Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 275
19
Optimizing Timber Yields in
New Brunswick Forests
Management of commercially important species to increase yields is
usually a local matter, independent of conditions and management actions
undertaken elsewhere. That is especially true in forestry, in which indi-
vidual trees are stationary, long-lived, and easily countable. Nevertheless,
plot-by-plot management can create long-term supply problems on a larger
scale. These larger-scale effects are the most important ones that influence
the vitality of economic activity based on exploitation of the resource.
This case study illustrates an attempt to link local decisions with regional
ones where the valued ecosystem component (wood) is readily identified
and the processes producing wood are well known. The major uncertainties
are associated with changes in scale, and they require a balancing of
scientific and social issues.
275
OCR for page 276
Case Study
THOM A. ERDLE, Forest Management Branch, New Brunswick
Department of Natural Resources, Fredericton, Canada
GORDON L. BASKERVILLE, Faculty of Forestry, University of New
Brunswick, Fredericton, Canada
INTRODUCTION
Silviculture and wood technology are highly developed, and the factors
that influence wood production rates are well known, despite the very
long cycle of the crop. Like many other areas, New Brunswick depends
heavily on industrial forestry. However, the forest base probably cannot
continue indefinitely to supply the quantity and quality of wood required
by industry without a large and expensive effort to develop the forest
resources. Recognition of this need by government and industry forestry
decision-makers has resulted in a strong commitment to carry out the
necessary development, which entails legislative changes in forest ad-
ministration, changes in the allocation of raw material to industrial wood
consumers, and intensified action aimed at making the forest biologically
more productive (Bird, 19801.
Increasing the productivity of the forest requires special consideration,
because it includes direct intervention in the development of a natural
system. Tools for direct biological intervention in forest development
include harvesting, protection, and silviculture; and the use of those tools
requires explicit decisions on how much, where, and when to implement
them. Their implementation initiates a biological response that alters forest
development and production of wood, and the response must be forecast
to permit the design of useful intervention strategies. This is all made
difficult by disparities of scale, in both time and space, between the stand
level at which the actions are taken and the forest level at which the
response must be assessed.
Stands, considered here as small homogeneous communities of trees
(10-50 hectares), collectively constitute a forest. A forest can be enor-
mously complex, encompassing many thousands of stands and many
hundreds of thousands of hectares. Forest management attempts to deal
with complexity of this scale by orchestrating the implementation of har-
vesting, silviculture, and protection at the stand level.
The direct results of local interventions are immediate and visible, as
stands are harvested, established, and protected. But the interventions
276
OCR for page 277
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS
277
alter local dynamics that govern the development of stands and that con-
tribute to the unfolding of forest dynamics. Because of the complexity of
the forest and the long periods associated with stand growth, forest-level
responses triggered by local interventions are not readily apparent and can
accumulate into unforeseen, undesirable, and essentially unalterable pro-
portions.
How can the time and scale differences between the stand level at which
actions are taken and the forest level at which management is planned
and evaluated be bridged to permit the formulation of local intervention
tactics that address the forest management problem in the most appropriate,
biologically realistic manner? The case described here, involving a 300,000-
ha forest management license in New Brunswick, is an attempt to construct
a part of that local-global bridge. The results of similar analyses have led
to management decisions that affect a large fraction of the forests of New
Brunswick.
THE ENVIRONMENTAL PROBLEMS
The issue in this case is impact assessment, i.e., designing local inter-
ventions that change the development of a biological system and fore-
casting the nature and extent of the change (impact) to permit evaluation
of the desirability of proposed interventions.
Three entities must be considered with respect to wood supply: the
quantity of wood, the quality of available wood, and the timing of avail-
ability. Analyses have revealed that development of New Brunswick's
forest under management restricted solely to harvesting and protection,
as has been the practice, yields wood whose volume is insufficient to meet
industrial demand, whose quality is below minimal standards, and whose
availability is discontinuous and erratic (Baskerville, 1982~. Studies have
shown that incorporation of silviculture, particularly tree planting and
tending in the present case, can mitigate these supply problems (Basker-
ville, 19831. This case study addresses the specific questions: What are
the probable forest-level responses to an array of local, stand-level de-
cisions regarding the amount of planting and density at which plantations
are established, and what are the immediate and long-term impacts of the
decisions on wood supply?
The restriction of the present analysis to planting is solely for illustrative
purposes. Stand spacing, thinning, and fertilization are other powerful
silvicultural tools. The procedure described here could well be extended
to those other tactics.
In planting, two major points are at issue. First, at what rate (hectares
per year) should plantations be established? Tree planting, an expensive
OCR for page 278
278
SELECTED CASE STUDIES
undertaking initially, forces a continuing commitment to tend planted
stands as they develop. Second, at what tree density (trees per hectare)
should plantations be established? High volume per unit area and large
trees are both desirable in a stand. However, the density dependence of
individual tree growth in a stand is such that volume per stand and volume
per tree cannot be maximized simultaneously. The former tends to be
directly related to stand density, the latter inversely related to it. Thus,
any chosen planting density represents a trade-off between volume per
unit area and volume per tree, and the impact of this trade-off must be
forecast for the proposed strategy and expressed in a form suitable for
review by a decision-maker.
Sophisticated routines have been developed that help to optimize stand
performance (Brodie and Kao, 1979; Hann et al., 19831. But they con-
centrate on stand-level responses and ignore the interplay of stand dy-
namics, governed by the interactions among trees in a stand, and forest
dynamics, governed by the collective development of stands in the forest.
Failure to place stand development in the context of forest development
has separated the elegant silvicultural solutions from the forest manage-
ment problem and has impeded evaluation and selection of efficient means
to address overall wood-supply concerns.
THE APPROACH
The approach presented here is an analytical framework that presents
probable outcomes of a wide array of decisions and yet retains the linkages
between trees, stands, and the forest. The framework is built around six
steps, as follows:
Step 1. Select a specific forest holding and determine the future wood
supply and the precise nature of the biological limitations on wood supply
produced by restricting management actions to harvesting and protection.
This is the biological definition of the economic problem of wood supply.
Step 2. Identify an array of remedial measures that can be applied to
the strategic problem of wood supply. These are tactical approaches and,
in this case study, are limited to the rate and density at which plantations
are established.
Step 3. Forecast the stand-level outcomes of each alternative in Step 2.
Step 4. Link the stand-level responses into the forest mosaic and forecast
the new forest dynamics that would result from implementing each of the
planting interventions of Step 2.
Step 5. Translate forest-level outcomes into performance indicators rel-
evant to the wood-supply problem (identified in Step 11.
OCR for page 279
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS
279
Step 6. Assemble the performance indicators of all the alternatives in
Step 2 to create a response surface (impact statement) that displays re-
lationships between stand-level decision variables and forest-level per-
formance indicators.
USES OF KNOWLEDGE AND UNDERSTANDING
Valued Ecosystem Component
Within the six-step framework described above, the valued ecosystem
component is the wood supply, as described by the performance indicators
at the stand and forest levels. These indicators can be divided into two
sets. The first set, at the forest level, includes short-term indicators of
maximal sustainable harvest and associated unit cost and long-term in-
dicators of potential harvest expansion and associated unit cost. These are
the most important measures, because they are related directly to the wood
supply. The second set, at the stand level, includes accumulated salable
volume and average tree size at any point in the life of the stand. Although
these are not the direct basis of decision-mal~ing, they collectively con-
tribute to the forest-level indicators.
Significance of Impacts
The significance attached to changes in forest-level indicators brought
about by the exercising of a set of stand-level control options is largely
a function of the goals of the decision-making agency. Interpretations of
significance can vary considerably between and within public and private
agencies. Therefore, interpretations of significance are deliberately omit-
ted here; "significance" has meaning peculiar to each agency, given its
current perspective on the problem. The magnitude of reforestation effects
on wood supply must be shown in a form that enables decision-makers
to impose their own values and to draw their own conclusions regarding
the significance of the impacts of a wide variety of strategies.
Bounding the Problem
The scale differences between forest-management control of stands and
silvicultural control of trees necessitate different spatial bounding. At the
forest level, spatial bounds can be logically established around the wood-
supply base for a mill or group of mills controlled by one management
agency. In this case, that supply base comprises 300,00(J ha in north-
western New Brunswick. At the stand level, the spatial bounds must be
OCR for page 280
280
SELECTED CASE STUDIES
expressed with a resolution consistent with silvicultural intervention (plan-
tations), usually about 20 ha. The density dependence of tree growth
operates at a very local level, so there is little interaction among non-
neighbors. For a given stand structure, this leads to a perfectly linear
relationship between stand production and size of stand (or group of stands
of a given size).
Two considerations were involved in imposing temporal bounds at the
stand and forest management levels. Because of the long period required
for stand development and the societal importance of the forest as a long-
term renewable resource, the appropriate period for examining wood flow
is necessarily long. Furthermore, the forest age structure has a biological
memory: effects of disturbances, such as harvesting and planting, will
surface and persist as the forest develops. A time horizon of 80 years was
therefore deemed appropriate. In wood-supply problems, it is not sufficient
merely to know the wood yield of a plantation or the volume of wood
available from a forest. The timing of wood availability, and therefore
the timing of supply problems, is of utmost importance. Detailed time
resolution within the long-term horizon is necessary, to address timing in
design of strategies. Consequently, the 80-year horizon was divided into
2-year steps for forecasting forest development and for tracking perform-
ance indicators.
For purposes of this chapter, it was decided not to consider the impact
of spruce budworm on the forest and not to consider the numerous other
issues that pervade real forest management. The budworm has played a
major role in the development of the New Brunswick forest, as reflected
in the present age-class structure of the forest. Yields from forest lands
also are influenced by the sizes of budworm populations. Nonetheless,
attempting to capture the complexities of the budworm-forest interaction
would so dominate the analysis as to obscure the central issue considered
here assessing the cumulative effects of harvesting practices. (For an
overview of the budworm problem, see Baskerville, 1976.)
Study Strategy Development
A clear understanding of population dynamics at two levels is required
for strategy formulation. At the stand level, dynamics are governed by
the density-dependent forces that influence growth of individual trees. At
the forest level, dynamics are governed by the structure and development
patterns of the various types of stands that make up a forest. Density-
dependent stand development and age-dependent forest development must
be functionally integrated to provide a path for local intervention to be
projected to the forest level.
OCR for page 281
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS
2s0
200
- ~ 150
~ '
c E
~ _
~100
a,
50 1
14
,,, 1 2
E
In
10
8
6
Sta nd age ( yea rs )
281
initial
Trees /ha
25
/
/
50 75 100
Stand age (years)
\\
4000
\ 2000
'500
\\\ Initial
\ \ \ Trees/ha
my\ 4000
t25 ~
500
75 100
2000
FIGURE 1 Stand development over time characterized by merchantable volume
(above) and stems/m3 (below) for various initial stand densities.
The results of density-dependent tree competition in plantation devel-
opment can be captured in two stand variables and their change over time:
salable volume per hectare (m3/-ha) and average tree size (trees/m31. Typ-
ical patterns for stands of three different initial densities are shown in
Figure 1. Within limits, the density dependence of tree growth shifts the
patterns up in higher-density stands and down in lower-density stands.
For each stand type or plantation in the forest, these relationships must
be quantitative.
For a stand to be considered economically available for harvest, minimal
thresholds in salable volume per hectare and average tree size must be
satisfied. These thresholds are the criteria that stands must meet to be
OCR for page 282
282
SELECTED CASE STUDIES
250
E
~200
o
~ c 1 50
c'
c ~
100
50
14
,,, 12
~10
in
8
6 \
Starr, hip
wi ndow
~ \
/ I
/ 1
/ 1
--to l-~
/
1
1
25
\
50
Stand age (years)
-
\
100 150
-
50
S1a nd age ( yea rs )
100 1 50
FIGURE 2 Determination of operable window for stand that defines its time of
availability for harvest. Broken horizontal lines represent operability constraints of
100 m3/ha (above) and 9 stems/m3 (below). Broken vertical lines indicate earliest age
at which both constraints are satisfied.
recruited into the operable (or available) volume inventory. The age range
over which these thresholds are satisfied defines the timing of a stand's
availability for harvest (Figure 2~. For a given set of volume and tree-size
thresholds, initial stand density can be silviculturally controlled to influ-
ence the timing of the stand's availability for harvest (Figure 31.
The potential power of stand-level control of availability becomes readily
apparent in the context of forest-level dynamics. Because of forces of
origin, species mixture, site fertility, and stocking, each of the many stands
in a forest has its own pattern of development. A description of a forest
shows the stage (age class) of each stand in its development pattern. Forest
OCR for page 283
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS
250
E
-
o
a'
D 150
c'
tic
c)
200
100
__
14
2
10
in
E 8
-
u)
6
~_1__
_ _
l
I ~
1 1
5~0
l
25
Stand age ( years )
\\
~"\~
\i
Initial
Trees /ha
4000
- 2000
500
75 100
1
1
1
1
\
-
-
500
Initial
Trees/ ha
~ 4000
2000
25 50 75 100
Stand age (years )
283
FIGURE 3 Effect of initial stand density on timing of stand availability for harvest.
Operability thresholds, indicated by broken horizontal lines, are 100 m3/ha (above)
and 9 stems/m3 (below). Broken vertical lines represent earliest time of availability
for stands at each initial density.
dynamics are generated by the collective development of the constituent
stands as they progress (age) along their own patterns of development.
Stands do not influence one another's growth, but they are linked at other
levels. First, there is an analytical relationship between stands, in that the
abundance of each developmental stage collectively constitutes an age-
class structure for the forest. Second, interventions, such as harvesting
and silviculture, performed in one place can force the withholding of such
interventions in some other place.
Managing the forest for wood supply is primarily a matter of regulating
OCR for page 284
284
SELECTED CASE STUDIES
the availability of stands. That is achieved through a balancing of the
liquidation of mature stands with the recruitment of immature ones across
the lower threshold of operability. It requires a harvest schedule that
defines the rate and sequence at which the harvest will proceed through
the age structure of the forest. The rate of harvest is constrained by the
timing of the availability of replacement stands as mature, operable ones
are harvested. Availability of wood over time is a function of the initial
forest age-class structure, particularly the relative abundance of stands in
each age class, and the rate of development of these stands to operability.
Consequently, any actions that hasten or retard availability of young stands
for harvest have obvious implications for the rate at which mature ones
can be harvested. That is the path by which density control at the stand
level influences wood supply at the forest level.
The forest age-class structure both determines availability and constrains
the manner in which implemented stand-level actions translate into forest-
level responses. As a result of the natural and man-made forces behind
their development histories, New Brunswick forests have irregular and
different age structures. A fixed set of stand actions, carried out on similar
sites, will yield identical stand-level results that are wholly independent
of the forest in which they are applied. However, the same set of stand
actions might generate dramatically different responses, depending on the
age structure of the forest. This case study deals with one such initial
structure.
Hypotheses
The use of these conceptual tools to manage forest-level responses
requires testing and evaluation of three hypotheses.
· A quantitatively specific hypothesis must be made about how stand
. ultimately, stand growth. It must be
density affects tree growth and
comprehensive enough to address such a question as: What will be the
pattern of volume per hectare and average tree size over the life of a black
spruce (Picea mariana (Mill.) B.S.P.) plantation established in northern
New Brunswick at an initial density of 500, 1,000, 2,500, or 4,000 stems/
ha? The variables expressed in the forecast must be, at least, average tree
size and volume per hectare, because of their important role in defining
stand availability for harvest.
· A hypothesis is necessary with respect to change over time in the
minimal thresholds of tree size and stand volume required for stand op
OCR for page 285
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS
285
erability (utilization standards). Rapid changes in use indicate that oper-
ability constraints change as a result of better technology in harvesting
and product development. Instead of attempts to forecast how utilization
standards would change, a range of reasonable future values was estab-
lished, and the impact of these values on the forest-level indicators were
analyzed. All the outcomes were incorporated into the impact statement
in the form of a response surface, so that decision-makers could locate
their own expectations with respect to changing operability limits and
assess their significance.
· A third hypothesis must relate development of the whole forest quan-
titatively to that of the component stands, particularly those for which the
plantation effort is contemplated. This hypothesis is also the basis for a
forecast. The following are relevant and representative groups of questions:
(1) How will the forest structure and resulting total growing stock change
over time if no harvest is performed? If harvesting is carried out in the
oldest stands first at a rate of, say, 400,000 m3/year? SOO,OOO m3/year?
(2) How will forest structure and resulting growing stock change, if the
same harvest schedules are attempted, but 4,000 stems are planted per
hectare at a rate of 1,500 ha/year? 3,000 stems/ha at 3,000 ha/year? This
second type of question is essential in designing a strategy of plantation
use, and it highlights the necessary linkage between the local decision
variables (planting rate and density) and the forest performance indicators
(wood supply from the forest).
Cumulative Effects
To evaluate the cumulative effects of stand-level tactics on forest-level
performance, the hypotheses with respect to stand development, opera-
bility limits, and forest development were systematically knitted together.
Plantation performance was forecast for each density alternative between
500 and 4,000 stems/ha. Various sets of operability constraints were im-
posed on each of these, to establish the pattern of stand availability for
harvest. Forest dynamics were then forecast iteratively as plantation tactics
were systematically varied over all combinations of 0-4,000 hectares planted
per year and 500-4,000 trees planted per hectare. For this purpose, the
planting rate and density variables were changed at intervals of 500 ha/
year and 500 trees/ha, respectively, to generate 72 unique planting strat-
egies, each with a pattern of future wood availability. Performance in-
dicators were tracked, through a model of forest development, to describe
the cumulative wood-supply effects resulting from each strategy.
OCR for page 290
290
SELECTED CASE STUDIES
for the forest model. Through this linkage, stand-level tactics are carried
through to the forest-level performance indicators.
CONTRIBUTION OF RESULTS TO ECOLOGICAL KNOWLEDGE
The study shows that limited preliminary data, combined with basic
ecological principles, can be systematically used to display the impacts
of a range of management strategies. Several hundred simulations were
performed by linking the plantation-development model with the forest-
development model to evaluate the forest-level outcomes of various com-
binations of planting rate and density. The results are presented as a set
of response surfaces, or nomograms, in Figure 6. This format is partic-
ularly useful when two decision variables are to be assessed via their
control over the valued ecosystem components through several indicator
variables (Peterman, 19751. Decision-makers can easily review outcomes
of a wide range of strategies before setting a decision process (or optim-
ization) in motion. In this case, the decision variables are local (stand-
level) implementation options of plantation density and planting rate,
shown on the Y and X axes, respectively. The surface evaluation, or Z
variable, represents a forest-level performance indicator associated with
each stand-level decision combination. This surface is a type of prediction
containing not only absolute outcomes of a host of alternative combinations
of stand tactics, but also the sensitivity of the forest indicators to those
tactics.
Figure 6 contains a number of noteworthy relationships between local
stand actions and overall forest outcomes that have important implications
for selection of a strategy. For example, Figure bA shows that increasing
the area planted annually increases the sustainable harvest immediately
available from the forest, regardless of planting density. The increase in
forest response surface is nonlinear, however, as evidenced by the wid-
ening intercontour gaps associated with increasing the planting rate. At
the onset of horizontal contours (e.g., at 2,000 ha/year and 2,500 stems/
ha), the positive immediate impact of planting on wood supply disappears
altogether. Obviously, at a given density, each additional hectare planted
is growing identically with the rest, but the effect of the marginal hectare
on total forest performance changes dramatically with the degree of ac-
tivity. Each additional hectare is progressively less "productive" in pro-
moting immediate increased harvest. Thus, increases in planting are not
accompanied by proportional harvest increases, and, beyond particular
amounts of planting, there is no immediate harvest increase at all. At this
point, harvesting has so drastically altered the forest structure that the
resulting low abundance of mature stands constrains further immediate
OCR for page 291
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 291
A. Maximum sustainable harvest B. Cost per m3 of increased
(103 m3/yr. ) harvest (#)
;
3000
Cl
u,
E 2000
(A
000
4m)
. ___
3000
U) 2000
E
-
~n
1000 _
144 ~
~ I_
_380 4000
3000
c'
E 20a)
-
~n
1000
000 200C 3000 4000 10002000 3000 4000
ha/year ha /year
C. Opera ble growing stock at D.
year 60 ( I o6 ma)
~o
Cost per ma of growing
stock at year 60 ( C )
30 ~ / ~ ~:
coo 2000 3000 4000 1000 2000 3000 4000
ho /yea r ha ~ yea r
FIGURE 6 Impact of planting rate (X axis) and plantation density (Y axis) on four
indicators of wood supply. Indicators A and B are related to immediate impacts;
indicators C and D are related to impacts 60 years later.
harvest increases. Essentially, the biological limitation on wood supply
switches from scarcity of young stands at low planting rates to scarcity
of mature stands at high planting rates. The time to stand operability
associated with each plantation density regulates the planting rate at which
this switch occurs. This has two important implications. First, the powerful
influence of forest structure on the effectiveness of planting is not evident
OCR for page 292
292
SELECTED CASE STUDIES
if planting is evaluated solely on the basis of local, stand-level perfor-
mance, as is common in investment approaches. Stand-level assessment
would indicate that, if some planting is good, then more is better, and it
is better in direct proportion to the extent of effort. Clearly, that is not
the case: the linearity breaks down when plantations are viewed in the
context of the whole forest. Second, the forest-structure constraint on
planting effectiveness necessitates plantation tactic design specific to the
forest in question. The surface in Figure 6A would be different if the same
stand response were applied to a forest of different initial age structure
from the one used here. Thus, management strategies embodying design
of stand tactics are not freely transportable between forests.
The maximal sustainable harvest for this forest is 480,000 m3/year
(Figure 6A) and can be attained by planting 1,500 stems/ha at a rate of
3,000 ha/year. This combination, of course, pertains to a specific set of
operability constraints and an 80-year horizon. The maximum resides in
a very sensitive area in Figure 6A. The tightly packed contours in the
lower third of the figure form a "cliff" that represents a high-risk zone
for the decision-maker. If strategies at the edge of the cliff are pursued
to maximize the harvest, then substandard plantation performance, poor
competition control, ineffective protection, or bad model forecasts could
have disastrous consequences. Such occurrences would mean that the
sustainable harvest was over the brink and either a severely reduced sus-
tainable harvest or an unexpected disruption in wood flow would result
from implementing a harvest rate set in accordance with plantation yields
that never materialize. Strategies in this high-sensitivity zone might be
chosen by the risk-taking decision-maker, whereas one with a more con-
servative approach to risk might opt for a higher plantation density (2,000-
3,000 stems/ha) and accept a slightly lower sustainable harvest to ensure
a less sensitive response.
At the stand level, the highest-yield-per-hectare option is planting at a
density of 4,000/ha (Figure 51. This does indeed keep stands in the most
productive state, which is seen as a desirable goal by some foresters, but,
as shown in Figure 6A, devastates the productivity of the forest as a
whole, because it severely delays availability of plantations by reducing
growth rates of individual trees. Again, this exposes the important dif-
ference between local stand response and overall forest response to selected
sets of tactics.
Figure 6C contains information relevant to possible future increases in
harvest, showing the growing stock available 60 years from now. Appre-
ciable gains in growing stock occur at planting rates higher than those
needed to achieve the maximal immediate harvest. Furthermore, with
respect to future increases, the greatest gains are realized at higher densities
OCR for page 293
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS
293
than those which generate the maximal immediate harvest. Thus, high
present harvests and high future harvests cannot be ensured simultaneously
through the implementation of one plantation tactic. The decision-maker
faces a trade-off between immediate gains and future gains. It can be
resolved by striking a compromise between the two or by using a strategy
with two simultaneous tactics: a low-density one aimed at immediate
harvest gains and a high-density one aimed at future harvest gains. Re-
gardless of the choice, failure to link stand and forest performance would
not even reveal the problem to the decision-maker, let alone suggest
potential solutions.
To illustrate the impact of operability limits on the decision-maker,
several stand- and forest-level simulations were performed in which the
minimal acceptable operability standards were systematically altered. The
annual planting rates were fixed at the number of hectares required to
maximize the current harvest for each planting density (e.g., 1,500 ha/
year at 4,000 stems/ha and 3,000 ha/year at 1,500 stems/ha). The results
are presented in nomogram form in Figure 7. The X, Y. and Z axes show
minimal Or. tress. ~i7.e plantation density, and maximal sustainable
lllillAlAl~$ ~ ~ _ it_ ~, ~ ~ r-
harvest, respectively.
Each of the four surfaces in Figure 7 represents a different minimal-
volume-per-hectare requirement. It is clear that the desirability of alter-
native densities, with respect to wood supply, varies considerably with
the tree-size constraints that will be in effect. In Figure 7B, for example,
under the stringent requirement of 6 stems/m3, plantation densities of
1,000-1,500 stems/ha provide the maximal harvest. However, as the tree-
size constraint is relaxed, both the maximal sustainable harvest and the
plantation density at which the maximal harvest is realized increase. This
is evidenced by the ridge that runs through the surface in the figure. That
the surface elevation increases with relaxation of tree-size constraint is a
reflection of the decreased time to stand operability and the consequent
increased availability of stands when harvest of smaller trees is acceptable.
That the ridge has a positive slope in the X-Y plane is indicative of the
earlier achievement of these lower operability thresholds by higher-density
(and consequently higher-volume) plantations when the tree-size constraint
is reduced.
Examination of all four surfaces reveals that the increase in harvest
associated with relaxed tree-size constraints holds for the minimal-volume-
per-hectare constraint as well. The maximal possible harvest increases as
the minimal volume per hectare is decreased from 205 m3/ha (Figure 7D)
to 70 m3/ha (Figure 7A). Similarly, the planting density that yields the
greatest harvest increases as the volume-per-hectare constraint increases.
That is consistent with Figure 5, which shows the advantages of low and
OCR for page 294
294
A. Maximum harvest (103m3/yr)
at 70 m3/ha volume constraint
40a) - An/ /
o~ Jl1117° ~,~
E 2000
-
cn
-
-
//// /~/// ~o
1 OOt)
4000
3000
C,
~ ID
E
-
1 000
SELECTED CASE STUDIES
B. Maximum harvest (103m3/yr)
at 115m~ha volume constraint
4003
3000
/// ~
6 8 10 12
Required stems /m3
C. Maximum harvest (103m3/yr)
at 160 m3/ha volume constraints
J
in
E 2000
in
ooo
14 6
4009
101~ / /
~/j2/ /
J/
460- 1COO
/
c
6 8 10 12 14
Required stems / ma
-500 - 3C~
E
On
360 /~/ 440 /
Jl74
/
480
/~
8 10 12 14
Required stems/ ma
D. Maximum harvest (103rn3/yr)
at 205 m3/ha volume constraint
10/ - Ad/ /
1 ~/420/ /
//~
6 8 10 12 14
Required stems /m3
FIGURE 7 Impact on wood supply of plantation density (Y axis) and minimal ac-
ceptable tree size (X axis) at four constraints on minimal volume per hectare.
high densities to be rapid tree-size development and rapid volume-per-
hectare development, respectively. Of importance to decision-making in
Figure 7 is the powerful influence that utilization constraints can have
over forest-level outcomes. The actual biological response underlying each
surface in Figure 7 is identical. What generates the marked differences in
forest outcome is a logistical harvesting constraint that the decision-maker
imposes on the biological system, which regulates the availability of stands
OCR for page 295
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS
295
for harvest- this is no mere detail, inasmuch as utilization constraints are
largely under the decision-maker's control. Figure 7 reveals likely out-
comes that the decision-maker might achieve by exercising that control.
APPLICATION OF THE TOOL
The modeling process captured in Figures 6 and 7 constitutes a useful
tool in policy design. Nomograms like those shown are powerful aids in
structuring policy questions so that detailed modeling and analysis can be
directed incisively. The diagrams are not used to choose a "best" planting
policy. Rather, they are displayed to senior decision-melters, first to dis-
cover how they weight various indicators and second to discover where
the decision-makers would like to be within the possible policy domain.
Detailed analyses can then proceed with the indicators of choice within
limits in terms of the policy variables. Thus, the nomograms are not used
to answer operational questions, but rather to structure management ques-
tions. In this context, they make the scientific advisor more efficient both
in the use of his own time and in establishing an understanding of the
problems between advisor and decision-maker.
CONCLUSION
The case study presented here shows how a wide range of stand-level
actions would influence forest-level outcomes. The purpose is to make
the decision-maker aware of the importance of linking the two levels of
consideration. No recommendations of optimal planting densities or plant-
ing rates are presented, because such decisions are strongly influenced by
specific industrial strategies and objectives and by the degree of risk
aversion of decision-makers. Furthermore, the decision-making picture is
incomplete. Analysis becomes more complex as wood value, harvesting
cost, and additional silvicultural tactics (such as spacing and thinning) are
considered. Once the range of impacts is understood, more sophisticated
analytical tools, like mathematical programing, might be effectively brought
to bear on strategy design.
Different planting alternatives, applied at the stand level, generate dif-
ferent forest-level outcomes, because of the interaction of stand and forest
dynamics. These differences highlight risks, sensitivities, and trade-offs
that, although of prime importance in decision-making, would not be
readily evident from stand-level or forest-level analyses alone. As a pack-
age, the analytical framework and the results it provides form a rich body
of information relevant to the forest management problem of evaluating
the impact of interventions on wood supply.
OCR for page 296
296
SELECTED CASE STUDIES
The techniques described in this study have been applied in the design
of New Brunswick's silvicultural program. Effective control over planting
is possible, because about 50% of forest land in the province is Crown
land and the large timber companies need access to Crown land to maintain
economical operations. Of the remaining land, half is owned by large
companies and half has varied ownership. Thus, about three-fourths of
the forest land in the province has been readily incorporated into a har-
vesting program suggested by the analysis as necessary to even the flow
of timber and counteract the unfavorable age distribution of the stands.
There is not yet a firm link between stand dynamics and budworm
population dynamics. Therefore, it is not possible to predict the effects
of the harvesting and planting program now being implemented on pop-
ulation dynamics of the insect. Whether the program will have to be
modified to accommodate problems generated by the budworm remains
to be seen.
REFERENCES
Baskerville, G. L. 1965. Dry matter production in immature balsam fir stands. For. Sci.
Monogr. 9:1-42.
Baskerville, G. L. 1976. Report of the Task Force for Evaluation of Spruce Budworm
Control Alternatives. New Brunswick Cabinet Committee on Economic Development,
Fredericton, N.B.
Baskerville, G. L. 1982. The Spruce/Fir Wood Supply in New Brunswick. New Brunswick
Department of Natural Resources, Fredericton, N.B.
Baskerville, G. L. 1983. Good Forest Management-A Commitment to Action. New
Brunswick Department of Natural Resources, Fredericton, N.B.
Bird, J. W. 1980. Forest management A provincial perspective. Pp. 33-37 in The Forest
Imperative. Proc. Can. For. Congr., Toronto, Ont., September 22-23, 1979. Canadian
Pulp and Paper Association, Montreal.
Brodie, J. D., and C. Kao. 1979. Optimizing thinning in Douglas-fir with three descriptor
dynamic programming to account for accelerated diameter growth. For. Soc. 25:665-
674.
Cuff, W., and G. L. Baskerville. 1982. Ecological Modelling and Management of Spruce
Budworm Infested Fir-Spruce Forest of New Brunswick, Canada. Paper presented at 3rd
Int. Conf. on State-of-the Art in Ecological Modelling, Colo. State Univ., May 24-28,
1982.
Hall, T. H. 1978. Toward a Framework for Forest Management Decision-Making in New
Brunswick. Report TRI-78. New Brunswick Department of Natural Resources, Fred-
ericton, N.B.
Hann, D. W., J. D. Brodie, and K. H. Riitters. 1983. Optimum stand prescriptions for
ponderosa pine. J. For. 81:595-598.
Ker, M. F. 1981. Early Response of Balsam Fir to Spacing in Northwestern New Brunswick.
Maritime Forest Research Center Information Report M-X-129. Canadian Forest Service,
Fredericton, N.B.
Lundgren, A. L. 1981. The Effects of Initial Number of Trees Per Acre and Thinning
OCR for page 297
\
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS
297
Densities on Timber Yields from Red Pine Plantations in the Lake States. Forest Service
Research Paper NS-193. U.S. Department of Agriculture, Washington, D.C.
Peterman, R. M. 1975. New techniques for policy evaluation in ecological systems: Meth-
odology for a case study of Pacific salmon fishenes. J. Fish. Res. Bd. Can. 32:2179-
2188.
Stiell, W., and A. B. Berry. 1973. Development of Unthinned White Spruce Plantations
to Age 50 at Petawawa Forest Experiment Station. Publication 1317. Canadian Forest
Service, Ottawa, Ont.
Commi`;tee Comment
The work reported in this case study provides a concrete example of
how to assess the long-term, regional consequences of immediate, local
forest-management actions in a manner useful for the design of sustainable
strategies for resource development. The story is interesting as a partic-
ularly simple and clear case of effective assessment of cumulative effects,
and it has proved useful in actual practice by the Department of Natural
Resources (DNR) of New Brunswick. Indeed, Baskerville initiated the
analysis described here to deal with the practical difficulties he encountered
as assistant deputy minister in DNR; Erdle is now responsible for applying
the results of the analysis in DNR's Forest Management Branch.
The problem addressed here is widespread in forest management. A
forest is a large-scale mosaic of individual stands of trees. Different stands
can generally be characterized by age distributions, species mixtures, and
habitat. The forest therefore is also a mosaic of stand-development tra-
jectories, different stands reaching maturity at different times and rates.
Management actions can alter those trajectories.
Although wood is considered the only valued ecosystem component in
this model, additional components, such as control of spruce budworm
and maintenance of an aesthetically pleasing mosaic of forest patches,
have been considered in other analyses of the system (Baskerville, 1976;
Clark et at., 19791. The focus developed in this study was the result of
a consensus that emerged among government, industry, and academic
participants in the forest-management debate that, if the wood-supply
problem were solved, most other concerns would be met automatically.
Interestingly, the supply of wood is from the outset defined in cumulative
terms over the regional scale and long periods relevant to the economics
of the provincial forest industry. What matters ultimately is not the pro-
duction from an individual stand of trees, but rather how the production
from all the stands of the forest, taken together, can best be managed to
meet society's needs.
OCR for page 298
298
SELECTED CASE STUDIES
Erdle and Baskerville studied the cumulative effects of alternative man-
agement actions by integrating local models of stand growth, regional
models of forest age-structure dynamics, and a detailed inventory of the
forest's existing age structure. They used available models of the relations
among site quality, tree density, and growth rates of individual trees as
the major sources of ecological knowledge. In essence, a commonly used
wood-supply and forest-productivity model was taken "off the shelf" and
provided with data relevant to the New Brunswick situation. This pro-
cedure made their task simpler and made their results more acceptable to
managers already familiar with the models.
The forest model used by Erdle and Baskerville is essentially a book-
keeping one that tracks the aging of forest stands as they respond to
additions to wood (through planting and natural regeneration) and removals
of wood (through harvest, insect damage, and natural death). The under-
lying ecological theories are simple. Growth rates are assumed to be
density-dependent, and the yield curve has a maximum. These results
have been empirically determined, but they can also be derived theoret-
ically from basic principles of plant competition. The simplicity of these
demographic accounting models is also their main strength. They are
robust, and, if solid biological data are available, they can be readily used
for a wide variety of situations (see also Chapter 121. More complex,
general, and rigorously tested stand models (Shugart, 1984) could have
been used in the analysis and probably would have afforded improved
credibility in academic circles. But there is no reason to believe that these
more elegant models would have appreciably improved the results obtained
by Erdle and Baskerville.
An additional aspect of ecological knowledge central to the success of
the Erdle and Baskerville analysis was the existence of an accurate de-
scription of the present age distribution of provincial forests. Such baseline
data on the heterogeneity of tree stands with respect to age distribution,
species mixture, and growth potential are necessary to coordinate local,
short-term management actions in a way that achieves desired regional
cumulative consequences. Unfortunately, such data are extremely rare,
especially in the case of age distribution, in inventories of forests and
other renewable resources.
New Brunswick has a forest inventory data base that is one of the most
accurate and useful for forecasting purposes in North America. This is in
part because of the data-base shortcomings uncovered in Baskerville's
earlier modelings and analyses and his later tenure in DNR. Generally
speaking, however, useful background data are difficult and unglamorous
to obtain, monitor, and update, although a characteristic feature of the
OCR for page 299
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS
299
few success stories in resource management is the existence and intelligent
use of such data bases (see, for example, Chapter 12~.
One final aspect of the Erdle and Baskerville analysis, the method of
presenting results, is relevant not only to the assessment of cumulative
effects, but also to a wide variety of efforts to provide usable ecological
knowledge. They use nomograms to illustrate the trade-offs in valued
ecosystem components that are likely to result from alternative manage-
ment actions or other development interventions. In this forest-manage-
ment case, as in many other environmental problems, interpretations of
what is important vary considerably within and among government agen-
cies, private interest groups, and the ecological profession. Moreover, as
Erdle and Baskerville point out, these perceptions of significance change
with time. Rather than adopting a single definition that would render their
analysis usable from only a single perspective, Erdle and Baskerville have
devised a framework in which users can specify (and indeed often discover)
their own definitions of significance and explore the implications of their
definitions for a wide range of possible decisions.
An added benefit of the nomogram approach, used to good effect in
this case study, is that the spacing of the nomograms' contours of valued
ecosystem components indicates the sensitivity of a predicted outcome to
errors or incompleteness in implementing the decisions. Peterman (1981)
has shown how minor modifications of the nomogram technique can be
used to show the significance of uncertainties in the ecological models
for the expected effects on valued ecosystem components. An analysis of
the uncertainties inherent in the regional demographic and local compe-
tition models underlying this case study would have increased its useful-
ness even more.
Nomograms of this sort have been applied to environmental problem-
solving in cases of renewable-resource management (e.g., Holling, 1978;
Peterman, 1975, 1977; Regier, 1976), river-basin planning (e.g., Rabi-
novich, 1978), and regional development (e.g., Miller, 19821. They have
a long history of application in business and industrial management. Nom-
ograms are not a cure-all, especially given their restriction to two or at
most three simultaneous trade-offs. But they are useful in appropriate
circumstances, and they deserve wider application.
Not all the approaches and techniques for cumulative-effects assessment
described by Erdle and Baskerville are relevant to the more complicated
situations of, say, river-basin planning, or even to the more closely anal-
ogous situations of regional fisheries management. As difficult as forest
age structure is to deal with, it is nonetheless measurable, and it provides
a firm handle on the assessment of future forest development a handle
OCR for page 300
300
SELECTED CASE STUDIES
that most cumulative-effects studies will not have. In addition, the mere
existence of a forest-level "planning authority" (the province's Depart-
ment of Natural Resources) makes the case reviewed here much easier
than the more typical one in which no agency is either solely responsible
for or has sole power over the long-term, regional impacts of a class of
related development decisions. More generally, it is worth emphasizing
with the authors that a great proportion of their analysis involved the
careful "tuning" to local conditions of a few relatively simple ecological
concepts and models. No general precr~ptions for cumulative-effects as-
sessment or management emerged from this study, nor are they likely to
emerge from others.
References
Baskerville, G. L. 1976. Report of the Task Force for Evaluation of Spruce Budworm
Control Alternatives. New Brunswick Cabinet Committee on Economic Development,
Fredericton, N.B.
Clark, W. C., D. D. Jones, and C. S. Holling. 1979. Lessons for ecological policy design:
A case study of ecosystem management. Ecol. Model. 7:1-53.
Holling, C. S., ed. 1978. Adaptive Environmental Assessment and Management. John
Wiley & Sons, Chichester, Eng.
Miller, P. C. 1982. Simulation of socio-ecological impacts. Environ. Manage. 6:123-144.
Peterman, R. M. 1975. New techniques for policy evaluation in ecological systems: Meth-
odology for a case study of Pacific salmon fisheries. J. Fish. Res. Bd. Can. 32:2179-
2188.
Peterman, R. M. 1977. Graphical evaluation of environmental management options: Ex-
amples from a forest-insect pest system. Ecol. Model. 3:133-148.
Peterman, R. M. 1981. Form of random variation in salmon smolt-to-adult relations and
its influence on production estimates. Can. J. Fish. Aquat. Sci. 38:1113-1119.
Rabinovich, J. E. 1978. An analysis of regional development in Venezuela. Pp. 243-278
in C. S. Holling, ed. Adaptive Environmental Assessment and Management. John Wiley
& Sons, Chichester, Eng.
Regier, H. A. 1976. Science for scattered fisheries of the Canadian interior. J. Fish. Res.
Bd. Can. 33:1213-1232.
Shugart, H. H. 1984. A Theory of Forest Dynamics. Springer, New York.
Representative terms from entire chapter:
brunswick forests
