Forest Mensuration. Brack and Wood


Index
Overview
Help

Quantifying site ©


In forest mensuration, the term site generally refers to the totality of environmental conditions that exist at a specified location. Site is an abstract concept which combines a multitude of environmental factors affecting tree growth into a unified classification.

The environmental factors that influence growth include:

  1. climatic factors, e.g. air temperature, humidity, radiant energy, precipitation, wind;
  2. soil factors, e.g. physical and chemical properties, soil moisture, soil microorganisms;
  3. topographic factors, e.g. slope, elevation, aspect;
  4. competitive factors, e.g. other trees and lesser vegetation, animals, man.
All the above factors combine within one climatic region to create sites.

Site survey and classification are important in forest management for several reasons:

Site is a useful concept in plantation forestry where it is used to delineate areas of different productivity. The factors which have the greatest bearing on productivity in a region are used in the delineation. As these factors may vary from region to region, the variables used to define site may also vary between regions.

In the context of timber management, site quality can be defined as:

the potential of the site to produce timber given a particular species or forest type.
Note in this definition that site quality has meaning only for the species or species composition occurring and being managed at the particular location. Of course, extremes exist which provide absolute limits for all species, e.g. latitudinal and altitudinal range limits for trees.

It is vital that forest managers measure and interpret site reliably because together with stand density, the quality of the site largely controls product size, quantity and value. Thus, it partly determines the investment justified in managing a forest stand. With respect to afforestation policy, reliable site evaluation is of importance for three major purposes:

  1. Evaluation of land for purchase.
  2. Siting of species.
  3. Forecasting the productivity capabilities of planted stands.
Accurate and reliable site maps are essential under intensive forest management. These maps are the basis of all yield forecasting and yield regulation, and all thinning, pruning, fertiliser application and other tending and protection strategies relate to them.

As site maps are used to determine thinning and related logging priorities, they must be made before the time of first thinning, i.e. at age approximately 9-10 years.

Site quality is not easy to assess. The factors of the site and the plants themselves are interacting and interdependent making it difficult to assign cause and effect relationships.

Much effort has been directed towards investigating environmental characteristics in an attempt to find some single environmental factor to serve as a reliable indicator of site quality. Though this approach is practical, it frequently leaves unassigned a sizable amount of the variation in site.

Site can be evaluated in two general ways:

Measurement of site factors

Of the numerous environmental factors affecting tree growth, the important relationship between soils and growth is most apparent. However, the soil characteristics significantly related to growth are not the same everywhere. Relative wetness, sandiness, depth, amount of clay in the A and B horizons, nutrient levels, soil temperature, etc., have different proportionate effects depending on the kind of soil and species involved.

An evaluation of site from soil characteristics has several advantages:

Unfortunately, it is difficult to establish for a particular forest the most important soil characteristics affecting growth. For this reason, use of soil characteristics per se as an index of productivity has limitations. Hence, soil/site investigations usually consist of multiple regression analyses using stand height (or site index) as the dependent variable and a number of soil and other environmental characteristics as the independent variables.

A common approach is to set up the form equation:

Log H = b0 + b1 x (1/A) + b2 x B + b3 x C + b4 x D + ........+ bn-1 x N                                         

where H denotes height;
      A denotes age;
      B, C, D, ... N represent soil or other environmental factors 
        (e.g. aspect, slope, elevation, photoperiod, air temperature,
        thickness of the A horizon, depth to a fine-textured horizon, 
        i.e. any climatic or topographic factor);
      b0..... bn-1   are the constants to be solved by least squares.

If the standard age for site index (normally 20, 25 or 50 years) is substituted for A, then the equation becomes a site index equation:

Log SI = b1 + b2 x B + b3 x C + b4 x D + .....+ bn-1 x N

where SI denotes site index, i.e. Height at standard age
      the other variables are as above, except that
         b0  + b1 (1/A) is a constant because age is now constant.

Czarnowski, Humphreys and Gentle (1971) examined dominant site factors for planted Pinus radiata across a range of geographical regions. They derived a complex equation defining site index as a function of humidity, soil cations, mean soil particle size, the proportion of large soil particles, and foliar (and indirectly soil) levels of phosphorus. They considered that their equation embraced the essential and most dominant characteristics determining site index, with the complexity of the functions reflecting the complexity of factors controlling site index. However, this complex approach has not been extended into forest management and yield prediction.

Soil profile descriptions alone can be a useful basis for defining and predicting site particularly in an area of diverse soils (Turvey, 1983). The growth of plantation trees is often observed to vary with changes in the soil profile. In Queensland, Pegg (1967) and Van Altena (1979) used soil physical and hydrological characteristics and soil colour to classify sites for Pinus elliottii and P. caribaea plantations. Turvey (1980) showed in Gippsland, Victoria, that soil map units based on soil profile criteria are strong indicators of significant differences in stand volume production at age 10 years.

Criteria of climate (rainfall, number of cloudy days, number of frosts, mean summer maximum temperature, mean winter minimum temperature, etc.) have been used successfully for broad qualitative classifications of forests, but these are unsuitable for the quantitative classification of forest stands. This simply implies that the effects of soil factors cannot be ignored in site classification.

Measurement of characteristics of the vegetation

Characteristics of the vegetation which can be used are:

Quantity of wood produced

Since the concept of site refers to productivity, its most direct measure is the quantity of wood grown on an area of land in a given period. However, evaluating site in this way is of limited practical value because estimates of wood quantity are time consuming and expensive. In addition, wood quantity varies with species according to variety or ecotype, stand density, and past management practices as well as with the inherent productivity of the site.

In forests where a standard, well-established management regime is consistently applied, it is possible to use stand volume information as an indication of site quality. The system used in the Pinus radiata forests of South Australia (Lewis et al., 1976) is an excellent example of such a procedure. The specific statistic used by the South Australians as a measure of site is volume per hectare at age 10 years as determined by systematic survey.

Volume has proved useful for site classification in fully stocked or (what are described as) 'normal' stands. For example, volume MAI at 100 years is suitable for classifying site in some forests of Western Europe which have been intensively managed for a long period.

The capability of a forest site to produce wood (plant matter) at any given time depends on:

Thus, a plantation site has a productive capacity only in terms of a given species, and the productive capacity in terms of wood is the most wood that good silviculture of a given kind can obtain from a given site with a given species. Hence, the objective of stratifying a forest into site qualities is to distinguish between the areas which are growing and will grow wood at different rates (e.g. South Australia). The South Australian site qualities represent not only stands of different growth rates but also stands of recognisably different characteristics and thinning and tending needs.

Species of plants naturally occurring on the area

This system is based on the theory of Cajander (1926) that certain key species (indicator plants) in the forest reflect the overall quality of the site for a tree species or forest type. This is not an unreasonable expectation. After all, the species composition of understorey vegetation present on a given site commonly reflects the fertility of the soil, and is often a good indicator of the availability of soil moisture in the upper soil horizons.

Site classification systems based on this concept have proved most successful in relatively undisturbed stands, e.g. northern forest types in Finland and Eastern Canada (extensive forests of a few species), Kaingaroa Plains in N.Z. (Ure, 1959), etc. Ure's system can be used to predict site index for native forest land being evaluated for planting and for inexpensively estimating the site index of established plantations if adjacent undisturbed native vegetation is available, e.g. on fire breaks.

The general applicability of indicator plant systems has several limitations:

Nevertheless, indicator plants can be of general assistance in site evaluation. As pointed out by Cajander, the statistical indices of the main crop (e.g. stand volume and stand height) are effective for classifying the productivity of a specific crop but they do not classify the land according to its potential productivity under other species as can indicator plants.

In Australia, lesser vegetation is used:-

Once coniferous plantations have become established, the lesser vegetation is of little use as an indicator because its presence is discouraged both naturally (changed microenvironment) and artificially (tending, etc.).

Size characteristic of trees (crop statistic)

The most useful tree size characteristic is height. Diameter is unreliable because it is sensitive to stand density. For many species, height growth of the tallest trees in a stand is correlated with the productive capacity of a site and is little affected by stand density and intermediate cuttings (except for thinnings 'from above'). Therefore stand predominant height and equivalent expressions (top height, mean dominant height, etc.) are most commonly used today to classify site.

The relationship of tree height to age, called SITE INDEX, is widely used in evaluating site for even-aged stands or for stands of nearly pure composition.

The popularity of the height/age relationship as a site indicator is that:

NOTE: Site quality rates the productivity of a site in terms of wood. Site index refers only to some factor of the site or stand which is correlated to some degree with its site quality. Inter-relationships of site productivity (quality) and site index generally have been ably reviewed by Spurr (1952).

The height/age relationship for a site is a sigmoidal curve. The procedure for compiling height/age curves as a basis for classifying site depends on circumstances: If there are no permanent sample plots in the forest:

  1. Establish temporary sample plots, covering the range of site and age, in all stands greater than 8 years of age (not less than 8 years as factors of the site have not fully expressed themselves)
  2. Determine stand predominant height in each plot
  3. Plot predominant height vs age. The result is a comet-like scatter of points
  4. Fit curves to the scatter of points. This may be difficult if the trend of height for age is not known.
One variation of the procedure outlined above is to establish semi-permanent sampling units and measure predominant height several times in a period of up to 10 years to give some idea of the trend of the height/age relationship over a range of site and age. This procedure provides better evidence than a simple scatter of points on which to gauge the trend of the height/age relationships over the complete range of age.

Another variation, if the trees have annual rings, and if one can assume the present tallest trees have always been the tallest (a reasonable assumption), is to compile height/age histories for the full range of stands by stem analysis.

If the forest is well served with permanent sampling units, and has been for some time, then height/age data should be readily available and the height/age trends should already be known. Therefore, curves can be fitted with confidence to the scatter of points.

The number of site classes to fit to the splay of points depends on the number of classes required or recognised and the range of height at a given age for the range of sites represented.

The 'given age' conventionally is termed standard age (also index age or base age). Standard age must be such that the factors of the site have had an opportunity to express themselves. On the other hand, age must not be so advanced that available data will be too restricted. 20 to 25 years is a common standard age for coniferous plantations in Australia whereas 80 to 100 years is common in Europe. A useful criterion for selecting standard age is 2/3 of rotation age.

Class interval may be determined indirectly by the width of splay at standard age and the number of classes, e.g. South Australia recognises seven classes from SQ I to SQ 7. Alternatively, intervals of some specific width may be nominated, e.g. 5 m. The classes are then called site index classes, e.g. SI 20 covers the range 17.5 m to 22.4 m. The range must be specified.

In the past, site index were fitted manually using one of two methods:

but with the advent of relatively freely available computers and statistical packages, this is rarely done today.

Site index equations may be classified into three types according to the nature of the height/age curves they generate (Clutter et al. 1983), viz.:
Anamorphic. In this type, the height of any one of the family of curves at any age is a constant proportion of the height of any other curve at the same age.
Polymorphic-disjoint. In this type, the proportionality relationship does not hold but the curves do not intersect within the age range of interest.
Polymorphic-nondisjoint. In this type, there is no constant proportionality relationship and at least some of the curves cross within the age range of interest.
In recent years, nearly all sets of published site index curves have been derived using statistical curve-fitting procedures. Most of these are special cases of three general methods of developing equations discussed by Clutter et al. (1983). The models commonly used in these methods are:

The Chapman-Richards function

H = n x [1 - EXP(-pA)]^(1 / (1 - m))

The von Bertalanffy model

H = [n / p x (1 - EXP(-p x (1 - m) x A) )]^(1 / (1 - m))

where H = predominant height (m) A = age (years) n, p, m = coefficients estimated by nonlinear regression.

Both of these models can be constrained so that predominant height must equal the value of site index at a specified (= standard) age. For example, constraining the von Bertalanffy model for age 20 years gives:
               1 - EXP -p (1-m) A            		
     H = S x [  _______________  ] ^ (1/(1-m))
               1 - EXP-p (1-m) 20
	
where S = site index (predominant height at age 20 years).
Transposing this equation enables one to predict site index directly from predominant height and age once coefficients p and m are known:
               1 - EXP -p (1-m) A            		
     S = H x [  ________________  ] ^ (1/(1-m))
              1 - EXP -p (1-m) 20
Ferguson (1979) used the constrained von Bertalanffy model to produce site index curves for P. radiata in the A.C.T. His solution of the model using data available from permanent yield plots was:
		H  =  1.736 x S x (1-EXP-0.066 x A )^1.792

	i.e.	S  =  H / [1.736 x (1-EXP-0.066 x A )^1.792]
One question which must be faced in developing site index curves is how many sample stands to select and measure. There is no simple answer to this question. The sample size used in any given situation should reflect the variability of the forest resource (height for age) and the cost of measurement.

Growth intercept method of evaluating site

An alternative to evaluating site based on current data for stand height is to use information on height growth over a defined short period during the life of the stand. Such a technique is called the Growth Intercept Method (Wakeley 1954) . It is practical only for uninodal species that display annual branch whorls or for multinodal species in which the spring whorls are clearly identifiable.

All growth intercept methods involve measuring the length of a specified number of successive annual internodes beginning at some well-defined point on the stem, e.g. the first whorl above breast height or above some other specified height, a whorl of known age, etc. If the measurement is over five internodes, it is termed a 5-year intercept.

Growth intercept values can be used directly to assess site quality (called growth intercept indices) or they can be used to calculate site index estimates, e.g.:

Site index = a + b x (5-year intercept).

Growth intercept indices have been found to be particularly useful for classifying site in young stands (Ferree et al. 1958, Wakeley and Marrero 1958, Beck 1971). In older stands, site index is more meaningful Clutter et al. (1983). A stand value for growth intercept is obtained by averaging the individual growth intercept measurements from selected sample trees in the stand. The advantage of the growth intercept method is that stand age is not needed and the length of the required number of internodes is usually easier to obtain than the heights of dominant trees.

Site index classification

Once the height/age relationship for a forest is known, site index classification can proceed. It involves either partial or complete coverage of the forest.

A procedure for coniferous plantations

  1. Lay down a systematic grid on the area (parallel intersecting strips), the intensity of sampling (distance between strips) being guided by the observed variation over the forest.
  2. Estimate predominant height at each grid intersection point. Suppose, for example, nominal spacing at planting was 3 m x 3 m, i.e. 1110 stems ha-1, and predominant height is based on 40 stems ha-1. Then, at each intersection point, take 4 rows x 7 tree spaces (initial spacing) and measure the height of the tallest tree in the group. This gives predominant height at that point. Alternatively, measure the tallest (or 2 tallest) tree(s) within 8.92 m (12.62 m) of the point.
  3. Refer to the height/age curves for the forest and locate the position where predominant height at current age falls. The corresponding height at standard age is the site index at that point.
  4. Repeat the procedure for all intersection points and then map in the boundaries of areas of similar site index class (class interval nominated). These boundaries may or may not be easy to define.
In unthinned and unpruned stands, the tallest trees may be difficult to locate and measure and assessing site index as described above may be impractical (South Australian experience) or predominant height may not be a good index of site. In these cases, other stand characteristics must be used to classify site, e.g. tree vigour and form, crown density, needle length and colour, bark colour, height to base of the green crown, etc. Mapping of site on this basis can be done using a strip-line system, e.g. in South Australia, parallel strips 60 m apart are used and the forest is assessed 30 m on either side of each line. This amounts to 100% coverage of the forest. Temporary sample plots of area 0.02 ha are established at 100 m intervals along each line. Volumes on these plots are determined and the assessed characteristics of the stand related to them to ensure consistency of assessment. Thus, these plots become the yardstick of site assessment.

A procedure for uneven-aged forests

Height in relation to age cannot be used to express site quality in uneven-aged forests of mixed species. Height growth of a species in such a forest is not closely related to age, but more to the varying stand conditions by which it has been affected during its life.

The relationship between height and dbh of the dominant trees is sometimes a sensitive and reliable measure of site quality (McLintock and Bickford 1957). Site index, according to this concept, is then defined as the height attained by dominant trees at a standard dbhob. A variant of this concept was applied successfully by Vanclay and Henry (1988) to classify site in indigenous cypress pine (Callitris sp.) in southern Queensland. They defined their index of site as the expected height of a 25 cm dbhob tree predicted from a stand height curve.

There seems little doubt that, in the future, a combination of detailed plot data (CFI), soil data, and phytoecological data (Cajander/Ilvessalo approach) will permit site to be classified more accurately and usefully than is possible at present.

[Index][Help][Authors]

http://online.anu.edu.au/Forestry/mensuration/SITE.HTM
Cris.Brack@anu.edu.au
Mon, 14 Apr. 1997