| [ HOME | Forest Faunal Systems Home | Table of Contents | Gamma Home | The Finder | Glossary ] |
Permanent Forest Openings
The manager may plan on developing food patches or "clearings" in forests. Intensive work within these areas is not unlike faunal space management within cities. The literature, supplies, equipment, and plant materials of the garden world and urban wildlife management are relevant. These are usually areas of less than 1 hectare. In some forests, trees are cut (or killed and left standing) and the native grasses and forbs underneath encouraged or grasses planted. These are for wildlife feed. The cost-conscious wildlife manager will estimate the animals in the area that might possibly use the food produced, estimate how many animals can be fed for one day or one month (the animal-unit-month used in range analyses and allocations) (CAP9067) and then estimate the cost of producing the food. The tree grower will be standing by observing and asking how much wood growth was foregone by allowing the area to remain in wildlife food? The manager will be replying with another question such as what is the combination of resources on these hectares that will produce the most money and social good over the next 100 years considering markets, transportation, interest rates, land value, and especially risks? There are no general answers; an answer must be calculated for each area. The question of fertilization costs and benefits for wildlife and trees may be asked. It is clear why there is a continuing discussion among foresters and wildlife managers. The objectives are rarely clear and the assumptions about costs, interest rates, and risks are rarely the same in the two groups of resource managers.
Proximity of potential food to trails and use centers (Hamilton and Watt 1970) e.g., waterholes or den sites, may be a relevant factor in analyzing food. The real nature of animals' areas is so poorly known and use rates vary so much that I suspect proximity may be a quantifiable, highly relevant observation.
Many foresters are accustomed to working with risks. The worth of planted trees is based on their expected value, the product of the estimated value of wood and the probability of getting the potential product marketed (e.g., the expected value of a $100,000 forest just planted with risks of loss of about 0.20 is [100,000 x 0.80] or 80,000). The risk of loss is from fire, disease, insects, poachers, storms, etc. The risk is a factor in the managerial equation that cannot be ignored. Many foresters include it (as should faunal system managers) in computations by adjusting the interest rate used in estimating the present value of an investment. A current, secure bank interest rate would be used when there is almost no risk. Increasing the rate by 2 to 4 times is not unusual in forestry or wildlife operations where many years and long-term investments are often required and risks of loss are high. The concept is to equate likely returns from investing in a risky situation with those from investing in a secure (low-risk) situation. Risk needs to be included in estimates of tree value foregone for wildlife benefits from permanent clearings.
The wildlife clearing in the forest is a special management practice that can be useful to many species. Specific recommendations:
1. Be sure of the species and life groups for which it is designed.
2. Costs of development as well as wood production that is foregone are high. They must be well developed and well placed.
3. In many cases, edge, not area, is needed. The doughnut, not the hole, is important. Therefore, the area should be as small as possible and edge as great as possible. Mowing and planting within the area can diversify the areas and edges for insect production. Edge will be discussed later in this chapter. See Figs. 7.7-7.11.
 |
| Fig. 7.7. Small clearings made or preserved for wildlife in forested areas can be very beneficial. They provide seasonal foods and other needs. For some species, area and food per unit area are to be increased. For others, the edge is important for foods and nesting sites. Depending on objectives, species, and available resources, individual wildlife clearings can be made very productive of a variety of forest and edge-dwelling animals. Raptor food and nesting places for turkeys and grouse are evident objectives. Mowing and fertilizing strips provides vegetative diversity and reduces maintenance costs per clearing. Without maintenance, clearings will revert to forests. |
 |
| Fig.7.8. Edge per unit area needs to be studied. The amount is usually very low. Some interior forest species may be harmed by edges but many are benefitted. Managed edges have little effect on the interior species. Management is not a one-time operation. |
|
|
| Fig. 7.10. Faunal edge may not be a line (solid line shown as E) but a tunnel, " a volume", to be estimated by the manager. Animals are likely to respond to the mix or juxtaposition index (the quality of 2 types coming together), and three other factors, namely length, width, and height. Faunal edge has a width of a zone of influence into one type (e.g., a grassy area) W2, that may be different from that width into another land unit (e.g., a forest stand) W1. E is the place of contiguity. Edges are life-group specific and changes in them may result in population decrease or increase. The volume changes by season and over time. Creating and maintaining such volumes is the manager's job in the forest. |
4. Place clearings along contours and in level areas. Minimize shade; maximize ease of maintenance and use.
5. A forest wildlife clearing can be considered a center of managerial activity. Putting many developments in one place simplifies management, reduces costs, and increases the probabilities that benefit opportunities produced will be utilized. Placing these centers, and enough of them so as to meet the needs of animals that must be increased, is one of the manager's decisions. A plan can be created to get orderly development and low-cost maintenance of these areas over many years. Virginia Biologist James Engel once spoke of "intensive management on an extensive basis."
6. Create a diagonally closed canopy at the edges of the clearings. This creates a "wide edge," allows fruiting understory species to be favored, and reduces winds that can cause the forest at the edge to be very dry and to have reduced wood production (i.e., reduces the site index).
7. By planting a few fruit or nut trees at the center, the needs of local animals can be met and the advantages of full-sunlight gained.
8. Where users will be using telescopes or binoculars, long straight clearings may be useful. Where an effort is made to prevent poaching and to reduce the apparent density of hunters or users, then long angled areas are best. Avian predators are suspected to be very efficient when they hunt down the length of long, unbroken clearings.
9. Clearings are especially useful for prey management, including production of insects that are an important food of the young of edge-nesting forest birds.
10. Clearings can be created by cutting the bark around trees (frilling), letting them die, then progressively planting grass or encouraging native plants under their canopy. Often the costs are low if time is not counted. Costs are much lower than when heavy equipment is used.
11. Access to clearings is needed to assure that the wildlife produced there or attracted to them can be used. Where animals need protection (e.g., the young of game birds while feeding on insects) these areas can be closed temporarily to humans. Activities can be encouraged elsewhere.
12. The importance of wildlife clearings created and maintained for many years can be reduced to nothing if someone cuts all trees around it. For at least this reason, a special preserved zone for trees is needed around the clearing. These may eventually be the only examples of old-growth forests. They are little wildernesses, natural areas. Where some trees must be cut, then this is where selective cutting, a well known silvicultural practice, can be followed. Usually the clearing will be used for loading logs so some damage to it can be expected. Because of this, roads and loading decks might be placed and developed at the very beginning of forest planning.
For more about edge and its role, see the section later in this chapter.
Food
"Provide food and cover" is simply to simplistic a phrase for the faunal manager to use. Increasingly, limits are seen; more food will not produce more animals. Cover as well as food are highly area- and time-relevant. Animals are mobile digestive systems. They are as much a function of area as any variable; the larger the forest area, the more the animals. Species richness is related to animal weight; the smaller the animals the greater the richness. Density is strongly related to animal weight. The animals adjust, substitute, adapt to and use up the resources available. Individual components of food (e.g., phosphorus) vary by species and seasonally as do apparent sex-specific, season-specific animal needs (Smith et al. 1978). Maiorana (1978) discovered that California plethodontid salamanders were not regulated by food but by spatial resources. All of these observations are made not to deny the importance of food, only to elevate other factors to equal importance in the mind of the faunal system manager. Other things are important, conditional upon food being present at the right time.
Forests are faunal food. Animals are entropic. Everything falls or goes away. Managers may be negentropic. Perhaps these are the only concepts needed by the forest faunal system manager.
All animals depend on plants for food, the primary component of which is energy. Animals need plant energy. Hundreds of plant forms have evolved to collect energy, store it, and then to reduce other losses from wind and other erosive forces. "Erosive forces" are all of the things that tend to reduce storage, that cause energy to dissipate, to go on its cosmic path to outer nothingness. Entropy is the name of the process leading to maximum dispersion. Every success of plants is in their ability to collect energy and to suppress its outgoingness. They engage in an entropy war, negative to its intrinsic loss, thus negentropic.
Animals, including people, are decomposers. Usually "decomposer" suggests fungi and bacteria, but whether a leaf or some of its cells is dead or alive, besides difficult to determine, is irrelevant. At all times and at all levels there are animals consuming plant parts. "Grazers" call forth visions of range cattle but to the forest faunal systems person it suggests herds of caterpillars, packs of cell-piercing and sap-sucking insects, flights of leaf cutters, and gangs of cambium consumers. Deer eat leaves and twigs. People take the boles. All is temporary. The forest fire is fast respiration; the alternative is slow respiration, animal-aided oxygenation, animal-aided energy-bond breaking, all heading to carbon dioxide and electrons leaving their temporary bondage at the speed of light.
Some trees cannot overcome the attack of a population of insects. They die. Others can overcome because growth or present conditions exceed loss of critical structure or process. It is useful to think of all plants as being decomposed - from seed to final condition. The dead and recently unattached materials (cut limbs, etc.) can be considered detritus and the plants and animals consuming this second plant stage which is no longer fixing energy, as detritivores. The second stage is called litter.
Litter undergoes a complex but predictable set of stages. It, as most other forest processes, displays equifinality - many different pathways to the same nominal states. The tree cavity as well as forest leaf and stem litter, depending on the mix and moisture content, undergoes decomposition, progressive size reduction, thus surface area increases and with it an increase is an attachment surface for fungi, bacteria, and other forms as well as a leaching surface from which minerals move to the soil solution. Mammals and birds break and mix the larger detritus as does wind and gravity. Insects (e.g., termites) further reduce the size and change the characteristics of the pieces. Terrestrial snails, rasp then pass enormous amounts through their systems, reducing the particle size to that suitable for processing by microfauna, themselves fed upon by the micro beasts of the forest floor. All is the same. The scale is different. The system works to release energy and to cycle nutrients. No purpose is implied, only the pervasive, tireless law of flow and cycle.
In the wet stream bed where gravity, wind, hikers, and horses move the litter, other organisms (e.g., crayfish and snails) are early shredders, preparing materials for the last stages of bond breakage and ionization. The forest manager realizes the massive ubiquitous, tireless second opponent - gravity. "Everything winds up in the ocean" is an interesting but imprecise educators' phrase but difficult to relate to the daily affairs of managing forest fauna. The crayfish and large aquatic insects can be instructive as they process litter for lower forms that eventually become fungi food (symbiotic with bacteria feeding at the molecular level) and then - it is washed downstream! The manager's minerals have left the forest system!
The faunal system manager realizes from basic biology and physics that with every transformation of energy from one stored form to another, a large amount is lost. Every plant consumed that results in animal flesh results in enormous losses. The animal producer is wed to such wastes. It is natural, expected, so commonplace that it gets no notice or thought. A similar thought pattern does not have to exist for minerals. The pattern for the manager must become one of: How can I keep my nutrients within the functioning system?
Nutrients are minerals useable by plants or animals. Herein, energy is not included as a nutrient. Nutrients are only useable by plants when in solution, thus the emphasis on soil solution. However, there are plants that live on nutrients in rain water and those that have developed relations for living on plant bark and rocks (epiphytes).
Of course nutrients can be added to a system, but these are costs the manager must bear. There are other energy-use costs that society must bear because now all movement of nutrients is largely uphill, by powered vehicles. Much nitrogen fertilizer is natural gas related, and fertilizer processing is energy expensive.
Keeping nutrients within the functioning system requires a state of mind that includes a tentative and always-failing:
"Prevent erosion" is a battle cry in most forestry and agricultural circles. Besides the overt damage, costs of repair, and water system problems, water and soil erosion are displays of nutrient loss and loss of the "container," the potential volume of the soil solution. Some river and stream communities are said to depend on the annual or periodic addition of nutrients in flood waters. Dams can block these nutrient and water pulses. It seems reasonable to attempt to sustain water levels where natural forested communities are desired. Rarely can a forest utilize all of the nutrients provided by such flood waters. The nutrients serve best over the long run when retained high on the land.
Energy is clearly a central theme of effective faunal system management (Moen 1973) but as elsewhere, efforts to simplify often violate concepts of total and highly interactive systems and are not sufficient.
Not minerals but nutrients; not nutrients but specific nutrients are needed in each subsystem. Available nitrogen is often in short supply in systems. Much of the biologically useful N is "on the hoof," bound as protein, and the system is thereby limited. Nitrogen-fixing bacteria are ponderous creatures, specific to a few plants and sensitively excluded from many sites. In many natural systems, N is the substance in short supply. To change that shortage will require imports - at very high costs. Other systems are similarly limited by other nutrients. I hypothesize that shortage of biologically-available calcium limits many forest faunal systems. I suspect understanding the calcium-mediated carbon:nitrogen ratio will help us understand the basic system and suggest points where the manager may intrude.
Space, not just area, but volume, is a factor with which managers may work. Animal behavior is another. "Food" is the topic when the manager produces energy for animals. Just as important to the population manager is preventing consumed energy from being lost before it can be effectively used by the population. Human disturbance can case energy losses from animals. It is often less costly to control energy losses than to produce energy in foods. Similarly one type of cover, that which prevents convective heat losses, can often be more readily produced or retained than can palatable food supplies.
Food is an animal's "resource." Food may be an animal itself, prey. A forest manager may plant grass in an opening or along a logging road, thus producing food. A snag left in an area provides a surface area for insects (CAP9057) which are food for one species as surely as blades of grass are for another. A stabilized road bank prevents siltation; feeding-surfaces on stream rubble are thereby retained for fish as surely as snag bark surfaces are retained for birds, small mammals, and tree frogs. Food produced or food retained - at what cost - to meet prespecified demands for a set of faunal life groups is the manager's question.
There have been hundreds of animal food studies. Hundreds of thousands of plots have been clipped. From them we know that plots are variable, weight of plants changes seasonally and over time, some plants are never eaten, some plants are infrequently eaten, and some are never eaten in one plot but always in others. We also know that a gram of dry food has about 4 kilograms of energy, no matter what the composition of plant species. Some plants are digested well, others poorly. Some plants substances (e.g., phenols and tannins) retard digestion. It is a lucky faunal system manager who is responsible for only one life-group. Few are so lucky; most must concentrate on 10; most should deal with a minimum of 200 and all creatures have different food needs and food-getting capabilities. The variability is overpowering; the magnitude equally so. Research can help, but hardly more than a drop of water to a person dying of thirst. Given the finding that equifinality stalks every study, the manager must act, aware of the kindly, caring, adaptive feedback that will be used in the instant after action.
Knowing that a pound of dry food has about 1800 kilocalories (a kilogram has about 4000) suggests that managers need to estimate the food available using plots or transects. They need to clip and weigh samples of food-like plant parts. There have been many techniques to simulate foraging. Clippers with a gauge that clips stems of a known diameter determined from twigs observed to have been eaten by deer is a good example. Good for research, such measures are too precise for the manager. Usually needing to draw a sound conclusion only about 2 times out of 3 with an error of no more than 10 or 15% (see Wiant 1987), the manager can take a sample of n based on s2/d2 as explained above. When low confidence is suitable, t2 = 1, so is not included. Once the mean and variance of a few samples is determined (CAP164), then n can be quickly determined. The weight is critical, mean, median, and variance. The weight in a small number of plots will rarely be normally distributed. The median is of greatest value. It is possible (and desirable) to run the samples through a lab to get digestible energy and protein but in the final use of the results lies the reason for the suggestion here. Do not do it! Objectives are poorly known, values poorly conceived, "risk level" unheard of, and substitutability likely. Precise data will not solve a grossly formulated problem.
One part of my advocated rational sloppiness is the concept that when all of the computations are completed for a system, the last step is to round the answer down to the lowest number of significant figures (a fundamental rule of arithmetic). For example 3/1.4321 is 2, not 2.0948. When objectives are rarely expressed with values more precise than 0 to 9, then one significant figure is all that is allowed in the output of the final systems analysis. This suggests that quick analyses of small samples may be managerially appropriate, especially if repeated often. Managers, as statisticians advise, need to stratify, to sample differently where vegetation is different. Random sampling is advised primarily to prevent bias. The manager is caught between so many contending forces and likely counterintuitive results from faunal space manipulation that he or she would not dare risk biasing a sample! The need is to take representative samples. Any manager knows how much of a clump of grass deer usually eat. That is what should be clipped! Then count clumps! Managers know what size twigs moose eat; clip that size! (So what if in bad years moose do "walk down" shrubs between their legs allowing them to eat excessively large stems from the tops. These are the outlier events, the rare ones, and have little bearing on analyzing the forage in an area. Even highway designers do not prepare for peak traffic conditions! No one can afford to be prepared for all extreme conditions!)
After clipping, drying, weighing, and analyzing weight per unit area of food in mapped areas (and this will be continued, unmindful of the above arguement), a manager can estimate how many animals or life groups such weight might support. With home range data (H) (CAP629, CAP630), it is possible to compare an estimate of the number of animals N1 that might be in an area (A) by simply dividing A by H. Based on weighed food (f) per unit area (a), a similar number can be estimated since total food (F) is Sfa and since annual food consumption (k) for most animals is known or can be estimated, then N2 is F/k. Comparing N1 with N2 in a simulation mode of "what if I change this factor, then what... ?" questioning can suggest whether potential animals have an adequate food supply or whether food may possibly be limiting.
Detailed energy analyses can and should be done but in the penultimate analysis, the real question is: how is food available in the field related to harvested animals per hunter hour or animals seen by resource users per hour of effort. The question can be as simple as what are the approximate values of the coefficients a and b in the regression
Y = a + b log (F + 1.0)
where Y is harvest or animals seen and F is available food (See Green 1979). By studying various areas, watersheds, hunt units, etc., a regression can be readily developed (CAP110 or CAP71). The results is called the manager's black box. In goes food and out comes animals, but the intricate interplay of all of the factors is hidden, perhaps not even known. It does not appear very sophisticated, very "scientific", but given the extreme variations, poorly articulated objectives, likely dynamics of environment, population, and people, it is a rational step to take. Given that an equation can be created, perhaps with three or four profound variables (and an R2 of greater than 0.6), then the manager can proceed to manipulate the right-hand-side variables (or explain that foresters are doing so) and the consequences of doing so to the desired life groups.
Not without other information, the manager can use techniques outlined by Kirkpatrick (1980) because the animals themselves are integrators of the environment and express it in their weight, fat, blood, etc.
A useful technique that can help determine a theoretical maximum amount of forage is in listing forage by plots, low to high. A leveling off usually occurs, one much like a maximum flood depth occurrence curve, and the upper amounts can then be used to simulate the effects of the occurrence of such amounts area wide. These outer-bound estimates (including a non-zero minimum) should be useful to the manager in producing estimates of animals that can be supported, when populations may consume excessive amounts of available food, and when populations can be expected to "explode" or level off.
There is hardly any more complex issue in forest wildlife resource management than food production. It starts with the economic premise that the manager can only "count" what he or she produced above that which would be produced naturally. The second premise is that extra produced food needed to meet demand can be counted. No success points are gained by over production or by counting "Nature's gifts."
The additional problems in understanding and working with food supplies for forest fauna are numerous. They include:
 |
| Fig. 7.11. Seasonal occurrence of mast (fruits) can be shown by a line of "presence" (A,B, and C) or more explicitly by curves of abundance for each species over a year. |
 |
| Grasses and forbs in wildlife clearings and roadsides have predictable change within a year. |
Gallina et al. (1981:141) showed that deer food consumption shifts among trees, shrubs, and herbs seasonally. Reasons may be interesting but the managerial reality is that the interplay of innate animal physiology, plant succulence and availability, plant nutrition, and the presence of secondary "harmful" substances result in a food supply that probably has little relationship to the quantities of plants measured in conventional sample plots. Deer (Hagen 1953, Klein 1970), lab animals, and presumably other animals (Miller 1968) show clear preferences and can select plants and plant parts with high nutritive value.
Food quality is limiting to deer (and most other animal forms) and may be more important than is quantity (Short 1972). Substance, such as nitrogen, is the factor of managerial interest. Amount is the measure, but it must be balanced with availability and the cost of collecting and processing it by the animal. Amount in a unit weight of plant matter is the observation. Weight per unit area and total area are the coefficients, the weighting variables. "Food" is an example of how a commonly used word can lead to deviant concepts and to mismanagement. The conspicuous plant or stands of plants may bear little relation to the needs of animals measured as weight of available metabolizable nutrient or energy per unit area. "Metabolizable" is an important modifier. Even though nutrients may be present and available, and consumed, they may not be metabolized because they are bound up in lignin, mastication is poor, animal health may be poor.
The complications continue to build. What is required and what function each nutrient plays is related to animal age, sex, and status (e.g., growth, pregnancy, lactation). Short (1972) observed that female deer gain approximately 80 percent of their mature body weight in their first year. Nutrients ingested afterwards are largely for maintenance and production of young.
As a result of these sex-age-status differences in forage needed and consumed, an area may be good, poor, or bad for a particular population at a particular time. Food may be in surplus for 100 animals, but for another 100 of different sex, age, or status, it could be inadequate.
Nutritional value or contribution of food to an animal must account for the condition of the animal. Where blood urea nitrogen may be a good index from animals suggesting a high protein diet, the same results can be gotten from the blood of starving animals. The body mobilizes its own protein for use as energy. A non-linear relationship is evident. Interpreting it is difficult and requires consideration of other factors.
Forests are faunal food.
Animals of all sizes and types consume all of the plants and plant parts before, during, and after the dominants, the trees, grow. Then animals eat each other. Usually the faunal system manager will give most attention to animals' energy intake. This is the one factor or phenomenon to which all population systems are extremely sensitive. There are many other phenomena, all interrelated, but when it comes to a managerial decision with limited time, money, staff, or research, the rule must be, at least 85 percent of the time, supply needed energy. See Moen (1973). Next, after energy, the managerial rule is to supply protein, but realizing that protein is catabolized to energy in stressful conditions, the answer is the same ... supply energy. Next I believe more attention needs to be given to available calcium, especially the rate of calcium in the thermoregulated brain and to the calcium-to-phosphorous ratio. High levels of Ca or iron reduce the availability of phosphorus to plants, even if present in the soil. Soil N needed by plants in protein building is inversely related to mean temperature, hence soils in hotter areas have lower levels of N. Forages on these areas have lower protein and phosphorus levels.
Food has many dimensions. It, itself, is a subsystem. The analysis begins with presence or absence) as in a plant or prey species, then may proceed to numbers of species or forms present (richness), then to abundance of each species present, then to total dry weight of each species present, then to the statistical variance in weights among species groups. The total amount of food present must be weighted by (1) richness and by (2) the variance. The species-specific or life-group-specific (i) food (Fi) model becomes:
Fi = a Bi + b Ri + c Vi
with B being total dry matter biomass of life groups known to be consumed by species or life-group i; R is the richness or count of species consumed or known present, and V is the variance in dry weight among all food types. These are kept separate (not put into some synthetic "diversity index") for analyses. A non-linear form is likely to emerge after more study but here the concept is important. The food supply for the manager to produce is not simply biomass but that biomass distributed among many life groups and having great variance (low evenness; extremely small and extremely large supplies) - all to meet the changing needs of a population or life group over time (say 5 times the ecological longevity of a species). The manager must discard a one-year view of their work; adopt decisions made for at least more than a decade, preferably 50 or more years.
The food concept becomes more complex. Food present may not be available. "Behind a fence" is a sharp example of unavailability but not very meaningful in the forest. Steep slopes, river barriers, and extensive blow-downs can make forage unavailable or costly to animals. Availability can modify B, R, and V in the above equation.
A manager may plant a new species to influence B and increase R by 1.0. Because of habit or knowledge, the fauna may not consume it. Known-consumption needs to be included in both B and R above.
Managers need to resist trail-side observations of foraging since animals are known to consume things there not needed and that are unpalatable. Some animals are known to have pointless behavior. Well-sampled vegetation will provide a more meaningful expression of what is eaten and probably needed. The manager needs to assess all plants or fauna, then give each a probability of consumption by life group i. Such probabilities may vary from zero (never), through 0.1 (only along trails) to 1.0 (always evidence of some foraging if i is present). This is a quantification of palatability.
Once consumed (ingested), food is digested (or not), and metabolized (or not). Rabbits, for example, pass large amounts of un-utilized energy through their bodies. Eating their own feces, caprophagy, is a reasonable strategy in getting energy. Ungulates "chew their cud" getting energy and other food contents through a prolonged period of mastication, help from a large internal population of microbes, and metabolism. Chemical blocking factors in some plants materials (e.g., tannins) make foods ingested unutilizable (e.g., Servello and Kirkpatrick 1987).
The food concept becomes one of measuring the totals present (e.g., phytomass), than subtracting from it that which is unavailable, too distant, undigestable, or blocked. This does not provide an estimate of the animals of a species or life groups, only the number that an area can support in a season (Hanley and Rogers 1989). Most managers hold the assumptions that "nature abhors a vacuum", that "areas fill to their limit" (Allee et al. 1949) and thus knowledge of the limit and how to get to it (or to know when it has been reached) is of paramount managerial interest.
Food Production
Faunal managers work to achieve a stable-over-many-years, net, seasonal, quality-weighted-amount of food for a life group of a desired size. This complex task can be done by manipulating the ages and conditions of forest stands.
"Succession" in this book means the highly predictable sequence of gradual changes over time following disturbance of a site, either natural or that caused by or encouraged by people. There are trends in that sequence, thus making predictable future states highly probable. The last stages occur over a long period and tend to be self-replacing if there are no major disturbances by people. A major disturbance of a land area "sets back" succession to earlier stages or to the very first stage. There is great variability in the pathways and stages due to how far back succession was set, seed and vegetative reproduction structures present, time, size of the area, and influence of grazers. These pathways are limited in number. For example, at any node in a set of pathways, an area may take a slightly different pathway. There are alternatives (e.g., soil exposed for seeds or soil not exposed). The stages differ by SAF type but the name tends to be that recognized as the mature or final, most stable stage. Forests "grow up"; the idea is about as simple as that. In growing up, there are physical changes in root, bole, and crown size, shape, structure, height, stem density, light on the ground, and moisture in litter. Species vary, but eventually become homogeneous with other stands of the same type in old age. Every spot is different; every stand is unique; but there are many (but not infinite) pathways to the same recognizable, consistently nameable end state.
There is little future certainty in anything, thus the future forest stand conditions can be predicted but a particular future condition can only be stated with a certain probability. In the forest, people deal with stands. These are acceptable by some ecologists as ecosystems or communities. That they are is denied by others. I see no need to engage the debate here. Ecosystem theorists suggest "ecosystem development" implying that there is some concept of an end-state or condition toward which development proceeds. A very long-term view, say over about 1000 years, results in a perceived end-condition likely to be different than that of someone with a 200-year managerial view. Perhaps climatic forces over more than 2000 years will predominate, resulting in a named forest of specified composition. Interesting, the premise does not influence current managerial decisions. Succession herein is used to mean the evident changes, possibly categorized for teaching or study as stages, through which the assemblages in an area tend to go over time. The discreteness of each stage is related to the units used, the area, and the magnitude of the range of difference between the disturbed and the final condition. The end state is a relatively homogeneous mixture that achieves a large, stable biomass (having the appearance of purposive self-regeneration). There is no insinuation here of a superorganism, no pre-design notion, no requirement for a singular dominant process such as energy flow, no precise uniform end composition or structure. "Biomass" is the most unifying of terms and meets some of the needs of the wood industry.
By using the word "stage" attention has been shifted from resources of interest to words and phrases relating to dominant things perceived. Lane (1978) observed that the degree of structure (species, biomass, etc.) noted in an ecological system will vary with the taxonomic unit that is considered. Connell and Slatyer (1977) noted that early ecologists were interested in the species invading sites and then others later began to study characteristics of the perceived stages (e.g., biomass, productivity, and diversity, and composition (proportions of species) Drury and Nisbet 1973). A mixture of studies of processes and results became scrambled with things of specific interest such as rangeland forage and the change in density of pines over time. R.P. McIntosh (1980) noted the elusive synthesis of the opposing concepts of succession.
No synthesis is attempted here. "Succession" is used in a general way to mean time-related production of a resource of interest. The causes are assumed known (an indescribably complex natural system that is the essence of the ecology concept - one of innumerable reactions, changing inputs, and feedback loops including plants and animals as well as the abiotic factors of a site - pure "black-box theory" grounds). Each estimate is strictly site specific. It is difficult enough to deal with the factors for a site; only slightly moreso the unlimited potential "multiple pathways" of succession (Cattelino et al. 1979, Arno et al. 1986); impossible to move into operation the interesting notions of "patch dynamics" (Denslow 1980); )
The time used within the concept is debatable, but is assumed to be simple chronological time, usually in 5-year segments as used in forest inventories (making need for knowledge of any specific year for populations of interest to be discounted. Year-classes or age-groups are used. The thing being observed is the proportion of the maximum attained in any year-class. 
The maximum attained over a 100 to 200 year scope of study has an assigned value of 1.0. Proportions compared to this maximum are estimated for each age-class.
A curve is imagined (or computed; Bailey and Dell 1973). It is expressive of estimates for each age-class, the sketch across the histogram. The vertical axis? Sketch of what? Any resource - ruffed grouse density, quail flushes (Gavitt and Giles 1973) occurrence of warbler X, soil erosion, stem-count, campsite quality, board feet. (It can be used with minimum effort to describe the change in tendencies (with age) of populations of people to hunt, trap, or to engage in dispersed outdoor recreation. Perhaps it can be used to relate to antisocial behavior (Foa 1971) and game law-violator age classes.
From time-zero (the major fire, the bull-dozed condition, the herbicided plot) change is likely and the direction can be estimated. There are real data on some areas (e.g., Stickney 1980, Ehrenreich and Murphy 1962). Wallmo and Schoen (1980) were able to generalize succession for two periods:
Abbreviated, this synopsis of secondary succession of western hemlock-Sitka spruce forest in southeast Alaska depicts a 15- to 20-year postlogging period with abundant but largely inaccessible forage, followed by perhaps two centuries of usable habitat with sparse forage, and a silviculturally overmature stage of indefinite duration with usable habitat and good forage supplies.
Likewise, Lima et al. (1978) were building the picture of the changing forests in West Virginia in their comparisons of data from 1933 to 1971 and were able to generalize about changes in erosion and suitability of the areas for different species of animals. Cunningham et al. (1980) developed the equation:
y = 103 - 2.59x + 0.0267x2
expressive of the percentage of tree snags standing (y) in a series of age classes (x = years) in a ponderosa forest. These are little more than "yield curves", (Curtis 1972) expressive of some resource ... perhaps including cubic feet of pulp.
A future forest is predictable in the same sense that a hydrologist predicts a flood saying that if it rains x, if the ground is not frozen, and if there is no new construction, then the river will be at flood stage F. If a flood does not occur, is the prediction wrong? If the rainfall is y, not x, does a mismatch between the actual stage and F negate the prediction? There are many possible nodes or "if ..." statements in the post-disturbance trends of forests, but in the forest biomes, it is a certainty that there will again be forests, and there is equal certainty these forests will not be "just any forest." They will be of a kind usually found on such sites. Existing vegetation expresses: seedling establishment is more important than conditions during the seedling stage; factors in the young forest determine the future forest, not conditions of the adult trees. In some areas it may appear that "Nature cannot make up its mind"; the end type of stand cannot be predicted. It will usually be one of only 2 forest types and the odds are not 50-50 on the type. Horn (1975) observed that "the most dramatic property of succession is its repeatable convergence on the same climax community from any of many different starting points." The reader may recall previous comments about equifinality. The prediction is: There will be a forest; it will finally be only of type A or B; the probability is greatest that it will be type B. Of course certainty is desired ... but that is for the determinists.
The long view is advocated elsewhere in this book but a relevant concept of succession is limited to several hundred years. I can imagine a nearly homogeneous stand (or a pond) in a forest that has been in mature trees for 2000 years and I can theorize about a universal end state but then I imagine the dynamics of soil, geomorphology, and a set of recurring natural disturbances over that period and suspect that the hypothetical forest has little theoretical utility.
Johnson and Fryer (1989) found that even though a lodgepole, Englemann spruce stand was not self-producing, the population of trees was not transient since fire frequency was sufficient to assure regeneration.
The result of the above observations is a fairly mechanistic concept of changes in an area, one animal or plant or factor influencing the next, a condition influencing the next ("relaying conditions", Egler 1954), all steps along the multiple pathways influencing the constraints of the space (not in "competition"), increasing some, relaxing others, in which certain dominant plants may grow.
| Note: General roles of individual plant or animal populations, at a site, usually simultaneous, and changing over time that result in the changes called "succession" |
| invade facilitate tolerate inhibit die |
The end result after many years is an old stand, an assemblage of plants and animals. We see the dominant plant(s) and call the stand, for example, a red oak stand. It will be unlike other red oak stands; it will be unique! Yet we predicted it 80 years ago. It could have been different if we had drawn the stand boundary lines differently, thereby influencing the percent of the area suitable for red oak, and if moisture and wind direction had been different and it increased the hemlock regeneration from a single adjacent seed tree.
The stand goes through fairly predictable stages. The manager can predict the conditions of the stand better than the occurrence of change-events, the disturbances or, as some prefer, "the perturbations". The manager can influence the stages and the paths through the network. There are costs to such influence (e.g., thinning, replanting) but there are costs associated with having forest Type A when the manager needed Type B for the managed fauna.
Unlike many ecological studies of succession (e.g., what plant and animal species occur after area disturbance and trends in replacement by others over time), the questions of the faunal manager must be more general. The questions may be: how much food (all species) is likely to be available in each stage, or what life stage or species will be present (and often in what density)? The questions are about sites or forest stands but behind the questions are the singular question: what is (and will be) the total resource over the next few years, the planning period.
The manager can influence succession, either by increasing the rate at which it progresses on a site or slowing it down or preventing it. "Preventing succession" is a bit of human arrogance, because even though humans may delay it, the change is temporary. Eventually Nature will have its way. The advanced forest community will one day "win out"; the manager may resist it or encourage it. The old-growth or ancient forest is an example of the advanced stage of development of a site. The recently harvested and chopped pine stand is an example of conditions of the forest stand (though now relatively bare ground) at time zero-plus-1 in the succession.
Succession is readily seen in changing plants. The emphasis needed and implied herein is that many, many changes are occurring in soil, water, faunal light, temperature, etc. - the structural relations, and dynamics of all components. Succession is a summary word for the many predictable change phenomena and pathways (a transition table) that may occur for an area, better yet, for any land volume.
A forest yield curve relating tree volume to age is a succession curve. These curves can be called "production functions" but this is a phrase from economics where it usually has some factor (such as labor) on the bottom axis, not "time." An example of the latter is the occurrence of eagles as a function of increasing elevation or latitude (Grubb and Kennedy 1982).
 |
| Fig. 7.12. A succession curve or production function. A succession curve may be of almost any shape. The verticle axis may be anything! q' can be almost any factor such as available seasonal food, probable sightings of a game bird, or hiding cover. These all change as a stand grows older. At zero age in this graph, a complete forest removal (as from fire or site preparation) is assumed. At the maximum of any curve, m, a value of 1.0 is assigned and other conditions related to this value. These curves might represent deer or elk forage as well as conventional forest wood volumes, the yield curves.. |
Each forest stand has its own succession curve. What is the value for q' on the vertical axis in Fig. 7.12? Succession of what? It can be of almost anything of interest that has been observed to change predictably over time (e.g., wood, animal abundance, runoff). These are estimates and vary so the curve is relatively fuzzy but that is the nature of our knowledge that many things vary in forest stands, year by year (e.g., due to insects, ground fires, precipitation). The values for q' are usually species- or life-group-specific. Food available to tree squirrels will have a curve. There may also be one showing available deer browse. Succession is not a new idea at all. I discovered how it could be used in work on Idaho's forested elk ranges (Giles and Snyder 1970). Since then, the concept and related techniques have been of immeasurable conceptual and practical utility. Here are the concepts as they relate to forest management for a single species or life group. A similar multi-resource procedure will be discussed later.
Succession-Based Management: Steps 1 to 3
 |
| Fig. 7.14. Three curves for different forest resources are shown (from the original report by Dr. Jack Lyon, USFS Handbook, Wildlife Surveys Handbook 110-2, June 1967) |
1. Create a succession curve (Fig. 7.14) for some phenomenon of great importance (such as available late-winter or early-spring forage). The curve can be created from long term studies of tracts but these are rare and costly. Byrnes and Miller (1973) produces similar curves for vegetation on the surface of earth that was cast aside in mining. (See also O'Connell and Brown, 1972:1194) for similar curves for water, wood, herbage, and sediment.) Usually stands of various ages (e.g., as determined by tree-ring analyses or records) are studied and the value of q' for each age class plotted. Then a curve is fit to the points. An alternative way to get such a curve is to interview experts and "old-timers." The questions should be about initial values, time at which maximum values occur, and percentage changes relative to the maximum. Even 3 or 4 points from experts can provide a sufficient "curve" (first of broken lines then rounded to be consistent with natural processes) to allow work to begin and the concept being discussed here to be used. Once the system is developed and its use demonstrated, then refined studies may be commissioned to improve knowledge of the shape of the curves.(In some cases, more such knowledge will be shown to be unnecessary and research-study funds better allocated to other topics.)
2. Determine the maximum point on the curve (m) then call this 1.0 and re-scale all units to this amount from zero to 1.0. Call these values q. In effect, the curve is now an efficiency curve. It expresses the proportion in any year of the average condition of a stand to achieve the maximum production. Potential profits may be relevant to these values of q on some forests. (Probabilities for these proportions may be assigned later.)
3. Multiply this curve (the values for each year) by the number of acres or hectares in each stand.
4. Plot each stand on a graph starting at its known or estimated date of origin. (Use the dominant tree age class in an uneven-aged stand.) Only 3 stands are shown for example in Fig. 7.15. The faunal system manager for a large forest or multi-county area may have to deal with thousands of such stands. A computer is used. The technique can be handled manually for small tracts of about 10 stands.
 |
| Fig. 7.15. Given a small 3-stand management area , each stand area has successional production of food (etc.) as shown. The height of the curve is a function of the size of each stand. Here all curves have the same approximate shape. A tail of a curve from the past is seen coming from the left. A unit is cut. The food for animals is being produced from the three areas. The total food is that produced from the sum of all three curves.The summation curve is not shown here |
5. Add the areas under all of the curves (Fig. 7.16). This produces a curve with values of Q. Because most of the separate stand curves are irregular and are difficult to describe mathematically, integration is very difficult. I discount parsimony here. Numerical computer methods work well and after years of struggling, the curves may be trashed and the entire system yield to a spread-sheet formulation. These curves are little more than bar graphs.
 |
| Note: Two simple curves (perhaps symbolizing one for sparrows, one for woodpeckers). Adding these two curves would produce a straight line. (For an ancient parallel: Mendelssohn 1971).When 10-40 species are analyzed, each with a slightly different shaped curve, it is impossible readily to predict the shape of the final curve. |
 |
| Fig. 7.16. Adding the area under all three curves (from Fig. 7.15) produces a total curve (the dark, top curve) expressive of the total food (or other resources) on the management area over the relevant period. |
As a manager of many tracts in a region, the available food supply is now, with such curves, known historically and the future can be predicted with ease, assuming nothing changes. The following picture available shows the post-war convergence of returning soldiers with high expectations (from 1935 conditions) and low game (populations having responded to low food supplies). It shows why there were low tree replacement rates in the early 1960's due to high deer populations responding to the large area burn (B) in 1945. The pattern in Q is, if it was a graph of a deer or elk population, would be called "eruptive."
 |
| Fig. 7.16a The curve to the left is the sum of many curves from stands. The objective is determined. The trend in forage or other life-support conditions is estimated. The difference (the deficit) is noted and managerial steps taken to reduce the deficit. Often it can be done by cutting timber (starting new curves). |
6. Determine an objective expressed in terms of Q. How many tons of food are needed to meet the needs of your population in your total area over the next 10 to 50 years? The answer is Q*. You cannot decide? Stop until you can do so! How can management be evaluated if you are producing an undetermined amount of food (etc.) for an unspecified population? One gross way to make an estimate is to say "The hunters seemed happy in 1965 (at A above). I'd like to product that much food for my animals." This line can be plotted as the objective Q* in the figure above.
7. The manager must now create succession curves to eliminate the deficit between Q and Q*. A match, no significant difference between Q and Q*, is the perfect managerial condition. Some curves will cause excesses in some few years but long-term good fits to Q*. Others will not be sufficient. (In a small area, such as in the above example of only three stands , it is very difficult to achieve a very good fit of Q to Q*.) Seeking the minimum squared deviation or difference between Q and Q* is a sound strategy to be employed as discussed under Objectives (Chapter 4).
8. Computer programs can be developed to state
 |
| Fig 7.16b. Given a residual forage production (R) throughout an hypothetical region it is possible to start production function curves (samples are shown) using timber harvests to fit approximately a line which is the objective function for a region. Using many very small curves can assure a very good fit. Without computer assistance, great imbalances and a poor fit seem likely. |
 |
| Fig 17.16c. When constant-shaped forage-related curves (single curves highlighted as green) for variable-size small areas are assigned at random for a region, highly variable total forage supplies (red) result. |
After many computer runs, it becomes evident that the summation curve may achieve a regular pattern - at least in period, somewhat in amplitude. For many studies which I have conducted, this matches well with observations of so-called cyclic behavior in animals. Over large areas the periods are probably largely a function of pi (3.14), reflective of the characteristic curvilinear nature of each major succession curve. If so, this would accommodate the principle that animal populations respond to their available resources and space, that regular periods of population abundance do occur (3-9 years), that amplitude varies, and relates well to other theories about the causes of cyclic behavior, namely: sunspots (as affecting plant foods); random number processes; predation (of course related to prey); prey (of course related to plant food); disease (related to foods of both types), etc. As May (1985) simplified the Lotka and Volterra relationships, the period of predation T is approximately
T = 2
(ab)0.5
 |
| Fig. 17.16d. Timing of timber harvests and changing the area of each harvest can produce forage supplies within a forest that are declining. This may be an objective or simply a picture of historical changes. |
This procedure outlined so far,the concept of the aggregates, is all grounded in knowledge of ecological succession, economics, objectives, statistics, seasonal food needs, history, and computer capability - a nearly total system.
9. Not yet total, the concept of succession curves and aggregate production can be used in multi-species management, for managing many life groups, and in multi-resource management. These are the quest of most forest faunal system managers and those interested in policies called "ecosystem management" and "an ecosystem approach." Each species or life-group has a different succession curve. Fig. 7.17 shows extremes. A recently clear-cut forest is good for
 |
| Fig. 7.17. Generalized succession curves for extreme faunal types and one forest. When a stand is brought to time zero, it typically is good for some creatures or resources, bad for others. Relative goodness can be described by the curves. |
10. After re-scaling each curve to 1.0 as described under 2 above, then the relative value or importance of each animal species as described under "Objectives" can be multiplied by the curve. If some group of decision makers judged tree squirrels to be 3 times more important than sparrows, then the curves would appear as in Fig. 7.18 and the sum of
|
| Fig. 7.18. The two weighted q curves (squirrels 3 times more important than sparrows (as a group)) are shown |
 |
| Fig 7.19 The sum of the 2 curves, the probable benefits over time, are shown next. In this illustration, benefits are to be experienced years later. |
Fifty or more fuanal species clearly related to forest succession and importance-weighted by citizens (see "values, risks, and demand" in Chapter 4) can be processed in this way. Even more may be used, but it is rare that people can discriminate in importance among the rodents, the salamanders, and frogs. They discriminate among some birds but rarely among the woods warblers and field sparrows. There are clear species values (e.g., the pileated woodpecker, the barred owl, the wild turkey, the ruffed grouse, the woodcock, etc.). Each species or life group has its own curve. The wild turkey poult, for example, representative of a life group, has a curve that is conspicuously different from the adult turkey. Each can be assigned a value relative to each other. (Assign the most important a value of 1.0, the others equal or lesser amounts. Eliminate all having zero value. If a species has negative value (e.g., a pest), define the succession curve as its inverse value. The species will remain (since all are protected, and richness, "ecological balance", and such stuff are desired) but its "good" curve is expressed as, for example, (1.0/population density)).
Species with uniform abundance over successional time (there are few) may be eliminated from some analyses since they will not participate in discriminating about how much or when to cut stands. When multiple types of forests are used, these species should be retained. Species for which no information on succession curves or those for which no known technique for changing succession to harm or benefit them should be eliminated from the analysis. They are by definition of "analysis." (It is not that we do not care about them or wish for data, merely that without information about them and their curves, the analyses and resulting prescriptions will not change.)
The results of summing all weighted species or life group curves (those produced by the manager harvesting wood in just the right year over just the right areas) is usually a peculiarly-shaped and usually area- and human-population-specific curve.By carefully selecting the timing, area (height of the curve), and shape (based on type and site conditions), a very good fit of the line Q and Q* can be achieved. The computer uses the composite curve with the weights for each animal to decide timing and size of area to be cut to fill in the deficit space on the graph between the current forest conditions in producing the faunal resource and the desired condition. The above procedure has been used for 40 species using small clearcuts and is manageable on a personal computer (Waldon 1987). See HAB10.
11. The manager must live with a concept that I believe is unique to those working with faunal systems. The concept is that of action taken to create a system that has potential for meeting human needs or achieving their objectives. The assumptions are that succession is relatively fixed, that animals will fill the spaces provided, that home range, territory, and similar known faunal behaviors are changing only slightly over the 50 to 100 year period in which the consequences of a harvesting (or burning, herbiciding, bulldozing, etc.) act are to be expressed. The forest faunal manager temporarily ignores poaching, predators, and catastrophes. The act of bring an area to successional time-zero is made for a probabilistic 100 years or more. Micro adjustments can be made and are expected later. The time-zero act, the decided-upon system disturbance to bring it back to some pioneer or early stage is the profound managerial act.
12. Making adjustments in the shape of the curves can be done and is significant. Some argue that the shapes vary so much that "succession" is not a useful word or concept. It is unpredictable. I disagree, but argue merely for a general standard concept, a "base datum of normalcy" as found in wilderness (Leopold 1949), and a means for discussing general concepts that can be made explicit in the curves described above. I contend that an "intended curve" can be drawn by most forest managers for their units that are reflective of soil and other conditions and can even reflect the saw-tooth changes likely from thinning, pruning, fertilizing, and other practices over a 50 to 100 year period. These are the curves to use.
There are 13 interactive major ways to alter succession, to change the shape of the curves , speeding up, slowing down, and changing the transitions within the system. (We return to Step 13 below.)
 |
| Fig. 7.20. A standard forest yield curve may have several site index curves showing predicted growth based on differences in soil, moisture, and other factors. A succession curve can be changed by using one or more of the F actions; the results is a difference in site, thus difference in faunal conditions. Here, in B, fertilizer speeds up the food production, but causes it to decline more rapidly due to shading and other factors. The larger area under the curve may answer: What is the best action to produce food over the next 100 years? |
"Habitat management" can be seen primarily as working with the F13 interaction set.
With little forcing, the classes of actions all begin with the letter f. See CAP157.
Returning to the manager's steps in influencing a system of succession curves, we come to Step No. 13.
13. The shape of curves (Fig. 7.20 above) can also be influenced by several major phenomena of spacing. These are discussed in other texts as "edge", "interspersion", and "juxtaposition", and more recently under the gross banner of "landscape ecology." Each is discussed, then we return to the process of managing succession.
Edge
Edge (Fig. 7.10). is an important factor in analyzing faunal space (Harris 1988). Certain animals are positively related to it but not all animals. Unfortunately, "edge effect", a phrase for increased game density of a few species at edges, has been badly used and overextended. Edge of certain types (e.g., between a grassy field, the permanent openings discussed above, and a forest) may be good for some species, neutral, even bad for others. Some species are most abundant in edge communities but "edge" cannot be said to be good for "wildlife." Edge is certainly bad for a few forest-interior-dwelling species and may be excellent for predators, increasing predation locally. It may also favor a few nest-parasitic species, but as always, there is a value implication to such statements, suggesting interior-loving birds are better than others, or that birds with a notable nesting strategy are less desirable than others. There are many observations about edges and wildlife but there are so many variables and the estimates of populations are so limited that it is very difficult to establish a sure relationship between edges and life group abundance. It may be that observers frequent edges, thus biasing observations. The following dimensions need to be analyzed and their effects clarified and separated.
Any manager will always be humbled by the awareness that at the outside border of the map or designated management area there is a vast unknown or at least a vast domain out of direct managerial control. The boundary or system context problem remains, and it may be dominant!
Edge is a line on a map between two stands or land unit types (e.g., between a pond and a forest, between a white oak stand and a white pine stand). It is a simple word for a line created where two significantly different areas are contiguous. For the faunal system manager, edge quickly becomes a complex idea. It is a volume, an imaginary box or tunnel, having length, width, height, and quality and its own dynamics (Fig. 7.10). (Edge is so frequently considered as a vertical phenomenon (a plane between two forest types) that I shall not attempt to discuss horizontal edge, the theoretical surface between forest layers, the effect of which I hypothesize influences species richness, abundance, and species-area relations much as does vertical edge.) The manager, convinced that edge influences a species of interest, may manage to maximize this edge volume (CAP9064).
Quantitative measures and estimates of edges can provide faunal system managers numbers to which many animals are correlated. Correlations are both positive and negative; some animals seem unrelated to it. "Edge effect" is a loosely used phrase suggesting that edges produce animals. It was originally used to suggest that greater species richness and abundance occur at edges than in either type comprising the edge. This is not a universal phenomena. Some edges have fewer species and less abundance than either type. Edges need to be addressed as physical, structural elements of a forest. Life-group occurrence, abundance, and dynamics may be evaluated relative to each element.
Edges involve many phenomena and they need to be separated by the manager. Edges can be readily created by managers. They are one of the easiest things to see, quantify, and change in the forest. Knowing what is really done by timber harvest, blow down, or fire is needed for a manager to make precise decisions and to debate effectively false claims by people asserting the desirable "effects" of ill-advised harvest and other land use practices.
It is no longer a secret that some habitat work is done for people, not the animals. People need to see things to which they can readily relate. Openings in the forest and the false over-generalized notion of edge effect is one conspicuous phenomenon to which the uninformed may relate.
Length
The length of edge around a polygon can be determined from segments of the surveyor's equation (CAP132), from a map using various instruments, and directly in the field. Brunt and Conley (1990) studied edge and patch relations in a mapped area using different scales and "windows" for evaluating edges. The context of the system, its size, shape and precision used is very much related to how edges for faunal areas should be evaluated. The shape of any area with minimum edge is a circle. (Here circumference is edge, E.) The relation is
E = 
D or 2 R, where D is diameter, R the radius. The importance of this relationship is that if a manager is creating areas (e.g., by clearcutting) and maximum edge is desired, then a standard for comparison is an area conceived as a circle. The algebra is straight forward:
A = R2
thus
R = (A/ )0.5
and since
C = 2 R
then
C = 2 (A/ )0.5
or
C = (A/0.07958)0.5
or
C = (12.566A)0.5
and C = E since we are calling the circumference, C, edge, E.
Where one acre is 43,560 square feet, then the circumference of a circle is 739.8 feet. The edge length of an acre (A) as a square is 834.8 feet. The square acre has an edge 1.13 times greater than a circular acre (CAP9062). Computer units allow comparisons of circles (CAP2007, CAP2008), ellipses (CAP670), squares and rectangles (CAP2006, CAP2011), triangles (CAP642), and hexagons (CAP172, CAP506).
Several people have studied habitat shape (e.g., Miller 1969, Fried 1975). Patton (1975:172) used the perimeter index saying it was "the ratio of circumference to area of a circle" and is given an index of 1.0. The larger the ratio is above 1.0, the more irregular is the area. The ratio from Patton (also quoted by Thomas et al. 1978:96) is the same as that used by geomorphologists (the circularity index) and by limnologists to express lake shoreline irregularity (Orth 1983). It has been used by fire fighters to estimate the perimeter of forest fires (i.e., fire line length) (Forbes 1956: Chapt. 7, p. 29). Probable fire perimeter indices are 1.5; maximum ameboid-like burned areas have indices of about 2.0.
Patton (1975) proceeded to define a spatial diversity index from the shape relations where
D = e / (12.566 A)0.5.
(See C above.) The relative diversity index is the ratio of all edges of the outside perimeter and interior edges to "perfect simplicity", the circle using this relationship of D. D* is an expression of the extra percentage of edge in the area over that of a circle of the same area and is
D* = (D-1) 100,
where if D is equal to 1.3, then D* = 30 percent.
Thomas et al. (1978) attempted to distinguish between inherent and induced edges. Inherent are those places where plant community meets plant community. These are "given" to the area. Induced edges occur when successional stage meets successional stage or condition meets condition within communities. These may be produced by management action. They are largely a function of age of the stand. I prefer to emphasize merely significantly contrasting places and to describe all dimensions of each edge volume.
Note that D is computed and may be very large. The area is bounded by an ownership boundary, an invisible "edge" in most cases, but this is the area being studied. Other indices will likely be more useful.
A related measure useful in some comparisons is that of some area to the area of a circle having the same circumference (C) as the measured perimeter. Where C or the edge is measured, then
A = (C/2 )2.
(See CAP9062.) If a 1-hectare area is to be placed within a forest as a rectangle, its ratio of edge to that same area as a circle might be 1.66 resulting from
E1 = Eactual/Eminimum.
In general, where edge is desired for the animals under management, then E1 should be maximized. (CAP9060 and CAP9062 allow experiments with area and E1.)
All areas are not smoothly circular, hexagonal, or rectangular. Table 7.6 shows the relations of the common structures. These can be used similarly to circles as standards or bases of comparison. Some people like to have a maximum (rather than a minimum) as a basis for comparison, e.g., so a score of 100 might be used in communicating results. This standard is difficult to obtain. Arriving at an evaluation index requires deciding on a minimum width of a
| Table 7.6. Basic geometric relations in forests, clearings, and water bodies. |
|
Symbols A = area d = diameter r = radius of circumscribed circle s = length of a side g = one axis (diameter) of an ellipse h = the other axis of the ellipse c = circumference or perimeter = 3.14159265
Circle A = r2c= dC = 2r = C/dC = 2( A)0.5 = 3.545(A)0.5A = nsr/2 A = (Cd)/4 = 0.25(Cd) r = C/(2 d = 2(A/ )0.5 = 1.128(A)0.5
Area of an annulus, "the doughnut" or exterior zone Ellipse = perimeter Polygon = perimeter of a regular polygon |
| Sides of Regular Polygons (n) | |||
|---|---|---|---|
| 3 Triangle |
4 Square |
6 Hexagon |
|
| A / s2 | 0.4330 | 1.0000 | 2.5981 |
| A / R2 | 1.2990 | 2.0000 | 2.3776 |
| A / r2 | 5.1962 | 4.0000 | 3.4641 |
| R / s | 0.5774 | 0.7071 | 1.0000 |
| R / sr | 2.0000 | 1.4142 | 1.0000 |
| s / R | 1.7321 | 1.4142 | 1.0000 |
| s / r | 3.4641 | 2.0000 | 1.1547 |
| r / R | 0.5000 | 0.7071 | 0.8660 |
| r / s | 0.2887 | 0.5000 | 0.8660 |
E = 2 (Area/W1 + W1).
In some regions, W1 is decided based on the height of the trees in the area and 1/2 or 1 times is often used for easy application in the field. See CAP9061.
Since managers may desire a specific amount of additional edge length, then the area to be treated to obtain this needs to be known. Where a minimum clearing width (W1) has been set, then since
E = 2 W1 + 2 W2,
solving for W2,
then
W2 = (E - 2 W1)/2,
and used in the above equations will give the size of the forest clearing to produce the desired edge. Cost per linear unit of edge may be evaluated as total dollars for the acre or hectares cleared and developed per unit of E (CAP9065).
The maximum amount of edge for an area is determined from a theoretical configuration of two alternating different vegetative strips of width W1 through an area or approximately
Emax = 2 A/W1 + 2 W1.
End widths may be included or not depending on outside-area types of vegetation, corners, equipment turning radii, and vegetation inside the area. These maximum-edge areas appear to be utility corridors, contour-farmed fields, or partially-mowed (brush-hogged) fields found on some areas managed for wildlife or for users' convenience and the quality of the hunting or observation experience.
In the field there are few perfect circles. Forest stands or tracts in a large area fit together. They do not overlap. They take on various sizes. If an average size was calculated for all stands and that mapped, a hexagonal pattern would have to be used to achieve minimum edge (Buckley and Buckley 1977). The pattern is commonly seen in nature. The hexagon has the relations to a circle shown in Table 7.6. In working in multiple stands and computing edges in a forest, I recommend replacing the circle with the hexagon as the basis for comparison.
Managers need to have well in mind the non-linear nature of the edge-to-area relations. "The more edge the better" seems intuitive but the more area, the more edge is not a cost-effective slogan. If edge is desired, then managers need to evaluate cost per unit or cost per unit change in the edge indexes, E. Costs are usually measured in units of area, e.g., as in bulldozing, clearing, seeding, etc. Thus, it is important to have a general feeling for how edge length changes as area changes. (See CAP666 and CAP9062.) Observed edge lengths, for example of Fig. 7.21, are entered into a table (e.g., Table 7.7).
|
| Fig. 7.21. An hypothetical map of a few areas to show the source of Table 7.7. |
A small representative management area with five stands or significantly different land units is shown. Length of edge and other measures will be described. The symbols A to E represent different forest types or ages. The unknown area around these tracts is called area F and assigned an average value. (Other comments will be made later about this unknown area and its effects on accuracy of estimates about interior conditions.)
| Table 7.7 The approximate units of length of each edge in Fig 7.21 are entered into a table. | |||||
|---|---|---|---|---|---|
| A | B | C | D | E | |
| A | 0 | 3000 | 4000 | 0 | 0 |
| B | 0 | 0 | 4000 | 0 | 1000 |
| C | 0 | 0 | 0 | 5000 | 6000 |
| D | 0 | 0 | 0 | 0 | 0 |
| E | 0 | 0 | 0 | 0 | 0 |
Edge-Volume Width
Just as edge length has been analyzed, similar analyses may be done for the areas next to the edge line. These are the zones of influence of the line, the areas in which animal abundance or richness, if it is to occur, will occur. The width can be viewed from two perspectives. One is that of a wide area of influence. The other is simply that in the field, when an edge is seen, it has width. There is almost no way to draw a line on the ground where the edge really is. It is a belt, a zone, and may be 1 meter wide (in a few very precise situations like a mowed field adjacent to a forest) or 10 meters wide (an extreme) where special conditions are managed for wildlife. The area is too small to map on conventional maps but it is a distinctively different area, unlike either contiguous areas.
Where two different areas are contiguous, there are two widths. They may be the same; they usually differ. Width may be estimated by experts or determined by radio telemetry, track and feces counts, direct observations, and instruments monitoring activity.
 |
| Whether woodlots are "forests" may be debated. Factors outside the forest may strongly influence the forest. How far into the forest from the edge reamins an issue for some species (and people). Whether "pattern" influences species within forests remains a question. |
| Table 7.8. The width (in the same units of measure as length) of the influence of contiguity of one type (column) into each stand or area (designated in the row) is estimated. (Read: There is an intrusion from the edge center line from area A a distance of about 0.05 units into area B.) These values may be regional averages or tract-specific. | ||||||
|---|---|---|---|---|---|---|
| Intrusion from Area | Intrusion into Area | |||||
| A | B | C | D | E | ||
| A | 0.05 | 0.06 | 0 | 0 | ||
| B | 0.02 | 0.06 | 0 | 0 | ||
| C | 0.05 | 0.03 | 0.06 | 0.03 | ||
| D | 0 | 0 | 0.05 | 0 | ||
| E | 0 | 0 | 0.04 | 0 | 0 | |
Certain birds nest near edges. Thus the more edge, the more zone in which nesting may take place. Creating forest edge for such birds is evaluated based on Z where
Z = E x w ,
and where E is the total length of edge and w the median width into one of the types. The observations of width can be entered into a table such as Table 7.8 for the area in Fig. 7.21. Some people hypothesize that this zone itself is not as important as the presence of an insect foraging area for the newly-hatched brood. The quality (q) of segments of this zone can be separately evaluated so that
Z = 
Ei Wi qi .
Quality, to be discussed later, allows the computation of Z for the manager to compare management areas over time (e.g., birds harvested, flushed, observed) as affected by treatments. Later we shall add height to form the "edge volume" or tunnel discussed earlier (Fig. 7.18).
Since a circle has the smallest edge-to-area ratio of any polygon, then when a zone has width, w, the area of the zone is the theoretical minimum area of interest around a landscape feature. The equation for the circular zone area, A*, is
A* = (R2 - r2)
and analyses are provided by CAP65 and CAP1008. The interior zone area, one extending into the opening or adjacent stand, must never have a width greater than W/2. If this occurs, an overlap in the influence of edge occurs. This overlap effect, if any, has never, to my knowledge, been studied. Often the ratio of the zone obtained per unit area developed (as by clearcutting) may be used in cost comparisons.
In working with zones, the "doughnut principle" becomes clear. The area of the clearing (the hole) is not important per se; the area in the zone of influence of the edge is what the manager seeks to influence. In zone work, the zone-area to clearing-area ratio is often useful (CAP9062). The higher the ratio the better. The area gained per unit cost is also a useful analysis (CAP9065).
A similar equation for the area of a zone around a linear feature (like a pipeline, stream segment or trail segment, or linear food patch) is
A = w (2L + w),
where w is the width of the zone from one side. The end "semicircles" may be so small as to be irrelevant when L is very long (CAP2011).
Height
Height of the edge box also requires a matrix. A 10-year-old hemlock stand adjacent to a 60-year-old white pine stand does not have the same edge as a 60-year hemlock contiguous to a 60-year pine stand. Heights vary, layers of the forest are involved. A 5-meter stand contiguous to a 20-meter stand gets an edge height of 5 meters. A 1-meter grass-forb clearing for wildlife adjacent to a 40-meter forest only gets 1 meter but if there is shrub growth at the edge, adventitious tree branches, and the beginning of a diagonally closed canopy, I would assign an average height of 2 meters. See Fig. 7.18. Given the height of each stand in Fig. 7.21, the minimum height at the union can be used. The dynamics of edges is exciting to observe over large areas as trees grow, fields revert to forests, or as stands are harvested. The height as well as the type of each unit influences the quality. Areawide edge height dynamics is a major part of faunal dynamics in management areas with long edge lengths. The product of the three Tables 7.7, 7.8, 7.9, is shown in Table 7.10. For each edge-related life group, it is likely that length, width, height or volume will most strongly explain the differences in populations among areas.
Edge Quality
 |
| Clearcuts produced edge length of variable quality or goodness for different faunal species |
An entry of 1 into the table means that areas are contiguous; zeros mean areas do not touch or share a common boundary. Contiguity at a point is excluded in edge analysis. Point contacts are analyzed as corners (described later). A is assumed adjacent to itself. This is a convention. Later when comparisons are made and juxtaposition values estimated, the areas at edges will be compared to similar areas interior to each stand or land use type. The 1's in the diagonal column allow this operation to be performed when a contiguity matrix is multiplied by a juxtaposition matrix (discussed next).
| Table 7.9. Observed or estimated average height of the contiguity plane. Bold values are actual heights of each area. A height of zero as in E indicates a pond or intensively used area.. | |||||
|---|---|---|---|---|---|
| A | B | C | D | E | |
| A | 30 | 30 | 3 | 1 | 0 |
| B | 40 | 3 | 1 | 0 | |
| C | 3 | 1 | 0 | ||
| D | 1 | 0 | |||
| E | 0 | ||||
| Table 7.10. The product of the tables or matrices for length, width, and height produce edge volumes. The value 4500 for A, B is from (3000 x 0.5 x 30). The total edge volume here is 9,520 cubic units. Similar volumes may be computed using CAP9064. | |||||
| A | B | C | D | E | |
|---|---|---|---|---|---|
| A | - | 4500 | 720 | - | - |
| B | 1800 | - | 720 | - | - |
| C | 450 | 360 | - | 300 | 180 |
| D | - | 250 | - | - | - |
| E | - | 240 | - | - | - |
| Table 7.11. A contiguity matrix shown for the management area in Fig. 7.21. Notice how an interior-stand index can be derived from column 6 and how stands with great contiguity can be selected from the rows. The juxtaposition matrix is general purpose and not necessarily specific to the tracts being analyzed on this map. The heights and widths (the distance from the edge into a stand or land unit) are tract specific. CAP9064 computes the interior edge volume and the outside volume. Area F, all outside the area of interest and about which there is usually little information, also has an edge volume. | ||||||
| A | B | C | D | E | ||
|---|---|---|---|---|---|---|
| A | 1 | 1 | 0 | 0 | 1 | |
| B | 1 | 0 | 0 | 1 | ||
| C | 1 | 1 | 1 | |||
| D | 0 | 0 | ||||
| E | 1 | |||||
| F | ||||||
All lands beyond the boundary of the study or management area are indicated as one class, here called F. There will always be boundary problems as well as the question of where to stop an analysis.
The matrix is triangular because B adjacent to C is the same as C adjacent to B.
There are several practical things that can be done with the contiguity matrix in addition to its uses with other matrices previously developed. The count of the rows produces the first index,
C1 = n.
It is an expression of the richness of types in the area. (Some call it spatial biodiversity.) Animal richness relates to it and local regressions can be developed. Managers can increase the types. The sum of the 1 's in the matrix is C2
C2 = cij.
When working with many species and life-group objectives, in general, the more contiguity the better. Thus, one contiguity index is a count of the filled cells compared to the potential contiguities:
C3 = cij / (n (n-1)/2).
In the example shown in Fig. 7.21, the index is C3 =6/10 or 60%
The divisor may be reduced by the number of cells in the contiguous-to-self diagonal and/or the cells in the outside boundary. The index, C3, can be used as an independent variable in analyzing area relations to population abundance, richness, or changes in them. Managers may study the map to see what one or two changes will increase the index the most for the least money. What minimum reasonable changes in the same area (fencing, planting, cultivation, etc.) will result in a C3 of 1.0 or 100%?
Next, the manager doing an edge analysis will load the matrix with lengths of all types (as done above in Table 7.7). Similarly, the products of lengths, heights, and widths will be loaded as shown above. The total tally of the volume of the physical edge tunnel in a management area is possible. It will be a large number, so expression in terms of 1000's or as its cube root may be used.
The missing ingredient is the quality of each cell in the contiguity matrix.
Juxtaposition is a word that has come to mean the quality of the edge tunnel or volume, its goodness for some life group. Fifty-meters of edge between a grassy opening contiguous to an old white oak forest does not have the same value to some animals as 50 meters of corn field contiguous to a fallow field. Edge is not edge. It has relative value for species and life groups.Edge is life-group specific. That numerical estimate, J, the expression of the goodness or quality of the edge or point volume at some time is juxtaposition. Contiguity is merely a geometric condition; juxtaposition is the assigned goodness to an animal life group of the condition. It is a multiplier, a coefficient, indicating how many times better the edge volume is than a similar volume near the center of one of the types of land unit being studied. The valuation is for the tunnel.
Contiguity can be changed by management. It is now possible to add constraints within linear programming work (CAP5009) to assure that timber areas harvested for wood or wildlife are contiguous or not depending on the perceived objectives (Meneghin et al. 1988). An optimum pattern of harvests for achieving many objectives within large areas is now possible. See also Lipscomb et al. (1987:317-318). Small developments (like ponds, clearings, conifer patches, fruit tree groves, etc.) can have drastic effects on area juxtaposition analyses. Linear strips can tie together many areas, improving an area index by hooking up types so that the index increases. This is the so-called"corridor effect" without the misplaced emphasis on travel routes and other alleged superior values of any thin stand (Stolzenburg 1991). The very same resources (types and areas) can be re-arranged to produce large changes in an area juxtaposition analysis and thus probable animal responses in an area.
| Table 7.12. The relative goodness, the juxtaposition value, of each contiguity is assessed. These values are region wide and expressive of the condition for a life group if any 2 land use types were contiguous. The types shown here are those of Fig. 7.21. Here, F has been assigned a gross average value. | ||||||
|---|---|---|---|---|---|---|
| - | A | B | C | D | E | F |
| A | 3 | 4.5 | 5 | 2 | 2 | |
| B | 2.5 | 4 | 3.5 | 2 | ||
| C | 1 | 3 | 2 | |||
| D | 1.5 | 2 | ||||
| E | 2 | |||||
| F(median) | 1 | 1 | 1 | 1 | 1 | 1 |
 |
| A diagonally-closed canopy to the left; shrub lespedeza in the lower right; opportunities to "day-light" the road side shrubs at the right (to dry the road and improve access and maintenance); and questions about edge width ...Is the road itself the edge? |
Quantifying juxtaposition is difficult. Field people readily recognize "good areas" for this bird, that flower, those mice. These are expressions of the relative value of certain types of edges for life groups. Getting water contiguous to summer foods makes sense. Getting conifers contiguous to or near winter foods also makes sense because of intrinsic knowledge of juxtaposition.
How is juxtaposition, J, measured?
J is the relative estimate usually on a scale of 0 to 100, of the goodness of a condition expressed in one cell of a contiguity matrix.
The sum of the products of the edge-volume matrix and the juxtaposition matrix cells is an area juxtaposition index (Table 7.13). A stand-specific edge volume determines whether the value of a stand to a life group is enhanced by or degraded by edge. See Fig. 7.31. The total edge volume for an area provides a manager a performance measure or an independent variable. Is it increasing over time? Which area is better for species X needing volume? How has or will
Table 7.13. The area juxtaposition index is computed as the sum of the products of all entries for the area. Here the sum is 29,545, expressive of the quality edge volume.
| A | B | C | D | E | |
|---|---|---|---|---|---|
| 4500 x 3 | 720 x 4.5 | 0 x 5 | 0 x 2 | ||
| B | 1800 x 3 | 720 x 2.5 | 0 x 4 | 0 x 3.5 | |
| C | 450 x 4.5 | 360 x 2.5 | 300 x 1 | 180 x 3 | |
| D | 0 x 5 | 250 x 4 | 0 x 1.5 | ||
| E | 0 x 2 | 240 x 3.5 | 0 x 3 | 0 x 1.5 |
A forest stand surrounded by other stands or land use types may have an edge volume with a faunal value greater than that of the stand or one that is less. A stand may have one edge enhanced, another degraded. The size of the stand, thus edge-to-area ratio, becomes significant.
The forest harvest strategy (or plan) influences the edge-volume (since it changes with height, length, and the zone (a product of length and relatively-fixed widths). The edge volumes along streams, rivers, and lakes can be similarly analyzed.
 |
| Note: Corners where 3 or more cover types or significantly different stand ages are contiguous (A) usually create desirable conditions for some species or life groups. These are called coverts. |
Corners
A faunal corner or covert is where three or more cover types come together. Fields adjacent to forests may have corners with only two types present (as at B, but these do not have exceptional influence on any forest fauna. Maximum corners throughout an area can be gained from working with stands of approximately equilateral triangular shapes. This configuration is rarely possible in the field (but has advantages in some timber clearcut and haul-to-center operations). Nevertheless, used as a conceptual standard (just as a circle was used as a minimum for comparing edge lengths of polygons) the triangles can be used to compute forest edges as well as corners. If the average stand size (of whatever shape) in a management unit is assumed to be an equilateral triangle, and they were all packed together, then how many corners would result? How does this differ from the current number? Can corners be increased in order to gain certain (specified) life-group increases? Comparison of actual corners to the hypothetical standard can provide a managerial score and suggest improvements (or losses) made with each managerial decision (Powers 1979).
An interesting problem in the interaction of management decisions is that as area triangles become larger, the greater becomes the edge length but the fewer become the corners in a large management area. Determining the proper (optimum) local relationship of the two can be a valuable standard allowing feedback to work.
Like the contribution of the edge volume to area succession curves, the corner volume can be used to modify the size or shape of the succession curve forr a forest stand or tract. Just as the volume of an edge can be calculated and used, so can volumes of corners be computed and used with succession curves (described next). All edge is not positive for fauna, certainly not corners such as seen in B, unless they have a unique condition due to mowing or equipment use, are not usually considered as beneficial to fauna. Maximum quality corners over a large area are gained by strategically placed triangles of different types and ages for certain life groups. Roads may dry some sites, make others non-productive. Campers may reduce fauna by disturbance around a site. Dogs and children may reduce animals where forests are adjacent to residences. Air pollution plume effects are well known, mappable, and can modify succession curves.
Edge has been abundantly discussed. The older concept, juxtaposition (King 1938), is largely replaced by the concept of the quality of a particular edge volume for a life group. How good is area A when contiguous to or nearby to B? Juxtaposition is an animal-based concept, related to energy budgeting. distance of travel, time required, and risks from predation are all parts of micro site planning relating water, dens, food, and other life requirements. For the long-term, larger-area manager, estimates of the edge volume can now be handled (CAP9064). Managers need to work for contiguity that will gain high indexes in the habitat, never forgetting that moving something closer to X may move it farther away from Y. A good arrangement of things makes sense; knowing what is best and when it will occur and how it will change over time are the manager's real problems. Wildlife clearings (permanent forest openings of 1/10 to 1 hectare planted in grasses and grains) have been a useful technique in forest faunal management. They have been judged harshly as techniques for producing deer forage (Larson 1966) (not their design intent). They were intended to produce soft mast, insects, and edge for game bird nesting.
Interspersion
Interspersion as a concept is more messy than edge or juxtaposition. All three terms have all been loosely used. Game seems related to interspersed lands. How can such interspersion be measured? Which is more interspersed, the black and red squares on a checkerboard or the same squares randomly distributed on such a board? A checkerboard seems to me to be perfectly interspersed; mathematical convention and most authors on the subject claim "random" to be the most interspersed. The point to be made is that fauna, at least some life groups, are a function of interspersion, high or low, however it is measured. Trying to get a good measure of it and then an excellent mathematical relationship to each desired faunal group is the essence of management. If
F = f (I)
for example
F = a Iz, with limits
where F is faunal abundance, I is a useful index, and a and z are derived coefficients, then the manager is in control. If he or she increases I, then F is likely to increase. Feasibility, knowledge, costs, or available budgets may all influence the changing of I, but at least the manager knows what must be done-and the limits-when no more change in I will make any difference in F. There is no one definition of interspersion. Randomly distributed areas have their mean size equal to the variance of their sizes, or their mean distance apart equal to the variance of their distance apart. Systematically distributed areas (like the checkerboard or orchards) have a low variance, thus their mean/variance ratio is very large. Clumped areas have a small ratio since the variance is very large. Table 7.14 suggests possible ways to describe interspersion. The ways are highly correlated; the most readily available or easily computed one is the place for the manager to begin; one with a variable likely influenced by a manager is the next best one to use. Precision in this estimate needs to be carefully considered. It can be costly to compute, impossible to change, and it may contribute little to the overall understanding of why Q (the sum of the succession curves) is consistently higher on one forest than another. If the theory of interspersion is correct, the same areas of the same type stands should produce more or less fauna depending on how these tracts are interspersed. Perhaps this is just saying: The larger the high quality edge volume and the more high quality corner volumes for a life group, the better.
| Table 7.14. Methods of computing an index to or expression of interspersion. |
|---|
|
1. Count the total cover types in an area. (This area of about the same size as that has twice as many cover types.)
2. Count types by ages. (This area, however, has more types of significantly different ages, thereby implying a gross estimate of greater interspersion.) 3. Compute and compare mean size of each stand. (The smaller, the more interspersed.) 4. Compute the variance of the area in each stand or type. (The larger the more interspersed.) 5. Divide the mean of the sizes or nearest-neighbor distances apart by the variance. (If randomly distributed, believed by some to be"interspersed," then a ratio of 1.0 will be approximated due to 1.0 being the random condition of the poisson distribution.) 6. Measure the length of edge on a map or photograph. (The longer the length, grossly, the more likely a high interspersion.) 7. Express edge/area. (The larger, the more interspersed.) 8. Make 2 diagonal lines across the area of concern (corner to corner). Count intercepts of these lines with any type edges. (The more the intercepts, the longer the edge, the greater the likely interspersion.) 9. Count the corners. (The more corners, the more interspersed.) 10. Determine the ratio of actual contiguity to potential theoretical contiguity. (When all stands or tracts are of equal average size and shaped as hexagons.) 11. Develop a contiguity matrix; compare the count of contiguous tracts to the number of potential tract contiguities (determined by I* = n (n-1)/2 where n is the number of tracts). |
Succession-Based Management: Steps 14-15
14. Returning to the steps of the process of succession management: Certain areas are set aside for wilderness, future ponds, etc., and are not included in the forest-wide analysis to achieve Q*. They do, however, produce various benefits that can contribute to achieving Q*. They are areas not to be manipulated or treated specifically (e.g., thinning to augment tree species that produce autumn color). The remainder are processed just as above with succession curves for wood (yield curves), inverse erosion curves, or runoff-any forest product or service (e.g., cooling, air pollution amelioration, noise attenuation, camp site quality. These are all succession related).
15. Many species and their likely abundance and human value can be combined using the procedure described above. Similarly, other forest system outputs can be analyzed.
Assuming that a reasonable unit of measure can be agreed upon for each resource (e.g., a board foot of wood, a cubic foot of soil, 1000 liters of water), then relative weights can be assigned
