CNCPS Model Biology
Observed vs predicted and some simple stats for CNCPS 6.1 biology.
Courtesy of Dr. Mike Van Amburgh, Cornell University, Dec 2009.
NOTICE: This accuracy is only available from programs based on CNCPS 6.1 licensed by Cornell University.
Models like the CNCPS are mechanistic models where, primarily the rumen, is described by a series of research derived non-linear equations. Instead of having an NEl and a RUP value, feeds are described by their chemical fractions and fermentation characteristics. Then these “pools” are degraded (kd / (kd + kp)) and a microbial yield is predicted. Given differences in animal types, diets, etc., the total ME and MP is calculated for a diet. This is then partitioned to the animal requirements (maintenance, pregnancy, milk, growth, and reserves). What this allows us to do is more accurately define the diets. In today’s economics (and the importance of environmental losses), this increased accuracy allows us to remove ration ‘safety factors’. Proper use of the model has been shown (via research) to:
- improve cow health via
- decreased acidosis (clinical and subclinical)
- decreased laminitis and other foot problems
- decreased metabolic diseases incidence and
- improve income over feed costs
- reduce nitrogen and phosphorus excretion
- improve animal performance (milk, meat, heifer growth, etc.)
- improve milk components
- place a higher farm awareness on forage quality
Where are the founders/developers of CNCPS/CPM now?
AMTS LLC, whose origins are the Fox modeling group at Cornell, obtained a license from Cornell to utilize the CNCPS core biological model. With this license, AMTS can guarantee that the model is implemented exactly as developed.
Contributors to the CNCPS effort have been many throughout the years. While many graduate students came and went, there was a core of people that spent more than 7 years working at Cornell and other places. The following lists the major contributors and where they are now:
-
D.G. Fox, One of founding fathers that continued developement until 2005 as leader of CNCPS project. Retired
- P. Van Soest, One of founding fathers. Degradation and fiber kinetics. Retired
- J. Russel, One of founding fathers, rumen sub-model. Deceased
- C. Sniffen, One of founding fathers, feed library, rumen, integration. Retired, private consultant
- W. Chalupa, Head of CPM project and is retired, private consultant
- R. Boston, CPM development that has now moved to a different project
- M. Van Amburgh, CNCPS developer and contributor, Currently leading model research team and co-chairing CPMb Dairy support team. Faculty member at Cornell
- R. Grant, CPM co-chair and is President RH Miner
- L. Chase, CNCPS and CPM contributor. Faculty at Cornell and does Training and support for CNCPS and CPM
- T. Overton, CNCPS and CPM contributor. Faculty at Cornell and does Training and support for CNCPS and CPM
- L. Tedeschi, CNCPS contributor, 1995-2005. Currently faculty at Texas A&M working on other projects
- V. Durbal, CNCPS programmer at Cornell, 1997-2005 VP for Development and CDO, AMTS LLC
- T.P. Tylutki, CNCPS contributor and developer. 1990-2005 at Cornell President and CEO AMTS LLC. Continuing model development and training/support under AMTS versions
6.1 Biology Changes from Previous Versions of CNCPS and CPM
We have been asked by several people doing evaluations/comparisons between CPM and 6.1 biology (.Dairy) for some guidelines since so many of the numbers have changed.
The files below will answer those questions. It is too big to print on 1 page (it would require 4 or 6 pages to print) but you can easily open the pdf, zoom in to various sections, and work with this document on your screen.
Did we miss things on the list? Probably! Email your requests to add to this evaluation matrix and we will get it done!
Evaluation Matrix
Moving from CPM to 6.1 biology
Model Accuracy
We are often asked about the accuracy of the core biological model, especially when people have had good success using previous versions of the biology.
In the file below, Dr. Van Amburgh calculated stats on accuracy and some herd level experiences from CNCPS 6.1 biology. I think the data clearly speaks for itself. NOTICE: This is 6.1 biology specifically and is only available from licensed groups. As for accuracy of earlier biological versions, it was significantly lower (while still good, it just had bias and more scatter).
Most limiting slides v6.1 biology
