Hi Raphaël

I am not sure to understand your question.

If most French dairy farmers cull a cow for each heifer calving, that does not mean they do at random, but according to some criteria such as lameness, mastitis, reproduction failure, poor milk production, so probably culled cows are the less profitable as dairy cows. Culling rapidly cows and replacing them with heifers is a way to improve quickly the genetic merit of the herd. More over a 4-5 year cow is probably easier to sell for meat than a 10-year one.

To my opinion, the risk to get mastitis, lameness, reproductive failure and so on is a characteristic of each farming system, so modifying the culling rules does not seem to be the better approach. It would be more efficient to address the causes of lameness and so on, and, why not, to model the bio-economic interest of changing the farming system (for ex shifting from permanent stabling to pasture, or from corn silage to hay…).

Once again, the average parity of culled cows is an output of the farming system, not a lever as you suggest with "the optimal parity of culled cows"

If you look at longevity as a heritable trait, then the primary factor is whether the cow stays in the herd for the first 60 days of first lactation. The sire proofs focused on longevity use primarily first-lactation records and this seems to be  highly reliable. So a model could use national breed data to look at why first lactation cows leave their herds in the first two months. Generally, first lactation cows comprise 25 to 35% of the herd, and this is the P1 is the mode for parity. In modeling this, one needs to use national data with some probability estimates for losses at various stages of 1st and 2nd lactation. Actually, mastitis, lameness and reproductive failure ARE NOT characteristic of farming systems, but are more related to level of management and level of production. At least in the USA, higher producing herds generally have lower mastitis and lameness, therefore can cull lower producers more quickly.

Sir,

I am working with HF X Kankrej crossbred herd at Livestock Research Station, Navsari Agricultural University, Navsari, Gujarat, India. We are also facing same problem with this herd. Therefore, we brought pure Kankrej cattle from another research station of Gujarat. This animals looked master of longevity. They are performing well even in 10th lactation. Mastitis is not much problem. Moreover, this animals are having good habit to sit on dry surface only which prevents teats from infection.

I recommend to add some blood of Kankrej animals to HF animals. The lactation yield of good Kankrej is near to 3000 litres with 4.5% of fat. 

I looked at heritability estimates for the traits that you mentioned. All are very low, so these traits are primarily a function of herd management. Therefore, it is not possible to predict one herd's performance from another herd's data on these traits. Basically you need to use records from previous years in a herd to predict future performance. As managment improves, the risks will change. One could use national data and look at various levels of these traits among herds and create some probabilities based on the national data. For example, what is the mastitis culling rate for herds with an average SCC of 50,000 versus  herds with an average SCC of 350,000? Alternatively, you can create a spreadsheet model that includes various levels of all of these traits and the culling that occurs. Please note that high SCC is associated with lower fertility and lameness is associated with lower fertility.  So there is some confounding among these traits.

Thanks for your answers,

The aim of the model is to know if the right decisions of culling are taken in the herd.

For exemple, if a breeding has a culling rate of 40% without big problems in the herd (mastitis, lameness and reproductive failure), why the culling rate is 40% ? What will be the consequences of changing the culling rate from 40% to 35% on the herd (demography, milk quality, reproductive performances, economic performances …) ? How many heifers have to be reared ? For an average cow in the herd, how many liters of milk have to be produce to absorb the rearing cost of the heifer, of operating expenses and overheads of his career ?

We want, with our model, explore the consistency of the herd management and propose improvements of the management in the farms we visit.

We will use the model to compare the management of the herd with an optimal management given by the model. We have chosen to study the culling management and the rearing management because it can be improve a lot in our farms.  That’is why I’m searching for solutions to estimate the probabilities of a cow, per parity, to have problems (to be culled for biological reasons). We distinguish two reasons of culling : biologic culling (because of problems, the cow is not able to produce milk anymore ; ore very few or very bad quality) and economic culling (the breeder chose to cull the cow because of an economic reason : for instance, to improve the genic level of the herd, to improve the quality of the milk tank …). We saw that most of our farmers have problems to manage the risk of biological culling. That is why they breed all of their heifers. The consequence is that they cull a cow for each heifer calving while the cow was probably still profitable (or while the cow as not damped his costs). We want to know the economic impact of a better management (given by the model), to sensibilize our breeders on this point.

In addition, working on the management of culling  makes us working on the causes of culling.

I’m already using the method proposed by Mr Britt. My problem is that data about this problem are difficult to get.

If you have more elements, you can react.

Thanks all

Dr. Victor Cabrera at the University of Wisconsin in the USA has a number of models that he provides freely to address issues like this. You can access the models at the link shown.

http://dairymgt.uwex.edu/tools.php