Enviromental Geology (1993)
note: figures and tables have been deleted at this time due to reproduction problems.
Effects of deforestation upon slopes in limestones and in volcanic rocks in the Benson River valley, northern Vancouver Island, have been investigated quantitatively. Post logging soil erosion and vegetal regeneration success were assessed by measuring soil depth, percent bare rock and moss cover, and the numbers and diversity of trees, shrubs, and plants on 25 sampling sites, each containing ten measuring quadrats selected at random.
Sixteen sites were on the Quatsino Formation, a well-karstified lime-stone, and nine on the Karmutsen Formation of basaltic lavas. Eight sites were of virgin forest, 16 were logged between 1970 and 1983, and one (on limestone) was logged in 1911. Both bedrock types were significantly affected by the cutting.
There was greater loss of soil and an increase in bare rock on the limestones. Erosion was increased significantly by burning on the limestones but not on the volcanics. Within-group comparisons on the limestones determined that steeper slopes and harder burned areas suffered the most and are slowest to regenerate. Volume of timber on the 1911 site was 19 percent of that in similar uncut forest sites. It appears that complete recovery on the barren limestone slopes will require at least some centuries.
In the northern half of Vancouver Island, British Columbia, Canada, there are rugged limestone karst terrains that are similar in many respects to those around the northern Mediterranean. Great coniferous forests growing upon them are in the process of being cleared for the first time, using modern technology based on heavy equipment. This paper reports on a study of the effects of clear-cutting upon the karst and, more particularly, upon features of the recovery and regrowth of the forest afterwards. It was what may be termed a baseline study, aiming to provide quantitative data on the state of sample forest before and after clearance, to compare the states on limestone with those on nonkarstic (or control) rocks, and to study the effects of deliberate burning of the forest waste (slash).
As far as the authors are aware, there have been no previous quantitative studies directed specifically at impacts of primary deforestation on karst slopes. However, there are many studies of deforestation on slopes in general. In the Pacific Northwest of North America, as else-where, it is found that runoff increases after logging, roughly in proportion to the size of the area cleared. Soil erosion is the principal damaging consequence, being in proportion to hillslope gradients and amounts of precipitation (runoff). Erosion rates are greatest in the first year or two after cutting (e.g., Roberts and Church 1986). There are distinct successional stages in regrowth; an early herbaceous stage is followed by shrub and young tree stages (Dyrness 1973; Klinka and others 1985). Species diversity tends to increase when an area has been harvested (Hix and Barnes 1984). Light burning may be advantageous but high-intensity burns tend to reduce the rates of recovery (Jones 1983).
Northern Vancouver Island is a mountainous region with relief up to 2000 m asl. It is developed in faulted and folded volcanic and sedimentary rocks of Triassic and later age. karst occurs on the Quatsino Formation (Triassic), a very uniform, micritic deep-water limestone that is 800 m in thickness (Fig. 1). It is underlain and overlain by marine volcanics that comprise the control (nonkarstic) rocks in this study. The sampling area was the valley of the Benson River, where strata form regular scarp and dip slopes but the crests of the cuestas have been reduced to more gentle, undulating plateaus by glacial action. Limestones in the Benson valley rise to 900 m asl. The region was intensely glaciated and mantled with deposits of till or outwash when the ice receded approximately 14,000 yr. BP (Howes 1981).
Climatic data for Port Alice, a station at sea level 25 km from the study area, are shown in Fig. 2. The temperature regime is mild. Soil may be eroded during the winter because it is not frozen. Mean annual precipitation is 3250 mm at sea level and is estimated to increase to about 5000 mm on the highest limestones that we studied. These are values for a superhumid regime essentially similar to that reported by James in this volume (pages 144-151). There is much runoff, but it should be noted that, as in the Mediterranean region, there is a distinct summer dry season when droughts may occur.
The area is defined botanically as Western Hemlock Forest. Western hemlock (Tsuga heterophylla) itself, silver fir (Abies amabilis), and western red cedar (Thuda plicata) are the dominant species. They are said to grow best on well-developed ferric or humic podsols (Edgell 1979). In the Benson valley trees grow to much greater dimensions on the limestones than on comparable sites on the volcanics (Fig. 3); dense stands of magnificent trees up to 2 m in diameter and 30 m in height can be found even on the skeletal soils of the steeper limestone slopes. The undergrowth is lush. Deadfall and organic litter are almost completely overgrown by thick mosses, herbaceous plants, shrubs, and saplings. There is no bare soil or rock exposed except in stream channels or where a large tree has been uprooted.
The aboriginal Indian population used the forests for collecting and some construction timber, but with negligible impact. Modern logging commenced around 1900 AD. One extensive, undulating limestone plateau, the Old Re-growth, was cleared in 1911 for mining purposes. Logging for timber began elsewhere in the Benson valley in 1970 and was fully modern in method, i.e., obtaining access by dense networks of bulldozed dirt roads, felling by chainsaw, extracting by crane and lumber truck. After clearing an area, it was standard practice to burn the slash and waste wood in order to provide some fertilizer, setting the fire with a match when the slash had dried (light burning). In the 1970s there were experiments with much hotter burning with petroleum (intense burning); ground surface temperatures may then have momentarily approached 900°C, close to the liming temperature for limestone. Once cut and burned, areas were left to reseed naturally. According to Barker (1977),80 years is the industry standard time required to obtain a harvestable second crop of timber following clear-cutting.
Fieldwork was undertaken in the summer of 1986 (Harding 1987). It was designed (I) to compare clear-cutting on the karstified limestones with the (nonkarstified) volcanic rocks; (2) to test effects of differing slope type, gradient, and aspect (scarp slopes, dip slopes, undulating plateaus) on both rocks; (3) to test for attitudinal effects (above or below 500 m asl); (4) to test effects of burning (not burned, burned, intensely burned); and (5) to study forest cutting status and regrowth [virgin timber; old regrowth-1911; new cuts (I) 1970-1974, (ii) 1975-1979, (iii) 1980-1986).
Twenty-five sample plots 50 x 50 m each were selected for detailed measurements, 16 on limestone and nine on volcanics, with further division by slope type as indicated in Fig. 4. This sample was also stratified by altitude (upper/lower), by cutting status, and by burning treatment (Fig. 4) Ten quadrats of 5 x 5 m were then selected at random in each 50- x 50-m plot, for a total of 250 sampling quadrats. The following variables were measured in each quadrat: number and species of trees, their height and basal diameter; cut tree stumps and their basal diameter; shrub species, numbers and height; mean depth of soil; and percentage bare rock area, soil cover, and moss cover. In each quadrat a 1.5- x 1.5-m subplot was chosen to identify and count species of small herbaceous plants (such as wild flowers) and their numbers.
All data were entered in computer master files. Frequency tables were produced to evaluate general trends. Relationships were investigated by cross-tabulation, chi-square tests, and Pearson's correlation. Scheffe multiple comparison procedures were used to determine which population means differed from each other. Analysis of variance was used only as a descriptive statistic because most data were nor normally distributed.
Our qualitative observations in the Benson valley and discussions with workers there tend to confirm the general findings in other logged regions-that there are great in-creases in run off and soil erosion, especially during the first one or two years. As expected, erosion is greater on steeper slopes. The quantitative studies revealed no significant differences between the upper and lower valley (i.e., effects of altitude); these are evidently overwhelmed by the general wetness of the climate.
Soil losses were much more severe on the karst (as shown quantitatively below), with most soil passing into the small cavities of the epikarst zone. Soil intrusion and plugging were noted in caves but there had been insufficient exploration and mapping before the clearance to permit proper estimation of its contributions. Some shallow cave roofs, speleothems, etc., were shattered by road blasting shock or load from bulldozers and other heavy machinery. Dolines were obstructed by waste timber and karren crushed where heavy machines had passed over them. Upon exposure by soil erosion after logging, the smooth, gleaming surfaces of karren quickly lose their patina; they are visibly roughened by dissolutional micro-pitting within a few years. Limestone surfaces that had experienced intense burning displayed many flaking or shattered patches.
The most important findings of the quantitative studies are summarized in Table 1, which presents results of Kolmogorov-Smirnov two-sample tests where a significant (one star) difference between any two data sets implies difference at a 0.1 level, rising to a highly significant (three star) difference at a 0.001 level.
Table 1 A compares the sum of all clear-cut quadrats on the limestone with the sum of all virgin forest quadrats on the limestone, and likewise on the control volcanic rocks. The impact of logging was significant to highly significant with respect to seven of nine variables on each rock type. However, there was no significant impact on mean soil depth on the volcanic rocks.
Table I B compares a given condition (e.g., virgin forest) on the limestone with the same condition on the volcanics in order to determine whether the karst is significantly more (or less) affected by logging and burning than the control sample. In the natural forest before logging (row 1, Table 1 B), it is seen that there were no significant differences in the important variables of soil depth, percent bare rock and moss cover, and number of trees. Cutting without any burning (row 2) reduced the moss cover on the volcanics significantly more than it did on the limestones, although both forests suffered substantially. Burning (row 3) and intense burning (row 4) introduced significant to highly significant differences in soil depth and percent bare rock, affecting the karst more severely in each case.
Some quantitative examples of the differences are shown in Fig. 5: histograms of soil depth, bare rock, and moss cover that compare limestones with volcanics in virgin forest, on lands logged between 1970 and 1985 (but not differentiated by measure of burning), and the Old Regrowth that was logged in 1911. Under the virgin forest, soil depths are essentially identical when averaged over very large samples such as these. Clear-cutting produces negligible loss on the volcanics (when averaged), but there is a mean reduction of 11 cm (or 40 percent) on the limestones. Soil was lost entirely on steeper slopes, as evidenced by the increase of percent bare rock area from 3 percent to 9 percent on the volcanics and from 2 percent to 23 percent on the karst.
In the study area, 75 percent of the sample sites logged between 1970 and 1985 were also burned, one third of them (on the limestone only) being subjected to intense, high- temperature burning. Slash is burned to reduce fire hazard, to fertilize the soil for natural reseeding or artificial planting, and to reduce competition from shrubs. It is a common practice in British Columbia, although intense burning has been discontinued in the Benson valley. Table 2 illustrates the most significant effects established by our investigation. Burning reduces mean soil depths and increases the extent of barren rock quite drastically on the limestone. Numbers of shrubs and of flowers and other herbaceous plants are also reduced when compared with the logged but unburned sites, although the diversity of all species and the number of trees are not affected. In contrast, on the volcanic rocks burning appears to be doing what it was intended to do; there is no increase in the amount of soil lost but numbers of flowers, shrubs, and seedling trees are significantly increased. The difference must be attributed to presence of epikarst cavities on the limestone. It ensures that soils will be thinner on many slopes. Burning (especially intense burning) destroys the litter and moss cover and even the root mesh within the soil that is binding it together. The residue of the burn is washed into karren pits and guttering by the first rains. The sample mean value of 40 percent bare rock on intensely burned limestones (Table 2) implies that on the steepest slopes and in other adverse situations nearly 100 percent of the surface will be barren; Fig. 6 is an example from a dip slope in the lower valley.
The recovery of the natural vegetation after logging and burning is also of interest. Figure 7 displays some effects on the herbaceous plant populations growing on the limestone. Three particular plants common in all of the virgin forest sites remained similarly abundant for 15 years or more but are absent from the 75-year site. Another selection of plants that thrive only in shady forests are able to maintain themselves for ten years after clearance but then are overcome. Certain moisture-loving species are not affected, but there is a radical increase in the numbers and dispersion of species associated with dry soils, indicating the increased droughtiness that logging produces in patches of the karst even in this very wet climate.
The Old Regrowth site was an optimal site for recovery because it was on gently undulating land with few steep slopes. There is well-developed epikarst and some deep dolines. It was not burned. It is fully recolonized by plants, shrubs, and young trees (western hemlock and red cedar) today. However, the total numbers of trees and their basal areas in our sample quadrats are only 15 percent and 19 percent, respectively, of the mean values in limestone virgin forest, making them comparable with the most impoverished natural vegetation (the steepest sites) on the less productive volcanic rocks. The impression is that further centuries will be required before the forest regains its past magnificence here.
This paper has documented the effects of clear-cut logging and burning on karst slopes by comparing them with adjoining, undisturbed natural forest, and with forested and logged lands on volcanic rocks nearby. Statistical analyses of large data sets revealed many significant effects. The karst was more adversely affected, displaying considerable losses of soil, moss, and litter, especially where cutting was followed by deliberate burning. Erosion was most extreme on the steeper slopes, many of which are left quite barren. Recovery is slow; even at a physiographically advantageous site only 20 percent of the original volume of timber has grown after 75 years. Some species of plants that are common in the natural forest do not survive the regrowth, while others proliferate.
In Ontario, Canada, there are very extensive tracts of karstified limestone and dolomite plains (pavements; Ford 1987). These were also densely forested until clear-cut between 100 and 170 years ago. All soil and litter was lost over wide areas. Without artificial aid the forest has re-established itself, although the trees are of undesirable species or much smaller than were there formerly. The experience is similar around the Mediterranean, as Gams reports in this volume (pages 144-151). in contrast to the situation that prevails on almost all other rocks, even the complete loss of soil and litter from the surface of limestone and dolomite will not prohibit the reestablishment of forest if there is a well-developed epikarst. This is because much soil and litter, plus other nutrients and water are retained in the karren troughs and microwaves that constitute the epikarst zone. It is a paradox that surface soil erosion is more severe on karst rocks than it is others (because of the high density of cavities that catch any detached soil and remove it from sight), yet vegetal recovery on severely eroded sites can be much more complete.
The situation does not appear to be so advantageous on many of the limestone slopes left bare by logging and burning in the Benson valley. These slopes are too steep to have developed (in the time available since the last glaciation) a dense epikarst of microcaves in underlying bedding planes, that are linked to the surface by karren troughs, pits, or shafts. Instead, shallow pits and dissolutional gutters (rinnenkarren) predominate. They are not adequate traps for soil and litter following logging. Trees that grew on these slopes (often to magnificent girth) were rooted in thin soils that have washed away down the gutters, plus litter, deadfall, and mosses that had accumulated to great depth. The trees cannot return until the litter and moss base has built up again, a process that seems likely to require at least some centuries. On long bare slopes such as shown in Fig. 7 it is possible that the forest cannot return until the next glaciers have deposited a new layer of till. Surfaces that are both glaciated and karstified are vulnerable to soil erosion and desertification that, in terms of human history, is permanent.
Kathy Gladysz, Joyce Lundberg, and Brenda Stephan are thanked for their assistance in the field. The Macmillan-Bloedel Company courteously permitted access to their cutting areas in the upper Benson Valley. This work was funded by a research grant to Ford from the Natural Sciences and Engineering Research Council of Canada.
Barker ML (1977) Natural resources of British Columbia and the Yukon. Vancouver: Douglas, David and Charles, Ltd.
Dyrness CT (1973) Early stages of plant succession following logging and burning in the western Cascades of Oregon. Ecology 54: 57-69
Edgell MCR (1979) Forest Industry; Vegetation. In Forward CN (Ed.), Vancouver Island: Land of contrasts, vol 17. Victoria: Western Geographical Series. pp 105-139
Ford DC (1987) Effects of glaciation and permafrost on the development of karst in Canada. Earth Surface Processes and Landforms 12(5):507-521
Harding KA (1987) Deforestation of limestone slopes on northern Vancouver Island. Unpublished Msc thesis. McMaster University, Hamilton, Ontario. 188 p
Hix DM, Barnes BV (1984) Effects of clearcutting on the vegetation and soil of an Eastern Hemlock-dominated ecosystem, upper Michigan. Can J For Res 14:914-923
Howes EE (1981) Terrain inventory and geological hazards northern Vancouver Island. APD Bulletin 5, British Columbia Ministry of the Environment.
Jones T (1983) Wilderness or logging? Case studies of two conSicts in B.C. Federation of Mountain Clubs of British Columbia, Vancouver, BC
Klinka K, Scagel AM, Courtin Pa (1985) Vegetation relationships among several ecosystems in southwestern British Columbia. Can J For Res 15: 561-569
Roberts RG, Church M (1986) The sediment budget in severely disturbed watersheds, Queen Charlotte Ranges, British Columbia. Can J For Res 16: 1092- 1106
Department of Geography,
Department of Geography,
RETURN TO KARST MANAGEMENT