Governor's Advisory Committee on Chip Mills
Final Report
August 1, 2000
A. SUSTAINABLE TIMBERLAND RESOURCE BASE
Timber Resource Setting2
On an overall basis, slightly less than one-third of Missouri -- some 14 million acres -- is covered by forest land (Table 1). Almost 96% of that area, or about 13.4 million acres, is classified as timberland.3 About 15% of this area is public land, most of which is administered by either the U.S. Forest Service or the Missouri Department of Conservation. Forest industry owns only 2% of the state's timberland area. The vast majority of these lands -- some 83% or 11.1 million acres -- is controlled by nonindustrial private forestland (NIPF) owners.4 The number of such owners in Missouri has increased dramatically over the past two decades, from about 81,000 in 1978 to 307,000 in 1994 (Birch 1996a).
Many of these individuals or groups own relatively small acreages. In Missouri, 48% of NIPF owners have tracts of less than 10 acres, and 79% own tracts of less than 50 acres in size (Table 2). On the other hand, when viewed from the perspective of size of holdings (as opposed to number of owners), 57% of the NIPF acres owned are in tracts of 100 acres or more, and 77% are in tracts of 50 acres or more. Thus much of the acreage is in larger tracts, while most of the owners own small tracts.
The most recent estimates of NIPF ownership turnover rates suggests that the average turnover rate for a given NIPF acre is every 28 years. One assumption that was adopted in the following analysis of growth and drain projections from Missouri forests is that an acreage equivalent to most of the NIPF parcels in the state will change ownership within the next 50 years. When extending that assumption to the notion of present and future timber availability in the state, this implies that an acreage equivalent to the total acres of NIPF lands in Missouri could potentially be available for timber harvest during that time.
_______________________________________________________
Table 1. Missouri forest resource and Missouri
timberland by ownership
( Source: Hahn and Spencer 1991 )
A. Missouri Forest Resource (1989)
(millions of acres)
Total Land 44.1
Total Forest 14.0
Total Timberland 13.4
B. Timberland by Ownership (1989)
| (thousand acres) | % | |
| National Forest | 1321 | 10.0 |
| Other federal | 246 | 2.0 |
| State | 403 | 3.0 |
| County | 42 | <1.0 |
| Industry | 222 | 1.7 |
| Other Corporate | 929 | 7.0 |
| Farmer | 5024 | 37.6 |
| Miscellaneous Private Individual | 5184 | 38.8 |
| Total : 13371 |
_______________________________________________________
When considering the two existing high-capacity chip mills in Missouri, the area of potential impact shifts from the above statewide focus to a smaller region in southeastern part of the state. For each of these operations, it is economically feasible to procure wood from within a 60-mile radius of the mill site (Figure 1). Thus it is this combined area, some of which is overlapping, in which the impacts of the chip mills on the state's timberland resources will be experienced. With respect to the Willamette facility at Mill Spring, about 3/5 of the area within a 60-mile radius of the mill is timberland, and 73% of that land is under NIPF ownership, with the rest under federal (22%) and state (5%) control. For the Canal Industries facility at Scott City, slightly less than 1/3 of the land base within a 60-mile radius of the mill site is timberland; and 81% of these lands are under NIPF ownership, with the rest administered by federal (15%) and state (4%) agencies. It should also be noted only 51% of all lands within the 60-mile radius of the Scott City mill are in Missouri..
Timber Growth and Harvest
The primary source of information on the status of Missouri's forest resources in terms of land area, volume of standing trees, and rates of timber growth and harvesting is the periodic inventory conducted by the U.S. Forest Service, in cooperation with the Missouri Department of Conservation. Data is analyzed through the USFS's Forest Inventory and Analysis (FIA) program at the North Central Forest Experiment Station located at St. Paul, MN. For Missouri, the two most recent comprehensive forest inventories were completed in 1972 and 1989, respectively. A new statewide inventory of forest resources is underway and will be completed in 2003. This inventory utilizes a new process that will sample approximately 20% of the forest land in Missouri on a continuing basis. However, at present it will be several years until results comparable to those from the 1989 inventory are available for the area around the chip mills. This obviously complicates the task of understanding current and potential impacts of chip mills in Missouri since, among other things, the two high-capacity mills currently operating did not arrive in the state until the late1990's. Their impacts, therefore, along with the impacts of any future mills, can at present only be addressed through the use of inferences based on existing inventory data. This is addressed further below.
When focusing on the quality of Missouri's forest resources for wood production, it is helpful to differentiate these resources into two broad categories -- growing stock and non-growing stock, often referred to as cull. Growing stock trees are live trees of commercial species that meet specified standards of size, quality, and merchantability. Noncommercial species, rough, and rotten trees are excluded from this category. The growing stock volume in Missouri's forests in 1989 was 9 billion cubic feet. In addition to this amount, the state's forests also contain 4.8 billion cubic feet of trees (or portions thereof) that are non-growing stock because of rot, form, or other defects. This non-growing stock volume has traditionally been termed cull material, because it is unusable for lumber production; and it accounts for about 35% of the total standing volume in Missouri forests, with growing stock accounting for the other 65%. Both in total volume and as a percent of standing volume, this cull material exceeds that of any other state in the nation. Moreover, the total volume of cull trees on nonindustrial private lands in the state is more than six times greater than the total cull volume for all other ownerships (i.e., federal, state, industry) combined (Table 3).
Table 3. Total and per acre volume of growing stock and
cull by ownership, Missouri, 1989
( Source: Hahn and Spencer 1991 )
A number of factors contribute to the above situation. Among the most significant is the combined effect of the relatively poor site quality of much of the Ozark forestlands and the cultural attitudes fostered over the years that led to 'highgrading' -- i.e., harvesting only the best quality trees and ignoring other factors such as stand structure and composition -- influenced in part by the lack of a market for low-grade materials. Nonetheless, while cull material has no current or potential market value for sawtimber, it can and does serve as raw material for other kinds of wood products such as railroad ties, pallets, novelties, etc. Missouri's forest industries have traditionally produced marketable products from wood by-products such as slabs, edgings and sawdust for charcoal, and bolts (four-foot sections with small-end diameters of greater than six inches) and re-sawed slabs for pallet stock Trees containing only one 8-foot sawlog, post or bolt are considered harvestable, which implies that at least 95% of the volume (65% in growing stock and 30% in rough trees) found growing on an average acre of timberland in Missouri in 1989 could be utilized for commercial products (Law 2000). Of particular interest here is that chip mills are capable of utilizing this low quality material for processing into inputs for pulp and paper operations. At the same time, while these mills constitute a potentially viable market for cull material, this does not imply that the mere potential for such a market is necessarily related to the kinds of forest practices undertaken to supply that market.
When considering timber growth and drain in the aforementioned areas impacted by the two chip mills, the figures in the top part of Table 4 depict the annualized difference between growth and removals of growing stock in the chip mill source areas. In the Mill Spring area, annual growth and removals totaled 96 and 49 million cubic feet, respectively, leaving a net annual growth of 48 million cubic feet. For the Scott City mill's source area, annual growth and removals totaled 63 million and 38 million cubic feet, respectively, yielding a net annual growth of 25 million cubic feet. Thus when considering the growing stock volume on all timberlands in both mill sourcing areas (taken individually), there existed a substantial excess of growth over removals between 1972 and 1989.
These figures have been adjusted, however, to more accurately reflect growing stock on timberlands that are actually accessible and otherwise available for marketable wood products (Shifley 1999a). In this analysis, two basic adjustments were applied to the net annual growth figures for each source area. First, the net annual growth figures were reduced by 45% to account for four factors that are presumed to significantly affect actual timber availability for chipping : a) the presence of species unsuitable for chipping, such as conifers, hickory, blackjack oak, etc.; b) slopes of greater than 40% grade; c) acres dedicated to riparian and road buffers; and d) all public lands within the two mill sourcing areas. In this latter regard, it was assumed that while both federal and state public lands are dedicated to sustained yield management of all resources, including timber, many resource outputs produced from these lands are not market driven. Secondly, the growing stock available for harvest remaining after this first set of adjustments was further reduced by 16% to avoid double-counting growing stock volumes in the area where the Mill Spring and Scott City sourcing areas overlap (Figure 1). The overall results for adjusted net annual growth of growing stock in each of the sourcing areas as of 1989 are 22 million cubic feet and 13 million cubic feet for Mill Spring and Scott City, respectively.
Table 4. Relevant data on annual growth, removals, net annual growth, and expected impact of chip mills on annual harvest levels. Missouri: 1972-1989. (Source : Shifley 1999a ; 2000)
Growing Stock
(cubic feet )
|
Missouri |
Mill Spring Area | Scott City Area | |
| Growing stock volume | 9 billion | 3 billion | 2 billion |
| Annual growth | 267 million | 96 million | 63 million |
| Annual removals | 117 million | 49 million | 38 million |
| Net annual growth | 150 million | 48 million | 25 million |
| Adjusted net annual growtha | 22 million | 13 million | |
| Projected annual removal for chips | 6.7 million | 11.7 million | |
| Annual harvest from growing stock required to meet chip mill and other demand | 56 million | 50 million | |
| Increase in current annual growing stock harvest to meet chip mill demand | 16% | 14% | 31% |
__________________________________________________________________________________
Cull Material (Non growing stock)
(cubic feet))
Mill Spring Area Scott City Area
Current cull volume
1026 million
517 million
Adjusted available cull volumea
474 million
265 million
Projected annual removal for chips 6.7 million 11.7 million
Estimated supply of chips from cull
at anticipated rate of utilization
71 years
23 years
______________________________________________________________________________
a Adjustments to net annual growth and cull volume involve 2 stages
: 1) a 45% (Mill Spring) or 39% (Scott City) reduction for unusable
species, riparian and road buffers, slopes greater than 40%, and public
lands, which are assumed to be dedicated primarily to non-timber uses
(although timber production is not presumed to be categorically excluded);
and b) a further reduction of 16% (from the above adjusted figures)
for source area overlap.
When comparing the above figures for adjusted net annual
growth in the two source areas to the anticipated wood utilization of
the two mills, the following picture emerges. For the Mill Spring
facility, projected production at average capacity will require about
200,000 tons (6.7 million cubic feet) of wood per year at the gate.
The Scott City facility will require 350,000 tons (11.7 million cubic
feet) per year (again noting that not all of the source area for this
mill is in Missouri). When relating adjusted net annual growth of growing
stock and anticipated wood demands by the mills, it is evident from
Table 4 that, for the Mill Spring area, to supply the 200,00 tons projected
demand for wood exclusively from growing stock, annual harvest will
increase by 14%, from 49 to 56 million cubic feet. To supply the Scott
City facility with its expected annual demand of 350,000 tons of
wood, again with the entire source being growing stock, annual harvest
in its source area will have to increase by 31%, from 38 to 50 million
cubic feet.
The lower portion of Table 4 presents an overview of the estimated available volume of cull material, or equivalently, non-growing stock in the source areas (i.e., within a 60-mile radius) of the chip mills at Mill Spring and Scott City. As noted earlier, cull material is not suitable for sawtimber products and thus not accounted for as part of the growing stock. Estimated available cull volumes, after adjusting total volumes in the same manner as was applied to annual growth (Table 4), are 14.2 and 7.9 million tons for the Mill Spring and Scott City source areas, respectively. At the expected rate of wood utilization by the mills, this translates into a 71-year supply of chips for the Mill Spring facility and a 23-year supply for the Scott City mill. Although all of this available cull volume will certainly not be harvested, these figures do suggest that there is a significant amount of low quality wood resource in the vicinity of the two mills to supply their expected demands for wood for at least two decades, in the case of the Scott City Mill, and substantially longer for the Mill Spring facility. The extent to which such resources will actually be utilized by the chip mills is a separate question which the above figures are not intended to answer.
Additional Scenarios for Projecting Growth and Drain
The above methodology provides an analysis of the impacts of Missouri's two high capacity chip mills on timber resources in the source areas surrounding the mill sites (Shifley 1999a ; 2000). Given that the next comprehensive statewide forest inventory will not be completed for several years, additional analyses were conducted which included a series of scenarios projecting wood utilization by the two chip mills and impacts on growing stock in the source areas over the next several decades (Shifley 1999b ; 2000). One set of scenarios was based on the above 'baseline' analysis of 1989 inventory data, to which additional more restrictive assumptions regarding wood availability were applied. A second set of scenarios, covering the period from 1989 to 2030, was based on both the 1989 statewide inventory data, and on periodic data on wood products production in Missouri.
Alternative Scenarios Based on 1989 Statewide Inventory Data. The first part of this supplementary analysis employed the same methodology for assessing timber supply in the mill source areas as described above. These scenarios examined how results from the baseline analysis would differ if there were: a) increasing wood demands and associated harvest levels generated by the chip mills; and b) a reduction in the volume of wood available to the mills due to landowner unwillingness to sell timber. Scenarios with this latter assumption also included an increase in total land in buffer areas to include all streams and roads, a certain aggregate percentage of which had been assumed in the baseline analysis to be without buffers (Shifley 2000).
One set of scenarios started with the baseline analysis described above and incorporated potential increases in the demand for wood chips by the two mills (Figures 2a and 2c; Appendix tables A1-A4). Annual chip demand levels for the Mill Spring facility were increased from 200 to 350, 600, and 1000 thousand tons; and Scott City demand levels were increased from 350 to 600 and 1000 thousand tons. The 5% volume reduction for stream buffers was also included in each of these scenarios. As in the baseline analysis, results of these additional assumptions may be interpreted both for situations where the entire supply of chips is obtained from growing stock (Figures 2a-2d) and for where it is derived exclusively from cull material (included in Appendix A tables). For growing stock, the differences from using these additional assumptions would be manifest in the amount of growing stock remaining after demands for both chips and other wood products were satisfied. In Figures 2a-2d, this 'residual' annual net growth is the difference between "excess growth available for chips" and "assumed chip harvest levels." For cull material, the difference produced by incorporating the scenario assumptions is depicted in the final row of Appendix tables A-1 - A4 as 'equivalent supply for chip harvest.' (This is also identical to the last entry in Table 4 : 'estimated supply of chips from cull at anticipated rate of utilization.')
Not surprisingly, the results of this set of scenarios indicate that as utilization of chips by the two mills increases, the 'residual' annual net growth in growing stock in the source areas declines. The scenarios provide a picture of how adjusted net growth (i.e., excess growth available for chips - chip removals) declines with different rates of wood utilization by the chip mills. Thus, for example, in translating the graphics of Figure 2a into numbers (see tables A1-A4), when projected annual removals for chips are increased from 200 thousand to 600 thousand tons for the Mill Spring area, the residual net annual growth suitable for chips declines from 436 to 36 thousand tons. Similarly, when projected chip removals are increased from 350 to 600 thousand tons for the Scott City area, the residual net annual growth suitable for chips declines from 25 thousand to a negative 225 thousand tons. That is, for the Scott City area, such a change in demand would lead to an overall excess of removals over growth under this set of assumptions. The implications of this set of scenarios is that if wood demand for chips increases significantly, and if all the wood for chips is taken from the growing stock, then the residual annual net growth -- i.e., that remaining after all removals, including for chips, have been accounted for -- will decline, and more rapidly so with greater volumes of removals for chips. Since there is less timber volume to begin with surrounding the Scott City mill, along with higher chip removals (350 thousand tons per year), total chip harvest will exceed annual net growth at lower harvest levels in the Scott City area than for Mill Spring as the wood demanded by the mills increases.
Figure 2. Growth and removals with different levels of chip harvest from growing stock in source areas for Mill Spring and Scott City under alternative assumptions regarding private land availability and the presence of buffers on all roads and bodies of water (Source : Shifley 2000)
From an alternative perspective, if the chip mills demands were met exclusively through non-growing stock or cull material (i.e., non-growing as opposed to growing stock), and the same increases in annual removals for chips occurred (i.e., from 200 thousand to 600 thousand tons for Mill Spring and from 350 thousand to 600 thousand tons for Scott City), the supply of cull material to meet the Mill Spring demand would decrease from 69 to 23 years, and to meet the Scott City demand, from 22 to 13 years (Table A-3).
A second set of four scenarios is identical to the above, but adds an additional constraint that reduces land available for timber harvest and thus wood supply to meet total demand (i.e., for chips and other wood products). In these scenarios, it is assumed that, for whatever reasons, private landowners are more reluctant to sell their timber than assumed above and, as a result, the land base (and associated wood volume) available for timber harvest is reduced by 20%. In addition, a further reduction in lands available for harvest was incorporated to account for buffers on all roads and streams, not all of which had been subject to the buffer restriction in the baseline analysis. With the same increases in levels of chip demands as considered above, and with only 80% of the land base available, it is not surprising that the above trends are accelerated. Figures 2b. and 2d. depict these scenarios where all removals are taken from growing stock. Appendix Tables A-5 through A-8 contain the numbers from which these figures are derived, along with projections depicting the situation where all chips are obtained from harvesting of cull material, as opposed to growing stock, in the source areas for the mills.
It is important to note that the likelihood of the assumptions regarding increases in wood demand (and associated chip removals) and the decrease in private land availability actually happening is an entirely different question from the information provided by these scenarios. The latter are intended to provide a picture of how, if any of these harvest levels did occur, the results would be translated into changes in two critical variables -- the 'residual' annual net growth for growing stock in the source areas when the demands from chip mills are fully accommodated (i.e., in Figure 2, the differences between 'excess growth suitable for chips' and 'assumed chip harvest levels'); and the supply of chips from cull material available to the mills if these demands were accommodated exclusively through the use of cull materials in the source areas (i.e., in Appendix A tables, the "equivalent supply for chip harvests). With the above kinds of information available, attention may focus more directly on the reasonableness of the assumptions.
Alternative Scenarios Based on Wood Products Data. In the past five decades, a number of timber products output (TPO) surveys have been conducted in Missouri (1946, 1958, 1969, 1980, 1987, 1991, 1994, and 1997). These surveys compiled information from primary processing plants (e.g., sawmills, etc.) to estimate the volume of wood products produced in the state and estimate associated harvest levels.. Over time the detail of these reports has been enhanced; and beginning in 1980, some statistics became available for individual inventory units in the state. However, unlike the statewide forest inventory, which is based on permanent plots on which trees may or may not be harvested, the TPO reports have been based on where trees are processed into products. The source areas for individual mills are identified by county. Thus with respect to the specific forest locations in which harvests occur, the TPO data is linked to county, while the statewide forest inventory data is linked to permanent plots. With respect to the actual time of harvest, however, timber products data has the advantage of being traceable to the year in which the wood is processed (which in turn is presumably linked more closely to the time of harvest); and this is monitored on a much more frequent basis for TPO data in comparison with the frequency with which FIA permanent plots are re-surveyed. In short, permanent plot and TPO use two different methodologies to arrive at estimates of timber removals. In so doing, permanent plots accumulate information about forest growth but say nothing about products, while the reverse is true for TPO information.
The long-term trends in timber product outputs indicate the general rate at which wood utilization is increasing or decreasing over time. This information on the rate of change in timber product outputs can be used to project proportional changes in future removals. Likewise, past rates of forest growth derived from forest inventory data can be used to project future rates of forest growth. However, all such projections must be viewed with a full understanding of the associated assumptions. The longer the projection period, the less likely the projected outcomes will be correct. Much like forecasts of the weather or the economy, this is true of any model used to forecast the future.
Figure 3 presents trajectories for industrial roundwood production in Missouri from 1946 through 2030. It is first evident that if the statewide average trend beginning in 1980 were to continue, the projected timber products output level would increase substantially. Similarly, if mean timber output from 1946 through 1997 were to continue, a much lower trajectory would occur. Neither of these trajectories is certain. On the one hand, timber outputs rarely if ever increase (or decrease) in a linear fashion over an extended period, but fluctuate up and down due to market price and other demand-related factors. This statewide picture, therefore, presents two sideboards reflecting the usual starting point for any projection efforts -- i.e., a linear extrapolation of historical trends. Moreover, it is clear that the extent of time encompassed by such trends will likely have a significant effect on the resultant projections. It is likely that the more recent trends are closer to present-day reality than those stretching into the far distant past. In any case, both these trends (i.e., 1980-1997 and 1946-1997) may be viewed as 'sideboards' within which the actual level of timber products output will likely fall over the next 30 years. The scenarios which follow are based on the 1980-1997 trend for the Eastern Ozark and Riverborder units as shown in Figure 3. In this light, they are inherently conservative in nature -- i.e., timber production may not increase at this same rate for the next 30 years. and more wood may be available for harvest than the scenarios will suggest.
Figure 3. Industrial roundwood production in Missouri
(1946-1997) and projections through 2030 based on the historical trend
from 1980-1997, prior to the establishment of chip mills in Missouri.
(Source : Piva et al. 2000)
( Note: Actual timber product outputs over the next 30 years are likely
to fluctuate rather than increase linearly.)
For the Eastern Ozarks, it is clear that an upward trend in timber products output, not as steep as the aggregate statewide pattern, commenced in 1980; it did, however, level off in the mid-1990's, and it is unclear exactly which trend will emerge over the next several decades. It should be noted, however, that the timber products output generated by the two chip mills since their arrival in southeast Missouri in the latter 1990's has not as yet entered into the TPO figures, and this could, of course, affect the future trajectory of product output in the Eastern Ozarks. Wood products output for the chip mills will be accounted for in the next TPO survey.
Based on the wood products data described above, a set of 14 additional scenarios was constructed depicting possible long-term trends in timber resource availability over the period from 1989 to 2030 (Shifley 1999b; 2000). From this perspective, two key questions include: Can Missouri forests meet the requirements for chip supply? Can Missouri forests do so while meeting the assumed increase in overall (non-chip) wood demand in the future? These are not the only important questions in the 'chip mill issue;' they are, however, the ones that these projections are designed to address.
The scenarios based on wood products data incorporated combinations of the assumptions used in the previous inventory-based analysis. For all scenarios, the supply of wood potentially available for chip was first reduced by 29% for road and stream buffers, steep slopes, and unusable species; this result was then further reduced by 27% (Mill Spring) and 19% (Scott City) for assumed unavailability of public lands as a timber source; and these combined reductions to the available wood supply were then reduced by 16% to account for source area overlap. For each source area, several of the scenarios also included an additional 20% reduction of the available land base (and associated available wood supply) to reflect private landowners who might not want to sell timber. Finally, the scenarios include increasing levels of demand for chips by the two mills, ranging from 200 thousand to 1 million tons per year for the Mill Spring facility, and from 350 thousand to 1 million tons per year for the Scott City mill. The scenarios are depicted in Figures 4 through 6, and individual scenarios are identified in Table 5.
Table 5. Scenarios for chip mill demands under two levels of private land availability
In examining the scenarios, it is first worthwhile to
consider what may happen if the current anticipated wood utilization
patterns by the two chip mills were to continue (i.e., 200 thousand
and 350 thousand tons of chips per year by the Mill Spring and Scott
City facilities, respectively), and
100% of private lands in the mill sourcing areas were available for
harvest (Figures 4a and 4c). The
other assumptions involving reductions in available wood supply for
chipping as described above are assumed to be in place.
For the Mill Spring facility, over the next 30 years (2000-2030), standing growing stock increases from 3.6 billion cubic feet and begins to level off at about 4.6 billion cubic feet; while standing non-growing stock (cull) rises at a slightly decreasing rate from just over one billion to slightly under 1.5 billion cubic feet. Over this period, annual growth of growing stock increases at a decreasing rate from about 116 million cubic feet to about 150 million cubic feet; while annual harvest for growing stock products other than chips rises at an increasing rate from 64 to 133 million cubic feet by 2030. Volume available for chips -- i.e., net annual growth adjusted for the various assumptions described above -- initially exceeds the 200 thousand tons (7 million cubic feet) chip harvest level and gradually declines over the 30-year period from 23 to 7 million cubic feet, thus equaling removals for chips.
From the above we may conclude that, given the assumptions underlying all scenarios, and with a continuation of anticipated present volume of chip utilization by the Mill Spring facility, if the entire chip harvest were taken from the growing stock, harvest from growing stock in the sourcing area would eventually equal annual growth in about three and one-half decades (2000-2035). By that time, total standing growing stock would be leveling off at around 4.5 billion cubic feet.
In the more restricted scenario in which only 80% of private lands are assumed to be available for chip harvest and buffers are assumed to be left on all roads and water bodies in the sourcing area
Scott City
Figure 4. Annual growth and removals, and excess growth suitable for chips, within a 60-mile radius of the Mill Spring and Scott City mills, under current anticipated volumes of chip utilization at two levels of wood availability from private lands (1990-2030) (Source: Shifley 1999b; 2000)
(Figure 4b), it is evident that annual growth of growing stock in the sourcing area begins to decline from a peak of about 81 million cubic feet in 2010, at which time harvest of growing stock for other than chips passes (i.e., begins to exceed) annual growth and continues to increase at an increasing rate. By 2030, annual growth is 56 million cubic feet and nonchip harvest is 135 million cubic feet. In this scenario, chip harvest is already equal to available volume for chips in the year 2000, and the difference becomes negative thereafter. Also, standing growing stock remains far less than the case where all private lands are assumed to be available for chip harvest and the assumption of buffers on all water and roads is removed (i.e., Figure 4a). Moreover, this lower level of standing growing stock itself reaches a peak of 2.5 billion until about 2005 and then gradually declines. In both of the above scenarios (Figures 4a and 4b), volume of non-growing stock (cull) is sufficient to meet the anticipated demand for chips.
When attention turns to a continuation of the anticipated level for annual chip harvest from the sourcing area for the mill at Scott City (i.e., 350 thousand tons or 12 million cubic feet), the following scenarios emerge (Figures 4c and 4d). Given the lower total growing stock in the source area and the higher level of annual harvest for chips required by the Scott City facility in contrast to that of Mill Spring, it is not surprising that projected annual growth of growing stock in the sourcing area under this scenario has already leveled off at 63 million cubic feet in the year 2000 and begins to drop within five years thereafter, ultimately declining by two-thirds to 20 million cubic feet by 2030. By the year 2005, harvest of growing stock other than chips passes (i.e., begins to exceed) annual growth and continues to increase at an increasing rate. Moreover, growing stock volume available for chips had already begun to decline by 1990 and annual chip harvest exceeded this volume when the Scott City mill began operating. Under this scenario, standing growing stock in the source area is projected to have already peaked at 2.6 billion cubic feet in the year 2000 and to decline to 800 million cubic feet by 2030. Standing non-growing stock (i.e., cull) remains steady at about 500 million cubic feet through 2005 and then gradually declines. Volume of non-growing stock trees is sufficient to meet the demand for chips under this scenario.
When the more restrictive assumptions regarding private land availability and buffer areas are incorporated into the above scenario (Figure 4d), the above pattern is substantially accelerated. Annual growth of growing stock has already dropped below the non-chip harvest level by 1994 and continues to decline at an increasing rate, reaching zero in about 2025. By the time the mill begins operating in 1997, any excess growing stock volume potentially available for chips has disappeared. Under this scenario, standing growing stock in the Scott City source area begins to decline from a peak of about 1.8 billion cubic feet in the year 2000 to zero by 2025. Volume of non-growing stock (cull) is sufficient to meet the demand for chips throughout most of this projection period, although it too approaches zero by the year 2025.
Given the inherent uncertainty in long-range projections, as well as the restrictive assumptions that ensure a conservative projection of future wood supply, the long-term implications of this analysis are tentative at best. However, they do reflect an educated estimate, using the best resource data available, regarding possible answers to the two key questions posed above. When focusing on growing stock (as opposed to cull material) as a potential source of chips, and with 100% availability of private lands for wood products, there is probably enough wood as growing stock to meet the 200 thousand ton annual demand for chips by the Mill Spring facility for at least the next three decades before volume available for chips declines to the level of chip harvests. For the Scott City facility, whose source area is only 31% timberland (as opposed to 58% for the Mill Spring source area), under these scenarios the growing stock volume available for chips has already dropped below the 1997 anticipated utilization level of 350,000 tons for the mill when it began operations. This in part reflects the fact that for the Scott City source area from 1992-1997, timber harvest was already increasing more rapidly on a percentage basis than was annual growth. During that time growth increased by 2.3% per year, while harvest increased by 3.7% annually. Thus even with no chip mill, potential wood supply from growing stock was declining in this area. If all of the Scott City facility's annual demand for chips (350 thousand tons) were to be supplied from growing stock in its source area, the volume of wood harvested for chips in the area would already exceed the growing stock volume available for chips as soon as the mill began operating.
Finally, when we impose the more restrictive assumption that only 80% of private lands in the source areas are available for timber harvesting, nonchip harvest exceeds annual growth much sooner (by 2010) in the Mill Spring area and chip harvest exceeds available chip volume shortly after the year 2000; while for the Scott City area, nonchip harvest had already exceeded annual growth in the early 1990', and the latter, along with standing growing stock, drop to zero by 2025. In both cases, but especially that of the Scott City mill, the sustainability of the forest resource in the source areas would obviously be greatly enhanced via the mills' utilization of the substantial volumes of cull materials within the forestlands surrounding the mills, in contrast to relying exclusively on growing stock in harvesting for chips.
Figures 5 and 6 present scenarios for Mill Springs and Scott City in which levels of chip harvest are assumed to increased above the baseline current anticipated levels of 200,000 and 350,000 cubic feet, respectively. In the top row of each figure, it is assumed that 100% of private lands will be available for timber harvesting in the source areas at each level of chip utilization; while the bottom rows incorporate an assumed 20% reduction in the availability of these lands. The assumptions restricting land (and wood available for chipping) common to all scenarios (Table 5) are also in place in these projections.
For the Mill Spring facility, with all private lands assumed to be available for timber harvest (Figures 5a-5c), an increase in chip utilization to 300 thousand, 600 thousand, and one million tons annually is reflected in the lower trajectories for standing growing stock, along with earlier peaks and subsequent declines. This is mirrored in similar trajectories and peaks for annual growth, as well as earlier points in time when nonchip harvests meet and exceed annual growth. Finally, as would be expected with increasing levels of chip harvest, the latter exceed growing stock volume available for chips at increasingly earlier points in time as levels of chip utilization increase, from about 2027 in the 300 thousand ton scenario to the mill start-up time in 1997 for the one million ton scenario. When
Figure 5. Projected growth and harvest levels from growing stock within a 60-mile radius of Mill Spring facility under increasing volumes of chip utilization at two levels of wood availability from private lands (1990-2030). (Source: Shifley 1999b; 2000).
Figure 6. Projected growth and harvest levels from growing stock within a 60-mile radius of Scott City mill under increasing volumes of chip utilization at two levels of wood availability from private lands (1990-2030). (Source: Shifley 1999b; 2000).
only 80% of private lands are assumed to be available for timber harvest (Figures 5d-5f), in addition to lower overall levels and more rapid declines in standing growing stock and annual growth, growing stock volume available for chips disappears more rapidly as the gap between available volume and chip harvest levels widens.
With respect the mill at Scott City, projections derived from assumed increases in annual chip utilization to 600 thousand and one million tons are depicted in Figure 6. Given the lower volume of timber resource in the source area relative to that of Mill Spring, the impacts of increased chip utilization on standing growing stock and annual growth are both more immediate and dramatic. When 100% of private lands are assumed to be available for timber harvest (Figures 6a and 6b), standing growing stock declines from the point in time in which the mill begins operating (1998) and would actually be exhausted by 2030 in the one million ton chip utilization scenario. Annual growth also rapidly declines from the time the mill begins operating. Since chip harvest is well above the growing stock volume available for chips when the mill begins operating, the latter rapidly disappears in a few years. When only 80% of private lands are assumed to be available for timber production (Figures 6c and 6d), the above pattern is accelerated. This yields a situation in which standing growing stock (and annual growth) are exhausted by the year 2022, and growing stock available for chips had disappeared several years before the mill began operations.
The analysis of scenarios can also shed some light on, while not answering completely, the question of additional chip mills locating in the state relative to potential wood availability. Thus, for example, the increasing levels of demand for chips in the source areas could be interpreted not simply as arising from one mill, but from additional mills locating in the state. Effects of new mills on timber growth and removals in the source areas would obviously not likely emerge from a new mill locating immediately adjacent to an existing one, but would more likely be manifest in the overlapping of certain portions of different source areas. It is almost a truism, however, that in terms of wood supply, the location of any additional mills will be a vital consideration in terms of their possible effects on the forest resource. It is clear from this analysis, for example, that -- in terms of growing stock -- locating another mill in the overlapping source areas of the two existing mills would not bode well in terms of sustainability of growing stock from a net annual growth perspective, unless nongrowing stock (cull) were used to supply the raw material. Potential impacts become harder to assess as we move away from the source areas. On an overall basis, many of these questions regarding viable source areas, economic feasibility, environmental soundness, and so on, could be -- and are, in some states -- addressed in the permitting process when mills first apply for clearance to locate in a state. Finally, the above discussion has pertained primarily to growing stock volumes as a potential source of chips. Availability and likely utilization of cull material for chips are also critical aspects to be considered. Regarding the former, Missouri clearly has an ample supply; so attention turns to the question of how much of this material is likely to be utilized.
It has been emphasized that the assumptions common to all the scenarios described earlier lead to inherently conservative estimates of potential wood availability for harvest in the chip mill source areas. In the case of the Mill Spring area, for example, Table 4 indicates that the difference between annual growth and annual removals results in 48 million cubic feet of net annual growth. When the combined effects of the adjustments (i.e., reductions) to net annual growth arising from the assumptions listed at the bottom of the table are applied to the above figure, the net annual growth is reduced by 54% to 22 million cubic feet. This figure represents the adjusted excess growth (i.e., growth - removals) in the Mill Spring source area that is potentially available for chips (and, of course, other wood products as well). These same assumptions contributing to adjusted net growth underlie the scenarios listed in Table 5. The figures and graphics depicting the scenarios are direct results of projections made with these assumptions in place.
Several points merit emphasis with respect to the above. First, different groups with different visions for the management of Missouri forests may be more or less comfortable with different sets of assumptions. Secondly, it is possible to prepare any number of scenarios for growth, removals, chip harvest levels, and so on, in the mill source areas and elsewhere. Thus, for example, a 'liberal' scenario relative to wood availability could be constructed just as readily as the more conservative approach used in this analysis. Estimates of growth increases through better management could also be incorporated. Thirdly, efforts to ensure the sustainability of Missouri forests also depend on other kinds of assumptions and resultant scenarios in addition to those directly related to timber growth and yield. Many such assumptions or expectations focus on social actions that may have repercussions for Missouri forests, and their scope may extend well beyond the boundaries of the state. Thus, for example, based on estimates by the mills themselves, the baseline demand (expected utilization) levels for chips utilized in the preceding analysis have been 200,000 tons (7 million cubic feet ) and 350,000 tons (12 million cubic feet) for the Mill Spring and Scott City mills, respectively. This reflects the assumption that the above figures all represent demand levels for these mills operating at average capacity (e.g., one shift). However, it is not inconceivable that demand for chips stimulated by, for example, a recovering and/or expanding economy in the Pacific rim could encourage mills to expand beyond average capacity to a greater or lesser degree. Such a 'social scenario' could be translated into projected demand levels for chips, incorporated into the above kinds of projections, and interpreted in terms of potential impacts on the sustainability of Missouri forests.
The key point of all of the above is that it is the assumptions underlying the scenarios that will shape the results that are interpreted by those with differing perspectives on the long term viability of the state's forests. For some, the assumptions employed in this report may reflect a cautious yet reasonable approach toward sustaining the state's forestlands; for others they may seem too cautious and, therefore, unrealistic. Given the substantial degree of uncertainty inherent in any such effort, such differing perspectives are virtually inevitable. Like any aspect of public policy, the most fruitful pathway to reconciling these differences is to be found through open discussion of the logic, reasonableness, and desirability of the assumptions underlying our efforts to predict the vital factors that will ultimately determine the sustainability of Missouri forests.
In summary, understanding the relationships among availability of standing volume and annual growth -- both growing stock and cull material -- and removals for sawtimber, chips, and other wood products is one key ingredient for arriving at viable policy alternatives that encourage sound utilization of Missouri's forest resources in a sustainable manner. At the same time, all that is standing is not necessarily economically available (even after reductions for the assumptions used above); and what is available will not necessarily be utilized. Factors affecting these other aspects of the 'chip mill issue' are the focus of the following pages.
Forest Land Change Detection
Changes in the composition and structure of Missouri forestlands are at the heart of the question of the ecological and economic sustainability of the state's forest resource base. The USFS forest inventory and analysis data (FIA) provides important information in this regard -- collected from about 5600 points in forested landscapes in Missouri. But due to the density of the sampling, it is difficult to make inferences from this data for areas smaller than the county level. However, another technique for assessing changes in Missouri's forest land base is available that may complement the statewide forest inventory.
Satellite remote sensing (RS) is a methodology which can be used to classify areas on the basis of amount of biomass. Satellite RS systems record electromagnetic radiation (EMR) from target objects. The data produced is multi-spectral; sensors on the satellite collect information in six bands including, but extending beyond, the range of the human eye. The thematic sensor can 'see' well enough to detect changes in forest cover with a high degree of accuracy. RS data for monitoring forest cover has a spatial resolution (i.e., grain or pixel size) of 30 meters. With this it is possible to produce scenes that measure about 185 kilometers on a side.
In this way, data or images can be used to establish a
baseline for monitoring the area and location of change in the Missouri
Ozarks. Remotely-sensed data yields a snapshot in time, which cannot
be accomplished on the ground; it is not possible to sample large forest
areas and provide an instant picture of the results. Using remotely
sensed data, different kinds and intensities of harvests would appear
differently in terms of colors of images. It can reveal that x acres
of forest land is no longer forest land the next year, and it can pinpoint
the location of those acres. However, such data may not be able to answer
the question as to whether the land was harvested or bulldozed . What
it does provide is location-specific information that the land was forested
and now is not, based on the amount of biomass present.
Remotely sensed data also cannot directly reveal whether any growth
has occurred on the site (e.g., via replanting or otherwise) until such
growth reaches a certain size. Thus, for example, using satellite imagery
alone, it would likely take three to five years to determine that after
an area is cleared it is being allowed to grow back as forest land.
When seedlings are just emerging and immediately after, their color
and transparency are similar to that of grass. But after growing dense
and tall enough, they exert a cooling effect on the micro-areas immediately
around them; and this is perceived by the satellite sensors and expressed
in different colored images
When used as a tool to understand the pattern and extent of timber harvesting across the landscape, the following picture of what RS data can do emerges. Consider, for example, the case where two scenes of same area are produced -- one from 1986 and one from 1992 Also assume there is 50,000 acres less forest in 1992. How would one determine how much of that 50,000 acres is pasture and how much of that land has been recently harvested and is in the process of re-growing a forest? It would not be possible to ascertain this directly from the 1992 scene. However, there are ways this data can be incorporated into such efforts to detect forest land change. On the one hand, with a continuing series of updated images, this could be monitored more closely. Another approach would be to select a random sample of those cleared areas (in the 1992 scene) and obtain a statistical measure of what percentage of them are pasture, regenerating forestlands, and so on.
This kind of data and information can also indirectly have a positive impact on how privately-owned forestland is managed. It can be very useful for public agencies in developing programs for private forestry that provide incentives for landowners to manage in ways that are both ecologically sound and financially beneficial. It allows agencies to have a better landscape picture, so that in identifying areas at risk (e.g., bottomland hardwood forests) they can better tailor programs in ways that would be appealing to those landowners. Other advantages of RS data are related directly to its value in conducting ecological assessments. Remotely sensed data can be used to assess key variables essential to understanding large-scale relationships essential to ecosystem health. Thus, for example, it is valuable for forest edge and interior detection that may be critical habitat factors for the viability of certain wildlife species. It is also used for characterizing features of watersheds that capture their role in important hydrological processes.
Remotely sensed data can be archived; it does not have to be analyzed immediately. Moreover, a data bank already exists for Missouri from the period when satellite orbits began (1971) through the present. Thus, for example, data for 60-mile radii around Mill Spring and Scott City between 1971 and 1999 may be accessed from the archives. Of course, satellite images must be purchased from the agency or organization controlling the satellite's operation. A new satellite has recently gone up, and the U.S. Geological Survey is controlling the policy for the data distribution from that satellite. While the revised cost structure is yet to be finalized, the current cost of RS imagery (i.e.,for a 185 km2 image for before any work is done with it) could range from $900 per image at a minimum (probably more) to $2500 at a maximum (likely not that high). Nonetheless, it can be obtained quickly and relatively cheaply versus any other technique for large areas. Moreover, results can be combined with other spatial data layers in a geographic information system (GIS) for analysis. Finally, data depicting forest cover and other land characteristics can be updated relatively quickly, given that the satellite passes over every sixteen days. In Missouri, the organization most closely involved with use of satellite imagery in the field of natural resources is the Missouri Resource Assessment Partnership (MORAP), which is based at the University of Missouri-Columbia. The organization is funded entirely through partnerships with state and federal natural resource agencies. MORAP data is stored on a USGS server at the Columbia Environmental Research Center.
Both the statewide forest inventory (FIA) and the remote sensing capabilities of MORAP provide important tools for enhancing the sustainability of Missouri forests. Their most effective use will, however, come not separately but as part of an integrated resource assessment package for monitoring the status of and patterns of change in the state's forestlands. With the improvements in FIA methodology, it will now be possible to inventory (via sample plots) one-fifth of the state's forest lands each year, thus reducing the overall inventory cycle to five years. Remotely sensed snapshots of large contiguous forestland areas provides and added dimension to the FIA data derived from sample plots, allowing for the monitoring of many important landscape-level ecological variables. Moreover, such images of areas within, for example, an individual county provide a valuable source of information on characteristics of smaller areas of forestlands, compensating to some degree for the rather limited ability to make inferences from FIA data to sub-county areas. In summary, some powerful tools are available for understanding and assessing the diverse forestlands of Missouri; the challenge remains that of combining their distinctive capabilities in the most effective way to ensure a sustainable forest resource base.
