245
Trifolium thompsonii Stand Conditions Following a Wildfire Event in the Entiat Mountains of Central Washington
George Scberer1, Richard Everett• and Ben Zamora2
'USDA-Forest Science lab, 113:3 Western Ave., Wenatchee, WA 98801 Tel. (.509) 662-4:JJ.5; Far (.509) 664-2742
2 Department ofNatural Resource Sciences, Washington StaJe University, Pullman, WA
Abstract. TrifoUum thbmpsonii. (Thompson's clover) is of the Columbia River, about 6 km southeast of the listed as a threatened legume of the Eastern Cascades Thompson's Clover Research Narura! Area. where wildfire is lhedootinantdisturbanceagenL Previous These new sightings snggesi even wider range poten researchsuggesttdfiresupressionadverselyimpactspopo tials and pezhaps a different ecological amplitude than pre lations of!his species. Recent fires in partS ofThompson's clover range indicate that Thompson• s clover responds viously described {Kennison and Taylor 1919). Toe pur pose of this study was to clarify the habitat characieciscics favorable to fire distwbance. Four areas of Thompson's of Tri[olium thompsollii and seek evidence which might clover habitat were sampled for vegetacion composicion indicaie the successiooal role of this species among the and structure. Thompson's clover stands we,e evaluated different plant communities where it is found. 1be ob for species density and morphological expression. In areas jectives of this.research were: 1. to docwnent the vegeia ofrecentfire, Thompson'scloverindividualsw~signifi tive communities and sue factors associated with Trifo cantly taller (means of range 32-25 cm) and had gn,arer li.um thompsonii.; and 2. to compare moqihological fea pwnbe:s offlowerbeads (means of range 4 to 1 pee plant) tures and plant density of Tri[olium thbmpsonii popula than on sites wilh no =nt fue activity, suggesting a tions among these vegetative communities. Thompson's clover response to fire caused resource release. This infonnation combined with a species density· ranging 0.6 to 4.8 plants m·2 suggests that Thompson's Methods clover appea,:s to achieve opcimum stand conditions on sires wbtte perodic grass-shrub eliminaling fires occur. Study Area Re-establishment of historical fire effeas may be a key to improved management oflandscapes in fire.respon which Climate of the area is typical of eastem flank of the sive ~-pecies such as Thompson's clover are components. Cascade range. Swnmers aie warm and
since 1930 (WNF 1990, Canfield 1977). The most recent rain fearures were tabulated for comparison by plant com tireoccmrenc:ewasJuly 1983. I.andscapeswithln thisaiea munity (Table 2). DataonclensityofTrifoliwnrhompsonii were grazed by domestic sheep and caUle until 1965 and measwemenrs of morphoiogical f~ we.e tabu (Tiedemannet.al.1977). NolmownfireshaveOCCUI7Cdon lated by plant community and analysis of variance was the Douglas =mty study sites but !he area is presently used LO 1CSt for differences among the community types.. under seasonal grazing impact by domestic caale and portions of the area are cultivated for wheat crops (Gamon and Sprague 1988, E. Gotzwile. 1993, pm. comm.). Resulis F~ld Sampling Trifolium thompsonii occurred in three broad yegeta tive settings: forest patches, SLeppe-like parklands, and Sampling sites were selected by reviewing maps of semiarid shrub Stq)pe. The forest patches consisted of a known Trifolium rhompsonii locations from the multilayered community ofirees, shrubs, grasses and forbs Wenatchee National Forest Thompson's Clove. Manage dominated by Pseudotsuga menziesii occurring on mesic ment Guideline (Gamon and Sprague 1988). Criteria for IOpographic·seuings such as gullies and northerly slopes. sample site selection was: The appearance of distinct Steppe paddands were grass dominated landscapes with patches of Trifoiium thompsonii in the overall mosaic of occasional mature Pi.nus ponderosa . individuals. The vegeiation coverage and patch size greater than 50 M di shrub steppe was the extensive drought LOlerant tall sbrub ameter. bunchgrass (,Artemisia trkkn!alal Agropyron spiCIJlun:,) Vegetali.ve sampling was done by systematically plac vegelative associations typical of the central Columbia ing a 50cm x 20cm sampling frame in a4 X 5,20 frame basin. matrix at three positions (0, 30, SOM) along SO mew tape Wilhin these lhree settings four plant communities transect running trough Trijolium thompsonii patches. The were identified using cluster analysis of species compo polygonal cover class of each species eocoonreted in the sition and cover values wilhin stands. These four com P,lot frame was recorded (Daubenmire 1959). Taxonomic munity types probably represent various successional nomenclature follows Hitchcoclc and Cronquist (1973). stages of four plant associations as descn"bed by Franklin At the same time, soil surface condition was recorded at .and Dymess (1973) and Daubenmire (1970). They are 10 _pointS on Ille plot frame as either litter, bate ground, Psew:w1suga menz~sii/Calmagrosris rubescens (PSMEI plant basal area, crypwgamic crust, wood pieces, or rocks. CARU) of the forest setting, Pillus ponduosa/Agropy At each sampled community !he density of Trifoliwn ro" spicatum (PIPO/AGSP) and Artemisia rriderual.tJ rhompsor.ii was measured by counting individuals within · vaseyana/Agropyron spicaium (ARVA/AGSP) of I.be Len ll'vf plot frames in !he same 4 X 5 vegetation sam steppe setting and Artemis/a 1ridenta1a-1riden1ataJA,gr<> pling matrix. Frol)l each sampled community 10 ID 30 pyron spicarum (ARTR/ AGSP) of !he shrub Sleppe zone. Trifolium Ihompsonii individuals were randomly selected Physiognomic differences between the community types and measured for height, leaf number, shoot number, we.e distinctive and easily recognizable. These four com flowerhead number and root crown diameter. munity types are descn"bed below wilh refe.ence LO veg At each sampled community the following sire fac. etativecom_position and site fac!Or dala in Tables l and 2. tors we.e recorded: elevation, slope, aspect, slope posi tion, and a microterrain feature (flat, convex, concave). Planr Commwliry Descriptions The soil depth was measured and a soil sample ofthe up per 20cm was collected at each sample matrix and ana The PSME/CARU community occurred on noi:th fac lyzed for bulk density and% soil moisture. Toe soils dala iog steep slopes (40-60 %) at mid LO upper elevations (> we.e compared and suppleme.nied wilh SCS Soil survey HMO m) in Chelan county. This was the weucst and laboraLOry information available on the Study area (Beiler coolest environment wilh estimated annual precipitation 1969, 1978). ranging 508 LO 1016 mm and estimated annual iempera ture ofS-6 "C. Pseudo1suga menziesii was the overstory Anaiysis -dominant with Pinus ponderosa a codom.inant in some .sample units. Spiraea betulifolia and Pensremon Mean coverage and frequency values were calculat£d jnuicosus were prominent shrub layer componentS. 1be for all species in each sample unit and used IO consuuct a grass, Calmagrosris rubescens was a significant CCHlomi species list ofcover and a table of plant communities with nant of this communjty type. Balsamorhiza sagi1ra1a, species cove. and coosiancy (Table I). Dendograms pro AcMllia millefo/ium and Trijoliwn 1/wmpsonii were the duced by cluster analysis (Wilkinson 1992) we.e used IO most prominent forb species. Other imponant hcrbacrous identify relationships among Trifolium 1/rompsonii sam =ics were Festuca idalroensis, JJrornus tectorwn, Poa pling clusters based on species composition and cover. s~cundi, Collensia parvif/ora, Cryprhantha a/finis, Site fac!Or data on soil moisture, surface cover and ter- Epi/obium paniculara. In Ibis community fire recently • Trifolium tltompsclfii Stand Conditions FoUowing Wildfire • 247
Tablcl.Conmooymdmean~crofspecicoinfourTrifo/iJJmtliomp,,,niicommunityiypes.Tr-=%coverlessthanO.l. N=nnmbcr of sample unit&. Cot,,, """""""Y· = % canopy covcr. PSME P.ceo ARTR ARVA N=Z N=7 N"9 N=J Ca, Cov Ca, Cov ,... Ca, Cov Con Cov PWU p<>ltduMO 100 29.0 p p 0 0.0 Puw:louwga nvlfZiaii 100 0 0.0 32.0 0 0 0 • 0.0 Shrub 0 0.0 Arte,ni,io tr;,J,AJaJa-tr 0 0.0 14 Tr 77 8.S 0 o.o Nt~t4 trid,usl4J4-\t2 0 0.0 0 0.0 tripartiJa 0 0.0 67 1.2 ,1,..,,,.,.. 0 0.0 0 0.0 11 1.0 67 14.0 ClllNldut: wlutima so 6.0 0 0.0 . 0 o.o 0 0.0 C~,a vueidij!ona 0 0.0 0 0.0 II 0.2 0 0.0 Erigero1t luu.ari, 0 0.0 14 0.7 0 0.0 0 0.0 ~-1,uad.e;,iJu 0 0.0 43 2.2 ss 2.6 100 8.S fjiogo,wm ,uy~""" 0 0.0 14 0.4 0 .o.o 0 0.0 Erit>ga,aun .ttric:bM 0 0.0 0 0.0 22 0.2 0 0.0 1/aplcpopp,u ,u,.cphyl/,u 0 0.0 14 Tr ss 1.2 0 0.0 Pu,.sUlf'IOn /n,tlic.onu 100 ll.S 28 o.s 0 0.0 0 0.0 P/Joz io,,gifolio 100 1.8 57 1.7 100 13 34 0.8 Pwhhi.o lrkkn1a1a 0 0.0 0 0.0 66 4.9 0 0.0 Riba CU~wrt so 0.4 0 0.0 22 0.6 0 0.0 Spi,,ua bd"1i[Q/;., 100 12.7 0 0.0 0 0.0 0 0.0 s~an- 50 1.6 14 0.4 0 0.0 Vacc.ini.ian txtiducla.k 0 0.0 50 2.3 0 0.0 0 0.0 Crus 0 0.0 Agropyr Table l. Rm,g-,ofvalues fouiie f:iaors in .Usamplc units of fc1lr , ...... :iy i:ypes ,woci,w,d with oCTrifoii,,m tmmpsonu. Anwuuzrui ro~a, Acheilia millefoliJlln were tbe major forb components. Annoals were net prominet1t in this PSME PIPO ARTR Ml.VA community, except for a Bromus ~ctorum coverage of 4 Site V.vibl,. %, tbe highest of this study. Charred buochgrass clumps EIO'llllioo (m} 1115-1060 810-399 890-1096 816-11171 Slq,e('il\lv 40-50 30-60 15-,l'O 30-40 and shrub skelerons indicated recent (1988) fire disiur A,pea( 3(,() 360-235 045-270 3(,().235 bance. P=;ptt (mm) 508-1016 ~ ~ 31)4-406 In tbe ARTR/ AGSP community cype !he overstory is Temp(C) 5-6 8-10 8 • 10 8· 10 dominated by the tall shrubs Anemisia tridentata Soil Surface liuu(%) 44-63 40-60 19-69 37-$ lrideillata and Purshia lridemata, wiih mean coverage Ba,c Grouod (%) II 8-34 13-SS 8-40 valoes of 9% and 5%. The herbaceous dominants were Rocle(%) I 0 0.26 0 represenied by Agropyron spicatwn, F estuca idalwensis, Wood(%) 4-11 I 0.10 0 Baul Arca (%) 18 11-29 7.z, 71.71 and Poa secunda with op to 44% mean cover. 11lis com Ctyplogaau ('ii\) 1-4 1-4. 5 0.1.S munity was rich in forbs with nearly 20 recorded The Solllfllllt most important were Trifolium thompsonii,Balsamorhirza ~. 35-48 25.55 24-75 48-56 sagiltala, AJJium acuminallun, Agoseris grandiflora, (31«) 1.22-1.40 1.11).1.30 1.05-1.30 1.21-1.29 Loma1ium nudicale, Achellui millefolium.. and Lupi.mu Available W~ sericeus. There was no evidence of recent fire in this com (an-an) 0.3~.43 0.43--0.50 0.15--0Sl 0.30--0.35 6.5 6.4 7.1 6.4 municy. However, some forbs and bunch grasses appeared ~anic Manet to have been grazed. %) 2.5. 4.5 2.8-6.4 1.5-4.5 2.6-3.0 CEC(meq) 10-20 9. 15 8·18 8.4-12.0 Community Structure Cea= 10.40 10.20 5-45 10.20 Fr>gmcnu (%) S,md($) 7Ul3 65-'9 33,,59 67- 68 Table 3 shows the SlruCIW'C of the four communities Silt ('ii\) 15-21 19-31 36-53 25-32 C.y('l\) 1-3 1-4 Z. 13 1- 6 by life form. · Grasses and forbs were prominent in ille communities of PIPO, ARTR, and ARV A with forbs nearly equal in presence but more variable in cover and :species composition among ille groups. Tree cover was clearly distinctive in th1e PSME/ CARU community but almost of no importance in !he ihreeoihers. Noteworthy, was a low value of shrub cover but high value of forb destroyed lhe crown foliage of nearly all Pseud01suga cover in PTPO/ AGSP, suggesting recent disturbance in menziesii and Pinus ponderosa. lhose layers; lilcely by fires, since rree boles were charred Thedominantvegciation in PIPO/ AGSPwasanAgro but crown foliage was most! y unharmed. The structure [7Jron spica1um- Fes1uca idahoensis grass understory data shows ·a decline in shrub and tree components and (Table 1). Pim,s ponderosa was present as solitary ma increase in grass and forb components from the mesie ture individuals in mesic topographic settings wiill cover forests to the arid shrub steppe. rarely exceeding 2 % on ille entire study area of the Re ~h NaruraJ. Arca. Only a few low shrubs, Eriogonum Plant density- Trifolium thompsonii heracleoides,Phlox longifolia, Hap/opappus stenophyflus, were present wiih less than 5% overall cover. No tall Density of Trifolium thompsonii ranged from O to 16 shrubs were recorded. The most prominent species in the plants-m-' on 258 plotS from within ille four community forb component were A1Ue1111tJTia rosea, Balsomorhiza types. There were significant differences in the mean sagittata, Trifolium thompsonii and A chi/lea millefolium. density of Trifolium thompsonii among all ille commu The annuals Coiiensui parviflora, Epilobium paniculaLa, nity types. Generally. Trifolium thompsonii densities were Crypthanlha affinis were nearly ubiquitous in the samples greater in the steppe-like settings than those in !he PSME but with small cover values. Fire disturbance was gener setting and specifically they were highest in ille ARTR ally in tbe grass shrub understory with no tree crown dam shrub- Steppe of Douglas County where individual plot age. frame samples contained as many as 16 plants m•. By The AAVA/ AGSP was located on tbe Thompson's contras!, plot samples in the forested (PSME) had high Clove:r Research Nalllial Area (Figure !) and was adja vegetative coverage in trees. shallow soil depths, and cent to PIPO/ AGSP. Individuals ofArlemisia 1ridFestuca idahoensis cause of its low density value of 0.6 plants meter. rangjngin mean cover from4-27 % were vegetativedomi nant.s. Trifolium 1hompsonii, Balsanwrhiza sagittata, - Trifolium thompsonii Stand Conditions Following Wildf:u:e - 249 Columbia ltive;r Douglas County ,Douglas County- Sites Chelan County 0 I 2 3 0 2 3 4 ";;~ec mJ River ~ Pct$Cnt sund.G illCQfsm !~ccB • Fi.ra during 19S0'.2. ~ El Fires during 195()1s III] F'il'e.$ during t960"s [D Fitts during 1970's Nort.ht ~ Fires durinc 19SO's Fl2ure l. Present l:nowns!Jlnd.s.o(TrifoUum thompsonii in Ch.clan and Douglas countics_Washingron wuhknownareas of fire disturbance in 'l"rijolium thompsonii range since 1930 (ad.aptecl !rom Canfield 1977 and WcruuchccNational Forest 1990). 250 Scherer, G., Eve."'ett, R. and Zunora, B. Table 3. Proportion cfa,va by lifcfomtin follfcommwiity types 3SS .------·-- - Density Height PSME ARVA PIPO ARTR-·· PSME ARVA ARTR Hi Med Lo No I Hi Med No Fire Fite. fire fire Fire Fire r.ro '------·------___ _j 'I -· ·--·--- - ··-·- -·-·. ~---- ' -I 4.5 1 • 1' ~ Root Diameter Flowerheads E § c,, j T i 3 •" " 2$ E "' ! ? ": 1.6 &l :,: I • o,,' • i. ~ < 0 DJ1 PSME _D.. ARVA PIPO ARTR PSME ARVA PIPO ARTR HJ Med Lo No HI Med Lo N<> Fire Fire ffre n,. Fire! Fire flre fire C-ommunity type abbreviations are: PSME=- PSeGJdotsuga menziesii/ Calma-grostis rubesoens PIPO= Por.derosa pine/Agropyron spic-aium; ARTR=Artemisla bidentata/ Agropyron splcat>, m ARVA= Attemisia tridentata/ Agropyroc, spicatum ~re 2. Trifoliwn thompsonii plant density and morphological fearures in four community types. Vertical b.,.. 95% confidence znterVals. = sc.raJ stage in this landscape. If this community is seral, crowns by the 1988 wildfire. This lacl::oftreecrown per then the vigor of Trifoliwn thompsonli individuals sampled mitted considerable light penetration 10 the understory and here may support the proposition that this species is a se surface. With .these openings and reduced competition, ral opportunist in response to disturbance. light tolerant heliophytes such as Trifolium thompsonii It was in the PSME community, however, that Trifo may have opportunity to develop particularly vigorous liwn thompsollii appeared to be most opportunistic. lndi indh'iduals. Of course, this strategy depends on a taxoo vidaals here were tallest and had the greatest floral-~ being present in lhe community, either in the soil seed potential. This PSME community resembled the reserves or in a reduced vegetative form prior to the event Pseudolsuga menziesii I Symphoricarpus a/bus habitat of disturbance. The low density value of T~ifo/ium type as descnoed by Daubenmire and Daubenmire (1968), thompsonii recorded in the PSME community suggestS largely because of the presence of the shrubs Spirea. !hat its presence was probably limited in the /Jeru/ifolia, Ceanorhus velUD/lus and Vaccinium'spp. This predisrurbance phase. comparison, though. is confounded by the presence of The ARTR community of this study appeared to have Caimagrostis rubesce.ns tha! was sampled on slightly drier !he least disturbance, certainly from fires. A nearby pans (ridges) of this community. The shrubs were located rancher grazes cattle in the area_ Grazing was observed in slightly wetter gully locations. It is important to note on the bunch grasses and on Trifoliwn rhompsonii. There that the Pseudotsuga menziesii was burned in the tree was no indication of receatfires, indeed with such exten- 252 Schete.r, G., Everett, R. and Zamora, B. sive coverage of fire in10lcrant Artemisia 1ridenta1a Franklin.. I.P. and C.T. [)ymcso. 1973. Naruru VegeLUion of triden1ata, fire probably has not been a disturbance fac Oregon· and Washington. OSU Press, Cor,allis, OR. 452 tor for decades (Table l). This community is thoroughly pages. described in Daubenmire ( 1970) as a major climax asso Gamon. J. andN. Sprague.1988. Babitatmanag References Agee, I.K. 1993. Fire Ecology of Proceedings of: courtesy Organized by: TIONAL ASSOCUQ'ION & Washington OF lANDRRE Foundation for the Environment Proceedings: First Conference on Fire Effects on Rare and Endangered Species and Habitats Coeur d'Alene, Idaho November, 1995 Dr. Jason M. Greenlee Editor Conference organizer: .-.,;;;;;;,...TIIINAI. ASSOCllmON OF LA.NORRI! Co-sponsors: International Association of Wildland Fire The Nature Conservancy Wildlife Forever U.S. Fish and Wildlife Service, NIFC Washington Foundation for the Environment Abundant Life Seed Foundation Dr. Jon Keeley Montana Prescribed Fire Services, Inc. · Colorado Natural Heritage Program Environmental & Botanical Consultants USDA Forest Service (Reg-6) Proceedings courtesy of: -- · -:._~~- & Washington Foundation ,. ...:. ... ---~. . .. ";:-.,;.;.- for the Environment - Dedicated to preserving and enhancing our state's Wildlife environmental heritage by supporting educational • Forever and Innovative projects