Table Of Content

PLANT ASSOCIATIONS OF CASTLE CRAGS STATE PARK, SHASTA COUNTY, CALIFORNIA John D. Stuart, Tom Worley, and Ann C. Buell Department of Forestry, Humboldt State University, CA Areata, 95521 Abstract Vegetation types in Castle Crags State Park were classifiedusing TWINSPAN into 8 series and 15 associations from 237 plots. The series, in decreasing order ofabun- dance, were Pseudotsuga menziesii, Quercus kelloggii, Alnus rhombifolia, Quercus chrysolepis, Ceanothus cuneatus, Bromus tectorum, Quercus garryana, and Juncus patens. In addition, 2 types occurring in chronically disturbed habitats (powerline corridor and mowed meadow) were described. Each vegetation type was further described using physiography, soil characteristics, species diversity, tree density by height class, and basal area of tree species. Successional interpretations were made for plant associations. Introduction Castle Crags State Park has several important botanical and eco- logical roles: a) it serves as a landscape bridge and corridor across the Sacramento River between the southern Cascades and the Klam- ath Mountains; b) it is host to the rare Ivisia longibracteata', and c) it preserves low-elevation montane forest, shrub, riparian, and herbaceous communities. Adjacent lands along the Sacramento River corridor are heavily disturbed from logging, mining, development, and construction and maintenance of transportation routes. Plant associations in this ecologically diverse and vital park have not been adequately classified and described. In this paper we present a classification and description of the park's existing plant associations, ascribe their distribution to physiographic or soil characteristics, pro- vide data for the future management of these communities, and as- sess the successional potential of the forests. The park is recovering from a series of disturbances dating from the late nineteenth and early twentieth centuries (Beat and Gut 1981). Mining, logging, and wildfires consumed nearly all of the old-growth forest, leaving behind early successional communities that were apparently characterized by sprouting hardwoods and shrubs, shrub seedlings, early serai herbaceous species, and occa- sional surviving conifers. Conifer seedlings and some late serai shrub and herbaceous species were undoubtedly present in these early serai communities. Today, after approximately 125 years of succession, the plant communities in Castle Crags State Park are relatively well defined. Madrono, Vol. 43, No. 2, pp. 273-291, 1996 I 274 MADRONO [Vol. 43 Much of the park is dominated by Douglas-fir and mixed conifer forest with meadows and shrublands situated on alluvial, granitic, or serpentinitic parent materials in Castle Creek Valley. Interspersed within the conifer forest are stands of California black oak. Methods Study Area Castle Crags State Park encompasses 1681 ha both east and west of the Sacramento River in the Klamath Mountains Province, Cal- ifornia (approximately 41°9'N, 122°20'W). Elevations range from approximately 640 m to 1220 m. Slopes typically exceed 40% and have soils derived from alluvium, serpentine, granitic rock, and greenstone (Mallory and Powell 1978). The Crags lying to the north- west of the state park are what remains of a granitic pluton that has been extensively eroded and glaciated. The McCloud weather station, situated 17 km to the northeast, receives an average of 130 cm precipitation per year. Mean high temperatures are 8°C in winter and 30°C in summer. Mean low temperatures are —7°C in winter and 8.5°C in summer (National Cli- matic Data Center 1990). Forested communities conform to the So—ciety of American For- esters cover types ofPacific ponderosa pine Douglas-fir, California black oak, and the Sierra Nevada mixed conifer (Eyre 1980). Study Site Selection We used 1:12,000 color stereo aerial photography in conjunction with a U.S.G.S. orthophoto of the Dunsmuir Quadrangle to identify areas (polygons) with apparently homogeneous vegetation composition and structure. Two hundred and thirty polygons were delim- ited. Ofthese we sampled all polygons that appeared to have unique structural or species composition characteristics. Forty-eight polygons were selected for ground sampling; the remaining polygons were indistinguishable from those selected for sampling. We established approximately 5 plots per selected polygon for a total of 237 plots. Field Measurements Plots were subjectively placed and centrally replicated to best characterize vegetation in each polygon (Mueller-Dombois and El- lenberg 1974). No ecotone plots were established. Forested plots were 0.1 ha in area, shrub-dominated plots were 0.05 ha, and herb- dominated plots 0.01 ha. Vascular plant cover was ocularly esti- mated to the nearest 5% for canopy trees, 1% for small trees and shrubs, and 0.01% for herbs. On forested plots, we recorded tree 1996] STUART ET AL.: CASTLE CRAGS PLANT ASSOCIATIONS 275 basal area and tree density by species for the 0-3 m, 3-10 m, and >10 m height classes. Data Analysis Species percent cover was averaged over all plots within a polygon to avoid pseudoreplication. Subsequent statistical analyses were based on the 48 polygons, and not on the 237 individual plots. Rare species, occurring in less than 3 plots, were deleted from the data matrix prior to classification because they act like outliers in multivariate classification algorithms, adding variability that ob- scures central tendencies (Gauch 1982). We used TWINSPAN (two-way indicator species analysis) (Hill 1979) to classify samples and species into similar groups. Pseudo- species (a type of differential species based on predetermined abun- dance levels) cut levels were 0, 2, 5, 10, 20, 35, and 50. Pseudo- species cut levels of 20, 35, and 50 received twice the weighting of the others in order to favor dominants. Plant associations were described using percent cover and constancy. Species richness was determined for each classified plant association. Richness estimates were based on the cumulative number of species in 5 randomly selected plots from each plant association. We did this to avoid sampling area bias and to ensure comparability between plant associations (Magurran 1988). Successional Inferences Successional inferences were based on the life history attributes of potential canopy dominants: mode of persistence, species tolerance of understory conditions, fire tolerance, longevity, stature, and reproductive success in the understory. Results and Discussion Classification and Description of Plant Associations We interpreted 8 series and 15 plant associations from the TWIN- SPAN output (Fig. 1). Nine associations were forested (Table 1); 3 were shrublands (Table 2); and 3 were herb-dominated (Table 3). In addition, two communities occurred in chronically disturbed sites: beneath a powerline, and in a mowed meadow. Level zero in the dendrogram (Fig. 1) divided forested from unforested plots. Level 1 separated plots according to life form, and levels 2 and 3 according to dominant overstory species, physiognomy, or parent material. Level 4 further divided plots according to physiography or parent material. Physiographic and soil characteristics of each plant association 276 MADRONO [Vol. 43 ^ ELEL-DACA ° JUPA-POPR-LUCO BRTE-POPR QUGA/BRTE 1^ CECU/VUBR-BRTE CECU-RHCA-ARPA CECU/ERLA NAPU-SABI QUKE/NAPU QUKE/TODI PSME-CADE/NAPU PSME-CADE-QUKE/NAPU o Zo o3 PSME/CONU-TODI ii PSME/LIDEE ALRH-PSME-CADE/RUDI ALRH-ACMA/ACCI QUCH-ARVI-ARPA 1996] STUART ET AL.: CASTLE CRAGS PLANT ASSOCIATIONS 277 are presented in Table 4. Basal area, species richness, and tree density for each plant association are presented in Table 5. Pseudotsuga menziesii Series All plant associations within the Pseudotsuga menziesii series had moderately deep soils, gravely loam soil textures, and were found on greenstone, alluvium, or basaltic rock. Differences betweenPseu- % % dotsuga menziesii associations were seen in aspect, slope, and rock and bare soil. 1. The Pseudotsuga menziesii!Lithocarpus densiflorus var. echinoides (PSME/LIDEE) association was dominated by Pseu- dotsuga menziesii with 40% cover and 100% constancy. The shrub form of Lithocarpus densiflorus, var. echinoides, was dominant in the understory with 24% cover and 94% constancy. Species richness was 27 and there were 39.7 mVha ofbasal area. This association typically occurred on north aspects. Slopes averaged 27%. 2. Pseudotsuga menziesii was predominant in the Pseudotsuga menziesii/Cornus nuttallii-Toxicodendron diversilobum (PSME/CONU-TODI) association with 51% cover and 100% constancy. Calocedrus decurrens and Quercus kelloggii were minor canopy associates. Cornus nuttallii and Toxicodendron diversilobum had the greatest cover and constancy in the un- = derstory. This association had the most biomass (basal area 57.2 m^/ha) and the least exposed rock and bare soil. No other Fig. L Twinspan dendrogram with interpretative annotations. Numbers beneath linesrepresentthenumberofplotspriortodivision. Numbersinboxesarethenumber ofplots in each classified plant association. Labels beneath numberedboxes areplant association acronyms. QUCH/ARVI-ARPA = Quercus garryanalArctostaphylos viscida-Arctostaphylospatula, ALRH-ACMA/ACCI = Alnus rhombifolia-Acerma- crophyllumlAcercircinatum, ALRH-PSME-CADE/RUDI =Alnusrhombifolia-Pseu- dotsuga menziesii-Calocedrus decurrens/Rubus discolor, PSME/LIDEE = Pseudo- tsugamenziesiilLithocarpusdensiflorusvar. echinoides, PSME/CONU-TODI = Pseu- dotsuga menziesii!Cornus nuttallii-Toxicodendron diversilobum, PSME-CADE- QUKE/NAPU = Pseudotsuga menziesii-Calocedrus decurrens-Quercus kelloggii/ Nassellapulchra, PSME-CADE/NAPU = Pseudotsuga menziesii-Calocedrusdecur- renslNassellapulchra, QUKE/TODI = Quercus kelloggii/Toxicodendron diversilobum, QUKE/NAPU = Quercus kelloggiilNassella pulchra, NAPU-SABI = Nassella pulchra-Sanicula bipinnatiifida, CECU/ERLA = Ceanothus cuneatusiEriophyllum lanatum, CECU-RHCA-ARPA = Ceanothus cuneatus-Rhamnus californica—Arctos- taphylospatula, CECU/VUBR-BRTE = CeanothuscuneatuslVulpiabromoides-Bro- mus tectorum, QUGA/BRTE = QuercusgarryanalBromustectorum, BRTE-POPR = Bromus tectorum-Poa pratensis, JUPA-POPR-LUCO = Juncuspatens-Poapraten- sis-Luzula comosa, ELEL-DACA = Elymus elymoides-Danthonia californica. MADRONO 278 [Vol. 43 Table 1. Average cover and constancy for forest associations. Species re- ported are those with >50% constancy. ' See Figure 1 for plant association acronyms. 2COV = average cover (%). ^CON = constancy (%). ALRH- QUCH/ARVI- ALRH- PSME- ARPA ACMA/ACCI CADE/RUDI Plant association' _ (5) (5) (8) (number ofplots) COV^ CON^ COV CON COV CON Arctostciphylospotulci OZ'JX 1on Arctostaphylos viscida 'XA IUU Qucrcus chrysolcpis 1Q i1nUnU z j'^Un Smilax califomica <1 orv Rubus ursinus <1 oorU\ Rhododendron occidentQle I sonU RhcimnuspuKshicinci 7 1lnUnU Montiaperfoliata 1IrU\cU\ LiHum pardalinum <1 ooUn DicentrofofmosQ <^J AOnU Athyriumfilix-fcminci <^J sOnU Aralia californica L 1InUnU Adiantumpedatum <1 oO Alnus rhonibifolia /•4i-nU 1InUnU 1VJc 1InUnU Acer circinatum 1o 1InUnU 1lnU ADiJ Trientcilis lotifolio ^1 osUn Z /J Pteridium aquilinum <1 oU Acer macrophyllurn oznu oonU Q A1 Symphoricarpos albus Corylus cornuta 1 AonU OQ A1 Rosa gymnocarpa 3 jU Rubus discolor zz Rubus leucodermis o jU Fraxinus latifolia 1irUk OQCO Ecjuisetum laevigatum D'^Un Danthoniapilosa / jU Carex subfusca 4 63 Agropyron spicatum 5 75 Lithrocarpus densiflorus var. echinoides Cornus nuttallii Pseudotsuga menziesii 10 80 13 75 Calocedrus decurrens 1 80 12 88 Quercus kelloggii 6 50 Pinus lambertiana Pinusponderosa 4 63 Toxicodendron diversilobum Viola lobata Luzula comosa Galium aparine Clarkia gracilis Carex cusickii Nassellapulchra Rhamnus californica Sanicula bipinnatiifida 1996] STUART ET AL.: CASTLE CRAGS PLANT ASSOCIATIONS 279 Table I. Extended. PSME/ PSME- PSME- PSME/ CONU- QUKE/ CADE- QUKE/ CADE/ LIDEE TODI TODI QUKE/NAP NAPU NAPU (32) (38) (31) (30) (6) (6) COV CON COV CON COV CON COV CON COV CON COV CON 2 67 1 33 1 56 3 63 9 66 1 50 4 63 2 72 24 94 2 50 7 75 5 50 40 100 51 100 10 94 24 100 50 22 100 6 75 11 87 6 65 12 97 21 100 11 84 8 74 32 94 15 83 63 100 4 56 4 50 5 59 13 87 6 50 67 1 61 4 68 <1 50 1 50 <1 52 <1 50 <1 50 6 67 11 67 18 67 6 50 10 83 4 67 MADRONO 280 [Vol. 43 Table 1. Continued. ALRH- QUCH/ARVI- ALRH- PSME- ARPA ACMA/ACCI CADE/RUDI (5) (5) (8) Plant association' (number ofplots) COV2 CON^ COV CON COV CON Pinusjejfreyi Phlox speciosa Monardella odoratissima Lomatium macrocarpum Festuca californica Epilobium brachycarpum Dichelostemma mulitflorum Cheilanthes gracilis Ceanothus cuneatus Agoseris retrorsa association had as many trees taller than 10 m (208/ha). Unlike the north-loving PSME/LIDEE, PSME/CONU-TODI was found on all aspects. 3. The Pseudotsuga menziesii-Calocedrus decurrens-Quercus kelloggiilNassella pulchra (PSME-CADE-QUKE/NAPU) association differed from the first two Pseudotsuga menziesii associations in that it had several co-dominant tree species in the canopy. Pseudotsuga menziesii was still dominant with 24% cover (100% constancy), followed by Calocedrus decurrens with 12% cover (97% constancy), and then by Quercus kelloggii with 15% cover (83% constancy). Nassella pulchra had the greatest understory cover. This type had more rock and bare soil (6%) than the first two Pseudotsuga menziesii types and was found on all aspects. It had the second highest total tree density (756 trees/ha), an average basal area of 35.8 m^/ha, and species richness of 29. 4. The Pseudotsuga menziesii-Calocedrus decurrens/Nassella pulchra (PSME-CADE/NAPU) association resembled PSME- CADE-QUKE/NAPU in dominant species composition, but was found at higher elevations, western aspects, and on steeper (52% versus 26%), rockier slopes (24%). This association had less biomass, with only 25.3 m^/ha of basal area and 46 trees/ ha taller than 10 m. Pinus jejfreyi shared the sparse canopy with 9% cover and 67% constancy. Vegetation dynamics. Pseudotsuga menziesii and Calocedrus decurrens were the most abundant reproducing species in PSME/ CONU-TODI, PSME-CADE-QUKE/NAPU, AND PSME-CADE/ NAPU. Lithocarpus densiflorus and Pseudotsuga menziesii were 1996] STUART ET AL.: CASTLE CRAGS PLANT ASSOCIATIONS 281 Table 1. Extended. Continued. PSME/ PSME- PSME- PSME/ CONU- QUKE/ CADE- QUKE/ CADE/ LIDEE TODI TODI QUKE/NAP NAPU NAPU (32) (38) (31) (30) (6) (6) COV CON COV CON COV CON COV CON COV CON COV CON 8 40 9 67 <1 50 2 50 1 67 10 50 <1 50 1 67 1 50 9 67 <1 50 Table 2. Average Cover and Constancy for Shrub-Dominated Associations. Speciesreportedarethosewith >50% constancy. ' SeeFigure 1 forplantassociation acronyms. ^COV = average cover (%). ^CON = constancy (%). CECU- CECU/ RHCA- CECU/ VUBR- NAPU- ARPA ERLA BRTE SABI (14) (10) (5) (7) Plant association^ (number ofplots) Cov^ Con"* Cov Con Cov Con Cov Con Arctostaphylospatula 10 79 Hypericumperforatum 1 50 Pinusjejfreyi 2 79 Quercus kelloggii 9 64 Rhamnus californica 13 86 Calocedrus decurrens 2 60 Eriophyllum lanatum 2 90 Monardella odoratissima 2 70 Elymus elymoides 2 60 Ceanothus cuneatus 39 100 22 90 10 60 Bromus mollis 1 80 Bromus tectorum 4 80 Draba vema <1 80 Epilobium brachycarpum <1 100 Gnaphalium sp. 1 80 Rumex acetosella 1 60 Scleranthus annuus 2 80 Vulpia bromoides 6 80 Dichelostemma mulitflorum 1 60 1 57 Nassellapulchra 100 50 22 100 Agoseris retrorsa 71 Eriogonum umbellatum 71 Sanicula bipinnatiifida 57 100 2 282 MADRONO [Vol. 43 Table 3. Average Cover and Constancy for Herb-Dominated Associations. Speciesreportedarethosewith >50%constancy. ' SeeFigure 1 forplantassociation acronyms. ^COV = average cover (%). ^CON = constancy (%). JUPA- POPR- BRTE- QUGA/ ELEL- LUCO POPR BRTE DACA (10) (5) (20) (5) Plant association' (number ofplots) COV2 CON^ COV CON COV CON COV CON LiUzuIg cotiiosci 1 JuncHSpatens '2.'2, 80 Rubus discolor 5 60 Ncissdlapulchfo 1 Poapratcnsis 12 11^J ou Dactylis glomerata 1 60 Vulpia bromoides 3 60 Aria caryophyllea 2 55 Bromus tectorum 15 100 12 90 Clarkia gracilis <1 55 Lupinus argenteus 1 55 Quercus garryana 10 50 Rumex acetosella 1 60 1 50 Scleranthus annuus 2 50 Danthonia californica 18 80 Lotus humistratus 4 100 Plantago lanceolata 10 80 Ranunculus occidentalis <1 60 Elymus elymoides 21 80 Tragopogon dubius <1 60 Trifolium macraei 3 60 Triteleia hyacinthina 7 60 Verbascum blattaria <1 60 most abundant in PSME/LIDEE. These species were more shade tolerant than other associated tree species (FrankUn and Dryness 1973; Powers and Oliver 1990) and, in the absence of fire or other disturbance, should become self-perpetuating (McDonald 1980; Sawyer 1980; Hermann and Lavender 1990). With fire, Pinus ponderosa, Quercus kelloggii, and Pinus lambertiana should remain within the mix of overstory dominants and co-dominants (Franklin and Dryness 1973, and Sawyer et al. 1977). Relation to other vegetation types. PSME/CONU-TODI, PSME- CADE-QUKE/NAPU, and PSME-CADE/NAPU share elements of the warm, dry component of the mixed conifer forests found in southwestern Oregon (Franklin and Dryness 1973); of the mesic, low elevation type in the Sierra Nevada/Cascade mixed conifer forest (Rundel et al. 1977; Holland 1986); and ofthe Pacific ponderosa pine-Douglas-fir cover type (McDonald 1980). A Pseudotsuga menziesii-Calocedrus decurrens type has been described in southern