ECOLOGICAL CHARACTER OF A SMALL-SCALE LINEAR WIND EVENT IN AN OLD-GROWTH BOTTOMLAND HARDWOOD STAND IN SOUTH-CENTRAL ARKANSAS James M. Guldin Research Forest Ecologist, USDA Forest Service Federal Building, P.O. Box 1270, Hot Springs, AR 71902 Brian R. Lockhart Assistant Professor, School of Forest Resources University of Arkansas at Monticello P.O. Box 3468, Monticello, AR 71656 Lance Peacock Arkansas Nature Conservancy 300 Spring Street, Suite 717, Little Rock, AR 72201 ABSTRACT: The Moro Bottoms Research Natural Area, a 45-ha bottomland hardwood stand in the Moro Creek flowage near Fordyce, Arkansas, is a relict old-growth stand of oaks, sweetgum, and cypress which the Arkansas Nature Conservancy acquired from Georgia-Pacific Corporation in 1986. As an element of that acquisition, a 100% inventory of the stand was obtained on the area. During a summer evening in 1989, a line of thunderstorms moved across south-central Arkansas, affecting forest vegetation in a localized manner. These winds promoted windthrow in a number of forested landscapes including the Moro Bottoms Research Natural Area. Following the windstorm, a 100% inventory of all windthrown trees was taken, providing a unique data set to scrutinize the ecological effects of the windstorm. There were 47 different species of woody stems on the tract, of which ten were uprooted. Sweetgum and the several white oak species were proportionally resistant to uprooting, whereas the red oaks were disproportionally uprooted. A larger proportion of overstory trees of intermediate diameter (22-30" d.b.h.) were uprooted than of small (12-20" d.b.h.) or large (32+ d.b.h.) diameters. Analyses focused upon the influence of species identity, diameter, and azimuth of uprooting, and on the fortuitous happenstance of a 100% pre-disturbance inventory in characterizing the ecological pattern of windthrow. INTRODUCTION The frequency and magnitude of natural disturbances such as windthrow in southern bottomland deciduous forests exert an important influence on forest structure. Catastrophic wind disturbances such as hurricanes and tornados (Turner, 1935; White, 1979) have long been recognized as a major force in shaping the structure of forest stands (Curtis, 1943; Oliver and Stephens, 1977; Oliver, 1981). To a lesser extent, the ecological values of winds that are not catastrophic but merely damaging have received less attention (Stephens, 1956; Webb, 1989), especially in southern United States forests. The opportunities for ecological study of such catastrophes is further limited by the nature of land ownership. In the major and minor river bottoms that dissect the upper Coastal Plain of the West Gulf region, the richly productive bottomland hardwood stands are typically owned by forest industry and non-industrial private landowners. Government ownership in the area has been the result of recent acquisition. In the event that catastrophic events occur in these forests, the past actions of landowners have been to salvage the windthrown trees rapidly. When disturbance affects bottomland forests set aside as ecological reserves or natural areas, the opportunities to study the ecological effects of such disturbances increase. In this paper, we report on the effects of a less-than-catastrophic disturbance that occurred within one of the rare ecological reserves in the region -- from a research perspective, a lucky situation. The study objective was to investigate the effect of a small-scale linear wind associated with the passing of a line of summer thunderstorms on the the composition, structure, and direction felled in an old-growth bottomland hardwood stand in Cleveland County, Arkansas (U.S.A.). METHODS The Moro Bottoms Research Natural Area (RNA) encompasses 45 hectares along Moro Creek in south-central Arkansas. This stand was considered old-growth based on the large diameters of the overstory trees and ages of selected trees. Two red oak trees were bored and each was determined to be approximately 150 years old, while three white oak trees averaged about 250 years old. There was little visual evidence of past harvesting practices and no evidence of other large disturbances. The Moro Bottoms RNA is one of the best remaining examples of a typical southern wet-mesic floodplain forest presently known in Arkansas (unpublished report by the Arkansas Nature Conservancy; copy on file at the University of Arkansas at Monticello). Due to the rarity and unique character of this old-growth bottomland hardwood stand, the Arkansas Nature Conservancy, a land conservation organization, purchased this stand from Georgia-Pacific Corporation in 1985. The area was then acquired by the Arkansas Natural Heritage Commission for continued preservation. As part of the acquisition process, a consulting forester with extensive experience in bottomland hardwood management was hired to conduct a 100% inventory of all trees greater than approximately 25 cm d.b.h. This consulting forester noted that this stand was unique, and to the best of his knowledge, "the only old-growth stand with substantial volumes of red and figured redgum remaining in the mid-south." Gum and redgum refers to sweetgum (Liquidambar styraciflua L.). One summer evening in July, 1989, a line of thunderstorms moved across south-central Arkansas. These storms were accompanied by unusually high winds and caused windthrow in many areas, among them the Moro Bottoms RNA. The storm uprooted 207 trees, comprising ten different species. Since this area was designated a research natural area, no salvage operations were conducted and the windthrown trees were left to decompose naturally. In August, 1989, a 100% inventory of the windthrown trees was conducted by species, diameter (d.b.h. for trees >25 cm), and azimuth direction of each fallen tree. Information from this post-disturbance inventory was then compared to the pre-disturbance inventory to determine post-disturbance stand composition. Two sources of error were possible in comparing pre- and post-disturbance stands. First, since three years had passed since the initial inventory, trees may have increased in diameter enough to have moved into larger diameter classes. A second source of error involved measurement error. Trees measured during the post-disturbance inventory were in contact with the soil surface, in some cases with the bole half-buried in soil. Therefore, calipers were used to measure tree diameters instead of measuring tapes, as was done in the first inventory. Based on analyses of diameter distributions, we felt that these sources of error were small, especially in light of the value of the 100% pre- and post-disturbance inventories. Data from both inventories were summarized into importance values for each species representing the summation of relative density and relative dominance (i.e., basal area; Spurr and Barnes, 1980). RESULTS Species Group Designations A total of 47 woody species were tallied during the pre-disturbance inventory. Overstory species were placed into the following species or species groups (scientific names follow Duncan and Duncan, 1988): (1) sweetgum; (2) white oaks, including white oak (Q. abla L.), overcup oak (Q. lyrata Walt.), and swamp chestnut oak (Q. michauxii Nutt.); (3) cherrybark oak (Q. pagoda Raf.); (4) other red oaks including water oak (Q. nigra L.) and willow oak (Q. phellos L.); (5) hickory species including pignut hickory (Carya glabra (Mill.) Sweet), shagbard hickory (C. ovata Mill.), and mockernut hickory (C. tomentosa (Poir.) Nutt.); (6) baldcypress (Taxodium distichum L.); (7) two ash species, green ash (Fraxinus pennsylvanica Marsh.) and white ash (F. americana L.); (8) other hardwoods including red maple (Acer rubrun L.), common persimmon (Diospyros virginiana L.), American beech (Fagus grandifolia Ehrh.), American holly (Ilex opaca Ait.), blackgum (Nyssa sylvatica Marsh.), and winged elm (Ulmus alata Michx.); and (9) loblolly pine (Pinus taeda L.). Species groups were designated by the consulting forester during the pre- disturbance inventory; therefore, the same classification was used during the post-disturbance inventory. Pre-disturbance Conditions A total of 3,456 trees >25 cm d.b.h. were tallied during the pre-disturbance inventory. Based on the pre-disturbance diameter distribution, the 40-cm diameter class contained the greatest number of trees with 454 stems ( Figure 1 ). The largest diameter tree was a 135-cm sweetgum. Beginning with the 40-cm diameter class, the diameter distribution conformed to the classic reverse-J curve. This distribution may represent an uneven- aged stand, although this may or may not be case since only five trees were actually aged. Sweetgum had the greatest density among species or species groups with 30%, or 1036 of the trees >25 cm d.b.h. Sweetgum also had the greatest dominance at 33% of total basal area giving this species an importance value of 63 (Table 1). White oaks had the second greatest importance among species or species groups at 33 composed of a relative density of 14% and a relative dominance of 19%. Other red oak species had the third greatest importance at 26. Hickory species had the fourth greatest importance at 22. This value was composed of a relative density of 13% but a relative dominance of only 9%, indicating that the mean stand diameter of the hickories was considerabley less than that of sweetgum and the oaks. Cherrybark oak had the fifth greatest importance at 21. Finally, importance values for each of the remaining species or species groups was >16. Post-disturbance Conditions The severe winds came from the northwest; therefore, the greatest number of trees fell southeast (100-180 degrees; Figure 2 ). But some trees fell in a different direction than the way the wind was blowing. This may have been due to localized vortices of wind developing on a occasional basis, or may have been due to physical deflection of trees by associates that fell on them. The post-disturbance inventory did not attempt to measure these 'domino' effects. A total of 207 trees, or 6.0% of the total number of trees >25 cm d.b.h., blew over during the wind event. Based on the post-disturbance diameter distribution, trees from the 35-120 cm diameter classes fell during the wind event ( Figure 1 ). Most windthrow occurred among the 55-100 cm diameter classes. The largest tree felled was a 110 cm white oak, although few trees larger than 105 cm were lost. Table 1. Pre-disturbance density, relative density, basal area, relative basal area, and importance values by species or species groups for the Moro Bottoms RNA. (Density = stems/ha; relative =% basal area = sq.m./ha) Species or Density Relative Basal Relative Importance species group density area basal area value ------------- ------- -------- ------- ---------- ---------- Sweetgum 1,036 29.98 280.78 33.36 63 White oaks 488 14.12 155.90 18.52 33 Cherrybark oak 369 10.68 85.80 10.19 21 Other red oaks 456 13.19 106.45 12.65 26 Hickories 466 13.48 72.95 8.67 22 Baldcypress 252 7.29 61.29 7.28 15 Ashes 48 1.39 8.04 0.95 2 Other hardwoods 238 6.89 44.11 5.24 12 Loblolly pine 103 2.98 26.43 3.14 6 Total 3,456 100.00 846.75 100.00 200 In terms of post-disturbance importance, sweetgum gained in importance from 63 to 67 out of a possible 200 (Table 2). White oaks also had a slight increase in importance from 32.6 to 32.9. These increases in importance for both sweetgum and white oaks, even though each lost trees during the wind event, was due to the far greater number of red oaks lost during the storm, both cherrybark oak and other red oaks. Cherrybark oak decreased in importance by 3 (from 21 to 18) while the other red oaks group also decreased by 3 (from 26 to 23). Hickory species also had a slight decrease in importance, from 22 to 21, while the remaining species or species groups had slight increases in importance. Based on post-disturbance diameter distributions by species, only 2% of the sweetgum trees were felled, even though sweetgum composed 30% of the trees. Trees were felled from the 40-105 cm diameter classes with greatest losses (4 trees) occurring in the 40 and 65 diameter classes ( Figure 3 ). A total of 22 white oak trees were felled from the 40-110 cm diameter classes with greatest losses occurring in the 65-85 cm diameter classes ( Figure 4 ). Fifty-four cherrybark oak trees were felled from the 35-105 cm diameter classes. Greatest losses (at least 4 trees) occurred from the 45-95 cm diameter classes ( Figure 5 ). In several places the post-disturbance diameter dropped below zero. This was due to the previously mentioned sources of error between the two inventories. Table 2. Post-disturbance density, relative density, basal area, relative basal area, and importance values by species or species groups for the Moro Bottoms RNA. (Density = stems/ha; relative =% basal area = sq.m./ha) Species or Density Relative Basal Relative Importance species group density area basal area value ------------- ------- -------- ------- ---------- ---------- Sweetgum 1,014 31.21 270.99 35.56 67 White oaks 466 14.34 144.49 18.96 33 Cherrybark oak 315 9.70 65.01 8.53 18 Other red oaks 390 12.00 80.07 10.51 23 Hickories 424 13.05 61.98 8.13 21 Baldcypress 251 7.73 61.05 8.01 16 Ashes 48 1.48 8.04 1.05 3 Other hardwoods 238 7.33 44.11 5.78 13 Loblolly pine 103 3.17 26.43 3.47 7 Total 3,249 100.01 762.17 100.00 201 In the other red oak species group, 66 trees or 14.5% of the original number of other red oaks were lost during the wind event. Losses occurred from the 35-110 cm diameter classes with greatest losses (at least 4 trees) occurring in the 50-90 cm diameter classes, similar to cherrybark oak ( Figure 6). In addition to the losses among cherrybark oak and the other red oaks, 42 hickory trees were lost during the severe wind. Losses occurred from the 35-100 cm diameter classes with greatest losses (at least 6 trees) occurring in the 40, 55, 60, and 65 cm diameter classes ( Figure 7 ). These values were similar to cherrybark oak and the other red oaks. As for the remaining species and species groups, few trees were felled during the severe wind. Only one 55-cm baldcypress was felled ( Figure 8 ). No ash trees or other hardwood trees >25 cm d.b.h. were found blown down or toppled by blown over trees ( Figure 9 s 9 & Figure 10, respectively). Finally, no loblolly pine trees were found on the ground ( Figure 11 ). DISCUSSION One might ask why more trees were lost in the intermediate diameter classes compared to the smaller and larger diameter classes. Cherrybark oak is generally considered to be the most desirable tree species in southern United States bottomland hardwoods (Burns and Honkala, 1990). It has excellent growth and form, making it a highly desirable timber species. This desirability results from the sturdy nature of the lumber produced from red oak in general, and particularly from cherrybark oak. But this sturdiness may have ecological ramifications. We hypothesize that cherrybark oak is not as flexible a species when subjected to severe wind. That is, because the bole is stiffer, the crown of cherrybark oak buffets about less than that of sweetgum or the white oaks. Instead, the crowns of cherrybark oak resist wind by depending more on a sturdy bole. The disadvantage of this strategy is that, in cases of severe wind, cherrybark oaks of intermediate diameter may not be as resistant to windthrow. Since this species apparently does not sway as much as other species, an increased incidence of blowdown may occur. In addition to there being fewer larger trees, lower windthrow of cherrybark oaks >80 cm may be due to a higher ratio of diameter to crown surface area -- that is, a given crown area supported by a much thicker bole. A cursory examination of the crowns of cherrybark oak suggests little difference in crown size between intermediate and codominant trees. The larger stem diameter of the codominant trees may make them more resistant to blowdown than smaller associates that have similar crown size. These larger trees may also have had more expansive root systems and thus greater anchorage in the soil, but this is speculative. Similar arguments used to explain the loss of trees in the intermediate diameter classes in cherrybark oak may help explain losses in the other red oak and hickory species groups. A possible ecological advantage of this sturdiness is, that during near calm to mild wind events, cherrybark oak crowns are better able to compete against other species during episodes of crown abrasion. This has important stand development consequences in mixed-species hardwood stands (see Oliver, 1976; Clatterbuck & Hodges, 1986; and Kittredge, 1988). CONCLUSIONS An important point relevant to the objectives of this meeting involves the fortuitous happenstance of a 100% pre-disturbance inventory on a relatively large stand. Given the increasing number of natural areas being set aside, it is imperative that as complete an inventory as possible be made on these unique areas. While greater detail can be gathered from small permanent plots, additional information can be gained by supplementing permanent plot data with a 100% inventory of at least the overstory trees. Although several errors were present in comparing pre- and post- disturbance inventories, primarily the result of diameter growth of trees during the three years between inventories along with measurement error, these errors were deemed relatively small in comparison to the results obtained from these complete inventories. Finally, results showing a species-specific and a size-specific pattern to windthrow (i.e., cherrybark oak, other red oaks, and hickories of intermediate diameters being more likely to fall during severe wind events) would probably not have been possible with sampling techniques. LITERATURE CITED Burns, R. M. & B. H. Honkala. 1990. Silvics of North America. Volume 2: Hardwoods. USDA Forest Service, Agricultural Handbook 654. 877 p. Clatterbuck, W. K. & J. D. Hodges. 1988. Development of cherrybark oak and sweetgum in mixed, even-aged bottomland stands in central Mississippi, U.S.A. Canadian Journal of Forest Research 18:12-18. Curtis, J. D. 1943. Some observations on wind damage. Journal of Forestry 41:877-882. Duncan, W. H. & M. B. Duncan. 1988. Trees of the Southeastern United States. Athens, GA: The University of Georgia Press. 322 p. Kittredge, D. B. 1988. The influence of species composition on the growth of individual red oaks in mixed stands in southern New England. Canadian Journal of Forest Research 18:1550-1555. Oliver, C. D. 1981. Forest development in North America following major disturbances. Forest Ecology and Management 3:153-168. Oliver, C. D. 1978. Development of northern red oak in mixed species stands in central New England. Yale School of Forestry and Environmental Studies Bulletin No. 91. 63 p. Oliver, C. D. & E. P. Stephens. 1977. Reconstruction of a mixed-species forest in central New England. Ecology 58:562-572. Spurr, S. S. and B. V. Barnes. 1980. Forest ecology. New York: John Wiley & Sons. 687 p. Stephens, E. P. 1956. The uprooting of trees: a forest process. Proceedings of the Soil Science Society of America 20:113-116. Turner, L. M. 1935. Catastrophes and pure stands of southern shortleaf pine. Ecology 16:213-215. Webb, S. L. 1989. Contrasting windstorm consequences in two forests, Itasca State Park, Minnesota. Ecology 70:1167-1180. White, P. S. 1979. Pattern, process, and natural disturbance in vegetation. Botanical Review 45:229-299.