|
Rural System? Just Dreaming …
A For-Profit Conglomerate for
Meaningful Jobs
Healthful Communities
and Improved Natural Resource Management
by Robert H. Giles, Jr., Ph.D.
Professor Emeritus
Virginia Tech, Blacksburg, Virginia
2007
Chapter 4. Design and Rationally Robust Work
Dreaming: How Deep?
How many fish are there? Childhood curiosity
Curious
Fruitful curiosity.
What will you do with the answer?
Do?
What need will it meet?
What time is it?
What is time?
What is the real question? Just dreaming…
| The thought processes and proposed policies and principles behind Rural System's design are expanded and made more practical, emphasizing the dimensions of rationally robust work. |
There are new procedures that will be used in the operation of Rural System and those are part of the continual design process. In this chapter I continue with design but note some basic or lower-level differences. I have said that much of what has been proposed is not very new. In this chapter I suggest the newness is in the way of arranging things into a system and breaking some thought patterns. Within Rural System we have knowledge about where there is knowledge and we know special ways about how to use it. We have an almost-anti-science attitude about getting and using knowledge to make money and to reduce losses.
I've have been taught and have participated in science in the Sputnik era during which science was viewed with national pride and pursued with nearly religious zeal. I've debated 'basic' and 'applied' as if there was more at stake than a budgetary criterion of the National Science Foundation. With colleagues, I have been involved with the 'scientific method' and wrestled with the interplay of deduction and induction (see Chapter 5). I've created models, done 'curve fitting,' and advised people on a wide variety of quantitative questions, some of which could be aided by statistical analysis. This experience has
| Based in part on an article by Giles, R. H., R. G. Oderwald, and A. U. Ezealor. 1993. Toward a rationally robust paradigm for agroforestry systems. Agroforestry Systems 24:21-37. |
There is nothing tight and crisp that I can call our work like a paradigm or theory. The work has a set of characteristics, many interlocked, that together are significantly different than the approaches and techniques used elsewhere. The work starts with the imperative of recognizing "a situation" and moves to achieve a "satisfactory condition." Rejecting the status quo, it moves toward optimization, it having demanded precise objectives. Fuzzy objectives may lead to using consequence tables, the important "consequences" being rephrased objectives. Consequences are estimated from computer models built for GIS maps. They use often-rejected concepts of risk taking, relaxed confidence and precision, greater use of ranges and medians than the average statistic, and the knowledge of equifinality existing within natural and social systems. A dynamic knowledge base is created, managed, and maintained primarily for improving models leading to optimization. The knowledge is then used within decisions for a system to achieve a set of objectives for the 150-year future, all at very low expected costs. In a challenging reversal, operating a system for "making money" (Rural System itself) is seen as the cost of achieving those objectives. The following may help clarify the characteristics of rationally robust work leading to a satisfactory condition.
Anti- Science?
There is need for a strong, sustained effort for gaining and retaining and then using knowledge, parameters, distributions, rules, and procedures that are known with high confidence. I outline the epistemological bases in Chapter 5, and we know that induction and deduction are the cornerstones of science. These two ways of knowing have served people well. They are inadequate and overly simplistic for progress in Rural System and related fields. There is need for rationally robust work, a concept of decision-making and action-taking that is timely, tentative and, in a low-risk, high-influence domain, always accompanied by feedback and timely response to the perceived and believed-in future. It seems irrational to insist that Rural System work (and probably many other related fields of work) exclusively follow the scientific method.
Research, like the good doctor, has an aura about it of objectivity, formality, and rigor, but it is not an aura needed in all fields. Research has solved some problems, given us some advances, and has given many people a useful pattern of thought for over a century. Increasingly, that pattern is being shown wanting. Research is said to answer questions, but it is also said that if you ask the wrong questions you will get wrong answers. Research is said to be descriptive, but of what? There are many questions which cannot be answered by science. A few people say that research is predictive (or should be), but people themselves have plenty to say about what happens in the future. Predictions, themselves, can change the future. There are many, many problems faced for which research has neither the answer nor an approach. Induction, while good, is not sufficient. It has little to provide in knowing the unique or rare event. It is of little service in highly variable situations with few observations. It is infeasible in many situations (e.g., hypothesis: rabies virus inoculation is not always fatal). A substitute is needed, at least an alternative. The needs are conspicuous in rural resource management - throughout the world. We may yearn for research, for the specificity and confidence it seems to give. The hard lesson, not yet learned, is that it is very expensive, takes much time, requires specialists, and after the reports of results filed, risks remain and there are persistent delays between discoveries and their uses. We have not learned that we do not work with simple fruit flies in all cases. We cannot gain the controls upon which belief in science depends. We work with incredibly large, complex and changing systems. They are unique and their every sampling period is unique and they cannot be assumed as uniform as cloned white mice. They are about as predictable as the flight of a flock of pigeons. We are short on money, time, and skills and our risks are evident. Answers are needed. Rationally robust work is badly needed for all of the realms of natural resource management.
It is easy to be hypercritical about anything. I am not a devil's advocate, merely an observer with a little energy left over for suggestions. One problem with research came upon me like a hawk over my tree stand. Suppose there are about 300 important bird species in India. (For now, let us not quibble about the actual number or the meaning of "important"). There are needed about 200 observations about the characteristics or parameters for each bird to complete all entries in a wildlife information system. These 200 items are selected from a much longer list. It is a group of need to know vs. nice to know observations selected by a variety of wildlife workers. Many facts are known for deer among the 200 items but many are not; deer are one of the best known animals. Some factors needed for each species take years of study, others only a brief period. I round off my estimate at a very conservative estimate of one year needed for each observation, and then I suggest an even more conservative $50,000 required to pay and equip a scientist for a year. It includes all travel, rent, equipment, computers, support staff, and salaries but it has never been analyzed exclusively for wildlife research people (probably greater than $50,000.) While several observations will be made in a few days, I assume I can make one official entry in our database per year. The cost of doing all of this is $3,000,000,000! If there were 1000 scientific wildlife researchers, it would only take 60 years. We cannot meet the research needs of the birds of India alone, much less those of the world, by the conventional, accepted research pathways. We have not even mentioned the similar research needs of the mammals, reptiles, amphibians, mollusks … and, oh yes, the fish and, equally as important, the insects, whether we study them as disease vectors, critical food supply for some other animals, or object of specific management, such as for the butterflies and their enthusiasts.
Once there was the notion of "do basic research" and then publish it. It was a rule within graduate schools and the hidden assumption behind it was that one day your findings, in a process unlike that of your own discovery, would be re-discovered and put into practical and good use. In rural systems with many parts threatened and changing, one day may never come. "Irrelevant" is the perfect word for a discovery made for a species that has just become extinct.
In presumably the most logical of all areas, research, I now think I perceive an illogical underpinning. It is illogical for us to continue using the classical, experimental, inductive approach to gaining knowledge about rural resources. Wild faunal resource workers, for example, will never gain the budgets needed, the staffing and expertise, the time, or the requisite use rates of conclusions that are reached. It is irrational for us to proceed in the current classical fashion.
Research, although I prefer the more relaxed "studies," is said to be done to improve decision making and also, strangely, to "support" decisions made, (as in a "decision-support system"). Let me tell you a beaver story, one about rural decisions, their number, complexity, users, and impacts and consequences. Rural decisions do not fit the textbook simplicity (or complexity) of decision making or so-called "decision support systems."
* * *
A Beaver Tale
A forester and laborer drove down a dusty road in late summer. Rounding a sharp corner, they were faced with leaves and brush in the road. "Tree down across the road!" was the observation that almost anyone could make. No decision needed there!
"I wonder how that happened." They both got out of the truck and followed the tree trunk to the base. Any TV-watcher over the age of 4 could tell it had been gnawed through by a beaver.
"I didn't think we had beavers around here." (Within the following parentheses are decision or knowledge base types, then the total count): (Presence or absence - 1)
"What is their local range?" (Local range - 2) "Are we on the edge of their range and are they expanding? I heard that they had been very widespread but that they disappeared with settlement." (National range - 3) (Continental and world range - 4)(Brief history - 5)
"Do you have a saw in the truck?"
"No. It is always a problem deciding what to bring on these trips. I could fill up the truck and still wish for one more thing. (Daily service equipment - 6) I do have an axe."
"Thank goodness! Hey! It's a double-bitted one. Why did you bring that kind and not the single bit? "(Equipment type - 7)
"More versatile, I guess, but I grabbed it from next to the door when I left. What safety issues do we have here?" (Safety instruction and training time allocations - 8)
"None, just stand back! I'm wearing my steel-toed boots." (Requirement to wear safety clothing in the field vs. efficiencies and comfort - 9)
"What kind of tree is that?" (Species identification - 10)
"Does it always grow where it is damp?" (Silvics for 50 species of trees in a region - 11)
"Why do you think the beaver felled it into the road?" (Details of beaver life history and dam building - 12)
"It didn't have the road in mind, I assure you. It was cutting wood for the dam and using it to plug the culvert crossing the road down there."
"That's a small culvert." (Proper culvert size within a watershed - 13)
"How are we going to stop the beaver from cutting more trees?" (Should we try to stop it - 14?) (If yes, how - 15) (If no, consequences - 16)
"We can walk away and assume it is only cutting a few trees and that they are not very valuable and whoever finds the next tree will also have an axe or a saw. To walk is a big decision. "
"The beaver are more valuable that one or two of these kind and quality of trees." (Local stumpage estimates - 17) (Local fur prices - 18) (Local attitude toward wildlife not present but potentially recovering - 19)
"We ought to try to get help from a trapper." (List of trappers - 20)(Trappers with live traps available - 21)
"I wonder what kind of bait they use." (Trapping techniques - 22)
"None. But I worry about growing sentiment against trapping." (Probability of local offense - 23) (In-house policy on trapping as a serious profit-loss reduction operation - 24)
"Where would the trapper take the beaver if they caught one?" (Current range and "non-range" or places where losses would be tolerated - 25)
"Maybe we could make some money by encouraging beavers and managing them for fur, meat, photography, educational tours, and castor…and start worrying about diseases they harbor and then contaminate stream waters."(Total net financial benefit potentials 26)(Disease public health risk analyses - 27)
"What's castor?"(Commercial uses of animal organs - 28)
"The glands at the hind legs. The oil is used as a perfume fixative." (Commercial potential and development - 29)
"There's a lot of talk recently about exporting products. Maybe there is a potential for exporting glands from a well-regulated beaver population managed for fur and other products." (Commercial development and business plan - 30)
"Could we export castor?" (CITES (endangered species, etc.) laws and export and customs laws, regulations, and tariffs - 31)
"Let's not discuss exports before we chop through this tree. How many cuts can we get by with and still roll the logs?" (Local tree weights and efficiency - 32)
"I've seen pictures of elaborate structures that prevent beavers from plugging culverts and thus protect roads from flooding or washout." (History of such efforts and best current practice - 33)
"What if we do clear the tree and trap the beavers. Won't they return? I hear that they migrate upstream. I hear that young are driven out of their homes by their parents." (Restoration ecology - 34)
"Correct. We'll continue to fight them from here on."
"Is this a new and perpetual cost of forestry?" (Forest health, invasive species, etc. - 35)(Forest economics models - 36)
"Only on some areas. Maybe pest work can become like forest fire accounting." (Forest tax law - 37)
"The boss might make more money from his farm and these two forest tracts with the beaver losses than he would if he spent a lot of time and money on trapping and beaver removals." (Total present net return calculations - 38)
"It will take a computer to figure that out. What about these beavers? If we trap them and release then somewhere else, will they return?" (Homing behavior - 39)
"Maybe we could just trap and remove the females." (Beaver external sex characteristics - 40)
"We do not even know how many there are here at the culvert or in the watershed."(Population estimation techniques - 41)
"Maybe we have no worries. This could be like a random event. We can watch and see what happens next. Maybe predators will wipe them out." (Predator-prey population relations - 42)(Loss of former large predators such as lion, bobcat, or wolf - 43)
"They can cut only so many trees. Maybe "Nature knows best."
"Trees cut are not the issue. Trees that die after an area is flooded by the beaver dams are the problem. The potential value of the trees may be the real problem. Beavers work up the banks around their water bodies but the acreages flooded are some of the best tree-growing sites. One family of beaver can kill superior trees over many acres." (Silvics and flooding and riparian area ecology - 44)
"Beavers slow the water, it warms, and good trout water can be reduced in quality.…but small-mouth bass go crazy in such waters." (Trout fishery management - 45)(Trout and Bass tradeoffs - 46)
"I'd say let the trappers have at them&3!33;But maybe there are many trappers and allocating beavers among them can be a problem. In Canada there are allocations of trapping areas made by the government. Would that be necessary?"(Area allocation - 47)
After much chopping and log rolling they sat eating their lunch. (Policy: always bring lunch - 48) a man appears from the forest edge and starts walking down the road.
"What are you doing cutting up my logs?" he said.
The "warm" conversation that followed included property lines, old owners, questionable corners, who owned the water rights, and whether the logs were in the road or on the right-of-way and whose property they really were.
Brought from the truck, the recent GIS map was studied by the three men. (GIS ownership and watershed boundary and stream channel - 49)
They finally departed; the old guy continued to walk down the road.
"Should we have given him a lift?" (Local etiquette - 50)
"Do we have to get these land boundary lines re-surveyed to be sure about whose log that really was?" (Survey or re-survey contract - 51) Who will we contact? Who did the boundary (Contract history - 52) and when is it scheduled to be re-painted? " (Boundary marking scheduling 53)(Contract work or employees - 54) "Are these 'our' beavers and will we be sued by that guy if our beaver's water floods some of his trees?"
"Keep chopping; I don't know! Back to the beavers. If we have 3 kinds of traps, 2 kinds of poisons, 2 systems of trapping, and a treat/not-treat decision, then we have at minimum 24 different options from which we must select. If we do about 7 separate things in various sequences, then the permutations of those 7 things gives us 5040 options. Picking the very best option from among them is tough. It is very easy to be good but not correct. Usually 'close-enough' or satisfying seems to work. The land, in my opinion, has been destroyed by the relentless practice of well-meaning but poor decision making. Hundreds of C- and D-grade decisions, all following after each other, produce a failing system."
"If we knew who should be deciding.… the boss in the big house, me standing here chopping, or that old guy down the road… then the question remains." (The properly designated decision maker - 55)
"What question is that?"
"How would you know what is a good decision, one really close enough to the singular best one? " (The objective function, what to maximize, stabilize, or minimize...the desired net expression - 56). "Then if we knew, we would have to face: How can we get it all together?" (The integrative model for satisfaction - 57)
"Maybe, but accountability is not yet in the decision about how well the decision-made is implemented."
* * *
In rural system management, there are decisions that range from whether to go to work on a day even when having stomach cramps to whether a decision made during such a day may threatened the existence of a form or life or close a mill for an entire region forever. There are at least 57 such decisions (such as those just listed) related to the simple local issue of beaver trapping. Over and over techniques for making decisions have been developed but use has been poor or fragmented. Agencies have created tools and the agency has disappeared before the tool was marketed and benefits gotten from the investment. There have been failures. There have been simplistic systems and excessive systems. There have been educational systems that have failed in application. Field systems created for demonstrations have been costly and not very educational for many people. Many systems have been designed for the wrong decision maker. Many have had unclear objectives. There have been many systems lost when their creators changed jobs or retired. There have been a few evil people at work.
New technology now makes major decision techniques available in the field. Deciding on the decisions to be addressed is important. The funds available, the ideas, the agency, the people who decide what is important … all influence the decision. Techniques for solving hundreds of trivial decisions can be created. Years can be spent in inventing the single decision method for "the big decision" (which may never be made in the blinding winds of globalization.) Some difficult decisions must be made about how to improve decision making. We need the results of an excellent system to create a system.
Maybe we need well-educated people in decision positions undergoing continual education. With encouragement and protection, an integrative precise objective, collaborations, a backup team for information and review, clear applications of feedback, and a feedforward program, things could get better, very fast.
Design to Reduce the Gap between the Present Situation and the Desired One
While some things, like those technological, seem to get better, other things seem to be getting worse. The CIA has been ridiculed, the air transportation system is not as convenient or dependable as it once was, more than half of new marriages fail, and private businesses out-compete the post-office monopoly. The U.S. Forest Service, part of the Department of Agriculture, seems to many people to be in trouble like other natural resource and "environmental" agencies. Once the paragon of federal agencies, it is now under attack for many reasons such as flawed planning, staff conflicts, failed accountability, unresolved regional differences, and failures to use the knowledge base that they have built over the years. A book, Reforming the Forest Service (O'Toole, 1988) has been written as if the need was real and the effort and pain of such writing was worth it. After years of spiraling decent in delays, suits, frustrations, counterclaims, and conflict among practitioners and analysts, the difficulties of decision making within public natural resource management (particularly the U.S. Forest Service) have reached gridlock.
The following components (with the traditional caveats about overlap and limits) create the situation in which classic decision theory has no meaning and little relevance to significant rural resource decisions:
Rather than continuing to add to a long list of dimensions and developing an abstruse argument, I assert from years of experience and observation that most of the above items are true and that even if as many as half were flawed, the conclusion would still be the same. We do have a new situation and it is not subject to classical decision theory or reasoning from science. Scientific management is a misnomer. There is a need for an alternative and the only one on the horizon is rationally robust work toward a satisfactory condition.
The Satisfactory Condition
The rise of environmental interest, while favorable, has had negative unavoidable consequences within the realm of management within which forests, rangelands, the wildlife resource, fisheries, streamside zones, and soil are the major topics. These observations connote problems in resource agencies, but they seem common to most organizations and corporations. I'm pessimistic and respectful of the size of the problem. Bold thought, maybe silly seeming, is needed. Maybe I cannot solve all of the problems, maybe none of them, but I think we can call what we perceive to be a situation and that by analyzing it and applying some creative effort, perhaps some tentative better condition can be created. We have a situation and now we have to find a way out of it, a way to make a major change. Almost any place will do because the present one is too risky and unsatisfying. Call it "bad." Of course "any place" will not "do," will not suffice, so we also have to figure out what will. That's a satisfactory condition. Perhaps synergism can be gained in a small group of similar actions to achieve the new condition. We can analyze our present situation and, if any good, we may prepare the grounds for describing and implementing a way of significantly improving the way natural resource decisions influence the creation of new, more desirable conditions.
Naive people like to look for solutions, even "the" solution, but in very complex situations with long planning horizons, there is no singular solution. Even if one could be found, it will be judged inadequate the next day because conditions have changed. Rather than a solution, we are looking for a condition, a satisfactory condition. It will not be right, or perfect, or even optimum. It will be satisfactory if we work hard, acquire knowledge and build a knowledge base, use available knowledge, and create systems that utilize well things that we now know about the way that complex systems tend to work.
Early on we can reject the quaint phrase: we learn from history that we do not learn from history. History can be a wonderful teacher if we have the ability to hear. We also need a place of order where we can store what we hear and otherwise sense. We get too much noise. We focus on details and miss the messages. "There will be a flood!" This is near calamity, yet we concentrate on depth, flow rates, dollars lost, and other details. We need to sort out the things we now know, things like that floods and fires do occur, trees grow, epidemics occur and people need each other. People help each other. We know many acidity limits on plants, what will poison cattle, and that tomatoes will not grow well under walnut trees. We know a lot!
I remember well a skeptical student noting the impossibility of predicting the leader that emerged from a Vietnam village and turned the tide of the war in an area. He may have been right, but that "leaders will emerge" can be predicted. Viable models can probably predicted the occurrence of such a person. "Leader" is not an unnamed genus; "emerge" is not an unknown verb. The time and place are certainly problems, but in the future, such events will probably be at least as predictable as are flood and earthquake events.
When generalized and modeled and retrieved in conjunction with other things we know, we will achieve our objective. Our objective is to know, not to do research. We discuss the means of knowing and the notion of degrees of certainty in Chapter 5. One part of a potentially-growing knowledge base needs to be from tribals, villagers, and practical folks that have made daily outdoor observations as they have regularly tended cattle, poultry, bees, and their crops.
We have been held for years by the wisdom of the technical literature describing analyzing decisions. It varies, but it usually has the elements of general systems theory sketched in Chapter 6. Typically there are objectives, facts and figures are gathered, they are processed in several ways (from very simple to complex computer means), and in some instant (the tap of the gavel or a registered letter being placed in the mail slot), the decision is made. There may be feedback that improves the decision when it is next made. Major decisions are singular, almost by definition. I now believe, however, that classic decision theory is inappropriate for public rural-related natural resource decisions. While there are similarities between classic decisions and the events within the public so-called "decision arena," I now believe the differences are so great that an alternative analysis is needed.
Situations are not classical decision-making events. Classical decisions are made by a person or a small group. There is a decision maker, a time, a place, limits, and fairly clear objectives (though that is often discussed and debated). Current public decision making has few of these characteristics and now has many other characteristics never addressed (to my knowledge) by decision analysts. While there are many, many public agency decisions, there are even more (on over 60% of US land) on private lands.
N-Dimensions
If two topics such as water and temperature were discussed (as they might relate to tree growth), then we could say that we are discussing a two-dimensional system. We could display it on a 2-dimensional piece of paper, a graph. If we discussed three factors (water, temperature, and light), we might imagine trees responding and being displayed within a box, a 3-dimensional space. The above dimensions of the situation suggest an n-dimensional or many-dimensional volume. It changes in time and differs depending on the region of the people being discussed. The area may change due to shifts in ownership, flooding, and wildfires. It is a changing, n-dimensional entity. Difficult to imagine, the situation can be pictured in an elementary way as an ever-changing, moving cloud or blob. Thinking about a three-dimensional thing is easy; four-dimensional thought is difficult (except for a 3-d object tumbling through time, the fourth dimension). N-dimensional thought is available to a limited few. The natural resource situation typically requires n-dimensional thought … or aids to approximating it … thus help from computers.
The weakness in the footings of the present decision procedures has been presented above in order to begin to understand why an alternative means is needed to arrive at a satisfactory condition within rural systems and within many public natural resource agencies. The needs now are so different and great and the decision making process is so different from classical decision theory as to make that process irrelevant. It is a general situation where current "public participation" in natural resource decisions on private (or public land) is untenable; where alternative risk-taking takes on an imperative; where education for specific situations (not "more, general education") becomes required; where computer aids are used to present alternative and emergent situations; and where ombudsforce is needed to assure that the decision made is implemented. (An un-implemented or un-enforced "decision" is little more than an opinion.)
Understanding the situation seems necessary. An alternative is to ignore the present situation and creatively develop an image of a perfect one, then to compare the present to that one and make changes. That is only a dream if it ignores the power moves within and outside of the agencies (many affecting private land owners), or the strongly held value-system forces at work in rural lands and the urban fringe. It will be only a dream if actions are not informed by those generalized from history. A new condition can be created. I call the work to do so rationally robust work.
The Status Quo
There is another option of course. That option is not to change, to retain the status quo (The existing condition or state of affairs. [Latin; state in which]). This may be necessary if there are no means seen for change, no resources, no creative option. The status quo may be pleasing to some people. A generally bad situation may prevent an agency or individual from doing a particular "good" that is offensive to some group or individual. Some people have been said to "stir the pot" as a strategy to prevent action. Preventing action may be the intent; it can lead to analyses (such as those conducted by the Government Accounting Office) that can lead to major agency changes or their abolition …not intended.
It is difficult in some societies to admit that there are no solutions to a problem or situation and no hope for one emerging. Things may be perceived to be a good as they will ever get. Perhaps that is the case with large, complex problems with a long history and strongly felt needs but conflicting objectives. The natural resource domain seems to have its own breeding ground for problems. Herein, the underlying assumptions are that within rural systems there is just such a problem and that there is no singular solution. That condition, the status quo, however, is not acceptable and thus another condition will be sought. It is unlikely that it will be judged to be good, only judged better than the former condition, and well prepared for the next changes likely to come.
We must delay or go directly to epistemology (the next Chapter 5) and its question of how do we know… whether we have a problem, have a solution, or can fix it with the resources and ideas available within the time remaining. I suggest that we back away from the profound bias of 'scientist' and start with person qua person, then advance to the knowledgeable person, living within a state of tentative certainty grounded simultaneously on several epistemological bases, most importantly the heuristic base. The following are the parts, the major dimensions of rationally robust work for us all, together, on which we may work. Together, the parts become an effective, new, whole way to achieve the desired satisfactory condition.
Problems of Space
There are scant research papers that provide the mammable latitude, longitude, and elevation where studies were conducted. So many phenomena operate in this real three-dimensional space (e.g., electromagnetism, insolation, gravity fields) that additional controls may be gained on the variance that typically is observed from observations in spacees in the field. Besides this subtle point about data analysis, it is possible to begin to make new site-specific prediction. We once created in Virginia a database of about 50 factors in each of 1.1 million, 27-acre (1/9th kilometer) square map cells. (It had no backup system and was destroyed by a political storm. But it has been restored with more factors and greater precision.) Such a data base allowed, for example, computation of the likely impact of a high voltage power line, if it were in place, using 12 dimensions of impact, 42 critical characteristics of the cells, and a 30-year economic expectancy. A Rural System group, when developed, can supply a farmer or rancher information about similar impacts (defensive knowledge for protecting land from invaders of all types), but also about suitable crops, best grasses, likely forest site index, probable runoff, and holdings on request. The intent of such map-cell-specific data bases is to bring to bear, on site, the findings of science to make them relevant to the decision-making tasks of the owner.
We have demonstrated that we have knowledge about and can be very particular, very precise about land conditions. I now believe that the probability of any two spots (say within an area between the lines on a football field (10-m x 10-m), in any place in the world, being alike is almost zero. Thus, places for agriculture, the fishery, and forestry are unique. Because we now have or can cost-effectively create and manage such data bases, and have sufficient computational power, even on desktop computers, we no longer have a genuine need for a gross land statistic or classification category. Even in developing countries, the ability (if not the motivation) to develop such systems cost-effectively is now available. Ease of use increases rapidly. Classification was once needed by the manager who took samples and made maps in order to form general pictures as the basis for making site-specific decisions. Now we have the knowledge of each site, with sufficient detail to assert the uniqueness of each spot on Earth. We do not have to make the reverse trip to generality! Even though we cannot visit every spot in a region or large farm, it is possible to compute in reasonable time and at low cost the characteristics of every land unit, (called the alpha unit and described in Chapter 12) using relative elevation, slope, distance to streams, gross soil texture, past land use, primary land cover, even time that the site is witin shadow each day. If there are only two gross classes used for each of these factors, there are potentially 128 mapable classes. There are not that many map colors or symbols that can be discriminated! Small computers now allow such large numbers of unique classes (descriptors) to be handled. Analyses once unthinkable are now available in tribal centers of developing countries. Because each site and situation is unique and we can know of and return to it precisely because of GPS technology, gross generalizations about conditions on Earth are no longer needed. By more situation-specific work, some risks can be dodged. We must shift from generalized regions and ecological "types" to specific unique map cell studies (disregarding the difficulties for teachers who seek generalizations for their lectures and graphic presentations.) The shift will not occur rapidly, given the historical evidence for changes, but current general knowledge can be used to "fill the knowledge about each cell" and it can rapidlybe improved with models.
Each point or cell on the Earth may be characterized in hundreds of ways. Computers are now capable of storing and retrieving these data and putting them together in the best ways currently known. These are the intricate relations at any site. We can now take a new orientation to the three dimensional spot on Earth and we can produce huge gains in predictive capabilities. There is no way to visit each cell in Virginia for example for research (to do so even for one hour each would take over 60 working years). Idaho has 2.1 times the area of Virginia; there are a few states in between. Scientists have generalized, classified, and clumped data in the past to an amazing degree. There are regions and range maps of all types. Now there exists the technology to dis-aggregate and discriminate. It is time to start the reverse journey.
The spatial domain is not unrestricted. Certain life forms have altitudinal limits. Knowledge of these can be used to eliminate the grossness and unpredictability of many animal and plant range maps. Predictability can be improved by managers preventing certain areas from being used. Land use zoning by people is somewhat related. A new zoning based on prediction is possible. Because we know that certain plants will undergo moisture stress in their lifetime if planted in cell of coordinate x, y, z, then let managers be sure that they are willing to assume the risk of that loss (or pay the total long term costs of failure). Let society be sure pesticide use will not be required in a cell when that cell is near another one in which occurs a highly threatened life form. By such action and containment it is possible to reduce the mismatches in predictions and reduce the large number of alternatives that must be explored in struggles to see the future and prepare for it.
Nearness
A little-acknowledged dimension of land analysis and prescribing uses is that nearby features and forces have more influence on plants and animals in a spot than the on-site factors. Shadows and the presence of water in dry areas are examples. We can use the lessons of landscape ecology to relate "nearness-to" or "distance-from" ideas to an exact site. One alpha unit of land 5 miles from a National Park is a very different piece of land from the identical one inside a Park. A square meter of habitat located within a half-mile of water is different to an animal population located more than five miles from water. Progressively, we will be able to add to knowledge about each site a set of distant, but otherwise influential, factors. There are other factors that are invisible and not present on a sampling site, but we attempt to measure and note these with increasingly more perceptive and accurate technology. These include geomagnetic, solar, and tidal forces. These factors may play leading roles in the conditions or actions of things we now call ecosystems. There will always be other things that are active in our systems, at least within the alpha unit, and we attempt to accommodate them in our measures of statistical variance and to live with the unexplained or so-called random (sometimes even called "mystical") forces.
If site visits to the land are impossible in real time, satellite imagery of only limited usefulness, and funding unlikely to increase substantially, then what are the alternatives for the nation and its scientists? What are the consequences to the land and its inhabitants? Certainly better planning is one answer. Research direction and leadership, a past anathema, will be essential in the energy- and money-short future. Far more attention must be paid to sampling in time and space. No laboratory-scientist will add excess animals to cage experiments having carefully computed sample sizes. No nation can afford limited or excessive research projects; the value of n, the sample size, must be carefully computed. Attention must be given to holistic computer models, particularly simulations that allow planners and managers to ask "what if we change this…? " type questions assuming goal sets as well as certain action proposed on the land. When equations are not known, then subjective probability needs to be used, limits and ranges need to be emphasized and normal central tendency used less than now. In the rural community we need to abundantly use computing with the best current knowledge all within a system with abundant feedback over time. Within this development there is reason to be hopeful about the future.
The Resource Tetrahedron
 |
There are four major aspects of any natural resource (Watt 1973: xi). They can be depicted as being at the four interactive vertices of a tetrahedron. By seeing energy (and/or matter) as having associated weights, risks, and desired or expected quantities (valued energy), the tetrahedron as both a real as well as conceptual space can help unify the salient dimensions of all natural resource and land use issues. The tetrahedron may bring meaning (symbolically at least), order and unification to the chaos of the resource and land use issue. From such organization and clarification, people may gain additional hope. The role of rationally robust work is to develop these mathematics, revise the statistics, and continually unify knowledge.
The Energy-Matter Problems
Not enough effort has been spent on describing and analyzing the net energetics of systems. Adopting an energy metric provides an invaluable aid to modeling (Odum and Odum 1976). Integrating various researchers' work and making tradeoffs and comparisons between quite different concepts can be expedited among these who adapt the metric and become attuned more closely to energy transfer and its loss relations in many systems. We know there will be losses. That is the rule. Reducing them is the rule for modern society.
Occurrence
Dr. Byron Cooper, the late dean of Appalachian geologists, showed me a community water tank placed on a rock outcrop and told me with unusual confidence that the particular rock would fail and the tank be destroyed - but he could not tell when. The people below it lived in ignorance. Thousands of people live on flood plains, fully aware of flooding, wiling to do so with certain knowledge of its occurrence. They do not live in ignorance, only with uncertainty about when floods will occur. There are dozens of similar examples of the mixed personal and social calculus, and Starr (1969) suggested that people make conceptual third-power transformations when dealing with risk, i.e., they are prone to equate hazards to the third power of the benefits, real or imagined. Society has not sorted out these complexities. It probably operates intellectually in a linear domain where the worst imaginable risk is loss of a member of the family. This socio-intellectual state neither justifies nor excuses scientists' snipes at those who create models and cannot match temporal events very well. There is no way to avoid a risky world: uncertainty is one of the immutable laws with which people must live. I think that while risk taking can be investigated, it is ascientific. It is a personal human trait, a function of a physiological, psychological, sociological, theological, and economic milieu. It cannot be observed well directly, only behaviorally. Its expression in behavior can be manipulated. Thus, like assigning weights or expressing preference, assigning acceptable risk levels is a human act, essential, and, while ascientific, is appropriate itself for study.
There are scientific laws and common knowledge,and these form the basis for a belief that occurrence of a class of things can be predicted with near certainty. The time and place predictions are worth working on. I view estimating flood rates as a scientific activity, just as I do predicting weather and the occurrence of solar and planetary events. These are activities dealing with occurrence and at least somewhat with their temporal precision.
The precise details of the future are not needed even if it is possible to know them. It would be a very boring world if such were available. Instead, what are needed are general characteristics of the future, expressions of orders of magnitude, and the near-presence of thresholds of concern. As Starr and Rudman (1973: 360) said in a parallel vein for land use: "While it is obviously not possible to predict the content and time scale of specific technical achievements which may be important in future social change, it may be feasible to see the range of the general characteristics of growth of that societal resource encompassed by the common term 'technology'."
Similar negative comments have been made about biologists' inability to predict micro-events about wildlife models. Could the formation of an anti-hunter group have been predicted when law Q was modified? It could have. At least the option could have been explored, and strategies then developed for dealing with occurrences of high probability. Whether it would occur in a particular area at a particular time or with a particular intensity of feeling implies the existence of more knowledge than is had for even some of the better-known aspects of science. Such knowledge is not achievable at present rates of acquisition of knowledge, with present organizations, or at current funding over any reasonable future period, for example the next 200 years. It is unreasonable to continue to behave as if it could be achieved. An alternative will be suggested.
The Variety Problem
Variety is a general word for variance, juxtaposition, richness, various aggregation indices, and diversity. It is interactive with the above topics and discussion of diversity in the last chapter. Knowledge of it adds another dimension to and thus, increases the potential to predicting and controlling temporal as well as spatial occurrence. It allows such concepts as likely yield and site quality to be quantified.
Modern science tends to be probabilistic and thus can link directly with population theory. Variety or variance is a population characteristic. Inductive science has a role in predicting the future of a population. There is little it can do for the absolutely unique event or the individual. It is far easier to remember that ecosystems are unique than that animals are unique. This premise needs careful handling for it can be misleading. In the same way that every person is said to be unique, every animal is also. Every geographic cell on the Earth's surface is different, by at least one characteristic. Classical experimental procedures generally assume internal similarity, sameness, and work to achieve control over external variables. Abundant computer data storage is now available allowing reasonable study of individual animals and structures, avoiding aggregation into statistics and thus potential loss of information. Individual plants, animals, and ecosystems - even humans - may be allowed to retain their identity and uniqueness in a large matrix. They are assigned a place in a sequenced, scaled, n-dimensional topology. The observed individuals occupy space in a hypervolume.
Subsequent observations must fall within or near the volume. As conditions change over time due to chemical, physical, and human forces, the position in the matrix may change. The future is limited to nearby empty cells, under the assumptions of uniqueness and a largely continuous real world. The options are narrowed; decisions for the future are made with less uncertainty.
 |
Conventional decision making requires an objective. Optimization can occur only after such decision making about such an objective. It has been very difficult to formulate objectives within rural land management circles. There are many reasons Chapter xxx). Without a clear objective, then any solution or set of actions can be argued as satisfactory. With no destination in mind, any trip...or staying at home…is equally as good. "Good" has typically produced the response: "as compared to what?" and the answer, after much discussion, is usually "as compared to this set of objectives."
Some farms are said to be "marginal." They exist on the fence between profitable or not. Being profitable is the objective. One dollar, more or less, determines on which side of the financial fence they may exist. On one side, they fail. Some owners move to the cities. Because the fence edge is so thin, the balance so precarious, it is easy to imagine how that small changes in management, information and risk reduction can move people well past the margin. Perhaps Rural System work can be considered a counter-marginalization effort.
The Procedure
People live in a perceptual or mathematical space. Here, above, it is shown as a 3-dimensional box but it needs to be imagined as a complex volume, much like a many-facet jewel, tumbling along. The arrows suggest that appropriate conditions have been exceeded and are outside of the box. People want to stay within the box. This is what they know, where things are safe, where they know what their parents and their history and culture have taught. This is where their survival skills work. In some cases, the limits are laws. It is illegal to go outside of the box. A simple box has three dimensions, and for people these might be food, housing, and clothing. Outside the box might be inadequate food or poisoned or polluted food. A little pollution may not be too bad, but if it exceeds a threshold, then sickness or even death might result. A reasonable person or group wants to avoid the thresholds, the limits. The closer to the center of the box, the better. The sides of the box are not very precise. Many variables, including the variation in the health and abilities of the people within the group, influence the limits of the box. If a limit is threatened (e.g., effects of a toxicant), but no one knows the exact limit, then it is reasonable to make decisions to avoid coming close to such limits. The limits are fuzzy; the center is safe; avoiding the limits is conservative. The more the dimensions and the more complex the limits, the more like a sphere the space probably becomes (because of the uncertainties at the intersections of the planes and surfaces).
Gaining Stuff for the Knowledge Base
It is easy to understand and appreciate administrative, budgetary, and legalistic reasons why there needs to be taxonomic difference between basic and applied research. Only recently has it become evident how harmful that classification has been to science and to applications of research findings to rural problems.
Science is. It exists, multidimensional but continuous. The fundamental difference between basic and applied is that of when the conclusions which have been reached are applied. Basic research seems to take longer to conduct that other studies and much longer for its findings to be applied, a trivial distinction on a temporal continuum. Taxonomic and administrative problems arise when basic research is quickly applied and so-called applied research findings languish in the shade. There is no longer any meaningful difference between these research taxa; they are artificial and invalid under the rules of nomenclature and should be abolished as intellectually, personally, and organizationally divisive. They are the roots of great ineffectiveness in the scientific community -- especially those dealing with land use questions. In the future, we can stress wholeness, similarity, and generality. Then predictions will be more correctly made.
As an example of the rationally robust work, let scientists not engage in the debate over whether studies of the endocrinology of mid-line color changes in certain stream fish are basic. Such studies are the substance for interpreting the effects on fish of non-point water pollution from farming and forestry practices. When pollution disrupts the endocrine system and prevents color change, there is impact. When color change is a basic sequel in a courtship ritual, then its failure to change causes reproductive failure and changes in expected population abundance. The real land use and impact question is not whether pollution killed fish, but whether it resulted in a generation not being born. From research, such a question can be answered, understood, and corrective changes made. There is only one science. It needs to be cast as rationally robust work.
Sequences
It seems conspicuous when looked at directly that a major aspect of problem of speeding up research applications in the rural environment is the problem of the sequence of discovery. Perhaps it is obvious, but emphasis is needed to prevent losing sight of the sequence phenomenon in research and to avoid attributing more to the basic-applied dichotomy than it deserves. The apparent scientific successes are those that by chance or planning fall in a fortuitous sequence. The fate of absolutely equal quality research (by any criterion) is a function of the temporal environment in which the results are placed.
The analogy of a three-number lock combination is somewhat instructive. Three correct numbers will not allow entrance, only numbers and the proper sequence. The odds of the proper sequence are quite low. This realization can put into perspective efforts at technology transfer and can explain to those with over-expectations for science why progress may not be as swift as desired.
Such an explanation can easily become excuses for lack of progress. A sequential strategy of research is only appropriate if rigorous planning is done. Ackoff (1962) delineated sequential and simultaneous research strategies and their counter-balancing forces of costs, time, and risks. Sequential research has lower costs, takes longer, but involves less risk than simultaneous research. Simultaneous research is a broad, multi-worker, multi-lab approach usually taken in a short period. It can be very expensive.
I'm convinced that the rural community has little time to aid people significantly and to preserve current living standards, at least for U.S. citizens. Only simultaneous assaults on major research issues seem appropriate. That conviction arises from observations of a host of environmental problems, the increase in counterintuitive consequences of many of the most altruistic actions, and the rate at which thresholds of tolerance and supply are reached. However, I am most adamant about intense efforts of using the past research…to work on improving the sequences so that the next discovery, well announced, will find a rightful place … fast.
Although I advocate simultaneous, planned team assaults on major problems, in such projects there may be inefficiencies and partial failures. Nevertheless, such projects seem advantageous because they buy society time. They put conclusions and discoveries in the hands of decision makers and shapers of society. There exists today an economic order that appears unwilling to tolerate costly simultaneous research programs. The programs are needed, desperately, but they seem unlikely. Society will trade time for risk and time for cost. Instead of buying time, it will spend it. This is very saddening; it is a decision that can be reversed, but not likely. Sequential research therefore is most likely to be done because of cost constraints and the social ignorance that says: (1) we have unlimited time, (2) the burgeoning society with its demands is not at great risk, and (3) the costs of the alternative, simultaneous work, seem very high and the payoff is unsure. The only current hope that can counter this failure is in research planning that can achieve some of the advantages of the simultaneous strategy. Since sequential research seems inevitable, then planning can reduce its costs, and importantly, allow all possible haste. Planning can reduce its major disadvantage of time required and improve its sequencing.
Research planning has been advocated for years. I voice its need again, but perhaps in more meaningful terms than the past. The planning needs are for solving problems like: (1) how can people maximize the total costs of delivering minimum, adequate in-dish meals to a person of specified sex, age, and weight anywhere in the world? (2) How can people achieve a sure, high quality ground water resource for all the people of the U.S. (or any country)? (3) How can peop1e preserve for use the present gene pool in wild and domestic animals? (4) How can people plan and shape 200,000 hectares for optimal biotic production for 1000 years? (5) How can people plan ways to manage 5,000 long-lasting refugee camps throughout the rural areas of the now war-torn world?
These are problem statements appropriate for high science. They are timely, researchable, essential, and will require assiduous application of the scientific method - from the most esoteric and micro- to the most philosophical and macro-approaches. They cannot be achieved in any period of time that has relevance to the human condition without the most profound and scholarly thought, without at least one or more people thinking them all the way through and writing and developing computer simulations of such thought. Previously, there was not enough known or the technology was unavailable to do so; these conditions have changed. The plan that will result following such thought must exist, it must be charted, it must be a shared view, it must be begun, and it must be altered as need arises. With all this, the goal must remain and pressure and leadership must be exercised to achieve the goal. Of course every scientist does not have to "join up"; there can be enough programs to occupy all scientists and require more. There need to be "outliers," challengers, and those with the alternate hypotheses, and they should be supported. There are enough parallels in biology to be convincing that long-term survival is closely tied to energy spent on monitoring, dispersing, and diversifying, and that society needs to fund these mutant efforts. But there must be a plan; the risk of planning must be assumed. The risk of planning rationally robust work must replace the risk of no planning. There can be no experimental "controls," only largely-comparable situations. If these cannot be addressed by science then fie upon it and the exaggerations of its ken.
Planning advocated herein has no similarity to the typical agency document called a research plan, little more than an open palm to a state or federal Congress. Neither is it the gyrations of or mere presence of that glib, handsome cadre of employed planners. Neither is it glorified statistical services or platitudinous reports. Planning is seeing where we as a world society, as a nation, must likely be in 50 years, charting a minimum course to that destination, and creating decision aids to allow changes along the way. Planners can say: "At least we must know some value B or at least we must have greater precision in our estimates of some rate D." This is possible in land use; it is probably possible for most of science. Of course the living, dynamic mind of people will not be satisfied by having a 2005 - 2007 concept, but the course will be set. Minor adjustments along the route to any destination are expected.
The forester is well attuned to the site that is "perfect" for one species but is stocked with another. A timber stand exists if a seed-source was present, and if a fire occurred after seeding, and if the ground conditions were right for the seed, and if the rain fell before or after the fire. A stand is a function of sequence as much as factor. The forest scientist with complete knowledge (in the theoretical sense) of all forest factors cannot predict, a priori, a stand type because of the innumerable sequences. Yet foresters can predict a forest will occur and over time what forest will eventually exist-- and persist indefinitely. This knowledge (or lack of it) is not discouraging. It allows the forester to explain what he or she sees; it allows him to compute with various degrees of probability the future states of the forest on any land. People desire certainty; it does not exist. Such awareness allows people to operate with less entropy or frustration, more attuned to the probabilistic world.
Duration
The expanding confidence bounds around predictive graph lines are familiar. The farther into the future one projects, the less confident one tends to become. But prediction is not projection and the statement about increasing confidence bounds does not necessarily apply, especially if attention is given to the occurrence phenomenon above. An example in resource use may be instructive.
Elk forage that grows and becomes available following fire or clear cutting is known to follow certain rules of succession or transition theory, being irruptive, and then declining to a fairly constant state over time (about 50 years). There are difficulties in predicting the forage in the first 10 years (the confidence bounds are quite wide), but few later. Aggregating these production functions can yield a far truer picture of regional elk forage in the distant future than the near future.
The mental image of sweeping, expanding confidence bounds on linear regressions has confused research planners. The future will not be like the past. Most natural resource experts know far more about the independent variables in the equations than they admit and some things (like elk forage) can be very closely estimated for the practical and long-term future (at least 50-years). Society is probably still behaviorally or genetically operating as if people had a life expectancy of 20 to 30 years. Changing concepts of future reality have not matched gains in life expectancy.
To understand land use change, to predict, people must understand succession or transitions (Golley 1977 and discussed in detail in Chapter 15). Further advances in this area are needed, but they are sufficient to allow rural system workers to estimate now the long term consequences of almost any act such as a spilling toxic material, constructing a powerline, or building an airport (Giles and Snyder 1970). The interaction between sequence and duration is fraught with difficulty. A host of degenerating, poorly-made decisions of the past beset present society. Large dams, contaminated areas, exterminated species, and desert range overgrazing are examples. These are irrevocable. Their rate of occurrence has probably slowed, but it is still a positive rate.
Students once worked with my computer game called Waterloo, trying to achieve its objective of stabilizing the shrimp population abundance in a coastal estuary. The shrimp are a biological integrator of most of the factors of the watershed. Only late in the game do they usually realize that they cannot replace the silt lost to beach erosion by their watershed decisions. The replacement silt from the watershed is all trapped behind a dam that was built prior to their involvement and clearly stated as part of the game. They are saddened and frustrated by this discovery. The best of managerial knowledge -- perfection if it exists -- cannot overcome the constraints placed on their system by past generations.
The above role of past decisions can alter Leopold's useful analogy of land health and the creation of a dynamic environment in which people may discover their humanity. I fear that instead of such an environment, the land manager, at current pace, can become little more than, perhaps by analogy, a ward nurse, solely intent on surviving and keeping the patients from hurting each other.
Students once worked with my computer game called Waterloo, trying to stabilize the shrimp in a coastal estuary. The shrimp are a biological integrator of most of the factors of the watershed. Only late in the game do they usually realize that they cannot replace the river silt lost to beach erosion by their watershed decisions. The replacement silt from the watershed is all trapped behind a dam that was built prior to their involvement and clearly stated as part of the game. They are saddened and frustrated by this discovery. The best of managerial knowledge--perfection if it exists--cannot overcome the constraints placed on their system by past generations.
Retrospect
Herein, I've discussed some of the pathways to discover the role that science has in predicting futures. I've suggested a unified humanistic concept of science transformed into rationally robust work. It has within it a concern for the time when discoveries will be used for people, the concept that research can buy society time in this critical period, and that society is likely to opt for more sequential than simultaneous work. To reduce the impact of this decision, it is important that rigorous research planning be given higher importance than ever before. Contrary to some who contend that prediction is out of the realm of science, I hold that it is presently well within science, has historical roots in astronomy, and needs to be given more emphasis, not because of its shortcomings, but inclusive of them for the utility it has for shaping a reasonable environment for people. The limitations have been discussed under interactive topics of sequence, occurrence, and duration.
Consequences
Every action in the rural lands has many consequences. A tree is cut, the soil erodes; as the soil erodes nutrients are removed from the area. "Nutrients removed" is a consequence of cutting a tree. Each consequence can be estimated based on studies and experience. (We do not have to do a study to confirm that water runs downhill!) Every action has many consequences, some more than others. Some consequences are trivial or, at least with our present knowledge, we cannot imagine a significant measurable effect or consequence. The more we learn, the more connected the consequences will become. We decide that, admitting to consequences that we do not know or cannot measure well, we will deal only with a maximum of 5 levels of influence. Levels might be:
We can imagine several more levels, effects on plants, then effects on insects feeding on them, then effects on pollination, then effects on contributions to the mix in the litter layer, and on and on. These are the studies of and tales of ecologists that believe that everything is connected. Many things are, but within this concept of the satisfactory condition, everything cannot be known; there is no time or money for studying everything; many things have effects that are not significant. Decisions are to be made in a timely fashion. The time to develop a meaningful consequence table for every major action can be very long. The computer analysis can take only a little time but preparation for the run can be costly and delaying.
A rural resource knowledge base must eventually embrace plants as well as animals, soils as well as forests, geology as well as climatic factors. There is no logical separation for topics of wetlands, watersheds, coastal zones, and precipitation of the water budget. Is a plant in the gut of a deer a part of the animal or exclusively a part of the plant world? I think that "wildlife" is all life of the wilds. In order to manage plants well, a great amount of knowledge is needed. All factors about each plant cannot be learned in separate studies. The plants themselves remain enigmas. Where one species stops and another starts is still debated. Mobile plants, such as the liverworts, have animal characteristics. Plant forms and their characteristics differ on different sites. Trees of some species unite their roots, making clumps-of-tree-like-forms the relevant unit. A general knowledge base is needed, one that is rooted more in "expert systems" than in conventional taxonomic keys. So much has been learned of plants over time that many generalizations can be made. There are many fields of knowledge already in a computer information base filled with an expressed high degree of confidence. The entry has to be general because we do not have the time or the money to continue our studies, plant by plant, species by species.
A computer simulation is said to be a means to compute answers to "what if…?" questions, questions such as "what will the changes be and what will be the consequences if I change this factor, build this roadway." The consequence table is a report of multiple consequences of an action; multiple runs of a simulation. "What if I cut this stand of trees; what will be the consequences?" The consequence table is a means of listing the major significant areas for which a report is needed.
It is important to realize that the words used can lead us astray. It may be that "consequences" are categories of interest and maybe these are re-phrased objectives. We may want to know the consequences of an act on the calcium in the soil, but we selected calcium because we knew it is vital to plant and animals growth and health. Stabilizing or increasing the supply of calcium may be an objective. Maybe we are only approaching objectives through the backdoor.
If the consequences of an action seem bad, approaching an undesirable threshold, adding excessive costs, or requiring major capital developments, then the action can be viewed as bad and hopefully then, not undertaken. The answers suggest whether the person or group will be able to remain within "the box," even if the boundaries are fuzzy.
People want to know what will happen, what will be the consequences of proposals or actions. They know full well that precise statements are usually unwarranted, so they will ask for the "odds" or for probability statements. Progressively, rationally robust work engages in using computer simulation to produce consequence tables, expressions of the likely changes in the conditions of important objectives.
Risk Zones
Workers in agroforestry and other outdoor endeavors soon learn that classroom measurement precision is a myth. Tapes stretch and kink; maps shrink and swell; surveys are variously conducted; ancient survey corners are difficult to find or are missing; assumptions compound assumptions; data falsification is impossible to check; bias is rampant; care in making measurements varies with the temperature, precipitation, insect bites, and fear of snakes. Proceding "on the contour" may mean including a large rock or ignoring it. When land surveys of boundaries are made, invariably the polygon will not 'close.' Survey field notes often require major adjustments. Surveys of large areas may have errors of hundreds of hectares in the 'slivers', the thin mapped areas between actual and hypothesized boundaries. And boundary changes are commonplace, especially in riparian habitats and mined lands.
To estimate animals or stems to a number with precision to the nearest decimal knowing it will be multiplied by 7000 hectares (which may have 15% error) is irrational. The law of significant figures still exists and excessive precision needs to be avoided. An estimate of thousands of units will probably meet the needs of anyone interested in animal or stem abundance in an area.
All consequences in this formulation being described are estimates of future conditions directly related to objectives. Thus they are predictions of likely
 |
| A linear regression, a line that is fit to or is modeling sampled points over time, can be used to suggest future values. The estimates and their deviation from the line become greater the farther away from the sampled points. |
Equifinality
A concept within general systems theory, equifinality (Von Bertalanffy 1968,1975) , deals with the observation that there are often several ways to arrive at the same end state. In arithmetic, the example is clear. To get 9 we can multiple 3 by 3. We can also get 9 by dividing 27 by 3, similarly by adding 4 and 5. Different numbers and processes can lead to the exact same outcome. This is true (but rarely noted) in the rural land sciences. There are many pathways to a mature oak tree, an adult deer, a mossy rock. A lot of water and a little fertilizer can result in the same crop yield as a little water and a lot of fertilizer. The emphasis here is that there are many ways to get to a desired end state, a position near the center of "the box." Many different abilities, tastes, backgrounds, and experiences, even objectives, can exist within a group as long as they recognize the space that they occupy as suitable. They may complain, test the limits, step outside and return, or live a
 |
Analyzing pathways to determine equal crop or tree responses, or finding the "best" pathway (most cost-effective, etc.) are typical problems in agroforestry. Finding the absolute pathway or combination and sequence of factors may be time-consuming and expensive. Inputs to a system can change over a fairly broad range and still yield almost the best result. An important concept to be followed is that there are ecological thresholds in rural systems, e.g., most natural areas of the world will produce no more than a certain limit of phytomass.
The study of equifinality can provide new insight into the importance of objectives. If total tree fodder for farm animals is the objective, then there are many pathways to that condition. It may be that by re-defining the objective (e.g., total forage vs. percent digestibility) different conditions may produce the same particular end result. The need is to select a means that will maximize or minimize the results from among the permutations of these ways (recall that permutations are not only combinations but sequences of applications). In rural systems, each permutation is a potential pathway to the same end, one of many pathways of equal or often insignificant difference in costs or other criteria. (The number of permutations of 10 items is 3,628,800.) There is a vast area of financial and other indifference. The search for the best one, or more likely "ones," among the pathways will serve well.
When we do sensitivity analyses in the rural arena, we find many factors to which the system performance measure is insensitive. We can change inputs to the system over a fairly broad range before we reduce optimum conditions. This observation has evolutionary and survival-value roots, but the point is that it is irrational, counter to the available evidence, to believe that very great precision is needed or will be useful in work in the field with most factors. Of course, it will be irrational to fail to look for those factors to which the system is most sensitive or to fail to use those that are found with care.
Shopping
People "shop," not having a specific objective. They buy something, achieve their objective, then "shop." There is fun in looking around, seeing what is available, evaluating whether things seen as new can be used effectively within the life space to improve it, reduce risks, reduce costs, provide amusement, achieve an alternative social status, or provide power to do work later. Rarely are big purchases made while shopping. Information is gained, then after pondering a purchase, the big buy is made. The big buy takes more thought, more complex decision making. There may be things available in the shopping area that will expand the space, reduce costs, improve the quality of life within the living volume. Shopping probably has survival value.
The Enterprise Paradigm
How do we get to operating as if groups of resource users exist within the fuzzy space of the box above? What is the paradigm for the satisfactory condition and how do we get there from the present condition, the situation? Perhaps we cannot. Perhaps we are already outside the box and will only make incremental changes as opportunities seem to arise to fix things, make adjustments, and appease vocal or influential individuals. Some, however, believe that the situation has not been seen clearly, that options for gaining satisfaction are available, and that the benefits of effective efforts may exceed the costs.
That ground is where needs are so different that the decision making process itself is so different from classical decision theory as to make it almost irrelevant. It is a general situation where current "public participation" in natural resource decisions is untenable; where private landowners must have new incentives for improved socially responsible land management; where risk-taking takes on a planned imperative; where education for situations (not "more, general education") becomes required; and where computer aids are used to present as well as alternative emergent situations.
Perhaps I can build a bit of a structure from my little corner to help with a bridge (one thing needed) from out of the perceived difficulties. I fear, not for myself, but for rural resources, if I should be left in my gloomy, freshly-painted little corner where I have painted myself and most things rural.
Temporal Aggregates
If we can stop thinking that each 24-hours is a very precise number for our analyses of differences and change per unit time, we will improve our models, stop much awe over great variance, and reduce the need for saying "more research is needed." Time is a human construct, an accounting mechanism. A "day," however, is grossly amalgamated solar relations, cumulative lunar forces, average soil movements, etc. It is the intrusion of variance into the most fundamental assumption about time units that seem constant and controlled. We have to replace clock units with accumulated biomass or Langleys of energy received or food metabolized. Sunlight is strongly time related but it is not equivalent to time and as we study grass, crop or tree growth, we know the major differences among seasons, latitudes, slopes, and aspect as they each affect the meaning of a clock-unit of day length. A day is a way of coding and recording when ecosystem radiation starts and stops and each day is unique in its measure of energy received at a point. It may be a way to code lunar forces. It has no intrinsic meaning to knowledge of plants or animals. Convenient and unlikely to be replaced however, we need to substitute time (at least "days" and "years") with one or more appropriate fundamental units such as potentially accumulated or received solar radiation. What else was at work in the green house, the lab bench, or the forest between 6am each morning when the clock buzzes is the unit for study, not a named unit called "time" or "a day."
In agroforestry and other ecological research, long-term studies are desirable but rare. The duration of studies influences the perceptions of practitioners about the suitability of such indices. We believe that such studies are needed, can only be afforded if there are unique observations possible, and if so, their use must be as carefully planned as the recording of observations and analyses. Serial observations are recorded with emphasis on trying to gain the maximum and minimum possible values. These observations would go into Nature Seen, one of the proposed groups of Rural System.
Farmers discuss seasons being "late" or "early." Ecologists study phenology, the study of the timing of biological events such as grouse mating, leaves falling, select plants blooming. In rationally robust work, including phonological time will help clarify chronological time and give that classical measure a new dimension, reducing claims of excessive variance in studies and needs for more, expensive samples and their analyses. Staff of the Plant People group work for increasingly precise models, mathematical expressions or computer programs that allow us to explain past events and to estimate or predict those for the future when some of the variables are under our major control.
Few workers in the rural environment know that they can gain massive statistical control within systems by knowing two factors, elevation and latitude. Slope, aspect, land form (ridge, saddle, etc.), watershed boundaries, stream channel location, stream order (and 20 established relations), and topographic indexes (40 or more) can all be computed just from elevation in cells across a landscape. Knowing and working with these fundamental relationships give us great, rapidly-developing modeling power.
Day length and radiation estimates can be computed from latitude and slope. Precipitation and temperature records can be adjusted based on nearness to multiple observation centers. Temperature estimates can be adjusted by solar radiation and elevation. This list of models is extensive. Workers in rural systems need to gain a knowledge base of the key abiotic factors, the non-living 'things' to which plants and animals respond. With amazingly few values, great predictive power can be gained over major system performance measures.
With site-specific models, optimization can be done for crop, plant, livestock, tree site selection, and production units. Plantation failures and disease and insect epidemics, which are often the results of introducing a production unit into the wrong place, can be avoided. We can improve existing models and create tentative models with much that we already know. We know from past studies the form of equations, that numbers less than zero do not occur, and the likely maximum values. We can probably advance more rapidly by proposing, using, and adjusting theoretical models than by whining about excessive variance and decrying the lack of funds for curve fitting. The knowledge base that we build will be within the models, documented, and changing as we cast ahead curves, find limits, bracket in coefficients, add variables and delete insignificant ones. We are skeptical of models now because they have not been used well, given far too much promise. Expectations were not fulfilled and the procedures were ceased, not modified, recast, or allowed to grow in the light of new understandings, redefinitions, and re-formulations.
The Monetary Resource
The classical, incremental approach to gaining knowledge about our biological world (more importantly, ecological world) can never succeed. There are at least four reasons:
If these statements reflect the minimum condition, then it is irrational to act as if they were untrue.
Economic analyses are generally about allocating scarce resources, not necessarily about how people spend money or gain profits. People substitute resources readily, take risks in ways not yet well known, but definitely not in ways described as 'linear'. Human systems seem devoid of any collectively perceived or at least readily communicated goals or objectives, thus denying prospects for mathematical analyses. Whole agencies have difficulty in expressing their objectives. Stated needs for physical production are in conflict with physical yield or harvest; physical harvest in conflict with gross annual income; income poorly accounted with costs; risks rarely computed; and present discounting is in conflict with intergenerational concerns. Given such ambiguity, we argue again that great precision in estimating optimum tree spacing or fertilization rates, for example, is inappropriate. Success is almost impossible to establish; applying feedback to make corrections to an unspecified standard is impossible.
The needs in rationally robust work are for strong pressures to state objectives precisely, but this seems at odds with the above pessimism. At least as much work needs to be done in this arena as in field sampling, and it needs to be done quickly and rigorously. People within rural areas might, for example, consider maximizing a primitive benefit-to-cost (B/C) ratio where benefits are expressed as citizen-weighted objectives (to be maximized by Rural System practices) and costs (present-discounted) are to be minimized. Work on the land, over time and throughout a system, could then be directed to stabilize or increase this unusual B/C ratio.
Tentative Confidence
Almost everyone likes to make decisions with high confidence (or a high probability of 1.0 minus risk). People desire low probability of 'being wrong,' but there is ample evidence that they do not behave in a way that is consistent with such theory. People marry with fairly low confidence for success (by several criteria). They make household purchases with only modest amounts of information about best options. Farmers or foresters rarely farm or practice forestry as well as they know how. Making decisions at some high level of confidence seems reasonable, but it is often inconsistent with human behavior.
Risk-taking behavior is never singular. In drug-free people it is always a combination of a perceived probability and the effect of the consequence of being wrong … in the present instant but longer future. The consequence of being wrong is the combination of the effect on an individual, the number of individuals, and the magnitude (especially over an area), and duration of the effect. When confidence or the probability of an error is computed separately, meaning of risk for a farmer or a community is lost.
Scientists, the community, have adopted an arch-conservative, risk-averse paradigm in the standards for confidence in their micro-environment, tightly-controlled experimental decision making. That paradigm has been taught and widely accepted, insisted upon for human drugs, generalized for everything else, and thus the general educated public now has excessively high, excessively costly, excessively delayed contributions of "science" to decisions.
Most rural research is stuck in a 95%-probability rut. Taught in college as proper, the level influences decisions throughout many aspects of rural decision making. The perception is that we can tolerate an improper decision (e.g., whether crop production was significantly increased by underplanting vegetables within a grove of trees) no more than 1 time in 20. Given that 10,000 such experiments have been done in the past 10 years, then 500 erroneous decisions have probably been reached. People concluded that there was a significant difference due to fertilizer or irrigation when there was none. The reason that the possibility of 500 mistakes does not bother many people is that they seem to know at a high level of probability when something did have an effect. They are more confident than 95% even though the test statistic is only working at that level, the 0.05 level.
While we would all like to be absolutely certain of almost everything, i.e., decide at a confidence level of 0.9999, such a criterion is unreasonable and excessively demanding in most Rural System work. For corporate boards or directors, a mere 5 to 15% improvement is acceptable. Successful gamblers only need to win a little more than they lose (say 55% of the time). A stock portfolio must have only 5-12% gains as compared to losses or failures. A game population only needs to be relatively stable. The population being over-hunted, for example, can be easily restored in a few years (the post-hunt population often exceeding that of the pre-hunt because of surplus surplus food and without crowding) or by simple adjustment in future harvest regulations. These observations about how other people deal with confidence or with acceptable levels of accuracy argue for me that it is unwarranted to assume that farmers or foresters operate substantially differently than they do. Using high levels of confidence such as alpha of 0.05 (the 95% level) in most rural system work is inappropriate.
Achieving high confidence requires excessive sample size, thus, high data processing, storage, and analytical costs, and produces results that are often inappropriately used, not reported, not stored, or not critiqued, and thereby violate many of the premises of classical science. The use of the confidence level as a separate statistic is inappropriate; it must be unified with effect, people, and time. A much lower alpha level needs to be adopted on the grounds of appropriateness, high expense per sample (in all dimensions), inevitable alternative sources of knowledge, and on the grounds that rationally robust work involves a clinical approach, one with feedback over relevant time. Simple computations can demonstrate the high costs resulting from establishing inordinately high requirements for confidence and tolerable error in studies. The statistics of brief, controlled studies does not apply to the rural situation. Assuming that they do apply may more than double the costs of studies. Given the massive needs over vast areas, the extreme pressures on resources, the desire for answers as quickly as possible, the extreme shortages in money and expertise for studies, the complexity of the problems, and the relative adaptability and resilience of natural systems, confidence levels of 0.20 need to be used, followed with applications and adjustments.
Within rationally robust work, staff of Rural System (I continue to dream) seeks to find the fewest number of pieces of information (the system inputs) that, with regression and other models, give estimates for the greatest number of important rural system models possible. Acceptable control over the system is judged when model goodness indices of R2 values above 0.64 are observed. Continuing highly precise pursuits now seem inappropriate. Perhaps other people who continue classical studies may contribute to rural system work and knowledge, and knowledge, however gained, will surely be welcomed to improve estimates throughout the complex models that will be used.
Range-related Knowledge
Many natural phenomena are not normally (bell-shaped) distributed. Because of this, the statistical median often better reflects central tendency than does the mean or average. When one value must be used for a factor in a 50-factor model, then the median should be used. The median has been effectively estimated in engineering and military work for many years and, although used to develop estimates of time needed to complete a project, we believe it can be used with low risk in other aspects of rural work when experts are available. The principal advantage of estimating a median value is that estimates of parts of the equation for " high, low, and likely " can be obtained quickly from experts (village elder, etc.), available records, or observation. Intensive sampling and measurement is not required to make at least a rough estimate of the median. It is rational to use estimates such as this for crop production, tree yield, charcoal yield, animal weight, and other aspects of rural work. It is irrational to deny so-called sensory and authority epistemological bases of the people of the country and to ignore the growth, survival, and potential harvests and real benefits from forests, livestock, and croplands. We must deduce with feedback. Cost-effective development of dominant relations among all major biological and social factors seems reasonable.
Maximum and minimum values are often easier to establish, and may provide more information about natural phenomenon being investigated, than an estimate of an average or mean. It may seem unreasonable to use ranges, for such use seems to relax our efforts to achieve great precision and ability to discriminate. However, hope for gaining knowledge lies in using the ranges and also increasing the number of dimensions of an analysis, not, as in past studies, in emphasizing increased sample size and precision in only a few dimensions. The result of using ranges will be to limit the sample sizes and, thus, later time and costs.
The observations of the range include those from a global maximum (e.g., the maximum temperature ever recorded by any weather station in the world), to a regional maximum, to a stratum maximum. Bayesian analysis suggests the practical use of a priori or before-hand knowledge of such phenomena as maximum temperature in a study area. The probability of two states of nature, above or below the range limits, can lead to a set of values (perhaps in a uniform distribution) that can be used within computer models. By such use, sensitivity of the system performance measure or 'success score' to each variable can often be determined. Determining which are the variables that need further study or that must be expressed precisely can lead to major savings.
There is a feeling, generally, that knowing the range of values for some aspect of a system provides little information for decision-making. In ecological systems, functioning over very long periods, what is now observable as the system is really the "remainder" of plants, animals, etc. after extreme or episodic events. (Those surviving are said to be "fit.") Because of this, the range is probably the best values for use, especially while the 'long-term' and the 'sustainable' system phrases gain political and research-support vogue. Ecologists see a multi-dimensional space within which people or plants or other animals may exist. The walls, as in the three-dimensional boxes depicted above, are the outside limits to where they occur. The walls are the ranges. The space is called the creature's "niche." Endangered species have very small niches.
The ranges from field studies of rural factors of interest can be mapped within a geographic information system. All work to date using more than seven variables has produced maps of great complexity, detail, and counterintuitive patterns. These maps seem to provide at least as much resolution for decision-makers as conventional yields from statistical analyses. Even the results of combining seven factors into a model and using it to produce a map can result in more complexity, detail, and lack of pattern than can be handled by even very bright managers in the field.
Many factors that operate on crops, trees, or animals operative suggest a feasible space or defined hypervolume within which they may survive, production may occur, or profits may be obtained. My emphasis is that ranges are important; an optimum might be found, but the computer searches and field tests need to first address the space defined by feasible or reasonable upper and lower limits of all relevant factors.
Looking Back
This and the previous chapter are partially about heuristics. "Heuristics" is not a widely used word but an exciting one, full of subtlety and potentially quite rich. It carries information and has its own ambience. Roughly, it means the way one finds out or discovers. Chapter 3 and this chapter have been long and will probably be relegated to cosmic otherness, losses that might be tracked by learning-forgetfulness curves and probability functions for ideas accepted. The desire I have is that, the reader may later adopt and improve, perhaps reconstruct a personal rationale, a viable process of study and rationally robust work that feeds an improved design process. That process may materialize in plans for people discussed in Chapter 19 …and a satisfactory condition for them. The reader is encouraged along my tortuous, conceptual path toward the ground for hope for the future of rural people.
There is a fundamental epistemological question (Chapter 5) behind stating the role of anything. How do I know what to do? What is best, right? The entire Rural System enterprise can be viewed as being focused on a desired future and that is dependent upon rationally robust work. That includes prediction and as well, explaining the past, for the future is likely to function similarly. It involves more that this, for making decisions and implementing them, and assuring their performance, and then managing the results is the enormous work ahead. It is in the understanding of these functional relations, using them in synthetic models with higher deductive skills, that the future can be known, that consequences of acts can be evaluated before they are performed, and that the future world can then be shaped as a proper place for humankind.
| Table of Contents |
About Rural System |
About the Author | Glossary | Groups of Rural System |
The Country Store | Progress | Contact the Author |