Like so many other words in popular use these days, "sustainability" has more political than scholarly value. Scientists and managers grapple with its meaning, trying to find a precise meaning, some measure, some way to recognize when it has been achieved, some standard or target. There is none. The word rolled easily from an oration about "keeping on" in about the same way that is familiar to most voters, at least to those contributing to election campaigns. The word has entered laws, then administrative bodies have become involved, almost as if infected, and regulations and rules have been developed. Requests for proposals have been issued and scientists have lined up for scarce funds. They try to understand the word and, usually failing, give the word their own unique definition so they can proceed along reasonable lines of inquiry. Dominance of one definition over another may be gained temporarily, making research risky in political vague-word days. An accountability clerk my claim, ex post facto, that a scientist did the wrong studies relative to a "wrong" definition. Such is life in the worlds of "pollution", "multiple-use", and "biodiversity", to name but three examples of where wrongness may hang on a definition.

There is the difficulty in understanding "sustainability", a word now used actively in ecology, farming systems, fisheries, forestry, and throughout the "environmental" literature. I can imagine a peer reviewer for a technical paper, a jury member for the scientific community, or some prosecutor in a civil court arguing against presentations made on behalf of the word.

Q: Is a coal mine sustainable?

A: No.

Q: So perhaps the word only applies to the so-called "renewable natural resources?"

A: No; probably more. Renewable is not the same as sustainable.

Q: Perhaps the scale is wrong and instead of a mine we should speak only of the coal of industry. Sustaining it seems a reasonable request but available coal resources are said by many to be limited; even reserves have limits.

Q: Can coal resources be used forever?

A: No. There appears to be no way to sustain the coal industry indefinitely. This implies the need for some time limit, a planning horizon to express "... at least to the year _____." By increasing the price of coal, licensing users, and by using other techniques, the horizon can be pushed far into the future. This is a strategy, an emphasis, but it reflects on the question. It may be that the quality of life will be greatly impaired if we press for very-long-term sustainability of coal use and the industry. The scale does, however, seem more appropriate for the industry than for the mine.

Is the present condition best? If the present condition of an old forest or primitive grassland the way things ought to be? There have been radical changes in plants and animals since the Pleistocene at Peculiar Manor. There is a need to incorporate a concept of acceptable trends with sustainability.

Questions and answers, as above, do not serve us well. Each leads to more of the other. The concept of sustainability arises when there are interruptions in supplies and people begin to realize what the absence of something really means. From the Foreword of "Caring for the Earth: A Strategy for Sustainable Living" (1991), we realize that sustainability is not seen as the opposite of development and that by uniting the words as "sustainable development" an emphasis to both protection and rational use of resources was thought to be given. The phrase created problems because it presented the question of whether it was development that was to be sustained or was sustained merely a modifier? Perhaps "sustain" was to rein-in on what use could be made of resources.

Are we to sustain a specific output, or are we to sustain a set of processes judged to be good? What will be most telling, a continuing system or a stable supply of outputs? (for they are not dependent). An Ecological Society president (1992) advocated sustaining research. Perhaps, since we can never be certain of anything (Chapter ). People need probability-of-survival statements. Is sustainability possible, proper, even morally acceptable in countries where rapid improvements in food production are needed to prevent extensive starvation of increasing human populations? In developed countries, efforts to sustain some industries will limit or destroy others. Is a wasteful, irrational protection of a by-gone function or industry a part of a sustainability policy? Is that the intent? Protection and rational use have been strong components of resource rhetoric for over 80 years.

Discussions abut use and stability are still needed, but some people are bored by the repetition. Some are disgusted by "members" of the world citizen "committee" that have not read the minutes of the last meeting. Perhaps there is more history text than anyone can now master. We cannot sustain an observation, a report on a risk-filled experience, or a problem-producing situation. It is hard to imagine sustaining anything, including a productive ecosystem, a healthful facility, or a conversation about human needs.

We can imagine, in general, what people are discussing when using "sustainable development." They are hopeful of long-term care; reasonable tradeoffs between improving economic conditions and protecting the environment; of reducing risks, of increased certainty. They want development, but it must not threaten species (the structure of the system) or life support systems (the processes and relations of the system) on which people live and work.

I read Serageldin's (1994) rich article "Making Development Sustainable" in Finance and Development. I do no think we want to sustain "development", only a wonderful life for groups of people. Development may be a part of that life. We need to struggle to get to that condition, then maintain it. We want as few costs of development as possible and then conversion to long-term quality-of-life management. What is this condition? How will we recognize it?

People who seek a sustained-yield, or an algorithm for sustainability seek parsimony in a weedy field that has grown tall with mathematical thorns, flowery taxonomy, and semantic vines. Somewhere in there are answers to:

• If I sustain production of a variety of products over 100 years and clear cut the forest, have I succeeded in sustained yield?
• If I stabilize the total yield of five products but the proportions of each vary by as much as 80 percent, have I succeeded?
• If I stabilize the area cut each year but not the volume, do I succeed?
• If I stabilize the volume cut each year but the area cut over (and benefits associated with it like wildlife foods and water yields) fluctuates, do I succeed?
• What if my average production of 20 products is the same each year (never mind the variance)? Success?
• I'm at the low ebb; I want to increase production substantially. A rapid, great increase hardly seems sustaining. Success?
• I'll eliminate the variable products. The ones left are stable, sustainable. OK?
• I'll bring all creatures to a threatened-endangered condition, thus minimum fluctuations. I get great interest on a low budget, and because they are small, they can only fluctuate a little.
• I'll get a good score until one disappears. If that happens, I'll move out of the area ... or explain that my scope of activity was too small and the life form really didn't disappear. Sustained, I'll win in court.

Ecologists are seen as preservationists. Sustaining means preserving. I view myself (and many others) as being in an applied area and as a user of the -ologies, the studies. As a wild-animal-system manager (bushmeat, tourism, hunting, pest damage control, etc.), I use ecological concepts. I hope we can think more seriously of ecosystem management (Chapter ), but for it to be relevant it must be unified eventually with the disciplines of the economists, sociologists, and, hopefully, those who deal with all aspects of esthetics, as well as the unifying topic of energy budgeting in its broadest sense (e.g., the work of H.T. Odum). Ecosystem implies the natural world and places like a forest or a lake. We need to deal with holistic human systems, not ecosystems, with regions as communities, moving beyond watershed or basin studies (which have natural, but largely irrelevant boundaries) for many social, cultural, and resource-analytic and goal-achieving needs.

Sustainability is a political word, so it does little good to try to define it or quibble over it. It is a good word symbol for a complex idea. If I want to sustain resource use or yield from a forest, even profit from a stock portfolio, I do not have to add "development." I simply desire to sustain production, yield, or profit. Sustain is the emphasis in an often-used definition of "sustainable development": that which meets the needs of the present without compromising the ability of future generations to meet their own needs.

We do not want to sustain people living in sub-human conditions in a Dakar border town. We do not want to sustain board feet of timber production in a region where the price has dropped and unprofitable production will bankrupt the single industry. We do not even want to sustain oil production unless that means that it will be done at a tightly regulated price that regulates a sustained flow of income-minus-costs and produces a flow of research and technology innovations. Lemming populations are sustained over the eons and are cyclic. Is such sustainability tolerable; what is meant? How variable is tolerable? What is the future ... a 3-year discount period or 300 or at least 3,000 (until the next small ice age)?

There are problems aplenty in deciding about things sustainable, including what areas, what groups of people, and at what level of the human condition. We need to discuss these and write about them. They are difficult, very difficult questions, and at least full-perception (from all perspectives) is needed, if not leadership. I admire, however, Serageldin's focus on "issues with potentially important operational implications." I have learned, as an example of this, that doing programming will result in new insights, new concepts, and new paths never seen when I started out with an idea (which seems perfectly clear and elementary).

Here in this region, the chestnut trees, Castanea dentata, disappeared. Ecologists cannot describe the effects of the loss of this dominant tree. They surely cannot describe the effects of losing a salamander. The endangered species represent a challenge; we can work on preserving them; but we need to move ahead using them as a modest constraint as we develop comprehensive systems for people. Keeping a few species at a viable population size is one objective among 300. We need, then, to figure the costs of achieving the whole set; the rare plants and animals get only their fair share. And we need to be able to make the argument to those who will listen. The others have a belief certainty that must be dealt with just as society deals with terrorists and other subgroups with a single excessively-weighted objective.

I approved of "sustained yield" until I poked at it with a question like "how will you know when you have it?"; "what's the acceptable evidence for it?"; or "when do you know you are doing management that results in it?" Mercuric, it skitters away. It has to be chased. The questions are not apparent to many of our students. We sketch on the board (Figure 17.1) curve where R is some performance from a forest system. An example is board feet of quality timber or days of recreation. They all agree that the yield has been sustained.

Figure 17.1. A system performance measure, R₁, changing over time is shown in five different patterns. The question of which system is judged as being sustained or having sustainability is discussed in the text.

We draw curve B. They are not so sure. Some say that it definitely is not sustained, some say yes, but the condition is definitely what the law intended. Others add to the graph C and conclude in their imagination that when R reaches b it will probably reach zero, so the production is not sustained. [Frustrating how some people will complicate a simple problem!] Next we draw curve D and we get mixed response again. Production is sustained. It is not very stable but has predictable performance. It is sustained but probably at a level that is desirable. Many books are written about mining and lumber operations that have these so-called boom-and-bust conditions. Curve E, as for D, seems OK to most students, for they are familiar with variation in nature. Curve E gets the most discussion, partially because it best reflects conditions in forests and natural resource systems. There is great variation and fluctuation and shifting trends. In curve F at point a, we can see that the radical fluctuations expected in purely natural or wilderness-like situations are brought under control of husbandry. They are kept within the shaded limits for a long time. This seems to us like what is desired. How to express it succinctly is a problem, but certainly not as big as the problem of getting it done. Problems arise and the production is excessive (at b₁). To some people this seems OK and is sustained yield (but it is not). At b₂ the manager fails again. Any excessive deviation, that beyond the stated limits will produce unacceptable consequences (waste because mills cannot process raw material fast enough, production is followed by drops in supplies, excessive deer that eat crops or newly planted crop trees). Over production in a complex system like a forest (b₁) is as bad as underproduction (b₂) due to bad decisions, forest fires, or other phenomena.

"Sustained" is an idea that is too complex for one word. We can give it a code word like "gamma" but that is not very satisfying. We might try, but to do so we need to try to write it down so we all understand the entire concept. The concept includes "yield" or "development" and this is as much a problem as "sustained." Pattern E is unacceptable! The sharp dips are dangerous, probably bringing short-term unemployment or bankruptcy. We need a regular pattern. That's what sustain implies. We can't have a perfectly smooth line as in A or B. A little fluctuation makes sense. Clearly sustained, just variable, line F is success. Never zero, never a real threat of no development, the pattern is highly acceptable. We don't want stable development; we want a continuous increase without excessive fluctuation well into the next century as shown in F.

I know that "development" means many things; development by some definitions means lost quality of life, higher taxes, and lower-wage jobs. Sustained development in a depression or unwise development "at any cost" while achieving that objective may fail in most other tests. Not yield, not production, not buildings built or land sales ... but quality of life for citizens subject to a set of reasonable, needed constraints, policies, and laws, is what we need. We need to replace the political slogans and easily used phrases with local phrases, e.g., "the zeta strategy" of regional quality-of-life development. Sustaining (whatever that means) development (whatever that means) needs to be replaced. In the future, people need to discuss:

• A 100+ year planning period
• Citizens' objectives (about 300)
• Quality of life based on achieving objectives (the more and the better achieved, the higher the quality)
• An increasing or constant (within ten percent) quality-of-life score (on the graph's y axis; replace Development in Figure 17.1 with quality of life)
• Reducing costs of all types, direct and secondary

I wish economists would work more diligently with some of us on valuation of the above 5 items, a big problem. In normal markets, if one commodity price (value) changes, one or more are likely to change too. There is little sense in "to sustain." There is no single price or value. Almost all are (or are potentially) dynamic. In my Wildlife Management textbook (1978), I outlined over 24 ways to assign value to wildlife, but none addressed the topic of time- and site-specific valuation based on achievement of objectives. I contend that the price or value cannot be known until after an optimization is run. As an example, I achieve a large set of objectives in region A for $2 million. I achieve the same set (but with different values, risks, etc.) for the people in that region in the next year for $4 million. The price changes; for a rational investing person, didn't the value change? What is the value of a deer that eats vegetation, changes evapotranspiration, changes runoff, changes silt, and changes commercial fish production by 0.5 percent, but stabilizes a site for an endangered fish. We want deer, plants, soil, water and two kinds of fish. We state these as needed. There are over 300,000 reasonable ways (many more options) to get the desired results. We do it for the least cost because we state up-front our objectives ... for all people, over 50 years, and keep the 50 sliding forward in a dynamic system.

Objectives

In the first century BC (before computers), people could state their objectives as "more" or "less" or if they wanted stability, they could say "more or less the same." Such statements worked reasonably well when there were great surpluses and few competitors, regulations, or parasitic costs of doing business. Now, more than ever because we have more people with very different standards and wants, we need specific measurable objectives in almost everything we do to improve our decisions, to allow people to invest better, to achieve and demonstrate accountability. We'd like to reward our forest workers for good work. How would you know? By what criteria? What is good work in the woods? The music behind the song phrase "you gotta have heart" is right for "you gotta have goals." People have always had goals and objectives (I tend to use them synonymously) but only AC (after computers) could we really get serious about stating them. The "sustained yield" problem is, at its roots, a we-want-more-or-less-the-same-under-most-circumstances-unless-conditions-change statement. That was OK BC. There have been a few changes.

What is to be sustained? "Forest yield" is an example in current U.S. law. Others discuss "agriculture" itself. A 1992 policy document for the Ecological Society of America describes interest in sustainable ecosystems. Courses on farming taught in a Kenya university emphasize sustainable agriculture. In the past few years "sustainable development" abounds and raises other questions of increasing complexity. What is to be sustained? Where is the emphasis? What is the namable entity by which we would probably recognize success?

I suggest we call this yet-unknown thing a performance measure. We tend to work with systems (Chapter 3) and what we want to describe is something called R, which we can plot as in Figure 17.2 to show the system over time and its progress, hopefully not declining to nothing. The idea of R being plotted is that of a system performance measure. It is one expression of the system, usually inadequate, and usually the best we can "come up with", given the situation. It is often an index because we can never quite capture the total reality of our system. It is an approximation, a major decision parameter. Deciding on the performance measure is a critical decision - one of the first order and potentially perilous. What if we decided our measure was tons of corn produced in a region? Corn is important; corn farming is important. Why not use "corn produced" as a measure? Perhaps there are alternative crops that will grow better on lands now growing corn. Perhaps the price has changed and to continue corn production in this region will cause 50 percent of the farmers to go bankrupt. Perhaps we can sustain tons produced but watch profits dwindle over the next 20 years. Should tons or bushels of corn be the measure?

Sustained yield forestry faces the same issues and questions. Sustained yield of logs from the forest may place and keep a community in poverty. Sustaining a level of harvest for 100 years that exceeds forest growth will destroy a logging community. The desired system

Figure 17.2 A system performance measure R, may fluctuate over time, but when an objective is decided (R*), then managerial controls can reduce its fluctuations (rarely even to zero) and sustain that production.

measure is not waste chips, not logs, perhaps not even energy, but profits from the woodlands subject to a group of constraints relating to long-term quality of life, pollution from processing, and aesthetics. The performance measure is a net concept, not one of physical output alone, of sustained product yield. It is one of sustained conditions valued by people. I continue to seek a good definition of R, the performance measure.

People have complex objectives. It is difficult to imagine that one measure can capture the intent implicit in the concept of a "sustainable system." These indices are in common use. People describe progress in company sales, citizen approval scores, quarterly profits, the Dow index, even national GNP. The are used, though all will admit, with caution. A similar index to be used with caution for the same reasons is needed for people to become seriously and actively involved in sustainable agriculture. Such an index may be developed. Without it, those working in this field may never be able to assert progress, success, or the end toward which their work is directed. Without it, feedback cannot be applied to achieve adjustments and improvements. Without it, "accountability" for funds and investment in sustainability cannot be readily achieved.

I think there can be, should be, a clear statement of what groups of people want from the woods or other wildlands. I call it a system objective and name it R-Star (R*) and we can imagine it being shown by the CEO in a company's annual board meeting (Figure 17.2).

I've seen too many situations in which the report has been "we've doubled our production" when the production unit should have been scrapped. The people in it were working at five percent of capacity and moved to ten percent - big deal!? Similarly deceiving comments are common: "We continued excellent performance as in the past two years" (at an increased rate toward company failure as prices dropped and inflation and costs increased, and union frustration increased!).

If we can ever figure out what we really want from the woods and then put it in the computer and calculate it, then people will work toward improving the score. That only sounds naive because it's never been done, couldn't be done BC, and has to be seen for the future. There is trouble ahead for many people, say development psychologists, cannot or will not conceive of the long-term, more than 50 years. Deciding on a performance measure sounds naive because it is a simple idea with untold complexity like "let's go to the moon."

I'll describe an overall concept of the measure first, then explain the basis for each part of the formulation. Before that, however, note the major detractors. Selecting first topics is part of the complexity. Like describing a fallen sequoia, it is impossible to start in "the right place", one that suits every skeptic of such an idea. The major concepts that need to be addressed before the alternative (of a complex objective and the achievement of which we want to sustain) can be clearly considered are:

1. "Multiple-use" is process-oriented, not output-, consequences-, payoff-, or products-oriented. Even if it were, the formulation remains uncertain. Many people have tried alternative formulations. As I view change at Peculiar Manor I fear that the following is just another one. People argue and waver in their efforts to clarify objectives. There is probably no need! But clear objectives have the power in themselves to improve systems! But political forces are so dominant! But an informed citizenry can dominate politicians! But ... the wildland manager needs to continue personal and collegial debate, reining in on skepticism, and run-away pessimism.

2. Some have tried a "Benefits/Costs ratio equal to or greater than 1.0." This is a government option and benefits like scenery, wildlife, wilderness have been as difficult to describe as costs (not the direct costs of doing business, but the extra ones like pollution, disease, and reduced land value or scenic beauty). The alternative is to use the best estimates of benefits, the best estimates of costs, then to maximize a B/C ratio, working to increase B and decrease C. The index takes on relevance to its value in some previous time; the idea is to increase the value of the ratio.

3. "Produce the most timber-related profit without harming many other factors of the forest" seems reasonable. It is called "constrained maximization", fancy words for "get the most from your key resource, but within limits." The limits can be quite profound. These may be whole resource domains such as watershed, big game, and outdoor recreation. In some areas, recreation is a "key" resource, so the way the problem is formulated for the computer can get turned on its head: maximize recreation, subject to a little profit from timber sale, some profit from wildlife, and no more than X amount of likely erosion in the watershed. You maximize, subject to a set of constraints - which is just a definition of "optimize."

4. You can minimize, subject to constraints too. You can minimize losses in profit, not a bad government objective formulation and one likely to result in a system that is judged to be sustained as long as taxes flow in. In natural systems minimizing sediment in a stream at its mouth can be a demanding algorithm. "Do anything you want to in the watershed, lads, only don't increase the soil particles and sediment in the stream!" If your pay is decreased when sediment increases, you'll probably stay out of the woods except as a manager or monitor.

5. We might add up all the benefits and try to get as many as possible. Maximizing is not a bad idea if you can figure out how to add a bear, a board, a bluebird, and a bubble in a mossy-fern rimmed wilderness spring. Some claim apples and oranges make fruit-salad and we can add spoons of salad, but the pieces still seem to be needed and measured.

6. I think it is possible to name all of the major products of a forest.

Realize, in this complexity, I am trying, yet, to describe what it is that we want to sustain. I symbolize it as R*, then describe how well the present system achieves that. I express the progress of the present system as R. The problems are in the gap between R and R*.

I subtract R* from R to get an index of performance and take the absolute value of the difference.

b = |R - R*|

In the difference between the actual and desired is the dissonance, the problems, the dissatisfaction. Plus or minus, the difference is the topic for the manager - how to reduce it. When an index to benefits is seen as B, and R is less than the magnitude of R*, then a formulation suggesting a conventional score of 100 for perfection can be formulated, namely

B = (1.0 - (|R - R*| /R*)) x 100

When there is no difference, the score is 100. Average scores can be obtained but they are mathematically improper and no value is independent.Most likely a manager will try to get a system improving in a "sustained way" as seen in Figure 17.2. Rather than variance or significant differences between years, the manager will discuss percent changes. During the early period of any system development, one of great disequilibrium (e.g., Heedee 1991), the manager is moving the system as rapidly as feasible to a condition of reasonable equilibrium around the condition defined as R*. Forming the objectives may be the way to bring policymakers and experts together - asking how will you achieve this objective? Then further unity may result from developing the system that shows how they all are achieved together simultaneously. This is the R* paradigm I worked out in Idaho, with TVA, and even on a 40-species wildlife project with Jeff Waldon (1987).

I am of the view that the average reasonable forest land owner who expresses any interest in financial returns from the forest does not want wood, but profit. The desired yield is not pulp or plywood, but profit. (There are many landowners who do not want profit from their land or harvested wood.) We see people arguing for stable production of wood expressed in thousands of cords of wood per year. Plotted, these numbers had better produce a flat or increasing line or the manager is in trouble. Poor person; Poor company! Flat profit!

We can imagine producing fields and fields of tomatoes, deflating the price (below its cost) by producing excessive beautiful fruit, and not even being able to sell the last of the crop because of spoilage. What kind of farmer does this? Why produce into bankruptcy? I once proposed using computer maps and models to follow a boll-weevil control program. I suggested that if the weevil was really controlled, the production would go up, the price of cotton would go down (conventional supply-demand theory), and the control effort would be canceled because expenditures on control would no longer be demonstrably cost effective. Warning! The analysis project was canceled and another one assigned. A surprise? The point, lest it be lost, one expression of the yield from working with forests and resources must not be expressed as wood (whatever the units), not in physical units, but in returns, i.e., all benefits and all costs, a comprehensive high-quality human condition at low cost.

Even recent graduates of forestry schools do not understand how to estimate profits. Few understand that there are optional ways to do so. One of our grandchildren played at tiny-tot soccer for several months before he realized that the objective of the game was to score points by kicking the ball into the net. I think there are a bunch of resource managers running (walking) around out there not knowing the point of the game. You win by making income greater than costs over many years. This is called B/C analysis and trying to get a big B/C ratio. You compare benefits to costs and you win by when B/C is much greater than 1.0. That is what Congress has said about Federal projects (but don't start looking for success or real demonstrations of the concept there!). For the private land-owner, B/C is best used when the benefits are described as B in the above equation and then trying to make B as large as possible and C as small as possible, simultaneously.

Because systems, as I am suggesting here, are so inter-locked, efforts to make air clean reduce the needs to clean water, and thus the costs of the unified effort are less than actions taken separately. With realistic monitoring, adjustments can be made frequently in progressive programs rather than seeking high-cost quick-fix options (and putting them off because the money does not seem available).

I do not want to over-emphasize objectives but given the evidence over my career, they can hardly be over-emphasized. They need detailed work and there are so many elements that a computer is essential for describing and working with the descriptions. Here let me list briefly the pieces that need to be built into the concept of B. Thus how R and R* are defined.

The parts include:

1. responses from different publics or groups
2. a long planning period (say, 50 to 100 years)
3. a sliding horizon, always at least 50 years ahead
4. many objectives, probably more than 50
5. the unit in which each objective is measured
6. the demand (the proportion of units that can be produced over or up to that produced under natural conditions)
7. upper and lower limits set
8. the value or relative importance of each unit
9. the expectation (1.0 minus the probability of failure) of gaining the desired number
10. substitutability (what units of demand can be met by an alternative?), and
11. action in any year if the previous three-year rate of change toward the outer bands exceeds 0.33.

There are legal, enforcement, and variety constraints in the formulation of the objective, but they are rarely extreme. Trying to develop mathematical expression of the above value for B is difficult. Knowing it can be handled readily by modern computers gives encouragement to do a formerly impossible task.

Time

Sustainable connotes time. To sustain something requires the answer to "to what time?" or "for how long?" ... or else the agreement on the assumption of infinity, forever.

Those working with conventional economic theory will typically insist on the criterion of discounted present value for major financial decisions. This requirement typically results in anything of value being discounted to less than $500 if a period of over 50 years is used. For example, in simple discounting of $50,000 at ten percent for 50 years, the value is:

V = V_t/(1 + r)^t

V = $50,000/(1.10)⁵⁰

V = $425

(implying that if $425 was invested at 10 percent interest for 50 years, the owner will have at that time, the end, $50,000). Over 100 years, the discounted value is merely $3.65. (Invest $3.65 and with assumptions, in 100 years you will have $50,000.) One hundred years is not forever. Thirty years, a frequently used limit for discounting and depreciating capital investment, is hardly time to get most trees past the sapling stage.

The stark issue of the appropriate horizon for sustainability is seen as Figure 17.3. Assuming a system can be sustained as at A, then does it stop at B, or begin excessive fluctuations as at C?

A stable cyclic performance is probably viewed as a sustainable system if the period (time between peaks) is short. How short is also debatable. In Figure 17.1 can be seen the types of questions that arise when periods are discussed. In 17.1 E, some people would argue that the system is unstable, so much so as hardly to be called sustainable. Yet the existence does not seem to be in question (no collapse to zero). The system in D looks perfectly sustainable with equal period and amplitude. Figure G may be equivalent to D, but if the period is very long, knowledge about cyclic behavior may be missing. The system in C looks as if it might not be sustainable. Knowing whether it is cyclic is a problem; knowing whether to adjust the amplitude is a problem; knowing whether an adjusted population is more sustainable or has greater sustainability than are those not managed is a key question for those who insist on the goodness and desirability of sustainability.

I recommend a minimum time horizon of 50 years. Five- and ten-year plans or programs will tend to produce unsustainable systems because of excessively short discounting and depreciation computations and will never account for production, quality, or service gained after extended periods (e.g., the later results of planting fruit trees or forests, or building bridges serving expanded populations of workers each day).

Variance and Confidence

Observations made of most systems are expected to vary. That expectation can be called statistical variation, an expression of how much things vary about an average. If average

Figure 17.3. A system is apparently stable in the A zone. Whether it is sustained after B, or beyond C, remains a question for discussion. How low it must get, D, to no longer be sustainable also may be discussed.

sustainability is the desired expression (certainly not a harsh and perhaps an unachievable objective at the exact same performance measure each year), then some expression of the tolerable variability is needed. As indicated elsewhere in this book, the more variable the system, the generally less well known the system and the greater are the risks associated with it.

It is well known that systems with the mean equal to the variance are called poissonally distributed. In ecological and other systems when this condition exists it is viewed as a random distribution. My view is that systems with such variations are at high risk of not being sustainable. Whether they are managed or under control is in doubt. As a minimum standard or criterion, I suggest a Mean/Variance ratio of 2 or greater. As sustainable systems are studied, comparisons are made. Is this system more sustainable than that one? Is this system more sustainable than it was five years ago? These are questions answerable in only a limited way by classical statistical analyses of means and variances. To do so requires ample observations or data points and a decision about appropriate levels of confidence. We may not want to conclude from our analyses that a system is sustainable if it is not. Wishing for decision power is such situations does not make it appear. I recommend comparing three years with t + 1 projected (Figure 17.4). A rate of change of less than 0.33 (plus or minus) probably is ground for optimism. More than three years will be better to use than suggested here, but given past skimpy analyses, this may not be a bad minimum strategy.

Figure 17.4. Comparing periods to decide whether a system is stable or not is difficult. Usually thirty years of data are needed. Rarely are so many available. Here, comparing the present and two prior years to a projected year is suggested.

Many objectives, things to sustain, are said to be impossible to quantify and subject to a process like I have described here. I really believe that by identifying the topics (such as social cohesion and cultural identity), we can find a way that actually allows (cf., constrains) a system to tend to achieve these topics. There cannot be a guarantee; we're not programming people, but we can include the meeting places, the ceremony, the annual rewards, "money for the dance" (which is not a profitable thing to do annually but which sustains the community. "Providing" becomes a constraint on the system. It is identical to painting the building or adding roof thatch - it is part of maintaining life and the community.

I've found with much computer modeling that what appears to be a tough question has a broad range of answers that will work very well. Such analyses can be very useful, very practical. We try to be too precise about sustainability and worry too much about it.