A unit of Lasting Forests Sustained forests; sustained profits evolving since March 30, 1999

I labored under the premise that if computer models for resource issues could be built, then data would become available to fill in the boxes, to be the system input. I also preached that the program or model should be developed first, by experienced people, then the data collected based on analyses done about the sensitivity of the model to each type of data. Why collect come data at high cost when it contributes little or nothing to changing the final results (such as dollars from wool or bushes of clean soybeans)? I never got a model built first; funding groups wanted to get busy in the field. Another reason for failure (I should have seen it) was that I never got all the needed data - and never would!

Even the simplest functional population model requires at least sex (2 categories), age (3 categories), weights (2), reproductive status (2), average young, and breeding age. Therefore, solid numbers about 26 things are needed for such a model to produce a solid estimate. (Of course, there are many reasons for modeling not discussed here.) The chances of knowing 26 things about any population , even those few that are intensively studied, are very small. The reason is that the costs are high and they keep coming. The numbers change before they can be summarized and entered into the model.

I'm not complaining. I'm reporting a perceived failure and I want to suggest what I plan to include in my future efforts. Perhaps some will join me. I plan to continue to model, to use the computer as a way to "think through" the enormously complex and complicated wild animal damage problems we have. I plan to continue to write simple programs to help analyze data (for example, the daily catch, not delaying until the end of the month or the end of the project). I'll make more general models (for example, a population estimate, plus and minus 10 percent; a likely proportion of females, plus and minus 5 percent; and a likely birth rate, plus and minus 5 percent). This is not very sophisticated analyses, but it is tedious and very inefficient to do over and over . . . and therefore it won't be done and decisions will not be informed by the results. We're in the business of deciding and getting better at it. The computer can help. I can suggest what are minimum data and what data we should get first, given the need and likely cost of doing so.

Perhaps I was brought-up in the wrong era, the era of well-funded science and great confidence in science. That has changed and so the future seems more clear than in the recent muddled period. The elements of the new strategy are:

1. Model
2. Use available data and past studies
3. Respect expert observations
4. Use best estimates plus or minus a reasonable amount
5. Use data on areas, temperature, precipitation, etc. (the abiotic factors) to refine general principles (e.g., how animal weight varies with latitude)
6. Narrow areas of study so that data collection is manageable
7. Confine data collection to the minimum indicated by models, then expand in increments only as it seems necessary
8. Concentrate system performance measures that can be readily valued in the marketplace, then expand to include other outputs or factors only to the extent that they cover the risks or limits of uncertainty about the monetary estimates.
9. Feedforward is essential. Go beyond forecasting and prognosticating such as those by the World Future Societyand make changes today that reflect confidence in the things foreseen.

This Web site is maintained by R. H. Giles, Jr.
Last revision January 17, 2000.