 days and with a fat-protein ratio <1 during the first test day in that lactation divided by the number of cows that have calved in the past 365 days and a first test day in that lactation.
27	% Cows with FPR >1.3/1.5 at first test day
27	Health	This is calculated as the number of cows that have calved in the past 365 days and with a fat-protein ration >1.3 or 1.5 during the first test day in that lactation divided by the number of cows that have calved in the past 365 days and a first test day in that lactation. The threshold of 1.3 or 1.5 is breed specific (1.3 for Holstein recommended).
28	% Cows with lameness
28	Health	The percentage of cows with lameness is calculated as the number of cows with at least one case of lameness in the past 365 days divided by the average number of cows with at least one calving (dry and producing) present in the past 365 days.
29	% Cows with mastitis
29	Health	The percentage of cows with mastitis is calculated as the number of cows with at least one case of mastitis in the past 365 days divided by the average number of cows with at least one calving (dry and producing) present in the past 365 days.
30	% Cows with subclinical metabolic issue
30	Health	The percentage of cows with a subclinical metabolic issue (ketosis, acidosis, displaced abomasum etc.) is calculated as the number of cows with at least one case of subclinical metabolic issue in the past 365 days divided by the average number of cows with at least one calving (dry and producing) present in the past 365 days.
31	Age at culling (dairy cattle)
31	Longevity	Age at culling (AAC) is calculated as the date of culling minus the date of birth. The average AAC is calculated as the sum of AAC of animals that were culled in the past 365 days divided by the number of animals that were culled in the past 365 days. If the date of culling is unknown, the last date of milk recording could be used as alternative.
32	Average Daily Production of culled animals
32	Longevity	Average daily production (ADP) of culled animals is defined as the lifetime milk production expressed in ECM divided by the age at culling (date of culling minus date of birth). The ADP is calculated in two steps. Step 1: calculate the sum of lifetime milk production expressed in ECM divided by the age at culling of all animals that were culled in the past 365 days. Step 2: divide the results of step 1 by the number of animals that were culled in the past 365 days.
33	Average Lactation Number
33	Longevity	N = number of test days in the past 365 days. M = number of cows present (both cows in milk and dry cows) in the herd each test day. The annual average lactation number (LN) is calculated in two steps. Step 1: calculate per test day the average LN and the number of cows in the herd. Step 2: take the total of all test days * average LN on each test day and divide this by the sum of all cows on all test days in the past 365 days.
34	Average Lifetime Production of culled animals
34	Longevity	The average lifetime production (LTP) is defined as the sum of the lifetime milk production expressed in ECM of all animals that were culled in the past 365 days divided by the number of animals that were culled in the past 365 days.
35	% Cows died ≤ 60 days in milk
35	Longevity	The percentage cows that died ≤ 60 days in milk is calculated as the number of cows that died ≤ 60 days in milk in the past 60 to (365+60) days divided by the number of cows that calved > 60 and ≤ 60+365 days ago.
36	Age at first calving
36	Young stock	Age at calving (AFC) is defined as the date of first calving minus the date of birth. The average AFC is calculated as the sum of AFC of all animals that calved for the first time in the past 365 days divided by the number of animals that calved for the first time in the past 365 days. Exclude animals with unknown birth date or unknown first calving date.
37	Young stock EBV ranking	Ranking of female young stock in the herd on the national genetic index relative to all other herds in the country (e.g. 0-100 scale).
37	Young stock	GI is any (national) Genetic Index describing the Total Genetic Performance of all animal present in the herd at the moment of calculation. Only young stock (all animals without first calving yet) should be included, as lactating animals are already included in other performance traits.
38	Young stock sire EBV ranking	Ranking of female young stock in the herd on the national genetic index of the sire relative to all other herds in the country (e.g. 0-100 scale).
38	Young stock	SI is any (national) Index of the sire of the animal describing the Total Genetic Performance of the sire of all animals present in the herd at the moment of calculation. Only young stock (all animals without first calving yet) should be included, as lactating animals are already included in other performance traits. SI should only be used if the Genetic Index of the animal itself is not available.
39	% Female young stock involuntary culled
39	Young stock	The percentage female young stock involuntary culled after 90 days of age is calculated as the number of female young stock animals that is involuntary culled (for any other reason than anticipated low production) after 90 days of age in the past 365 days divided by the average number of female young stock (older than 90 days of age) present in the herd in the past 365 days.
40	% Calves born dead
40	Young stock	Calves born dead is defined as calves born dead including calves that died within 24 hours. Calves include both male and female calves. The percentage of calves born dead (CBD) is calculated as the number of calves born dead in the past 365 days divided by the number of calves born in the past 365 days.
41	% Female calves with diarrhea
41	Young stock	The percentage of female calves with diarrhea is calculated as the number of female calves that had diarrhea ≤ 90 days of age and that were born > 90 and ≤ 455 (365+90) days ago divided by all female calves that were born > 90 and ≤ 455 days ago.
42	% Female calves with respiratory diseases
42	Young stock	The percentage of female calves with respiratory diseases is calculated as the number of female calves that had a respiratory disease ≤ 90 days of age and were born > 90 and ≤ 455 (365+90) days ago divided by all female calves that were born > 90 and ≤ 455 days ago.
43	% Mortality of female calves until 90 days
43	Young stock	The percentage mortality of female calves until 90 days is calculated as the number of female calves born alive that died ≤ 90 days of age and that were born > 90 and ≤ 455 (365+90) days ago divided by all female calves that were born > 90 and ≤ 455 days ago. Percentage mortality does not include stillbirth.

Appendix 2: Prediction equations for feed intake, feed efficiency and methane for dairy cattle

Introduction

This appendix is part of the ICAR guidelines for sustainability recording traits, to assess sustainability at herd level. In this document, prediction formulas are described for the traits feed intake, feed efficiency and methane. These prediction formulas can be used to estimate values for these traits, in case no measured data is available for these traits.

Prediction formulas for feed intake

TMR Equation (North America based)

Farm Application:

DMI (kg/d) = [6.89 + 0.305 × MilkE (Mcal/d) + 0.022 × BW (kg) + (−0.689 + parity × −1.87) × BCS] × [1 –0.288× exp(−0.053 × DIM)]

Comprehensive:

DMI (kg/d) = [(3.7 + parity × 5.7) + 0.305 × MilkE (Mcal/d) + 0.022 × BW (kg) + (−0.689 + parity × −1.87) × BCS] × [1 – (0.212 + parity × 0.136) × exp(−0.053 × DIM)]

(mean bias = 0.021 kg, slope bias = 0.059, CCC = 0.72, and RMSEP = 2.89 kg),

where parity is equal to 1 if the animal is multiparous and 0 otherwise.

Pasture equation (North America Based)

Farm Application:

Holstein DMI (SE = 0.73 kg/d) was: DMI (kg/d) = 15.36×[1 - e^{(−0.00220×BW)}]

Crossbred DMI (SE = 0.81 kg/d) was: DMI (kg/d) = 12.91×[1 – e^{(−0.00295×BW)}].

Comprehensive:

Holsteins DMI (kg/d) = 15.79×[1 – e^{(−0.00210×BW)}]−0.0820×NDFdv, where NDFdv = (dietary neutral detergent fiber as a % of dry matter) – {22.07 + [0.08714×BW] – [0.00007383×(BW)²]}

Crossbred DMI (kg/d) = 13.48×[1 – e^{(−0.00271×BW)}]-0.0824×NDFdv where NDFdv = (dietary neutral detergent fiber as a % of dry matter) – {23.11 + [0.07968×BW] - [0.00006252×(BW)²]}.

DMI (kg DM/d) equations by Gruber et al. (2004)^[1]

Equation 1 considers the concentrate amount (kg DM) for feeding concentrate separately from forage. It can be used for pure forage diets (0 kg concentrate intake) and be adapted to other feeding systems like partial mixed rations (PMR) and TMR via additional mathematical equations. The methodical approach of including concentrate mixed with forage (PMR or TMR) or of including separately fed forage (kg DM) is explained in Ledinek et al. (2016)^[2].

Equation 5 is especially applicable for TMR. It was evaluated by Jensen et al. (2015)^[3]. It was evaluated by Jensen et al. (2015)^[3], results are shown in Table 3.

Table 3. Evaluation criteria, regression estimates, and significance level of the five models predicting dry matter intake (DM1) in dairy cows fed total mixed ration, evaluated across the 12 experiments.
		Evaluation criteria				Regression estimates
Models	n	MSPE^b	ECT^c	ER^d	ED^e	Intercept (kg DM/day)	Slope (kg DM/day)
NRC	94	3.20	2.07	0.13	1.00	-1.44***	-0.14***
NorFor^g	48	2.32	0.14	0.87	1.30	-0.38*	-0.37***
TDMI	94	2.91	0.01	0.65	2.25	0.10	-0.29***
Zom	94	9.97	2.62	2.78	4.57	-1.62***	-0.56***
Gruber	94	1.37	0.05	0.04	1.28	-0.23	-0.08

^a NRC (NRC, 2001), NorFor (Volden et al., 2011a), TDMI (Huhtanen et al., 2011), Zorn (Zorn et al., 2012^a) and Gruber (Gruber et al., 2004^[1]).^b Mean square prediction error (MSPE; Bibby and Toutenberg, 1977).

^c Error due to central tendency (ECT).

^d Error due to regression (ER).

^e Error due to disturbance (ED).

^f The regression estimates is retrieved from simple linear modelling.

^g Evaluated on 9 out of 12 experiments as the remaining 3 experiments had been used in the development of this model.

* p < 0.05

*** p<0.001

Table 4. Feed intake equations (total dry matter intake, kg DM per day), n = 25.482, Gruber et al. (2004)^[1].
Parameter	Unit		Equation 1	Equation 5
Intercept			3.878	2.274
Effect Country × Breed		FV [G+A]	-2.631	-2.169
		BS [G+A]	-1.826	-1.391
		HF m [G+A]	-2.720	-1.999
		HF h [G+A]	-1.667	-0.898
		FV [CH]	-0.275	-0.315
		BS [CH]	-0.882	-0.593
		HF [CH]	0.000	0.000
Effect of lactation number	n	1	-0.728	-0.658
		2 - 3	0.218	0.236
		≥ 4	0.000	0.000
Effect of DM Model: a + b × (1 – exp(-c×DIM))	d	a	-4.287	-5.445
		b	4.153	5.298
		c	0.01486	0.01838
Regression coefficient for body weight Model: a + b₁×DIM + b₂×DIM²	kg	a	0.0148	0.0173
		b₁	-0.0000474	-0.0000514
		b₂	0.0000000904	0.0000000999
Regression coefficient for milk performace Model: a + b₁×DIM + b₂×DIM²	kg	a	0.0825	0.2010
		b₁	0.0008098	0.0008080
		b₂	-0.000000966	-0.000001299
Regression coefficient for concentrate amount Model: a + b₁×DIM + b₂×DIM²	kg DM	a	0.6962	-
		b₁	-0.0023289	-
		b₂	0.0000040634	-
Regression coefficient for concentrate proportion Model: a + b₁×DIM + b₂×DIM²	% TMR	a	-	0.0631
		b₁	-	-0.0002096
		b₂	-	0.0000001213
Reg.coeff. NEL_Forage	MJ /kg DM	-	0.8580	0.6090
R²	%	-	86.7	83.5
RSD	kg DM	-	1.32	1.46
CV	%	-	7.1	7.9
Correction factor by validation	DMI = a + b×DMI_predicted		0.47+0.930×DMI_p.	0.71+0.920×DMI_p.
G Germany; A Austria; CH Switzerland; HF m HF h medium and high management level of farm

Model 1 (concentrate considered as amount, e. g. as kg DM):

DMI_predicted (kg/day) =

Intercept + Effect breed.country (kg) + Effect of lactation number (kg) + Effect of DIM (kg DMI, non-linear function depending on DIM) + b_body weight (kg DMI/kg BW) × body weight (kg) + b_milk yield (kg DMI/kg milk) × milk yield (daily milk, kg) + b_concentrate amount (kg DMI/kg concentrate in DM) × concentrate amount (daily amount, kg DM) + b_NEL_Forage (kg DMI/MJ NEL per kg DM) × Energy content of forage (MJ NEL per kg DM)

Correction factor by validation: DMI = 0.47+0.930×DMI_predicted

Model 5 (concentrate considered as proportion, e. g. as % of DM):

DMI_predicted (kg/day) =

Intercept + Effect breed.country (kg) + Effect of lactation number (kg) + Effect of DIM (kg DMI. non-linear function depending on DIM) + b_body weight (kg DMI/kg BW) × body weight (kg) + b_milk yield (kg DMI/kg milk) × milk yield (daily milk. kg) + b_concentrate proportion (% of TMR, DM basis) × concentrate proportion (% of TMR) + b_NEL_Forage (kg DMI/MJ NEL per kg DM) × Energy content of forage (MJ NEL per kg DM)

Correction factor by validation: DMI = 0.71+0.920×DMI_predicted

Regression coefficients (b_) for body weight, milk performance, concentrate amount and concentrate proportion are depending on DIM and are calculated using quadratic polynomials (see Table 3). The effect of DIM is described by a non-linear function.

Link to alternative equations based on regions

1. Development and evaluation of equations for prediction of feed Intake for lactating Holstein dairy cows, 1997. D.K. Roseler, D.G. Fox, L.E. Chase, A.N. Pell, W.C. Stone https://doi.org/10.3168/jds.S0022-0302(97)76010-7

2. NRC. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad. Press, Washington, DC tested by https://doi.org/10.3168/jds.2018-16166

3. https://doi.org/10.1007/s11250-022-03275-8

Prediction of voluntary feed intake in NorFor (Nordic Feed Evaluation System)

Feed intake is predicted from animal fill capacity and dietary fill values. In the intake is regulated both physically by the diet and metabolically by the energy demand of the animal. Each individual feed is assigned a basic fill value (FV; section 6.2) and the animal is assigned an intake capacity (IC) expressed in the same units as the feed When predicting DMI, the following general equation must be fulfilled:

IC= FV_intake

where IC is the animal intake capacity and FV_intake is the total feed intake expressed in fill units

The following equation is used to calculate IC in dairy cows:

where IC_cow is the intake capacity of lactating dairy cows; DIM is days in milk; ECM is the energy corrected milk, kg/d; BW is the animal body weight, kg; and a, b, c, d, e, f, g are regression coefficients presented in Table 5.

Table 5. Multiple regression coefficients used to predict dairy cow intake capacity (IC).
	Multiple regression coefficients ¹
Cow category	a	b	c	d	e	f	g
Primiparous large dairy breeds	2.59	0.134	-0.0006	0.55	0.091	500	0.006
Multiparous large dairy breeds	2.82	0.134	-0.0006	0.55	0.091	575	0.006
Jersey primiparous cows	2.25	0.134	-0.0004	0.25	0.110	360	0.006
Jersey multiparous cows	2.40	0.134	-0.0004	0.15	0.110	405	0.006
Icelandic primiparous cows	4.07	0.087	-0.0014	0.65	0.015	400	0.002
Icelandic multiparous cows	4.77	0.071	-0.0013	0.14	0.035	523	0.0013

Cornell Net Carbohydrate and Protein System (CNCPS, Cornell University Model), for lactating cows

DMI (kg/d) = (0.0185 * BW + 0.305 * FCM) * Lag

Where:

BW = Body Weight (kg),

FCM = fat corrected milk (kg/d) = (0.4 * milk + 15 * milk* fat %, Milk = milk yield (kg/d),

Fat = Fat test (%),

Lag = adjustment for stage of lactation: IF(WOL≤16; 1 − exp(−(0.564 − 0.124 × 2) × (WOL + 2.36)));1) and

WOL = Week of lactation.

NASEM equation (the updated NRC model)

DMI (kg/d) = (3.7 + parity * 5.7) + 0.305 * MilkEnergy (Mcal/d) + 0.022 * BW (kg) + (-0.689 – 1.87 * parity) * BCS) * (1-(0.212 + parity * 0.136) * e (-0.053 x DIM))

Where:

Parity = 0 for primiparous and 1 for multiparous cows,

BCS = Body condition score (1 to 5),

MilkEnergy = 9.29 * kg fat/kg milk + 5.85¹ * kg true Protein/kg milk + 3.95 * kg lactose/kg milk, ¹) if total protein is used: replace 5.85 by 5.5.

Prediction formula’s for feed efficiency traits

Residual feed intake (in dry matter, DM)

RFI = FI_observed – FI_predicted

Y = β0 + β1 X1 + β2 X2 + ε

Residual energy intake (in MJ ME or MJ NEL)

REI = EI_observed – EI_predicted

Y = β0 + β1 X1 + β2 X2 + ε

Explanation: Residual traits are mostly defined as difference between an observed and predicted input trait. The input trait can be feed intake for calculating residual feed intake (RFI, in dry matter) or energy intake for calculating residual energy intake (REI, in MJ ME or MJ NEL). The prediction of feed or energy intake considers at least milk performance, metabolic body size for maintenance requirements and body reserve change. Either a standardized formula is used [REI = Energy intake_observed minus Energy intake_predicted; predicted energy intake includes the requirements for milk, maintenance, body reserve change (and pregnancy) calculated based on the energy system of the country]. Or REI and RFI are defined as residuals of a regression model. As regression models and thus the regression coefficients vary very much in scientific literature, the comparability is reduced.

Feed efficiency

Feed efficiency = ECM (kg) /FI (kg DM)

Explanation: Efficiency is defined as ratio between output and input. Feed efficiency is herd average energy corrected milk (output) divided by herd average feed intake (input). Only milk producing cows are included. Data should be based on one year, in any other case, data should be adjusted for stage of lactation of the cow.

Energy efficiency

Energy efficiency = Lactation energy (MJ LE) / Energy intake (MJ ME or MJ NEL)

Explanation: Efficiency is defined as ratio between output and input. Energy efficiency is herd average energy in milk (output: lactation energy, LE) divided by herd average energy intake (input). Only milk producing cows are included. Data should be based on one year, in any other case, data should be adjusted for stage of lactation of the cow.

Energy efficiency should be used instead of feed efficiency if possible. Efficiency traits based on energy consider the energy content of diet. They do not rate herds or animals as poor or well performing as diet quality is lower or higher. Additionally they can be adjusted to body reserve change as known from residual traits. The shorter the calculation period, the more important is the consideration of body reserve change and of stage of lactation, because the mobilization of body reserves and a (too) low or late regeneration of body reserves fake a high efficiency.

Methane Emissions

Simplistic (IPPC Tier 2)

eCH4 (g d−1) = DMI (kg d−1) × 18.5 (MJ kg−1 DM) × Ym)/55.65 (MJ kg−1 eCH4), where Ym = 3.0% when dietary concentrate proportion is ≥90%, otherwise Ym = 6.5%

Comprehensive

P. Escobar-Bahamondes, M. Oba, and K.A. Beauchemin, 2016. Universally applicable methane prediction equations for beef cattle fed high- or low-forage diets. Canadian Journal of Animal Science. 97(1): 83-94. https://doi.org/10.1139/cjas-2016-0042.

Table 4. Methane prediction (g d-1) equations for beef cattle developed in this study.
Dataset	Equation ID	n	Equations	eCH₄ (g d^-1)	RMSE	P
High forage
Original high forage	[HF-OR]	123	eCH₄ = 715 (±11.45) + 0.12 (±0.03) x BW + 0.10 (±0.01) x DMI³ – 244.8 (±56.44) x fat³ x (NDF-ADF)² + 0.1 (±0.00)	156.4	27.0	<0.01
Monte Carlo high forage	[HF-MC]	100305	eCH₄ = 25.9 (±0.54) + 0.13 (±0.001) x BW + 145.4 (±1.31) x fat + 10.3 (±0.16) x DMI³ - 27.4 (±0.20) x (starch:NDF)	149.6	34.0	<0.01
IPCC (2006) Tier 2	IPCC 2006	123	eCH₄ = [DMI x 18.5 (MJ kg^-1 DM) x (6.5 x 10)] / 55.65 (MJ kg^-1 eCH₄)	156.1	-	-
Low Forage
Original low forage	[LF-OR]	34	eCH₄ = -26.4 (±20.17) + 0.21 (±0.04) x BW + 30.1 (±11.83) x CP - 70.5 (±25.48) x fat² + 10.1 (±5.12) x (NDF-ADF)³	98.3	22.2	<0.05
Monte Carlo low forage	[LF-MC]	27364	eCH₄ = -10.1 (±0.62) + 0.21 (±0.001) x BW + 0.36 (±0.003) x DMI² - 69.2 (±1.65) x fat³ + 13.0 (±0.45) x (CP:NDF) - 49 (±0.07) x (starch:NDF)	95.2	11.2	<0.001
IPCC (2006) Tier 2	IPCC 2006	34	eCH₄ = [DMI x 18.5 (MJ kg^-1 DM) x (3.0 x 10)] / 55.65 (MJ kg^-1 eCH₄)	89.6	-	-
All databases
Original complete database	[AL-OR]	194	eCH₄ = -35.0 (±17.03) + 0.08 (±0.03) x BW + 1.2 (±0.14) x dietary forage content - 69.8 (±14.4) x fat³ + 3.14 (±0.36) x GEI	154.9	154.9	<0.05

Link to alternative equations based on regions

1. https://doi.org/10.3168/jds.2011-4439

2. https://doi.org/10.1080/09064702.2013.851275

3. https://doi.org/10.3168/jds.2012-6095

↑ ^1.0 ^1.1 ^1.2 Gruber L., Schwarz F.J., Erdin D., Fischer B., Spiekers H., Steingass H., Meyer U., Chassot A., Jilg T., Obermaier A., Guggenberger T., 2004. Vorhersage der Futteraufnahme von Milchkühen – Datenbasis von 10 Forschungs- und Universitätsinstituten Deutschlands, Österreichs und der Schweiz [Feed intake prediction in dairy cows based on the data of 10 German, Austrian and Swiss research institutes and universities]. Proceedings of the 116th VDLUFA-Kongress, Sept 13–17, 2004; Rostock, Germany, 484-504.
↑ Ledinek M., Gruber L., Steininger F., Fuerst-Waltl B. Zottl K., Royer M. Krimberger K., Mayerhofer M., Egger-Danner C., 2016. Efficient Cow – estimation of feed intake for efficiency traits using on-farm recorded data. Animal Science Day, Ptuj, Acta agric. Slov, Supplement 5, 71-75.
↑ ^3.0 ^3.1 Jensen L.M., Nielsen N.I., Nadeau E., Markussen B., Nørgaard P., 2015. Evaluation of five models predicting feed intake by dairy cows fed total mixed rations. Livest Sci. 176, 91-103.

[:0-1] 1.0 ^1.1 ^1.2 Gruber L., Schwarz F.J., Erdin D., Fischer B., Spiekers H., Steingass H., Meyer U., Chassot A., Jilg T., Obermaier A., Guggenberger T., 2004. Vorhersage der Futteraufnahme von Milchkühen – Datenbasis von 10 Forschungs- und Universitätsinstituten Deutschlands, Österreichs und der Schweiz [Feed intake prediction in dairy cows based on the data of 10 German, Austrian and Swiss research institutes and universities]. Proceedings of the 116th VDLUFA-Kongress, Sept 13–17, 2004; Rostock, Germany, 484-504.

[2] Ledinek M., Gruber L., Steininger F., Fuerst-Waltl B. Zottl K., Royer M. Krimberger K., Mayerhofer M., Egger-Danner C., 2016. Efficient Cow – estimation of feed intake for efficiency traits using on-farm recorded data. Animal Science Day, Ptuj, Acta agric. Slov, Supplement 5, 71-75.

[:1-3] 3.0 ^3.1 Jensen L.M., Nielsen N.I., Nadeau E., Markussen B., Nørgaard P., 2015. Evaluation of five models predicting feed intake by dairy cows fed total mixed rations. Livest Sci. 176, 91-103.

[1]

[2]

[3]

Anonymous

Search

Section 22 – Sustainability recording traits

Namespaces

More

Page actions

Contents

Sustainability recording traits

Introduction to ICAR sustainability traits

Acknowledgements

List of ICAR sustainability recording traits

Appendix 1 : List of definitions of traits to assess sustainability at herd level

Appendix 2: Prediction equations for feed intake, feed efficiency and methane for dairy cattle

Introduction

Prediction formulas for feed intake

TMR Equation (North America based)

Pasture equation (North America Based)

DMI (kg DM/d) equations by Gruber et al. (2004)^[1]

Link to alternative equations based on regions

Prediction of voluntary feed intake in NorFor (Nordic Feed Evaluation System)

Cornell Net Carbohydrate and Protein System (CNCPS, Cornell University Model), for lactating cows

NASEM equation (the updated NRC model)

Prediction formula’s for feed efficiency traits

Residual feed intake (in dry matter, DM)

Residual energy intake (in MJ ME or MJ NEL)

Feed efficiency

Energy efficiency

Methane Emissions

Simplistic (IPPC Tier 2)

Comprehensive

Link to alternative equations based on regions

Navigation

Navigation

Wiki tools

Wiki tools

Anonymous

Search

Section 22 – Sustainability recording traits

Sustainability recording traits

Introduction to ICAR sustainability traits

Acknowledgements

List of ICAR sustainability recording traits

Appendix 1 : List of definitions of traits to assess sustainability at herd level

Appendix 2: Prediction equations for feed intake, feed efficiency and methane for dairy cattle

Introduction

Prediction formulas for feed intake

TMR Equation (North America based)

Pasture equation (North America Based)

DMI (kg DM/d) equations by Gruber et al. (2004)[1]

Link to alternative equations based on regions

Prediction of voluntary feed intake in NorFor (Nordic Feed Evaluation System)

Cornell Net Carbohydrate and Protein System (CNCPS, Cornell University Model), for lactating cows

NASEM equation (the updated NRC model)

Prediction formula’s for feed efficiency traits

Residual feed intake (in dry matter, DM)

Residual energy intake (in MJ ME or MJ NEL)

Feed efficiency

Energy efficiency

Methane Emissions

Simplistic (IPPC Tier 2)

Comprehensive

Link to alternative equations based on regions

Navigation

Wiki tools

Page tools

DMI (kg DM/d) equations by Gruber et al. (2004)^[1]