This file was created by scanning the printed publication. Text errors identified by the software have been corrected; however, some errors may remain. LANDSLIDE INITIATION, RUNOUT, AND DEPOSITION WITHIN CLEARCUTS AND OLD-GROWTH FORESTS OF ALASKA A. C. JOHNSON,D. N. SWANSTON,ANDK. E. MCGEE Made m United States of America Reprinted from JOURNALOF THE AMERICAN WATER RESOURCESASSOOATION VOI. 36, No. 1, February 2000 Copyright © 2000 by the American Water Resources Association VOL. 36, NO. 1 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION AMERICANWATERRESOURCES ASSOCIATION LANDSLIDE INITIATION, RUNOUT, CLEARCUTS AND OLD-GROWTH FEBRUARY2000 AND DEPOSITION WITHIN FORESTS OF ALASKA1 A. C. Johnson, D. N. S w a n s t o n , a n d K. E. McGee 2 ABSTRACT: More than 300 landslides and debris flows were triggered by an October 1993 storm on Prince of Wales Island, southeast Alaska. Initiation, runout, and deposition patterns of landslides that occurred within clearcuts, second-growth, and oldgrowth forests were examined. Blowdown and snags, associated with cedar decline and "normal" rates of mortality, were found adjacent to at least 75 percent of all failures regardless of land use. Nearly 50 percent of the landslides within clearcuts occurred within one year following timber harvest; more than 70 percent of these sites had hydrophytic vegetation directly above failures. In following the runout paths of failures, significantly more erosion per unit area occurred within clearcuts than in old-growth forests on slopes with gradients from 9 to 28" (16 to 54 percent). Runout length, controlled by hillslope position within deglaciated valleys, was typically longer in old-growth forests than in second growth and clearcuts (median values were 334, 201, and 153 m, respectively). Most landslides and debris flows deposited in first- and second-order channels before reaching the main stem channels used by anadromous fish. Slide deposits in old-growth forests were composed of a higher proportion of woody debris than deposits derived from slides in second growth or clearcuts. (KEY TERMS: landslides; debris flows; land use planning; erosion and deposition; woody debris; deglaciated valleys; Alaska; anadromous fish.) B e n d a et al., 1998), w i n d ( S w a n s t o n , 1979; H a r r i s , 1989; K r a m e r , 1997; N o w a c k i a n d Kramer, 1997), natural declines in forest species ( J o h n s o n a n d Wilcock, 1997), or by h u m a n factors such as c l e a r c u t t i n g a n d road building (Bishop a n d Stevens, 1964; Wu et al., 1979; Sidle, 1985; S w a n s t o n , 1991). Previous studies conducted in t h e glaciated t e r r a i n of s o u t h e a s t A l a s k a h a v e i n d i c a t e d t h a t t h e r e is a t w o - to f o u r - f o l d i n c r e a s e in l a n d s l i d e f r e q u e n c y a s s o c i a t e d w i t h timb e r h a r v e s t ( B i s h o p a n d S t e v e n s , 1964; S w a n s t o n , 1974, 1979; Sidle 1985; S w a n s t o n , 1991; S w a n s t o n and Marion, 1991) and decline of yellow-cedar (Chamaecyparis nootkatensis) (Johnson, 1993, 1997). In this study, c h a r a c t e r i s t i c s of landslide initiation are p r e s e n t e d , i n c l u d i n g an a s s e s s m e n t of forest h e a l t h a n d h y d r o l o g y as i n d i c a t e d by v e g e t a t i o n . I n addition, t h e v o l u m e of m a t e r i a l eroded or deposited along landslide flow p a t h s is examined. T h e s e factors are a s s e s s e d for l a n d s l i d e s i n i t i a t e d d u r i n g a single s t o r m e v e n t w i t h i n v a r i o u s l a n d use t y p e s i n c l u d i n g old-growth, second-growth, and clearcut forests. F r o m t h e s e a n a l y s e s , t h e u n d e r l y i n g factors t h a t m a y e x a c e r b a t e t h e effects of l a n d use on landslide occurrence a n d r u n o u t are e v a l u a t e d . INTRODUCTION The link between mass-wasting and valuable downstream resources, including salmonid habitats, s u g g e s t s a need to assess both t h e t r i g g e r i n g m e c h a n i s m s a n d r u n o u t p a t h s of t h e s e l a n d s l i d e e v e n t s . L a n d s l i d e s typically occur in conjunction with m a j o r s t o r m s in m o u n t a i n o u s t e r r a i n of t h e Pacific N o r t h w e s t of t h e U n i t e d States. T h e s e e v e n t s are exacerb a t e d by n a t u r a l factors such as fire ( K r a m m e r , 1965; Rice, 1973; S w a n s o n , 1981; B e n d a a n d D u n n e , 1997; STUDY AREA Prince of Wales I s l a n d is a b o u t 40 k i l o m e t e r s w e s t of K e t c h i k a n , A l a s k a a n d a t 5 7 8 , 0 0 0 h e c t a r e s is the l a r g e s t island of the A l e x a n d e r Archipelago. T h e landslides sampled occurred during a single storm on O c t o b e r 26-27, 1993. M o r e t h a n 300 i n d i v i d u a l 1Paper No. 98140 of the Journal of the American Water Resources Association. Discussions are open until October 1, 2000. 2Respectively, Hydrologist, USDA Forest Service, Pacific Northwest Research Station, 2770 Sherwood Lane, Suite 2a, Juneau, Alaska 99801; Consulting Engineering Geologist, P.O. Box 34255, Juneau, Alaska 99803; and Student/Hydrologist, Oregon State University, School of Forestry, College of Forest Engineering, Corvallis, Oregon 97331 (E-Mail/Johnson: ajohnson/rl0@fs.fed.us). JOURNALOF THE AMERICANWATER RESOURCESASSOCIATION 17 JAWRA Johnson, Swanston, and McGee granodiorites, greywackes, conglomerates, limestones, and s a n d s t o n e s • M e t a s e d i m e n t a r y m u d s t o n e s , greywackes, shales, slates, metavolcanic diorites and granodiorite dominate the study area. Most of the bedrock below 1000 m elevation is covered by till landslides occurred primarily in the central and eastern central portions of the island at approximately 55°30"N,133°E (Figure 1). The island is dominated by rocks of the Alexander Terrain, a large accretionary fragment, consisting of -- : J i . i .," •£ 7 /, •" ;! eJ ,,,l.;ral -.& " / • .., ~t~ ~jV 10 20 30 40 I + Kilometers Figure 1. The Two Regions of High Landslide FrequencyAssociated with the October 1993 Storm on Prince of Wales Island. JAWRA 18 JOURNALOF THE AMERICAN WATER RESOURCES ASSOCIATION Landslide Initiation, Runout, and Deposition Within Clearcuts and Old-Growth Forests of Alaska associated with the late Wisconsinan glacial advance (Swanston, 1969). L a n d s l i d e s were i n i t i a t e d in soils d e r i v e d from glacial till classified as Spodosols (87 percent), and regions of exposed rock where soils include Inceptisols (10 percent) and Entisols (3 percent). The spodisols are sandy with an internal angle of friction of 35 ° and an effective cohesion of 8 k P a (Shroeder, 1983). They are classified primarily as typic humicryods, which are moderately-well to well-drained soils, typically greater than 60 cm in depth. Precipitation was recorded in Hollis, central Prince of Wales Island. Precipitation for eight consecutive days preceding the storm event initiating the landslides t o t a l e d a p p r o x i m a t e l y 160 mm. D u r i n g the storm t h a t initiated these landslides, an additional 209 m m of rain fell. W e a t h e r stations at sea level recorded near freezing conditions, so some wet snowfall and associated rapid s n o w m e l t also m a y have occurred. In addition, a 2 percent increase in rainfall for every 30 m rise in elevation near Hollis has been d o c u m e n t e d in p r e v i o u s s t o r m s ( W a l k o t t e n a n d Patric, 1967). Personal accounts indicate local periods of high intensity rainfall and wind. Storm cells were d o c u m e n t e d in s e v e r a l a r e a s n e a r high l a n d s l i d e activity. The closest w e a t h e r station that m e a s u r e d wind speeds was located 128 km NE in Petersburg, Alaska, where wind speeds of 18.4 m/s were reported. The forests on Prince of Wales Island are dominated by stands of western hemlock (Tsuga heterophylla), S i t k a s p r u c e (Picea sitchensis), red c e d a r (Thuja plicata), A l a s k a y e l l o w - c e d a r (Chamaecyparis nootkatensis), and shore pine (Pinus contorta). Yellowc e d a r decline, a n a t u r a l m o r t a l i t y of o l d - g r o w t h forests exists in some of the areas. The decline has b e e n a s s o c i a t e d w i t h c h a n g e in soil t e m p e r a t u r e , chemistry, and/or change in hydrology as opposed to an association with pests or pathogens (Hennon et al., 1990; H e n n o n a n d S h a w , 1997). U n d e r s t o r y species include red h u c k l e b e r r y (Vaccinium parvifloium), A l a s k a b l u e b e r r y (Vaccinium alaskaense), r u s t y menziesia (Menziesia ferruginea), and devil's club (Opopanax horridus). Skunk cabbage (Lysichitom americanus), a n d f a l s e h e l b o r e (Veratrum eschscholtzii) grow in the wettest areas. Windthrow is the dominant forest disturbance (Harris et al., 1974; Harris, 1989; Nowacki and Kramer, 1997). Clearcutring of the old-growth forests has occurred since the 1950s r e s u l t i n g in a m o s a i c of c l e a r c u t s , second growth, and old growth patches. In this study, forests are i d e n t i f i e d as old g r o w t h if no h a r v e s t i n g h a s occurred, second growth if harvested before 1985, and clearcut if harvested after 1985. JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION METHODS A total of 45 landslides were labeled and randomly s e l e c t e d u s i n g a r a n d o m n u m b e r g e n e r a t o r from inventories conducted by the Ketchikan area of the Tongass National Forest (Johnson, 1993; Landwehr, 1993). These inventories were developed through lowelevation (less than 300 m) aerial reconnaissance surveys and road surveys. Fifteen landslides were randomly selected from clearcuts, second-growth, and old-growth forests, respectively. We use landslides as a descriptive term for all rapid soil failures. Debris flows are specified if it was apparent that the initial failure m a s s became liquefied as indicated by flow deposits. Each landslide and debris flow was examined from the point of initiation to the point of deposition to evaluate initiation, runout, and deposition characteristics. Characteristics of the initiation zones included land use, indicators of hydrology, plant disturbance, and terrain configuration controlling slope stability. Initiation zones were defined as the locations where groundwater seepage converged in subsurface depressions or "hollows." Head scarps of failures m a y have migrated upslope following failure, but for the most part, initiation zones were characterized by a clear headwall and rapid downslope channeling into preexisting gullies, which t r a n s p o r t e d resulting debris downslope. I n d i c a t o r v e g e t a t i o n conditions at t h e immediate vicinity of landslide initiation included the presence of blowdown, decayed stumps, standing dead trees, cedar decline, split trees, leaning or "pistol butted" trees (tree trunks bending in downhill direction at ground level), logging disturbance, and presence of hydrophytic vegetation such as skunk cabbage. Indicator terrain conditions analyzed included elevation at the landslide initiation zone, distance to the ridgeline, slope, mean soil depth of failure mass within the failure zone, width of failed soil mass, and eroded volume. The r u n o u t of l a n d s l i d e s and d e b r i s flows w a s defined as the affected region between the initiation zone and the final deposition zone. Analysis of runout included a d e s c r i p t i o n of the c h a n g e s in l a n d u s e encountered along the landslide or debris flow path, measurements of runout widths and lengths, and erosional and depositional trends. These t r e n d s were described for r u n o u t segments, divided and sorted according to slope gradient category. Segment length ranged from 3 to 100 m, averaging 28 m with 4 to 21 segments per landslide. The fraction of total runout length occurring per slope gradient category is presented in conjunction with the range of volumes of erosion or deposition per unit runout area. Analysis of variance and t-tests were used to statistically analyze 19 JAWRA Johnson, Swanston, and McGee differences in erosion or deposition per slope gradient category. Differences were judged to be statistically significant at P = 90 percent (p = 0.10). Final, or terminal deposits at the end of the landslide r u n o u t tracks were characterized by depositional slope, composition, volume, and location. For deposits in streams, order and stream type were determined. Stream type was defined as Class I if the stream supported anadromous fish populations, Class II if supp o r t i n g r e s i d e n t fish such as c u t t h r o a t t r o u t , and Class III and IV if without fish populations (USDA, 1997). Size and number of logs were counted at some of the landslide deposits. A comparison of woody debris to soil volume ratios for clearcuts and old-growth forests was e s t i m a t e d using the following equation: forests m a y have occurred. As a result, failure rates and volumes per area per landuse were not calculated. Initiation of landslides was closely associated with landforms. At least 40 out of 45 (89 percent) of the failures initiated in bedrock hollows (bedrock depressions where colluvium, woody debris and water collect), a phenomenon common in the Pacific Northwest (Dietrich and Dunne, 1978). The remainder occurred on planar slopes. Till was the primary failure material in at least 75 percent of the sites. Slope and soil depth at landslide initiation sites were similar for all land use types (Table 1). Initiation aspects differed. Nine out of 15 of landslides in old growth trended to the northeast and southeast, 11 out of 15 landslides in clearcuts occurred on west and s o u t h w e s t aspects, a n d 10 out of 15 of t h e l a n d s l i d e s w i t h i n second growth occurred on south-facing slopes. Springs a n d seepage were found at m o s t of the l a n d s l i d e scarps at the soil/till (bedrock) contacts (Table 2). Approximately 90 percent of the sites had two or three seeps t h a t converged at the base of the bedrock/till hollow. Hydrophytic vegetation was more common at clearcut sites (7 out of 15) t h a n in oldgrowth (3 out of 15) or second-growth (2 out of 15) sites. Five of the seven sites within clearcuts having h y d r o p h y t i c vegetation were h a r v e s t e d less t h a n a year before the landslides occurred. Root deterioration was apparent at 10 but of 15 sites within each of the land use types. Residual logging disturbance was apparent in 7 out of 15 second-growth sites and in 5 out of 15 sites in clearcuts, typically in the form of corridors where logs were dragged along and across the failure scarp. In addition, at several of the sites above the headwall, rotten logs directed water to the hollow where failure occurred. A (Tr) Vtr : A (d) where, A = area of landslide; Tr = average number of trees per area; d = average depth of soil; and Vtr = average tree volume. From this ratio, comparisons were made between old-growth and clearcut forests. Tree densities used in this calculation were 255 trees per ha with an average volume of 2.2 m 3 per tree, values typical for a productive, well drained forest in southeast Alaska (Rogers and van Hees, 1992). Within clearcuts, stump volume was assumed to be 1/20 the size of an entire tree, or 0.11 m 3. Woody debris on the forest floor was not included in this calculation because it varied considerably across sites. RESULTS Initiation Runout The areas of initiation (head scars) for 40 of the 45 l a n d s l i d e s and debris flows in t h e r a n d o m sample were greater t h a n 200 m 2 (volumes greater t h a n 77 m 3) except one slope failure within clearcut forests, and two failures w i t h i n old-growth and second-growth forests. This has been the m i n i m u m size consistently observed under the old-growth forest canopy of southe a s t A l a s k a at the p h o t o g r a p h i c scale of 1:12,000 ( S w a n s t o n a n d M a r i o n , 1991). O t h e r s t u d i e s in coastal regions of Oregon (Oregon D e p a r t m e n t of Forestry, 1998) have found t h a t the m i n i m u m observable size of landslide i n i t i a t i o n zones u n d e r forest canopy was 418 m 2 (roughly 15 by 27 m) using 1:6,000 and 1:12,000 scale aerial photographs. Although we believe our sample is unbiased, we acknowledge t h a t some bias toward larger landslides within old-growth JAWRA At least one-third of the sampled landslides traveled from one land use type into a n o t h e r land use type, and were t h e r e b y affected by road crossings, change in woody debris loading, and change in the size of standing timber (Figure 2). In lieu of this complexity, details of r u n o u t were described solely for the landslides t h a t remained consistently within the forest type in which they initiated (nine within each landuse type). The m a j o r i t y of f a i l u r e s (5 out of 9) w i t h i n old growth split j u s t below the zone of initiation. These landslides split when standing trees (greater than 0.3 m in diameter) impeded the flow path of the failure mass and forced it to shift into several channels. Two 20 JOURNAL OF THE AMERICAN WATER RESOURCESASSOCIATION Landslide Initiation, Runout, and Deposition Within Clearcuts and Old-Growth Forests of Alaska TABLE 1. Summary of Landslide Initiation Site Characteristics (n = 15 for each land use type). L a n d u s e Type Mean Value Maximum Standard Deviation Minimum OLD GROWTH Soil Depth (m) Width (m) Slope (degrees) Elevation (m) Distance to Ridgeline (m) Eroded Volume (m3) 1.2 19 38 370 234 382 2.0 32 55 608 701 855 0.5 10 31 172 50 25 0.5 6 7 115 198 274 SECOND Growth Soil Depth (m) Width (m) Slope (degrees) Elevation (m) Distance to Ridgeline (m) Eroded Volume (m3) 1.1 17 34 256 282 633 2.2 49 45 396 1037 3952 0.4 7 27 183 46 20 0.7 13 5 85 296 1037 CLEARCUTS Soil Depth (m) Width (m) Slope (degrees) Elevation (m) . Distance to Ridgeline (m) Eroded Volume (m3) 1.2 13 36 391 165 347 2.0 35 47 500 671 1332 0.6 5 25 195 61 104 0.4 9 6 83 157 322 TABLE 2. Conditions at Sites of Landslide Initiation (n = 15 per condition per land use type). Site C o n d i t i o n Clearcuts Second Growth Old G r o w t h 3 6 1 2 9 3 1 5 15 7 3 6 0 0 11 0 0 7 13 2 8 7 3 3 9 2 4 0 13 3 Blowdown Pistol Butted Trees Split Trees Leaning Trees Decayed Stumps and Decayed Snags Cedar Decline (CD) Standing Dead (other than CD) Yarding Disturbance Groundwater Seepage Hydrophytic Vegetation o f t h o s e five l a n d s l i d e s t u r n e d i n t o c h a n n e l i z e d d e b r i s flows and followed pre-existing gullies. Another two landslides flowed out onto planar slopes and upon splitting, appeared to undercut the adjacent slopes causing additional slides. One landslide split initially a n d t h e n r e t u r n e d to t h e o t h e r flow p a t h l e a v i n g a n i s l a n d of t r e e s . O n l y o n e o f t h e f i v e l a n d s l i d e s i n v e n t o r i e d h a d a m a j o r i t y o f d e p o s i t i o n in a s t e e p c h a n n e l . A t t h i s s i t e , t h e f l o w p a t h c r o s s e d a s t e e p (21 °, 37 p e r JOURNALOF THE AMERICANWATER RESOURCESASSOCIATION c e n t ) c h a n n e l w i t h a j u n c t i o n a n g l e o f 70 °, a n d t h e majority of the flow stopped in the channel. Of the remaining four landslides that did not split, two flowed into channels and two remained wide and unchannelized. Unlike landslides in old-growth forests, those in second growth and clearcuts did not split. Four out of nine landslides and debris flows in second growth f l o w e d i n t o s t e e p c h a n n e l s (> 20 °, 36 p e r c e n t ) , t h e r e s t 21 JAWRA Johnson, Swanston, and McGee Old arowlh Initiation Runout Second growth Deposition # slides Initiation ~ h Runout Deposition # slides 4 Sgroe~w~ltc['JS°econgnrdodwth51 5 Clearcuts 1 Initiation Runout Deposition # slides 1 4 1 0ro 1 ' 4 Figure 2. The Initiation and Runout of Landslides on Prince of Wales Island with Changes in Land Use Indicated. Note t h a t for at least one-third of the landslides, runout passed into a different land use. were wide and unchannelized. All of the landslides and debris flows in clearcuts flowed into steep firstorder t h r o u g h third- order c h a n n e l s (>20", 36 percent). The longest r u n o u t segments occurred within oldgrowth forests followed by those in clearcuts and second growth. Five out of nine of the failures in old growth had large widths (> 20 m) for the first 75 m w h e r e a s o n l y two l a r g e - w i d t h l a n d s l i d e w i t h i n clearcuts traveled 25 m before decreasing in width ( F i g u r e 3d). T h e m a j o r i t y of l a n d s l i d e s w i t h i n clearcuts h a d r u n o u t s t h a t i n i t i a l l y were m e d i u m widths (10-19 m) and did not travel distances exceeding 400 m (Figure 3c). Small-width landslides within second growth, clearcuts, and old growth traveled < 200, < 300, and < 600 m, respectively (Figure 3b). Two failures in second growth started with large widths and an a d d i t i o n a l two l a n d s l i d e s became large in width. Five r u n o u t segments from these four landslides had runout distances exceeding 101 m (Figure 3d). Deposition within the r u n o u t of debris flows and landslides occurred on slopes as steep as 38" (78 percent) in old growth and second growth (Figure 4a). Deposition occurred as berms along the flow path and piles of soil and woody debris t h a t had been backedup behind standing trees. For old-growth forests, the amount of erosion was nearly equal to the amount of JAWRA deposition on slopes from 19 to 28" (34 to 54 percent). These slopes accounted for thirty-three percent of the total runout distance occurring in old-growth forests (from Figure 4d). Although landslides and debris flows in old growth were generally larger in size t h a n landslides in second growth and clearcuts, there was more erosion per a r e a in clearcuts t h a n in old growth. Erosion was dominant, as indicated by median, upper and lower quartile values, on slope gradients greater t h a n 29" (56 percent) in old growth (Figure 4a) and second growth (Figure 4c), and greater t h a n 19" in clearcuts (Figure 4b). These slope gradients accounted for 38 percent, 29 percent, and 55 percent of the total runout l e n g t h s , r e s p e c t i v e l y (Figure 4d). In o t h e r words, there was a tendency for more erosion to occur along the majority of the runout lengths within clearcuts. Slope gradients from 19 to 23" (34 to 42 percent) had significantly more erosion per unit area in clearcuts t h a n in old growth (t-tests; p = 0.02). On these slope gradients, mean erosion was 0.28 m3/m 2 (volume per area) in clearcuts (n = 12) while in old-growth (n = 16), deposition was dominant with a mean volume of 0.16 m3/m 2. A significant difference was also found on 9 to 28" (16 to 54 percent) slope gradients (p = 0.08). Here, 0.03 m3/m 2 was eroded in clearcuts (n = 50), and in old growth (n = 57), 0.15 m3/m 2 was deposited. 22 JOURN OAFTLHA EMERC IW ANATERRESOURC AE SSOCA ITO IN Landslide Initiation, Runout, and Deposition Within Clearcuts and Old-Growth Forests of Alaska a) All slides together b) Small width landslides (< 10 m) c) Medium width landslides d) Large width landslides (10-19 m) (>20 m) m l m m m m m m mi J m m m m m m | m im m m m m | 2 6 - 50 51 -75 A 76-100 8 101-200 g LI I I I n l l , l l l l l l l l ,l l l T inl|l II r ...... j ....... ~ . LIJUMUMU , I t ....... . - I J .... T. . . . . . . nmn,mn i||l|lllrlllllll|ll ! p - - 201 -300 + ~ 3Ol- 4oo 401 - 500 m r a m BIB m m 501-600 1 ; [ ] Second growth I 701 -800 801 -900 0 6 9 0 2 4 6 Cumulative Frequency "0 2 4 6 2 Frequency Figure 3. Landslide Runout Width and Length for: (a) All Landslides Together, (b) Small Width Landslides, (c) Medium Width Landslides, and (d) Large Width Landslides. Deposition deposited in muskeg ponds (wetland bogs or fens) t h a t are classified as Class II due to use by resident fish. The remainder of the debris flows deposited in Class III and Class IV channels without fish. In contrast, all of t h e l a n d s l i d e s in c l e a r c u t a n d s e c o n d - g r o w t h forests deposited in Class III and Class IV channels, e x c e p t for one l a n d s l i d e in s e c o n d g r o w t h t h a t deposited in a Class II stream (Table 4). The relative proportion of woody debris to sediment in the landslide deposits was greater in old growth t h a n clearcuts. Deposits in clearcuts consisted of sediment, woody debris fragments and stumps with rootwads, as compared to slides in old growth which were composed of sediment, entire trees and woody debris fragments. Slide deposits in old growth often had a snout of woody debris on the downward side. Deposits in second growth deposits were similar to those in clearcuts if the second growth was relatively young, and more similar in composition to old growth with i n c r e a s i n g age. Deposits in old-growth forests h a d from 98 to 300 logs (with diameters from 0.2 to 2.0 m). When a large snout of wood existed within a single deposit, smaller woody debris and sediment was often backed-up behind it. Lacking a large woody debris snout, the unimpeded sediment freely moved greater distances in the downstream or downslope direction. The estimated percentage of woody debris derived from 0.25, 0.5, and 1.0 ha landslides and debris flows Final, or terminal deposits were located at the end of the runout tracks. Maximum and mean deposit volu m e s were l a r g e s t in o l d - g r o w t h f o r e s t s a n d t h e smallest in clearcuts (Table 3). The slope upon which the final deposits were located ranged from 2.5 to 19" (5 to 34 percent). Depositional processes appeared to be enhanced by large standing trees (> 0.3 m diameter), a lack of channelization, and r u n o u t onto road surfaces on slopes less t h a n 15" (27 percent). Four of the nine landslides and debris flows t h a t t r a v e l e d through second growth and clearcuts ended at or near roads (Figure 2). T w e n t y out of 27 t e r m i n a l deposits occurred in first-order and second-order stream channels on 4 to 19' slopes (7 to 34 percent) with a mean of 10" (16 percent) (Table 4). Two sites (in a clearcut and second growth) did not deposit in channels. The remaining five slides deposited in third- and fourth-order channels on slopes ranging from 2.5 to 9" (5 to 16 percent) slopes with a mean slope of 6 ° (10 percent). Two landslides/debris flows in old growth deposited in a fourthorder, Class I, or anadromous fish bearing channel. Only one of these failures had a major impact on the channel by introducing over 100 m 3 of sediment and woody debris, while the other slide deposited approximately 10 m 3 of sediment and wood. Two debris flows JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 23 JAWRA Johnson, Swanston, and McGee Old-growth f o r e s t Slope (degrees) <4 A ~ 4rob 9to13 14to 18 19 to 23 24 to 28 29 to 33 34 to 38 39 to 43 >44 2 E g - - o : -1 ._~ II -2 UJ a) -4 • m~ximum -- upper quartile • median -- lowerquartille • minimum Clearcuts 3 ~' ..---.,~ I 2 E E v ~ o -1 ..~ -2 W b) -4 Second growth 2 % 1 o 8 ==. -1 -2 "~ c) -3 -4 0.3 J= a 0.25 0.2 ~0.15. lS 0.1 . u. 0.05 o d) <4 4to8 9to13 14to 18 19 to 23 24 to 28 Slope (degroes) 29 to 33 34 to 38 39to43 >44 Figure 4. Distributions (box and whisker plots) of Erosion and Deposition in: (a) Old Growth, (b) Clearcuts, and (c) Second Growth by Slope Category, and (d) Fraction of Total Runout Length per Slope Category (deposition is indicated by positive volumes and erosion by negative volumes along the Y-axis). Note more erosion in clearcuts. JAWRA 24 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATtON Landslide Initiation, Runout, and Deposition Within Clearcuts and Old-Growth Forests of Alaska TABLE 3.Characteristics of Landslide Terminal Deposits (n = 9 for each land use type). M e a n Value Maximum Minimum Standard Deviation OLD GROWTH Slope (degrees) Width (m) Volume (m 3) 10 26 756 19 72 2649 2.5 5 10 5 24 958 SECOND GROWTH Slope (degrees) Width (m) Volume (m 3) 13 21 514 17 53 2268 4 7 100 10 14 692 CLEARCUTS Slope (degrees) Width (m) Volume (m 3) 7 11 187 12 20 724 3 5 13 3 5 197 Landuse Type TABLE 4. Number of Landslide Terminal Deposits Per Stream Order and Stream Class [Class I, anadromous fish; Class II, resident fish; Class III and IV, without resident fish populations (U.S.D.A., 1997)] with Average Slope Indicated in Parentheses (total n = 27). Landuse First Order C l a s s IV Old Growth 2 (14", 25%) Second Growth Clearcuts Totals Second Order C l a s s III Third Order C l a s s II Fourth Order Class I 3 (8", 12%) 2 (7 °, 12%) 2 (2.5",5%) 5 (15", 27%) 2 (6 °, 10%) 1 (7 °, 12%) 0 4 (12", 21%) 4 (10", 18%) 0 0 9 3 2 11 with average soil depths of 0.1, 0.2 and 0.5 m indicates t h a t the proportion of woody debris in a deposit was greatest when landslides derived from clearcuts and old growth were large and soils were shallow (Figure 5). Landslide deposits in old-growth forest typically h a d ten times the woody debris volume of deposits originating in clearcuts. t h a n clearcut or old-growth sites, due mainly to earlier removal of easily reachable, h i g h value timber. These land use pattern differences not only relate to d i f f e r e n c e s in f o r e s t h y d r o l o g y a n d f o r e s t h e a l t h which m a y be associated with the initiation of landslides, but also affect runout length. Although it is diff i c u l t to d i r e c t l y c o m p a r e l a n d s l i d e s w i t h i n t h e different landuse types because of differences in hillslope location, some trends are apparent. DISCUSSION Initiation While our selection of landslides was random, the pattern of land use distribution was not. In southeast Alaska, clearcuts are often located in mid-slope and upper-slope regions of wide, U-shaped glacial valleys whereas areas of remaining old growth typically occur in steep V-shaped valleys above clearcuts or in steep canyons. In addition, old growth is common on valley side-slopes due to economic and regulatory reasons. Second-growth sites are generally lower in elevation JOURNALOF THE AMERICANWATER RESOURCESASSOCIATION Seven of 15 landslides occurred within one year of forest harvest on clearcut sites, in contrast to the general 2-5 y e a r d e l a y f o u n d by o t h e r s ( Z e i m e r a n d Swanston,1977; Zeimer, 1981; O'Loughlin and Zeimer, 1982). This difference could be due to the magnitude of the storm and site conditions, including the condition of roots and preponderance of site s a t u r a t i o n . L a n d s l i d e s in c l e a r c u t s m a y h a v e been i n i t i a t e d 25 JA WR A Johnson, Swanston, and McGee 4O • Clearcuts °m o Q. 30 :i!iiii!il 25 iii~iiiiil 15 iiii?!iiil 10 ii'?:?:?:: [] Oldgrowth :!~!:!:! o o 5 ii~i!!iiil ii;i~i~i~i 0 i!i!iiiiil : , n . . . . , . . N Failure size (ha) (soil depth (m)) Figure 5. Estimated Proportion of Woody Debris in Clearcut and Old Growth Landslide Deposits. Calculations are plotted by size of landslide scar and by depth of erosion (in parenthesis). Woody debris diameters are greater than 23 cm. regardless of storm and harvest timing due to preharvest hydrologic and root conditions predisposing some sites to failure. Five out of seven landslide scarps within clearcuts having failures within one year of forest h a r v e s t h a d a p r e p o n d e r a n c e of rotted roots that pre-dated harvest activities and hydrophytic vegetation, specifically skunk cabbage. Skunk cabbage is indicative of a high water table (Minore, 1969) and is also w e l l - c o r r e l a t e d with t o p o g r a p h i c convergence (area where ground water concentrates). Topographic models such as TOPOG (O'Loughlin, 1986) and T O P M O D E L (Beven and Kirkby, 1979) quantify the degree to which ground saturation due to rising ground w a t e r is possible. These models have been linked to s t a b i l i t y models ( M o n t g o m e r y and Dietrich, 1994), but are dependent on detailed topographic scale. Lacking the appropriate topographic data, we suggest that the prevalence of wetland vegetation in conjunction with steep slopes, convergent topography, and bedrock hollows, are indicative of high landslide potential. These indicators of landslide hazard, when combined with a decrease in overstory canopy, m a y markedly alter the magnitude and timing of peak pore pressures in convergent topography following major rainstorms. For example, in southeast Alaska, as much as 22 percent of the rainfall (where rainfall for one storm was 203 mm) is intercepted by hemlock and spruce forests (Patric, 1966). Loss of forest not only results in a decrease in root strength, but also increases the a m o u n t of precipitation directly reaching the ground, decreases evapotranspiration, and m a y decrease the time it takes for the groundwater table to rise to the soil surface. JAWRA A reduction of root strength as well as a lack of overstory vegetation m a y have played a key role in triggering some landslides as blowdown, rotted roots, or cedar decline occurred on 93 percent of old-growth locations, 73 percent of clearcut sites, and 80 percent of second-growth forests. Within clearcuts, skunk cabbage was present on a minority of the landslide sites t h a t occurred 2-6 y e a r s post h a r v e s t (13 percent). Skunk cabbage was found in 20 percent of old-growth sites and in 13 percent of second-growth sites. Factors besides topographic control on hydrology, perhaps including y a r d i n g d i s t u r b a n c e and changes in root strength, m a y have triggered these landslides. If all other variables are similar, roots have a greater relative role in the stability of hillslopes when soil depths are shallow (Johnson and Wilcock, 1997). Runout and Deposition Wide valleys and trellis drainage systems of Prince of Wales Island, characteristic of glaciated systems, c o n t r o l l e d p a t t e r n s of r u n o u t and d e p o s i t i o n . In this recently deglaciated landscape, channel development is relatively young (generally less than 10,000 ybp), compared to older, nonglaciated landscapes in Oregon and California where drainage systems are dendritic. A comparison of m e a s u r e d and predicted runout lengths with corresponding change in elevation for both of these landscapes (Benda, personal c o m m u n i c a t i o n ; M o n t g o m e r y and Dietrich, 1994) indicates t h a t in nonglaciated systems the average 26 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION Landslide Initiation, Runout, and Deposition Within Clearcuts and Old-Growth Forests of Alaska runout slope is approximately 14 ° (25 percent) whereas it is approximately 23 ° (42 percent) in glaciated systems. Linear relationships for change in elevation a n d r u n o u t l e n g t h i n d i c a t e d i f f e r e n c e s in r u n o u t trends for landslide occurring within different geomorphic regions (Figure 6). C h a n n e l g r a d i e n t s at landslide deposits r a n g e d from 3 to 19 ° (5 to 34 percent); a higher range than t h a t found in Oregon, California, and J a p a n where depositional slopes range from 1 to 10 ° (2 to 18 percent) (Ikeya, 1981; Benda and Cundy, 1989). Benda and C u n d y (1989) found t h a t deposition typically occurs on 8 ° (14 percent) slope g r a d i e n t s in thirdorder channels and 1 ° slope gradients in fifth-order channels. In Alaska, deposition occurred on slopes averaging 10 ° (18 percent) in first-order and secondorder channels. Differences in deposition location m a y be due to the lack of channel confinement in Alaska as compared to other regions. The rapid decrease in low-order stream gradients found in Alaska as compared a gradual decrease in channel gradients elsew h e r e , m a y a l s o c o n t r i b u t e to h i g h e r s l o p e s of deposition. In both areas, higher depositional slopes m a y be attributed to lower water content (Johnson, 1984), increased woody debris content, higher angle tributary junctions, and/or smaller basin area (Ikeya, 1981; Benda and Cundy, 1990). B e c a u s e t h e location of l a n d s l i d e occurrence is associated with landscape form, the location of any particular land use within a landform affects both occurrence and deposition of landslides. Slides in oldgrowth forest tend to deposit within a range of stream orders whereas slides in second growth and clearcuts tended to deposit in lower order (higher class) channels within the valley floor. Slides in second growth and clearcuts tended to flow from the slopes of wide U-shaped valleys and deposit in valley floors, likely because they lacked the momentum to reach the higher order, lower gradient mainstem channels. All of the l a n d s l i d e s t h a t moved into h i g h e r o r d e r c h a n n e l s occurred in steeper portions of the landscape more dendritic in form. The difference in erosion and deposition processes within runouts was largely due to the effect of large standing trees in old-growth and second-growth forests that blocked the flow of oncoming debris. In addition, sediment and woody debris were b a c k e d - u p b e h i n d l a r g e fallen logs in o l d - g r o w t h derived deposits. The size and pattern of deposits of fine sediment and large woody debris can impact fisheries habitat ( S w a n s o n and L i e n k a e m p e r , 1978; Bisson et al., 1992). While t h e m a j o r i t y of l a n d s l i d e / d e b r i s flow deposits do not reach higher order, main stem channels (habitat for many anadromous fish populations), the composition of deposits h a s i m p o r t a n t implications for fisheries habitats. First, large woody debris retained temporarily in deposits and released during s u b s e q u e n t storms is an important source of woody d e b r i s r e c r u i t m e n t to s t r e a m c h a n n e l s . Secondly, steep tributaries that are contiguous with main stem channels provide debris and s e d i m e n t r e c r u i t m e n t conduits. Third, m a i n and t r i b u t a r y channels m a y be particularly sensitive to inputs of fine sediment introduced from channels flowing across and through 250 X °°°* x Deglaciated landscape , , . Y = 0.4172x - 5.7526 ~ 200 R 2 = 0.8807 E A = O .," 150 = A A ~ " ..-,t 100 % Xx X'x~ "x'~ C o X x ;,¢-" ..x > e .. • " . . " 50 .y~ X" 0 ×~xr ~" r "'X A Non-glaciated landscapes y = 0.2487x + 14.695 R2 = 0.9398 ^ ! I ! I : ; 100 200 300 400 500 600 700 Observed lengths (m) Figure 6. Change in Elevation Plotted as a Function of Obseiwed Length of Landslides Occurring on Prince of Wales Island (shown as X's), Oregon (average value from Benda, personal communication, shown as diamond), and in California (Montgomery and Dietrich, 1994, shown as triangles). JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 27 JAWRA Johnson, Swanston, and McGee landslide deposits - particularly if the sediment is not backed-up behind woody debris. This lack of back-up is the typical case for l a n d s l i d e s occurring within clearcuts. These i m p l i c a t i o n s s u g g e s t t h a t understanding the recruitment and storage of sediment and large woody debris in time and space at a landscape scale is e s s e n t i a l to i n t e r p r e t i n g the variability of landslide impacts within a watershed context (Benda et al., 1998). forests had five times the woody debris volumes of deposits originating from clearcuts. Lacking large woody debris volumes, fine sediment from clearcuts tended to migrate further down tributary channels to t h e receiving main stem channels, possibly to the detriment of anadromous fish habitats. Before an understanding of the effects of land mana g e m e n t on landslide initiation and r u n o u t can be developed, the underlying routing of sediment and large woody debris need to be understood in a landscape context. In the glaciated systems of Prince of Wales Island, most deposits from landslides or debris flows do not immediately reach main stem channels t h a t are i m p o r t a n t h a b i t a t s for a n a d r o m o u s fish. Hence, management of this habitat will benefit from an u n d e r s t a n d i n g t h a t d e s i r a b l e levels of w o o d y debris m a y occur as a result of recruitment from natural hillslope failure routed through the few steep r e a c h e s contiguous to the main stern channels. In addition to these inputs, woody debris r e c r u i t m e n t occurs as a result of direct inputs from riparian forest adjacent to the main stem channel and as a result from r e c r u i t m e n t from t r i b u t a r i e s . I n p u t m e c h a nisms, in addition to landslides, include blowdown, bank erosion, and inputs due to forest mortality; natural processes largely controlled by topography and c l i m a t e . Finally, m a n a g e m e n t p l a n s m a y b e n e f i t through the acknowledgement and integration of historical correlations between these naturally occurring disturbance processes and the habitats they create. CONCLUSIONS E x a m i n a t i o n of a r a n d o m s a m p l e of l a n d s l i d e s within a deglaciated landscape in southeast Alaska indicated t h a t n a t u r a l site conditions, e v i d e n c e of past blowdown, and hillslope position largely determined locations of landslide initiation and position of final deposition. We found that 89 percent of initiation sites had rotted roots, convergent topography, evidence of blowdown, and wetland vegetation which predated any human-related disturbances. Runout of landslides appeared to be affected by drainage pattern. A trellis form, typical of glaciated t e r r a i n in southeast Alaska, tended to induce deposition in firstor second-order channels in wide U-shaped valleys with slopes averaging 9" (18 percent), leaving main stem channels unaffected. Less typically, debris flows crossed mainstem channels at 90" angles. In contrast, the uppermost regions of glaciated systems, as well as t h e d e n d r i t i c d r a i n a g e s y s t e m s c h a r a c t e r i s t i c of nonglaciated systems, typically had longer runouts and a lower average depositional slope. This feature was due to the existence of acute angles at s t r e a m junctions and gradual slope change encountered as the debris flows moved into higher order channels from lower order channels. When effects of clearcutting were superimposed on n a t u r a l site conditions, several factors appeared to contribute to the enhancement or suppression of landslide and debris flow impacts. These factors included an increase in triggering mechanisms associated with loss of forest canopy and loss of soil strength due to root cohesion. In addition, there was a significant difference in erosion and deposition characteristics of runout due to loss of the roughness effects provided by s t a n d i n g trees. For example, on slopes ranging from 19 to 23" (34 to 42 percent), erosion (with a mean of 0.28 m3/m 2) w a s d o m i n a n t in r u n o u t p a t h s of debris flows in clearcuts while deposition (mean of 0.16 m3/m 2) was dominant in old growth. Different patterns of erosion and deposition may have affected c o m p o s i t i o n of f i n a l d e p o s i t s in old g r o w t h a n d clearcut stands. To illustrate, we estimated that the composition of deposits originating from old-growth JAWRA ACKNOWLEDGMENTS Funding for this study was provided by the U.S.D.A., Forest Service, Ketchikan Area of the Tongass National Forest and PNW Research Station, Juneau, Alaska. Soil scientists Roger Johnson and Dennis Landwehr produced the landslide inventories from which our sample was derived, and aided immensely in the field work. We also thank Richard Woodsmith, Rachel Baker, Dave Mongomery, Terry Shaw, and three anonymous reviewers for their extremely useful editing and critiques of the manuscript. LITERATURE CITED Benda, L. E. and T. W, Cundy, 1990. Predicting Deposition of Debris Flows in Mountain Channels. Can. Geotech. J. 27:409-417. Benda, L. E. and T. Dunne, 1997. Stoicastic Forcing of Sediment Supply to Channel Networks from Landsliding and Debris Flows. WRR 33(12):2849-2863. Benda, L. E., D. J. Miller, T. Durme, G. H. Reeves, and J. K. Agee, 1998. Landscape Dynamics. In: Ecology and Management of Streams in the Pacific Northwest Ecoregion, R. Naiman and R. Bilgy (Editors). Springer-Ver|eg, pp. 261-288. Beven, K. and M. J. Kirkby, 1979. A Physically Based, Variable Contributing Area Model of Basin Hydrology. Hydrel. Sci. Bull. 24:43-67. 28 JOURNAL OF THE AMERICANWATER RESOURCESASSOCIATION Landslide Initiation, Runout, and Deposition Within Clearcuts and Old-Growth Forests of Alaska Nowacki, G. J. and M. G. Krammer, 1997. The Effects of Wind Disturbance on Temperate Rain Forest Structure and Dynamics of Southeast Alaska. In: Conservation and Resource Assessments for the Tongass Land Management Plan Revision, U.S.D.A., PNW Research Station, Gen Tech. Report 421. O'Loughlin, E. M., 1986. Prediction of Surface Saturation Zones in Natural Catchments by Topographic Analysis. Water Resour. Res. 22(5):794-804. O'Loughlin, C. L. and R. R. Zeimer, 1982. The Importance of Root Strength and Deterioration Rates Upon Edaphic Stability in Steepland Forests. In: Carbon Uptake and Allocation in Subalpine Ecosystems as a Key to Management. Proceeding of an I.U.F.R.O. Workshop, August 2-3, 1982, pp. 70-78. Oregon Dept. of Forestry, 1998. 1996 Storm Impacts Monitoring Project - Preliminary Report. Patric, J. H., 1966. Rainfall Interception by Mature Coniferous Forests of Southeast Alaska. Journal of Soil and Water Conservation. Rice, R. M., 1973. The Hydrology of Chaparral Watersheds. Proc. Symp. Living with Chaparral. Sierra Club, Riverside, California, pp. 27-33. Rogers, G. and W. W. S. van Hees, 1992. Timber Resource Statistics for the Ketchikan Area of the Tongass National Forest, Alaska, U.S.D.A. Forest Service, PNW-RB-184, 18 pp. Schroeder, W. L., 1983. Geotechnical Properties of Southeast Alaskan Forest Soils. Oregon State University Civil Engineering Department, U.S.D.A. Forest Service, 46 pp. Sidle, R. C, 1985. Factors Influencing the Stability of Slopes. Proceedings of a Workshop on Slope Stability: Problems and Solutions in Forest Management. U.S.D.A. Forest Service, Gen. Tech. Rep. PNW 180, pp 17-25. Sidle, R. C. and D. N. Swanston, 1982. Analysis of a Small Debris Slide in Coastal Alaska. Can. Geotech. J. 19:167-174. Swanson, F. J., 1981. Fire and Geomorphic Processes. In: Fire Regimes and Ecosystem Properties. Proceedings of the Conference, H. A. Mooney, T. M. Bonnicksen, N. L.Christensen, J. E. Lotan, and W. A. Reinerset (Editors). USDA For. Serv. Gen. Tech. Rep. WO-26, pp.401-420. Swanson, F. J. and G. W. Lienkaemper, 1978. Physical Consequences of Large Organic Debris in Pacific Northwest Streams. U.S.F.S. Gen. Tech. Rep., PNW-69, PNW Forest and Range Experiment Station, Portland, Oregon. Swanston, D. N., 1969. Mass Wasting in Coastal Alaska. U.S.D.A. Forest Service Research Paper PNW-83, Institute of Northern Forestry, Juneau, Alaska, 15 pp. Swanston, D. N., 1970. Soil-Water Piezometry in a Southeast Alaska Landslide Area. Res. Pap. PNW-103, For. Serv., U.S.D.A., Portland, Oregon, 17 pp. Swanston, D. N., 1974. The Forest Ecosystem in Southeast Alaska, Soil Mass Movement. U.S.D.A., Forest Service Gen. Tech. Bull. PNW-17, 21 pp. Swanston, D. N., 1979. Landslide Prediction and Assessment Interpreting Stability Problems for the Land Manager. Proceedings of Workshop of Scheduling Timber Harvest for Hydrologic Concerns, U.S.D.A. Forest Service, Pacific Northwest Region, PNW Forest and Range Experiment Station. Swanston, D. N., 1991. Natural Processes. In: Influences of Forest and Rangeland Management on Salmonid Fishes and Their Habitats, W. Meehan (Editor). American Fisheries Society Special Pulblication 19:139-179. Swanston, D. N. and D. A. Marion, 1991. Landslide Response to Timber Harvest in Southeast Alaska. In: Proceedings of the Fifth Interagency Sedimentation Conference, Federal Energy Regulatory Commission, Las Vegas, Nevada. Bishop, D. M. and M. E. Stevens, 1964. Landslides in Logged Areas in Southeast Alaska. U.S.D.A. Forest Service Research Paper NOOR-1, Northern Forest Experimental Station, Juneau, Alaska, 18 pp. Bisson, P. A., T. P. Quinn, G. H. Reeves, and S. V. Gregory, 1992. Best Management Practices, Cumulative Effects, and LongTerm Trends in Fish Abundance in Pacific Northwest River Systems. In: Watershed Management, R. J. Naiman (Editor). Springer-Verlag, pp. 127-188. Dietrich, W. E. and T. Dunne, 1978. Sediment Budget for a Small Catchment in Mountainous Terrain. Zietshrift fur Geomorphologie, Suppl. Bd. 29:191-206. Harris, A. S., 1989. Wind in the Forests of SE Alaska and Guides for Reducing Damage. USFS-PNW-GTR-244. Harris, A. S., O. K. Hutchison, W. R. Meehan, D. N. Swanston, A. E. Helmers, J. C. Hendee, and T. M. Collins, 1974. The Forest Ecosystem of Southeast Alaska. 1. The Setting, U.S.D.A. For. Ser. Gen. Tech. Rep. PNW-12. Hennon, P. E., E. M. Hansen, and C. G. Shaw, 1990. Dynamics of Decline and Mortality of Chamaecyparis nootkatensis in Southeast Alaska. Canadian Journal of Botany 68(3):651-662. Hennon, P. E. and C. G. Shaw, 1997. The Enigma of Yellow-Cedar Decline. Journal of Forestry 95(12):4-10. Ikeya, H., 1981. A Method for Designation for Areas in Danger of Debris Flow,. In: Erosion and Sediment Transport in Pacific Rim Steeplands. IAHS Spec. Publ. 132, pp. 576-588. Johnson, A. C., 1993. The Association Between Landslides and YelLow Cedar Decline in Southeast Alaska. EOS, Trans. Amer. Geophysical Union, 74:315. Johnson, A. C., 1997. Hillslope Stability in Cedar Decline Forests of Southeast Alaska. Unpubl. Master's Thesis, Johns Hopkins Univ., Baltimore, Maryland, 52 pp. Johnson, A. C. and P. Wilcock, 1997. Effect of Root Strength and Soil Saturation on Hillslope Stability in Forests with Natural Cedar Decline in Headwater Regions in SE Alaska. In: Headwaters - Water Resources and Soil Conservation, M. Haigh, J. Krecek, G. S. Rajwar, and M. P. Kilmartin (Editors). A.A. Balkema, Rotterdam, Netherlands, pp. 381-390. Johnson, A. M., 1984. Debris Flow. In: Slope Stability, D. Brunsden and D. B. Prior (Editors). Wiley and Sons, New York, New York, pp. 257-361. Johnson, R., 1993. Inventory and Analysis of Landslides Caused by the October 25-26 Storm Event on the Craig Ranger District, Ketchikan Area Watershed Group. Unpublished U.S.D.A. Inventory. Kramer, M. G, 1997. Abiotic Controls on Windthrow and Forest Dynamics in a Coastal Temporate Rainforest, Kuiu Island, Southeast Alaska. Unpublished Master's Thesis, Montana State University, Bozeman, Montana, 79 pp. Krammer, J. S., 1965. Seasonal Debris Movement from Steep Mountain Side Slopes in Southern California. Proceedings, Federal Interagency Sediment Conference, 1965, Paper 12, in U.S.D.A. Misc. Pub. 970, pp .85-88. Landwehr, D. J., 1993. Inventory and Analysis of Landslides Caused by the October 25-26 Storm Event on the Thorne Bay Ranger District. Ketchikan Area Watershed Group, U.S.D.A., Forest Service, Ketchikan, Alaska, Unpublished Report. Minore, D., 1969. Yellow Skunk-Cabbage (Lysichitum americanum Hult and St. John) - An Indicator of Water-Table Depth. Ecology 50(4):737-739. Montgomery D. R. and W. E. Dietrich, 1994. A Physically Based Model for the Topographic Control on Shallow Landsliding. Water Resour. Res. 30(4):1153-1171. JOURNALOF THE AMERICANWATER RESOURCESASSOCIATION 29 JAWRA Johnson, Swanston, and McGee U.S.D.A., 1997. Land and Resource Management Plan, Tongass National Forest. Forest Service Publication, Alaska R10-MB338dd, pp. 4-8 to 4-12. Walkotten, W. J. and J. H.Patric, 1967. Elevation Effects of Rainfall Near Hollis, Alaska. USDA Forest Service Research Paper PNW-53, Institude of Northern Forestry, Juneau, Alaska, 8 pp. Wu, T. H., W. P. McKinnel, and D. N.Swanston, 1979. Strength of Tree Roots on Prince of Wales Island, Alaska. Can. Geotech. J. 16 (1):19-33. Zeimer, R. R., 1981. Roots and the Stability of Slopes. Int. Assoc. Hyd. Sci., Publ. 132:43-357. Zeimer, R. R. and D. N. Swanston, 1977. Root Strength Changes After Logging in SE Alaska. Res. Note PNW 306, For. Serv., U.S. Department of Agric., Portland, Oregon, 10 pp. JAWRA 30 JOURNAL OF THE AMERICANWATER RESOURCESASSOCIATION