A Proposed Planning System : GeorgA
GeorgA is a proposed prototype of a system for creating a Forest plan and allowing it to be updated as needed. It can likely be used on on other Forests. In addition to information for making decisions, it presents a scoring technique, one that shows how the system may be changing. The system is strongly focused on human objectives and on performance. Decisions implemented on the land (proposed or actual) are characterized by:
It is reasonable to be skeptical of such a model. The promise has often exceded the expectation. There are several reasons why this effort is different, reasons why such a system has not been created before:
No scientific "breakthroughs" are needed. The proposed work is based on a new formulation of work over the past 20 years to develop a practical tool with self-adaptive features.
The system is largely "non-linear." It operates with a complex set of thresholds and constraints, uses curvilinear, multidimensional ecological succession and state transition functions (forest yield curves being an example), uses information on spatial arrangements (e.g., erosion from uphill sites or influence of private cropland on forest deer forage), uses a concept of zone-of-influence of any action, and allows any optional action to be studied (a simulation) and compared to an optimization run (the standard). Not an energetics model," the system uses energy throughout in parallel with other analyses as a fundamental modeling theme and as a basis for examining "costs"
Because the project is large (appropriate for the magnitude of agency problems and relations with other individuals and groups and dodging the piecemeal approaches of the past), a simplistic sketch may be useful.
A computer-aided project management system allows all of these parts to proceed simultaneously to a rapid conclusion in 3 years. We hasten to add that as a living, continuing, continually revised and corrected system it must be used and maintained for full benefits to be derived. Some reasonable, tentative commitment to its longterm use is essential. If absent, it should not be developed for time, expertise, and funds will be wasted.
No one, no computer or system, "knows" the future. It must always be predicted. Predictions can be improved; the magnitude of the difference between the actual and previously-decided condition can usually be reduced by the system.
There are great gaps in human knowledge about all aspects of economics, ecology, and other areas of concern in the management of forested lands. Filling these gaps requires much modern research and new strategies for gaining knowledge.This system is built so that it contains all relevant components and topics, even if knowledge from detailed research is not yet available. Rather than wait for all knowledge in all components at high levels of confidence, this system attempts to improve decisions
The context of the system and the scale have prevented certain assumptions from being made because it was unreasonable to do so. (For example, excessive wood production may deflate wood prices so that profits to people of the area may decline, even though wood production increases.) These out-of-Forest, market level changes are not computed. Systems design is for the hardware currently available. Changes in hardware are expected.
Only people who could work harmoniously on the objectives of the system under the constraints are recruited.
Only a reasoned set of possible typical, "projects" or proposed types of land use changes were included although the entrepreneurial concepts of Rural System are encouraged for study and development. By working toward capabilities in the named fields,
The system will not make decisions.
The system will not take risks.
The system will not establish objectives (only elicit them from managers and the publics) or any of their parameters. (It does, however, provide a methodology for studying the effects of altering different assumptions and dimensions of the objectives.)
These limitations are not seen as weakness but are a clear transparent statement of part of the context of the system and thus expectations for it and its results.
The Major GeorgA Project Objectives
To create a comprehensive expert system for aiding in
1To develop a concept of and explanation of desired human benefits from single National Forests (the George Washington National Forest Prototype).
2To relate and estimate potential human benefits from National Forest and to these desired benefits.
3To develop a fundamental array of generalized land uses and practices with costs.
4To develop a system for estimating the physical consequences (expressed in major benefit categories) of engaging in any of the fundamental uses and practices.
5To develop an estimate of the optimum land use system and to make comparisons of the existing or proposed systems to the optimum.
6To produce decision guidance and planning "documents" showing the state of the system, proposed (or needed) changed, effects of proposed changes on the state of the system relative to the optimum, and a limited set of suggested optional actions.
Each objective will be discussed briefly. Details will be provided.
1Objective 1- Human Benefits
The concept of happiness is such an indeterminate one that even though everyone wishes to attain happiness, yet he can never say definitively and consistently what it is that he really wishes and wills.
Past modeling efforts in comprehensive forest and land management have failed because of poorly stated or overly simplistic objectives (e.g., maximize profits, minimize sedimentation). It has also seemed to fail because all people were assumed to have the same objectives or equally-valued objectives, or that objectives of different groups seemed to cancel out each other. In some cases people's objectives were assumed to be inestimable , i.e., that only computer simulation was used to estimate effects of projects with descriptions only of the physical effects of proposed actions. GeorgA emphasizes gaining clear, precise estimates of objectives (called by Nute et al. (2000:359) operationalizing them)...with continual testing and revision of the estimates.
|The general system of general systems theory shown here is used. Modified, it includes the elements of "start" and feedforward (modifying the present based on predictions).|
The proposed system is not an ecosystem model( Rauscher et al. 2000); it is more comprehensive and contains several other major components. It includes people, their economics and esthetics. It depends upon professionals to create a subsystem to quantify objectives with public help. Then it allows staff to seek realistic valuation of these objectives in terms of:
The composite expression of desired benefits is Q*
It is possible to evaluate the benefits currently being produced by or potentially derived from the present Forest system. These are symbolized as b. The longterm progress of the system is seen in B* computed as
B* = ((B - b) / B) x 100
The perfect score is 100 when the system is "right on" the objective (or very near it, say within no more than 2% or at least 80% of the stated objective). These are called the bounds of the objective and are a managerial policy decision. It is undesirable to exceed the objective (suggesting excessive allocation of funds). Under achievement may be budgetary as well as from fire or other assault on the system.
The complex details of obtaining B*, B, or b are available to the public, but most citizens will be pleased if a high score or progressive improvement in B* can be shown. B* as well as the system score, Q,of course, will reflect forest fires, tornadoes or hurricanes, mismanagement, adverse political intrusion, and outside edicts such as land confiscation for utility corridors.
Individual decisions usually have costs to the U. S. Forest Service. Alternatives can be evaluated in terms of maximizing the likely benefits over time (say a 150 year planning horizon, T) by selecting an option
Q* = ( Et B*t) / C*t
(a simple purchase or a complex program or project) by attempting to maximize the "expected benefits" (where Et is the probability or certainty of a benefit in a future year) per unit of "expected discounted cost", C*. The cost includes inflation, changing resource value, etc.(Overton and Hunt 1974).Costs in dollars will be expressed as current and 2000 dollars. Costs will also be evaluated in terms of expected kilocalories of energy or fossil-fuel equivalents. Q* is subject to a large set of constraints.
This is an unconventional benefit-to-cost ratio. Benefits can be translated with difficulty and debate; real current monetary values can be suggested by the proposed conglomerate. Properly distributing costs to benefits (or assigning impacts) when complex projects are developed (e.g., a road, a dam, a multi-acre wildlife-area development) is at least as problematic as evaluating the worth of wilderness or wildlife. Just as gross system performance measures have served well in education, in public economic policy, and in crime analyses, so can Q* serve the Forest.
The system benefit performance is judged in B* the total system performance is judged in Q*.
For every project or any interrogation of the system, the results will be reported in terms of total system benefits.
The system denies a single-purpose or single-interest orientation because there are hundreds of constraints and the large set of objectives requires allocation of some resources, space, etc., for them all to be achieved in some manner. The analyses are for marginal gains, those advancing beyond the natural or current capabilities of the system.
2Objective 2 - Potential Benefits
The methods to be used to estimate potential benefits are categorized in the resource dimensions of (1) energy or matter, (2) time, (3) space, and (4) variety.
The "energy-matter" category include the nominal resources within the highly related or overlapping components of:
|Relative atmospheric and temperature maps (in color), somewhat like the soil erosion map shown here for a county
can be produced for alpha units of the Forest
The potential in each of these areas will be analyzed by area or volume; over time (0-10, then 10-year increments to 150 years); and pattern (proximity ratios and pattern statistics).
|Funtional or Process Elements of the System
Fire Prevention, Management, and Supression
Safety and Security
Waste and Litter Disposal
Education, Information, and Technology Transfer
Vertebrate Damage Management
3Objective 3 - Alternatives and Costs
A major difficulty in past modeling has been in defining a project or the "thing" that may cause impacts. A system designed to solve only one action (e.g., to locate optimally a waste disposal site) is not likely to be able to solve another problem or type of action. One action may be a combination or permutation of many factors (water, fertilizer, fire, dozing, tree cutting, mowing, etc.). The permutations of merely 5 such factors is 120, implying 120 computer run to evaluate the effects of these combined "projects."
We shall develop a small set of generalized actions in the Forest such as (1) corridors (e.g., powerlines, roads, trails); (2) points (e.g., shopping centers, ponds, airports, recreation sites); (3) areas (e.g., clearcut, air pollution zone; and (4) volumes (e.g., ground water, airsheds, viewscapes).
The results, about 30 standard land use "acts," will have algorithms developed for each. The results will be a system that is flexible and one that can analyze an array of well known or "standard" actions such as "group selection cut" or "build 1 mile of arterial road."
Limited in the past by standard programs to solve unique problems, GeorgA provides opportunities for studying new and unique land uses. The computer is "fooled" to solve the new problem because the parallel or isomorphic general systems are available.
4Objective 4 - Consequence Simulations
A simulation is developed to produce the consequences of an action (one in Objective 3) on all potential human benefits (resource components shown in Objective 2). This is a "home made" simulation allowing answers to "what if action 12 was done in cell XY of a watershed at time t? What will be the physical consequences immediately (1 year), over 10, 50, and 100 years?" This is largely an ecosystem modeling effort. It includes a deterministic option as well as stochastic one. Units particular to each resource are used. Results are produced in tabular form. Comparisons are possible among present, potential and likely future states. A narrative report is presented by the system.
5Objective 5- Optimization
Given the Forest limits, the potentials, the stated demands or comprehensive set of desired benefits, and the limitations of present conditions (roads, streams, ownerships, etc.), what is the configuration of the Forest that best achieves the objectives? This is a gross land use assignment process -- involving proportions of the area in land use types and then as appropriate specific site locations of points (e.g., mines or hospitals), corridors, areas (e.g., wilderness and productive timber land). This concept of the best state of the Forest system over time becomes a standard. It is not necessarily "the way things should be."
Areas are protected as needed, patterns assured, economies recognized (e.g., straight roads), transitions included, present discounted costing used. Each allocation of land to a permanent use is done to achieve a set of objectives. After this is done, that land becomes a constraint on the system. Future expansion needs (e.g., for an equipment storage site that would minimize future public costs) also become limited constraints. Finally, large tracts of land are allocated for specific uses such as recreation and tree plantation.
An optimum forest harvest strategy is specified. The desired forest that achieves soil, water, animal, recreation and other objectives is thereby specified. The way to get desired forests is by harvests and reforestation. Costs of doing so (or profit) are not explicitly included since this can be manipulated and/or regulated outside of the system. In some cases, a cut and let-lie practice may be used to achieve a particular age in a desired space.Not profit but total Forest objectives drive the system. Probable profits (and a range) will be reported for comparisons, but optimization is done on benefits (Objective 1).
Not minimizing the cost issue, C*t, the system provides with each run expressions of cost as:
1.Estimated direct current costs of an action
2.Present discounted costs (150-year horizon)
3. I -- Modified discounted costs (Overton and Hunt 1974)
4.II -- Modified discounted costs (Giles, in preparation)
5.Net total system energy budget (Kcal - fossil energy equivalents cost)
6.Net total system water budget (public water costs)
7.Net total system water budget (kwh - hydropower equivalents)
8.Score foregone and cost differential between actual (or proposed condition) and the optimum condition.
9.Wilderness energy ratio - the ratio of current energy storage relative to that in the Forest as an hypothetical wilderness 1000 year ago.
10. Sequestered carbon
The fifth objective is conceptually the most difficult and it has resisted solution for decades. Several algorithms have been prepared. The one here is called the Q* Heuristic. The objective is simply to maximize Q*. Q* is achieved by minimizing the difference between the condition of the Forest in which the desired benefits can be obtained and the actual condition. This is done at the lowest total costs. The actual condition may be the present state or that likely if a certain action was taken. The formulation was shown under Objective 1.
The "desired condition," B, is obtained by the following:
1.Determining whether objectives can be achieved based on land potentials, then adjusting objectives to reflect ecological potentials.
2.Determing whether modified objectives are achieved by the present conditions in the Forest.
3.Rank ordering objectives not yet achieved.
4.Dedicating areas (map cells) to highest valued objective not yet achieved.
Optimum locations are based on primeness (Giles and Koe1n 1983). An array of cells is held in storage.
5.Each subsequent objective is analyzed. Spaces needed may be taken from the lower primeness areas determined (in 4 above).
6. When the last cell (or equivalent value cells) available (from 4) is used, the system drops that criterion for location. Residual cells are then available for locating projects that achieve subsequent objectives.The search is "total enumeration" or "numerical analysis" of all non-zero cells.
7. The objectives requiring fixed, longterm dedication of areas are achieved first.
8. All objectives are to be achieved, not just the highly weighted ones. This assures the minimum effort to achieving desired biodiversity.
9. Integer allocations are made in spaces such as for ponds and lakes, recreational areas, and waste disposal sites --- as required, to achieve highly weighted desired benefits in sequence of importance or weight. Remaining spaces (map cells or alpha units) are constrained based on the potential of each for achievement of any objective. Allocations of spaces (e.g., to achieve a corridor or a spatial index) radiate from the largest homogeneous tracts of the type for which there is a supply but a shortage.
10. Spaces (e.g., harvest cuts) are located based on succession curves (Giles and Snyder, Giles 1978, Waldon 1987) and state-transition functions (Williamson 1981). Details for continued development of the "yield curve" or "ecological production function" approach is now available. Previous work (Waldon 1987) shows how we can determine the area and timing of harvests to achieve optimum conditions for 40 bird species..thus for others. TVA work (Giles - WRAP, and Giles (1978) shows how multiple benefits (wildlife, fish, timber,recreation, etc.) can be aggregated and solved.
The present system extends that process to include approximate spatial Designations for land uses.Within the forest areas, an optimum road location pattern is created by computer.Planned harvests in each cell are a function of proximity to a road, present road condition (including constructed or not), and time during which it may be used. Harvest potential in each cell is determined based on the multi benefits potentially derived from each stand of trees (proportion in a map cell) characterized by slope, aspect, elevation, probable texture, nearness to ridge, nearness to stream, latitude, solar radiation, growing-period precipitation, and growing-period degree days.
There are two major types of questions usually posed for (or by) the land manager.
A third type is, occasionally and hopefully more so later is, "I have been allocated dollars.What should I do with them, where and when, in order to maximize the overall Forest land use score?" (The Opportunity Area is expanded to the District, then to the Forest. The option to address all 3 will be provided to the user of the system.
6Objective 6- Hypermedia
The answers to the above enormous complexity must be produced in brief form useful for plans, publications, prescription documents, and impact assessments and statements. A major aspect of this project is the creation of an easily read, pleasing appearance e-book, one on a computer monitor, in a CD, or printed hardcopy. (The options will be available to see any one or series of "pages" as computer-monitor pages -- maps, images, tables, etc.). The book will contain introductory text, summary tables, graphs and charts, maps, and recommendations. The output component of the system is composed of:
From the one common base of information and analyses, there are three levels of reports produced, all differing in detail and content.
Level 1 is for the field person working for the District Ranger.
Level 2 is for a Forest Staff Specialist, and
Level 3 is for the Forest Supervisor.
Economies will be discovered as GeorgA is used Region and nation- wide.
Objective 5 may suggest the optimum solution or condition but how to achieve it may be in question. Impact statements require alternatives to be presented. Often difficult decisions need to be explained or at least the pathway to be taken needs to be shown. (In one recent meeting opposition to a waste disposal site selection was silenced as map after map was presented saying: "certainly not in these areas... not in these areas... not in these areas ... etc."The only solution became quite evident! The opponents were silenced.)
A set of managerial alternatives or suggestions will be available. Suggestions from the literature will
One important aspect of systems work is feedforward, a concept that includes many steps but fundamentally means predicting or estimating the future and modifying the present system so that it will take that estimate into consideration and act on it so that it will be most effective or optimal over the long run. In an unusual logic, feedforward makes the system suboptimum or "wrong" today but increases the average longterm performance.
No one can "know" the future. Stating (1) the time of occurrence of future events is difficult but (2) stating the occurrence of events, of growth, and of stages are very predictable. Many sets of assumptions in systems work are statements about beliefs about the future. There are many feedforward techniques.GeorgA will employ:
1.Regression, time-series analyses
2.Ecological succession curves
4. Scenarios of local conditions
5.Simulations with maps of future conditions 6.Conferences on the future.
All of the work in achieving the above objectives leads to the ability to produce, on demand, a planning document.GeorgA, in one of its major uses, becomes a planning system. This objective seeks to suggest a new concept of planning, reduce time and costs of planning, reduce litigation, and make legal provisions and requirements more easily met.
May 13, 2012