THE FOREST ASSOCIATIONS OF THE LITTLE LOST MAN CREEK RESEARCH NATURAL AREA, REDWOOD NATIONAL PARK, CA

by

James M. Lenihan

A Thesis Presented to The Faculty of Humboldt State University

In Partial Fullfillment of the Requirements for the Degree Master of Science

June, 1986 THE FOREST ASSOCIATIONS OF THE

LITTLE LOST MAN CREEK RESEARCH NATURAL AREA,

REDWOOD NATIONAL PARK, CA

by

James M. Lenihan

Approved by the Master's Thesis Committee

Rudo 14,--W—Beeking,

Ger)firfd�Al lien

Stephen,Dr., Vgirs

Director, Natural Resources Graduate Program

Natural Resources Graduate Program Number

Approved by the Dean of Graduate Studies

Alba H. Gilleipiej AB STRACT

The old-growth redwood (Sequoia sempervirens) forest associations of the Little Lost Man Creek Research Natural Area in Redwood National Park, CA, were defined by this study. Eighty stands on a systematic sample grid were inventoried using releve-style sampling procedures of the Zurich-Montpellier (ZM) method of phytosociology. Traditional ZM synthetic procedures and modern multivariate techniques of classification (i.e. two-way indicator species analysis and discriminant analysis) were used to define the associations and to investigate the relationship between the distribution of the associations and topographic and edaphic factors in the natural area. A forest association map derived from the classified sample grid subdivided the vegetation into homogeneous subunits occupying uniform habitats. Three forest associations were defined: the Sequoia sempervirens /Blechnum spicant association at lower elevations on relatively moist, concave lower slopes; the Sequoia sempervirens/Arbutus menziesii association at higher elevations on relatively xeric, convex upper slopes and ridges; and the Sequoia sempervirens/Mahonia pervosa association at mid-elevations on intermediate, relatively mesic sites.

iii ACKNOWLEDGEMENTS

I wish to acknowledge with love and deeply-felt gratitude the many contributions of my major professor, Dr. Rudi Becking. I am also grateful for the support of Redwood National Park, and especially to Steve Veirs, Jr., Research Scientist, for providing employment, for suggesting this study, and for his timely review of draft manuscripts. I wish to thank Dr. Gerry Allen for encouraging my interest in multivariate statistics, for his careful review and constructive criticisms of draft manuscripts, and for his good humour. I also wish to thank Dr. Richard Golightly for his careful editing and helpful suggestions that greatly added to the quality of this thesis. My parents and family must be acknowledged for their love and continual support of my interests in natural sciences. And finally, I am grateful for the companionship and many skull sessions with many good friends in Humboldt County, especially Esteban, Bill, Lori, Neil, and Steve H.

iv TABLE OF CONTENTS

Page ABSTRACT �iii ACKNOWLEDGEMENTS �iv LIST OF TABLES �vii LIST OF FIGURES �ix INTRODUCTION �1 Review of Studies in the Redwood Forest Region �2 Review of Traditional Zurich-Montpellier Methodology and Its Relationship to Modern Multivariate Classi- fication Techniques �4 STUDY AREA �12 Location �12 Topography �12 Geology and Soils �15 Climate �17 Vegetation �23 MATERIALS AND METHODS �25 RESULTS �37 DISCUSSION �54 Ecological Relationships �54 Relationships to Regional Types �64 REFERENCES CITED �70 vi

TABLE OF CONTENTS (CONTINUED) page APPENDIXES A. A List of Vascular Plant Species Found in the Study Area �78 B. Structured Table Formed by TWINSPAN Showing the Two-Way Classification of Species and Sample Stands from Little Lost Man Creek, CA �82 C. Relationships of Little Lost Man Creek Forest Associations to Regional Redwood Seral Types ... �83 LIST OF TABLES

Table � Page 1 �Typical Soil Profile Descriptions for the Melbourne Clay Loam Series and the Hugo Loam Series (deLapp et al., 1961a; DeLapp � and Smith, 1977) 19 2 �Thirty-year Summary (1941-1970) of Temper- ature and Precipitation Data from the Orick-Prairie Creek Weather Station � (ElFord and McDonough, 1974) 20 3 �The Zurich-Montpellier Cover-Abundance Scale (Becking, 1957) Used for Esti­ mating Species Cover, the Scale Used to Transform Cover Data for Analysis (Maarel, 1979) , and the Abundance Levels Defined for TWINSPAN 28 4 �Scheme Used to Determine Species' Fidelity Classes (Becking, 1957) �35

5 �Species Distributions in TWINSPAN Stand Groups Selected to Represent the Seq­ uoia sempervirens/Blechnum spicant, Sequoia sempervirens/Mahonia nervosa, and sequoia sempervirens/Arbutus man­ ziesii Forest Types of Little Lost Man Creek, CA �38 6 �Species F-Ratios and Their Significance in Analysis of Variance Tests Among Repre- sentitive TWINSPAN Stand Groups �39 7 �Results of Discriminant Analysis Showing Sample Stands in TWINSPAN Groups and in Groups where Membership Was Predicted by Discriminant Analysis �41 8 �Species Distributions in Equal Groups of Typical Stands Selected to Represent the Sequoia sempervirens/Blechnum spicant, Sequoia sempervirens/Mahonia nervosa, and sequoia sempervirens/ Arbutus menziesii Forest Types of � Little Lostb Man Creek, CA 45

vii viii Page LIST OF TABLES (CONTINUED) 9 �Summary Synthesis Table Showing Species Constancy/Model Cover Values in the Sequoia sempervireps/Blechnum spicant, Sequoia zempervirens/Mahonia nervosa, and Sequoia sempervirens/Arbutus men­ ziesii Forest Associations of Little Lost Man Creek, CA �46 10 �Fidelity Table Showing Species Fidelity Classes for the Sequoia sempervirens/ Blechnum spicant, Sequoia sempervirens/ Mahonia nervosa, and Sequoia sempervi­ rens/Athutus menziesii Forest Associa­ tions of Little Lost Man Creek, CA 47 11 �Mean Basal Area (BA) (Square Meters/Hectare) and Relative Dominance (RD) (Percent) of Canopy Species in the Sequoia sempervi- rens/BitZhnilm spicant, Sequoia sempervi­ rens/Mahonia nervosa, and Sequoia semper­ virens/Arbutus menziesii Forest Associa- tions of Little Lost Man Creek, CA �48 12 �Structure Coefficents for First Discriminant Function Derived for Discriminant Analysis Relating Distribution of Forest Types to Topographic Factors in Little Lost Man Creek, CA �50 13 �Synoptic Table for the Sequoia sempervirens/ Blechnum spicant, Sequoia sempervirens/ Mahonia nervosa, and Sequoia zempervi­ /ens/Arbutus menziesii Forest Associa­ tions of Little Lost Man Creek, CA 53 LIST OF FIGURES

Figure � Page 1 �Regional Map Showing the Location of Red- wood National Park in California �13 2 �Subregional Map Showing Location of the Little Lost Man Creek Research Natural Area in Redwood National Park, CA �14 3 �Topographic Map of the Little Lost Man Creek Research Natural Area, Redwood National Park, CA Showing Routes of Access, Section Lines, and Watershed Boundary Indicate by the Dotted Line �16 4 �Soils Map with Topographic Contours for the Little Lost Man Creek Research Natural Area, Redwood National Park, CA (after DeLapp, et al., 1961a; b) �18 5 �Thornthwaite (1948) Water-Balance Diagram for the Orick-Prairie Creek Weather Sta- tion �22 6 �Topographic Map Showing the Systematic Lo- cation of Sample Stands and the Stand Numbering System in the Little Lost Man Creek Research Natural Area, Redwood National Park, CA �26 7 �Topographic Map Showing the Forest Type Classification of Sample Stands in the Little Lost Man Creek Research Natural � Area, Redwood National Park, CA 43 8 �Generalized Forest Type Map of the Little Lost Man Creek Research Natural Area, Redwood National Park, CA �44 9� Distribution of Sample Stands on the Axis Representing the First Discriminant Func­ tion Derived to Relate the Distribution of the Forest Associations to Topographic Factors in the Little Lost Man Creek Re­ search Natural Area, Redwood National Park, CA �51

ix LIST OF FIGURES (CONTINUED) � Page 10 �Photograph of Stand No. 3 Representing the sequoia sempervirens/Blechnum Zpi - cant Association of Little Lost Man Creek, CA �56 11 �Photograph of Stand No. 66 Representing the sequoia sempervirpns/Arbutus man­ ziesii Association of Little Lost Man Creek, CA 59 12 �Photograph of Stand No.38 Representing the sequoia sempervirens/Mahonia ner­ vosa Association of Little Lost Man Creek, CA �62 INTRODUCTION

This study described the old-growth redwood (Sequoia sempervireng) forest associations of the Little Lost Man Creek Research Natural Area in Redwood NationalI Park, California.I In 1973, the relatively undisturbed, 972 ha watershed was designated a research natural area by the National Park Service to represent virgin, upland, old-growth redwood forest in the northern part of its range. Little Lost Man Creek was part of a nation-wide system of natural areas on federal lands (U.S. Federal Committee on Ecological Reserves, 1977) that preserved representative ecosystems as baseline areas and provided, through scientific research, information about ecosystem components, processes, and comparisons with manipulated systems (Franklin and Trappe, 1968). This study provided an inventory of vascular plant species and a definition of old-growth redwood forest types in terms of their structure, composition, and distribution along environmental gradients. A forest type map subdivided the vegetation into homogeneous subunits occupying uniform habitats, and thus provided a framework for future studies in the natural area.

1 2

Review of Vegetation Studies in the Redwood Forest Region

Several authors listed dominant species associated with Sequoia sempervirens throughout its range (e.g. Clements, 1920; Clark, 1937; Jensen, 1947; Mason, 1947; Munz and Keck, 1968; Franklin and Dyrness, 1973; Ornduff, 1974; Eyre, 1980). These generally included Pseudotsug. menziesii, Lithocarpus densiflora, Rhododendron macrophyllum, Vaccinium ovatum, Gaultheria shallop, Polystichum munitum, and Oxalis oregana. Roy (1966) and Becking (1982) provided more complete redwood floras, but there were few detailed accounts of old-growth redwood vegetation. Kuchler (1977) recognized two redwood forest types on his vegetation map of California. Sequoia sempervirens and Pseudotsugd menziesii were canopy dominants in the redwood forest type mapped from the Oregon border south to Monterey County. Sequoia sempervirens, Lithocarpus densiflora, Arbutus menziesii, and Ouercus wislizenii dominated the mixed hardwood and redwood forest type south of Monterey. Parker and Matyas (1979) showed two redwood forest types on their vegetation map of California with a composition and distribution similar to Kuchler's types. �Zinke (1977) derived data from the Cooperative Soil-Vegetation Survey maps (Colwell, 1974) and �showed dominant overstory �species along a moisture gradient determined by latitude and distance from the ocean. Associates of Sequoia sempervirens were shown varying from mauga heterophylla and Abies grandia in moist conditions to 3

Pseudotsuga menziesii, Lithocarpus densiflora, and Arbutus menziesii in drier situations. In Del Norte and Humboldt Counties, California, Veirs (1982) described a similar sequence of species along a moisture gradient related to elevation and aspect. Waring and Major (1964) used an autecological approach to describing the vegetation of Humboldt Redwoods State Park, California. They recorded the distribution of individual species along environmental gradients of moisture, nutrients, light, and temperature, and then calculated vegetation indices along each gradient for five redwood cover types. The indices for moisture and nutrients showed a gradual progression from relatively moist and fertile scores in pure redwood stands on alluvial flats to relatively dry and infertile scores in mixed stands of Sequoia sempervirens, Pseudotsuga menziesii, and hardwoods on upper slopes and ridges. Light under the forest canopy gradually increased along this gradient, but temperatures did not vary significantly. Becking (1967) defined two forest alliances in Humboldt Redwoods State Park using the Zurich-Montpellier (ZM) method of phytosociology (Becking, 1957). The "Redwood-Oxalis" alliance (Sequoia sempervirens-Oxalis oregona) on moist, alluvial flats and lower slopes was dominated by Sequoia Bempervirens. The shrub layer was generally sparse, but the understory was rich in characteristic herbaceous species, including Oxalis oregana, Blechnum spicant, Adiantum lopedatum, and AthyriUM filix-femina. ��The "Redwood-Swordfern" alliance (Sequoia 4 sempervirens-Polystichum munitum) occurred on upper slopes and ridges where drier conditions supported a mixed stand of Sequoia sempervirens, Pseudotsuga menziesii, Abies grandis, and Lithocarpus densiflora. A dense shrub layer of Rhododendron maLLD11.42211m, Gaultheria Shallop, Vaccinium ovatum and Vaccinium parvifolium was characteristic of this alliance. Characteristic members of the sparse herbaceous layer included Viola sempervirens, Mahonia nervosa, and Clintonia andrewsiana. McBride and Jacobs (1978) mapped redwood cover types in Muir Woods National Monument, California which they assigned to the "Redwood-Swordfern" alliance. �Dyrness et al., (1973) described old-growth redwood vegetation at the northern extreme of �its distribution in southwest Oregon. They presented data from three reconnaissance plots that showed a range of species composition encompassed by Becking's "Redwood-Swordfern" alliance.

Review of Traditional Zurich- Montpellier Methodology and Its Relationship to Modern Multivariate Classification Techniques

The forest types in this study were classified and characterized using both traditional phytosociological methods and modern multivariate classification techniques. The traditional ZM method was generally acknowledged as the most fully-developed and widely-used approach to the classification of vegetation (Whittaker, 1962; Westhoff and 5

Maarel, 1978). In comparative tests with other methods, the ZM method was consistently favored due to its combination of efficiency and effectiveness (Moore et al., 1970; Frenkle and Harrison, 1974). The standardization of concepts and methods was a major advantage of the ZM approach because it permitted the results of local studies to be integrated into hierarchial classifications of regional vegetation types. Classification studies in the redwood forest (Becking 1957; Muldavin et al., 1981; Lenihan et al., 1982) and related forests in California (Sawyer et al., 1977) and the Pacific Northwest (Becking, 1954; Bailey, 1966; Franklin and Dyrness, 1973) used the ZM or closely related techniques. A major objective in this study was to relate the vegetation of the study area to existing classifications; thus the ZM approach to classification was also used in Little Lost Man Creek. The ZM method was founded on three basic precepts: 1) the classification of vegetation types was based on total floristic composition, 2) the preference of species for certain vegetation types characterized the types and indicated the environment, and 3) character species served to integrate local vegetation types into formal, hierarchial classification systems for regional vegetation. Such a phytsociological study proceeded in three phases: analytic, synthetic, and syntaxonomic (Becking, 1957; Westhoff and Maarel, 1978; Gauch, 1982). The analytic phase consisted of extensive reconnaissance and data collection. Sample stands were chosen subjectively to represent preliminary �

6

interpretations of the vegetation based on reconnaissance, and to satisfy the requirement of stand homogeneity. The dimensions of the sample stands were large enough to exceed minimal area (Mueller-Dombois and Ellenberg, 1974), and variously-shaped to include only homogeneous vegetation. Standard data records, or releves, consisted of notes on vegetation structure, complete lists of vascular and non-vascular plant species, estimates of species cover-abundance and sociability using standard five-point scales (Braun-Blanquet, 1932) , indications of species vitality, and notes on the physiographic and edaphic character of the sample site. The first step in the synthetic phase of �the traditional ZM method of vegetation classification involved rearrangement of a floristic data matrix to reveal a major aspect of the matrix structure. The objective of this rearrangement was to concentrate positive matrix entries into blocks along the matrix diagonal so that stands with similar species compositions were grouped together, and species with similar distributions in the stands were grouped together. Matrix rearrangement was an iterative process of selecting species that differentiated groups of stands and stands differentiated by groups of species. The finished product was called the "differentiated" table. The next step in the synthetic phase was the derivation of synthetic characters that floristically defined the vegetation types. Groups of stands with a relatively distinct species composition were 7 outlined on �the �differentiated table. �These stands represented the vegetation types in the determination of species constancy and fidelity. Constancy was a measure of a species' presence within a vegetation type. Fidelity was the degree to which a species was restricted to a given type. Braun-Blanquet (1932) developed a rather complex scheme for fidelity determination based on species presence, abundance, sociability, and vitality. Five fidelity classes and three fidelity catagories were distinguished. Character species (fidelity classes III-V) were those which exhibited a definite preference for a particular vegetation type. Character species were of prime importance in the floristic definition of the vegetation types. Companion species (fidelity class II) were those with about equal preference for all comparable types. Incidental species (fidelity class I) were those with a definite preference for vegetation types other than the type under consideration. The syntaxonomic stage of a phytosociological study based on traditional ZM methods proceeded only if regional classifications were sufficiently advanced. If so, new types defined by character species were placed in hierarchial classifications of regional units. The fundamental unit of the hierarchy was the association. A number of associations were grouped into alliances defined by alliance character species. Alliances were similarly grouped into orders, and orders into classes. 8

Despite the proven effectiveness and efficiency of the ZM approach, the methodology was subject to much criticism , especially concerning its inherent subjectivity (Poore 1955a, 1955b). In this study, more objective, multivariate techniques (Gauch, 1982) were substituted for subjective procedures without deviating from basic ZM principles. Perhaps the most criticized step in the traditional approach was the biased selection of sample stands. In response to this criticism, a more objective method was used to locate sample stands in the natural area. Additional criticism concerned procedures in the synthetic phase of the ZM approach. Deriving the differentiated table by iterative rearrangement of the data matrix was a slow, tedious process subject to error, and was partially an art learned by example and practice. The outcome of the rearrangement was somewhat dependent on subjective choices involving initial and problematic groupings. Two-way indicator species analysis (Hill, 1979) was substituted for traditional methods of matrix rearrangement in this study. Two-way indicator species analysis was a polythetic, divisive classification technique that arranged multivariate data in an ordered two-way table by classification of individuals and attributes. In general, polythetic divisive classification techniques were most advantageous because the classification process began with the total data in contrast to agglomerative methods that began with information relating only to pairs of entities. In a test of five common 9 multivariate classification techniques, Gauch and Whittaker (1981) found two-way indicator species analysis among the best analyses �due to a combination of effectiveness, robustness, and objectivity. �The technique was an improvement over the original method described by Hill et al., (1975) in that species were classified as well as stands. Both classifications were used to structure a two-way table that emphasized the positive diagonal of the data matrix. Matrix rearrangement with this intent resulted in a structured table much like the ZM differentiated table. Braun-Blanquet's scheme for fidelity determination served as a standard but in practice was rather difficult to apply. The fidelity of a species in a type was defined by reference to its presence in comparable types, and often the determination of comparable types was difficult, especially where regional vegetation classification systems were non-existent, underdeveloped, or uncomparable. Fidelity class determinations also presented some difficulties in that sociability and vitality data were often unavailable. A simplified method proposed by Becking (1957) was used in this study. Becking's method for fidelity determination was based solely on the presence of a species within a type expressed as a percentage of the species' presence throughout its total range. The method was easier to apply and more similar to the basic concept of fidelity because it focused on species presence and eliminated abundance, sociability, and vitality from the determination. Two difficulties remained however. 10

The first concerned the definition of the species' presence throughout its total range. In this study, a species' total range was defined as its presence throughout the study area as indicated by its presence in the total set of the sample stands. Therefore, the species fidelity classes reported here applied only within the geographical limits of the study area. The second difficulty was inherent in the fidelity value calculation. The unstated requirement for deriving comparable values was that each vegetation type was represented by an equal number of stands. If the types were represented by unequal numbers of stands, then species in larger groups would have received relatively higher fidelity values. In this study, a systematic sampling scheme and a uneven distribution of forest types in the study area led to the formation of unequal groups of stands in the stand classification derived by two-way indicator species analysis. Discriminant analysis (Tatsuoka, 1970; Williams, 1983) , a method of distinguishing among groups by forming linear combinations of discriminating variables, provided an objective criterion for selecting equal subsets of typical stands for the constancy and fidelity calculations. Discriminant analysis was also used to investigate the relationship between the distribution of the forest types and topographic factors in the study area. Environmental data was often included on ZM differentiated tables, but a 11 systematic correlation of vegetation types and environmental parameters was not attempted by the traditional methodology. This study identified three old-growth redwood forest types in the Little Lost Man Creek watershed: the relatively moist Sequoia sempervirens/Dlachn= spicant association, the mesic Sequoia sempervirens/Mahonia nervosa association, and the xeric sequoia sempervirens/Arbutus menziesii association. STUDY AREA

Location

The Little Lost Man Creek watershed was located in Redwood National Park in northern Humboldt County, California (Figure 1) at 41 degrees 18 minutes north latitude and 124 degrees 00 minutes west longitude. Primary access to the study area was from Orick, California (Figure 2) located approximately 76 km north of Eureka, California on U.S. Highway 101. The western edge of the study area at Lady Bird Johnson Grove was reached by traveling the Bald Hills Road east of its junction with U.S. 101. An unimproved park service road from the Bald Hills Road at the southern end of the study area provided limited access to the southeastern boundary. The northern end of the study area was reached by traveling north of Orick on U.S. 101, and then taking Geneva Road east to the bridge over Little Lost Man Creek. The northwest and northeast boundaries and interior portions of the study area were accessible only by cross-country travel.

Topography

Little Lost Man Creek was a perennial stream in the greater Redwood Creek drainage basin (California Dept. of Water Resources, 1965). The creek flowed southeast to

12 13

NOR TH

CRESCENT CRY REDWOOD NATIONAL PARK

EUREKA

Figure 1. Regional Map Showing the Location of Redwood National Park in California. Figure 2.Subregional MapShowingtheLocationof the NORTH Freshwater Lagoon Stone Lagoon

Pacific Ocean Little LostManCreek ResearchNaturalArea in RedwoodNational Park,CA. 0 0 1 �

I i � 1 I I 2 - 3km 2mi

14 15 northwest for 8.6 km and then joined Prairie Creek. Prairie Creek flowed south from this confluence and emptied into Redwood Creek 3.6 km downstream. The Little Lost Man Creek watershed was a mature landform with a deeply-incised ravine, steep lower slopes, more moderate upper slopes, and a well-defined ridgeline (Figure 3). Numerous intermittent streams dissected the slopes of the watershed, and there was one major perennial tributary in the upper southeast corner of the study area. Elevations ranged from 30.5 m (100 ft) at the mouth of the creek to 731.5 m (2400 ft) in the upper southeast corner of the watershed. Parts of the study area were found on three U.S. Geological Survey topographical quadrangles: the Orick, California sheet (7.5 minute series), and the Orick, California and Tectah Creek, California sheets (15 minute series).

Geology and Soils

The study area was located at the northern end of the Coast Range Province in the central belt of the Franciscan formation (Irwin, 1960). This belt was a complex assemblage of eugeosynclinal rocks deposited in the late part of the Mesozoic from the late Jurassic to the late Cretaceous (Bailey, 1966). The most common rocks in the study area were graywacke sandstones, shales, and conglomerates. In the Cooperative Soil-Vegetation Survey (DeLapp et al., 1961a, 1961b; DeLapp and Smith, 1977) the majority of 16

Figure 3. Topographic Map of the Little Lost Man Creek Research Natural Area, Redwood National Park, CA Showing Routes of Access, Section Lines, and Watershed Boundary Indicated by Dotted Line. Sections Lie in T1ON, R1E; T11N, R1E; T1ON, R2E; T11N, R2E; Humboldt Meridian. 17 soils in the study area were classified as Melbourne clay loams or Hugo loams. The soil-vegetation maps for the study area (Figure 4) also showed minor inclusions of Mendocino and Usal clay loams, but about 95 percent of the soils were identified as Melbourne, Melbourne-Hugo, or Hugo-Melbourne associations. Typical soil profile descriptions for the Melbourne and Hugo series are provided in Table 1. Soils in the Melbourne series had a moderately rapid to rapid permeability, good drainage, a medium erosion hazard, and high to very high timber productivity (DeLapp and Smith, 1977). Soils in the Hugo series had a moderately rapid to rapid permeability, good to excessive drainage, a moderate to very high erosion hazard, and moderate to very high timber productivity (DeLapp and Smith,1977). In the current soil classification nomenclature, Melbourne clay loams were classified as fine, mixed, mesic ultic haploxeralfs and Hugo loams as fine, loamy, mixed, mesic dystric xerochrepts.

Climate

The study area had a maritime Mediteranean climate with ��cool, rainy winters and cool, dry, summers. �A thirty-year ��summary (1941-1970) ��of temperature �and preciptation �data from the Orick-Prairie Creek weather station (Elford and McDonough, 1974) was compiled (Table 2). This was the station nearest to Little Lost Man Creek which 18

Melbourne clay loan 01.1 Hugo loan Melbourne-Hugo association 1 -1 Others

Figure 4. Soils Map with Topographic Contours for the Little Lost Man Creek Research Natural Area, Redwood National Park, CA (after DeLapp et al., 1961a; b). 19 Table 1. � Typical Soil Profile Descriptions for the Melbourne Clay Loam and Hugo Loam Series (DeLapp et al., 1961a; DeLapp and Smith, 1977)

Depth (cm) �Characteristics

Melbourne Clay Loam Series 0-20 � dark brown strongly granular clay loam moderately acid fairly high in organic matter 20-50 � strong brown weakly blocky clay loam strongly acid 51-90 � strong brown, gley mottled fine blocky heavy clay loam strongly acid few rock fragments 90 + � partially-weathered, fine-grained sandstone

Hugo Loam Series 0-25 � grayish brown granular loam slightly acid few small rock fragments present 25-65 � pale brown weakly granular loam strongly acid increasing rock fragments with depth 65 + � shattered, partially-weathered, fine-grained sandstone Table 2. � Thirty-year Summary (1941-1970) of Temperature and Precipitation Data from the Orick-Prairie Creek Weather Station, CA (Elford and McDonough, 1974).

Month

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual

Temperature ( C) Mean ���7 8 8 10 12 14 15 16 15 13 ���9 7 11 Maximum �22 21 25 31 32 37 35 36 35 32 22 19 �37 Mean Maximum 11 12 13 15 17 19 21 21 22 18 14 11 �16 Minimum ���-8 -5 -2 -1 -1 2 4 ����6 2 2 -3 -3 -8 Mean Minimum �������4 4 4 6 7 9 11 11 10 8 ���6 4 7 Precipitation (mm) Mean �302 272 242 119 101 37 �9 12 30 158 223 297 1801 21 also had long-term climatic records. It was located about 4.7 km north of the study area at an elevation of 50 m (165 ft). At higher elevations in the study area, rainfall totals were probably greater. Only trace amounts of snowfall were recorded by the weather station, but light snowfall was common in the study area at higher elevations during the winter months. Heavy, wet fogs produced by onshore movement of moist oceanic air were a common feature of the climate in the summer. The fog often persisted for all but a few hours in the afternoon at lower elevations in the study area. A warm inversion layer formed during the summer months at 457 m (1500 ft) (Freeman, 1971) , so fog was probably less common above this altitude. The data from the Orick-Prairie Creek station were used to construct a water balance diagram (Thornthwaite, 1948) to represent the regional climate of the study area (Figure 5). The climate was characterized by a long wet season with a large water surplus and a short dry season with a slight water deficit. These features were typical of the climate throughout the coastal coniferous forest belt of the Pacific Northwest (Becking, 1954; Franklin and Dyrness, 1973). Climate diagrams constructed by Walter (1967) showed a gradual increase in the duration and intensity of summer drought for coastal locations south from Vancouver, British Columbia to Santa Cruz, California. At the southernmost extent of this coastal belt, in California's redwood region, fog was an important factor moderating aridity in the summer 22

300 ACTUAL EVAPOTRANSPIRATION = 493 mm

250

200 150

100

40------4, Precipitation 4,-----111Potential (scale in mm) �Evapotranspiration Water Surplus Water Deficit

Water Used From Soil Recharge Soil Storage

Figure 5. Thornthwaite (1948) Water-Balance Diagram for the Orick-Prairie Creek Weather Station. Temperature and Precipitation Data from Elford and McDonough (1974). Calculations assumed 100 mm of available soil water storage. 23 months. Fog drip helped offset the summer drought (Azevedo and Morgan, 1974), but reduced evapotranspiration due to higher humidity and reduced insolation was the main effect of fog that moderated summer water deficits in the redwood forest (Byers, 1953).

Vegetation

The vegetation of Little Lost Man Creek was representative of upland redwood forest vegetation at the northern part of its range. The term "upland" was used to distinguish the vegetation from redwood forests on alluvial flats. Forest resource maps prepared for the Department of Interior (1967) showed the stands of the study area as Sequoia sempervirens or Sequoia sempervirens-Pseudotsuga menziesii. The summary of forest classes showed 27 percent of the stands were old-growth in large size classes (i.e. diameters greater than 18.2 cm), 57 percent were old-growth in smaller size classes, and 13 percent were young stands in cutover areas. �The species composition of the stands was shown in more detail on soil-vegetation maps for �the watershed (DeLapp et al., 1961a, 1961b). On the maps the distribution of the soil-vegetation units indicated a compositional gradient from Sequoia sempervirens and Picea sitchensia at lower elevations, to Sequoia sempervirens and Pseudotsuga menziesii at mid-elevations, to Lithocarpus 24 densiflorA and Arbutus Eenziesil in the upper reaches of the watershed. MATERIALS AND METHODS

Sample points were located by reference to a dot grid placed on a 7.5 minute topographic map of the watershed. Dots on the grid were evenly-spaced at intervals representitive of 300 m. Twenty of the 100 points falling within the boundary of the natural area (Figure 6) fell within in areas disturbed by logging or in non-forested areas (e.g. the narrow riparian zone along the major stem of the creek). These points were eliminated from the sample because the objective of the study was to describe only undisturbed forest vegetation. The ends of lines along the east-west axis of the sample grid were located in the field and the lines were flagged by following an east or west magnetic bearing read from a hand compass. Sample points were located along each line by reference to a pocket altimeter and a topographic map. Given this somewhat approximate method of locating the eighty sample points in the field, each point was identified only as a general area for sampling. The precise location of the sample stand in each area was guided by the criterion of stand homogeneity. The homogeneity of a sample stand was judged by assessing the uniformity of vegetation structure and species dominance (Westhoff and Maarel, 1978). The area of the 15x25 m (375 m2) sample stands was in accordance �with �Mueller-Dombois and Ellenberg's (1974)

25 26

Figure 6. Topographic Map Showing the Systematic Location of Sample Stands and the Stand Numbering System in the Little Lost Man Creek Research Natural Area, Redwood National Park, CA. Dotted Line Indicates Watershed Boundary. Diagonal Shading Indicates Areas Disturbed by Logging. 27 empirical values for minimal area for sampling in forest vegetation (i.e 300-500 m2). The stand size exceeded minimal area in all types of vegetation sampled in the study area. Very few species in addition to those in each stand were found outside stand boundaries in each general area for sampling. �The rectangular sample stands were positioned so that the longer side was parallel to slope contours. �Each stand corner was marked by an aluminum tag nailed to a tree or attached to a large, woody shrub. Each sample stand was carefully searched and every vascular plant species occurring within stand boundaries was identified (Munz and Reck, 1968) and recorded in Raunkier life-form categories (Mueller-Dombois and Ellenberg, 1974). Non-vascular plants were not recorded because of the difficulty of field identification. �The scale used to estimate the cover of each species was the ZM cover/abundance scale ����(Becking, 1957; Table 3). Also recorded were percentage estimates of the cover of the following life-form/height strata: canopy trees greater than 50 m tall; subcanopy trees greater than 5 m tall; small trees and tall shrubs 2-5 m tall; and the field layer including low shrubs and herbaceous species less than 2 m tall. The basal area of canopy species was estimated in ft 2 per acre by the Bitterlich method (Grosenbaugh, 1952). Two counts were made from the center of the stand using two angle gauges with basal area factors of 50 and 100. The basal area data were later converted to metric units (i.e. m 2 per ha). Several 28 Table 3. The Zurich-Montpellier Cover-Abundance Scale (Becking, 1957) Used for Estimating Species Cover, the Scale (Maarel, 1979) Used to Trans­ form Cover Data for Analysis, and the Abundance Levels Defined for TWINSPAN

Percent Cover Range (Midpoint)

1-5 1-5 5-25 25-50 50-75 75-100 (2.5) (2.5) (15.0) (37.5) (62.5) (87.5)

Cover-Abundance + 1 2 3 4 5 Transformation 1 2 3 5 7 9 Abundance Level 1 1 2 3 3 3 29 topographic factors were also measured or described. �The aspect was measured in degrees azimuth using a hand compass. The slope was read from a clinometer and recorded in percent. The stand elevation was estimated using a topographic map and a pocket altimeter. The topographic configuration of the stand was described as concave, even, or convex. The soil profile at the center of the sample stand was examined using a hand auger. The profile was described in terms of color, texture, and depth of horizonation. Any additional information that served to characterize the sample stand was added to the notes. The stand was then photograped in black and white from a representative angle. To improve the efficacy of the data analysis, species with frequencies of less than five percent in the total set of stands were removed from the data set. The cover values for the remaining species were transformed according to a scale by van der Maarel (1979) (Table 3). The reduced, transformed matrix was then entered into a computer program called TWINSPAN (Hill, 1979) and used for two-way indicator species analysis. The first step in the analysis was the construction of a stand ordination for the purpose of revealing the main trend of variation in the sample data. This "primary" ordination was derived by reciprocal averaging. The ordination was then divided at the mean score to form two groups of stands. Preference scores were determined by frequencies on either side of the dichotomized ordination. A preference score of one was assigned to 30 species that exceeded a certain minimum frequency and were at least three times more frequent on one side of the dichotomy as on the other. Rare and less preferential species received scores which were downweighted accordingly, and the scores were signed positive or negative to indicate the species' preference for one side of the dichotomized ordination. To incorporate quantitative data such as cover into the classification procedure, the program constructed "pseudospecies" that represented the species at different levels of abundance. These levels were chosen in accordance with Hill's (1978) recommendation to establish a few classes and not to overweigh the effect of dominance (Table 3). The derivation of a discriminant function called the "refined" ordination was based on the preference scores of �the pseudospecies. �This function strongly polarized the primary ordination and separated "borderline" stands (i.e stands with scores near the mean ordination score). This process of deriving an ordination, making a dichotomy, and refining the dichotomy was then repeated on each of the two groups of stands formed at the first division level to yield four groups at the second level, eight groups at the third level, and so on until each group had a specified minimum number of stands. After the stand classification was �completed, TWINSPAN then classified the species. The first step was calculating the degree of fidelity of the species for stand groups at each division level of the sample classification 31

hierarchy. Species fidelity values were based on the ratio between the species' mean abundance level in a particular group and its mean abundance level outside the group. The fidelity values became the species' attributes in the species classification process. The same process of deriving a primary ordination, making a dichotomy, and refining the dichotomy at successive division levels was used to classify the species.

The final step in TWINSPAN was the conversion of the species and stand classifications to orderings. The orderings were designed to structure a two-way tabular arrangement of the species and stand groups so that the positive diagonal of the data matrix was emphasized. The ordering criterion was based on a series of group comparisons evaluated by discriminant analysis. The stand classification on the structured TWINSPAN table was interpreted at a division level where the floristic composition of the stand groups was consistant with that of plant assemblages repeatedly observed in the field. Stands in typical groups formed at this level were selected to represent the recurrent assemblages, or forest types, in subsequent steps in the analysis. Discriminant analysis was used to accomplish three objectives: 1) to provide a criterion for the selection of equal subsets of typical stands for the constancy and fidelity calculations, 2) to classify "transitional" stands

(i.e. stands outside the typical groups formed by TWINSPAN) , 32 and 3) to relate the distribution of the forest types to topographic variables. The DISCRIMINANT procedure (Klecka, 1975) in the SPSS system of computer programs was used for discriminant analysis in this study. The linear combinations of discriminating variables, or discriminant functions, were derived so that the first function accounted for the greatest amount of group separation, the second accounted for an unique amount of separation given that described by the first, and so on until the number of functions was one less the number of groups or number of discriminating variables, whichever was smaller (Tatsuoka, 1970). It was desirable to limit the number of functions for purposes of interpretation and graphical representation. The importance of each successive discriminant function was assessed by reference to the relative percentage of the eigenvalue (Klecka, 1975). An eigenvalue was a measure of the variance of a function, and the sum of eigenvalues for all possible functions was a measure of the total variance of �the discriminanting variables. �An eigenvalue expressed as a percentage of the sum of eigenvalues thus provided an index to the relative importance of a function. A discriminant function may have accounted for a large percentage of the total variance and still not have exhibited a high degree of group separation if the variance of the discriminators was not strongly related to group differences, so an index of "discriminatory power" described by Tatsuoka (1970) was used to assess the effectiveness of a derived function. The index was a measure 33 of the percentage variability of a function due to group differentiation. In discriminant analysis the total sample size should be at least two or three times the number of variables (Tatsuoka, 1970). The full set of species entered into the TWINSPAN analysis was too large to use as a set of discriminanting variables given the number of typical stands representing the forest types. A smaller number of species variables were chosen to serve as discriminators in the discriminant analysis. Selected species had the highest F-ratios in analysis of variance tests for differences among groups of typical stands (Klecka, 1975). A set of functions based on these high F-ratio species was derived to discriminate among the groups of typical stands. �The discriminant scores of �the stands and the pooled within-groups covariance matrix for the functions were used to derive a classification function for each typical group (Klecka, 1975). All stands were scored on each classification function and the scores were converted to two probabilities of type membership. One was the probability of a stand being as distant from the typical group centroid as the stand under consideration. Stands with the highest "distance" probabilities were selected to form equal subsets of typical stands for constancy and fidelity determinations. Constancy was calculated as the species percentage frequency within a subset of typical stands representing a forest type. Fidelity was calculated as the species presence within a 34 subset of typical stands expressed as a percentage of its presence in the total set of sample stands. The fidelity values were reported in four of the five fidelity classes (Becking, 1957) shown in Table 4. Incidental species (fidelity class I) were not identified in this study. Assigning fidelity value to an erratic species with a definite preference outside of the vegetation type under consideration seemed contradictory. The second set of probabilities derived from the discriminant classification functions indicated the likelihood of each stand being a member of each type. All stands in the sample set, including those identified as transitionals, were assigned to the type where this "membership" probability was highest. The forest type membership of each stand was then plotted on the sample grid to reveal the spatial distribution of the final classification. Lines were drawn around groups of adjacent stands within the same type to produce a generalized forest type map of the study area (Crawford and Wishart, 1968). Seven variables served as discriminators in the set of functions derived to describe the relationship between the distribution of the forest types and the topography of the study area. The values of elevation and percentage slope were entered into the analysis as recorded in the field. Topography was coded as an ordinal variable (i.e. one for concave, two for even, and three for convex). The aspect in degrees azimuth was transformed into two directional vectors. 35 Table 4. Scheme Used to Determine Species Fidelity Classes (Becking, 1957). A Species' Fidelity Value Was Calculated as the Species' Occurrence in One of the Equal Sets of Representitive Stands Expressed as a Percentage of the Species' Occurrence in the Total Set of Stands.

Fidelity Classes �Fidelity Values (%)

Character species� Fidelity �V� 90-100 Fidelity IV� 75-89 Fidelity III 50-74 Companion species Fidelity II �5-49 Incidental species Fidelity�� I 0-4 36

The cosine of the aspect represented the north-south vector, and the sine represented the east-west vector (Kercher and Goldstein, 1977). The slope position of each sample stand was calculated as the ratio between the stand elevation above the creek and the ridge elevation above the creek. Elevations required by this calculation were taken from the topographic map of the study area. Solar radiation indices based on slope and aspect data were taken from tables presented by Frank and Lee (1966). Discriminant functions based on these variables were derived to discriminate among groups of typical stands. The pooled within-groups correlations between the functions and the topographic variables were examined to identify the variables that best distinguished the forest types. The use of these "structure coefficients" to assess the importance of individual discriminanting variables was recommended by Williams (1983). Topographic variables with large structure coefficients were identified as factors related to the distribution of the forest types. RESULTS

Eighty-three vascular plant species were encountered in the sample stands (Appendix A). Forty-three species with frequencies of five percent or more (i.e. 52 percent of the total species set) were included in the analysis. Species and stands were arranged in a two-way classification on the structured TWINSPAN table (Appendix B) , and the stand classification was interpreted at the second division level where three typical groups and one transitional group were formed. �The three typical groups were comprised of sixty-four stands (i.e. �80 percent of the total set of sample stands); these stands were arranged by their TWINSPAN species-orderings and within-group stand-orderings (Table 5). Twenty-one species (i.e. almost one-third the number of typical stands) had F-ratios with levels of significance greater than 0.0001 in analysis of variance tests for differences among typical groups (Table 6). These twenty-one species with high F-ratios were selected to serve as discriminators in two discriminant functions. These two functions accounted for 86 percent and 14 percent respectively of the total variance of the high F-ratio species. The index of discriminatory power for the set of functions indicated that 99 percent of their variability was due to differentiation among typical stand groups. All stands in the sample grid, including those identified as

37 IIIII

38

Table 5. Species Distributions in TWINSPAN Stand Groups Se­ lected to Represent the sequoia sempervirens/ spicant, Sequoia sempervirens/MahoniaBlechnum nervosa, and Sequoia sempervirens/Arbutus men­ ziesii Forest Types of Little Lost Man Creek, CA. Numbers and Plus Symbols Are ZM Cover-Abundance Values (Becking, 1957). Stand Identification Numbers Are Printed Vertically.

Stand No.I 100110000002110 2544332166654221877666554443321 775555477776664875 143536875213729 2275651095404744198732868608059 707543263218619051

Sequoia/ I Sequoia/ISequoia/I Species BlechnumIMahoniaIArbutus

Disporum smithii 111 111 Athyrium filix-femina 12 121 1 Adiantum pedatum 1 1 1 Rubus spectabilis 11 1 Menziesia ferruginea 112 1 Corylus cornuta 223 2 32 Rhamnus purshiana 11 1333222133 Picea sitchensis 1 1 11 Blechnum spicant 222223 1 11312 1I +121+ Stachys mexicana 21I1 + Adenocaulon bicolor 1I 11 Vancouveria hexandra 11 1I111 1+1I 1 Oxalis oregana 333333222223223 222 2112122 2221I1I2 1I + 11 Polystichum munitum 555555543225544 222121122231331122211I22 212211 11111 2 1 Vaccinium parvifolium 21 133113142411 211 1 +1 111 I11 I+ 2I11 1 + 12 1 + Trillium ovatum 11I+11+1 1I1 12I+ 1 1 1+ Disporum hookeri ++ 1 I+11I1I 1 Carex hendersonii 1 II 1 1 Actaea rubra Abies grandis 1I2122I 1 Iris douglasiana 11I1+1 II1 11I 1 1 Festuca subulata 1111I+ 1 1 1I 111+1211 Mahonia nervosa 1+I+1 2 22222222 22 122 21121222222 211 12 2 Galium aparine + 1I1 1I+ 11 11111111 11111 1 11111111111 1 1 Tsuga heterophylla 1I1I11111 1I1 1132 13 I1I1I1 1 11I 2 1 Sequoia sempervirens 344454333332423I 3232333333223333323333333333333 3222323322222 2222 Trientalis borealis +I+ 1 + 111I+11I1 1 +I 1 11 111 Gaultheria shallon 1121321+2 4332435433253342533432433232433 222131123443521352 Vaccinium ovatum 212121233421131 12312232113 135123344334344344 444544523233234315 Lithocarpus densiflora 312 3 21 4232 2245443353313342445334333443424 I 334544354443545544 Whipplea modesta 21 +111 111I1111+111111 1 1 + + Hierochloe occidentalis 2122222211 + 1 11 11112112 211 Viola sempervirens + 1 + 2222222112222221222221222121212I 211221121 1111+111 Festuca occidentalis I +I+ 1 1 Rhododendron macrophyllum 11 3211I1 1 2 232424 2123233143233233 444443333414224 3 Pseudotsuga menziesii 1 11 11 2221 3 122212123 232223322222333333233 222312312222221112 Lathyrus vestitus + 1I 1 + 111 11 1 I Cardamine integrifolia 1 11 +I+111 +11 2 1I11I 111 Pleuricospora fimbriolata +1 1+ 11 Pteridium aquilinum I 12111+1 1+ Goodyera oblongifolia 1I11 +I2121 11 11 211+ 1I 1 11 11 Corallorhiza maculata 1 +I2 1 1I 111 Arbutus menziesii 1I 11 2 121222 39

Table 6. Species F-Ratios and Their Significance in Analy­ sis of Variance Tests among Representitive TWIN­ SPAN Stand Groups. Asterisks Indicate Species Selected as Variables for Discriminant Analysis.

Species F- Signif- Discriminating ratio icance Variable

Sequoia sempervirens 18.94 0.0000 Pseudotsuga menziesii 11.27 0.0001 Tsuga heterophylla 0.96 0.3886 Abies grandis 3.60 0.0334 Picea sitchensis 8.49 0.0006 Lithocarpus densiflora 24.21 0.0000 * Arbutus menziesii 12.08 0.0000 * Rhododendron macrophyllum 11.95 0.0000 * Rhamnus purshiana 44.82 0.0000 * Corylus cornuta 14.02 0.0000 * Vaccinium ovatum 6.30 0.0033 Vaccinium parvifolium 17.37 0.0000 * Menziesia ferruginea 8.08 0.0008 Rubus spectabilis 5.84 0.0048 Gaultheria shallon 20.75 0.0000 * Mahonia nervosa 25.07 0.0000 * Polystichum munitum 90.37 0.0000 * Blechnum spicant 25.75 0.0000 * Actaea rubra 0.53 0.5947 Adenocaulon bicolor 3.46 0.0377 Adiantum pedatum 5.84 0.0048 Athyrium filix-femina 14.58 0.0000 * Carex hendersonii 1.69 0.1940 Corallorhiza maculata 12.96 0.0000 * Disporum hookeri 0.32 0.7272 Cardamine integrifolia 4.23 0.0191 Festuca occidentalis 0.85 0.4339 Festuca subulata 13.18 0.0000 * Galium aparine 26.06 0.0000 * Goodyera oblongifolia 6.11 0.0038 Hierchloe occidentalis 45.60 0.0000 * Iris douglasiana 7.49 0.0012 Lathyrus vestitus 7.49 0.0012 Oxalis oregana 58.62 0.0000 * Pteridium aquilinum 20.99 0.0000 * Stachys mexicana 3.69 0.0307 Trientalis borealis 1.93 0.1541 Trillium ovatum 2.79 0.0691 Vancouveria hexandra 5.83 0.0048 Viola sempervirens 62.17 0.0000 * Whipplea modesta 25.16 0.0000 * Disporum smithii 15.57 0.0000 * Pleuricospora fimbriolata 10.96 0.0001 40 transitionals, were assigned to the type where membership probability was the highest (Table 7). The type membership was plotted on the map showing the systematic location of sample stands (Figure 7) , and the spatial distribution of this classification was used to generate the forest type map (Figure 8). Fifteen stands with the highest distance probabilities (Table 7) were selected to represent each of the three forest types in the constancy and fidelity calculations. The equal subsets of typical stands were arranged by their TWINSPAN orderings (Table 8). Species modal cover-abundance, constancy (Table 9), and fidelity (Table 10) were calculated for the three forest types. The mean basal area and relative dominance of canopy species (Table 11) were also calculated by forest type. Mean basal area and relative dominance values were based on the subsets of typical stands used to determine fidelity and constancy. The relative dominance of a species was its basal area expressed as a percentage of the total basal area of canopy species in a stand. The forest types were named by appending the binomial of a character species with a modal cover value of two or more to the name Sequoia sempervirens. SequoiaThe three type names derived in this manner were the sempervirens/Blechnum spicant (Sequoia/Blechnum) type, the Sequoia sempervirens/Mahonia nervosa (Sequoia/Mahonia) type, and the Sequoia sempervirens/Arbutus menziesii (Sequoia/Arbutus) type. 41

Table 7. Results of Discriminant Analysis Showing Sample Stands in TWINSPAN Groups and in Groups where Mem­ bership Was Predicted by Discriminant Analysis. A Distance Probability Is the Likelihood of Any Stand Being As Distant from the Predicted Group Centroid As the Stand Under Consideration. A Mem­ bership Probability Is the Likelihood of a Stand Being a Member of the Predicted Group. Asterisks Indicate Stands Selected to Form Equal Groups of Typical Stands for Constancy and Fidelity Calcu­ lations. Group Codes: (1)Sequoia sempervirens/ Blechnum spicant, (2) Sequoia sempervirens/ Mahonia nervosa, (3) Sequoia sempervirens/ Arbutus menziesii, and (4) Transitional Stands.

Probabilities Stand Selected TWINSPAN Predicted No. Stand Group GroupIDistance Membership

1 * 1 1 .0159 1.0000 2 * 1 1 .1969 1.0000 3 * 1 1 .3159 1.0000 4 * 1 1 .5552 1.0000 5 * 1 1 .9966 1.0000 6 * 1 1 .1423 1.0000 7 * 1 1 .3713 1.0000 8 * 1 1 .6033 1.0000 9 * 1 1 .5913 1.0000 10 2 2 .2737 1.0000 11 * 1 1 .1410 1.0000 12 * 1 1 .6975 1.0000 13 * 1 1 .6202 1.0000 14 2 2 .2048 0.9996 15 * 1 1 .4388 1.0000 16 4 2 .0011 1.0000 17 * 1 1 .1926 1.0000 18 4 2 .0000 1.0000 19 * 2 2 .8946 1.0000 20 4 1 .0000 1.0000 21 2 2 .4016 1.0000 22 2 2 .1338 1.0000 23 * 1 1 .3108 1.0000 24 2 2 .1046 1.0000 25 * 2 2 .8912 1.0000 26 4 1 .0000 1.0000 27 * 2 2 .6399 1.0000 28 4 1 .0069 1.0000 29 4 2 .0725 1.0000 30 * 2 2 .4269 1.0000 31 4 1 .0000 1.0000 32 4 2 .0000 1.0000 33 4 2 .0000 1.0000 34 4 2 .0001 1.0000 35 2 2 .3335 1.0000 36 2 2 .1567 1.0000 37 4 2 .0000 0.9994 38 * 2 2 .4494 0.9998 39 4 2 .0001 1.0000 40 * 2 2 .7001 1.0000 42

Table 7. Results of Discriminant Analysis Showing Sample Stands in TWINSPAN Groups and in Groups where Mem­ bership Was Predicted by Discriminant Analysis. A Distance Probability Is the Likelihood of Any Stand Being As Distant from the Predicted Group Centroid As the Stand Under Consideration. A Mem­ bership Probability Is the Likelihood of a Stand Being a Member of the Predicted Group. Asterisks Indicate Stands Selected to Form Equal Groups of Typical Stands for Constancy and Fidelity Calcu­ lations. Group Codes: (1)Sequoia sempervirens/ Slechnum spicant, (2) Sequoia sempervirens/ Nahonid nervosa, (3) Sequoia sempervirens/ Arbutus menziesii, and (4) Transitional Stands. (continued)

Probabilities Stand Selected TWINSPAN Predicted No. Stand Group Group Distance Membership

41 4 2 0.0014 0.8758 42 * 3 3 0.7754 1.0000 43 4 2 0.0000 1.0000 44 * 2 2 0.9727 1.0000 45 2 2 0.1512 1.0000 46 * 2 2 0.9713 1.0000 47 * 2 2 0.4736 1.0000 48 2 2 0.0514 1.0000 49 * 3 3 0.8969 1.0000 50 2 2 0.3562 0.9999 51 * 3 3 0.7351 1.0000 52 * 2 2 0.6730 1.0000 53 * 3 3 0.6895 1.0000 54 * 3 3 0.4682 1.0000 55 * 3 3 0.9427 1.0000 56 2 2 0.2181 1.0000 57 3 3 0.1455 0.9932 58 * 2 2 0.4375 1.0000 60 4 2 0.0097 1.0000 61 * 3 3 0.5200 1.0000 62 2 2 0.3459 1.0000 63 2 2 0.7438 1.0000 64 * 2 2 0.3653 1.0000 65 * 2 2 0.5417 1.0000 66 * 3 3 0.5882 1.0000 67 * 2 2 0.4292 0.9999 68 3 3 0.1250 1.0000 69 2 2 0.1969 0.9968 70 * 3 3 0.8938 1.0000 71 * 3 3 0.6482 1.0000 72 * 3 3 0.8168 1.0000 73 * 3 3 0.7250 1.0000 74 4 2 0.0053 1.0000 75 * 3 3 0.6722 1.0000 76 * 3 3 0.9077 1.0000 77 3 3 0.2896 0.9993 78 * 2 2 0.7129 1.0000 79 2 2 0.0702 0.9793 80 * 3 3 0.5061 1.0000 81 2 2 0.3792 1.0000 1.1- 3

B- Sequoia sempervirens/Blechnum spicant

M- Sequoia sempervirens/Mahonia nervosa A- Sequoia sempervirens/Arbutus menziesii

Figure 7. Topographic Map Showing the Forest Type Classifi­ cation of Sample Stands in the Little Lost Man Creek Research Natural Area, Redwood National Park, CA. Dotted Line Indicates Watershed Boun da ry. 43

B- Sequoia sempervirens/Blechnum spicant M- Sequoia sempervirens/Mahonia nervosa A- Sequoia sempervirens/Arbutus menziesii

Figure 7. Topographic Map Showing the Forest Type Classifi­ cation of Sample Stands in the Little Lost Man Creek Research Natural Area, Redwood National Park, CA. Dotted Line Indicates Watershed Boundary. 144

S• uoia Ilmtalamem.chnum Idaaja Ss uoia enamix1211/Mahonia nervosa Se alt EmpAr/irsag/Arbutus esnzis sii

Figure 8. Generalized Forest Type Map of the Little Lost Man Creek Research Natural Area, Redwood National Park, CA. Dotted Line Indicates Watershed Boun­ dary. 45

Table 8. Species Distributions in Equal Groups of Typical Stands Selected to Represent the Sequoia Bemper­ virens/plechnum Bpicant, Sequoia sempervirens/ Mahonia nervosa, and Sequoia sempervirens/Arbutus menziesii Forest Types of Little Lost Man Creek, CA. Numbers and Symbols Are ZM Cover-Abundance Values (Becking, 1957). Stand Indentification Numbers Are Printed Vertically.

Stand No.I 100110000002110 546427665443321 755547777664875 143536875213729 275478738608059 054326321619051

I I Sequoia/ I Sequoia/ISequoia/ Species Blechnum MahoniaIArbutus

Disporum smithii 111I 111 Athyrium filix-femina 12 121I 1 Adiantum pedatum 1 1I1 Rubus spectabilis 11I 1 Menziesia ferruginea 112 1 Corylus cornuta 223 I2I 32 Rhamnus purshiana 11 1333222133 Picea sitchensis 1111 Blechnum spicant 222223 1 11312 11+ Stachys mexicana 21I1 Adenocaulon bicolor 1I 11 Vancouveria hexandra 11 1I111 +1 Oxalis oregana 333333222223223 222221 I2I + 11 Polystichum munitum 555555543225544 222332112212211 111I2 1I + + Vaccinium parvifolium 21 133113142411 111I+211 1 + 12 1 I+I + Trillium ovatum 11I+11+1 ' I1I+ 11 I1I 11+1 Disporum hookeri ++ + I1I 1 1 1 Carex hendersonii 1 Actaea rubra Abies grandis 1 Iris douglasiana 11 I1I 1 1 1 Festuca subulata + 11 1I 11211 Mahonia nervosa 1+I+1 2 22212121222 211 1I 2 Galium aparine + 1 I1 1I+I1 1111 11111111 1I 1 Tsuga heterophylla 1I1 11111 1 12 I1 1I1 1 1I 2 1 Sequoia sempervirens 344454333332423 233333333333333 2232332222 2222 Trientalis borealis +I+ +1I1 1+I1 1I 111 Gaultheria shallon 1121321+2 333333434232433 213112344521352 Vaccinium ovatum 212121233421131 121I 13343344344 454452323234315 Lithocarpus densiflora 312I3 21I 4232 243335333443424 354435444545544 Whipplea modesta 1 1I1+111I1 1 + + Hierochloe occidentalis 121I 11112 211 Viola sempervirens + 1I+I 222222222121212 1221121 111+111 Festuca occidentalis 11 Rhododendron macrophyllum 11I3211I1 13243231233233 444333341224I 3 Pseudotsuga menziesii 1 11 11I2221 3 22 222223333233 231231222221112 Lathyrus vestitus + 111 1 Cardamine integrifolia 1I+1111 2 1 I1I 111 Pleuricospora fimbriolata + 1+ 11 Pteridium aquilinum 1211+1 1+ Goodyera oblongitolia 1I+ 21I11 1+I1I111I 11 Corallorhiza maculata 1+ I2 1I 111 Arbutus menziesii 1+1 1I11I 2121222 IIIIIIII

46

Table 9. Summary Synthesis Table Showing Species Constancy /Modal Cover Values in the Sequoia sempervirens/ .Blechnum spicant, Sequoia semperviISMz/Mahonia nervosa, and Sequoia Bempervirens/Arbutus menz­ ziesii Associations of Little Lost Man Creek, CA. Constancy Values Are Species Frequencies within Equal Groups of Representitive Stands. Cover Values Are from the ZM Cover-Abundance Scale (Becking, 1957).

I Species Sequoia/ISequoia/I Sequoia/ BlechnumIMahoniaI Arbutus

Canopy Trees Picea sitchensisI0.27/1 I Tsuga heterophyllaI0.47/1 0.47/1I 0.20/1 Sequoia sempervirensI1.00/3 1.00/3 I 0.93/2 Pseudotsuga menziesiiI0.67/1 0.93/2 1.00/2 Abies grandis 0.07/1 Subcanopy Trees Lithocarpus densifloraI0.67/2 1.00/3I 1.00/4 Arbutus menziesii 0.20/1I 0.67/2 Small Trees / Tall Shrubs Rubus spectabilis 0.20/1 Menziesia ferruginea 0.27/1 Corylus cornuta 0.40/2 Rhamnus purshiana 0.80/3 I Vaccinium parvifolium 0.93/1 0.60/1 I 0.33/+ Vaccinium ovatum 1.00/1 0.93/4I 1.00/4 Rhododendron macrophyllum 0.47/1 0.93/3 0.87/4 Low Shrubs / Herbs Disporum smithii 0.40/1 Athyrium filix-femina 0.40/1 Adiantum pedatum 0.20/1 Stachys mexicana 0.20/1 Adenocaulon bicolor 0.20/1 Blechnum spicant 0.80/2 0.20/1 Vancouveria hexandra 0.40/1 0.13/+ Oxalis oregana 1.00/3 0.67/2 Polystichum munitum 1.00/5 1.00/2 0.47/1 Trillium ovatum 0.47/1 0.13/+ 0.47/1 Disporum hookeri 0.13/+ 0.20/1 0.13/1 Gaultheria shallon 0.60/1 1.00/3 1.00/2 Trientalis borealis 0.13/+ 0.20/1 0.47/1 Viola sempervirens 0.27/+ 1.00/2 0.93/1 Mahonia nervosa 0.33/+ 0.92/2 0.13/1 Galium aparine 0.33/1 0.87/1 0.13/1 Festuca subulata 0.07/+ 0.53/1 Carex hendersonii 0.07/1 Actaea rubra 0.07/+ Iris douglasiana 0.40/1 Hierchloe occidentalis 0.73/1 Lathyrus vestitus 0.33/1 Whipplea modesta 0.60/1 0.13/+ Festuca occidentalis 0.07/+ 0.13/1 Cardamine integrifolia 0.47/1 0.33/1 Goodyera oblongifolia 0.40/1 0.53/1 Corallorhiza maculata 0.47/1 Pleuricospora fimbriolata 0.33/1 Pteridium aquilinum 0.53/1 47

Table 10. Fidelity Table Showing Species Fidelity Classes for the Sequoia sempervirens/Blechnum spicant, Sequoia sempervirens/Mahonia nervosa, and Sequoia sempervirens/Arbutus menziesii Associations of Little Lost Man Creek, CA. Fidelity Was Calcula­ ted as the Species Presence in an Equal Set of Representitive Stands Expressed as a Percentage the Species Presence in the Total Set of Stands. Fidelity Values Are Expressed in Fidelity Classes II-V (Becking, 1957). Character Species Are in Fidelity Classes III-V.

I Species Sequoia/ISequoia/ Sequoia/ Blechnum Mahonia Arbutus

Canopy Trees Picea sitchensisIV Tsuga heterophyllaIII IIII II Sequoia sempervirens�II IIII II Pseudotsuga menziesiiIII IIII II Abies grandisI V Subcanopy Trees Lithocarpus densifloraIIIIIIIII Arbutus menziesiiI IIIIV Small Trees / Tall Shrubs Rubus spectabilisIV Menziesia ferrugineaIV Corylus cornutaIV Rhamnus purshianaIV Vaccinium parvioliumIIII IIII II Vaccinium ovatumIII IIII II Rhododendron marophyllumIII IIII II Low Shrubs / Herbs Disporum smithiiIV Athyrium filix-feminaIV Adiantum pedatumIV Stachys mexicanaIV Adenocaulon bicolorIV Blechnum spicantIIVIII Vancouveria hexandraIIVI II Oxalis oreganaIIIIIII Polystichum munitumIII IIII II Trillium ovatumIII IIII II Disporum hookeriIII IIII II Gaultheria shallonIII IIII II Trientalis borealisIII IIII III Viola sempervirensIII IIII II Mahonia nervosaIII IIIII II Galium aparineIII IIIII II Festuca subulataIIIIIV Carex hendersoniiI V Actaea rubraI V Iris douglasianaI V Hierchloe occidentalisIV Lathyrus vestitusI V Whipplea modestaI IVIII Festuca occidentalisIIII III Cardamine integrifoliaIIIIIII Goodyera oblongifoliaIIII III Corallorhiza maculataI V Pleuricospora fimbriolata I V Pteridium aquilinumI V 48 Table 11. Mean Basal Area (BA) (M 2 /Ha) and Relative Dominance (RD) (Percent) of Canopy Species in the Sequoia sempervirens/Blechum spi­ cant, Sequoia sempervirens/Mahonia nervosa, and Sequoia sempervirens/Arbutus menziesii Associations in Little Lost Man Creek, CA. Relative Dominance is the Species' Basal Area Expressed as a Percentage of the Total Basal Area.

Sequoia/ ��Sequoia! Sequoia! Blechnum Mahonia �Arbutus

Canopy Species BA RD BA RD BA RD

Picea sitchensis 2 1 Tsuga heterophylla 4 3 7 5 2 3 Sequoia sempervirens 126 82 84 59 39 54 Pseudotsuga menziesii 22 14 52 35 31 43 Abies grandis 2 1 Total Basal Area 154 145 72 49

The first discriminant function based on topographic variables accounted for 98 percent of the total variance. The second function accounted for just two percent of the total variance and was therefore considered insignificant. The index of discriminatory power indicated 86 percent of the variability of the first function was due to differentiation among typical stand groups. The structure coefficients for the first function (Table 12) showed elevation, topographic configuration, and slope position were the topographic variables that best distinguished the forest types. A plot of the typical stands in the discriminant space formed by the scores on the first two functions (Figure 9) showed the forest types were well-separated along the axis representing the first function (the axis representing the insignificant second function was included for graphical purposes only). The mapped distribution of the forest types showed a relatively weak correspondence with the distribution of the soil series as mapped by the Soil-Vegetation Survey (DeLapp et al., 1961a, 1961b). Comparisons of soil profile descriptions from the groups of typical stands revealed a somewhat stronger correlation between forest types and soil series. Soils on the concave lower slopes at the northern end of the study area where the Sequoia/Blechnum type was best developed were generally dark brown to brown clay loams with gleyed subsurface horizons. �These were the most Melbourne-like soils in the study area. �Soils on convex upper slopes and ridgetops at the southern end of the study 50 Table 12. Structure Coefficents for First Discriminant Function Derived for Discriminant Analysis Relating Distribution of Forest Types to Topo­ graphic Factors in Little Lost Man Creek, CA. Large Coefficients Indicate Factors Most Re­ lated to Forest Type Distribution.

Variable Coefficient

Stand Elevation 0.61730 Topograpic Configuration 0.21580 Slope Position 0.11148 Radiation Index 0.10696 North-South -0.08558 East-West 0.06079 Slope -0.06028 Figure 9.DistributionofSample Stands ontheAxisRepre­

Seco nd Discr im inan t Func t ion nervosa association, (3)Sequoiasempervirens/Arb­ Associations toTopographicFactors intheLittle Analysis toRelatetheDistribution oftheForest National Park,CA.TheAxisforthe SecondNon- Lost ManCreekResearchNaturalArea, Redwood utus association. Group Codes:(1) Sequoia sempervirens/Blechnumspi­ 1 cant association, (2)Sequoiasempervirens/Mahonia 11 Purposes Only.Asterisks IndicateGroupCentroids. Significant FunctionWasIncluded for Graphical senting theFirstFunctionDerived byDiscriminant 1 11 1 1 1 1 * 1 1 1 1 First DiscriminantFunction 2 2 2 2 2 2 22 2 2 2 22 *22 2 2 2 2222 2 3 3 2 2 3 33* 2 3 3 3 3 33 32 3 3 3 3 51 52

area where the Sequoia/Arbutus type was best developed were generally grayish brown to pale brown gravelly loams and clay loams. These were the most Hugo-like soils in the study area. The soils occupied by the Sequoia/Mahonia type were generally intermediate in character between the Melbourne and Hugo extremes. The edaphic, topographic, and floristic character of the three forest types are summarized and contrasted in Table 13. 53

Table 13. Synoptic Table for the Sequoia sempervirens/ Blechnum spicant, Sequoia sempervirens/Mahonia nervosa, and Sequoia sempervirens/Arbutus lama­ ziesii Associations of Little Lost Man Creek, CA Showing Mean Cover of Each Life-form Stratum, Dominant Species with Modal Cover of 5-25 Percent or Greater, Character Species in Fidelity Classes III-V, and Typical Topographic and Edaphic Con­ ditions in Each Association.

� Sequoia / Blechnum Sequoia / Mahonii �Sequoia / Arbutus

Canopy Trees � Mean Cover (percent) 70 � 60 �40 Dominant Species Sequoia sempervirens � Sequoia sempervirens � Pseudotsuga menziesii Pseudotsuga menziesii Sequoia sempervirens Character Species Picea sitchensis � Abies grandis Subcanopy Trees Mean Cover (percent) 35 50 75 Dominant Species Lithocarpus densiflora Lithocarpus densiflora Character Species Arbutus menziesii Small Trees and Tall Shrubs Mean Cover (percent) 30 60 � 70 Dominant Species Rhamnus purshiana Vaccinium ovatum �Vaccinium ovatum Rhododendron macrophyllum Rhododendron macrophyllum Character Species Rhamnus purshiana Corylus cornuta Menziesia ferruginea Vaccinium parvifolium Rubus spectabilis Low Shrubs and Herbs Mean Cover (percent) 90 50 30 Dominant Species Polystichum munitum Gaultheria shallon Gaultheria shallon Oxalis oregana Viola sempervirens Blechnum spicant Mahonia nervosa Polystichum munitum Character Species Blechnum spicant Mahonia nervosa Pteridium aquilinum Oxalis oregana Hierchloe occidentalis Pleuricoapora fimbriolata Athyrium filix-femina Lathyrus vestitus Corallorhiza maculate Adiantum pedatum Whipplea modesta Goodyera oblongifolia Stachys mexicana Galium triflorum Festuca occidentalis Disporum smithii Festuca subulata Trientalis borealis Adenocaulon bicolor Iris douglaaiana Vancouveria hexandra Cardamine integrifolia Elevation Range (meters) � 60-310 240-555 430-620 Topographic Position � moist sites, esp. mesic sites, esp. xeric sites, esp. concave lower slopes even middle slopes convex upper slopes and ridgetope � Soil Series � Melbourne Melbourne-Hugo Hugo-Melbourne Soil Texture clay loam gravelly clay loam gravelly loam to clay loam Surface/Subsurface Color � dark brown / brown brown / brown to grayish brown / strong with gley mottling strong brown brown to pale brown DISCUSSION

Ecological Relationships

Elevation, ��slope position, ��and topographic configuration were the topographic variables that best discriminated among the forest associations of the study area. �The Sequoia/Blechnum association was found at lower elevations and primarily on lower �slopes where the topographic configuration was generally concave. �The Sequoia/Arbutus association was found at higher elevations and primarily on upper slopes and ridgetops where the topographic configuration was generally convex. The Sequoia/Mahonia association was found at intermediate positions in the study area (i.e. convex upper slopes and ridgetops at lower elevations and concave lower slopes at higher elevations). The distribution of the forest associations along this topographic gradient formed a "toposequence" (Barbour et al., 1980). Toposequences formed in regional vegetation types were commonly interpreted as expressions of variation in microclimate and soil moisture availability (e.g. Becking, 1967; Franklin and Dyrness, 1973; Zinke, 1977; Sawyer et al., 1977; Veirs, 1982). Concave lower slopes tend to be cooler because they accumulate cold air and thus radiate heat more rapidly at night, while convex upper slopes and ridgetops tend to be

54 55 warmer because they drain off cold air and thus radiate heat less rapidly at night (Spurr and Barnes, 1980). Soil moisture availability is generally greater on concave lower slopes because evapotranspiration rates are lower in cool microclimates and because both water and soil material is transported downslope towards these sites, while there is generally less soil moisture availability on convex upper slopes and ridgetops because evapotranspiration rates are higher in warm microclimates and because water and soil material is transported downhill away from these sites (Buol et al., 1980). In the study area, extremes in microclimate and soil moisture availability from one end of the toposequence to the other (i.e. from relatively cool, moist sites on concave lower slopes to relatively warm, xeric sites on convex upper slopes and ridgetops) were probably accentuated by the higher incidence of summer fog at lower elevations in the study area (Freeman, 1971). Fog cover reduces both insolation and evapotranspiration, and thus concave lower slopes at lower elevations where the incidence of fog was highest were probably the most cool and moist sites in the study area. Convex upper slopes and ridgetops at higher elevations where the incidence of fog was lowest were probably the most warm and xeric sites in the study area. The Sequoia/Blechnum association (Figure 10) attained its best development on lower slopes at lower elevations at the moist end of the toposequence. Here the fog was probably 56

Figure 10. Photograph of Stand No. 3 Representing the Sequoia sempervirens/Blechnum spicant Asso­ ciation of Little Lost Man Creek, CA. 57 frequent during the otherwise dry summer months (Freeman, 1971) , and consequently, the relative humidity was probably higher, especially in cool, sheltered sites. The fine-textured colluvial soils on concave topography also probably contributed to a relatively high moisture regime. These Melbourne-like soils were morphologically those with the highest moisture-holding capacity in the study area, and the gley-mottled subsurface horizons were indicative of downslope movement of groundwater towards the lower slope sites (Buol et al., 1980). Sequoia sempervirens dominance in the canopy of the Sequoia/Blechnum association was probably related to a favorable moisture regime and the effects of disturbance factors. Fire and mass-wasting of slopes were two major agents of disturbance in similar forest-types throughout the region (Becking, 1967; Franklin and Dyrness, 1973; Zinke, 1977; Sawyer et al. , 1977; Veirs, 1982; Lennox, 1982). Earthflows and slumps were the most common mass-wasting events on moist lower slopes near the study area (Coleman, 1973). In stands near the study area, Sequoia sempervirens attained dominance on moist mineral soils exposed after landsliding either directly through seedling dominance (Lennox, 1982) or indirectly through an Alnus rubra seral stage (Becking, 1967). In either case, Sequoia sempervirens probably gained dominance by virtue of its longevity and immense stature on sites with a favorable moisture status. Assuming a cool microclimate, relatively high humidity, and 58 moist soils, and given the lack of a dense woody understory in the Sequoia/Blechnum association, it is likely that the frequency of fire was relatively low and surface fires were most common on the lower slopes at lower elevations in the study area (Barbour et al., 1980). Sequoia sempervirens maintained its dominance under a light fire regime in old-growth redwood stands near the study area (Veirs, 1982). In the presence of infrequent surface fires on moist lower slopes, the longevity, stature, and fire-resistant characteristics of Sequoia sempervirens (Stone et al., 1972) probably favored the species over its shade-tolerant but less fire-resistant associates. The shade-intolerant Pseudotsuga menziesii (Fowells, 1965) was found in different size-classes widely scattered throughout lower slope stands in the study area. This mode of occurrence suggested establishment in association with small gaps in the canopy formed by windthrow or fire. Species comprising the subcanopy and small tree/large shrub layers in the Sequoia/Blechnum association probably needed more phytosynthetic energy to maintain their woody structure (Whittaker, 1970) , so these strata were not well-developed except in association with gaps in the dense canopy. 221yStichum munitum and other fern species in the dense low shrub/herb layer were indicative of low light and high moisture in old-growth redwood forests (Waring and Major, 1964). The Sequoia/Arbutus association (Figure 11) was best developed on upper slopes and ridgetops at higher elevations 59

Figure 11. Photograph of Stand No. 66 Representing the Sequoia sempervirena/Arbutus menziesii asso­ ciation of Little Lost Man Creek, CA. 60

at the xeric end of the toposequence. �At elevations above the 457 m (1500 ft) inversion layer, the incidence of summer fog was probably low (Freeman, 1971). Consequently, the relative humidity was probably lower, especially in warm, exposed positions. The coarse-textured residual soils on convex topography also probably contributed to a xeric moisture regime. The Hugo-like soils occupied by the Sequoia/Arbutus association were morphologically those with the lowest moisture-holding capacity in the study area (Buol et al., 1980). The co-dominance of �Sequoia sempervirens and Pseudotsuga menziesii and the dense subcanopy of sclerophyllous hardwoods in the Sequoia/Arbutus association was probably related to a relatively xeric moisture regime and the effects of fire. Assuming a warm microclimate, relatively high humidity, and relatively dry soils, and given the greater exposure to lightning strikes near ridgetops and the dense woody understory of the Sequoia/Arbutus association, it is likely that the frequency of fire was relatively high and destructive crown fires were more common on upper slopes and ridgetops in the study area (Barbour et al., 1980). �Sequoia sempervirens maintained co-dominance with Pseudotsuga menziesii, Lithocarpus densiflora, �and Arbutus menziesii under a relatively frequent and intense fire regime in old-growth redwood stands on upper slopes and ridgetops near the study area (Becking, 1967; Veirs, 1982). Here Sequoia sempervirens and hardwoods probably sprouted 61

from lignotubers and thus survived destructive crown fires which opened up the canopy and exposed dry mineral soils and thus �favored establishement of �Pseudotsuga menziesii seedlings. The light intensity under the canopy and subcanopy of the Sequoia/Arbutus association in the study area was evidently sufficient to support the dense, woody structure of the small tree/large shrub layer, but the dense shade, dry soils, and thick accumulations of Lithocarpus densiflora litter below this stratum probably inhibited the growth and germination of many species other than mycotrophic ones (Furman and Trappe, 1971; Franklin et al., 1981). The sparse low shrub/herb layer in the Sequoia/Arbutus association was characterized by several species which parasitize mycorrhizal fungi for nutrients, including Goocyera �oblongifolia, �Pleuricospora fimbriolata, �and Corallorhiza maculata. Sites occupied by the Sequoia/Mahonia association (Figure 12) formed the mesic portion of the toposequence between the moist and xeric extremes. This forest association seemed to occur throughout the watershed wherever a combination of topographic and edaphic factors formed a relatively mesic moisture regime. At lower elevations where fogs were probably more frequent, the Sequoia/Mahonia association occupied exposed, convex positions on ridgetops

and upper slopes where soils were often shallow and gravelly. At higher elevations where fogs were probably less frequent, the mesic association occupied cool, sheltered sites adjacent 62

Figure 12. Photograph of Stand No. 38 Representing the SequQia sempervirens/Mahonid nervosa asso­ ciation of Little Lost Man Creek, CA. 63 to the creek, especially in concave positions where soils were deeper and finer-textured. In the Sequoia/Mahonia association, �Pseudotsuga, menziesii was nearly co-dominant with Sequoia sempervirens as in the xeric association, but the total basal area and crown cover of canopy species were as great as those in the moist association. The subcanopy of Lithocarpus densiflora was better developed in the Sequoia/Mahonia association compared to the moist association, but the subcanopy had a lower mean cover than in the xeric association and lacked Arbutus menziesii. The structure and composition of stands in the Sequoia/Mahonia association seemed to indicate an intermediate moisture and fire regime relative to those described for the moist and xeric types. The canopy structure probably created intermediate light levels in the understory of the Sequoia/Mahonia association relative to those in the moist and xeric types. The light penetration was evidently sufficient to support a well-developed small tree/large shrub layer similar in composition to that in the xeric association but lower in cover. The low shrub/herb stratum in the Sequoia/Mahonia association was the most diverse in the study area. Dominant species in this layer included species common in the moist forest association (i.e. Polystichum mural= and Oxalis oregana) as well as those common in the xeric association (i.e. Gaultheria shallon and Viola sempervirens). The co-dominance of typically moist and xeric species and the presence of several characteristic 64

species in the low shrub/herb stratum seemed to indicate the mesic status of the Sequoia/Mahonia association. Character species such as Hierchloe occidentalis, Iris douglasiana, and Whipplea podesta were indicative of both intermediate light and mesic moisture conditions in old-growth redwood forests (Waring and Major, 1964).

Relationships to Regional Vegetation Types

The forest type classification scheme for Little Lost Man Creek was closely related to Becking's (1967) redwood alliance classification. The forest types of the study area were identified as associations within Becking's forest alliances based on shared floristic characteristics. The Sequoia/Blechnum association was placed in the "Redwood-Oxalis" alliance based on the shared characteristics of the strong dominance of Sequoia sempervirens in the overstory; the sparse small tree/large shrub layer characterized by Rhamnus purshiana, Vaccinium parvifoliuM, and Rubus spectabilis; the dense low shrub/herb layer

dominated by Polystichum =lit ► and Oxalis oregana and characterized by Blechnum spicant, Athyrium filix-femina, Adiantum pedatum, $tachys mexicana, and Disporum smithii. Both the Sequoia/Mahonia and Sequoia/Arbutus associations in the study area were placed in the "Redwood-Swordfern" alliance based on shared characteristics of the co-dominance of Sequoia sempervirens, Pseudotsuga menziesii, and 65

Lithocarpus densiflora and the characteristic presence of Arbutus menziesii in a two-layered overstory; the dense small tree/large shrub layer comprised of Rhododendron macrophyllum and Vaccinium ovatum; the low shrub/herb layer co-dominated by Polystichum munitum and Gaultheria shallop and characterized by Mahonia nervosa, Viola sempervirens, Actaea arguta, Hierchloe occidentalis, Lathyrus vestitus, Festuca occidentalis, Corallorhiza maculate, and Goodyera oblongifolia. The hierarchial classification of the associations within Becking's alliances was supported by the results of the TWINSPAN analysis. The two groups of stands formed at the first division level (Appendix B) represented the classification at the alliance level. Pseudospecies with high preference scores on the "Redwood-Oxalis" side of the dichotomized ordination included Psay_stichum munitum, Oxalis oregana, Rhamnus purshiana, and Blechnum spicant. Pseudotsuga menziesii and Viola sempervirens scored high on the "Redwood-Swordfern" side. The floristic composition of the forest associations in Little Lost Man Creek also indicated their relationship to redwood border forest types. Kuchler (1977) on his vegetation map of California showed the "Picea-Abies" forest type (Picea sitchensis-Abies grandis) forming a narrow strip between the ocean and the western edge of the redwood forest. Species characteristic of the Sequoia/Blechnum association including Picea sitchensis, Vaccinium parvifolium, Menzieisia ferruginea, Polystichum munitum, and Oxalis oregana were 66

common components of coastal "Picea-Abies" forests (Franklin and Dyrness, 1973). "Tsuga-Pseudotsuga" forests �(Tsuga heterophylla- Pseudotsuga menziesii) formed a border with redwood forest at the northern edge of Sequoia's range in southwest Oregon, about 98 km north of the study area. Franklin and Dyrness (1973) described Sequoia sempervirens forests as variants of the "Tsuga-Pseudotsuga" forests of western Washington and Oregon. "Tsuga-Pseudotsuga" forest types described by Becking (1954) , Bailey (1966) , and Franklin and Dyrness (1973) showed a close floristic relationship to the moist Sequoia/Blechnum and mesic Sequoia/Mahonia associations in the study area. Moist "Tsuga-Pseudotsuga" forest types were characterized by the dominance of _Polystichum munitum, Oxalis oregana and the presence of species such as Blechnum spicant, Athyrium filix-femina, Adiantum Daidatmm, Disporum smithii, and Vancouveria hexandra. Mesic types were characterized by the dominance of Rhododendron macrophyllum, Vaccinium ovatum, Gaultheria shallon, Nahonia nervosa, and Viola sempervirens. Kuchler's (1977) map showed mixed evergreen forest forming the eastern border forest in the region of the study area. The mixed evergreen forest types of California were described by Sawyer et al., (1977). Their "Psuedotsuga-hardwood" forest type (Pseudotsuga menziesii) showed a floristic relationship to the xeric Sequoia/Arbutus association in the study area. Both types had a two-level canopy co-dominated by Pseudotsuga menziesii and Lithocarpus 67 densiflora and characterized by the presence of Arbutus menziesii. The understories in both types were dominated by Rhododendron maLL2phyliam, Vaccinium ovatum, and Gaultheria shallop. Pteridium AgialiliMam and Goodyera oblongifolia were shared characteristic herbaceous species. Barrows (1984) noted the presence of several mycotrophic species in the understory of old-growth "Pseudotsuga-hardwood" forests, including species of Allotropa, Pityopus, and pyrola also found in the Sequoia/Arbutus association in the study area. There was no definite southern redwood border forest in the region of the study area. Kuchler (1977) mapped "Sequoia-Pseudotsuga" (Sequoia sempervirens-Pseudotsuga menziesii) as a near-continuous unit extending south from the Oregon border to Monterey county. Descriptions of redwood vegetation in southern Humboldt county (Becking, 1967), Mendocino county (Zinke, 1977) , Marin county (Howell, 1970, McBride and Jacobs, 1978) , Santa Cruz county (Cooney-Lazaneo and Lyons, 1981), and Monterey county (Becking, 1971) indicated a compositional gradient from conifer to hardwood dominance on a transect extending south from the study area. Moving south through the redwood forest belt, Lithocarpus densiflora and Arbutus menziesii were increasingly important stand components. The composition and structure of modal (i.e. mesic) redwood forests from Mendocino county south to Santa Cruz county were much like those in the xeric Sequoia/Arbutus association, except for the presence of Umbellularia californica and Ouercus species in the southern 68 stands. In Monterey county at the southern edge of Sequoia sempervirens' range, the species occurred as isolated groves within mixed hardwood forests dominated by Lithocarpus densiflora, Arbutus menziesii, Ouercus sps., and Pinus coulteri. Here Pseudotsuga menziesii and characteristic understory components of the northern redwood forest and related forest types were very scarcely represented (Becking, 1971). Together the old-growth forest associations of Little Lost Man Creek and the related seral types of the lower Redwood Creek basin (Muldavin et al., 1981; Lenihan et al., 1982; Appendix C) helped define regional redwood habitat-types. Both Stone et al., (1972) and Daubenmire (1975) recommended definition of redwood habitat-types for the purposes of vegetation management in Redwood National Park. Here National Park Service policy (Leopold et al. , 1963) mandated restoring about 5,700 ha of cutover parkland to a more pristine appearance (i.e. as they appeared before the advent of white man). A forest habitat-type includes all land areas potentially capable of producing the same old-growth forest community (Daubenmire, 1952). A habitat-type map based on the floristic relationships between old-growth forest associations and seral vegetation types would therefore represent the distribution of the old-growth associations under pristine conditions in the lower Redwood Creek basin. Management treatments of seral vegetation would probably produce similar within-type results, and these 69

results could be evaluated by reference to the standard composition and structure of corresponding old-growth forest associations in the Little Lost Man Research Natural Area. REFERENCES CITED

Azevedo J., and D.L. Morgan. 1974. Fog precipitation in coastal California forests. Ecology 55: 1135-1141. Bailey, A.W. 1966. Forest associations and secondary plant succession in the southern Oregon Coast Range. Ph.D. thesis, Oregon State University. Corvallis. 160 p. Bailey, E.H. 1966. Geology of northern California. Bulletin 190. California Division of Mines and Geology. San Francisco, California. 507 p. Barbour, M.G., J.H. Burke, and W.D. Pitts. 1980. Terrestrial plant ecology. Benjamin/Cummings Publishing Co. Menlo Park, California. 604 p. Barrows, K. 1984. Old-growth Douglas-fir forests. Fremontia 11: 20-23. Becking, R.W. 1954. Site indicators and forest types of the Douglas-fir region of western Washington and Oregon. Ph.D. thesis. University of Washington, Seattle. 133 p. �. 1957. The Zurich-Montpellier School of Phytosociology. Botanical Review 23: 411-488. �. 1967. The ecology of the coastal redwood forest and the impact of the 1964 floods upon redwood vegetation. Final Report. NSF Grant GB No. 3468. 91 p. (on file at Humboldt State University library, Arcata, California) �. 1971. The ecology of the coastal redwood forests of northern California and the impact of the 1964 flood upon redwood vegetation. Final Report. NSF Grant GB No. 6310. 158 p. (on file at Humboldt State University library, Arcata, California) �. 1982. A pocket flora of the redwood forest. Island Press, Covelo, Ca. 237 p. Braun-Blanquet, J. 1932. Plant sociology: the study of plant communities. Revised edition. Trans. by G.D. Fuller and H.S. Conrad (eds.). Haftner Publishing Co. New York. 439 p. Buol, S.W., F.D. Hole, and R.J. McCracken. 1980. Soil genesis and classification. Iowa State University Press. Ames. 404 p.

70 71

Byers, H.R. 1953. Coast redwood and fog drip. Ecology 34: 192-193. California Dept. of Water Resources. 1965. Land and water use in the Redwood Creek Hydrographic Unit. Bulletin 94-7. Sacramento. 80 p. Clark, H.W. 1937. Association types in the north coast ranges of California. Ecology 18: 214-230. Clements, F.E. 1920. Plant indicators: the relation of plant communities to process and practice. Carnegie Inst. Wash. Publ. 290. Washington, D.C. 388 p. Coleman, R.J. 1973. The history of mass-wasting processes in the lower Redwood Creek basin, Humboldt County, California. M.S. thesis, Pennsylvania State University. University Station. 120 p. Colwell, W.L., Jr. 1974. Soil-vegetation maps of California. U.S.D.A. Forest Service Res. Paper PSW-76. Pacific Southwest Forest and Range Experiment Station. Berkeley, California. 54 p. Cooney-Lazano, M.B. and K. Lyons. 1981. Plants of Big Basin Redwoods State Park and the coastal mountains of northern California. Mountain Press Publishing Co. Missoula, Montana. 142 p. Crawford, R.M. and D. Wishart. 1968. A rapid classification and ordination method and its application to vegetation mapping. Journal of Ecology 56: 385-404. Daubenmire, R.F. 1952. Forest vegetation of northern Idaho and adjacent Washington, and its bearing on concepts of vegetation classification. Ecological Monographs 22: 301-330. �. 1975. The community status of the coastal redwood, Sequoia sempervirens. Prepared for the U.S. Dept. of Interior, National Park Service, 15 p. (on file at the Redwood National Park office, Arcata, California) �. 1976. The use of vegetation in assessing the productivity of forest lands. Botanical Review 42: 115-143. DeLapp, J.A., R. Skolman, and B. Smith. 1961a. Timber stand and vegetation cover map, soil-vegetation map, legends and interpretation - SE 1/4 of Orick quadrangle (10A4). California Cooperative Soil-Vegetation Survey. U.S.D.A. Forest Service, Pacific Southwest Forest and Range Experiment Station. Berkeley, California. 40 p. 72

�. 1961b. Timber stand and vegetation cover map, soil-vegetation map, legends and interpretation - SW 1/4 of Tectah Creek quadrangle (11B3). California Cooperative Soil-Vegetation Survey. U.S.D.A. Forest Service, Pacific Southwest Forest and Range Experiment Station. Berkeley, California. 37 p. �and B.F. Smith. 1977. Tables for the soil-vegetation map of the NE 1/4 of Orick quadrangle (10A-1). USDA Forest Service, Pacific Southwest Forest and Range Experiment Station. Berkeley, California. 5p. Dyrness, C.T., J.F. Franklin, and C. Maser. 1973. Wheeler Creek Research Natural Area. Supplement No. 1 to Federal Research Natural Areas in Oregon and Washington: a guidebook for scientists and educators. U.S.D.A. Forest Service, Pacific Northwest and Range Experiment Station. Portland, Oregon. 16 p. Elford, C.R. and M.R. McDonough. 1974. The climate of Humboldt and Del Norte Counties. Humboldt and Del Norte Counties Agricultural Extension Services, University of California. Berkeley. 52 p. Eyre, F.H. 1980. Forest cover types of the United States and Canada. Society of American Foresters. Washington, D.C. 148 p. Fowells, H.A. 1965. Silvics of forest trees of the United States. U.S.D.A. Forest Service. Government Printing Office. Washington, D.C. 527 p. Frank, B. and J. Lee. 1966. Potential solar beam irradiation on slopes: tables for 30 to 50 degrees latitude. U. S.D.A. Forest Service Research Paper RM-18. Rocky Mountain Forest and Range Experiment Station. 116 p. Franklin, J.F. and J.M. Trappe. 1968. Natural areas: needs, concepts, and criteria. Journal of Forestry 66: 456-461. and C.T. Dyrness. 1973. Natural vegetation of Oregon and Washington. U.S.D.A. Forest Service General Technical Report PNW-8. Portland, Oregon. 417 p. �, K. Cromack, Jr., W. Denison, A. McKee, C. Maser, J. Sedell, F. Swanson, and G. Juday. 1981. Ecological characteristics of old-growth Douglas-fir forests. U.S.D.A. Forest Service General Technical Report PNW-118. Pacific Northwest Forest and Range Experiment Station. Portland, Oregon. 48 p. Freeman, G.J. 1971. Summer fog drip in the coastal redwood forest. M.S. thesis, Humboldt State University. Arcata, California. 100 p. 73

Frenkle, R.E. and C.M. Harrison. 1974. An assessment of the usefulness of phytosociological and numerical classificatory methods for the community biogeographer. Journal of Biogeography 1:27-56. Furman, T.E. and J.M. Trappe. 1971. Phylogeny and ecology of mycotrophic achlorophyllous angiosperms. Quarterly Review of Biology 46: 219-225. Gauch, H.G. 1982. Multivariate analysis in community ecology. Cambridge University Press. New York. 298 p. �and R.H. Whittaker. 1981. Hierarchial classification of community data. Journal of Ecology 69: 537-557. Gothier, K.C. 1980. The impact of recreational hiking on the redwood forest. M.S. thesis, Humboldt State University. Arcata, California. 46 p. Grosenbaugh, L.R. 1952. Plotless timber estimates: new, fast, easy. Journal of Forestry 50: 32-37. Hill, M.O. 1979. TWINSPAN: a FORTRAN program for arranging multivariate data in an ordered two-way table by classification of the individuals and the attributes. Section of Ecology and Systematics, Cornell University. Ithaca, New York. 90 p. �, R.G. Bunce, and M.W. Shaw. 1975. Indicator species analysis, a polythetic divisive method of classification, and it's application to a survey of native pinewoods in Scotland. Journal of Ecology 63: 597-613. Howell, J.T. 1970. Marin flora. University of California Press. Berkeley. 366 p. Irwin, W.P. 1960. Geologic reconnaissance of the northern Coast Ranges and Klamath Mountains, California, with a summary of mineral resources. Bulletin 179. California Division of Mines and Geology. San Francisco. 507 p. Jahn, G. 1982. Application of vegetation science to forestry. Handbook of vegetation science, vol. 12. Dr.W Junk Publishers. The Hague, The Netherlands. 405 p. Jensen, H.A. 1947. A system for classifying vegetation in California. California Fish and Game 33(4). 266 p. Kartesz, J.T. and R. Kartesz. 1980. A synonymized checklist of the vascular flora of the United States, Canada, and Greenland. Vol. II. The Biota of North America. The University of North Carolina Press. Chapel Hill. 498 p. 74

Kercher, J.R. and R.A. Goldstein. 1977. Analysis of an eastern Tennessee oak hickory forest by canonical correlation of species and environmental parameters. Vegetatio 35: 153-163. Klecka, W.R. 1975. Discriminant analysis. Pages 434-467 in N.H. Nie (ed.) , SPSS: Statistical Package for the Social Sciences. McGraw-Hill Book Company. New York. 675 p. Kuchler, A.W. 1977. Appendix: the map of the natural vegetation of California. Pages 909-938 in M.G. Barbour and J. Major (eds.) , Terrestrial vegetation of California. John Wiley and Sons Publishers. New York. 1002 p. Layser, E.F. 1974. Vegetative classification: its application to forestry in the northern Rocky Mountains. Journal of Forestry 72: 354-357. Lenihan, �W.S. Lennox, E.H. Muldavin, and S.D. Veirs, Jr. 1982. A handbook for classifying early post-logging vegetation in the lower Redwood Creek basin. Redwood National Park Technical Report no. 7. U.S. Department of Interior, National Park Service. Arcata, California. 40 p. Lennox, W.S. 1983. Stand composition and diameter distribution in sixty-year-old second-growth coast redwood stands. Pages 125-138 in C. van Riper, L.D. Whitting, and M.L. Murphy (eds.) , Proceedings of the first biennial conference of research in California's National Parks. Cooperative Park Studies Unit, University of California. Davis. 357p. Leopold, A.S., S.A. Cain, D.M. Cottam, I.N. Gabrielson, and T.L. Kimball. 1963. Wildlife management in the national parks. Trans. North Am. Wildl. Nat. Res. Conf. 28:28-45. Lewis, H.T. 1973. Patterns of indian burning in California: ecology and ethnohistory. Ballena Press. Ramona, California. 101 p. Maarel, E. van der. 1979. Transformation of cover- abundance values in phytosociology and it's effect on community similarity. Vegetatio 39: 97-114. Mason, H.L. 1947. Evolution of certain floristic associations in western North America. Ecological Monographs 17: 201-210. McBride, J. and D. Jacobs. 1978. Vegetation history of Muir Woods National Monument. Final Report. National Park Service Contract No. CX 8000-6-0035. Department of 75

Forestry and Conservation, University of California. Berkeley. 39 p. Moore, J., P. Fitzsimons, E. Lambe, and J. White. 1970. A comparison and evaluation of some phytosociological techniques. Vegetatio 20: 1-20. Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and methods of vegetation ecology. John Wiley and Sons Publishers. New York. 547 p. Muldavin, E.H., J.M. Lenihan, W.S. Lennox, and S.D. Veirs, Jr. 1981. Vegetation succession in the first ten years following logging of coast redwood forests. Redwood National Park Technical Rep. no. 6. U.S. Department of Interior, National Park Service. Arcata, California. 69 p. Munz, P.A. and D.D. Keck. 1968. A California flora and supplement. University of California Press. Berkeley. 1681 + 224 p. Ornduff, R. 1974. An introduction to California plant life. University of California Press. Berkeley. 152 p. Parker, I. and W. Matyas. 1979. CALVEG: a classification of California vegetation. U.S.D.A. Forest Service, Pacific Southwest Region. Berkeley, California. 159 p. Pfister, R.D. 1 982. Designing succession models to meet management needs. Pages 44-53 in J. Means (ed.) , Forest succession and stand development in the Pacific Northwest. Proc. 1981 Symp. at Oregon State University, Corvallis, Oregon. Forest Research Laboratory. Corvallis. 170 p. Poore, M.E. 1955a. The use of phytosociological methods in ecological investigations. I. The Braun-Blanquet system. Journal of Ecology 43: 226-244. �. 1955b. The use of phytosociological methods in ecological investigations. II. Practical issues involved in an attempt to apply the Braun-Blanquet system. Journal of Ecology 43: 245-269. Roy, D.F. 1966. Sivicultural characteristics of redwood. U.S.D.A. Forest Service Research Paper PSW-28. Pacific Southwest Forest and Range Experiment Station. Berkeley, California. 20 p. Sawyer, J.0., D.A. Thornburg, and J.R. Griffin. 1977. Mixed evergreen forest. Pages 359-382 in M.G. Barbour and J. Major (eds.) , Terrestrial vegetation of California. John Wiley and Sons Publishers. New York. 1002 p. 76

Spurr, S.H. and B.V.Barnes. 1980. Forest ecology. John Wiley and Sons Publishers. New York. 687 p. Stone, E.C., R.F. Grah, and P.J. Zinke. 1972. Preservation of the primival redwood forest in the Redwood National Park, part I. American Forests 78: 50-55. Sudworth, G.B. 1908. Forest trees of the Pacific Slope. U.S.D.A. Forest Service. Government Printing Office. Washington, D.C. 441 p. Tatsuoka, M.M. 1970. Selected topics in advanced statistics No. 6.: discriminant analysis. Institute for Personality and Ability Testing. Campaign, Illinois. 57 p. Thornthwaite, C.W. 1948. An approach towards a rational classification of climate. Geographical Review 38: 55-94. U.S. Dept. of Interior. 1967. Forest resource study for the proposed Redwood National Park in Humboldt and Del Norte counties, California. Prepared for the National Park Service by Hammon, Jensen, and Wallen Mapping and Forestry Services, Oakland, California. 159 p. (on file at the Redwood National Park office, Arcata, California.) U.S. Federal Committee on Ecological Reserves. 1977. A directory of research natural areas on federal lands of the United States of America. Superintendent of Documents. Washington, D.C. 280 p. Veirs, S.D., Jr. 1982. Coast redwood forest: stand dynamics, successional status, and the role of fire. Pages 119-141 in J. Means (ed.), Forest succession and stand development in the Pacific Northwest. Proc. 1981 Symp. at Oregon State University, Corvallis, Oregon. Forest Research Laboratory. Corvallis. 170 p. Walter, H., E. Harnickell, and D. Mueller-Dombois. 1975. Climate-diagram maps of the individual continents and the ecological climatic regions of the earth. Springer-Velag Publishers. New York. 36 p. Waring, R.M. and J. Major. 1964. Some vegetation of the California coastal redwood region in relation to gradients of moisture, nutrients, light, and temperature. Ecological Monographs 34: 167-215. Westhoff, V. and E. van der Maarel. 1978. The Braun- Blanquet approach. Pages 287-399 in R.H. Whittaker (ed.), Classification of plant communities. Dr. W. Junk Publishers. The Hague, The Netherlands. 408 p. Williams, B.K. 1983. Some observations on the use of discriminant analysis in ecology. Ecology 64: 1283-1291. 77

Zinke, P.J. 1977. The redwood forest and associated north coast forests. Pages 679-698 in M. G. Barbour and J. Major (eds.) , Terrestrial vegetation of California. John Wiley and Sons Publishers. New York. 1002 p. APPENDIX A. List of Vascular Plant Species Found in the Little Lost Man Creek Research Natural Area, California. Taxonomy Follows Munz and Keck (1968) with Nomenclature Updated from Kartesz and Kartesz (1980).

Aceraceae Acer circinatum Pursh Acer magLophylaum Pursh Adiantaceae Adiantum pedatum L. var. aleuticum Rupr. Aspleniaceae Athyrium filix-femina (L.) Roth var. cyclosurUM Polystichum munitum (Kaulfuss) Presl Aristolochiaceae Asarum caudatum Lindl. Asteraceae Adenocaulon bicolor Hook. Cirsium vulgare (Savi) Tenore Erechtitea minima (Poir.) DC. Hieracium albiflorum Hook. Berberidaceae Achlys triphylla (Sm.) DC. Nahonia nervosa (Pursh) Nutt. Vancouveria hexandra (Hook.) Morr. & Dec. Betulaceae ].nus rubra Bong. Corylus cornuta Marsh. var. californica (A. DC.) Sharp Blechnaceae Blechnum spicant (L.) Sm. Woodwardia fimbriata Sm. Brassicaceae Cardamine integrifolia (Nutt.) Greene var. sinuata (Greene) C.L. Hitchc. Campanul aceae Campanula prenanthoides Dur. Caprifoliaceae Sambucus racemosa L. ssp. pubens (Michx.) House var. aborescens (Torr. & Gray) Gray Cucurbitaceae Marah oreganus (Torr. & Gray) T.J. Howell

78 79

APPENDIX A. List of Vascular Plant Species Found in the Little Lost Man Creek Research Natural Area, California. (continued)

Cyperaceae Cares bendersonii Bailey Ericaceae Allotrope virgata Torr. & Gray ex Gray Arbutus menziesii Pursh Gaultheria shallop Pursh Hemitomes congestun Gray Menziesia ferruginea Sm. Monotropa uniflora L. Pleurocispora fimbriolata Gray Pityopus californica (Eastw.) Copeland f. pyrola bracteata Hook. pyrola picta Sm. Rhododendron macrophyllum D. Don ex G. Don Vaccinium ovatum Pursh var. ovatum Vaccinium parvifolium Sm. Fabaceae Lathyrus vestitus Nutt. ex Torr. & Gray var. bolanderi (S. Wats.) C.L. Hitchc. Lotus aboriginus Jepson Fagaceae Castanopsis chrysophylla (Doug. ex Hook.) A. DC. Lithocarpus densiflora (Hook. & Arn.) Rehd. var. densiflora Ir idaceae • douglasiana Herbert Juncaceae Luzula parviflora (Ehrh.) Desv. var. parviflora Lamiaceae Stachys mexicana. Benth. Lauraceae Umbellularia californica (Hook. & Arn) Nutt. var. californica Liliaceae Clintonia andrewsiana Torr. Disporum hookeri (Torr.) Nichols var. hookeri Disporum smithii (Hook.) Piper Lilium columbianum Hanson ex Baker Smilacina racemosa (L.) Desf. Trillium ovatum Pursh ssp. ovatum 80 APPENDIX A. List of Vascular Plant Species Found in the Little Lost Man Creek Research Natural Area, California. (continued)

Papaveraceae Dicentra formosa (Haw.) Walp. ssp. formosa Poaceae Festuca occidentalis Hook. Festuca zubulata Trin. Hierchloe occidentalis Buckl. Melica subulata (Griseb.) Scribn. var. subulata Orchidaceae Corallorhiza maculata (Raf.) Raf. Corallorhiza striata Lindl. var. striata Goodyera oblongifolia Raf. Listera caurina Piper Orobanchaceae poschniaka strobilacea Gray Oxal idaceae Oxalis oregana Nutt. ex Torr. & Gray Pinaceae Abies grandis (Dougl. ex D. Don) Lindl. Picea sitchensis (Bong.) Carr. Pseudotsuga menziesii (Mirbel) Franco Tsuga heterophylla (Raf.) Sarg. Polypodiaceae Folypodium scouleri Hook. and Grey. Portul acaceae Claytonia sibirica L. var. sibirica Primul aceae Trientalis borelis Raf. ssp. latifolia (Hook.) Hulten Pteridaceae Pteridium aguilinum L. Knuhn var. pubescens Underw. Ranuncul aceae Actaea rubra (Ait.) Willd. ssp. arguta (Nutt.) Hulten Anemone deltoidea Hook. Rhamnaceae Rhamnus purshiana DC. Rosaceae Fragaria vesca L. var. californica (Cham. & Schlecht.) Staudt 81 APPENDIX A. List of Vascular Plant Species found in the Little Lost Man Creek Research Natural Area, California. (continued)

Rosa nutkana Presl. var. nutkana Rubus spectabilis Pursh var. zpectabilis Rubus vitifolius Cham. & Schlecht. var. vitifolius Rubiaceae Galium aparine L. Galium triflorum Michx. Saxifragaceae Whipplea modesta Torr. Scrophulariaceae Synthyris reniformis (Dougl. ex Benth.) Benth. var. cordata Gray Taxodiaceae .Sequoia .$empervirens (Lamb. ex D. Don) Endl. Viol aceae Viola glabella Nutt. Viola sempervirens Greene APPENDIX B. Structured Table Formed by TWINSPAN Showing the Two-Way Classification of Species and Sample Stands from Little Lost Man Creek, CA. Hori­ zontal Lines Divide Species Groups and Vertical Lines Divide Stand Groups. Stand Numbers Are Printed Vertically. Asterisks at the End of Division Lines Indicate Classification Division Levels. One Asterisk Indicates where Species or Stands were Divided into Two Groups at the First Division Level. Two Asterisks Indicate where the Two Groups of Species or Stands Formed at the First Level Were Divided into Four Groups at the Second Division Level. Three Asterisks Indicate where the Four Groups of Species or Stands Formed at the Second Level Were Divided into Eight Groups at the Third Division Level. The Stand Classification Was Interpreted at the Second Division Level where Three Typical Groups (i.e. Stand Numbers 11-09, 22-19, and 77-51) and One Transitional Group (i.e. Stand Numbers 34-16) Were Formed. Numbers and Plus Symbols Are ZM Cover-Abundance Scale Values (Becking, 1957).

Stand No.I100110 000002110 332276 4433223311 25443321 66654221877666554443321 77555547 7776664875 Specie.I 143536 875213729 429840 3191607386 22756510 95404744198732868608059 70754326 3218619051 Diaporum .mithiiI111 111 1 Athyrium fil ix-feminaI12 121 1 1 Adiantum pedetumI1 1 1 1 Rubus spectabilisI11 1 1 Mensiesia ferrugineaI112 1 1 1 Corylus cornutsI223 2I32 1 Rhamnus purshianaI11 1 333222133 1 1 1111111 + + Picea sitchensisI1 1I11 Blechnum epiceneI222223 1I11312 2I211 221222 1 +121+ + ••• Stachys mexicanaI21 1 11 1 + Adenocaulon bicolorI1 11 1I11 1 •• Vancouveria hesandraI11 1 111 11 121 + I11I1 1.1 1 Oxalis oreganaI333333 222223223 322322 2222222323 222I2112 122 2221I1I2 1I + 11 Polystichum munitueI555555 543225544 433444 3353433332 22212112 2231331122211I22 212211 11111I2 1I . + Vaccinium parvifoliumI21 133 113142411 22 2 322313111 211 1I+1 111 I11I+ 2I11 II1 + 12 1 .I + Trillium ovatumI11 +11+1 1111 1I1I1I1 1I1I12I+ 1I1 1+I 11+1 ••• Disporum hookeri a+ 1I11 11I11I11 1I+11I1I1 1 1 Cares hendersonii 1+11 1 1I1 1 Actaea cobra 1I111 + Abies grandis 1222 . 123I1 2 1I2122 1 Iris douglasiana +112 1+ 11I1+1 1I11I 1 1 Pestuca .ubulata + 111I11 1111I1 1111 + 1I1 1I 111+1211 mahonia nervoea 1+I+1 2 12 121 111212 21 22222222 22 122I21121222222 211 12I 2 ••• Galium aparineI+ 1 1 1I+ 111111 1 11111111 11 11111 111 11111 1 11111111111 1I 1 Tsuga hetecophyllaI1 1I11111 1 I1I1 2+ 11I1 1I1 1132 13I1I1I1I1 11 I 2I1 Sequoia sempervireneI344454 333332423 323333 3333332333 32323333 33223333323333333333333 32223233 22222 2222 Trientalia borealis .I+ 1I1 1 I1I1 1 I+I11 1I+11I1 1+ 1 11I 111 Gaultheria shallon 1121321+2 234222 3223222233 43324354 33253342533432433232433 22213112 3443521352 Vaccinium ovatumI212121 233421131 1I1111 13I1111422 1231223 2113 135123344334344344 44454452 3233234315 Lithocarpus densifloraI312 3I21I4232 454332 2433433434 22454433 53313342445334333443424 33454435 4443545544 •• Whipplea modesta +11.1+ 21I+111 111I1111+111111I1 1 . + Bierochloe occidentalis 11+ 21 .I1 21222222 11 +I1 11 11112112 211 + Viola sempervirensI+ + 1I+ 111122 12 +111.11 22222221 12222221222221222121212 21122112 1 1111+111 Feature occidentalis + 1 1 Rhododendron macrophyllum I11 3211I1 1I1 1212234421 1I2 232424 2123233143233233 44444 333 3414224I 3 Pseudotsuga menxiesiiI1 11 1 1I2221I3 11 3112112 1 12221212 3 232223322222333333233 22231231 2222221112 ••• Lathyrus vestitus + + 1I1I + 111 11 1 Cardamine integrifolia + 1 11 +I+111 +11I2 1 11I 111 Pleurracospora fimbriolata +1I1+ 11 Pteridium aquilinum 1 2111+1 1+ Goodyera oblongifolia .I. 1 1 1 +I2121 1 1I11 211+I1 1I11I 11 Corallorhisa maculate + 1 +I2 1 1I 111 Arbutus mensiesii 1+1+ 1I1 1I2 121222 • •

82 APPENDIX C. Relationships of Little Lost Man Creek Forest Associations to Regional Redwood Seral Types.

A classification scheme forIearlyI post-logging vegetation was developed for harvested redwood forest in the lower Redwood Creek basin of Redwood National Park (Muldavin et al., 1981; Lenihan et al., 1982). The floristic composition and topographic setting of the seral vegetation types indicated their relationship to the old-growth forest associations in Little Lost Man Creek. The seral types were distributed along a moisture gradient from moist lower slopes at low elevations to xeric upper slopes at relatively high elevations. Seral types within a particular moisture regime were distributed along a secondary gradient of soil disturbance. On moist lower slopes the seral vegetation types represented successional stages of the old-growth Sequoia/Blechnum association. The "Polystichum-Oxalis" type occupied moist, cable-yarded sites were soils were least disturbed. Remnant old-growth understory species comprised the dominant herbaceous layer, including Esay.stichum munitum, Oxalis oregana, Blechnum spicant, lithyrium filix-femina, Vancouveria hexandra, and Stachys mexicana. The weedy Erechtites minima was strongly represented for the first few years after logging. Tractor-yarding, landsliding, and other logging impacts that exposed moist mineral soils promoted the development of the "Alnus" seral type with a dense overstory of Alnus rubra and a relatively sparse understory of 2alyatiLlum munitum and Oxalis oregana. 83 84 APPENDIX C. Relationships of Little Lost Man Creek Forest Associations to Regional Redwood Seral Types. (continued)

Seral types on relatively mesic sites represented successionalIstages ofI the old-growth Sequoia/Mahonia association. The "Rhododendron-Vaccinium" type and its "Gaultheria" phase occupied mesic sites where soils were least disturbed (i.e. cable-yarded sites or islands among tractor skid trails). These types were dominated by a tall shrub layer of Rhododendron macrophyllum and Vaccinium ovatum sprouts. The low shrub/herb layer was comprised of typically mesic species including Gaultheria shallop, Hierchloe occidentalis, Iris douglasiana, and Whipplea modest,. The "Baccharis/Whipplea" type occupied mesic sites more severely disrupted by tractor-yarding or slash burning. Baccharis pilularis, a coastal scrub species scarcely represented in old-growth redwood forest, formed the dominant shrub layer in this type. The herb layer was strongly dominated by Whipplea modesta, a characteristic species with much lower cover in the mesic old-growth association. On xeric upper slopes in Redwood Creek the seral vegetation types represented successional stages of the old-growth Sequoia/Arbutus association. The "Lithocarpus/Whipplea" seral type occupied islands among tractor skid trails on upper slopes with northerly aspects. The dominant shrub layer composed of Lithocarpus densiflora sprouts and scattered Arbutus menziesii indicated the seral type's relationship to the xeric old-growth association. 85 APPENDIX C. Relationships of Little Lost Man Creek Forest Associations to Regional Redwood Seral Types. (continued)

Whipplea modesta, a species with veryI limited occurrence in the xeric old-growth association, dominated the herb layer in the xeric "Lithocarpus/Whipplea" seral type. This was indicative of disrupted soils and higher light intensities on these cutover sites. The "Ceanothus-Arbutus" seral type occupied south-facing upper slopes heavily impacted by slash burning. The remnant Arbutus menziesii in the tree layer indicated this type's relationship to the xeric old-growth forest association. The most prominent structural component of this seral type was the dense shrub layer dominated by Ceanothus thrysiflorus. This species and Arctostaphylos columbiana, another shrub in the "Ceanothus-Arbutus" type, were chaparral species scarcely represented in the old-growth forests. Zinke (1977) described post-logging succession in xeric redwood forests, and he also noted that succession is shunted to Ceanothus chaparral by fire.