Shared Navigation Interface

Shared Disclaimers Reference Maps Tools Projects Photos Flowers Conferences
Members Crosswords FolkSongs MySpace GoogleVideo Weather Morgue Headlines  Editorials Alerets Links Genesis Cowfree Odds&Ends
Public Domain Photos Morgue MultiMedia Morgue          

Grazing Simulation Model
by: Larry Walker
(Posted to the web 5/6/00)

A few years back before I retired from the Bureau of Land Management (BLM), I ran a bunch of step-wise multiple regressions comparing trends from monitoring with a whole slew of combinations of season, frequency, degree, and so forth of grazing for something over 200 pastures. From that data and analysis, I developed a prototype model that I felt could be a tool for quickly screening proposed grazing systems and red-flagging those that did not appear to be in the ballpark for achieving stated objectives. Had I been able to generate enough interest, it would have been a great foundation for a research grant to some university to further test, refine, expand, and in general develop the modeling approach. Needless to say, the interest did not materialize.

While I no longer have the original data, I do have the model and the statistics in support of the model. You will note in those statistics that only 48 observations are plotted. That is because about 70 of the pastures reviewed did not have adequate information (less than 4 years of data, no discernible trend, etc.). The remaining pastures were averaged by 1% intervals of trend to normalize the skewed distribution of the data. This yielded the 48 observations that were used in the model.

If you have adequate software (Internet Explorer 4.01 or later and the Microsoft Office Web Components as of this writing), you can use the interactive online model. If you can't use the interactive model, a message should display informing you what you need - so you might as well go ahead and try it even if you are not certain about your browser. Otherwise, you can download the spreadsheet templates for Excel, Lotus, and Quattro Pro by right clicking and then saving . You will then need to "unzip" the file after downloading. If you do not have a zip/unzip program, a free trial version of PKZIP is available online for downloading.

The appearance may vary somewhat depending whether you are using the interactive online model, or one of the downloaded spreadsheets. In general, the appearance should resemble what is shown in Figure 1 which is not a functioning model.

Figure 1. Grazing Simulation Model
Cycle (YEARS)


Utilization Target


YEARS per cycle with any use


YEARS per cycle used during May (5/1-5/31)


DAYS per cycle used during June (6/1-6/30)


DAYS per cycle used during Aug-Dec (8/1-12/31)


Predicted Trend


Approximate Odds of Predicted Trend

5 to 1

Predicted average annual change


Minus 1 standard error


Plus 1 standard error


Data entry is done in the first six fields in the second column as follows:

  • Cycle (YEARS) - enter a whole number
    • Enter the number of years it takes to complete a grazing rotation. If use is the same every year, then enter 1.
  • Utilization Target - Enter the percent of the production by key forage species that you plan to remove through grazing each year. With some spreadsheets, you may need to enter this as a decimal fraction.
  • YEARS per cycle with any use - enter a whole number
  • YEARS per cycle used during May (5/1-5/31) - enter a whole number for the number of years in a cycle that there will be any use in May.
  • DAYS per cycle used during June (6/1-6/30) - enter a whole number
    • This is the TOTAL days used within this period during a complete grazing rotation. For example: If a pasture was used all of June (30 days) in the first year of a 3 year rotation, used from the middle of June through July (15 days) during the second year, and not used in June during the final year of the cycle, then you would enter 45 days.
  • DAYS per cycle used during Aug-Dec (8/1-12/31) - enter a whole number
    • This is the TOTAL days used within this period during a complete grazing rotation. For example: If a pasture was used during July and August of the first year (31 days in August) of a 3 year rotation, used during August, September, and October (92 days) of the second year, and used during May and June (0 days) of the final year of the cycle, then you would enter 123 days.

That's all there is to using the model, but note that the model functions at the pasture level. It is not intended for use at the allotment level (unless it is a single pasture allotment) where there may be several pastures used at different times.

The intent of the model is to provide a quick and easy "reality check" when considering grazing systems. It was developed from data for BLM lands in southeastern Oregon and should be applicable to other areas where bluebunch wheatgrass and Idaho fescue, or other cool season bunchgrasses commonly associated with them, are prominent in the present plant communities or were so before deterioration. While my intent was to provide quick and objective reality checks, beyond my personal judgment, on grazing systems being proposed by BLM, the model could be equally useful for private planning efforts.

When using the model, bear in mind that any model can produce highly improbable results when used outside of its development parameters. Furthermore, this model has not been "idiot proofed" to prevent doing things like entering more years of use in a cycle than there are in the cycle, entering more than 100% as the utilization target, and so forth.

Otherwise, have fun and let me know what you think. You can send your comments to <>. If any university types out there would be interested in pursuing further development of the concept, I would be happy to discuss my approach in more detail with them. BLM's monitoring program has produced adequate data to develop models specific to most regions and even stratified by things like elevation, major vegetation types, and so forth. They just don't know how to use the stuff.

I am posting some examples of actual BLM grazing systems and results from checking the systems against the model as time is available. In the meantime, you might note that the data and results in figure 1 represent a typical 3 pasture rest-rotation system for a 6 month season:

  • Year 1 - graze May 1 through July 31

  • Year 2 - graze August 1 through October 31

  • Year 3 - rest

As a side note, I want to say something about "rest-rotation". I have seen a lot of systems billed as "rest-rotation" that would be better described as "deferred-rotation", "rotational rest", "flip-flop" and so forth. As I see it, a grazing system should incorporate two minimum requirements in order to qualify for the "rest-rotation" label.

First, there should be two or more grazing periods that are not repeated or overlaped during a cycle. If, for example, the system calls for grazing a pasture in May and June of one year, there should not be any other years during the cycle when that pasture is grazed in either May or June.

Second, there should be at least 365 consecutive days of rest between periods of grazing.

In case it isn't obvious, at least three pastures are needed in a system to meet these requirements. I remember seeing one BLM grazing system billed as an 8-pasture rest-rotation system that sounded pretty advanced until I looked a little closer. What I actually found was that 4 herds were being used and each herd was doing a two pasture flip-flop. Oh well, I always did have this suspicion that range cons are binary. They have difficulty counting higher than two.