UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Rates of revegetation of gullies in coastal British Columbia : implications for sediment production Pellerin, Diane 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2000-0261.pdf [ 9.86MB ]
Metadata
JSON: 831-1.0089473.json
JSON-LD: 831-1.0089473-ld.json
RDF/XML (Pretty): 831-1.0089473-rdf.xml
RDF/JSON: 831-1.0089473-rdf.json
Turtle: 831-1.0089473-turtle.txt
N-Triples: 831-1.0089473-rdf-ntriples.txt
Original Record: 831-1.0089473-source.json
Full Text
831-1.0089473-fulltext.txt
Citation
831-1.0089473.ris

Full Text

R A T E S O F R E V E G E T A T I O N O F G U L L I E S I N C O A S T A L B R I T I S H C O L U M B I A : I M P L I C A T I O N S F O R S E D I M E N T P R O D U C T I O N by D I A N E P E L L E R T N B. Sc., Universite du Quebec a Rimouski, 1995 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F S C I E N C E in T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Department of Geography) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A May 2000 © Diane Pellerin In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2/88) 11 Abstract Steep forested terrain in coastal British Columbia is typically dissected by V-shaped debris flow channels, or gullies. These steep, unstable slopes are important sources of sediment. Debris flows scour gully sidewalls, strip vegetation cover and leave mineral soils vulnerable to erosion. Accelerated sediment production persists for several years following a debris flow. However, little is known about the time frame for gully sidewall re-stabilisation or the associated decline in sediment production over time. The objectives of this study were to compare rates of revegetation between logged and unlogged gullies, to estimate fine sediment discharge from gully sidewalls at different stages of vegetation recovery, and to examine the role of environmental factors in determining plant succession and fine sediment production. Eleven gullies were studied in the coastal western hemlock forest of Coquitlam basin, southwestern Coast Mountains. Vegetation quadrats were established at 92 sites and sediment yield was monitored at 18 of the sites for one year. Community analysis suggests that land management influences the availability of colonising species and gully light regimes. Consequently, logged and unlogged gullies show different community associations and vegetation recovery rates. Six years after a disturbance, vegetation cover was significantly lower on unlogged compared to logged sidewalls (58% and 78%, respectively). After 19 years vegetation cover was similar in both environments (111% and 114%, respectively). The logged environment was dominated by shrubs such as Rubus parviflorus and Thuja plicata, and likely will become a Thuja plicata community. Initially, the unlogged environment was dominated by ferns and forbs; however, 19 years later it showed signs of future dominance by Thuja plicata. Results suggest that debris flow disturbance in gullies promotes higher regional plant species diversity. Sediment yield declined over time in most gullies following revegetation. Sediment yields varied between 0.1 to 72 m /ha/yr in logged gullies, and 0.3 to 24 m /ha/yr in unlogged gullies. Sediment yields showed a similar rate of decline in both environments over time. Results indicate that plant cover on gully sidewalls accounted for most of the variation in sediment yield. Other variables such as percent bare soil, slope angle and adjacent vegetation stands (which intercept rainfall reducing rainsplash erosion) were also important determinants of sediment yield. Consequently, by determining composition of adjacent stands, land management significantly influences sediment yield from gully sidewalls. Table of contents Abstract 1 1 Table of contents iii List of Tables vi List of Figures vii List of Appendices ix Acknowledgement x Chapter 1: Introduction 1 1.1 Background to debris flow gullies in southwestern British Columbia 1 1.2 Objectives 2 Chapter 2: Overview of gully morphology and processes responsible for sediment production in gullies... 4 2.1 Gully morphology 4 2.2 Factors and processes controlling sediment input to a gully 5 2.2.1 Mass movement processes 6 2.2.2 Surface sediment production 7 2.3 Plant colonisation and development on landslide scars 11 2.4 Summary 12 Chapter 3: Study Area 14 3.1 Location of the research area 14 3.2 Physiographic setting 15 3.2.1 Geology 15 3.2.2 Climate 17 3.2.3 Vegetation and forest harvesting 17 3.3 Gully descriptions 18 Chapter 4: Methods and Measurements 21 iv 4.1 Field surveying of gullies 21 4.2 Experimental design 23 4.3 Debris flow dates 24 4.3.1 Vegetation inventory 24 4.4 Sediment trap monitoring 28 4.4.1 Method : 28 4.4.2 Evaluation of sediment yield..... 29 4.4.3 Evaluation of errors in the sediment yield estimation 31 4.4.4 Lab work on soil and eroded sediment samples 32 4.5 Analytical methods 32 4.5.1 Ordination methods for vegetation 32 Chapter 5: Results • 36 5.1 Gully attributes 36 5.2 Plant cover variations according to major treatment groups 37 5.2.1 Floristic comparisons 43 5.2.2 Species richness 43 5.2.3 Diversity-abundance curves 46 5.3 Ordination analysis results 49 5.4 Soil erosion results from sediment traps 52 5.5 Summary 60 Chapter 6: Discussion 61 6.1 Vegetation Recovery 61 6.2 Sediment production on gully sidewalls 68 6.2.1 Plant cover and sediment yield 68 6.2.2 Indicator species for erosion intensity 70 6.2.3 Role of vegetation stature in erosion rates 71 6.2.4 Soil loss prediction model 72 V Chapter 7: Conclusion 74 7.1 Plant re-establishment conclusions 74 7.2 Soil erosion conclusions 75 7.3 Summary of the main results: 75 Bibliography 77 Appendices 84 V I List of Tables Table 3.1. Climatic characteristics for the study area 17 Table 4.1. Domin-Krajina vegetation cover scale 26 Table 4.2. Quadrat and sediment trap numbers for the various gullies 26 Table 4.3. Organisation of gullies per treatment groups 27 Table 5.2. Means (u) and standard deviations (a) for percent bare soil, other cover ( C W D and rocks) total plant cover (%), and cover per vegetation strata: tree, shrub, herb and ground cover (cover lower than herb) across treatment and control groups 39 Table 5.3. One-tailed Student's t-test summary for percent bare soil, other cover ( C W D and rocks) total plant cover (%), and cover per vegetation strata: tree, shrub, herb and ground cover (cover lower than herb) across treatment and control groups 40 Table 5.4. Index of similarity in species composition 43 Table 5.5. Kolmogorov-Smirnov goodness-of-fit tests (for continuous distributions), indicating the level of similarity between the distributions of observed mean percent herb cover of treatment and control groups 47 Table 5.6. Variables associated with sediment traps and their contributing areas ; 53 Table 5.7. Soil textural properties for surface soil samples collected in gully quadrats and soil sampled in the sediment traps 55 Vll List of Figures Figure 2.1. Morphological components of a gully 5 Figure 3.1. Location map of Coquitlam basin, GVRD watershed, and Cedar creek 14 Figure 3.2. Field site locations along Cedar Creek, Coquitlam basin, after GVRD, (1999a) 16 Figure 3.3. Typical gully sidewalls of a) logged and b) unlogged gullies 19 Figure 4.1. Schematic view of a typical gully 22 Figure 4.2. Schematic representation of experimental design 23 Figure 4.3. a) Vertical projection of stratified forest stand and b) in scale, simplified plan view of nested vegetation sampling quadrats 25 Figure 4.4. Typical 1.2 m sediment trap, showing a prism of accumulated sediment 28 Figure 4.5. Evaluation of the height of the accumulation prism 30 Figure 4.6. Schematic showing the prism of accumulation and the different angles 30 Figure 4.7. Schematic showing the difference assumption a) and b) of sediment accumulation in the sediment trap and the difference in the angle of the prism 31 Figure 4.8 First and second axes of C C A diagram which show the biplot interpretation of species for the slope aspect gradient 35 Figure 5.1 Mean species richness of a) herb and b) shrub strata, for different treatment and control groups 45 Figure 5.2. Observed distribution of mean percent cover for herb species in all treatment (•) and control groups (•) 47 Figure 5.3. Abundance curves of treatment groups represented by their observed distribution compared with expected distributions of the Geometric Series model 48 Figure 5.4. Axis 1 and 2 of Detrended Correspondence Analysis diagram 50 Figure 5.5. Axis 1 and 2 of Canonical Correspondence Analysis diagram 51 Figure 5.6. Bar graph comparing, for all treatment groups, textural analysis results of soil samples from sidewalls and sediment traps 56 Figure 5.7. Sediment yield determined from sedimentation traps vs bare soil, including elapse time since last debris flow 57 Figure 5.8. Sediment yields (m3/ha/yr) as a function of vegetation cover (%) for logged and unlogged land treatments 58 Figure 5.9. Axis 1 and 2 of C C A triplots 59 Vlll 3/ Figure 6.1. Sediment yields (m3/ha/yr) as function of vegetation cover (%) for land 'treatments' of high and low adjacent vegetation 72 Figure 6.2. Percent bare soil times sine of the slope as a function of soil loss, compared with Bovis (1982) and Millard (1993) 73 ix List of Appendices Appendix 1: List of species Latin names, abbreviations and common names 84 Appendix 2: Mean variables measured per sediment trap during monitoring, and used to determine yielded volumes 86 Appendix 3: Fine material texture analysis and mineral and organic content. 89 Appendix 4a: Species cover - bar graph sumarising the following species cover per treatment groups 1 91 Appendix 4b: Species cover - plant species cover per quadrat 93 Appendix 5: RDA analysis 105 Appendix 6: DCA analysis 108 Appendix 7: CCA analysis 109 Appendix 8: a) Forward and b) Manual selection 110 Appendix 10. Variables used to evaluate sediment yield 112 Appendix 11. Correlation matrix of the variables associated with the sediment traps 113 Appendix 12. Sediment yield CCA analysis 114 Acknowledgements This thesis summarises several years of work, which includes two summers of field work, and a whole year of hacking my way through multivariate analysis and the realm of ordinations, in addition to the every day frustrations of working in a second language. This work would not have been possible without the help of my supervisor, Mike Bovis, whom I sincerely thank for his constant support and patience, and for all his long hours of editing. I would like to acknowledge Tom Millard for initiating the idea that lead to this project and for field information, and Greg Henry, my second reader, for taking the time to overview this work and answer the ordination uncertainties. This work was funded by Forest Renewal BC (FRBC), with logistical help from Greater Vancouver Water District (GVWD). I want to specifically thank Derek Bonin, Lome Gilmour and Dave Dunkley who answered my questions and provided me with maps, aerial and satellite photos, and ecological inventory information. I also would like to say how much I appreciated working in the Coquitlam watershed, which is a beautiful area staffed by very nice people: thanks to the watershed staff for making sure we were safe and for putting-up with our late working hours. The support of friends is necessary in this kind of work and without Wendy Hales, Jenny Salmond, Andrew Murphy, Rodrigue Cloutier and Brian Menounos this thesis would not exist. Thanks to all of you. No applied research is done without data collection, in this case, many people agreed to walk the steep and difficult road of gully terrain with me, to count grasses, to capture rocks, and to spy on the weather (which got us at end!!). For all of this, I must say thanks to Kelly Hishon, Rebecca Brown, Wendy Hales, Margaret Bacon, Kristy Steckler, Matt Saskals and Etienne Rivard. Also, thank you to Jennifer Hales, Daphne Hales, Adam Gillespie, Catherine Griffiths, Mike Trebberg and Bernard Gourdeau for comments, editing, mathematics, statistics, and computer hints. A very special thanks to two persons who mean a lot to me, and who, in different ways, encouraged me to undertake this work. Chris Barnes, you were right, it is possible to do a M.Sc. in English. Bernard Hetu, merci de m'avoir transmis, a travers ta passion pour ton travail, la graine qui a fait germer l'idee d'integrer la geographie a mes connaissances botaniques. Enfin, pour le support et l'encouragement des derniers milles, mais surtout pour avoir retenu le ciel gris de me tomber sur la tete, par Toutatisl, merci Etienne et a charge de revanche. D.P. 1 Chapter 1: Introduction 1.1 Background to debris flow gullies in southwestern British Columbia Most steep forested slopes in the southern Coast Mountains of British Columbia are dissected by gullies, which are steep, v-notch drainage channels of first or second order (Chatwin et al, 1994). Gullies are generally less than one kilometre in length and vary in depth from 3-30 m (Bovis et al, 1998). Gullies are points of convergence on hillslopes where water and woody debris are concentrated, stored, and transported subsequently to higher-order stream channels. This accumulation of water and debris in gullies, and their inherent steepness, makes them susceptible to mass movement processes, especially debris flows. Channelised debris flows are responsible for large episodic deliveries of sediment to higher-order stream channels. Debris flows, travelling at speeds of 5-10 m/s, and with depths of 2-5 m, can transport volumes of up to 50,000 m 3 of sediment and woody debris (Bovis, 1993). This large sediment delivery is acknowledged to have significant impacts on both downstream water quality and stream ecology (Hogan et al, 1998). Consequently, more and more attention has been not only drawn to the debris flow process itself, but also to its ecological impacts. Studies range in scope from the geotechnical aspects of debris flows (Chatwin et al, 1994; Fannin and Rollerson, 1993; Jordan, 1994; and Bovis and Jakob, 1999) to those dealing with associated downstream impacts (Wilford and Schwab, 1982; Roberts and Church, 1986; Jordan and Slaymaker, 1991; and Bovis et al. 1998). In the 1980's, there was an increasing emphasis on the relations between forest harvesting and the increased occurrence of debris slides and debris flows in gullies (Sidle et al. 1985; Bovis et al, 1998). It was observed that debris flows in clear-cut areas were 2.2 times more frequent than in undisturbed forests of similar characteristics (O'Loughlin, 1972). Wilford and Schwab (1982) also noted an increase of mass movement activities in logged gullies caused by a reduction of the root tensile strength due to decay following logging. During important storms, Schwab (1983) observed in the Queen Charlotte Island a 41 fold increase in landslide activity in clear-cut areas relative to undisturbed forests. All these studies document the large impact of forest harvesting on debris flow frequency, but do not address the issue of gully recovery following a debris flow. 2 Debris flows scour a gully and leave large areas of bare and unstable soil exposed to erosive agencies along the sidewalls and channel. As a consequence, sediment delivery to the gully and to higher order streams increases after a debris flow event. Very few studies have focused on this aftermath period of sediment production. Millard (1993) and Bovis et al. (1998) investigated fine-sediment production in real-time, and the role of coarse woody debris ( C W D ) in trapping sediment during the gully recharge following a debris flow. Nistor (1996) attempted to isolate the main processes involved in real-time fine-sediment production in logged gullies. Although these studies provided information on fine-sediment production processes in gullies, they did not consider the longer term variability in sediment production related to the recovery of gullies following debris flows. After a debris flow, bare gully sidewalls are slowly recolonised by vegetation, which provides cover and protection against erosion processes, and thus progressively reduces sediment production. Very little is known about the processes of plant invasion and development in this specific environment and, especially how sediment production varies according to plant successional stages. In this study a biogeographical survey of gully vegetation is combined with the monitoring of soil erosion to provide estimates of sediment production in relation to the re-establishment of vegetation on gully sidewalls. The study is conducted at the plot-scale with plots systematically nested within gullies. Given the acknowledged impact of forest harvesting on sediment production, comparisons are made between logged and unlogged gullies. The gullies investigated were fairly typical of those found in the southern Coastal Mountains (Bovis et al. 1998), thereby increasing the prospects for applying the study results to wider region. 1.2 Objectives The four main research objectives were: (1) to document the rate of plant invasion on gully sidewalls following a recent debris flow of known date. (2) to compare the variation in vegetation recovery rate between logged and unlogged gullies. (3) to determine the role of environmental factors, such as slope angle, slope aspect, and soil characteristics on plant succession. (4) to conduct short-term monitoring of soil erosion rates from gully sidewalls at various stages of plant succession. 3 The main working objective can be stated as a hypothesis: gully sidewalls should become progressively more stable with time as a result of plant succession. In this study, the term succession is intended to refer to the development of plant communities through time after a disturbance. One of the major goals of my research is to determine the rate of re-stabilisation with plant succession. The second objective concerns the effect of land treatment on gully revegetation, and is best presented as a question rather than a hypothesis: do gullies in clear-cuts recovery faster or slower than those in old-growth environments? At present, there is no information which allows predictions to determine which areas should re-stabilise most rapidly. On the one hand, clear-cuts have a higher light regime than old-growth areas, allowing shade-intolerant successional species to gain an early advantage. This would imply a more rapid stabilisation of sidewalls in clear-cuts. However, old-growth gullies have a moisture advantage over clear-cut gullies during the long summer drought in southwestern British Columbia, which might offset their disadvantage in solar energy inputs. Two main variables are used in the experimental design: (1) Land-treatment, corresponding to clear-cut or old-growth conditions, is used to determine if clear-cutting influences the rate and type of revegetation in gullies. (2) Time elapsed since the last debris flow; this allows for the effect of time to be isolated from land treatment. Since debris flows scour a gully of its vegetation, the plant succession clock is reset to zero since an almost bare mineral soil environment is produced. By selecting different elapsed times since last debris flow (in other words, a space for time substitution), it is possible to observe the pioneer stages of vegetation and later the full plant community development. The third objective is self-explanatory and proposes to test environmental variables collected in the field, using gradient analyses, to determine the factors controlling the plant succession and soil loss. Sediment production, the subject of the forth objective, is evaluated using soil erosion plots, and follows the same experimental design by monitoring the sidewalls of clear-cut or old-growth gullies at different elapsed times since disturbance. Given the large spatial variability in soil erosion rates, the erosion values obtained here are regarded as indices of sediment production on gully sidewalls rather than accurate average values of erosion rate at the various plant successional stages. 4 Chapter 2: Overview of gully morphology and processes responsible for sediment production in gullies. 2.1 Gully morphology Gullies or steep v- notch drainage channels are usually first, but occasionally sometimes second order channels that dissect steep hillslopes in the forested belt of Coastal British Columbia (Chatwin et al, 1994). Compared to lower gradient stream channels, gullies have greater depth relative to width, and are characterised by steep headwalls and sidewalls. They carry a proportionally larger sediment load relative to their discharge (Heed, 1975). Gullies combine hillslope and stream channel features, and can be divided into four main morphological components: headwall, sidewall, channel and fan (Figure 2.1). Each of these morphological components and its associated sediment production and delivery processes are defined and explained in this section. The gully headwall marks the upper limit of the gully. The headwall has a triangular to trapezoidal shape and a slope generally greater than 30°. Sidewalls flank the gully channel and, like the headwall, often have steep gradients but slope angles can vary from less than 15° to more than 50°. This degree of variability can occur in a single gully and depends on the amount of sidewall undercutting by the gully channel. Some of the lower gradient sidewall slopes are due to deposition by debris flows rather than erosional undercutting. Gully headwall and sidewall composition vary regionally. In the southern Coast Mountains, the most common headwall and sidewall materials are coarse colluvium deposits, glacial till, and bedrock. In general, sidewall and headwall materials tend to be cohesionless, especially within the granitic southern Coast Mountains. Lack of cohesion renders sidewalls and headwall areas very susceptible to mass instability from debris slides, as well as ravelling and rainsplash erosion. The loose nature of sidewall materials also renders them prone to deep scour when debris flows occur. As a result, there is reasonable assurance that virtually all of the pre-existing shrub and herbaceous cover is removed by such an event. This is an important pre-requisite for the studies of plant succession reported in this study. Headwall Figure 2.1. Morphological components of a gully. Sediment produced on headwalls and sidewalls is concentrated in the gully channel and, depending on the water discharge and gradient of the channel, may be transferred to a lower channel reach or stored in the channel. Gully channels are often have an ephemeral to seasonal flow regime, and do not have surface flow along their whole length year round. This implies an episodic movement of sediment during major storms. Both logging slash, and large woody debris tend to accumulate in the gully as well and may strongly influence the storage and delivery of mineral material (Bovis et al, 1998). The fan is a flattened cone-shaped depositional zone at the lower end of a gully. Fans may be composed of alluvial or colluvial material, or a mixture of both, depending on the relative importance of stream flow and debris flow processes in the gully. Alluvial material is deposited on the fan as a result of stream competency changes when the water passes from the gully channel to the gentler fan stream channel. 2.2 Factors and processes controlling sediment input to a gully Erosion processes correspond to the detachment and transport of soil or rock by the action of water, wind, frost action or gravity (Selby, 1993). The intensity of erosion is controlled by numerous factors, principally slope angle and length, material erodibility, rainstorm and 6 snowmelt events, vegetation cover and land management. These factors have long been recognised as important in predictive schemes such as the Universal Soil Loss Equation ( U S L E ) (Selby, 1993). Erosional processes delivering sediment to gully systems can be classified under two distinct categories. The first corresponds to episodic events that generate large amounts of material by mass movement. Debris slide, debris avalanche, small slump and debris flow are mass movements most commonly found in gullies, and tree throw is also locally significant. Soil creep is also considered a mass movement process, but is not episodic. The second erosion process category refers to surface sediment production by rainsplash, sheet wash, frost heave, and dry ravel. The factors and processes that control sediment production are described and explained in more detail in the following sub-sections. 2.2.1 Mass movement processes A combination of variables, such as steep headwalls and sidewalls, relatively loose materials, and a humid climate make the gullies of southwestern British Columbia very prone to mass movements. One of the most devastating gully mass movements, responsible for large sediment transfers, is the debris flow. As mentioned in the introduction, debris flows can rapidly transport thousands of cubic metres of material over large distances in just a few minutes. This rate of doing geomorphic work is thus millions of times greater than that of surface erosion processes, but is episodic in nature. Confinement of a debris flow within a gully channel results in a high sediment transfer velocity and a high flow depth. Consequently, the flow can deeply scour the sidewalls, entraining more material. High velocity flow also means farther travel, which has important outcomes on the amount of sediment delivered to higher-order streams (Chatwin, et al. 1994). Debris slides are movements of large blocks of non-deformed soil whereas debris avalanches involve complete disintegration of the detached mass and a higher moisture content (Selby, 1993). Small shallow rotational slumps are also common failures observed on steep slopes. All of these movements can occur on gully headwalls and sidewalls and can trigger debris flows by suddenly increasing the shear load in a saturated channel (VanDine, 1985; Bovis and Dagg, 1992). Debris flows can also be triggered by excessive water accumulation in a channel due to rainstorms, snowmelt or temporary damming of the channel (Bovis et al, 1985). Wind tree throw, the uprooting of tree due to wind stress, and high soil moisture level can result 7 in debris slides or debris avalanches, forms of mass movements involving shallow soil failure on steep slopes. The accumulation of debris in a debris flow gully is an important prerequisite for debris flow initiation as observed by Wilford and Schwab (1982). Research by Millard (1993), Oden (1994), and Bovis et al. (1998) attempted to clarify the role of woody debris accumulation in gully mass movement initiation. They concluded that slash removal after logging could reduce the volume of subsequent debris flows by reducing sediment storage potential. Smaller debris flows could imply a lesser scouring of the channel and sidewalls. However, there is no evidence that slash removal reduces the frequency of debris flows or other types of mass movement these studies indirectly showed how yarding methods influence gully stability. The selection of better yarding methods can minimise soil and underbrush disturbances resulting in better root cohesion as well as a better environment for regrowth, which in turn will reduce soil erosion (Sidle et al., 1985). In a review of literature on the effect of timber harvesting in the Coast Mountains of British Columbia, O'Loughlin (1972) observed higher rates of debris avalanches in logged areas relative to undisturbed forest. Rood (1984) and Roberts and Church (1986) also demonstrated mush higher sedimentation from mass wasting in clear-cut than undisturbed forest areas. Higher rates of mass movements after timber harvesting are due to: (1) reduced root reinforcement and sites disturbance after logging (O'Loughlin, 1974; Gray and Megahan, 1981; Ziemer, 1981; Riestenberg and Sovonick-Dunford, 1983; Sidle, 1992); and (2) increased water inputs, thus higher soil moisture, as a consequence of reduced evapotranspiration, changes in the snow pack volumes and rate of snowmelt (Sidle et al, 1985). Other research has been undertaken to evaluate the relationship between forest harvesting and debris flow frequency (Rood, 1984, 1990; Krag et al, 1986; and Young, 1992), and generally agrees with the fact that landslides and debris flow activities are accelerated by logging activities and road construction. 2.2.2 Surface sediment production Processes such as rainsplash, sheet wash, frost heave and dry ravel act on the soil surface and produce fine to moderately coarse sediment. Even though they act at a smaller scale than mass movements, their frequency is higher and, therefore, the amount of sediment produced from these processes may be significant in the long term. Variables such as soil properties, slope morphology (gradient and length), climate and land cover (also referred to as land management or vegetation cover), control the sediment production. These variables are considered in soil 8 management and are integrated into surface erosion models such as the U S L E (1965 and 1978) developed by the United State Department of Agriculture ( U S D A ) (Selby, 1993). However, these types of models are used mostly on croplands with Hortonian overland flow and may not apply to steep forested hillslopes such as those found in southwestern British Columbia. Each of these variables is discussed briefly below. Note that the land management factor is divided into a description of the vegetation factor and the forested land management practices. Soil properties Soil erodibility is a measure of how bare soil resists detachment and transport. Since erodibility depends on many factors it is very difficult to quantify. Resistance to detachment is determined by surface shear strength, principally the cohesion and the internal angle of friction. Cohesion corresponds to particle bonding, either by clay content or vegetation root systems. The internal angle of friction is influenced by the shape, roundness, size and packing arrangement of the soil particles (Sidle et al. 1985). Although it is possible to measure average values of cohesion and internal angle of friction in the soil layer to evaluate mass slope stability, there is no ideal method to determine the same values for a soil surface and to evaluate the likelihood of surface erosion. Researchers have developed empirical models from results obtained with erosion plots. The U S L E (Dissmeyer and Foster, 1985), is an example of this kind of work. The difficulty with using these models is that they are designed for specific environmental factors. Most researchers agree that, whatever the environment, soil loss is correlated with rainfall intensity, slope angle, and splash susceptibility of soils, the latter a function of texture and cohesion (Selby, 1993). Rainsplash erosion is the detachment of soil due to raindrop impact. Once soil is detached, slope processes determine the subsequent transport of the particles. Slope morphology Slope angle and slope length control the distance soil particles move. Particles move greater distances on a steeply inclined plane, and slope length determines the potential for runoff accumulation from rainfall or snowmelt. The combination of both variables implies that overland flow processes will move soil faster and further on long steep slopes. Sheet flow erosion or overland flow erosion is the transport of soil by laminar flow of water running over the ground. The erosive ability of sheet flow to transport soil is related to the velocity and the depth of the flow (Horton, 1945; and Selby 1993). 9 Climate Precipitation, temperature, and wind are climatic factors that influence sediment production (Selby, 1993). Precipitation is responsible for rainsplash, sheet flow erosion processes and for increasing soil moisture levels that modify the soil mass shear strength. Temperature acts indirectly by changing the type of precipitation and by causing the ground to freeze, which limits rainsplash and sheet flow erosion. On the other hand, the formation of needle ice or frost heave is a direct erosion processes that can be very effective at detaching soil particles through the formation of ice crystals which are then transported down by gravity, rain, or wind (Selby, 1993). Wind can transport fine material and can cause tree throw. Tree throw mobilises significant amounts of material and leaves the ground exposed to further surface erosion. Strong winds can also modify the angle and the speed at which precipitation hits the ground, determining the outcome of rainsplash erosion. Whatever climatic processes are responsible for soil loosening, there is a large body of evidence showing correlations between rainfall energy and rate, and soil detachment. An important factor moderating rainfall impact energy is plant cover. Land management: vegetation Vegetation is a factor that influences most erosional processes by providing a protective cover for the soil and by adding shear strength through root systems. The cover provided by vegetation reduces rainsplash and sheet flow erosion by intercepting rain in the canopy and by increasing soil porosity, which in turn decreases the potential for water accumulation at the soil surface. If sheet flow still occurs, vegetation will decrease its velocity by providing greater surface roughness, thus decreasing the ability of sheet flow to detach and entrain sediment. Vegetation also thermally insulates the ground, reducing the occurrence of frost heave. However, vegetation is responsible for leaf-drip. Water drops falling from the canopy may be larger than those from direct rainfall, and consequently may produce similar, if not worse, surface erosion than unmodified rainsplash. This process is often tempered by the presence of litter on the ground, which intercepts water drops. Source of fine-sediment in forested areas Since vegetation plays an important role in stabilising, and protecting soils, surface erosion is not as common on forested land as it is on cleared land. Nistor (1996) demonstrated in a small basin of coastal British Columbia, that sediment transfer activity occurs in well defined areas of bare soil and is driven by episodic climatic events. Other literature on the topic shows 10 that riparian areas including gully sidewalls, road cuts, and landslide scars are the major sources of sediment from surface erosion in forested land (Carr, 1985; Roberts and Church, 1986; Nistor, 1996). In these areas, the main surface erosion processes are rainsplash erosion, sheet flow, dry ravel and frost heave (Nistor 1996). Although landslides constitute an important source of sediment, few studies have focussed on the production of sediment from landslide scars in British Columbia. Landslide scars in gullies, which are the erosional forms considered in this study, are often avoided due to their complexity and inaccessibility. The result of this is that very few data are available on the input of sediment to streams such as gullies. Furthermore, there are even fewer data on temporal variations in erosion rates initiated by debris flow scour. Recently, Millard (1993) and Bovis et al. (1998) have considered the problem of chronic fine-sediment production in gullies following debris flow events in British Columbia. These two studies were interested in the importance of fine-sediment production in gully recharge processes, and the implications of this for future debris flow events. Their work compared logged gullies that were slash full, slash clear, and recently torrented, with unlogged gullies that had not recently torrented (at least in the past 15 years, M . Bovis, pers. com., 2000). Hereafter, the term "torrented" is used to refer to gullies that have experienced debris flow. They observed that production of fine-sediment was highest in torrented gullies, least in slash full gullies, and fairly low in old-growth gullies. The clear maximum observed in torrented gullies indicates that following a debris flow, surface erosion is much higher for a few years after the event, then slowly decreases. They argued that the importance of surface erosion in recently torrented logged gullies is directly related to the percent of bare soil found in these gullies. However they did not conduct detailed assessments of erosion rates under different vegetation covers. Since it is demonstrated that forest harvesting increases the rate of mass movements in gullies, and thus higher fine-sediment production occurs following failures, Hogan and Millard (1998) proposed a gully assessment procedure (GAP), the main goal of which was to reduce the risk of failure, after harvesting, and thus sediment delivery to lower streams. The G A P stresses the fact that erosion, channel stability, and sediment and debris transport should not be greatly modified by forest harvesting under best practices. This is achieved by eliminating disturbance that results in exposing mineral soil in the gully. One recommendation was for hydroseeding and shrub planting for destabilised or scoured gullies. However, as shown by Carr and Wright (1992) these procedures can be relatively expensive and the need for bio-geotechnical work can only be assessed from the potential downstream impacts of chronic erosion problems within a gully. 11 Often the expected gully sediment yield is based on assumed values, since the natural rate of vegetation recovery on scoured gully sidewalls is unknown. Consequently, accurate predictions of the fine-material production in gullies have hitherto not been possible. The only local study on the topic of vegetation regrowth on landslide scars was conducted by Smith et al. (1986), who looked at merchantible forest regrowth on debris avalanche scars compared with adjacent slopes in the Queen Charlotte Islands. Moreover, no attempt was made to relate vegetation recovery rate to sediment production from those sites. Studies on the vegetation regrowth on landslide scars have been attempted elsewhere, however, and the next section concentrates on the general findings concerning revegetation in non-British Columbia environments. 2.3 Plant colonisation and development on landslide scars. New sites for vegetation to colonise are periodically created by landslides and other catastrophic events at various scales. The conditions that prevail during vegetation re-establishment at a specific site are mostly a result of the changes which produced the development of bare soil and of the new environmental conditions prevailing after the soil denuding event (van Hulst, 1978). To a large degree, the pioneer colonisation and the course of plant succession at a disturbed site are controlled by the local site conditions such as soil characteristics, slope angle, light and water availability, local seed and spore sources and the inherent abilities of plants to colonise new sites. Very few studies have been conducted specifically on plant succession following rapid mass movement disturbances. Studies conducted on the vegetation present on avalanche paths showed that plant cover is not modified in the same way by avalanches compared with debris flows. The literature in this field was not used for direct comparison with this study since soil transformation processes and the amount of surviving vegetation after each type of event are very different. From the available studies on landslide scar, leading factors emerge as dominant controls of plant recovery: plant substrates, and the proximity to seed sources. The importance of the plant substrate, or soil, is recognised as a major factor in plant establishment, and in the case of landslide scar sites, was identified as the most important factor. On a landslide scar, the level of soil scouring controls the presence or absence of developed soil profiles and their quality, for example, deep-seated landslides would remove all trace of a former soil. In the case of disturbance by rapid mass movement, depending on the intensity of the event, soil can be partially or totally scoured from the site. Veblen and Ashton (1978) and Hull and Scott (1982) have observed that vegetation establishment is greatly influenced by the degree of 12 regolith removal. Significant differences in revegetation rate have also been found between accumulation and scour zones of the landslide scars (Flaccus, 1959; Veblen and Ashton, 1978; Hull and Scott, 1982; Miles and Swanson, 1986; Smith et al, 1986; Gecy and Wilson, 1990). Rates are generally higher in the deposition areas, relative to the scour areas. The second factor to stand out is seed availability, which is greatly influenced by the adjacent vegetation stands. In the case of deep scouring of the slope, all the remaining vegetation is assumed to have been removed (Flaccus, 1959; Hupp, 1983; and Sousa, 1984) and the surviving adjacent vegetation is therefore the main seed source for regrowth (Hull and Scott, 1982; Miles and Swanson, 1986; Smith et al. 1986; and Gecy and Wilson, 1990). This underscores the importance of land management history in determining future vegetation communities at a slide site (Gecy and Wilson, 1990). Adjacent vegetation may interfere with community development on a slide by indirectly modifying the environmental characteristics of the slide area. Gecy and Wilson (1990) observed this when comparing regrowth on torrented clear-cut stream reaches with regrowth on torrented stream reaches surrounded by hardwood or coniferous forest. Trees surrounding stream reaches filtered most of the light and reduced the available light resources at ground level. The reduction in light would favour more shade tolerant species communities. Flaccus (1959) noted that other factors such as slope angle, and light and water availability were not as significant during revegetation, simply because most sites are not differentiated in these respects. Slide areas generally have steep slopes, and are relatively open to light and experience the similar weather conditions to undisturbed sites. Other studies observed in addition to the influence of the substrate and the adjacent vegetation, the influence of slope shape, soil moisture, and slope aspect (Miles and Swanson, 1986). However, Gecy and Wilson (1990) did not detect a moisture effect across vegetation community patterns, although a moisture gradient is observed in riparian studies. 2.4 Summary Gullies are linear erosional features which are acknowledged to be important sources of mass movements. Debris slides, debris avalanches, small slumps and tree throw are important processes of mass instability on gully headwalls and sidewalls. Debris flows are channel processes initiated by headwall or sidewall failures that also subsequently affect the headwall and sidewalls by scouring them. All these processes deliver important amounts of sediment to the gully channel and higher-order streams. The inherent wetness and steepness of gullies means 13 that tree throw, higher soil moisture, and reduced root cohesion are all factors contributing to an increased risk of mass movement events. Surface erosion processes in gullies are very much controlled by area of bare soil, which is a function of recent slope movement history as well as forest harvesting activities. Any events capable of removing the protective layer of vegetation and opening areas of bare soil increase the area susceptible to surface erosion, and thus increase fine-sediment production. In gullies the scouring effect of debris flows on the headwall and sidewalls is the main agent exposing large areas of soil to erosive agents. Timber harvesting promotes mass movement erosion by reducing root strength and increasing the soil moisture content. It also contributes to surface sediment production by opening an area to the direct effects of raindrop impact. Work on the revegetation of landslide scars has shown that the establishment of the new plant communities on landslide scars is greatly affected by the depth of the landslide scour and proximity of local seed sources. However, very little is known on the rate of revegetation of gullies and the decrease of sediment production due to the increased protection contributed by the developing vegetation canopy 14 Chapter 3: Study Area 3.1 Location of the research area This study was conducted in Coquitlam basin, within the southern Coast Mountain of British Columbia, approximately 30 km north-east of Vancouver, British Columbia. (Figure 3.1). Coquitlam basin is managed by the Greater Vancouver Water District (GVWD) and is used as a water supply reservoir for Greater Vancouver. Figure 3.1. Location map of Coquitlam basin, GVRD watershed, and Cedar creek. The study sites are all located in the Cedar Creek sub-basin, one of the main tributaries of Coquitlam Lake (Figure 3.2). Cedar Creek valley varies in elevation from 160 m 1300 m above sea level. The Cedar Creek is a typical example of a steep sided valley of southwestern British Columbia. The valley sides are cut by dozens of gullies. Such a terrain is prone to mass movements and past studies conducted in this basin, Thurber Engineering Ltd (1991) and Millard (1993) reported debris flow activities during the 1990 winter storm. Past debris flow activities were also observed from aerial photos which, combined with the more recent ones 15 (1990 and 1993), provided potential sampling areas. The studies plus the availability of visual documents (maps and aerial photos) and the accessibility of the sites motivated the selection of this area. 3.2 Physiographic setting 3.2.1 Geology According to Roddick (1965) the dominant lithology of the area is gabbro with some quartz diorite and small sulfide exposures. Within the field sites, however, the dominant rock exposed at the surface is quartz diorite. Basal till and colluvium are the common surficial materials found in the sub-basin. Very consolidated basal till covers a large portion of the sub-basin, varying in exposure thickness from over 5 m along road-cut exposures along Branch Road 230 (Figure 3.2) (Millard, 1993) to less than 1 m on steep gradients. Bedrock and till are covered in many areas by colluvium, typically less than 2 m deep, but thicker on fan and lower slope areas (Millard, 1993). The steep slopes forming Cedar Creek valley are cut by v-notch gullies. Most gullies originate in old-growth forest areas on upper slopes where they are incised into thin drift and bedrock. Their middle to lower sections are cut into much thicker surficial materials such as till and colluvium. Colluvial fans are found at the mouths of gullies, and comprise fluvial and debris flow materials transported through gully channels. 16 S 1 : 25 00G\ Legend L1 L o g g e d gully U1 Un logged gully Contours (100m interval) S t r e a m s Cutb locks and number R o a d s : B R 1 3 0 Fie ld s i tes Cutblock treatment* and year: • 1 * 1. L78 P80 2 . L83 P85 3 . L76 P78 4. L78 B79 P80 5 . L76 P78 6 . L83-84 B87 P88 7 . L79 8. L83 P84.86 9. L79 P80 1 0 . L83 P84.87 1 1 . L77 B78 P80 1 2 . L89 P90 1 3 . L89 P90 1 4 . L89 P90 1 5 . L83-84 P84.86 1 6 . L78 B79 P80.82 *L= logged B= burned P= planted Figure 3.2. Field site locations along Cedar Creek, Coquitlam basin, after G V R D , (1999a). 17 3.2.2 Climate Coquitlam basin has a cool, humid temperate climate. Due to its location on the south-west side of the Coast Mountains, orographic precipitation is relatively important (Hay and Oke, 1973). For comparison, the study area receives about three times the 1,260 mm of precipitation recorded at sea level in Vancouver over a typical year (Canadian Climate Program, 1993). Average precipitation and temperature values from the Atmospheric Environment Service (AES) are presented in Table 3.1. Snow cover was observed to persist between December and March. However, snow precipitation is recorded in the period November to April for an annual mean of 160 cm. High rain precipitation is also recorded in the winter and peaks in December with 569 mm. Less than 200 mm of rain is reported per month from May to September. Table 3.1 provides necessary information to determine the biogeoclimatic subzone (see below). Table 3.1. Climatic characteristics for the study area. Coquitlam Lake Elevation (m) 161.0 Mean annual precipitation (mm) 3616.0 Mean precipitation of the driest month (mm) 86.0 Mean precipitation of the wettest month (mm) 569.0 Mean annual Temperature (°C) 8.4 Mean temperature of the warmest month (°C) 16.6 Mean temperature of the coldest month (°C) 0.6 Note: Data are from Canadian Climate Program, 1993, Environment Canada 3.2.3 Vegetation and forest harvesting Cedar Creek sub-basin is in the coastal western hemlock wet hypermaritime (CWHvh) biogeoclimatic zone (GVRD (1999b); and Klinka et al, 1991). In old-growth stands, the dominant tree species observed are Thuja plicata (western redcedar) and Tsuga heterophylla (western hemlock) followed by Abies amabilis (amabilis fir), Chamaecyparis nootkatensis (yellow-cedar), and Pseudostuga menziesii ssp. menziesii (Douglas-fir)1. The age of veteran trees in these stands is around 250 years (Arsenault and Bradfield, 1995). Forest clear cutting occurred in this basin between 1976 and 1989 (Figure 3.2). The early period of logging used high-lead yarding and was later replaced by a Wyssen skyline system 1 For plant species common names see Appendix 1 18 (Millard, 1993). After logging, post-harvest treatments to sites were not uniform. A few sites were prescribed for slash burning and most sites were replanted with Pseudostuga menziesii ssp. menziesii (Figure 3.2), which is not the dominant species in this biogeoclimatic area. Planted stands, or second growth stands, show differing levels of maturity. They vary from young Pseudostuga menziesii ssp. menziesii stands combined with shrub forms, through older stands that have reached canopy closure and have sparse underbrush. 3.3 Gully descriptions The Cedar Creek basin is very steep and dissected by a number of gullies. Gully selected for this study were assigned names according to their surrounding vegetation either clear-cut or old-growth forest. Gullies such as 1, 3, 4, 7, 8 and 9 include both environments. These gullies were divided into two reaches, with a letter indicating their local environment, e.g. LI for clear-cut section of Gully 1 and UI for the old-growth section of Gully 1. The main morphological attributes of the gullies sampled are given in Table 5.1. Millard (1993), Thurber Engineering Ltd (1991), and the information gathered from the air photos indicates that all selected gullies have experienced debris flow disturbance in the 22 years prior the inception of the study. In many cases, debris flows initiated in the old-growth section of a gully and affected both the old-growth and clear-cut reaches. Since debris flow scouring destroys the gully sidewall vegetation, the elapsed time since debris flow is identical in both reaches. This allows direct comparison of recovery rates between old-growth and clear-cut environments. Newly established vegetation was an indication that scouring was relatively important in all surveyed gullies. Figure 3.3 illustrates typical gully sidewalls in logged and unlogged gullies. Exposed surficial material on gully sidewalls varied between sites. For example, the upper parts of gullies LI , UI, L3 and L5 showed basal till exposed in some sections of the sidewalls. The other gullies were largely cut into colluvium. Percent of surficial material exposed on the gully sidewalls is expressed by the percent bare soil variable, and listed in Table 3.2. Total plant cover is also listed in this table, and a more detail description of the vegetation is presented in section 5.1. In addition, it should be mentioned that gullies 2 and 5 were hydroseeded. This vegetation has now mixed with, or been replaced by, local vegetation. The hydroseeding probably took place following the logging operations in 1989. Finally, another type of ground cover found on gully sidewalls was coarse woody debris ( C W D ) , which includes both slash and large woody debris (Table 3.2). Figure 3.3.Typical gully sidewalls of a) logged and b) unlogged gullies. 20 In gullies LI , L6 and L8, debris flows flowed out of the channel in the clear-cut gully sections. This resulted in large scoured areas beyond the main channel, since both the adjacent open slope and the gully channels were involved. However, the scoured open slopes showed larger amounts of CWD and clastic material deposited than the gully sidewalls. This is a function of the lesser depth, and hence lower velocity, of debris on open slope areas. These areas were also re-planted with Pseudostuga menziesii spp menziesii to help recovery, but gully sidewalls were not. Many gullies are located on the NE flanks of Cedar Creek valley and therefore are well exposed to solar radiation. By contrast, gullies 3, 6 and 7, are located on the opposite side of the valley and receive the least amount of solar radiation (Figure 3.2 and 3.3). However, during the study, moisture levels along gully sidewall soils were not observed to be significantly different between the two slope aspects. The greatest difference was observed between gully reaches in clear-cut and old-growth areas. In the old-growth forest, large veteran trees and complex vegetation stratification intercepted a large amount of the solar radiation and thus served to maintain higher moisture levels. All surveyed gullies were classed as ephemeral streams and therefore convey flow periodically during the winter wet season and are dry during the summer. Gullies often exhibit surface flow in the old-growth environment and shift to sub-surface flow in the clear-cut. This is caused by the coarse texture and high infiltrability of channel and fan materials. Channel material was generally bouldery with a higher level of CWD near the clear-cut - old-growth boundary where tree-throw is more frequent. 21 Chapter 4: Methods and Measurements 4.1 Field surveying of gullies Gully surveys consisted of morphological, soil, and vegetation measurements. Gully length was measured with a hip chain, and every ten to twenty metres, depending on total gully length, slope angles and vegetation cover were noted. Transects were established perpendicular to the gully channel at equal intervals (Figure 4.1), generally fifty metres or less from each other depending on gully length. A minimum of three transects was established per gully. Along each transect, sidewall aspect and slope angles were measured, vegetation quadrats (plots) were established on the sidewalls (Figure 4.1: Ql) and on slopes directly adjacent to the gully sidewalls (Figure 4.1: Q2 and Q3). Quadrats located outside the gully provided undisturbed control sites within both logged and unlogged portions of each gully. The different environments sampled on adjacent slopes were classified into four groups. (1) The "Torrented open slopes" corresponds to an area found only in the logged portion of the gully where open slope vegetation had been scoured by the same debris flow which scoured the gully channel and sidewalls (Figure 4.1: Q2). Since elapsed time and land treatment are constant, these plots allow isolation of a gully versus open slope effect in revegetation. (2) Adjacent slopes that were recently logged which have not yet reverted to closed-canopy second growth («10 years) are referred to as "10-year second growth". With some reservation, these sites allow isolation of a debris flow effect in cases where elapsed time since logging is similar to elapsed time since debris flow disturbance. (3) Older logged areas, which were harvested between 1977 to 1978, and planted not long after 1980, are called the "20-year second growth". Again, these allow isolation of a debris flow effect. Finally, (4) the "Old-growth" forest represents the oldest successional stage in this Coastal Western Hemlock forest, and provides an ultimate background control for undisturbed sites of "infinite" age in terms of the time frame applied to logged and torrented gully sites. 22 Figure 4.1. Schematic view of a typical gully showing locations of transects (T) and vegetation quadrats on gully sidewalls (Ql), on the torrented adjacent open slope (Q2), and control undisturbed by torrent (Q3). 23 4.2 Experimental design By documenting the vegetation recovery in gullies that have experienced debris flows at different periods in the past, one can determined a rate of vegetation recovery as a function of elapsed time after such disturbance. One objective of this study was to verify whether forest harvesting exerts a significant effect on the time required for a gully to revegetate and re-stabilise. Accordingly, gullies were selected from two contrasting environments: the undisturbed old-growth forests referred to as unlogged areas, and the clear-cut areas referred to as logged areas. In many cases, the old-growth and clear-cut segments of a gully had been affected by a single debris torrent. In these cases, elapsed times since disturbance are identical and comparisons are possible on the separate effect of land treatment. Elapsed time and land treatment constituted the two main factors in the experimental design (Figure 4.2). Time factor Land treatment (Time elapsed since last debris flow) Plant community development on gully sidewalls Logged Unlogged Logged Unlogged Figure 4.2. Schematic representation of experimental design, showing that land treatment is nested into elapsed time since last disturbance by debris flow. The gullies selected in this study experienced their last debris flows between 1975 and 1993 (Table 5.1). Gullies were grouped to represent two recolonising time frames (Table 4.3): 1) averaging 19 years, representing gullies torrented between 1975 and 1980; and 2) averaging 6 years, representing the gullies that torrented between 1990 and 1993. Gullies were divided according to time and land treatments into four distinct groups, called "treatment groups". Each of these was given a name composed from time and land treatment. For example the logged section of Gully 1, torrented in 1990, belongs to treatment group L6. The organisation of gullies within their respective treatment groups is given in Table 4.3. Henceforth, the abbreviated group names are used, as detailed in Table 4.3. Throughout the study area, gullies were carefully chosen to avoid domination by bedrock outcrops, since the objective was to monitor vegetation development and temporal variation in fine sediment production from colluvial and till sidewalls. Outcropping bedrock became more 24 frequent at higher elevation and with increasing slope angles, which complicated and limited gully selection. 4.3 Debris flow dates An important prerequisite in this study is knowledge of debris flow dates to allow assignment of elapsed times since last disturbance. Since gully selection had to incorporate both logged and unlogged gullies for different elapsed times since last debris flow, air photos, forest cover maps and previous work conducted in Coquitlam basin (Thurber Engineering Ltd, 1991, and Millard, 1993) were used to guide gully selection. Some dating of the most recent debris flows was determined using air photos. Air photos from 1979, 1982, 1987, 1992 and 1997 were available through the Greater Vancouver Water District (GVWD) and the Geographic Information Centre (UBC Geography Department). It was assumed that a newly torrented gully would show visible signs on air photos of destroyed vegetation, a large bare soil area, and possibly a new sediment fan. With these criteria in mind, it was possible, to determine, within two or three years, the date of the last torrent. Work from Thurber Engineering Ltd (1991) and Millard (1993) concentrated on the impact of the 1990 storm in Coquitlam Watershed, confirming the dating for the 1990 torrents in gullies: 1, 3, 5, 6, and 7. Forest cover maps were used to determine the date of logging and the treatment applied to cut blocks after logging activities, which was also very useful in characterising the areas adjacent to gullies. The elapsed times since last torrent for the surveyed gullies are listed in Table 5.1. 4.3.1 Vegetation inventory Vegetation sampling area was based on square quadrats of variable size. Guidance as to size was based on the work of Smith et al. (1986), who sampled similar vegetation communities, and on the BC Ministry of Environment guidelines provided in "Describing Ecosystems in the Field" (Luttmerding et al, 1990). Vegetation cover was divided into three main strata, according to the most simplified scheme suggested by Luttmerding et al. (1990). Stratum A (Figure 4.3), the tree stratum, comprises woody plants taller than 10 m. Stratum B, the shrub layer, corresponds to woody plants with a height of 10 m or less. Finally, stratum C, the herb and ground cover stratum includes herbaceous species which divides into forbs and ferns, some very low woody species, mosses, and seedlings. The maximum average height of this stratum is approximately 0.30 m. Each stratum was sampled with different sized quadrats that were nested within each other (Figure 4.3-b). 25 a) i i , b) A a o — Approx. 10 m B -Approx. 0.3 m c 10m A 4 m B B 2 m * c Figure 4.3. a) Vertical projection of stratified forest stand and b) in scale, simplified plan view of nested vegetation sampling quadrats, A: tree stratum, B: shrub stratum, and C: ground cover stratum. On the slopes adjacent to each gully, that were unaffected by debris flows, vegetation was often taller than in the gully areas. In these cases, 10 x 10 m (100 m 2) quadrats were used (Figure 4.3 b): A). In undisturbed stands or old-growth forests, tree diameters were considerably larger and under this type of cover a 20 x 20 m (400 m 2) quadrat was used to record the tree cover. For both the 100 m 2 and the 400 m 2 quadrats, nested smaller quadrats of 16 m 2 and 2 m 2 were used to complete the survey and record the lower vegetal strata. A quadrat of 1 x 2 m (2 m 2) was used to record herb and ground cover (Figure 4.3 b):C). This was nested within a 4 x 4 m (16 m2) quadrat used for the shrub layer (Figure 4.3 b): B). Since trees were nearly always absent from gully sidewalls, only the 16 m and 2 m quadrats were used for plant inventory on sidewalls. Within each stratum, all plant species were identified and the total percentage cover by each species was recorded using the Domin-Krajina scale (Table 4.1). The resultant cover values are semi-ordinal and were converted to their mid-class values for quantitative analysis. In this study, relative cover refers to the percent area of a plot covered by the vertical projection of the crowns of each plant species. The relative area occupied by bare soil, large rock, and woody debris was sometimes large enough to interfere with the vegetation growth and was noted and estimated in the 16 m 2 quadrats. Other observations at each quadrat included slope angle, slope aspect, moisture regime, and a surface soil sample for texture analysis. Vegetation cover in open slopes adjacent to the gullies was originally intended to be sampled using quadrats located along each transect line. However, due to the lack of variation in the open slope vegetation and the large quadrats required to accurately sample forest stands, quadrats were established on open slopes only when there was a visible change in the vegetation cover. Overall, two plots were established in the old-growth forest, six in second growth planted stands, six in relatively young and immature planted and naturally grown stands, and five on torrented areas outside the gullies. 26 Table 4.1. Domin-Krajina vegetation cover scale (from Mueller-Dombois and Ellenberg, 1974). Code Description Cover (%) Mid-class 10 complete cover s 100 100.0 9 more than 75% cover but less than complete >75 87.5 8 50% - 75% cover 50-75 62.5 7 33%> - 50% cover 33-50 42.0 6 25%> - 33%> cover 25-33 29.0 5 10% - 25% cover 10-25 17.5 4 5% - 10% cover 5-10 7.5 3 scattered cover < 5% 1-5 2.5 2 widely scattered < 1% <1 0.1 1 rare, insignificant cover + solitary Measurement of the non-vegetal attributes of quadrats, such as slope aspect, slope angle, moisture regime, area covered by rock, coarse woody debris, and bare soil, was conducted to permit analysis of the relations between vegetation communities and the underlying environmental gradients. Determination of the environmental factors controlling vegetation recovery in the gullies following torrent disturbance corresponds to the third objective of this study. Table 4.2. Numbers of quadrats and sediment traps in the various gullies. Total number of gullies 12 gullies Total Number of gullies per time treatment 6 years 19 years 18 Logged Unlogged Logged Unlogged Number of gullies per land treatment 6 4 5 3 18 Gully quadrats 25 21 20 8 74 Control quadrats 12 2 4 - 18 Sediment traps 4 6 6 2 18 27 Table 4.3. Organisation of gullies per treatment groups. Gully numbers Treatment Groups Gullies that torrented between 1975 and 1985 Logged gullies Unlogged gullies LI L3 L5 L6 L7 L8 UI U3 U7 U8a Logged 6 years Unlogged 6 years Gullies that torrented between 1990 and 1993 Logged gullies Unlogged gullies L2 L4 L9 L10 L13 U4 U8b U9 Logged 19 years Unlogged 19 years A total of twelve gullies were surveyed, five were logged, one was completely unlogged and six had been logged in their lower reaches only. This provided a total of 18 gully segments for land treatment comparisons, not counting control plots located outside of gullies (Table 4.2). The selection of partly logged gullies permitted a comparison of vegetation regrowth between land treatments for the same debris flow event, and for the same elapsed time since torrenting. This was the case for gullies 1, 3, 4, 7, 8 and 9. The total number of quadrats established in each treatment group is summarised in Table 4.2. A total of 46 different plant species was observed in the 92 quadrats. Appendix 1 lists all the species identified and their abbreviations. 28 4.4 Sediment trap monitoring 4.4.1 Method An important objective of this study was to evaluate sediment production from gully sidewalls at different stages of vegetation recovery. This was done by collecting eroded sediment on gully sidewalls with sediment traps under each of the major types of vegetation cover. Sediment traps, were 1.2 m in width built from three pieces of 10" x 1" lumber assembled in an open-box arrangement on the sidewalls (Figure 4.4). Traps were secured to the slopes with metal re-enforcement bars hammered into the ground, down-slope from the main board. In some cases, wire tie-backs were used. To maximise the contributing area of the sidewalls, traps were installed close to the base of a sidewall, but above the water channel's highest flow erosion lines. Figure 4.4. Typical 1.2 m sediment trap, showing a prism of accumulated sediment. A total of 18 sediment traps were installed under various cover types. A special effort was made to install replicate pairs of sediment traps to obtain some local variance of soil loss in environments. However, these apparently similar sites had somewhat different plant covers (see Table 5.5). Sediment traps were all established in October 1997, then re-surveyed in November 1997, May, August and November of 1998. 29 Initially, erosion pins were installed upslope of the trap to assist in monitoring the accumulation of sediment. When using erosion pins, the monitoring of soil loss or gain is obtained by measuring the change in length from the head of the pin to the ground surface. The change in length of the pin corresponds to either an accumulation or erosion of the surrounding soil, depending on the direction of change. In the present case, soil was often found accumulating primarily on the upslope sides of the pins. This is a common problem encountered with the method, and can be avoided by measuring the sides of each pin. However, frost heave also acted on many of the pins and moved them up or totally out of the ground. This made readings difficult or outright unreliable. Thus a decision was made to abandon monitoring by erosion pins. The initial installation of erosion pins upslope of the trap may have contributed to detachment of soil particles but was judged to be very insignificant in the total outcome. 4.4.2 Evaluation of sediment yield Sediment accumulating behind the main board formed a roughly triangular prism of accumulation (Figure 4.4 and 4.5). The volume of sediment was determined by evaluating the cross-sectional area of the accumulation prism multiplied by the trap width. Since the side traps were installed at 90° to the slope, the height of the triangular prism corresponds to the distance occupied by the accumulation prism along the front board of the trap (Figure 4.5). In the field, the distance between the top of the front board and the surface of soil accumulation was measured (x,). Height of the accumulation prism (X) was obtained by subtracting this length from an initial measurement, x0 (Figure 4.5). The change in height of the accumulation prism was measured at pre-determined points along the board, every 5 cm, and then averaged for the entire trap. No physical boundary or marker allows the distinction between newly accumulated and original soil at the base of prism of accumulation. Therefore, this length could not be measured accurately in the field, but angles measured on the accumulation surface provide enough information to determine this distance by trigonometric manipulations (Figure 4.6). The base of the accumulation prism can be evaluated with the trigonometric equation: base = height / tan (B) Equation 4.1 where height corresponds to X and B is the angle between the surface of accumulation and the sidewall (Figure 4.6). This angle can be evaluated with the relation: B = 0 - a, Equation 4.2 30 Figure 4.5. Evaluation of the height of the accumulation prism. Both angles 0 and a were measured in the field (Figure 4.6). The angle 0, was measured on the up-slope side of the front board of the trap, at pre-determined points, every 10 cm along the board, and was averaged for the trap. The angle 0 corresponds to the initial slope angle taken when the trap was installed. This initial angle is equivalent to the gully sidewall angle. All angles were measured using the clinometer of a "Brunton compass" with a precision of ±1° . As sediment accumulates within the trap, this angle slowly changes and a new angle called a, is measured at the same pre-determined points along the board. Once sediment had accumulated within the trap, the new sediment surface was measured using the angle a, which was measured at the same pre-determined points along the board. Figure 4.6. Schematic showing the prism of accumulation and the different angles. Angles 0 and a can be integrated in Equation 4.2 to obtain B which is used to find the base (Equation 4.2). Base and height of the prism of accumulation are then known and the area of prism can be averaged. This last value is then multiplied by the total width of the collecting area, which corresponds to the width of the trap, 1.20 cm, thus giving the total volume of eroded material collected in the trap. Appendix 2 presents the average angles (0, B) and height of 31 sediment (X), measured throughout the field season. A table summarising sediment yield is presented in section 5.4 (Table 5.5). 4.4.3 Evaluation of errors in the sediment yield estimation In the field, the height of the prism of accumulation, X, was measured with a ruler, with a precision of ± 1 mm. Besides the potential for error inherent in reading the ruler, the angle of the ruler along the board and small stones moving along the board introduced uncertainties in the final reading. To overcome this problem, measurements at one trap were triplicated in order to determine precision. The evaluation of the uncertainties in the measurement was determined at ± 5 mm. This means that imprecision on X is approximately ± 1 0 mm, since X corresponds to the subtraction of x, from x0. The initial error of ± 1° in the reading of the accumulation angle was re-evaluated in the same manner as for the height measurement and showed a mean error of ± 2°. Thus the B accuracy is ± 4°. The implication of these uncertainties in the measurements of the calculated sediment yields, and changes in measured values, within the limits of uncertainty, can modify the final sediment yield answer by a factor of one hundred. To avoid over-estimation of sediment yield by relying on imprecise angles a more conservative method for evaluating yield was adopted. This method assumed that the surface of the accumulation area in the trap was approximately level, therefore that a is equal to zero, and B = 6. Thus the error on B = ± 2 ° (Figure 4.7-b). This assumption was observed to be more realistic in many cases where the sediment collected was small. However where sediment production was large, the surface of accumulation sloped gently as represented in Figure 4.7a. a) V — b) Figure 4.7. Schematic showing the difference assumption a) and b) of sediment accumulation in the sediment trap and the difference in the angle of the prism, 9 and a measured angles and the angle determined from them, B. 32 Again, by eliminating the area of the sloping part of the accumulation prism, the sediment yield result is conservative, and will be underestimated in the same cases. However, this was preferable to a gross over-estimation of yield based on uncertain upslope extrapolation of angle a and 6. The fact that very few data of sediment yield exist for this terrain makes comparison of the values obtained in this study rather problematic. 4.4.4 Lab work on soil and eroded sediment samples Textural analyses were conducted on all soil samples following methods similar to those described in Gee and Bauder (1982). Initially the whole sample was air dried in the laboratory to reduce soil agglomerates and to obtain comparable bulk weights. The latter were obtained immediately after the drying process and the whole sample was then separated into fine and coarse components, separated at 2 mm. Sub-samples of approximately 30 g were taken for organic content analysis and more detailed textural analysis. Sub-samples were oven-dried overnight (approximately 12 hours) at 105°C, then weighed. Organic contents were based on loss on ignition tests at 550°C for 3 hours then re-weighed. The standard test time is 2 hours; however, very high organic contents in some of the samples necessitated an increase in the length of the ignition test. Sub-samples were sieved to separate very coarse sand (1000pm), coarse sand (500pm), medium sand (250um), fine sand (125pm), very fine sand (63pm) and silt and clay (<63um). Each part was weighed and stored. Appendix 3 gives a detailed listing of these textural results. 4.5 Analytical methods 4.5.1 Ordination methods for vegetation Multivariate analyses were used to determine the importance of environmental variables, as stated in the objective (1.3), in the plant re-establishment on gully sidewalls. Ordinations, or gradient analyses, are multivariate analyses commonly used to help generate hypotheses about the relations between the species composition at sites and their underlying environment controls (Digby and Kempton, 1987). Ordination analyses arrange sites along axes on the basis of species compositions (terBraak, 1995). Ordination results are often presented as a two- or three-variable diagrams. The closer together the points are, the more similar their species composition, and vice versa (ter Braak, 1995). Interpretation of an ordination diagram is based on knowledge of the environmental 33 controls. When controls are not explicit, interpretation is said to be indirect, and interpretation is based on the distribution of the species along the axes. On the other hand, if environmental variables were measured at each plot, direct ordination analysis is possible. This type of analysis is also known as canonical analysis, and uses species and environmental variables simultaneously to determine the positions of the sites along the axes (ter Braak, 1995). Correspondence Analysis Correspondence analysis, CA, is an indirect ordination analysis which assumes a unimodal distribution of a species along an environmental gradient. Species and site scores are based on the weighted averages of initial randomly allocated species and site scores, where weights are assigned according to the abundance of a species at each site. The whole process is iterative and new scores will be assigned to species and sites until small changes are detected in the newly generated scores. These results give the first axis of the ordination diagram, displaying sites and species having the highest correlations with each other along the axis (Gauch, 1982 and Pielou, 1984). The second and other axes are evaluated in the same way, but include an extra step to ensure that they are uncorrelated to any other axes (ter Braak, 1995 and Palmer, 1993). A value, called an eigenvalue, measures how well the axes (or eigenvectors) explain the species dispersion. Eigenvalues vary from 1 to 0, 1 being the highest value of correspondence, zero indicating no correspondence between sites and species (Palmer, 1993). Any value above 0.5 is normally considered to indicate acceptable correspondence (ter Braak, 1995), and on this basis, the most significant axes are then subjected to substantive interpretation. In general, the first axis usually shows the highest correspondence, followed by the second axis and so on. Detrended Correspondence Analysis Correspondence analysis is subject to two problems that often render interpretations problematical. The first is the presence of an "arch effect" in the data distribution on the diagram. The "arch" is the distribution taken by the scores along the first axis, and appears when the second axis values show a quadratic relation with the first axis. In reality, the second axis is only a folded version of the first axis and hides the true dispersion along the second axis (ter Braak, 1995). The second problem is due to compression at the end of the axes. Detrended Correspondence Analysis, or DCA, (Hill and Gauch, 1980) was developed to overcome these problems by detrending and rescaling the axes. Both procedures are described at length in Gauch (1982), Pielou (1984), Digby and Kempton (1987) and Kent and Coker (1992). The resulting DCA ordination diagram can be interpreted in the same way as CA diagram. DCA can be used to determine the gradient lengths in standard deviation units of species turnover (ter Braak and 34 Smilauer, 1998). Gradients of length greater than 4 standard deviations indicate a strong unimodal response (ter Braak and Smilauer, 1998) and meet the main analysis assumptions. Canonical Correspondence Analysis Canonical Correspondence Analysis, C C A , is a restricted form of C A because it integrates environmental variables into the determination of the ordination axes. Thus, the eigenvalues of the C C A are expected to be smaller than those obtained from C A . The use of environmental variables makes C C A a direct gradient analysis. C C A uses, in addition to the weighted averaging of species, a multiple linear least-squares regression which creates linear combinations of site scores (environmental variables and species scores) to determine the final site and species scores (Palmer 1993). Sites and species on the diagram have the same interpretation as C A and D C A , and environmental variables are indicated by arrows (ter Braak 1987) that point in the direction of maximum change. The length of an arrow corresponds to the rate of change of the weighted average, and measures how much the species distribution changes along an environmental gradient. Therefore, an environmental variable that plays an important role in species distribution would be represented by a longer arrow (ter Braak, 1986). A long arrow also indicates a good relationship between the environmental variable and the ordination axis, hence between sites and species (ter Braak, 1987). As in C A and D C A , the usual assumption in C C A is that each species follows a unimodal distribution when plotted along an environmental gradient, implying that a single optimal environmental condition exists for each species (Palmer 1993). C C A is a robust test and performs well even if the assumptions do not hold fully (ter Braak, 1987, Palmer 1993). Moreover, Palmer (1993) has demonstrated that C C A also performs well in conditions where environmental gradients or species are highly correlated, and that C C A is not affected by the arch effect which affects C A , which suggests that there is no need to use D C C A , the detrended form of C C A . Interpretation of C A , D C A and C C A diagrams can be done either with the environmental centroids method or the environmental biplot-scores method. For short gradients (shorter than 3 standard deviations), ter Braak and Smilauer (1998) suggest the biplot method as more appropriate. When using this method, one can draw a line from the species or site point intersecting the arrow at 90° with the environmental gradient arrow (Figure 4.8). To ensure that the point falls on the arrow, the latter can be extended on each side to form a line. The species or site weighted average rank along the environmental gradient arrow is also determined in this way (ter Braak, 1987). A high position on the gradient indicates a higher weighted average for this species along that gradient. For example, in Figure 4.8, Smils. shows a closer relation to the 35 gradient "aspect" than does Poly m. C A , D C A , and C C A analyses are explained at length in Gauch, 1982; Pielou, 1984; Palmer, 1993; Jongman et al, 1995; and Palmer, 2000. o m + Com c Petap • aspect ^Tiiuj p» moisture level 1 Smil s • I Maia d ^ , Care m 1 — # Tsug h -Anap m * \ Epil a J • Dice f • iGymn D Gram Arun d \ I Blecs • Gali t 9 • \ . — — 1» slope Dryo e * Aspl v t Q t a | C Q V e r Circ a * Tiar t 1 • Lact m 1 b-Samb 9 J Cryof • Taxub A thy f 1 • L Linn b • Thel p Poly m | Crip c • Aden b o • Pseu m n i i i 1 1 1 . 1 Figure 4.8 First and second axes of C C A diagram which show the biplot interpretation of species for the slope aspect gradient. For abbreviations see Appendix 1. 36 Chapter 5: Results This chapter is divided into four main sections. Section 5.1, briefly summarise the surveyed gully attributes. Section, 5.2, and its sub-sections, compare the different vegetation covers according to time and land treatment. Analyses are based on average values from groups representing the different treatments (Table 4.3). Percent bare soil and vegetation covers are compared using the Student's t-test, Jaccard's Similarity Index, species richness comparisons, and abundance curves using the Geometric model series (Pastor, 1995). Ordination analyses are used in section 5.3 to test whether vegetation recovery on sidewalls is controlled by specific environmental variables. Finally, the last section of the chapter, 5.4, presents results from sediment monitoring conducted in the gullies. Sediment yield values, their related environmental variables, and textural analyses of soil samples are presented. An ordination analysis is also conducted on the soil erosion results to determine whether species distribution explains sediment production. 5.1 Gully attributes Attributes include morphological data such as mean gully channel slope angles, lengths, gully sidewall mean slopes and mean lengths. In this study, gully channel length refers to the surveyed sections of each gully. These begin upslope of the access roads and extend well into the old-growth forest, or in some cases end close to it (Figure 3.2). Since some of the surveyed gullies do not exhibit surface flow over the whole year, they are not all integrated into the mapped drainage system obtained from the base map used to make Figure 3.2. The upper reaches of gullies are often incised into bedrock, and accordingly were not included in the study since they lacked colluvial materials on their sidewalls. 37 Table 5.1. Summary table of gully dimensions and covers for the surveyed section. Gully Channel Channel Sidewall Sidewall Sidewall Sidewall Sidewall Time Length slope length slope vegetation cover bare soil CDW cover elapse since last (mean) (mean) (mean) (mean) (mean) (mean) torrent (m) (°) (m) (°) (%) (%) (%) (year) Gully 1 410 26 4.8 39 72 29 2 Ll 250 18 5.1 36 99 14 0 1990-92 UI 160 33 4.5 41 44 45 4 1990 Gully 2 350 21 14.3 35 73 6 0 L2 350 21 14.3 35 73 6 0 1975 Gully 3 320 20 5.4 35 52 41 1 L3 220 19 3.7 35 46 44 0 1990 U3 100 21 7.1 35 59 38 1 1990 Gully 4 240 19 4.0 28 105 18 4 L4 160 14 4.0 23 111 29 1 1980 U4 80 24 4.4 32 99 7 7 1980 Gully 5 210 15 5.0 42 89 0 4 L5 210 15 5.0 42 89 0 4 1990 Gully 6 230 12 4.0 31 76 16 14 L6 230 12 4.0 31 76 16 14 1990 Gully 7 340 16 7.1 37 81 14 0 L7 240 16 6.2 44 103 13 0 1990 U7 100 15 8.0 30 59 15 0 1990 Gully 8 250 20 4.1 38 65 31 0 L8 140 14 2.6 44 75 25 0 1993 U8 110 25 5.6 31 55 38 0 1993 Gully 8b 150 25 5.4 45 128 0 6 U8b 150 25 5.4. 45 128 0 6 1980 Gully 9 210 25 6.1 25 112 10 12 L9 130 24 5.1 23 130 1 20 1980 U9 80 26 7.0 27 94 18 3 Gully 10 190 22 3.3 16 124 1 7 L10 190 22 3.3 16 124 1 7 1980 Gully 13 120 18 3.4 31 132 4 14 L13 120 18 3.4 31 132 4 14 1980 5.2 Plant cover variations according to major treatment groups Calculated means and standard deviations of each treatment and control group are summarised in Table 5.2. The raw data associated with these groups are presented in Appendix 4. Table 5.3 presents the Student's t-tests results based on Table 5.2. The pooled samples of logged and unlogged plots are compared across both time periods in Table 5.3a to d. Table 5.3e to h compares the treatment to the control groups. In Table 5.3, bold-face figures indicate that the tested variables are significantly different. Considering the high level of natural variability in the 38 gully environment, and the difficulty of obtaining vegetation samples representative of a large area, the probability of Type 1 error, a, was set equal to 0.10 in these tests. Comparisons over time within logged and unlogged groups are given in Table 5.3a and b. These results indicate significant changes in plant cover over time: as percent bare soil decreases, total plant cover increases (Compare Table 5.2a with b, and c with d). Recently torrented logged and unlogged gullies are compared in Table 5.3c. For both groups, results show that bare soil and total cover are significantly different in the two land treatments. Finally, Table 5.3d shows that the vegetation covers in the LI9 (Logged 19 years) category are not significantly different from those found in U19 (Unlogged 19 years). This result suggests that the impact of land treatment on plant cover progressively declines over time. A more detail discussion of the different cover types observed in the vertical stratification of the treatment and control groups is presented in section 6.1. 39 Table 5.2. Means (u) and standard deviations (a) for percent bare soil, other cover ( C W D and rocks) total plant cover (%), and cover per vegetation strata: tree, shrub, herb and ground cover (cover lower than herb) across treatment and control groups. Treatment groups Variables CT (a) Bare soil 21.5 28.4 Logged 6 yrs Other cover 17.2 17.2 (N=25) Total cover 77.4 43.0 Tree - -Shrub 27.1 29.6 Herb 28.4 17.8 Ground 21.8 24.5 (b) Bare soil 8.8 13.7 Logged 19 yrs Other cover 12.5 14.5 (N=20) Total cover 113.8 44.0 Tree 13.8 28.0 Shrub 48.8 32.9 Herb 38.8 20.9 Ground 12.5 19.9 (c) Bare soil 33.8 32.0 Unlogged 6 yrs Other cover 10.0 8.0 (N=18) Total cover 57.4 33.4 Tree - -Shrub 18.7 20.3 Herb 32.2 19.5 Ground 6.5 9.4 (d) Bare soil 6.8 15.1 Unlogged, 19 yrs Other cover 6.7 6.4 (N=ll) Total cover 111.0 34.3 Tree 9.5 26.6 Shrub 36.2 35.7 Herb 54.6 21.4 Ground 10.7 16.3 Control groups Variables CT (e) Bare soil 6.4 8.20 Torrented Other cover 20.4 18.9 open slope Total cover 99.8 27.8 (N=5) Tree - -Shrub 64.3 21.1 Herb 31.4 10.5 Ground 4.1 1.31 (f) Bare soil 4.33 7.9 10-year Other cover 14.0 10.9 second growth Total cover 110.4 59.0 (N=6) Tree - -Shrub 79.7 38.7 Herb 22.0 19.1 Ground 8.8 16.5 (g) Bare soil _ . 20-year Other cover 33.2 22.7 second growth Total cover 114.7 47.4 (N=5) Tree 64.5 24.0 Shrub 39.3 23.5 Herb 6.42 5.1 Ground 4.52 2.8 (h) Bare soil - -Old-growth Other cover 0.5 0.7 (N=2) Total cover 213.0 36.9 Tree 62.0 25.5 Shrub 73.8 51.3 Herb 68.4 1.1 Ground 8.85 12.2 40 Table 5.3. One-tailed Student's t-test summary for percent bare soil, other cover ( C W D and rocks) total plant cover (%), and cover per vegetation strata: tree, shrub, herb and ground cover (cover lower than herb) across treatment and control groups. Only significant P-values are shown in bolded-face type. Treatment Comparison Variables df t-value P-value fa) Bare soil 36 1.97 0.03 Logged 6 yrs vs. Other cover 43 1.00 0.16 Logged 19 yrs Total cover 40 -2.79 0.00 Tree 19 -2.20 0.02 Shrub 39 -2.29 0.14 Herb 37 -1.77 0.04 Ground 43 1.42 0.08 (b) Bare soil 26 3.07 0.00 Unlogged 6 yrs vs. Other cover 25 1.22 0.12 Unlogged 19 yrs Total cover 21 -4.12 0.00 Tree 10 -1.46 0.09 Shrub 14 -1.48 0.08 Herb 20 -2.83 0.01 Ground 14 -0.78 0.22 (c) Bare soil 34 -1.30 0.10 Logged 6 yrs vs. Other cover 36 1.85 0.03 Unlogged 6yrs Total cover 41 1.71 0.05 Tree 0 _ Shrub 41 1.10 0.14 Herb 35 -0.64 0.27 Ground 33 2.85 0.03 fd) Bare soil 19 0.36 0.36 Logged 19 yrs vs. Other cover 28 1.54 0.07 Unlogged 19 yrs Total cover 25 0.20 0.42 Tree 26 0.47 0.32 Shrub 19 0.97 0.17 Herb 20 -1.99 0.03 Ground 24 0.26 0.40 Table 5.3 continues next page Table 5.3. (cont'd) Treatment Comparison Variables df t-value P-value Ce) %bare soil 24 2.24 0.02 Logged 6 yrs vs. Other cover 5 -0.34 0.37 Torrented open slope Total cover 8 -1.48 0.09 Tree 0 Shrub 8 -3.33 0.01 Herb 9 -0.50 0.31 Ground 25 3.59 0.00 ff) %bare soil 28 2.63 0.01 Logged 6 yrs vs. Other cover 12 0.58 0.28 Ten years second growth Total cover 6 -1.29 0.12 Tree 0 -Shrub 6 -3.11 0.01 Herb 7 0.75 0.24 Ground 11 1.56 0.07 (K) %bare soil 10 0.50 0.31 Logged 19 yrs vs. Other cover 5 -0.87 0.21 Torrented open slope Total cover 10 0.89 0.20 Tree 19 2.20 0.02 Shrub 10 -1.30 0.11 Herb 13 1.12 0.14 Ground 20 1.87 0.04 (h) %bare soil 15 1.00 0.17 Logged 19 yrs vs. Other cover 11 -0.26 0.40 Ten year second growth Total cover 7 0.13 0.45 Tree 19 2.20 0.02 Shrub 7 -1.77 0.06 Herb 9 1.85 0.05 Ground 10 0.45 0.33 (i) %bare soil 19 2.86 0.00 Logged 19 yrs vs. Other cover 5 -1.94 0.06 Twenty year second growth Total cover 6 -0.04 0.49 Tree 7 -4.09 0.00 Shrub 8 0.74 0.24 Herb 23 6.24 0.00 Ground 21 1.73 0.05 (j) %bare soil 10 1.50 0.08 Unlogged 19 yrs vs. Other cover 11 3.13 0.00 Old-growth forest Total cover 1 -3.63 0.09 Tree 1 -2.74 0.11 Shrub 1 -0.99 0.25 Herb 10 -2.12 0.03 Ground 2 0.19 0.43 42 A debris flow responsible for scouring the gully sidewalls might also scour the adjacent slope if the torrent spills out of the channel (Figure 4.1:Q2). Results comparing L6 (Logged 6 years) gullies and TOS (torrented open slopes) (Table 5.2a and e, and Table 5.3e) showed that the adjacent open slopes had a lower percentage cover of bare soil and higher vegetation cover, indicating that they recover more rapidly than the sidewalls. In L19 gullies, the difference between torrented open slopes and sidewalls was less notable (compare Table 5.2b with e and Table 5.3g). Overall, for similar elapsed times since debris flow, gully sidewalls show higher percent bare soil than the adjacent slopes. (Table 5.3e and f). Total percent bare soil and vegetation cover in the 10 II (ten-year second growth) and L19 were similar. However, the vertical structure of the vegetation types differs (Table 5.3h). The 20 II (twenty-year second growth) and OG (old-growth) compare very poorly with their corresponding gully groups (Table 5.2b, g with 5.3i and 5.2d, h with 5.3j). The twenty-year second growth group represents a more mature stand, even though the elapsed time is similar for this group and the L19 group. Old-growth plots were not expected to be comparable with the U19 gullies, but were sampled to obtain an indication of the type of plant cover found in a final successional stage. To better compare the communities observed in gullies, associations can be defined from the species dominating the cover in each treatment group (see Appendix 4a). Hence, L6 shows a Rubus parviflorus - Thuja plicata - Boykinia elata association, which evolves into a Thuja plicata - Rubus parviflorus - Boykinia elata association in L19. This represents a very small change in the community with only a shift in the dominant and co-dominant species. The U6 the community is dominated by a Boykinia elata - Rubus spectabilis-Tsuga heterophylla association to become in U19 a Blecchnum spicant - Tiarella trifoliata - Dryopteris expansa association. This environment is clearly dominated by herb species rather than the shrubs of the logged gullies. Tables 5.2 and 5.3 give information about, "other cover", which refers to large rocks or bedrock and coarse woody debris ( C W D ) . This type of cover is not expected to change rapidly over time but might exhibit variation between land treatments. This other cover category shows larger values in the logged gullies where disturbance from logging has left wood slash and other C W D . Note, that toward the headwall area, often in the unlogged section of a gully, bedrock is more common on the sidewalls and in the channel, but survey of a gully stopped before, or as these areas were encountered, since these gully sections did not meet the requirements of the experimental design (see section 4.2). 43 5.2.1 Floristic comparisons The similarity in species composition between sites was tested with Jaccard's Similarity Index (Mueller-Dombois and Ellenberg, 1974) (Table 5.4). The index is obtained by dividing the number of similar species in two sites by the total number of species. The values obtained in this study vary between 26 and 70 %. As in Hull and Scott (1982), similarity is determined by a level exceeding 50%. The highest similarity values are found between the different land treatments, within the same time interval, with 70% after 6 years and 69% after 19 years. The lowest similarity is observed between L6 and U19, where only 48 % of the plant species are common to both sites. Except for the torrented open slope (TOS) category, which shows a 67% similarity with L6, there is poor similarity between the gully and control groups. Note that the old-growth ( O G ) category shows a slightly higher similarity level (52 % and 55 %), with the unlogged gullies than with any other groups. Within the controls, there is a low degree of similarity, with values ranging from 26 to 45 %. Table 5.4. Index of similarity in species composition, in percentage, according to Jaccard Index of Similarity (Mueller-Dombois and Ellenberg, 1974). Treatment and control groups L6 L6 100 L19 L19 63 100 U6 U6 70 62 100 U19 U19 48 62 69 100 TOS TOS 67 48 54 41 100 10II 10 II 32 35 36 40 26 100 20 II 20 II 35 41 36 44 33 46 100 O G O G 42 41 52 55 27 45 45 100 5.2.2 Species richness Species richness refers to the average number of species inventoried per treatment group. In Figure 5.1, richness was averaged over treatment and control groups for the herb and shrub strata. Tree cover was not considered because of its negligible presence in gullies, and in any 44 case does not add to richness. A total of 30 herb species and 16 shrub species were identified on the different sites. Although total plant cover was shown to be higher in L6 gullies than in U6 gullies (see Table 5.2), herb richness was marginally lower in L6. In the shrub stratum, species richness is comparable between the L6 and U6 gullies. L6 gullies have an average of 22 herb species and 12 shrub species compared with an average of 24 herb species and 11 shrubs species in the U6 gullies. Knowing that a total of 30 herb species and 16 shrub species were found in all plots, there is a difference of 7 % in herb richness between L6 and U6, and 6 % in shrub richness. Herb richness changes differently between logged and unlogged gullies over time. In the L19 gullies, herb richness increases slightly from 22 to 26 species, a change of 13%. On the other hand, in the unlogged gullies herb richness decreases by 17%. Shrub richness increases in both logged and unlogged gullies over time. In logged gullies, 12 species are recorded 6 years after disturbance, while 14 species are found 19 years after disturbance. For unlogged gullies, the change is smaller with a gain of one species after 19 years. For the older gullies, results indicate that L19 gullies have a higher richness in the herb and shrub strata than do U19 gullies. In the control groups, richness differs between the herb and the shrub strata. The highest herb richness is observed on the TOS, and the highest shrub richness occurs in 10II. The poorest herb environment is observed in second growth stands planted 10 and 20 years after logging. The poorest shrub environment is found on the torrented slopes 7 years after disturbance by debris flow. Figure 5.1 Mean species richness of a) herb and b) shrub strata, for different treatment and control groups. The mean number of species in each group is indicated above each column. 46 5.2.3 Diversity-abundance curves Abundance curves, or dominance-diversity curves, show the abundance of plant species forming a community and are used to compare community structures (Bastow, 1991). Graphs are obtained by plotting relative abundance of species against the sequence of species, which is the ranking of species when ordered from most abundant to least abundant. In this study, abundance graphs were created using the total mean cover of the herb stratum. Two factors motivated the use of this stratum. Firstly, since species cover is used as the measure of abundance, it is not possible to compare herb species to shrubs or trees due to their differences in size. Thus comparisons must be done within a stratum. Secondly, of all the vertical strata, the herb stratum is the first to colonise new sites, hence responds most quickly to disturbance. The dominance-diversity curves for all treatment and control groups are illustrated in Figure 5.2. A Kolmogorov-Smirnov goodness of fit test was conducted to compare level of similarity between the species distributions in each group (Table 5.5). The results indicate no significant differences when comparing species distributions of L6 to other logged and unlogged gullies. However, comparisons of the other gullies do not show very similar species distribution or a significant difference. Community distributions are very similar between treatment and control groups with the exception of the second growth 10- and 20-year control groups. This is probably due to the fact that these latter groups are planted stands. Comparisons of the treatment groups with the controls shows a low to very poor fit for almost all control environments except in the case of TOS, which shows a similar distribution to the U19 gullies. Figure 5.3 uses the Geometric Series model to compare the changes of the species distributions in time as outlined by Tokeshi (1993). The Geometric Series model applies to a situation where few species dominate a large portion of the available resources. This model is based on preemption of available resources by species (Pastor, 1995). In Figure 5.3, the Geometric Series distributions follow straight-line patterns, and over time, the slopes of the lines becomes flatter in logged and unlogged gullies. This indicates a better sharing of the resources by different species in the older gullies. There is a few other abundance model that can be compared to the species distributions. The other most commonly used models are the Log Series, Log-Normal and Broken Stick model. Attempt to fit these models to the changing communities can be done and equation for the different model can be found in Magguran, 1988 and Bastow, 1991. However, in this case, only the Geometric model 47 was used as a comparative tools in between communities of the different treatments (Bazzaz, 1975; and Tokeshi, 1993). > o o C O a. o 100 -i 10 H o.i H 0.01 13 •• n U6 E! " 'ill ' Second growth (20yrs) Torrented open U19 (lOyrs) slopes (lOyrs) O a L6 sL19 10 15 20 Sequence of species 25 30 Figure 5.2. Observed distribution of mean percent cover for herb species in all treatment (a) and control groups (•). Table 5.5. Kolmogorov-Smirnov goodness-of-fit tests (for continuous distributions), indicating the level of similarity between the distributions of observed mean percent herb cover of treatment and control groups. Treatment and control groups L6 L6 1.000 L19 L19 0.921 1.000 U6 U6 0.740 0.522 1.000 U19 U19 0.921 0.197 0.334 1.000 TOS TOS 0.522 0.108 0.197 0.740 1.000 10 II 10 II 0.001 0.000 0.000 0.026 0.055 1.000 20 II 20 ii 0.002 0.000 0.000 0.026 0.026 0.334 1.000 OG OG 0.026 0.001 0.002 0.334 0.522 0.522 0.197 1.000 48 0 5 10 15 20 25 30 Species sequence 0 5 10 15 20 25 30 Species sequence a Observed 6 yrs Geometric 6 yrs 0 Observed 19 yrs Geometric 19 yrs Figure 5.3. Abundance curves of treatment groups represented by their observed distribution compared with expected distributions of the Geometric Series model (Pastor, 1995). 49 5.3 Ordination analysis results The program CANOCO 4.0 (ter Braak and Smilauer, 1998) was used for the ordination analyses. The 92 vegetation quadrats were averaged per gully and treatment group in order to simplify the interpretation of the gradient analysis. Bare soil (soil in Figure 5.4) was included in the ordination analysis with species to observe its multivariate "location" and its relation with sites and species. Results indicate that this variable is associated with logged sites. D C A results are presented in Figure 5.4. and Appendix 3. The eigenvalue of the first axis was 0.445 and for the second axis was 0.307. The total sum of the eigenvalues was 2.500. Hence the first two axes explain 30% of the site and species variation. The D C A diagram shows that site scores are relatively close together, which suggests a very low site differentiation by species composition. Hence, small differences exist in the community composition for the different land and time treatments. However, if gullies are grouped by treatments (logged and unlogged) and elapsed time (6 and 19 years) (see polygons in Figure 5.4), it is possible to observe some migration of groups along the axes. Logged sites are closer to the origin of the first axis than are unlogged sites. Similarly, along the second axis, the most recently disturbed gullies (L6 and U6) are closer to the origin than are older gully stands. Control sites show a greater dispersion, reflecting their environmental variability. Note that the TOS control is close to the L6 gully sites, suggesting similar composition (Table 5.4). A plant species succession is visible along the first axis (Figure 5.4). Going from the left to right along the axis, one sees that fast colonising species (found on the left of the diagram) are slowly replaced by climax species and/or better competitors (found to the right). Climax species refers to the final stage of vegetation development in the biogeoclimatic zone of coastal western hemlock. A similar successional pattern in shade tolerance is also visible, with shade intolerant species on the left giving way to more shade tolerant species on the right. This pattern is reflected in land treatment, with logged gullies on the left and unlogged on the right. The distribution along the second axis is not as clear; however species that appear at each end of this axis are can be differentiated by their life cycle. Species distributions seem to suggest a small effect of "colonising time", in other words, that species distributions might reflect the time for species to become established at sites. The position of logged and unlogged gullies reflects this, with the earlier disturbed gullies located above the more recently disturbed ones on the diagram. Overall, the distribution of sites along both axes suggests a progression from more open stands to a condition where less light reaches the ground due to interception by a high canopy. 5 0 Axis 2, Aden b b-Acer d-Tsug d-Thuj • Galit Circ a ~ „ „ „ Clays • C a r e m . Petap Trifr Maia d Smil s Thuj p o Tsug h Dryo e b-Kr-'Ble.cs -bcre b-Abie • C o m c Stre spp Viol spp • Tiart • 2< • Linnb ° b-Vacp Gymn D b-Tsug Thelp # Taxu b Bryo» Crip c ^ • « b - A l n u ^ C r y o C v B o y k e O ° • o " Gram" •Soil "Epi l a O . ~* »Lact m b-RubS D _ A ™ d * * b - R o s a n Anap m • * b-Vac spp Axis 1 b-RuPa Dice f A t . h y f . . P o l y m ' ^ H i Aspl v b-Samb •b-Pseu » b-Cham Pseu m 2 . 5 +4 . 5 Figure 5.4. Axis 1 and 2 of Detrended Correspondence Analysis diagram. L6 (o), L19 (O), U6 (•), U19 ( • ), TOS (+), 10II ), 20II ), OG ( X) , and species scores (•). For abbreviations see Appendix 1. Details of the analysis are presented in Appendix 6. The D C A results also provide information on the gradient lengths, 2.87 for the first axis and 2.43 for the second axis. The gradient length indicates that there is no great differentiation of sites by species composition (ter Braak, 1995). Such-gradients indicate low unimodality in the species distribution and suggest the use of analysis which assumes linear responses, such as Redundancy Analysis, R D A (the canonical version of Principal Component Analysis, P C A ) . A test analysis using R D A (Appendix 6) showed that the site and species distributions were very similar to the ones obtained in the C C A analysis (Figure 5.5 and Appendix 7). Explanations of the variance by the C C A and R D A are relatively similar; however C C A shows a larger spread in the data than does R D A which renders it easier to interpret. However, C C A eigenvalues and the spread of the site and species scores are significantly reduced when compared to those of R D A . 51 Furthermore, Palmer (1993) showed that C C A was a robust analysis and performed well even when the unimodal assumption does not hold. For these reasons, it was decided that C C A presented comparable results in a diagram that was easier to interpret. C C A analysis (Figure 5.5) eigenvalues are for the first axis was 0.32 and for the second axis was 0.23. The sum of canonical axes was 0.64, and therefore those two axes explained 86% of the variance of the canonical diagram. CN + IT) d-Tsug d-Thuj Axis 2 Dicef Soil • b-Vac spp Crip c Boyk e SLOPE Cryo f Care m • Bryo Stre spp , b - ° P l 0 Taxnb Thelp b-Rosa • b-Taxu Maia d Smil s D i y o c Violspp Gymn D b-RuPe Poly m Thuj p T s u S h • Alhyf Blecs b-Tsug TIME b-Alnu c l a y s Anap m b-RubS LOGGING b-Acer b . R u P a Gali t . • Epil a Com c b-Samb " L i n n b b-Vacp b-Abie Axis 1 0 Petap b-Thuj b-Pseu Aspl v b-Cham Pseum 2.0 +4 . 0 Figure 5.5. Axis 1 and 2 of Canonical Correspondence Analysis diagram. Arrows indicate the main environmental gradients. L6 (o), L19 (O), U6 (•), U19 ( • ) , T O S (+) , 10 n ( ) , 20 ii ( ) , O G ( X) , and Species scores (•). For abbreviations see Appendix 1. Details of the analysis are presented in Appendix 7. The selection of significant environmental variables is done with manual and forward selection of variables in the program C A N O C O 4.0 (ter Braak and Smilauer, 1998). The level of significance of the environmental variables is found with a Monte Carlo Permutation test, which identifies the environmental variables that best explain the species distributions. This test 52 provides a probability of Type 1 error, a, set at 0.10 and a calculated F value. The test proceeds by randomly selecting sub-samples of all quadrats to test the relations between environmental variables and species. Selected significant environmental variables are then used in the ordination to explain the species distributions. The variables T I M E , which corresponds to elapsed time since disturbance, L O G G I N G (the land treatment), and S L O P E corresponding to sidewall slope angle, provided the best explanation for site and species score distributions (Figure 5.5). The T I M E gradient was correlated at 0.87 with the first axis, the L O G G I N G gradient at -0.83 with the second axis, and the S L O P E gradient at 0.59 with the second axis (Appendix 7). T I M E and L O G G I N G effects were observed and discussed in the D C A analysis (Figure5.4). The S L O P E variable plays a role in species distributions, and consequently affects site scores. Although the S L O P E gradient direction is influenced by the low slope angles associated with the adjacent areas, unlogged sites that recently experienced a debris flow display high weighted averages along this gradient. This is mostly due to scouring from debris flows, which would be less as one gets closer to the depositional area. In this study, the mean sidewall slope angle in logged gullies is 32°, and 38° in unlogged gullies, which is significantly different with a P = 0.03. This is partially explained by the experimental design, where the unlogged sites were mostly found in upper section of the gullies, in contrast to logged gullies which where located meanly downstream. Therefore, it is possible that within the pool of logged gullies some did not experience scouring as deep as most the unlogged gullies. It is also possible that deposition of debris flow material occurred in the lower reaches of many of the logged gullies. Other variables such as slope aspect, soil moisture, C W D (coarse woody debris), and rock cover were not identified as significant controls during the forward and manual selection in the Monte Carlo Permutation Test (Appendix 8). As for the bare soil variable, it was integrated with the species matrix to test its association with the different sites. The soil score showed the closest association with the U6 group. This score is also situated in the upper section of the S L O P E gradient; however, there was no correlation between these two variables (Appendix 9). 5.4 Soil erosion results from sediment traps As mentioned and described in sections 4.4.1 to 4.4.4, four gullies were instrumented to measure soil loss from their sidewalls. A total of 18 plots were used including site replicates. The objective was to obtain characteristic soil erosion values for each plant successional stage. Sediment trap locations were therefore systematically designated, to obtain sediment yield values 53 under various vegetation covers, while encompassing the different time and land treatment combinations. Table 5.6 presents the site variables measured at the eighteen sediment traps and the sediment yields. Sediment yields are the total sediment collected in each sediment trap divided by the estimated contributing area, re-expressed as m 3 per hectare (ha) per year equivalents. The contributing area of a trap corresponds to the area of sidewall upslope of the trap (Appendix 10). Percent bare soil and vegetation cover were determined from quadrats established over the contributing areas. Gravimetric moisture was measured from soil samples collected at the trap sites in August, 1998. Measurements of all other variables are described in Chapter 4. Table 5.6. Variables associated with sediment traps and their contributing areas. Sediment trap sites+ Sediment yield (m3/ha/yr) Standard error (m3/ha/yr) Elapse time * Slope (°) Bare soil (%) Aspect o Plant cover (%) Gravimetric moisture factor (%) L2-1 1.24 0.036 22 45 7 135 99 6 L2-2 0.19 0.155 22 45 3 135 101 6 L2-3 8.62 0.015 22 42 10 220 50 2 L2-4 2.39 0.013 22 38 7 220 58 4 L3-5 4.02 0.050 7 36 30 5 71 5 L3-6 0.09 0.022 7 40 10 6 58 5 L3-7 26.68 0.002 7 37 100 220 0 21 L3-8 71.51 0.001 7 36 100 240 0 10 L4-1 4.30 0.007 12 48 35 15 28 7 L4-2 0.61 0.008 12 41 30 24 97 6 U3-1 13.22 0.002 7 35 100 18 0 3 U3-2 24.39 0.008 7 45 100 245 0 3 U3-3 2.42 0.001 7 57 5 100 112 7 U3-4 0.26 0.010 7 39 25 60 60 6 U4-3 1.14 0.004 12 41 15 130 80 6 U4-4 2.62 0.004 12 48 7 190 124 8 U8-1 1.16 0.004 4 41 25 120 48 7 U8-2 1.30 0.007 4 44 33 270 40 7 Sediment trap names are derived from the gully number (L2) and a trap number in the gully (-1), hence L2-1. * Years elapsed since last debris flow event and sampling year, 1997 Sediment yield varies greatly from one trap to another (Table 5.6). The highest yield values (L3-1, L3-2, U3-7, and U3-8) are associated with 100% bare soil areas in logged and unlogged gullies. The lowest yield values (L2-2 and L3-6) correspond to very low percent bare 54 soil. However, other traps showing similar percent bare soil, yield significantly larger amounts of sediment (L2-1, L2-3, L2-4, U3-3 and U4-4). Like percent bare soil, total vegetation cover does not show a simple relationship with sediment yield. Although most traps associated with lower yield values have good vegetation cover, the opposite is not true: traps with similar percent vegetation cover can yield widely varying quantities of sediment (e.g.: compare traps L2-4 and L3-6). Other variables clearly play an important role in determining soil erodibility at specific sites. Slope is an important variable in soil loss variability (Bovis, 1982), however in the current study, sidewall slope angles were not greatly variable, and hence did not correlate well with sediment yield (Appendix 11). Another potentially important variable is slope aspect, which acts indirectly on vegetation cover and soil moisture, by determining the slope exposure to solar energy. High levels of moisture in the ground can provide better cohesion of sediment on the slope, and thus limit the ravelling of material. Table 5.6 indicates fairly constant values of gravimetric soil moisture that were mostly between 2 % and 10 %. The highest value of 21 % was from L3-7, which had a large sediment yield. For other traps the variation in the moisture factor does not explain much of the sediment yield variability. Soil characteristics, such as soil texture, organic and mineral content, are also important variables to consider when assessing soil erodibility. The upper section of Table 5.7 presents mean values for samples collected from sidewalls, and the lower portion shows the results for samples collected from the sediment traps. Overall, sediment trap material had a higher organic content, perhaps because such material tends to be trapped and concentrated in them. Results from analyses of soil from sidewalls and sediment traps (Table 5.7 and Figure 5.6) show that logged gullies had a higher proportion of mineral soil than unlogged gullies. This difference was small in recently torrented gullies, but increased over time. Since organic content values are obtained from the weight lost after ignition of organics, they correlate inversely with mineral content. For both surface soil and sediment traps, coarser material concentrations were higher in the logged gullies. The tendency is for coarser material to be higher in the 6 elapsed year gullies of each treatment group, but then decreases at a faster rate over time in the unlogged gullies (see 1000pm per opposition to <63um). Direct comparisons of soil sampled on gully sidewalls with that in the traps indicated that materials smaller than 63 pm are less mobile than the coaser fractions (Figure 5.6). This is not surprising, since silt and clay materials have better cohesion than sand (2000 - 63pm). Also, finer material is more prone to being trapped in small voids, whereas coarser particles can roll downslope more easily. 55 Table 5.7. Soil textural properties for surface soil samples collected in gully quadrats and soil sampled in the sediment traps. Young refers to 6 years and Older to 19 years after disturbance. Organic Content Mineral <2000um <1000um content >1000um >500um <500um >250um <250um >125um <125um >63um <63um Sample from : (%) (%) (%) (%) (%) (%) (%) (%) Gully quadrat Young 11.0 89.0 20.5 20.6 17.3 16.6 9.4 14.5 Logged Older 18.1 81.9 12.4 16.9 17.0 21.6 11.8 18.9 Total 14.5 85.5 16.4 18.7 17.2 19.1 10.6 16.7 Unlogged Young 16.8 83.2 13.6 17.2 17.6 21.8 10.2 18.0 Older 27.6 72.4 7.6 12.2 14.2 20.8 16.2 27.2 Total 22.2 77.8 10.6 14.7 15.9 21.3 13.2 22.6 Sediment trap L2-1 16.9 83.1 5.0 10.3 16.1 20.4 22.8 22.0 Logged L2-3 6.5 93.5 23.4 25.2 19.3 12.4 9.1 8.7 L2-4 14.0 86.0 27.3 22.0 17.3 12.1 8.3 9.3 L3-5 16.9 83.1 13.9 17.2 22.6 17.2 13.7 13.7 L3-7 4.1 95.9 26.9 25.8 17.3 11.5 7.9 9.3 L3-8 3.8 96.2 25.5 21.1 16.3 13.7 12.4 9.8 L4-1 18.5 81.5 5.8 15.4 18.4 18.4 16.6 19.1 L4-2 13.7 86.3 24.4 23.3 19.0 12.2 9.7 10.7 Unlogged U3-1 10.2 89.8 19.2 20.6 17.9 15.6 13.0 13.4 U3-2 21.1 78.9 13.2 20.9 21.8 18.5 13.6 12.0 U3-3 61.7 38.3 2.6 20.5 21.6 9.8 8.1 13.6 U3-4 34.4 65.6 20.9 22.9 17.8 10.8 6.9 7.3 U4-3 30.0 70.0 8.4 10.7 17.5 14.2 17.0 24.0 U4-4 57.2 42.8 17.7 32.1 12.5 9.6 4.6 3.6 U8-1 35.2 64.8 14.5 29.4 19.5 10.6 9.1 7.4 U8-2 38.2 61.8 23.7 26.7 10.0 7.2 5.6 5.7 Figure 5.7 shows a semi-log graph of sediment yield as a function of percent bare soil. In the graph, different symbols were used to illustrate the different times elapsed since the last debris flow. The variation in the data for all times demonstrates the difficulty of grouping values per time period. Moreover, the lowest yield values do not necessarily correspond with the sites that experienced debris flow the earliest (22 and 12 years compared to 4 and 6 years). Even though the data exhibit variation, the correlation of percent bare soil and sediment yield is relatively high with a value of 0.74. Correlation between vegetation cover and sediment yield is lower than with bare soil, with a value of -0.62. Other variables do not show good correlations with sediment yield (Appendix 11). Figure 5.8 shows sediment yield as a function of vegetation cover. Sediment yield is higher in sediment traps located in the logged gullies. The land treatment variable within vegetation cover explains sediment yield relatively well with R 2 i o g g e d = 0.59 and R 2 u n i ogged = 0-76. Two traps (Figure 5.8: outliers) were not included in the regression of the unlogged sites. These 56 traps showed a large amount of organic material which was not observed in other traps (see Table 5.7, U3-3 and U4-4). 100.01 90.0 -I 80.0 70.0 • 60.0 • 50.0 40.0 i 30.0 20.0 • 10.0 -0.0 V ** *** **** Logged 6 years I Sampled ^Trapped Logged 19 years H Sampled FJ Trapped Unlogged 6 years • Sampled D Trapped Unlogged 19 years CD Sampled O Trapped **** Mineral Organic 1000pm 500pm content content 250pm 125pm 63 pm <63pm Figure 5.6. Bar graph comparing, for all treatment groups, textural analysis results of soil samples from sidewalls and sediment traps. T-tests were conducted on samples from sidewalls and sediment traps. Asterix indicates a significant difference between the samples, * =0.1,** = 0.05, *** = 0.01, **** = 0.001. CCA was used to verify whether species distributions could explain the observed sediment yield gradients (Figure 5.9 and Appendix 12). The CCA was used to analyse the sediment yield data to see i f it was possible to associate vegetation with sediment production, or if high sediment yield sites were associated with another environmental gradient. In this analysis, sediment production values were treated as an environmental variable. The variables that best explained the species and site distributions were organic content, sediment yield, slope aspect, stone cover, and herb-moss cover, coniferous shrub cover, and deciduous shrub cover. The last three variables are obtained from the breakdown of total vegetation cover into vertical strata. These variables were used because they provided a better explanation of the total variance than did total vegetation cover. 57 100.00 © y = 0.67e' ,0.04x A 4 years ^ 7 years n 20 years 22 years 0.10 4 0.01 0 20 40 60 80 100 120 Bare soil (%) Figure 5.7. Sediment yield determined from sedimentation traps vs bare soil, including elapse Organic content was the variable that best explained the sites and species distributions along the first axis of the diagram. Species associated with disturbance are found at the left end of this gradient. The gradient was independent of the sediment yield gradient. This last gradient was correlated with the second axis and plant species that were strongly associated with sites of high sediment yield (Figure 5.9). Presence of Smilacinia stellata, Pseudotsuga menziesii ssp. menziesii and Thuja plicata at the highest position on this gradient indicates some skewing in the species scores due the fact that these species were recorded only at sites with high sediment yield. Plant species common to open disturbed areas also show a high weighted average with the Sediment yield gradient. Finally the species found at the ends of the gradient are mostly successional species associated with longer times of development. Note that coniferous shrub and moss-herb gradients are the vegetation covers that show the highest negative correlations with sediment yield (Figure 5.9 and Appendix 12). time since last debris flow. 58 100.0 10.0 1 3 iS 00 • Logged • Unbgged Outliers 20 40 60 80 100 Vegetation cover (%) 120 140 Figure 5.8. Sediment yields (m3/ha/yr) as a function of vegetation cover (%) for logged and unlogged land treatments. 59 Thuj sediment yield Axis 2, •Pseu m Smil s aspect rocks V b-RuPa Anap m Epi l a* b-Alnu dec-shrub ° r g a m C b-Acer T h e l p • T a x u b Poly m •Stre spp L3TR6 R r v n ^ A r a n d Boyk e < . C l a y s Bryo» # M *U8TR1 Lact m Axis 1 d - T h u j » b ^ h " J Peta p d-Tsug Violspp U3TR4 L2TR1 U3TR3 -Tiar t Dryo e b-Vaccp , C r y o f Circ a b-Vacc spp / *b-Tsug J P B l e c s T r i f r herb-moss \ con-shrub G y m n D b - T a x u » Corn c Maia d t b-Oplo 2.5 + 4.5 Figure 5.9. Axis 1 and 2 of C C A triplots illustrating the environmental gradients that best explain species and sites distribution. For abbreviation see Appendix 1. Details of the analysis are presented in Appendix 12. 60 5.5 Summary Overall, plant cover was initially higher in L6 than in U6. However, total percent cover in both land treatments was roughly equalised after 19 years (Tables 5.2 and 5.3), although these total covers represent different plant associations (Table 5.4). The species abundance distributions were very similar overall, although the Geometric Series model indicated a tendency towards a more even dominance-diversity relations (Figures 5.2 and 5.3). Species richness was highest in the logged gullies 19 years following disturbance (Figure 5.1). Land treatment, elapsed time, and sidewall slope were identified as the major environmental variables controlling the species occurrence by CCA. Species dispersion on the diagram illustrated that light regime is an important control in the community succession of each land treatment. The associated sediment yield values varied from 0.1 to 71 m3/ha/yr in the logged gullies and 0.3 to 24 m3/ha/yr in the unlogged gullies (Table 5.6). Yields tended to be higher in the logged gullies (Figure 5.8). Soil texture differed between land treatments, with finer material in the unlogged gullies. Organic content also was higher in these gullies (Table 5.7 and Figure 5.6). Sediment yield showed a positive relationship with bare soil and a negative relationship with vegetation cover (Figures 5.7 and 5.8). Higher yield was also associated with younger, less developed plant communities (Figure 5.9). 61 Chapter 6: Discussion This discussion is divided into two main sections. The first section specifically deals with vegetation recovery of gully sidewalls following debris flow events. This section is subdivided to facilitate a discussion of vegetation recovery according to elapsed time (6 and 19 years). Rates of recovery, cover dominance, community differentiation, and species diversity are compared between land treatments for each elapsed time. The second section discusses the values of the sediment yield data as a function of vegetation recovery per land treatment. The implications of bare soil and slope angle in soil loss variability are also briefly discussed. Finally, the land treatment division used in this study is critically examined. Regarding sampling "representativeness" of the vegetation in this environment: it is important to bear in mind that the gully environment is very heterogeneous, and that soil conditions on the sidewalls change continuously, whether they are actively eroding or sustaining plant communities. The larger picture that as been drawn here from the vegetation sampling is based on a certain number of samples which are considered representative of gully sidewall environments. The main aim of the data collection was to describe and account for the vegetation present on the gully sidewalls. Hence, this sampling may have under-represented the sidewall sections that are actively eroding, and not supporting vegetation. Nonetheless, if these areas are underestimated in the study, this bias occurred equally in all treatment and control groups. 6.1 Vegetation Recovery Early community recovery It was observed that after a debris flow disturbance, vegetation cover increased at a faster rate in logged gullies than in unlogged gullies. A significant difference of 20 % is recorded in the total plant cover between the two groups (Table 5.2a, c, and Table 5.3a). This difference in cover is due mostly to the high moss cover encountered in logged gullies at the time of the survey. Mosses are also present in the early plant communities of unlogged gullies but not to the same extent. Smith et al. (1986) reported an average cover of 17.5% (N = 2 and a = 21, where N is the number of samples and cr is the standard deviation) for open slope slides of similar elapsed time, on the Queen Charlotte Islands, although samples in both logged and unlogged areas are not differentiated. In the Cascade Mountains of western Oregon, Miles and Swanson (1986) reported a cover of 33% (N = 2, and a = 9), again for open slope slides, with no differentiation of logged 62 and unlogged surrounding areas, averaging 8 years elapsed since slide disturbance. However, unlike the surveying methods followed here, areas covered by bedrock were included in the Smith et al. and Miles and Swanson's survey, which reduces the relative importance of the vegetation cover. Still in the western Oregon Cascades, Gecy and Wilson (1990) reported an average of 24 % (N=2, and a = unknown) for gullies that torrented in 1986. Their method does not differentiate clearly between logged and unlogged areas; however, they showed that total cover is significantly higher in clear-cut reaches, than in hardwood or coniferous reaches. All of these surveys show plant cover to be much lower than the figures reported here. The main difference between the logged and unlogged gullies at the early stage of development is partly due to their species composition and partly to the relative importance that these species occupy in the whole community. At this stage, logged gullies show a dominance of Rubus parviflorus-Thuja plicata-Boykinia elata. In unlogged gullies, the dominance belongs to herb species and the initial community is aBoykinia elata-Rubus spectabilis-Tsuga heterophylla. It is difficult to compare species associations with other studies, since they are not conducted in the same areas, but factors influencing landslide colonisation can be compared. Two factors appear to be responsible for most of the differences in early plant communities between logged and unlogged gullies, and both are due to the type of vegetation adjacent to them. First, adjacent stands can act as a seed source or provide vegetative fragments or roots from which new plants can grow. Many authors agree on the role of the adjacent stand during vegetation recovery of landslide scars (Kellman, 1974; Hull and Scott, 1982; Swanson et al. 1982; Hupp, 1983; and Miles and Swanson, 1986). In gullies, one expects to find that species colonising scoured sidewalls would occupy a large edaphic range, demonstrating their ability to grow and develop on poor soils. However, plant species of the late to final stages of community development are also found growing in newly scoured gullies, despite the fact that these species usually require a prior soil transformation by pioneer plants (Hupp, 1983). This was also observed in the Queen Charlotte Islands by Smith et al. (1986), who studied environments very similar to those in this study. Comparison of species composition between gullies and their adjacent stands using Jaccard's Similarity Index indicated some but in no case was this strong. Unlogged gullies showed a similarity of 52% with old-growth stands, and climax species such as Thuja plicata, Tsuga heterophylla, Maianthemum dilatatum, Smilacina stellata, Streptotus sp., and many species of ferns are found on the sidewalls. Except for Thuja plicata, these plants do not display 63 the expected ecological aptitude of a large edaphic range initially expected. Flaccus (1959); Bormann and Likens (1979); Hull and Scott (1982) and Hupp (1983) also observed the importance of the prominent canopy species on slide scars, even though this environment appears barely suitable for these species. Veblen and Ashton (1978) concluded for similar observations of tree species that they were not dependent on prior modification of the site for their establishment, but that the development of other vegetation probably increased the rate of stand development. In logged gullies, the influence of the adjacent stands was not as easy to observe as the unlogged. The adjacent stands were more or less "mono-cultural": in other words, they were planted areas of Pseudotsuga menziesii ssp. menziesii. This possibly contributed to the relatively low similarity index values of 32-35% observed between the L6 gullies and the second growth adjacent stands (Table 5.3). These similarity values also indicated that species contribution from adjacent stands was much lower in logged gullies. Nonetheless, climax stage species were also found to grow in L6 gullies. Seeds for these species probably originated in upslope old-growth forest stands, and Thuja plicata and Tsuga heterophylla are often well established on the sidewalls. It was also observed that logging road side-slopes could contribute to the species pool of logged gullies. These slopes host many pioneer forb and shrub species common to this highly disturbed environment, which were seeded or naturally grew there. The most common species observed on these road side slopes and also encountered on gully sidewalls were Rubus parviflorus, Anaphalis margaritacea, Carex mertensii, graminoids, and Epilobium angustifolium. These are fast colonisers and grow well on disturbed sites. Some of these species show very high cover in the logged gullies but were also observed in the unlogged gullies (Appendix 3). They are relatively shade intolerant species (Klinka et al, 1995), which would explain their generally low cover in the unlogged gullies. Management of clear-cut stands such as logging and replanting possibly reduces species richness, which suggests a limitation in the number of species available for gully recovery. Results indicated that the richness of both 10 and 20 year stands were very similar, with 20 different species recorded in the 10-year second growth and 22 species in the 20-year second growth. However gully sidewalls supported a richer plant community of 30 different species in the gullies near the 10-year stands and 26 species in the gullies near 20-year stands. This compares to a total of 34 species in the logged gullies. The unlogged gully richness reached a total of 35 species, and there were 21 different species in the old-growth forest. Consequently, 64 this does not indicate that logging and stand manipulation have influenced species diversity in logged gullies. Better results would be obtained from a comparison with naturally re-establishing stands. The higher richness observed in unlogged gullies is likely due to species contribution by old-growth adjacent stands, and also possibly by the logged gullies. Since most old-growth species are shade tolerant, they will generally perform better in a shaded environment which is generally absent in logged gullies. Hence, shade species are poorly represented in the latter environment. On the other hand, some species that favour the logged gully environment are found in the unlogged gullies, but in lower proportions. This mobility of some species between the logged and unlogged environments possibly contributes to greater gully richness relative to the adjacent open slope stands. Higher richness values observed in gullies compared to the adjacent stands were also observed in other diversity studies (Franklin, 1982; Halpern 1988, 1989; and Halpern and Spies,1995). Although vegetation recovery follows a continuous development toward increasing plant cover, the diversity of sampled communities hides a higher level of complexity. For example, in logged Doulgas fir forest, Halpern and Spies (1995) confirmed the presence of a diversity pattern (initially observed by Franklin, 1982), whereby plant diversity increases to a maximum level before the canopy closes and then decreases, only to increase again with reopening of the canopy. This model is a simplified version of a larger scale diversity oscillation. Halpern (1988,1989) observed that at a finer resolution, diversity might peak several times before canopy closure. Yet, even in the limited time scale of this sampling (referring to the recovery times), the diversity of the compared sites displays a similar pattern. Highest species richness are observed in the open canopy environment, represented by torrented gullies, and lower diversities are recorded in the adjacent closed canopy stands. This leads to the second factor responsible for plant community differences between the land treatments, namely modification of the gully light regime by the adjacent stand. This factor was also observed in other studies on vegetation recovery on landslides scars (Hull and Scott, 1982; Swanson et al. 1982; and Hupp, 1983). Even though logged and unlogged gullies have similar slope aspects, unlogged gullies are more shaded by the adjacent veteran conifers (Gecy and Wilson, 1990). This factor has direct implications for the species richness as demonstrated above and on the abundance and vigour exhibited by the species. Since gullies create an opening in the canopy, where some light can reach the ground, one can argue that unlogged gully environments would be affected by a nearly similar light 65 regime as in the logged gullies. However, the height of trees and the stratified vegetation of the old-growth stand bordering the unlogged gullies are not comparable to what is observed in the adjacent stands of the logged gullies. Hence the light regime of the logged gullies is much higher than what is observed in unlogged gullies, especially in the first years immediately following debris flow. Some species responsible for the large cover observed in L6 are dependent on a high light regime to maintain this abundance and vigour. This is a well known fact in plant ecology, given that species distribution is controlled strongly by light and that some species have developed a tolerance to shade. However some trade-offs were made to acquire this advantage (over other species) and the shade tolerant species will generally colonise the new available sites less rapidly. The very abundant moss cover in L6 is a good example of this. Moss cover, even if present in other land and time treatments, does not reach this extent anywhere else. Since these species perform well in logged gullies, it is possible that these mosses are some species of Polytrichum that are shade tolerant - intolerant (Klinka et al. 1995), and they are generally associated with disturbed areas (Klinka et al, 1995). Plant species such as Alnus sinuata, Rubus parviflorus, Anaphalis margaritacea, and graminoids occupy a large part of the coverage in the logged gullies, and are poorly represented in the unlogged environment. These species are very common in open areas and can grow and proliferate quite rapidly. Alnus sinuata, Anaphalis margaritacea and graminoids species grow very well on exposed mineral soil (Klinka et al, 1995). Rubus parviflorus is common to the early-seral stage in logged or burned sites (Klinka et al, 1995). These species appear to develop well in torrented gullies and were also observed on road-cuts. As mentioned earlier this species cover dominates in gully group L6. Although these species are all found in unlogged environments (Klinka et al, 1995), they do not exhibit similar cover due to their lower shade tolerance. Enough light can filter into unlogged gullies to allow some of the less shade tolerant species to colonise. The data implies that many species found in logged gullies are also found in the unlogged gullies, though they do not appear to grow and reproduce as well. Other species colonising the unlogged gullies, probably coming from the adjacent stand, are better adapted to these light conditions, but grow at slower rates. These species are often better competitors in this environment and will slowly replace the less shade tolerant ones. Sampling of torrented open slopes ( T O S ) allowed comparison between growth on gully sidewalls and areas outside the gully. T O S plant cover is about 20% higher than in L6 (Table 5.1). 66 These areas had a higher deposition of material with a lower level of scouring compared to what is experienced on the sidewalls. Hull and Scott (1982), Smith et al. (1986), Miles and Swanson (1986), and Gecy and Wilson (1990) found the level of scouring by torrents to be the most significant variable influencing vegetation recovery rate and successional pattern. A higher level of organic soil remaining on the slope, which likely includes seeds and organic soil fragments, will promote faster regrowth of vegetation than in gullies. One other probable explanation of the rapid growth on these sites is the rapid stabilisation of the adjacent slope (Miles et al. 1984). Plant community development In the second 'time window', represented by gullies that experienced debris flows between 1975 and 1980, total plant cover is substantially higher and reaches similar values in logged and unlogged gullies. Vertical stratification of vegetation is also very similar in the tow cases after this elapsed time (Table 5.2d). In both environments, the tree stratum had started to show cover, indicating the presence of vegetation taller than 10 m. In logged gullies, shrub and herb cover increased, yet only the herb layer recorded a statistically significant change from 28% to 38%. Although not significant, the shrub layer still showed substantial change from 27% to 48%. Finally, the ground cover decreased significantly from 22% to 13%. In unlogged gullies there was a significant increase in cover in all strata, except at the ground level. Smith et al. (1986) recorded a mean total cover of 82%o (N = 2, a = 6) for 20-year old slides while Miles and Swanson report an average cover of 52%(N=19ando- = 24). In logged gullies at 19 elapsed years the dominance changed to a Thuja plicata-Rubus parviflorus-Boykinia elata community. This represents a shift of dominance and co-dominance from the early successional stage, but the herb third-dominance is maintained. The unlogged gullies were entirely dominated by herb species with Blecchnum spicant-Tiarella trifoliata-Dryopteris expansa. Tsuga heterophylla has slowed its cover growth and Thuja plicata showed a higher cover, but still lower than that of the herbs. Alnus rubra (red alder), reported in many studies as the prime re-coloniser on slide scars (Smith et al. 1986; Swanson et al, 1987; and Gecy and Wilson, 1990), was not observed in the surveyed gullies, but instead was present downslope near Cedar Creek. This tree species was also observed in much older torrented gullies in Seymour basin, another GVWD water reservoir. The reasons why Alnus rubra was not found as prime coloniser remain unclear, however, the elevation or dryness of this area during the summer months may account for its absence. 67 There was a reasonably strong similarity between older gully communities and initial recovery populations. However the similarity was lower between logged gullies and unlogged gullies. Species forming the 19-year communities in logged and unlogged gullies showed a level of similarity of 62%, which is slightly lower than that of the initial recovery stage. This is to be expected since logged gullies are still gaining in species richness: the number of species increased from 34 at the early stage to 40 after 19 years. Unlogged gullies showed the opposite trend, with the plant community losing species, from 35 to 31 different species over a similar time frame. Relative abundance curves indicate very small changes over time in species distribution, and thus in resource usage among species. According to Kolmogorov-Smirnov goodness-of-fit tests, the distribution of species in the L6 and L19 groups is very similar, hence no large changes occur in resource usage. Although not statistically significant it does appear that species distribution changes slightly over time in unlogged gullies. Geometric series models (Pastor, 1995) calculated for each treatment group indicate a small progression over time toward a more even environment, or a more equal sharing of the resources (Figure 5.3). These changes in slope agree with the simple fact that if more individuals are added to an environment, resources have to be better shared by the different plant species for them to survive. Bazzaz (1975) observed that changes in slope on such diagrams correspond to variations in successional communities, which often show steeper model slopes at the initial stages. The species distributions in D C A and C C A (Figure 5.4 and 5.5) confirm the working hypothesis that both elapsed time and land treatment influence the structure of plant communities on gully sidewalls. The ordination results reinforce the preceding analyses that underlined the differences between logged and unlogged communities. In the C C A diagram species distributions show very strongly the progression of an open-area colonising community, toward a more shaded environment near climax development. The graph is divided by the first axis into logged or unlogged environments. Under the first axis, logged sites start on the left of the diagram with gully sites, and end with the logged adjacent plots. A similar progression is observed for unlogged gullies and old-growth plots, above the first axis. Hence a very clear difference exists between the logged and unlogged environments inside and outside the gullies. The C C A diagram suggests an evolution of gully plant communities towards the composition of adjacent open-slope communities. However, the current data set does not permit positive confirmation of this. It was also observed in the field that vegetation growing at the sidewall-open slope boundary often differed from that of the closed canopy forest stand. This vegetation can play a key role in gully revegetation, and is possibly more important than the main stand itself. Future studies of this kind should use a plotless sampling method such as a line-intercept method (Mueller-Dombois and Ellenberg, 1974). This method would improve sampling of the plant distributions on sidewalls and their progression from the adjacent open-slope stand into the channel. This would include the transitional area between them. In this study, the effect of this area on plant re-establishment on gully sidewalls may be underestimated. Gullies are unstable landforms and, although partially colonised by vegetation, they have recurring mass movement activities. Sites are susceptible to micro- or macro-failures, and debris flows that continually provide new sites for colonisation and prevent canopy closure. This suggests that gully environments will probably always show higher diversity due to a higher light regime and differences in substrate. Moss and Rosenfeld (1978) considered mass movement disturbances in forests to be considerably different from disturbances by fire or tree-throw because of the extent to which they modify the soil, which forces new communities to remain different from the climax stand for longer periods. Another type of disturbance results from snow avalanches, which also contribute to community diversity and fragmentation of the vegetation mosaic. Studies by Patten and Knight (1994) and Butler (1979) suggest that the level of soil disturbance by avalanches is generally lower than that by debris flow events. Avalanche frequency typically is greater than that of debris flows, and the resistance of some vegetation to avalanche disturbance create unique plant communities in avalanche paths, quite dissimilar to those created by debris flow disturbance. Overall, these types of disturbance regimes are not a negative factor and disturbance communities contribute to the general floristic diversity of the forest community. 6.2 Sediment production on gully sidewalls 6.2.1 Plant cover and sediment yield. As identified in section 2.2.2, rainsplash, sheetwash, frost heave and dry ravel are the dominant surface erosion processes in the south coastal climate of British Columbia (Nistor 1996). On vegetated slopes, these processes occur at different intensities, since the cover provided by vegetation tempers their impact. Moreover, in a natural setting, this cover varies 69 both vertically and horizontally, making this effect difficult to isolate or control. This is why soil loss research is often based on bare soil areas, which are easier to compare. Vegetation cover, as discussed earlier, generally increases with time at all sites; however, this tendency was not well observed in the sediment trap data sub-set, and sediment production was not significantly negatively correlated with time (r = -0.2392) as expected. Attempts to select different vegetation cover in the sampling possibly masked the effect of the elapsed time variable, which might explain this poor result. Another important characteristic of the sediment traps is that total vegetation cover comparisons between land treatments differ from that which was observed in the whole vegetation sampling. Hence, vegetation cover in the gullies with sediment traps, excluding the bare soil plots, was similar between recently disturbed sidewalls of both land treatments, but was higher in the older disturbed group of unlogged gullies (Table 5.6). For the whole data set, recently logged gullies showed higher cover values than unlogged gullies, yet cover values were roughly equal in both land treatments after 19 years. The sediment yield values obtained in this survey showed some relation to vegetation cover values (correlation of -0.6214). Sediment yield was higher in logged gullies and regression analysis indicated that yield decreases similarly in both land treatments as plant cover increases over time. Vegetation cover explains 59% of the sediment yield variance in the logged gullies, and 76% of the variance in the unlogged gullies (Figure 5.7). Unfortunately, vegetation cover in these calculations are limited to that immediately associated with the sedimentation traps, and consequently are not necessarily representative of the vegetation cover of the whole gully. Increased sediment trap density might have shown different yields for the two treatments, logged and unlogged. The soil composition was another distinction observed between the gully sidewalls of the different land treatments. Logged gullies showed coarser soil (1000 um) on the sidewalls and in the traps, compared to unlogged gullies which have a higher level of silt and clay (<63um) (Figure 5.6). Fine textured soil is very susceptible to rainsplash unless cover is provided to intercept raindrops. Material larger than sand (gravelly) has reduced susceptibility to rainsplash erosion (Luce and Black, 1999). Clay materials have an inherent cohesiveness and are, therefore, not as susceptible to rainsplash erosion, unlike coarser materials such as sand, which lack this property (Farres and Cousen, 1985 ). The rainsplash resistance properties of the different textural classes partially explains the higher sediment yield observed in the logged gullies, especially in cases where soils contain the more erodible sandy materials. 70 Higher organic contents observed in unlogged gullies, both on the sidewalls and in the traps (Table 5.6), would influence vegetation regrowth by enriching the soil. Even if a fraction of this organic material corresponds to a higher litter presence, this litter can be an important agent interfering with rainsplash erosion (Benkobi et al, 1993), giving more soil protection in unlogged gullies. Moreover, organic matter can also form stable soil aggregates, which act as rainsplash erosion inhibitors (Farres and Cousen, 1985). 6.2.2 Indicator species for erosion intensity The ordination analysis which included sediment yield illustrates the fact that high sediment yield values are to be expected in areas of low cover or newly disturbed areas. Although all variables used in the ordination diagram significantly explain the data dispersion, the sediment yield variable has an artificially high explanatory power since it was selected as the first variable in the model. When this variable is selected later its level of explanation to the model is lower. The idea here was to determine how plant communities relate to areas of high or low sediment production, since such "cover communities" would be useful in erosion management. The community dominated by Rubus parviflorus-Thuja plicata, which was also associated with graminoids and Anaphalis margaritacea (a disturbance indicator species (Klinka et al, 1995)), showed the highest association with mineral sediment production. This closely corresponds to the plant association previously identified in the logged gullies. Figure 5.8 also illustrates that another group of plants has a high weighted-average on the sediment yield gradient, but is also associated with the organic gradient. This group includes old-growth plants such as Taxus brevifolia, which actually grows outside gullies, but hangs significantly over the sidewall. The Rubus parvifloru-Thuja plicata association is composed of plants that would provide good cover if the sidewalls were densely populated; however, this was not what was observed on recently disturbed gully sidewalls. Moreover, these plants provide a seasonal cover since Rubus parviflorus is a deciduous shrub, and may not protect the ground against autumn rainstorms, although leaf litter may compensate for the lack of vegetation. As for plots under the old growth association cover, the sediment yield from these sites had a high organic content relative to debris falling from the forest above. Although ferns and lilies, found in this group, do not grow in full mats, mineral sediment yield is not believed to be high. Organic debris and old fern leaves 71 will also create litter that reduces soil loss (Benkobi et al. 1993). Other species were associated to low sediment yields and probably represent communities that having good ground cover. 6.2.3 Role of vegetation stature in erosion rates In general logged and unlogged treatments appear to explain relatively well the differences in sediment production. However, this explanation only takes into account the vegetation present on the sidewalls. The vegetation present on the adjacent open slopes also shields the gully sidewalls from direct rain impact, and thus from rainsplash erosion. This also accounts for the lower rate of sediment production in the unlogged gullies. Vegetation tall enough to intercept a significant amount of rain is found in the logged area within the 20-year second growth stands (20 II), and in unlogged areas with the old-growth stands ( O G ) . Ten-year second growth stands (10 ii) have lower stature vegetation which does not provide the same amount of shelter to the gully. Figure 6.1 uses a different land treatment division that replaces logged and unlogged gullies by two land categories based on the height of the adjacent stands. This modification probably test the impact of land treatment on the sediment yield results. The "high adjacent" vegetation is represented by old-growth and 20-year second growth stands, while the "low adjacent" corresponds to the 10-year adjacent stands. This re-allocation of the sites into the new land categories only involves two sites, which represent about 10 % of the sites. In this graph the low and high vegetation divisions provide very similar erosional responses to the logged - unlogged land treatments. Differences in the regression coefficients are not significant, although slightly lower ( R 2 i o w = 0.58 and R 2 h i g h = 0.73; compared with Figure 5.7, R 2 iog g ed = 0.59 and R 2 u n i o g g e d = 0.76). This suggests that the logged - unlogged land treatments provide a slightly better explanation, but the stature effect is worthy of more investigation. Especially when considering the importance of keeping bands of undisturbed standing vegetation (buffer strip) along active or potentially active erosion features during logging. 72 100.0 T3 C/3 10.0 0.1 • Low adjacent • Higher adjacent X Outliers -0.05x 20 40 60 80 100 Vegetation cover (%) 120 140 Figure 6.1. Sediment yields (m3/ha/yr) as function of vegetation cover (%) for land 'treatments' of high and low adjacent vegetation. 6.2.4 Soil loss prediction model Bovis (1982) demonstrated that the percent of bare soil multiplied by the sine of the slope angle provides a relatively good prediction for soil loss, with a R 2 = 0.67 (Figure 6.2). This study was conducted in the montane forest, sub-alpine forest, and alpine tundra of the Colorado Front Range, which represents a much drier environment than the Coquitlam basin and has very different vegetation. However, it is interesting to observe that the erosion data observed in the present study follows the same trend suggested by Bovis' (1982) data (Figure 6.2) Further, they also strengthen the regression with a new overall R coefficient of 0.71. This result emphasises the importance of bare soil and slope angle in sediment production. But one should observe that in this study the results appear to be heteroscedastic, meaning that the regression errors increase as the soil loss increases. 73 Q. O + Pellerin(2000) - Bovis (1982) -Millard (1993) l.E+07 l.E+09 Soil loss (grams) Figure 6.2. Percent bare soil times sine of the slope as a function of soil loss, compared with Bovis (1982) and Millard (1993). Millard's (1993: 70) rainsplash erosion data, evaluated in the Cedar Creek gullies, also show a relation with the product of bare soil percent and sine of slope angle, but follows a different trend. A possible explanation for this is that Millard used erosion pins to measure sediment yield. As stated in section 4.4.1, pins are very difficult to use and may overestimate sediment yield. This study and Bovis' (1982) study used sediment traps. Millard's (1993) data also show a very high rate of erosion, possibly because pins were deliberately installed on actively eroding areas. This study and Bovis' reported vegetation on the surveyed areas. Although two plots in this study can be considered as very actively eroding, they do not approach Millard's rate of erosion of 10 mm/yr, having mean yields of only 5 mm/yr. 74 Chapter 7: Conclusion 7.1 Plant re-establishment conclusions The objectives of this study were to observe and document vegetation recovery in gully sidewalls after they had experienced debris flow disturbance. The experimental design also allowed comparison of these observations between clear-cut and undisturbed forest areas. Furthermore, analyses to determine the leading environmental factors controlling vegetation re-establishment were conducted. The results have demonstrated that vegetation re-establishment in gullies is influenced in two ways by land management history. The first concerns the availability of species for revegetation of the scoured gully sidewalls, and the second involves the gully light regime, which is a key variable in vegetation development on sidewalls. Plants of late community succession were observed on the sidewalls six years after debris flow disturbance, indicating that they do not require land modification by pioneer plants prior to their establishment. Many of these species are directly contributed by stands adjacent to the gullies. Pioneer and early-seral species were also found in varying numbers, depending on the gully light regime, which is largely controlled by the adjacent stand. Consequently, logged gullies, having a higher light regime, had a larger number of pioneer and early-seral species than unlogged gullies. These plants generally spread rapidly, which gives an advantage to logged gullies over light-limited unlogged gullies. Therefore, recently torrented logged gullies have a higher initial plant cover than those in unlogged environments. However, this situation is only temporary and after 19 years of development both logged and unlogged gullies showed a similar vegetation cover with only 10 % bare soil. Higher species diversity was observed in the gully communities compared to the adjacent stands. This suggests that the disturbance regime induced by debris flow activities contributes to the general plant diversity of the area. In the logged gullies the initial dominant species cover association following disturbance is Rubus parviflorus - Thuja plicata - Boykinia elata, which evolves into a Thuja plicata - Rubus parviflorus - Boykinia elata association after nineteen years. This environment is dominated by shrubs and will likely become a Thuja plicata community, unless disturbed again. In unlogged gullies, community dominance is very different from that of logged gullies. The early recovery stage is dominated by Boykinia elata - Rubus spectabilis - Tsuga 75 heterophylla, which later becomes a Blecchnum spicant - Tiarella trifoliata - Dryopteris expansa community. This environment is initially dominated by herb species, however there is also a good shrub cover. Although not represented in the dominant community species, Thuja plicata replaces Tsuga heterphylla in abundance after nineteen years. 7.2 Soil erosion conclusions Another objective of this study was to determine sediment yield from gully sidewalls and to attempt to associate sediment yield with specific vegetation cover. It was observed that sediment yields decreased with increasing vegetation cover, and showed similar decreasing rates for both land treatments at similar elapse time since disturbance. However, yield is generally higher in the logged gullies. Sediment yields varied between 0.1 to 72 m3/ha/yr in logged gullies, and 0.3 to 24 m3/ha/yr in unlogged gullies. This result highlights the role of the adjacent gully stands in rain interception, which tend to reduce rainsplash erosion. Unfortunately, limitations in the sampling design do not allow clear identification of which is the more important of the two, undisturbed forest or vegetation height. Either way, the sediment yield results have important implications in gully management and should be taken into account in gully rehabilitation work and clear cut logging in the proximity of gullies. 7.3 Summary of the main results: Plant re-establishment: (1) Vegetation cover areas were significantly lower on unlogged sidewalls, relative to logged sidewalls, after 6 elapsed years (58% and 78%, respectively). After 19 years, vegetation covers were quite similar in both environments (111%) and 114%, respectively). (2) Bare soil areas were significantly higher on unlogged sidewalls, relative to logged sidewalls, after 6 elapsed years (34% and 22%, respectively). After 19 years, bare soil values were quite similar in both environments (7% and 9%, respectively). (3) Logged and unlogged gullies showed different community associations and different rates of recovery, likely due to land management influences and light regimes of disturbed areas. 76 (4) Vegetation communities differed between logged and unlogged gullies at both 6 years and 19 years following disturbance. Thuja plicata, however, seems to be the primary tree species in the undergrowth community in both gully types, suggesting that it will be the dominant tree in the future, barring further debris flows. (5) Debris flow disturbance regimes in gullies promoted higher plant species diversity, and enriched the floristic panorama of the area. Soil erosion: (1) Sediment yields varied between 0.1 to 72 m3/ha/yr in logged gullies, and 0.3 to 24 m3/ha/yr in unlogged gullies. Regression curves show a similar rate of sediment yield decline in both environments, albeit at higher absolute values in logged environments. (2) Results indicate that plant cover on gully sidewalls significantly explains the variation in sediment yield. (3) Land management influences sediment yield in two ways: i) as an important factor in determining the rate of growth in vegetation stature, which helps limit rainsplash erosion. ii) by determining the incoming plant communities, which affects recovery rates, and thus the sediment yield rate. (4) Variables such as percent bare soil and slope angles were shown to be important in determining sediment yield in steep, partially vegetated gully environments. 77 Bibliography Books and Serials Arsenault, A. and Bradfield, E. , 1995. Structural-compositional variation in three age-classes of temperate rainforests in southern coastal British Columbia. Canadian Journal of Botany, 73: 54-64. Bastow, W.J., 1991. Methods for fitting dominance/diversity curves. Journal of Vegetal Science, 2: 35-46. Bazzaz, F.A., 1975. Plant species diversity in old-field successional ecosystems in southern Illinois. Ecology, 56: 485-488. Benkobi, L. , Trlica, M J . and Smith J.L., 1993. Soil loss as affected by different combinations of surface litter and rock. Journal of Environmental Quality, 22: 657-661. Bormann, F.H. and Likens G.E. 1979. Patter and Process in a Forest Ecosystem. Springer-Verlag, New York, 253 p. Bovis, M.J., 1993. Hillslope geomorphology and geotechnique. Progress in Physical Geography, 17(2): 173-189. Bovis, M.J., 1982. The spatial variation of soil loss and soil loss controls. In C E . Thorn, (ed.). Space and Time in Geomorphology, Binghamton Symposia in Geomorphology, no. 12. George Allen and Unwin, Binghamton, p. 1-24. Bovis, M.J. and Dagg, B.R., 1992. Debris flow by impulsive loading: Mechanical modelling and case studies. Canadian Technical Journal, 29: 345-352. Bovis, M.J., Dagg, B.R., and Kaye, D., 1985. Debris flows and debris torrents in the southern Canadian Cordillera. Discussion of paper by D.F. VanDine. Canadian Geotechnical Journal, 22: 608. Bovis, M.J. and Jakob, M . , 1999. The role of debris supply conditions in predicting debris flow activity. Earth Surface Processes and Landforms, 24(11): 1039-1054. Bovis, M.J. Millard, T.H., and Oden, M.E. , 1998. Gully processes in coastal British Columbia: The role of woody debris. In D.L. Hogan, P.J. Tschaplinski, and S. Chatwin, (eds.). Carnation Creek and Queen Charlotte Islands Fish/Forestry Workshop: Applying 20 Years of Coastal Research to Management Solutions. British Columbia Ministry of Forests Research Program, Land Management Handbook, no. 41, p. 49-75. Butler, D.R., 1979. Snow avalanche path terrain and vegetation, Glacier National Park, Montana. Arctic and Alpine Research, 11(1): 17-32. Canadian Climate Program, 1993. Canadian Climate Normals, 1961-1990, British Columbia. Environment Canada, Atmospheric Environment Service, Ottawa, p. 36. 78 Carr W.W., 1985. Watershed Rehabilitation Options for Disturbed Slopes on the Queen Charlotte Islands. Land Management Report no. 36, British Columbia Ministry of Forests, Victoria, 34 p. Carr, W.W. and Wright, I., 1992. A Soil Erosion and Sediment Control Planning System for Managed Forest Land: A Case Study of Shomar Creek, Queen Charlotte Islands, British Columbia. Land Management Report no. 75, British Columbia Ministry of Forests, Victoria, 23 p. Chatwin, S.C., Howes, D.E., Schwab, J.W., and Swanston, D.N., 1994. A Guide for Management of Landslide-Prone Terrain in the Pacific Northwest, 2 n d ed. Land Management Handbook no. 34, British Columbia Ministry of Forests, Victoria, 212 p. Digby, P.G.N., and Kempton, R.A., 1987. Population and Community Biology Series: Multivariate Analysis of Ecological Communities. Chapman and Hall, London, 206 p. Dissmeyer, G.E. and Foster, G.R., 1985. Modifying the universal soil loss equation for forest land. In S.A. El-Swaify, W.C. Moldenhauer and A. Lo, (eds.), Soil Erosion and Conservation, Soil Conservation Society of America, Ankeny, Iowa, p. 480-495. Fannin, R.J. and Rollerson, T.P., 1993. Debris flows: some physical characteristics and behaviour. Canadian Geotechnical Journal, 30: 71-81. Farres, P. and Cousen, S.M., 1985. An improved method of aggregate stability measurement. Earth Surface Processes and Landforms, 10: 321-329. Flaccus E. , 1959. Revegetation of landslides in the White Mountains of New Hampshire. Ecology, 40: 692-703. Franklin, J.F., 1982. Old-growth forests in the Pacific Northwest: An ecological view. In Bureau of Governmental Research and Service. Old-Growth Forest: A Balanced Perspective. University of Oregon, Eugene, p. 5-27. Gauch, H.G., 1982. Multivariate Analysis in Community Ecology. University of Cambridge Press, Cambridge, 298 p. Gecy, J.L. and Wilson, M.V. , 1990. Initial establishment of riparian vegetation after disturbance by debris flows in Oregon. American Midland Naturalist, 123: 282-291 Gee, G.W. and Bauder, J.W., 1982. Particle-size analysis. In A .L . Page, R.H. Miller and R.H. Keeney, (eds.). Methods of soil analysis. Part I. Physical and mineralogical methods, 2nd ed. Agronomy no. 9, American Society of Agronomy, Madison, p. 383-411. Gray, D.H. and Megahan, W.F., 1981. Forest vegetation removal and slope stability in the Idaho batholith. USDA, Forest Service, Intermountain Forest and Range Experiment Station, Research Paper, INT-271, p. 1-23. 79 Halpern, C.B., 1989. Early successional patterns of forest species: Interactions of life history traits and disturbance. Ecology, 70(3): 704-720. Halpern, C.B., 1988. Early successional pathways and the resistance and resilience of forest communities. Ecology, 69(6): 1703-1715. Halpern, C.B. and Spies, T.A., 1995. Plant species diversity in natural and managed forests of the Pacific Northwest. Ecological Applications, 5(4): 913-934. Hay, J.E. and Oke, T.R., 1994. The Climate of Vancouver, 2 n d edition. British Columbia Geographical Series no. 50, University of British Columbia, Vancouver, 84 p. Heed, B.H., 1975. Watershed indicators of landform development. In Arizona Section of America Water Resources Association and Hydrology Section of Arizona Academy of Science, Proceedings of Hydrology Water Resources in Arizona and the South-west, vol. 5, p. 43-46. Hill, M.O. and Gauch, H.G., 1980. Detrended correspondence analysis, an improved ordination technique. Vegetatio, 42: 47-58. Hogan, D.L., Tschaplinski, P.J. and Chatwin, S., ( eds.), 1998. Carnation Creek and Queen Charlotte Islands Fish/Forestry Workshop: Applying 20 Years of Coastal Research to Management Solutions. Land Management Handbook no. 41, British Columbia Ministry of Forests, Victoria, 275 p. Hogan, D.L. and Millard T.H., 1998. Gully Assessment Methods. In D.L. Hogan, P.J. Tschaplinski, and S. Chatwin, ( eds.). Carnation Creek and Queen Charlotte Islands Fish/Forestry Workshop: Applying 20 Years of Coastal Research to Management Solutions. Land Management Handbook no. 41, British Columbia Ministry of Forests, Victoria, p. 183-188. Horton R.E., 1945. Erosional Development of Streams and their drainage basins: Hydrophysical approach to quantitative morphology. Bulletin of the Geological Society of America, 56: 275-370. Hull, J.C. and Scott, R.C., 1982. Plant succession on debris avalanches of Nelson County, Virginia. Castanea, 47: 158-176. Hungr, O., Morgan G.C. and Kellerhals, R.,1984. Quantitative analysis of debris torrent hazards for design of remedial measures. Canadian Geotechnical Journal, 21(4): 633-677. Hupp. R.C., 1983. Seedling establishment on a landslide site. Castanea, 48(2): 89-98. Jakob, M . , 1996. Morphometric and geotechnical controls of debris flow frequency and magnitude in southwestern British Columbia. Unpublished Ph.D. thesis, University of British Columbia, Vancouver, 232 p. Jongman, R.H., ter Braak, C.J.F., and van Tongeren, O.F.R., ( eds.), 1995. Data Analysis in Community Ecology. Cambridge University Press, Cambridge, 299 p. 80 Jordan, P., 1994. Debris flows in the southern Coast Mountains, British Columbia: Dynamic behaviour and physical properties. Unpublished Ph.D. thesis, University of British Columbia, Vancouver, 258 p. Jordan, P. and Slaymaker, O., 1991. Holocene sediment production in Lillooet River basin, British Columbia: A sediment budget approach. Geographic Physique et Quaternaire, 45(1): 45-57. Kellman, M . , 1974. Preliminary seed budgets for two plant communities in coastal British Columbia. Journal of Biogeography, 1: 123-133. Kent, M . and Coker, P.,1992. Vegetation Description and Analysis: A Practical Approach. CRC Press, Boca Raton, Fla, 363 p. Klinka, K., Krajina, V.J. , Ceska, A. and Scagel, A . M . , 1995. Indicator Plants of Coastal British Columbia. University of British Columbia Press, Vancouver, 288 p. Klinka, K., Pojar, J., and Meidinger, D.V., 1991. Revision of Biogeoclimatic Units of Coastal British Columbia. Northwest Science, 65(1): 32-47. Krag, R.K., Sauder, E.A. and Wellburn, G.V., 1986. A Forest Engineering Analysis of Landslides in Logged Areas on the Queen Charlotte Islands, British Columbia. Land Management Report no. 43, British Columbia Ministry of Forests, Victoria, 138 p. Luce, C.H. and Black, T.A., 1999. Sediment production from forest roads in Oregon. Water Resources Research, 35(8): 2561-2570. Luttmerding H.A., Demarchi, D.A., Lea, E.C. , Meidinger, D.V., and Void, T., 1990. Describing Ecosystems in the Field, 2 n d edition. M O E Manual no. 11, British Columbia Ministry of Environment, Victoria, 213 p. Magurran, A . E , 1988. Ecological Diversity and its Measurement. Princeton University Press, New Jersey, 179 p. Miles, D.W.R., and Swanson, F.J., 1986. Vegetation composition on recent landslides in the Cascade Mountains of western Oregon. Canadian Journal of Forest Research, 16: 739-744. Millard, T.H. , 1993. Sediment in forested and logged gullies, coastal British Columbia. Unpublished M.Sc. thesis, University of British Columbia, Vancouver. 202 p. Moss, M.R. and Rosenfeld, C.L. , 1978. Morphology, mass wasting and forest ecology of a post-glacial re-entrant valley in the Niagara escarpment. Geografiska Annaler, 60A(3-4): 161-174. Mueller-Dombois, D. and Ellenberg, H., 1974. Aims and Methods of Vegetation Ecology. Wiley, New York, 547 p. 81 Nistor, C , 1996. Temporal patterns in the normal regime fine-sediment cascade in Russell Creek basin, Vancouver Island. Unpublished M.Sc. thesis, University of British Columbia, Vancouver, 236 p. Oden, M.E. , 1994. Debris recharge rates in torrented gullies on the Queen Charlotte Islands. Unpublished M.Sc. thesis, University of British Columbia, Vancouver, 100 p. O'Loughlin, C.L. , 1974. A study of tree root strength deterioration following clearfelling. Canadian Journal Forest Research, 4: 107-113. O'Loughlin, C.L. , 1972. The stability of steepland forest soils in the Coast Mountains, Southwest British Columbia. Unpublished Ph.D. thesis, University of British Columbia, Vancouver, 147 P-Palmer, M.W., 1993. Putting things in even better order: The advantages of canonical correspondence analysis. Ecology, 74(8): 2215-2230. Pastor, J., 1995. Diversity of biomass and nitrogen distribution among plant species in arctic and alpine ecosystem. In F.S. Chapin, III and C. Kdrner, ( eds.). Arctic and Alpine Biodiversity. Springer-Verlag, Berlin, p. 255-269. Patten, R.S. and Knight, D.H., 1994. Snow avalanches and vegetation pattern in Cascade Canyon, Grand Teton National Park, Wyoming, U.S.A. Arctic and Alpine Research, 26(1): 35-41. Pielou, E.C. , 1984. The Interpretation of Ecological Data: A Primer on Classification and Ordination. Wiley, New York, 263 p. Pojar, J. and MacKinnon, A., 1994. Plants of Coastal British Columbia: Including Washington, Oregon & Alaska. Lone Pine, Vancouver, 527 p. Riestenberg, M . M . , and Sovonick-Dunford, S., 1983. The role of woody vegetation in stabilizing slopes in the Cincinnati area, Ohio. Geological Society of America Bulletin, 94:506-518. Robert, D.W., 1986. Ordination on the basis of fuzzy set theory. Vegetatio, 66: 123-131. Roberts, R.G. and Church, M . , 1986. The sediment budget of severely disturbed watersheds, Queen Charlotte Islands, British Columbia. Canadian Journal of Forestry Research, 16: 1092-1106. Roddick, J.A., 1965. Vancouver North, Coquitlam, and Pitt Lake Map Aeras, British Columbia. Geological Survey of Canada, Memoir, no. 335, 276 p. Rood, K . M . , 1990. Site Characteristics and Landsliding in Forested and Clearcut Terrain, Queen Charlotte Islands. Land Management Report no. 64, British Columbia Ministry of Forests, Victoria, 46 p. 82 Rood, K . M . , 1984. An Aerial Photograph Inventory of the Frequency and Yield of Mass Wasting Events on the Queen Charlotte Islands, British Columbia. Land Management Report no. 34, British Columbia Ministry of Forests, Victoria, 55 p. Schwab, J.W., 1983. Mass wasting: October-November 1978, Rennell Sound, Queen Charlotte Islands, British Columbia. Report no. 91, British Columbia Ministry of Forests, Victoria, 23 P-Selby, M.J., 1993. Hillslope Materials and Processes, 2 n d edition. Oxford University Press, Oxford, 451 p. Sidle, R.C., 1992. A theoretical model of the effects of timber harvesting on slope stability. Water Resources Research, 28(7): 1897-1910. Sidle, R.C., Pearce A.J. and O'Loughlin, C.L. , 1985. Hillslope Stability and Land Use. Water Resources Monograph no. 11, American Geophysical Union, Washington D.C., 140 p. Smith, R.B., Commandeur, P.R., and Ryan, M.W., 1986. Soils, Vegetation, and Forest Growth on Landslides and Surrounding Logged and Old-Growth Areas on the Queen Charlotte Islands, British Columbia. Land Management Report no. 41, British Columbia Ministry of Forests, Victoria, 95 p. Sousa, W.P., 1984. The role of disturbance in natural communities. Annual Review of Ecology andSystematics, 15: 353-391. Swanson, F.J., Benda, L . E . , Duncan, S.H, Grant, G.E., Megahan, W.F., Reid, L . M . and Ziemer, R.R., 1987. Mass failures and other processes of sediment production in Pacific Northwest forest landscapes. In E.O. Salo and T.W. Cundy. Streamside Management: Forestry and Fishery Interactions. Institute of Forest Resources, Contribution no. 57, University of Washington, Seattle, p. 9-34. Swanson, F.J., Gregory, S., Sedell, J. and Campbell, A., 1982. Land-water interactions: The riparian zone. In R.L. Edmonds (ed.). Analysis of Coniferous Forest Ecosystems in the Western United States. Hutchinson Ross Publishing Co., Stroudsburg, p. 267-291. ter Braak, C.J.F., 1995. Ordination. In R.H., Jongman, C.J.F. ter Braak and O.F.R. van Tongeren, (eds.). Data Analysis in Community Ecology. Pudoc, Wageningen, Netherlands, p. 91-173 ter Braak, C.J.F., 1987. The Analysis of vegetation-environment relationships by canonical correspondence analysis. Vegetatio, 69: 69-77. ter Braak, C.J.F., 1986. Canonical correspondence analysis: A new eigenvector technique for multivariate direct gradient analysis. Ecology, 67(5): 1167-1179. ter Braak, C.J.F. and Smilauer, P., 1998. CANOCO Reference Manual and User's Guide to Canoco for Windows. Centre for Biometry, Wageningen, Netherlands, 351 p. 83 Thurber Engineering Ltd., 1991. Geotechnical assessment of 1990-1991 landslide events in Greater Vancouver Water District Watersheds. Unpublished report to Greater Vancouver Water District, Vancouver, 26 p. Tokeshi, M . , 1993. Species abundance patterns and community structure. Advances in Ecological Research, 24: 11-186. VanDine, D.F., 1985. Debris flow and debris torrents in the Southern Canadian Cordillera. Canadian Geotechnical Journal, 22: 44-68. van Hulst, R., 1978. On the dynamics of vegetation: Patterns of environmental and vegetational change. Vegetatio, 38(2): 65-75. Veblen, T.T and Ashton, D.H., 1978. Catastrophic influences on the vegetation of the Valdivian Andes, Chile. Vegetatio, 36(3): 149-167. Wilford, D.J. and Schwab, J.W., 1982. Soil mass movements in the Rennell Sound area, Queen Charlotte Islands, British Columbia. Canadian Hydrology Symposium '82, p. 521-541. Young, S., 1992. Slope stability prediction techniques for forest management purposes - A case study from the Queen Charlotte Islands, British Columbia. Unpublished M.Sc. thesis, University of British Columbia, Vancouver, 150 p. Ziemer, R.R., 1981. Roots stability of forested slopes. In T.R.H. Davies, A.J. Pearce, (eds.) Erosion and Sediment Transport in Pacific Rim Steeplands, Proceedings. USDA Forest Services, International Association of Hydrological Sciences, no.132, p. 343-361. Maps: GVRD, 1999a (May). Coquitlam Watershed Topographic Detail, 1:25 000. Arc/INFO 7.2.1 datafile. GVRD, 1999b (May). Coquitlam Watershed Forest Inventory Series, map 8 of10,1:10 000. Datum NAD83 U T M , Inventory source photo 1992, 1:15 000 B&W, Timberline Starter Kit. Arc/INFO 7.2.1 datafile. Specialised Software: ter Braak, C.J.F. and Smilauer, P., 1998. CANOCOfor Windows Version 4.0. Centre for Biometry Wageningen, CPRO-DLO, Wageningen, Netherlands. Website: Palmer, M.W., 2000. Ordination Web page, http://www.okstate.edu/artsci/botany/ordinated, (last visited April 2000). 84 Appendix 1: List of species Latin names, abbreviations and common names (Pojar and MacKinnon, 1994) Species Latin name Abbreviation Species Latin name Abbreviation Tree species Herb species : forbs Pseudotsuga menziesii ssp.menziesii Pseu m Adenocaulon bicolor Aden b Taxus brevifolia Taxu b Anaphalis margaritacea Anap m Thuja plicata Thuj p Aruncus dioicus Arun d Tsuga heterophylla Tsug h Boykinia elata Boyk e Circaea alpina Circ a Tree seedlings Claytonia sibirica Clay s Thuja plicata d-Thuj Cornus canadensis Corn c Tsuga heterophylla d-Tsug Dicentra formosa Dice f Epilobium angustifolium Epil a Shrubs species : Young trees Galium trijlorum Gali t Abies amabilis b-Abie Lactuca muralis Lact m Chamaecyparis nootkatensis b-Cham Linnea borealis Linn b Pseudotsuga menziesii ssp.menziesii b-Pseu Maianthemum dilatatum Maia d Taxus brevifolia b-Taxu Petasites palmatus Peta p Thuja plicata b-Thuj Smilacina stellata Smil s Tsuga heterophylla b-Tsug Streptopus spp.** Stre s Acer circinatum b-Acer Tiarella trifoliata Tiar t Alnus crispa ssp. sinuata b-Alnu Trifolium repens Trif r Viola spp.*** Viol s Shrubs species Gramineae Gram Oplopanax horridus b-Oplo Carex mertensii Care m Rosa gymnocarpa b-Rosa Rubus parviflorus b-Ru pa Herb species : ferns Rubus pedatus b-Ru pe Asplenium viride Aspl v Rubus spectabilis b-Rub s Athyrium filix-femina Athy f Sambucus racemosa ssp. pubens b-Samb Blechnum spicant Blec s Vaccinium spp.* b-Vac sp Criptogramma crispa Crip c Vaccinium parvifolium b-Vac p Cystopteris fragilis Cryo f Dryopteris expansa Dryo e Gymnocarpium dryopteris Gymn d Polystichum munitum Poly m Thelypteris phegopteris Thel p Bryophytes Bryo * Vaccinium spp. = Vaccinium alaskaense, Vaccinium ovalifolium, Vaccinium membranaceum, ** Streptopus spp. = Streptopus amplexifolius, Streptopus roseus, ***Viola spp.*** = unknown Viola spp. 85 Appendix 1 (cont'd): List of species Latin names, abbreviations and common names. Species Latin name Species common name Abies amabilis Acer circinatum Adenocaulon bicolor Alnus crispa ssp. sinuata Anaphalis margaritacea Aruncus dioicus Asplenium viride Athyrium filix-femina Blechnum spicant Boykinia elata Bryophytes Carex mertensii Chamaecyparis nootkatensis Circaea alpina Claytonia sibirica Cornus canadensis Criptogramma crispa Cystopteris fragilis Dicentra formosa Dryopteris expansa Epilobium angustifolium Galium triflorum Gramineae Gymnocarpium dryopteris Lactuca muralis Linnea borealis Maianthemum dilatatum Oplopanax horridus Petasites palmatus Polystichum munitum Pseudotsuga menziesii ssp.menziesii Rosa gymnocarpa Rubus parviflorus Rubus pedatus Rubus spectabilis Sambucus racemosa ssp. pubens Smilacina stellata Streptotus spp.** Taxus brevifolia Thelypteris phegopteris Thuja plicata Tiarella trifoliata Trifolium repens Tsuga heterophylla Vaccinium parvifolium Vaccinium spp.* Viola spp.*** Amabilis fir Vine maple Pathfinder Sitka alder Pearly everlasting Goat's beard Green spleenwort Lady fern Deer fem Coast boykinia Mosses Mertens' sedge Yellow-cedar Enchanter's nightshade Siberian miner's-lettuce Bunchberry Parsley fem Fragile fem Pacific bleeding heart Spiny wood fem Fireweed Sweet-scented bedstraw Grasses Oak fem Wall lettuce Twinflower False lily of the valley Devil's club Palmate coltsfoot Sword fem Douglas-fir Baldhip rose Thimbleberry Five-leaved bramble Salmonberry Red elderberry False solomon's-seal Twinsted stalk spp. Western yew Narrow beech fem Western redcedar Foamflower With clover Western hemlock Huckleberry spp. Huckleberry spp. or Blueberry spp. Violet spp. * Vaccinium spp. = Vaccinium alaskaense, Vaccinium ovalifolium, Vaccinium membranaceum. ** Streptopus spp. Streptopus roseus, ***Viola spp.*** = unknown Viola spp. : Streptopus amplexifolius, Appendix 2: Mean variables measured per sediment trap during monitoring, and used to determine yielded volumes. Sites Depth, X Accumulation Base (cm) = Area (cm2) Volume (cm3) Days of (cm) Angle, C (°) (X * tan (fi )) - ioase~A> 2 = (area * 120) accumulation L2-1 0.5 40 0.6 0.2 20.2 29 4.3 40 5.1 11.1 1328.3 186 4.9 40 5.7 14.0 1675.8 215 * 5.6 40 6.6 18.5 2225.9 305 -0.2 21 -0.6 0.1 9.1 +83 Total accumulation 2235.0 388 L2-2 0.1 37 0.2 0.0 1.3 29 -0.1 37 -0.1 0.0 0.5 186 0.1 37 0.1 0.0 0.7 215 * 0.4 37 0.5 0.1 12.3 305 0.6 25 1.3 0.4 44.6 +83 Total accumulation 56.9 388 L2-3 0.6 30 1.0 0.3 35.9 29 9.8 30 17.0 83.5 10014.7 186 10.4 30 18.0 93.4 11213.3 215 * 11.8 30 20.4 120.1 14409.0 305 0.6 17 2.1 0.7 80.4 +83 Total accumulation 14489.4 388 L2-4 0.6 28 1.1 0.3 36.5 29 4.6 28 8.5 19.3 2319.8 186 5.1 28 9.5 24.4 2928.9 215 * 6.3 28 11.8 37.6 4506.5 305 0.6 13 2.6 0.7 88.9 +83 Total accumulation 4595.5 388 U3-1 2.1 22.7 5.0 5.3 638.6 28 5.1 22.7 12.1 30.8 3694.2 184 7.2 22.7 17.2 61.7 7404.8 212 * 7.3 22.7 17.6 64.6 7752.6 302 2.2 16.0 7.8 8.8 1051.5 +081 Total accumulation 8804.0 383 U3-2 1.0 27.2 1.9 0.9 113.0 28 7.1 27.3 13.8 49.2 5899.9 184 8.2 27.3 15.8 64.3 7710.5 212 * 9.7 27.3 18.7 90.3 10830.2 302 0.0 5.7 0.3 0.0 0.5 +081 Total accumulation 10830.7 383 U3-3 2.0 54.0 1.4 1.4 167.9 28 2.2 54.0 1.6 1.7 209.4 184 4.2 54.0 3.0 6.3 752.3 212 * 5.0 54.0 3.6 9.0 1078.9 302 0.1 12.2 0.5 0.0 2.8 +081 Total accumulation 1081.7 383 87 Appendix 2 (cont'd): Mean variables measured per sediment trap during monitoring, and used Sites Depth, X Accumulation Base (cm) A r e a ( c m 2) Volume (cm3) Days of (cm) Angle, B (°) = (X * tan (B )) ~ ( oase-A ) 2 = (area * 120) accumulation U3-4 0.6 32.2 0.9 0.3 31.1 28 0.2 32.2 0.3 0.0 3.2 184 0.8 32.2 1.3 0.5 61.1 212 * 1.1 32.2 1.8 1.0 123.4 302 0.2 14.3 0.7 0.1 7.9 +081 Total accumulation 131.3 383 L3-5 0.1 32.0 0.2 0.0 1.0 28 3.3 32.0 5.3 8.6 1037.8 184 3.4 32.0 5.4 9.2 1107.3 212 * 4.6 32.0 7.4 17.1 2050.2 302 -0.5 11.8 -2.4 0.6 73.3 +081 Total accumulation 2123.6 383 L3-6 0.0 37.5 0.0 0.0 0.1 28 0.2 37.5 0.2 0.0 2.7 184 0.2 37.5 0.3 0.0 3.0 212 * 0.7 37.5 0.9 0.3 37.4 302 -0.3 37.5 -0.4 0.1 6.3 +081 Total accumulation 43.7 383 L3-7 0.7 41.1 0.7 0.24 29.1 28 ** 184 ** 212 * 10909.2 302 3.6 24.6 8.0 14.5 1710.2 +081 Total accumulation 12619.4 383 L3-8 0.4 26.8 0.7 0.1 16.3 28 9.9 26.8 19.5 96.1 11536.9 184 10.2 26.8 20.2 103.3 12400.6 212 * 11.7 26.8 23.2 136.4 16363.7 302 3.5 12.7 15.6 27.4 3285.9 +081 Total accumulation 19649.7 383 L4-1 1.0 49.9 0.8 0.4 50.9 27 2.9 49.9 2.4 3.5 422.2 186 3.9 49.9 3.3 6.4 766.4 213 * 5.4 49.9 4.5 12.1 1447.5 302 4.2 49.9 3.5 7.3 876.6 +82 Total accumulation 2324.1 384 L4-2 0.1 37.5 0.1 0.0 0.3 27 1.3 37.5 1.7 1.1 133.0 186 1.3 37.5 1.7 1.1 136.4 213 * 1.7 37.5 2.3 2.0 236.1 302 1.2 49.9 1.0 0.6 69.2 +82 Total accumulation 305.3 384 88 Appendix 2 (cont'd): Mean variables measured per sediment trap during monitoring, and used Sites Depth, X Accumulation Base (cm) = A r e a ( c m 2) Volume (cm3) Days of (cm) Angle, B (°) (X * tan (B)) (Dase-A) 2 = (area * 120) accumulation U4-3 0.5 48.9 0.4 0.1 13.7 27 0.7 48.9 0.6 0.2 28.1 186 1.2 48.9 1.1 0.7 81.2 213 * 2.4 48.9 2.1 2.5 297.1 302 2.0 49.9 1.7 1.8 210.5 +82 Total accumulation 507.6 384 U4-4 1.0 47.3 0.9 0.4 53.0 27 1.9 47.3 1.8 1.7 202.3 186 2.9 47.3 2.7 3.9 462.4 213 * 3.9 47.3 3.6 6.9 828.6 302 3.3 49.9 2.8 4.6 556.9 +82 Total accumulation 1385.5 384 U8-1 0.8 33.5 1.2 0.5 58.7 27 1.7 33.5 2.6 2.3 270.2 184 2.5 33.5 3.8 4.8 580.9 211 * 3.2 33.5 4.8 7.6 908.2 302 0.6 7.8 4.2 1.2 143.3 +82 Total accumulation 1051.5 384 U8-2 1.1 29.2 2.0 1.2 140.1 27 0.5 29.2 0.9 0.2 26.0 184 1.7 29.2 3.0 2.4 292.7 211 * 2.2 29.2 4.0 4.5 538.2 302 0.0 6.0 0.3 0.0 0.5 +82 Total accumulation 538.7 384 * trap emptied, and days of accumulation reset ** Damaged trap, yield volume after for 302 days is estimated to be the two-third of L3-8 89 c CD o o o T 3 CD fi co CO 13 CD X ca •c CD t3 s CD fi •3 S3 CD a, a co SO V =L a C N so A V A a O V ) m C N V A a a o o O W IT) C N V A a a o o o o O Ti — i A V a a o o o o o o C N rl V A So" V .2° V •- s «* .5 o 3 M oo H ro oo C N oo N oo n io r-; so >-; <-< C N m ^ m cn © 00 C N C N os in C N t-- m in v~> -t CNi q C N <-< C N C N CN T f in — Os r— os os in in —' ~ t so C<10O « * oo ON oo cn © o oo r~-m in TJ- os os in o\ so vo C N so — so m r-O O ^ f l M T r- r~- oo os m r- m so os co ^- —• in r- i—1 oo i—I O oo r- CN CN in oo OS cn m r-- m CN cn in cn Os in cn CO cn cn OS CN oo SO Os r- SO O 00 i—1 co m o so oo CN SO so CN in Os © SO in in © r-; oo so © in rn in Os sq cn q sq CN 00 oo CN sq C<i CN CN CN CO CO CN co cn CN in CN co CN CN CN CN cn rn cn cn rn in •sr CN cn 00 in OS 00 oo © CO so Ti- i—1 TT r- oo so r- CN o 00 oo Os m OS cn Ti- T t OS i-H o OS 00 00 r- OS r- o CN OS in CN O T f r~ r~ o so Os CN en sq in oo r-^  CN os OS Os rn in q cn oo in in Os cn q oo CN CN i—i 1—< rn CN i—1 CN 1 - 1 CN r—< © * " - CN CN CN 00 S O m O S O S O CN O S r-~ t~- in o CN oo m © m m 00 CN CN S O S O o O s O S 00 o r- CN cn O S r- © O s CN © CN r~ S O O S O S oo q o CN rn S O r~; q cn CN S O cn s q CN O s S O © in CN S O CN ~* in rn in S O in in CN -1 CN CN cn cn CN S O in s d cn i t S O 00 so CN oo m © OS r- t r- SO so in in Os so SO © OS OS m © © OS © in so CN m m in m m CN so SO r- © OS in in © CN r-. CN in OS © CN © m © CN CN in so r-; © sq rn Os © sq Os OS CN SO CN sq cn CN CN CN CN in rn in ro CN CN CN rn cn rn rn in r^' SO r^" ro so SO OS CN © m m CN cn oo CO OO in cn r- Os in SO m SO Os Os © SO 00 Os CN oo r~- SO oo SO CN OS CN CN © 00 00 1—1 oo in in cn in m TT 00 so oo m p~ rn cn OS CN in r-; OS co © so © © OO Os © oq Os in OS rn CN CN SO in rn in cn CN CN CN sd in in sd SO cn SO rn in m r- OS CN m OS o cn oo >n 00 oo © cn Os oo Tt- oo OS OS Os CN m m © P- OS SO in CN Os r- CO CN © OS r~- so ^-i m in SO CN in Os in in SO CN oo rn cn cn i—t 00 00 CN cn oo sq cn so so sq © © Os sq rn © in rn in CN in CN CN ©' •-< rn CN •<t oo so © in SO CN CO oo ro CN © m m © _ © CN r- CN Os r~ CN © cn r- © Os m rn Os CN oo cn O O cn OS SO m oo in 00 OS cn © CN CN m Os © © sq Os oo in Os oo CN r-; OS cn cn OS rn m sq CN CN CN CN CN in < CN rt CN CN CN cs CN rn CN in in cn in o cn CN —< sq sq cn ©' so cn Os © SO oo OS m © SO -3" cn OS in oo oo OS *—i m © © © cn CN CN SO © m SO © cn oo OS © © © m rn CN oo TJ" © CN m CN sq 00 CN r- m Os cn 00 q 00 cn in rn 00 cn in in cn CN cn © SO © r~ CN © cn in in CN in so SO SO SO 00 rn in cn cn CN CN CN CN CN CN CN *"H CN CN CN CN CN CN CN CN CN CN CN CN CN -<oo ( s | o - ' © m s o i n m © r - r ~ - o s t - ~ o \ i n T t c N i n © r - o o m o s © o o SO CN SO oo CN CN m © 00 in cn © CN cn SO CN SO SO m Os 00 so cn in in cn so CN oo CN cn cn m m CN oo CN •—1 in f—< in r~- Tf, in CN Tt- in sq SO Os © CN r-~ © r~ CN sq Os oo OS 00 Os oo 00 CN SO sd 00 oo Os Os OS OS Os oo OS OS Os oo CN CN CN CN CN CN CN *—1 .—i CN CN CN CN CN CN CN CN CN CN CN CN CN CN Os in Os SO CN in CN oq cn 00 in ro © CO CN CN © so q © © © © q © r-~ © © SO oo in" in © in OS CN Os in ©' CO so OS CO in SO od cn —^  1-^  CN •<t 00 Os oo CN m cn © © m SO m CN 1—1 ro r-- 1—1 T f 00 © SO 1-^  © m i—i oo 1—1 oo CN Os oo ro CN m m CN m m r- CN so so m CN r- r- CN CO CN in CN m m OS so © © so m CN m i-H 1—1 CN Os m Os so _ OO in OS CN 1-H r- oq CN OS © © © q © © © © © © in © © •*r od OS OS rn © r-' od Os Os SO CN od in Os 00 co OS oo Os OS Os OS Os in in d *-^ 00 m so cn cn ro CN 00 ro 00 SO Os © SO oo © CN 00 SO CN SO CO so CN in SO oo in so Os cn r- Os i—< CO in CN Os © 1—1 Os © © oo cn in m CN oo rn 1—1 CN —"' '—, OS cn cn CO SO SO © CO in oo C N C N in in in SO cn in in od CN d d Os SO cn in Os © in so SO 00 in oo © 00 so oo cn o m © Os m in so C O - - CN —* i—• CN © o © © CN m — CN cn ^- m ~p t-H H H H H H H J J J J J 5 S 00 00 C O © S O in CN >n O S t-^  CN d O s S O sd CO* C O so d r-" 1—1 m m O s in r- r- © O s so m O s CN O S CN in m S O O S m CN C O in m CN CN cn CN (N CM —"* CN —* CN m ,—1 (N m H H H H H CN CN m m cn —1 -J in SO H H H m m cn D D a x> a r—1 CN m •—i CN CN m r—* CN cn in -—• CN m H H H E- H H H E— H H E— H H H H H >n m S O S O S O so S O r- r- C O C O C O i—1 D —1 J —1 —1 J J D D —1 i j J (D bO at CU -*-» X (D fi CD _ f i fi o o 90 a. ro vo V E £ a. a. vi ro rs N© * H A V E £ a. a. o m C N r-H V A a. a. o o o ui in CM V A E S a. a. o m '-< A V E E a. a. o o o o o o C N r-H V A •- JS 6 * « J 3 .E | V 6B Q s-U S I a V S E <u S3 E 60 60 ~ H i—. _« © m t OO OO (N ^ Oj 0\ Tt ^  Tf- IT) CN 00 N O CN CN CN O ON ro oo in rO ~H •rr ON OO -rr; 00 CN CN — ' CN OO N O — © co — O N — i rn* n oo t~-in in -Tf' co 0O *o o >n O N O O ro ^ / I « O ro ro ro ro ro NO O ro v© O "< m « q « -m N N o m c~ O N c-oo in r-; ~ H T J -N m r - ' d O N f - N O 00 VO CN N O VO N O vo C N C N r> oo M ON O O 1 CO CN CN t- in o CN O 1—H <n CO CN ^H CO oo ON m O VO CN O N r» O CN O 1—< O N r-- vo oo ON t-- CN co •/I r- oo m O ON ON r—< in CN VO ON ro 00 CN CO VO 00 in q u-i CO vd CN ro CN r-< CN O ro r-H d d CN CN CN d CN CN ON m — oo CN ON r-o >o t—i N O ON O N m in O VO O CN O N >n CN CO >n CO in »-H CN O VO -<t NO o CO OO 00 r~- NO 00 VO ^1 oo O N t-; r-; N O CN in co in oo CN q o CN q CO CO CN CO CN 1~~t CN CN CN d CO CN d d CN CN CO d CN d O CN ON ON r- •<f O N CN NO oo ON CO in CO OO o -<1- r-~ r—H ON vo CN co CO >n CO oo CN o vo o CN m >n CN ON oo CN m in O N N O rO in >n CN ON OO CN oo O N oo O N >/-> >n vd in CN vd >n r—i CO ' " H rO CN d d CN CN CO d CN d CN NO >n O CO o f~ CN oo N O r~ CN CO oo O o VO ON in CO f- in CN 00 CO O N CN CN CO ON O >n CO O N vq CO ro q ON 00 oq CN oq CO CN N O oo in CN CN CO t ' CO CO CO CO <-< in —< CO d d CO d d O — C N m vo oo O C N O o d d q co o o d d CO -^t-co o — CO o d d in >n >n oo O N co NO in co in CO CO NO o o o CN CO CN CN 00 t-H m 00 ON oo >n oo VO i N O oo m vo N O VO •sr. O N vq ON »—J CO y—t CN m oo >n in CO CN CN CN ro CO CO d CN in CO d d CN N O in d >n d C N in in o f- o CN1 C \ o CO CN oo ON >n rH O N CN CN o o VO O N ON O N CN CN r~- o VO CO T—t OO CN CO CN r- CN N O CO »n r—1 N O CN O N CN r-; O N CO VO O N oo o o m CN oo ON CN oo CO CN r—1 r—1 CO in CN d vd CN CN d d CN vd vd d in d o CO 6 - H 6 N O ON o o o o vo ro oo O N C N >n co O N O N VO OO C - ro NO N C ^ r~ t~- T f -H -H rs vo in O ^ - O v B O O r-H r~- q r-< 00 vi T t « vi d CN CN CN CN CN O N ro o r—< r- CN o O N CN OO ON r^ m CN in >n O N ro oo r—t o CN o r- r-H oo O N VO CO in ON OO .—< oo r-- m OO O m CN in q vq t-- vq 00 >n >n O N r-; 00 o NO CN CO o O o oo oo ON vd in vd oo in r-" in >-< r-! CN r - » r^ CO r^ ro — ; —) VO o CN P~- r ^ o ON ON CN o CN oo oo oo CN N O in ON ON CO r — < O N o ro oo r- CO N O m ro O N t-; f-; ON vq in ON CN vo >n in ro NO CN O CN CO CN vd CO oo ro CN CN r~: oo d >n in CN CN CN CN CN CN r—i CN CN CN CN CN m o CO m oo _^ •r. CN CN ON CO ro o O N r -H ro CO -3; CN N O ON CN ro in r O NO in r> ro r J ro r ^ " — 1 CN CN CN o CI oo ro r-H O ON OO 0O NO o v-i r- 3; ON O CN >n r^ ON r^ N O 00 NO O N CN CN O N VO CN O ON O o o in m o CO vo ON CN CO oo ON NO CN O N CN CN CO oo O N CO >n CN o N O ON CN r-H 00 ro ro ro ON r—i CN CN ON CN CN VO O N r-H co ro O N ro r- r-; N O m ro vq o 00 O N N O NO 00 O N ON ON ON ON NO 00 O N d CN r-' O N CN ro ON vd oo in vd CN CN CN CN CN CN CN CN CN CN CN CN CN CN CN co r-< CN CN CN CN CN NO >n 00 q r-H >n rt vq q q m q r- q ro q oo oo ON O N oo CN NO d NO d oo ON in co in ON O N N O vd d r—< r-H in ro r^ d d CN o 00 o ON CN ON ON ON CN CO o oo CN r^ ro in ro r—i ro ro CN CO co m >o CO co N O NO O N N O o o q CN q o o - H in O r-i d d O N ON r^ d O N •<t d ON 00 co in r^ O N ro oo o >n oo ON ro CN CN ro CN CN N O •<* 00 t ro O r~; O N q CN ro r-H N O 00 CN NO ON ON O N ON 00 ON ON oo vd O N ro oo O N oo 00 oo O N O N ON vd oo m ON ro r—1 oo O N in oo in CN m ON oo H q vq vq f- ; q q O N o CN in oo O N ON CN O N d d in vd ON d CN r—1 in NO in d oo d ON d vd ON NO 00 r—t m ON r-- ON in 00 r-H CN TP r-H in r» r~ ON 00 00 ON r—1 rH r->n r-- m NO vo •3- r- r-. 00 CN r- »-H r—t CO r—i NO ro ON VO O CN o CN r-H CN CN rH CO CN vo O N r~-— CN ro 2 2 g H h H -t-" r - r-1 r> n o H ON H m CO CN CN CO O CO 00 in v i o co vd t^ -t -<t i^-in vo oo r D - O H O oo oo oo 00 00 oo D D D D P O CN ro J J J m VO H E-ON ON ON D D O CN ro CN ro H H H O O O CO CO -J —1 —1 •J r - CO r - CN ro m r- oo — CN ro •<!• CN L2-L2-L2-L3-L3-L3-L3-L3-L3- ro -J 3 3 3 3 L8-L8-- O C3 -o t 3 O o •a a Si > T 3 r Q V O ro -J 1 3 0 a CN -J PH u -f-» a a> 6 •3 u t>0 CD fi x CD CD l-H c3 cfl a o '3 > CD VH •8 ID 'o CD a, & o 6b CD s CD CD r-C CD > O o ai _<D 'o CD OH CO CJ) fi o I CD XI 60 fi cfl &, CvJ ed JO V-C CD > o CD CO CD C 'o CD pe Spi Ap •a fi CD OH OH < OO 00 O 5 a • • CN CN o <N (% ) J3A0D 3 A U C ] 3 J UB3JAJ 92 d <u OH 9 m CN CN o CN U 00 00 o 13 O 00 00 o IS tkkkklklllllllWSSSSSSSSSSSl i,kkikkSkkkkkksmssssssssssssisssss| ksssssssssssiiisiiiiwiss^ sssssisissississiiiiiw •svssssvssssskkkkllsllllllSSSSSkkSkkSkSkSSSSSSSSS| in r-" —I— m CN in r--in CN (%) I9AO0 3AIJEJ3J UEajAJ 8nsx-p fnuj,-p d|=iU. U1<|0J a uiu^o 3 0A1Q J»*D 3dii3 J *mv A (dsy UJ 3 J B ^ E O J M°!A JJ"I SMS S[rais drod pBIBt\J q UUJT H|E0 E|ida 3 IUO3 p un jy ui deuy q i u p y d3EA-q lU3Bj\-q quics -q B j n ^ - q esoy -q oido-q nujv -q j a o y q 8nsi-q •q nxcx-q n M j - q ureiQ-q 3 i q y - q q S n s i dfnuj. qnxBX Appendix 4b: Species cover - plant species cover per quadrat. Pseu m Taxu b Thuj p Tsug h b-Abie b-Cham b-Pseu b-Taxu b-Thuj b-Tsug b-Ac L1T1Q1 0 0 0 0 0 0 17.5 0 2.5 2.5 0 L1T2Q1 0 0 0 0 0 0 0 0 62.5 7.5 0 L1T3Q1 0 0 0 0 0 0 0 0 0.1 0 0 L1T4Q1 0 0 0 0 0 0 0 0 0 0 0 L1T5Q1 0 0 0 0 0 0 0 0 2.5 7.5 0 U1T6Q1 0 0 0 0 0 0 0 0 0 0 0 U1T7Q1 0 0 0 0 0 0 0 0 2.5 2.5 0 U1T8Q1 0 0 0 0 0 0 0 0 0 0.1 0 L1T1Q2 0 0 0 0 0 0 0 0 2.5 0 0 L1T3Q2 0 0 0 0 0 0 2.5 0 7.5 0 0 L1T1Q3 42 0 0 0 0 0 2.5 0 7.5 2.5 0 L1T3Q3 42 0 0 0 0 0 0 0 2.5 17.5 0 L4T3Q3 87.5 0 0 0 0 0 17.5 0 7.5 0 0 U1T6Q3 0 0 17.5 62.5 2.5 0 0 2.5 2.5 17.5 0 L2T1Q1 0 0 0 0 0 0 0 0 0 17.5 0 L2T2Q1 0 0 0 0 0 0 0 0 0.1 0 0 L2T3Q1 0 0 0 0 0 0 0 0 0 0 0 L2T1Q3 0 0 0 0 2.5 0 0 0.1 7.5 7.5 0 L2T2Q3 0 0 0 0 0 0 0 0 17.5 17.5 0 L2T3Q3 0 0 0 0 17.5 0 0 2.5 7.5 42 0 L3T1Q1 0 0 0 0 0 0 0 0 2.5 0 0 L3T2Q1 0 0 0 0 0 0 0 0 2.5 0.1 0 L3T3Q1 0 0 0 0 0 0 0 0 0 2.5 0 U3T4Q1 0 0 0 0 0 0 0 0 0 0 0 U3T5Q1 0 0 0 0 0 0 17.5 0 0 0 0 U3T6Q1 0 0 0 0 0 0 0 0 0.1 0.1 0 L3T2Q3 0 0 0 0 0 0 0 0 42 17.5 0 L4T1Q1 0 0 0 0 0 0 17.5 0 ' 17.5 0 0 L4T2Q1 87.5 0 0 0 0 17.5 2.5 0 29 42 0 L4T3Q1 87.5 0 0 0 0 0 0 0 7.5 0 0 L1T4Q3 42 0 0 17.5 0 0 2.5 0 29 17.5 0 U4T4Q1 0 0 0 0 0 0 0 0 17.5 7.5 0 U4T5Q1 0 0 0 0 0 0 0 0 2.5 2.5 0 U4T6Q1 0 0 0 0 0 0 0 0 0 0 0 L5T1Q1 0 0 0 0 0 0 0 0 2.5 0 0 L5T3Q1 0 0 0 0 0 0 0 0 7.5 0 0 L 6 T l Q l a 0 0 0 0 0 0 0 0 2.5 7.5 0 L 6 T l Q l b 0 0 0 0 0 0 0 0 2.5 7.5 0 L 6 T 2 Q l a 0 0 0 o. 0 0 0 0 0 0 0 L 6 T 2 Q l b 0 0 0 0 2.5 0 0 0 0.1 2.5 0 L6T3Q1 0 0 0 0 2.5 0 0 0 0.1 2.5 0 L6T3Q2 0 0 0 0 0 0 0 0 2.5 0.1 29 L6T2Q3 0 0 0 0 17.5 2.5 0 0 42 29 0 L6T3Q3 0 0 0 0 0 0 0 0 29 17.5 0 L7T1Q1 0 0 0 0 0 0 0 0 17.5 42 0 L7T2Q1 0 0 0 0 0 0 0 0 2.5 0 0 L7T3Q1 0 0 0 0 0 0 17.5 0 29 17.5 0 U7T4Q1 0 0 0 0 0 0 0 0 7.5 17.5 0 U7T5Q1 0 0 0 0 0 0 2.5 0 7.5 7.5 0 U7T6Q1 0 0 0 0 0 0 0 0.1 2.5 7.5 0 L7T2Q3 62.5 0 0 29 0 0 0 0 42 17.5 0 U7T5Q3 0 7.5 7.5 29 7.5 0 0 2 2.5 87.5 3 L8T1Q1 0 0 0 0 0 0 0 0 2.5 0 42 L8T2Q1 0 0 0 0 0 0 0 0 0 0 0 L8T3Q1 0 0 0 0 0 0 0 0 0 0 0 L8T2Q2 0 0 0 0 0 0 0 0 0 0 0 L8T3Q2 0 0 0 0 0 0 0 0 0 0 0 94 Appendix 4b (cont'd): Species cover - plant species cover per quadrat. b-Alnu b-Oplo b-Rosa b-RuPa b-RuPe b-RubS b-Samb b-Vac spp b-Vac p Aden L1T1Q1 0 0 0 2.5 0 2.5 0 0 0 0 L1T2Q1 0 0 0 7.5 0 0 0 0 0 0 L1T3Q1 0 0 0 29 0 0 0 0 0 0 L1T4Q1 0 0 0 2.5 0 0 0 0 0 0 L1T5Q1 0 0 0 2.5 0 2.5 0 0 0 0 U1T6Q1 0 0 0 0 0 0.1 0 0 0 0 U1T7Q1 0 0 0 17.5 0 7.5 0 0 0 0 U1T8Q1 0 0 0 0.1 0 2.5 0 0 0 0 L1T1Q2 0 0 0 42 0 17.5 0 0 0 0 L1T3Q2 0 0 0 62.5 0 17.5 0 0 0 0 L1T1Q3 0.1 0 0 0 0 2.5 0.1 0 0 0 L1T3Q3 0 0 0 ' . 0 0 2.5 0.1 0 0 0 L4T3Q3 0 2.5 0 0 0 2.5 0 0 0 0 U1T6Q3 0 2.5 0 0 0 2.5 0 0 7.5 0 L2T1Q1 0 0 0 0 0 0 0 0 0 0 L2T2Q1 0 0 0 0 0 0.1 0 0 0 0 L2T3Q1 0 2.5 0 17.5 0 17.5 0 0 2.5 0 L2T1Q3 0 0 0 2.5 0 17.5 0.1 2.5 7.5 0 L2T2Q3 0 0 0 0 0 2.5 0 0 0 0 L2T3Q3 0 0 0 0 0 2.5 2.5 0 7.5 0 L3T1Q1 2.5 0 0 0 0 0 0 0 0 0 L3T2Q1 0 2.5 0 0 0 0 0 7.5 0 0 L3T3Q1 7.5 0 0 0 0.1 0 0 0 2.5 0 U3T4Q1 0 0 0 0 0 7.5 0 0 0 0 U3T5Q1 0 2.5 0 0 0 17.5 0 0 0 0 U3T6Q1 0 0 0 0 17.5 17.5 0 17.5 0 0 L3T2Q3 > o 0 0 0 0.1 0.1 0 0 0 0 L4T1Q1 0 0 0 17.5 0 2.5 0 0 0 0 L4T2Q1 0 0 0 0 0 0 0 0 0 0 L4T3Q1 0 0.1 0 2.5 0 2.5 2.5 0 0 0 L1T4Q3 0 0 0 2.5 0 7.5 0 2.5 0 0 U4T4Q1 0 7.5 0 0 0 29 0 0 0 0 U4T5Q1 0 29 0 0 0 17.5 17.5 0 0 0 U4T6Q1 0 0 17.5 62.5 0 0 0 17.5 0.1 0 L5T1Q1 0 0.01 0 42 0 7.5 0 0 0 0 L5T3Q1 0 2.5 0 7.5 0 2.5 0 0 0 0 L 6 T l Q l a 0 0 0 0 0 2.5 0 0 0 0 L 6 T l Q l b 17.5 0 0 0 0 2.5 0 0 2.5 0 L 6 T 2 Q l a 0 0 0 17.5 0 7.5 0 0 0 0 L 6 T 2 Q l b 0 0 0 0 0 0 0 2.5 0 0 L6T3Q1 0 0 0 0 0 0 0 0 0 0 L6T3Q2 0 0 0 29 0 7.5 0 0 0 0 L6T2Q3 0 0 0 0 0 17.5 0 2.5 7.5 0 L6T3Q3 0 0 0 0 42 29 0 7.5 7.5 0 L7T1Q1 29 0 0 2.5 0 7.5 . 0 0 0 0 L7T2Q1 0 0 0 7.5 0 7.5 0 0 0 0 L7T3Q1 17.5 0 0 0 0 7.5 0 0 7.5 0 U7T4Q1 0 0 0 0 0 2.5 0 0 0 0 U7T5Q1 0 0 0 0 0 7.5 0 0 0 0 U7T6Q1 0 0 0 0 0 0 0 0 0 0 L7T2Q3 0 0 0 2.5 0 2.5 2.5 0 0 0 U7T5Q3 0 0 0 0 2.5 2.5 0 0 2.5 0 L8T1Q1 0 0 0 29 0 0 0 0 0 0 L8T2Q1 0 0 0 17.5 0 17.5 0 0 0 0 L8T3Q1 0 0 0 17.5 0 0 0 0 0 0 L8T2Q2 0 0 0 7.5 0 62.5 0 0 0 0 L8T3Q2 0 0 0 29 0 0 2.5 0 0 0 Appendix 4b (cont'd): Species cover - plant species cover per quadrat. Anap m Arun d Boyk e C i r c a Clay s Corn c Dice f E p i l a G a l i t Lact L1T1Q1 17.5 0.1 2.5 0 0 0 0 2.5 0 0 L1T2Q1 17.5 0 2.5 0 0 0 0 0 0 0 L1T3Q1 2.5 0 2.5 0 0 0 0 2.5 0 0 L1T4Q1 17.5 0 0 0 0 0 0 0.1 0 17.5 L1T5Q1 7.5 0 42 0 0 0 0 0 0 0 U1T6Q1 0 2.5 2.5 2.5 0 0 0 0 0 0 U1T7Q1 0 2.5 17.5 2.5 0 0 2.5 0 0 2.5 U1T8Q1 0 2.5 0.1 0 0 0 0 0 0 2.5 L1T1Q2 0 0 0 0 0 0 0 17.5 0 0 L1T3Q2 0 0 2.5 0 0 0 0 2.5 0 0 L1T1Q3 0 0 0 0 0.1 0 0 0 0 0 L1T3Q3 0 0 0 0 0.1 0 0 0 0 0 L4T3Q3 0 0 0 0 0 0 0 0 0 0 U1T6Q3 0 0 0 0 0 2.5 0 0 0 0.1 L2T1Q1 2.5 0 29 0 0 0 0 0 0 0.1 L2T2Q1 2.5 0 2.5 0 0 0 0 0 0 0.1 L2T3Q1 2.5 0 17.5 0 0 2.5 0 2.5 0 2.5 L2T1Q3 0 0 0 0 0 2.5 0 2.5 0 0 L2T2Q3 0 0 0 0 0 17.5 0 2.5 0 0 L2T3Q3 0 0 0 0 0 17.5 0 0 0 0 L3T1Q1 2.5 0 2.5 0 0 0 0 0 0 0 L3T2Q1 7.5 2.5 7.5 0 0 2.5 0 2.5 0 0 L3T3Q1 0 7:5 7.5 0 0 7.5 0 0 0 0 U3T4Q1 0 17.5 0 0 0 0 0 0 0 0 U3T5Q1 0 7.5 2.5 0 0 0 0 0 0 0 U3T6Q1 0 7.5 2.5 0 0 0 0 0 0 0.1 L3T2Q3 0 0 0 0 0 0.1 . 0 • 0 0 0 L4T1Q1 7.5 7.5 0 0 0 0 2.5 2.5 0 2.5 L4T2Q1 0 7.5 7.5 0 0 0 0 0 0 2.5 L4T3Q1 0 0 0 0 0 0 0 0 0 0 L1T4Q3 0 0 0 0 0 2.5 0 0 0 0 U4T4Q1 2.5 0 7.5 0 0 0 0 0 0.1 7.5 U4T5Q1 0 0 7.5 0 0 0 0 0 0 2.5 U4T6Q1 0 0 2.5 0 0 2.5 0 0 0 0.1 L5T1Q1 7.5 2.5 0 0 0 0 0 0 0 2.5 L5T3Q1 2.5 7.5 17.5 0 0 0 0 0 0 2.5 L 6 T l Q l a 7.5 17.5 2.5 0 0 0 0.1 2.5 0.1 2.5 L 6 T l Q l b 7.5 2.5 29 0 0 0 0 2.5 0 0 L 6 T 2 Q l a 0.1 0 0 0 0 2.5 0 0 2.5 0.1 L 6 T 2 Q l b 2.5 7.5 17.5 0 0 0 0 0 0 0 L6T3Q1 2.5 29 7.5 0 0 0 2.5 0.1 2.5 7.5 L6T3Q2 7.5 2.5 7.5 0 0 0 0 0 7.5 7.5 L6T2Q3 0 0 0 0 0 7.5 0 2.5 0 0 L6T3Q3 0 0 0 0 0 29 0 2.5 0 0 L7T1Q1 2.5 2.5 17.5 0 0 0 0 0 0 0 L7T2Q1 7-5 2.5 2.5 0 0 0 0 0 0 0 L7T3Q1 0 2.5 7.5 0 0 0 0 0 0 0 U7T4Q1 0 7.5 7.5 0 0 0 0 0 0 2.5 U7T5Q1 2.5 2.5 17.5 0 0 0 0 0 0 0 U7T6Q1 0 0 17.5 0 0 0 0 0 0 7.5 L7T2Q3 0 0 0 0 0 0 0 0 0 0 U7T5Q3 0 0 0 0 0 17.5 0 0 0 0 L8T1Q1 2.5 17.5 7.5 0 0 0 0 2.5 0 2.5 L8T2Q1 2.5 2.5 2.5 7.5 0 0 0 2.5 2.5 0 L8T3Q1 0 2.5 0 0 2.5 0 0 7.5 2.5 0 L8T2Q2 2.5 2.5 2.5 0 0.1 0 0 2.5 0 0 L8T3Q2 2.5 o' 7.5 2.5 2.5 0 0 0 0 2.5 96 Appendix 4b (cont'd): Species cover - plant species cover per quadrat Linn b Maia d Peta p Smil s Stre spp T i a r t T r i f r V i o l spp Gram Care m L1T1Q1 0 0 0 0 0 0 0 0 2.5 2.5 L1T2Q1 0 0 0 0 0 0 0 0 2.5 2.5 L1T3Q1 0 0 0 0 0 0 0 0 2.5 2.5 L1T4Q1 0 0 0 0 0 0 0 0 2.5 0 L1T5Q1 0 0 0 0 0 0 7.5 0 2.5 0 U1T6Q1 0 0 0 0 0 2.5 0 0 0.1 0 U1T7Q1 0 0 0 0 0 2.5 0 2.5 2.5 2.5 U1T8Q1 0 0 0 0 0 0 0 0.1 2.5 2.5 L1T1Q2 0 0 0 0 0 0 0 2.5 7.5 2.5 L1T3Q2 0 0 0 0 0 0 0 0 29 2.5 L1T1Q3 0 0 0 0 0 0.1 0 0.1 0 0 L1T3Q3 0 0 0 0.1 0 0.1 0 0.1 0 0 L4T3Q3 0 0.1 0 2.5 0 0 0 0.1 0 0 U1T6Q3 0 2.5 0 29 0 7.5 0 2.5 0 0 L2T1Q1 0 0 0 0 0 0 0 0.1 2.5 2.5 L2T2Q1 0 0 42 0 0 0 29 0 7.5 0 L2T3Q1 0 0 0 0 0 0 0 0 2.5 0 L2T1Q3 0 0 0 0 0 0 0 0 0 0 L2T2Q3 0 0 0 0 0 0 0 0 0 0 L2T3Q3 0 0 0 0 0 0 0 0 0 0 L3T1Q1 0 0 0 0 0 0 0 0.1 0.1 2.5 L3T2Q1 0 0 0 0 0 2.5 0 2.5 2.5 2.5 L3T3Q1 0 0 0 0 0 2.5 0 2.5 7.5 0.1 U3T4Q1 0 0 0 0 0 2.5 0 2.5 7.5 0 U3T5Q1 0 0 0 0 0 2.5 0 2.5 0 0 U3T6Q1 0 7.5 0 2.5 0 2.5 0 2.5 0 0 L3T2Q3 0 0 0 0 0 0.1 0 0 0 0 L4T1Q1 0 0 0 0 0 2.5 0 0 2.5 0.1 L4T2Q1 0 0 0 0 0 0 0 0 0 0.1 L4T3Q1 0 0 0 0 0 0 0 0 0 0 L1T4Q3 0 0.1 0 2.5 0 0.1 0 0.1 0 0 U4T4Q1 0 0 0 0 0 2.5 0 2.5 7.5 0 U4T5Q1 . 0 0 0 0 0 7.5 0 2.5 0.1 0 U4T6Q1 0 0 0 2.5 0 7.5 0 2.5 7.5 0 L5T1Q1 0 0 0 0 0 29 0 0 7.5 2.5 L5T3Q1 0 0 0 0 0 0 0 0 2.5 0 L 6 T l Q l a 0 0 0 0 0 0 0 2.5 2.5 0 L 6 T l Q l b 0 0 0 0 0 0 0 0 2.5 0 L 6 T 2 Q l a 0 0 0 0 0 0 0 0 2.5 0 L 6 T 2 Q l b 0 0 0 0 0 0 0 0 2.5 0 L6T3Q1 0 0 0 0 0 0 0 0.1 7.5 0 L6T3Q2 0 0 0 0 0 0 0 0 2.5 0 L6T2Q3 0 2.5 0 0.1 0 2.5 0 2.5 0 0 L6T3Q3 0 0 0 0 0 17.5 0 0 0 0 L7T1Q1 0 0 0 0 0 0 0 2.5 0 2.5 L7T2Q1 0 0 0 0 0 0 0 0 0 0 L7T3Q1 0 0 0 0 0 0 0 0 0 0 U7T4Q1 0 0 0 0 0 2.5 0 2.5 0 0 U7T5Q1 0 0 0 0 0 2.5 0 0 0 0 U7T6Q1 0 0 0 0 0 0 0 2.5 0 0 L7T2Q3 0 0 0 0.1 0 0.1 0 0 0 0 U7T5Q3 0 . 7.5 0 7.5 0 2.5 0 7.5 0 0 L8T1Q1 0 0 0 0 0 0 0 0 2.5 0 L8T2Q1 0 0 0 0 0 0 0 0 2.5 0 L8T3Q1 0 0 0 0 0 0 0 0 2.5 0 L8T2Q2 0 0 0 0 0 2.5 0 0.1 2.5 0 L8T3Q2 0 0 0 0 0 2.5 0 0.1 2.5 0 97 Appendix 4b (cont'd): Species cover - plant species cover per quadrat L1T1Q1 L1T2Q1 L1T3Q1 L1T4Q1 L1T5Q1 U1T6Q1 U1T7Q1 U1T8Q1 L1T1Q2 L1T3Q2 L1T1Q3 L1T3Q3 L4T3Q3 U1T6Q3 L2T1Q1 L2T2Q1 L2T3Q1 L2T1Q3 L2T2Q3 L2T3Q3 L3T1Q1 L3T2Q1 L3T3Q1 U3T4Q1 U3T5Q1 U3T6Q1 L3T2Q3 L4T1Q1 L4T2Q1 L4T3Q1 L1T4Q3 U4T4Q1 U4T5Q1 U4T6Q1 L5T1Q1 L5T3Q1 L 6 T l Q l a L6T1Q1D L 6 T 2 Q l a L 6 T 2 Q l b L6T3Q1 L6T3Q2 L6T2Q3 L6T3Q3 L7T1Q1 L7T2Q1 L7T3Q1 U7T4Q1 U7T5Q1 U7T6Q1 L7T2Q3 U7T5Q3 L8T1Q1 L8T2Q1 L8T3Q1 L8T2Q2 L8T3Q2 Aspl v Athy f 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.1 0 0.1 0 0 0 0 0 0 0 0 0 0 0.1 0 0 0 0 0 0 0 0 0.1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.1 0 0 2.5 17.5 0 0 0 0 0 0 0 0 0 0 0 0 2.5 17.5 17.5 0 0 17.5 7.5 0 17.5 0 0 2.5 0 0 0 0 0 0 0 0 0 0 0 2.5 0 7.5 0 0 0 2.5 Blec s 2.5 0 0 0 0 0.1 2.5 0.1 2.5 0 0 0 2.5 17.5 0 0 0.1 2.5 0 7.5 0 2.5 2.5 7.5 7.5 • 17.5 0 2.5 7.5 0 2.5 0 0 0 0 0 0 2.5 0 0.1 0 7.5 2.5 7.5 0 2.5 7.5 2.5 7.5 0.1 0 17.5 2.5 0 0 0 0 C r i p c 0.1 0 0 0 0 0.1 0 0.1 0 0 0.1 0.1 0 0.1 0 0 0 0.1 0 0 0 0 0 0 0.1 0 0 0.1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.5 0 0 0 0 0.1 0 0 0 0 0 Cryo 0 0 0 0 0 0 0 0.1 0 0 0.1 0.1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.5 0 0 0 0 2.5 0 0 0 0 0 0 0 0 0 0 Dryo 0 0 0 0 0 0 0 0 0 0 0 0 0 2.5 0 0 0 2.5 0 0 0 0 0 0 2.5 2.5 0.1 2.5 0 0 0 0 0 0.1 0 0 0 0 0 0 0 0 0.1 0 0 0 0 0 0 0 2.5 2.5 0 0 0 0 0 Gymn 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.5 0 7.5 7.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.5 0 0 0 2.5 0 0 0 0 0 D Poly m 0 0 0 0 0 0 0 0 0 0 0 2.5 0 2.5 0 0 0 0 0 0 0 0 0 17.5 0 2.5 0 0 0 17.5 2.5 0 7.5 0 0 2.5 0 0 2.5 0 0 0 0 0 0 0 0.1 0 0 0 7.5 2.5 0 2.5 . 0 ' 2.5 0 Thel p 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.5 2.5 0 0 2.5 0 0 0 0 0 0 0 0 0.1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 d-Thuj 0.1 0.1 0.1 0.1 2.5 0.1 2.5 17.5 2.5 2.5 0.1 0 0 0.1 0.1 0.1 0 0 0 0 0.1 0 2.5 0 0.1 2.5 0 0.1 0 0 0 2.5 0 0 0 0 0 2.5 0 0.1 2.5 2.5 0.1 0 2.5 0 2.5 0 2.5 2.5 0 0 0.1 2.5 0 0.1 0.1 Appendix 4b (cont'd): Species cover - plant species cover per quadrat d - T s u g B r y o S o i l L1T1Q1 0.1 17.5 25 L1T2Q1 0 62.5 7 L1T3Q1 0.1 62.5 10 L1T4Q1 0.1 2.5 20 L1T5Q1 2.5 29 7 U1T6Q1 0.1 0.1 85 U1T7Q1 2.5 2.5 5 U1T8Q1 17.5 2.5 20 L1T1Q2 0 2.5 0 L1T3Q2 0.1 2.5 0 L1T1Q3 0 7.5 0 L1T3Q3 0 2.5 0 L4T3Q3 0 2.5 0 U1T6Q3 0 0.1 0 L2T1Q1 0.1 7.5 0 L2T2Q1 0 87.5 15 L2T3Q1 0.1 7.5 6 L2T1Q3 0 0.1 1 L2T2Q3 0.1 0.1 20 L2T3Q3 0 42 5 L3T1Q1 0.1 62.5 25 L3T2Q1 0 7.5 5 L3T3Q1 2.5 29 2 U3T4Q1 0.1 17.5 20 U3T5Q1 0.1 7.5 15 U3T6Q1 2.5 7.5 1 L3T2Q3 0 0.1 0 L4T1Q1 0.1 2.5 50 L4T2Q1 0.01 2.5 0 L4T3Q1 0 2.5 0 L1T4Q3 0 2.5 0 U4T4Q1 0 0 0 U4T5Q1 0 2.5 0 U4T6Q1 0 0.1 0 L5T1Q1 0 2.5 0 L5T3Q1 0 0 0 L 6 T l Q l a 0 2.5 25 L 6 T l Q l b 2.5 62.5 0 L 6 T 2 Q l a 0 0 2 L 6 T 2 Q l b 0.1 29 5 L6T3Q1 0.1 7.5 50 L6T3Q2 2.5 0 5 L6T2Q3 0.1 7.5 0 L6T3Q3 0 2.5 0 L7T1Q1 2.5 2.5 0 L7T2Q1 0 2.5 40 L7T3Q1 2.5 17.5 0 U7T4Q1 0.1 0 5 U7T5Q1 2.5 0 25 U7T6Q1 2.5 2.5 15 L7T2Q3 0 7.5 0 U7T5Q3 0 17.5 0 L8T1Q1 0.1 2.5 5 L8T2Q1 0 7.5 15 L8T3Q1 0.1 0 55 L8T2Q2 0 2.5 7 L8T3Q2 0.1 2.5 20 99 Appendix 4b (cont'd): Species cover - plant species cover per quadrat Pseu m Taxu b Thuj p Tsug h b-Abie b-Cham b-Pseu b-Taxu b-Thuj b-Tsug b-Ace U8aT4Ql 0 0 0 0 0 0 0 0 0 0 17.5 U8aT5Ql 0 0 0 0 0 0 0 0 0 0 0 U8aT6Ql 0 0 0 0 0 0 0 0 0 0 0 U8bT4Ql 0 0 0 0 0 0 0 0 0 0 0 U8bT5Ql 0 0 0 0 0 0 0 0 2.5 7.5 2.5 U8bT6Ql 0 0 0 0 0 0 0 7.5 42 7.5 0 L9T1Q1 0 0 0 0 0 0 0 0 7.5 0 0 L9T2Q1 0 0 0 0 0 0 0 0 17.5 0 0 L9T3Q1 0 0 0 0 0 0 0 2.5 29 7.5 0 U9T4Q1 0 0 0 0 0 0 0 17.5 0 2.5 0 U9T5Q1 0 0 0 0 0 0 0 2.5 0 7.5 0 U9T6Q1 0 0 0 0 0 0 0 0 0 0 0 L10T1Q1 7.5 0 0 0 0 2.5 0 0 17.5 2.5 0 L10T2Q1 17.5 ' 0 0 0 0 0 0 0 62.5 7.5 2.5 L10T3Q1 0 0 0 0 0 2.5 42 2.5 17.5 7.5 0 L13T1Q1 0 0 0 0 0 0 17.5 0 29 29 0 L13T2Q1 0 0 0 0 0 0 0 0 62.5 0 0 L2TR1 0 0 0 0 0 0 0 0 2.5 0 0 L2TR2 0 0 0 0 0 0 0 0 2.5 0.1 0 L2TR3 0 0 0 0 0 0 2.5 0 2.5 0.1 0 L2TR4 0 0 0 0 0 0 0 0 0 0 0 U3TR1 0 0 0 0 0 0 0 0 0 0 0 U3TR2 0 0 0 0 0 0 0 0 0 0 0 U3TR3 0 0 0 0 0 0 0 7.5 7.5 17.5 0 U3TR4 0 0 0 0 0 0 0 0 2.5 7.5 0 L3TR5 0 0 0 0 0 0 0 0 0 0 0 L3TR6 0 0 0 0 0 0 0 0 0.1 0 0 L3TR7 0 0 0 0 0 0 0 0 0 0 0 L3TR8 0 0 0 0 0 0 0 0 0 0 0 L4TR1 29 0 0 0 0 0 0 0 0 0 0 L4TR2 29 0 17.5 0 0 0 0 0 0 0 0 U4TR3 0 42 0 0 0 0 0 0 0 0 0 U4TR4 0 62.5 0 0 0 0 0 0 0 0 0.1 U8TR1 0 0 0 0 0 0 0 0 2.5 2.5 0 U8TR2 0 0 0 0 0 0 0 0 0 0 0 100 Appendix 4b (cont'd): Species cover - plant species cover per quadrat b-Alnu b-Oplo b-Rosa b-RuPa b-RuPe b-RubS b-Samb b-Vac b-Vac p Aden U8aT4Ql 0 0 0 0 0 0.1 0 spp 0 0 2.5 U8aT5Ql 0 0 0 0 0 17.5 0 0 0 0 U8aT6Ql 0 0 0 2.5 0 0 0 0 0 2.5 U8bT4Ql 0 0.1 0 0 0 0 0 0 0 0 U8bT5Ql 0 17.5 0 0 17.5 0 0 0 0 0 U8bT6Ql 0 17.5 0 0 0 0 0 2.5 0.1 0 L9T1Q1 0 0 0 62.5 0 29 0 0 0 0 L9T2Q1 0 0 0 17.5 0 2.5 7.5 0 0 0 L9T3Q1 0 0 0 0 0 17.5 0 0 0 0 U9T4Q1 0 0 0 2.5 0 0 0 0 0 0 U9T5Q1 0 2.5 0 2.5 0 2.5 0 0 0 0 U9T6Q1 0 2.5 0 0 0 0 2.5 0 0 0 L10T1Q1 2.5 0 0 17.5 0 7.5 7.5 7.5 0 0 L10T2Q1 0 0 0 0 0 17.5 0 0 0 0 L10T3Q1 0 0 2.5 29 0 0 0 0 0 0 L13T1Q1 0 0 0 0 0 7.5 0 0 7.5 0 L13T2Q1 0 0 0 0 0 2.5 0 0 0 0 L2TR1 0 0 0 0 0 0 0 0 2.5 0 L2TR2 0 0 0 42 0 0 0 0 0 0 L2TR3 0 0 0 2.5 0 17.5 0 0 0 0 L2TR4 0 0 0 7.5 0 17.5 0 0 0 0 U3TR1 0 0 0 0 0 0 0 0 0 0 U3TR2 0 0 0 0 0 0 0 0 0 0 U3TR3 0 29 0 0 0 2.5 0 2.5 7.5 0 U3TR4 0 2.5 0 0 0 0 0 7.5 2.7 0 L3TR5 0 0 0 0 0 0 0 0.1 0 0 L3TR6 2.5 0 0 0 0 0 0 0 0 0 L3TR7 0 0 0 0 0 0 0 0 0 0 L3TR8 0 0 0 0 0 0 0 0 0 0 L4TR1 0 0 0 7.5 0 7.5 0 0 2.5 0 L4TR2 0 0 0 17.5 0 7.5 0 0 0 0 U4TR3 0 0 0 0 0 0 0 0 0 0 U4TR4 0 0 0 0 0 0 0 0 0 0 U8TR1 0 0 0 0 0 2.5 0 0 0 0 U8TR2 0 0 0 0 0 2.5 0 0 0 0 101 Appendix 4b (cont'd): Species cover - plant species cover per quadrat Anap m Arun d Boyk e C i r c a Clay s Corn c Dice f E p i l a G a l i t Lact m U8aT4Ql 0 2.5 7.5 0 0 0 0 0 0 0 U8aT5Ql 0 0 2.5 0 0 0 0 0 0 0 U8aT6Ql 0 0 17.5 0 0 0 0 0 0 0 U8bT4Ql 0 0 0 0 0 0 0 0 0 0 U8bT5Ql 0 0 0 0 0 7.5 0 0 0 0 U8bT6Ql 0 0 0 0.1 0 0 0 0 0 0 L9T1Q1 0 0 7.5 0 0 0 0 0 0 0 L9T2Q1 0 2.5 7.5 7.5 0 0 0 2.5 0 2.5 L9T3Q1 0 7.5 2.5 0 0 0 0 2.5 0 2.5 U9T4Q1 0 0 17.5 0 0 0 0 0 0 2.5 U9T5Q1 0 0 17.5 0 0 2.5 0 0 0 0 U9T6Q1 0 0 0 0 0 0 0 0 0 0 L10T1Q1 2.5 2.5 7.5 0 0 0 0 0 0 2.5 L10T2Q1 0 2.5 0.1 0 0 0 0 0 0 0 L10T3Q1 0 0 2.5 0 0 0 0 0 0 0 L13T1Q1 2.5 0 29 0 0 0 0 2.5 2.5 0 L13T2Q1 2.5 0 7.5 0 0 0 0 2.5 0 2.5 L2TR1 2.5 2.5 17.5 0 0 0 0 0 0 0 L2TR2 2.5 2.5 2.5 0 0 0 0 0 0 0 L2TR3 17.5 0.001 0 0 0 0 0 0 0 0 L2TR4 17.5 0 0 0 0 0 0 0 0 2.5 U3TR1 0 0 0 0 0 0 0 0 0 0 U3TR2 0 0 0 0 0 0 0 0 0 0 U3TR3 0 0.1 2.5 0 0 2.5 0 0 0 2.5 U3TR4 0 7.5 7.5 0 0 0 0 0 0 0 L3TR5 2.5 0 2.5 0 0 0 0 0.1 0 0 L3TR6 2.5 2.5 2.5 0 0 0 0 0 0 0 L3TR7 0 0 0 0 0 0 0 0 0 0 L3TR8 0 0 0 0 0 0 0 0 0 0 L4TR1 0 0.1 2.5 0 0 0 0 0 0 0 L4TR2 0 2.5 2.5 0 0 0 0 0 0 0 U4TR3 0 0 2.5 0 0 0 0 0 0 7.5 U4TR4 0 0 0 0 0 0 0 0 0 2.5 U8TR1 0 2.5 2.5 2.5 0 0 0 0 0 0 U8TR2 2.5 17.5 0 2.5 0 0 0 0 2.5 0 Appendix 4b (cont'd): Species cover • Linn b Maia d Peta p Smil U8aT4Ql 0 0 0 0 U8aT5Ql 0 0 0 0 U8aT6Ql 0 2.5 0 0 U8bT4Ql 0 0 0 0.1 U8bT5Ql 7.5 0 0 2.5 U8bT6Ql 0 2.5 0 0 L9T1Q1 0 0 0 0 L9T2Q1 0 0 0 0 L9T3Q1 0 0 0 0 U9T4Q1 0 2.5 0 2.5 U9T5Q1 0 0 0 0 U9T6Q1 0 0 0 0 L10T1Q1 0 0 0 0 L10T2Q1 2.5 0 0 0 L10T3Q1 7.5 0 0 0 L13T1Q1 0 0 0 0 L13T2Q1 0 0 0 0 L2TR1 0 0 29 0 L2TR2 0 0 17.5 0 L2TR3 0 0 0 0 L2TR4 0 0 0 0 U3TR1 0 0 0 0 U3TR2 0 0 0 0 U3TR3 0 2.5 0 0 U3TR4 0 2.5 0 0 L3TR5 0 0 0 0 L3TR6 0 0 0 0 L3TR7 0 0 0 0 L3TR8 0 0 0 0 L4TR1 0 0 0 2.5 L4TR2 0 0 0 0 U4TR3 0 0 0 0 U4TR4 0 0 0 0 U8TR1 0 0 0 0 U8TR2 0 0 0 0 102 plant species cover per quadrat Stre T i a r t T r i f r V i o l Gram Care spp spp 0 7.5 0 7.5 0 0 0 7.5 0 7.5 0 0 7.5 2.5 0 2.5 0.1 0 0 7.5 0 0 0 0 2.5 7.5 0 0 0 0 2.5 7.5 0 2.5 0 0 2.5 2.5 0 2.5 0 0 0 0 0 0 2.5 0 0 2.5 0 0 2.5 0 2.5 17.5 0 2.5 0 0 0 17.5 0 0 0 0 0 2.5 0 0 0 0 0 0 0 2.5 2.5 0 0 2.5 0 0 0 0 0 0 0 7.5 0 0 0 0 0 2.5 0 0 0 0 0 0 0 0 0 2.5 17.5 0.1 2.5 0 0 0 0 0.1 2.5 0 0 0 0 2.5 2.5 2.5 0 0 0 2.5 2.5 7.5 0 0 0 0 0 0 0 0 0 0 0 0 0 2.5 0 2.5 0 0 0 2.5 0 2.5 2.5 0 0 0 0 0 0 2.5 0 0 0 0.1 2.5 2.5 0 0 0 0 0 0 0 0 0 0 0 0 0 2.5 0 0.1 2.5 0 0 0 0 0 2.5 0 2.5 2.5 0 2.5 0 0 7.5 17.5 0 0 2.5 0 2.5 2.5 0 17.5 0 0 2.5 0 0 7.5 0 0 103 Appendix 4b (cont'd): Species cover - plant species cover per quadrat Aspl v Athy f Blec s C r i p c Cryo f Dryo e Gymn D Poly m Thel p d-Thuj U8aT4Ql 0 7.5 2.5 0 0 17.5 0 0 0 0.1 U8aT5Ql 0 2.5 0 0 0 0 0 2.5 0 0.1 U8aT6Ql 0 17.5 7.5 2.5 0 0 0 0 7.5 0.1 U8bT4Ql 0 0 17.5 0 0 42 0 0 0 0.1 U8bT5Ql 0 0 17.5 0 0 17.5 2.5 0 0 2.5 U8bT6Ql 0 7.5 29 0 0 7.5 0 0 0 0 L9T1Q1 0 17.5 17.5 0 0 0 0 7.5 0 0 L9T2Q1 0 0 0 0 0 29 0 0 0 0 L9T3Q1 2.5 7.5 2.5 0 0 0 0 2.5 0 0 U9T4Q1 0 17.5 2.5 0 0 0 0 7.5 17.5 0.1 U9T5Q1 0 7.5 17.5 0 0 0 2.5 0 17.5 0 U9T6Q1 0 17.5 0 0 0 0 0 7.5 2.5 0 L10T1Q1 0 0 0 0 0 0 0 0 0 0 L10T2Q1 0 0 7.5 0 0 0 0 7.5 0 0 L10T3Q1 0 0 0 0 0 0 0 0 0 0 L13T1Q1 0 0 17.5 0 0 0 7.5 0 0 0 L13T2Q1 0 7.5 17.5 0 0 2.5 0 0 0 0 L2TR1 0 0 2.5 0 0 0 0 0 0 0.1 L2TR2 0 0 0 0 0 0 0 0 0 0.1 L2TR3 0 0 0 0 0 0 0 0 0 0 L2TR4 0 0 0 0 0 0 0 0 0 0 U3TR1 0 0 0 0 0 0 0 0 0 0 U3TR2 0 0 0 0 0 0 0 0 0 0 U3TR3 0 0 17.5 0 0 0 2.5 0 0 0.1 U3TR4 0 0 0.1 0 0 0 0 0 0 2.5 L3TR5 0 0 0.1 0 0 0 0 0 0 0.1 L3TR6 0 0 0.1 0 0 0 0 0 0 0.1 L3TR7 0 0 0 0 0 0 0 o 0 0 L3TR8 0 0 0 0 0 0 0 0 0 0 L4TR1 0 0 0 0 0 0 . 0 0 0 0 L4TR2 0 0 0 0 0 0 0 0 0 0 U4TR3 0 0 0 0 0 2.5 0 17.5 0.1 0 U4TR4 0 0 0 0 0 0 0 29 0 0 U8TR1 0 0 2.5 0 2.5 2.5 0 2.5 0 0 U8TR2 0 0 0 0 0 0 0 0 0 0.1 Appendix 4b (cont'd): Species cover - plant species cover per quadrat d - T s u g B r y o S o i l U8aT4Ql 0.1 0.1 20 U8aT5Ql 0.1 0 60 U8aT6Ql 0.1 2.5 50 U8bT4Ql 0 29 0 U8bT5Ql 2.5 42 0 U8bT6Ql 0 2.5 0 L9T1Q1 0 7.5 3 L9T2Q1 0 7.5 0 L9T3Q1 0 29 1 U9T4Q1 0 29 50 U9T5Q1 0 0 1 U9T6Q1 0 2.5 2 L10T1Q1 0 2.5 2 L10T2Q1 0 17.5 0 L10T3Q1 0 7.5 0 L13T1Q1 0 0.1 2 L13T2Q1 0 2.5 5 L2TR1 0.1 17.5 7 L2TR2 0 29 3 L2TR3 0 0.1 10 L2TR4 0 0.1 7 U3TR1 0 0 100 U3TR2 0 0 100 U3TR3 0.1 2.5 5 U3TR4 2.5 7.5 25 L3TR5 0.1 62.5 30 L3TR6 0.1 42 10 L3TR7 0 0 100 L3TR8 0 0 100 L4TR1 0 0 35 L4TR2 0 17.5 30 U4TR3 0 0.1 15 U4TR4 0 2.5 7 U8TR1 0 0.1 25 U8TR2 0.1 2.5 33 105 o o o O LO O 00 O CN o o o o * * * * - H u 4J ID n a o - r l 4J 10 rH <D U U 0 CJ * « © O O o X < u w OH o LD o TP o m o CN H 1 o CN TP o H cn o r> CN o H TP 1-1 1 '" o o o o o o o o o o o o o o o o o o m in CTl o o o TP in o o 1-1 so o o CN 1-1 so H o O o CN TP o o o ro CO TP o o o so so cn o o o cn ro TP H 1 o o o o H co t> o o o o r> CO 00 o o o o SD H in o o o o rH CTl in rH i O o o o o o o o O o o o o o o o O o o o o o o o O o o o o o o o i—1 o H o o ro o 00 ro H o CTl o o OS. o cn CN cn o r- o o SO o H oo r-o H o o LO o H o to H t> H o in o o r- oo in CTl 00 o CO o o LT) r- CN CTl ro o in o o CN r> r> O H i o r> o o t> 1 CN ro CO CO cn o o o in CN 00 TP o o o o o o CN cn CN CN TP o o o TP r> SD O o 00 o o o 1—1 r- TP CN CO TP H CN CO TP et <t <i <i e£ < a •H d) U U U H H H H tJIW ft H H W > > > > ens 0 W W W W W W H r H H O j ft o B •H cn C •H W •J cj W 0. CO a u a w u a a, co u w Cu co tt) u a cd -H r4 cd > rd 4J o EH o o o in TP CO cn m o CN so O m 00 o rs) o o o rH o CO m m so in CO so o oo cn o H cn o CN co o CO CN CN ro CN cn 00 in • II ll ll ll II II O r> VD H CN cn tt) rH CS 0 u 0 <u > -H 3 -H •3 a 4J rH 1) 4J i—1 <D CC rd u rd rfl 01 !H > ni U > -H 1 i SH 1 i O H O CM <u Cn EH 00 TP • • H 00 CO CO H so DO CO C 0 -H 4J rd rH 0) U u o u 4-1 c <u e o !H •H El aj i (0 CO to cn a) d) -H ft X a co u G rd -H !H cd > d) 01 cd 4J cd a 4J dJ cd u t ) HI !H I tt) CO CO ft d) d) •H -H d) CJ CJ > d) d) •H ft ft 4-1 CO CO id i—I MH MH 3 0 0 £ o a o •H cd rH d) !H 4J c tt) E a o in •H > a CO CO tt) tt) •3 3 rH rH (0 Cd > > 0 CI tt) tt) 01 01 -H -H 0) d) TJ d) c •H Cd rH SH fd 4J U CO -H a a rH rH Cd Cd MH MH 0 o e 6 CO to « * 4J ca d> 4J •H X cd rH cd cj -H . a o c cd o 4-> co SH -H MH MH 0 d) o a cd u -H MH •H C 01 -H CO MH 0 4J CO d) CO D X cd rd U •H C 0 C rd U MH o d) o c cd u c 01 •H CO MH 0 4J CO d) EH 106 X _c -4—' O CD O LLI ro <D >> 6 _ CN ° CO ^ -id "* CO CO ZD 1— CD 0 5*i CM • o -Q CO 3 • • o O •o Q) 00 U3 C O u X s ( X < o _ l CO co c? i n co o 3 • co oo 3 r--9'0+ Z S L X V o + CO o I vo-p oo 0, •«-» o a p so ja 3 'C B D. -a J3 •4—I c o w •c +-» •9 '3 u ex on • —• < 107 to • F-1 id c aj ex 108 (G •H 4-1 U <D fi O O i n o o in rd 4-> O E H <tf n ro CTl o r> co ro ro ID CN o •HI m H H 10 ro CN CN r - H rH O ro ro •sj" o ro CN m n •Hi r-CO co co J>N 13 < u Q NO *3 a CU a CD U fi rd •H U crj > ca cu rH rd > fi CD 01 -H CD CD <D fi 01 - H rd rd 4-1 rd M fi 4-1 4J CD rd U X) CJ X) rd r4 CD CO CJ fi 3 4J fi <D DI a ID fi o u CD CD u rd 0 > CD r - l • H ft rH rd CO 4J to rd 3 > -fi rd w fi 4-1 r - l to CD 01 0 0 •X CD 01 fl e * X - H E * <d w u * co 109 o o o o 0) ft o O 00 o rn ID SO O rH O SO o OS o H H l o SO ro o 00 r> o so rH o rsi TP H i 1 o o o o o o o o o o o o o o o o o (M o 01 o 00 CN o TP so CTl o rH H SO 01 a •H 01 01 o J pa o o o o\ in cn o o o oo so H O O O CN TP SO O O O 0-1 o so H I I O O O O CN t> rH O O O O H TP so O O O O TP 00 00 O O O O CO CTl CN o o o o o o o o o o o o oo < u u X 0 co OH * a - H iH 4 J cd 13 a 0 - H J J id rH CD u u o o o o o o o o r- oo r- C N SO 00 O C N I I cd o E H T P O T P C N O co o oo ro H TP o 00 so in so TP m o co H rH . ro in o as o • SO O so o o o oo in o • CN o CN CN rH II II II ll II ll LCI t> CM CN CTl CN 00 3 cd 0 <L> 0 OJ in > •iH 3 •H 00 d •U rH 0 4J rH <L> cd cd u cd cd cn u > cd SH > • H 1 i ! H i <u b C M E H fa CM TP in CN CTl CN LO CTl CTl m ro CN oo O O rH 00 o X C J < * CJ * w * CM * CO u u u H W H CM CM Oi 1 1 w o xis CO T* H H CTl • rd OJ o o o o o a ro ro CN CTl X o o o o o • H TP rH rd o o o o o CJ cr* o o o o o SPE es es *** nic; cal r o IX! fl 3 * 0 -H 0 r-l rH d d TP o o rH TP CO -H rd rd 4 J rd 0 TP o so in r> c J J > > ca CJ d o o H ro in CJ 0 OJ rd d d (U rd 00 o H H in pa o< co -iH u rH OJ OJ 4 J 4-1 u J J d OJ 01 01 CO la cd -H SH -H OJ -rH OJ lo SH -H IT CD U J J M MH rd o o CTl t> m X rH cd Cl TJ id o o CN H ro rH > OJ <U CJ MH MH o o ro TP CTl PEC 0 E a 0 0 o o 00 o in PEC o OJ a -H ID PEC 01 0 rd rH •U OJ OJ 1 1 0] J J cd SH SH rd fi CJ u c J J rd -H J J CJ o d d * d J J > CO •H 2 rd rd o o so CN so H * E OJ rd a C u u o o CI SO TP * d CJ T) <u 0 0 M H •H -H o o o r> in * 0 u i u d 0 MH MH o o ro CO CN CJ SH OJ CO CO a rd -H -rH SPE •H ft OJ OJ 3 U |>i d d i i SPE SH id CO > •H -rH 17 01 01 SPE <U d OJ CJ CJ rH rH id -H -rH 6 01 > OJ OJ rH rH CO CQ co TP Cn H rH I -H ft ft rd rd cd CO J J CQ CO p MH MH < < d CQ > <L> cd MH MH ca 0 0 •H <L) C! •rH rH MH MH 0 0 H H 01 OJ ft * CO OJ CJ 3 0 0 * J J 4-1 {> 01 E 0 * dJ 01 OJ E E E * CO CO H w o -H rH * ^ • H ft 3 3 3 * aj 01 H co * a a co u CO CO * H 110 Appendix 8: a) Forward and b) Manual selection, where N refers to number of different entry for a variable, P refers to the type 1 error, and F correspond to F c a i c u , a t e d of Fisher. a) Forward selection tested with Monte Carlo Permutation Test Variable1 N P F Torrent 2 0.005 3.04 Logging 1 0.005 2.22 Time 3 0.04 1.67 Slope 4 0.26 1.12 CWD 5 0.425 1.06 Moist 7 0.5 0.98 Aspect 8 0.625 0.84 Rock 6 0.91 0.56 b) Manual selection tested with Monte Carlo Permutation Test Variable1 N P F Slope 4 0.1000 1.46 Time 3 0.0050 2.76 Logging 1 0.0200 1.67 Torrented 2 0.0200 1.97 CWD 5 0.4250 1.06 Moist 7 0.5100 0.98 Aspect 8 0.5450 0.94 Rock 6 0.7450 0.75 'Slope: gully sidewall slope angles Torrent: debris flow activities Logging: presence or absence Time: elapsed time since last torrent C D W : coarse woody debris cover Moist: observed moisture factor Aspect: Slope aspect (Cosine of the angle, Roberts, 1986) Rock: rock cover. 03 to " > O o 1(2 '1 O is "•e 0 0 1) 1) IS co co CO >> "S3 CO CU C O * CXI CJ , E 3 o rM o T3 <U co 3 co CO "S > C+H O X 1 a 3 O 'M o U 73 3 CD a o 1) a, a § CQ o 55 O CO T-3 © O r-H IT) o o o ^  _ r-^' o o o O r-H r-H CN ^ ^ o o o o O r-H O r - i r-H r-H r-H © © © O © O CN r-H r—I CO r-H CN >-H ©' o © © o © O r n t N t N H H H t s ' - H © ' © © ' © ' © © ' © O r-H r ^ © ' O CN CN O CN r-H r-H o ©' o o ©' © o O C N r - H C N C N C N i n c O r ^ v o r ^ © ' © " © ' © © ' © ' © - © ' © ' OMnHqtNHlf lHHCfl r ^ ' O O O O O O O O O O <0 '—< -^ r; •—; rn CN >JO rn -—< ijO" r ^ © ' © © ' © © ' © ' © © © © ' © U > u o u o > -a ° o +J * H O Appendix 10. Variables used to evaluate sediment yield. Sites Sediment yield (cm3/yr) Standard error (cm3) Slope length (m) Contributing area * (m2) Sediment yield (cm3/m2/yr) Standardard error (cm3/m2/yr) L2-1 2235 452 15 18 124 3.6 L2-2 57 293 2.5 3 19 15.5 L2-3 14489 1313 14 16.8 862 1.5 L2-4 4595 310 16 19.2 239 1.3 L3-5 2124 2016 4.4 5.28 402 5.0 L3-6 44 21 3.9 4.68 9 2.2 L3-7 12648 657 4 4.74 2668 0.2 L3-8 19650 650 2.3 2.748 7151 0.1 L4-1 2324 312 4.5 5.4 430 0.7 L4-2 305 50 4.2 5.04 61 0.8 U3-1 8804 231 5.6 6.66 1322 0.2 U3-2 10831 2020 3.7 4.44 2439 0.8 U3-3 1082 30 3.7 4.464 242 0.1 U3-4 131 25 4.2 5.052 26 1.0 U4-3 508 41 3.7 4.44 114 0.4 U4-4 1385 106 4.4 5.28 262 0.4 U8-1 1051 43 7.5 9.048 116 0.4 U8-2 539 95 3.5 4.14 130 0.7 * Contributing area = slope length times trap width ** Yield = Sediment yield / contributing area / year 11 Appendix 11. Correlation matrix of the variables associated with the sediment traps. Yields (m3/ha/yr) Elapse time Slope %Bare soil aspect total Moisture cover Yields (m3/ha/yr) 1.0 Elapse time -0.3 1.0 Slope -0.3 0.2 1.0 % Bare soil 0.7 -0.5 -0.4 1.0 Aspect 0.2 -0.5 -0.4 0.6 1.0 Total cover -0.6 0.5 0.5 -0.8 -0.5 1.0 Moisture 0.4 -0.2 -0.1 0.4 0.1 -0.3 1.0 O T P r~- ro H o ro O O Ol 01 O ri IN O > o r> so ro cn m co H o o co ro o so so o 1 0 TP c o CO 0 in 0 Tf 0 0 T P r- 0 CO in r> 0 0 T P C N C N cn C N in 0 so 1-1 r o 0 C N T P H o o o o 3 u CO I u CD Tj 0 0 CO H H CN CN so H 0 0 H CN t> CN T P so cn 0 0 in CO t> O SO co r o 0 0 in r o CO CN m so CN 0 0 0 0 TP CN m m H CN cn 0 0 0 0 t> SO r> r o H O r> 0 0 0 0 s o rH SO 0 r o CO 0 0 0 0 0 0 i n r o c o H O TP O O O GO TP H r> so CN SO 0 O O O cn in H c o s o cn CO H O O O TP r> in rH 00 H cn CO O O O cn 0 r o CO 0 SO in s o 3 CJ w PH co O C N o r> o o o so o o o o SO rH rH t> i-H cn CN CO I I O T P o r> o co o H rH I r> C N C N CO 0 r-rH r o 1 I r o r o CO H O CO r o O 3 U Xi CO I a o 0 CO CO o g I rQ iH CD CJ - H a rd Cn u o I 0 0 O O in 0 H CN c-so in O CO 0 H O O cn 0 CO T P T P 00 r> m so 0 rH O O CN 0 in CO O H t> cn T P 0 O O O cn 0 in H CO 0 CN CO rH CO >> < U u *c3 +-> G CD T3 sU C N X c CD OH 0 CN SO O TP 0 0 CO CN CN TP CO so TP 0 r> rH O SO 0 0 co TP CO r- 0 H CO 0 rH TP O r> 0 0 CO CN s o cn m CN CO 0 O O O c n 0 0 in r o CO H in SO CN O cn cn CO in 0 0 0 TP 0 0 so TP TP O O CO CO in H 0 0 0 co so so so CO 00 TP O H CN so TP 0 0 0 so CO TP in CN CN 00 O 0 0 0 cn 0 0 0 0 TP co r> H O CO u w PH CO 3 CJ w PH co 3 CJ W PH CO O o O O O O O o TJ rH OJ •H 00 so CN CN m so C O C O I r o 00 so cn cn cn o so o t> o so o o H I o o in r-T P T P CN T P I C N r-00 00 r o C N C N H I r> o o o so o O C N I CN rH rH r> o r> CO CQ H CN r o TP rH CN CO TP J J X c CJ OS ru ru nt CJ OS ru ru < < < <C < < < < CD J J •H 6 X ! OJ J J •rH E A A 6 TJ CO CJ c 1 CO CO E CQ CJ C 1 CO CO CJ C J u CJ H H H H - H rH r* OJ rd rQ 1 1 - H OJ rd Si 1 1 a w w W > |> > TJ OJ CJ ft cn u a CJ TJ U ft Cn u a CJ PJ PJ p< PH Z 2 ; OJ - H 0 CO rH <D 0 OJ OJ 0 CO VH OJ 0 <D CO to CO CO w w w W CO >i rH rd 0 CJ TJ CO u rd 0 J4 CJ TJ J J CJ OJ ft CQ rd CQ M CJ 0 r l J J a <u g - H TJ <U CO 115 rd •H 4J U CD c ON VO ON 4J o ON ro VD ON cri CN rp- rH o en VD LD . . ri m ^ CN i n co r - ? C I U u >> d CD CD C/J d o C D CN T3 d C D CM 2 CO CD CO o CN CN ro • CN ON VO o • o o o ON i n ro ON m VD ro ON CN o ro CN o CN r> o CN rH CN H vo vo CN • II II II II II II ro ON CN ro m CD r» rH rd 0 CD 0 CD > •rH •H a 4-) rH CD 4J rH (D rd rd u rd rd en u > rd u > •H 1 1 u I 1 o CN ro ON CD Pn PH En fa OH VD •rf • • ro ON GO H CN1 CO w a o <D -H u 4J Pi rd rd •H M rd > CD U U O U CD 01 rd C O •H 4J rd rH CU U 4J c CD g c o u fi CD 4J rd -H fi V > <D rd J U Tj rl i CD 10 CO ft CD CD •H -H ( D U O > CD <D CO CO CD CD rH rH rd rd fi fi CD <D oi cn -r| -r| CD CD X ) CD fi •H rd rH U rd I -H ft ft rd CO 4-1 CO CO > CD rd C -H « ) U 3 01 <U g •H ft 3 W CO U <U 4H o o rH rH rd rd 4H 4  0 0 g g 3 3 co co to 0) 4J •H X rd rd O -H fi o fl rd U 4J co U -H l*H IH 0 CD O fl rd O •H UH -H fi 01 l(H o 4J CO CD E H CO a> X rd rd U -H fl 0 fi rd U MH o CD CJ fi rd CJ -rl MH -H fi 01 MH 0 4J CO CD E H 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0089473/manifest

Comment

Related Items