Quick viewing(Text Mode)

Fire and the Forest History of the North Cascade Range Author(S): Les C

Fire and the Forest History of the North Cascade Range Author(S): Les C

Fire and the Forest History of the North Cascade Range Author(s): Les C. Cwynar Reviewed work(s): Source: Ecology, Vol. 68, No. 4 (Aug., 1987), pp. 791-802 Published by: Ecological Society of America Stable URL: http://www.jstor.org/stable/1938350 . Accessed: 29/11/2011 12:08

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Ecological Society of America is collaborating with JSTOR to digitize, preserve and extend access to Ecology.

http://www.jstor.org Ecology, 68(4), 1987, pp. 791-802 © 1987 by the Ecological Society of America

FIRE AND THE FOREST HISTORY OF THE NORTH CASCADE RANGE1

LES C. CWYNAR2 Departmentof Botany, Universityof Toronto,Toronto, Ontario M5S IA1, Canada

Abstract. Postglacial vegetation changes are often ascribed to the direct effects of climate change. I studied pollen, plant macrofossils, and sediment charcoal in order to determine the potential role of changes in the disturbance (fire) regime in the postglacial development of local vegetation at Kirk Lake in the foothills of the North Cascade Range in northwestern . Five pollen assemblage zones are recognized: a Pinus-Populus zone > 12 000 BP, a Picea-Alnus sinuata zone from > 12 000 to 11 030 BP, an -Pteridium zone from 11 030 to 6830 BP, a Cupressaceae zone beginning at 6830 BP, and a late Holocene Pinus-Alnus rubra zone from 2400 to 900 BP. The first forests (> 12 000 BP) were an open mixture of and deciduous trees, chiefly , Abies, , and Populus on a landscape subject to erosion. Just before 12 000 BP, the pioneer species Picea sitchensis, Alnus rubra, and A. sinuata became important constituents of the forest. Although pollen accumulation rates were high, the abundance of Alnus sinuata indicates an open-canopy forest. Beginning 11 200 BP, climatic warming initiated major changes in forest composititon and the fire regime. migrated into the region, rapidly expanded, then declined shortly thereafter, while Pseudotsuga menziesii, Alnus rubra, and Pteridium expanded, and Pinus contorta, Picea sitchensis, Populus, and Alnus sinuata declined. The abundance of Pseudotsuga, Alnus rubra, and Pteridium between 11 030 and 6830 BP corresponds with increased influxes of charcoal into the sediment; this zone is interpreted as a closed forest with a relatively high fire frequency and composed of a mosaic of postfire successional communities in which fire-adapted Pseudotsuga and Alnus rubra predominated over fire- sensitive Tsuga heterophylla. Pinus monticola became locally important m8000 BP. Be- tween 6800 and 6400 BP arrived, Tsuga heterophylla expanded, and Alnus rubra, Pseudotsuga, and Pteridium declined. These changes are accompanied by a reduced fire frequency, inferred from lower charcoal accumulation rates, and they indicate a shift to wet-temperate climate similar to today's. The late Holocene fossil record shows the development of the adjacent peatland, which Pinus contorta eventually invaded. Key words: CascadeRange; charcoal;fire; forest history;paleoecology; palynology; plant macro- fossils; Quaternarystudies.

INTRODUCTION teraction of physical and biological processes on three time x Recent studies of vegetation emphasize the impor- scales: micro (1-5000 yr, 1-106 m2), macro (5 tance of disturbance (White 1979, Mooney and God- 103-106 yr, 106-1012 m2), and mega (>106 yr, >10'2 ron 1983, Pickett and White 1985). The composition m2). They concluded that "the Quaternary paleoeco- of forests is partly determined by the disturbance re- logical record demonstrates the importance of climatic the gime, which includes average rates, distribution in change as dominant influence upon vegetational space, and severity of disturbances (Loucks 1970, processes and patterns at the macro-scale of spatial- Mooney and Godron 1983, Shugart 1984, Pickett and temporal resolution" and that "even at the micro-scale, White 1985, Runkle 1985). In forests, disturbances prevailing disturbance regimes that affect plant succes- occur on various scales, ranging from small gaps cre- sion must be viewed within the context of their pre- ated by the death of one or a few trees, leading to gap- vailing macroclimate." When macroclimate changes, phase regeneration, to large gaps created by a variety vegetation may therefore be expected to respond not of external factors, such as fire or wind, leading to only to the direct effects of change in climate, but also secondary succession (Runkle 1985). to the indirect effects of change in the disturbance re- Delcourt et al. (1983) developed an hierarchical gime. model of how vegetation patterns arise from the in- Although fire is an important agent of disturbance in the modern evergreen coniferous forests of the Pa- Manuscriptreceived 8 April 1986; revised 15 September cific Northwest and some forest 19 1986. (PNW) types depend 1986; accepted September on it for their and 2 Presentaddress: Department of Biology,Sir Wilfred Gren- regeneration (Franklin Dyrness 1973, fell College, Memorial University of Newfoundland, Corer Agee 1981, Hemstrom and Franklin 1982), the post- Brook, NewfoundlandA2H 6P9, Canada. glacial development of PNW forests is generally inter- 792 LES C. CWYNAR Ecology, Vol. 68, No. 4 preted as resulting from the direct effects of climate the maximum extent of the Cordilleran Ice Sheet (Waitt change. The general sequence of postglacial vegetation and Thorson 1983) of the Fraser Glaciation (circa change at lowland sites in the northern Puget Trough 25 000-10 000 BP). The North near (Hansen 1938, 1947, 1948, Hansen and Easterbrook Peak were deglaciated before 12 250 BP (Beget 1984). 1974, Baker 1983, Heusser 1983, Barnosky 1984) be- The vicinity of Kirk Lake was deglaciated before gins with open forests of various admixtures of tem- I 1 500 BP, and the lake probably formed as a kettle perate lowland (Pinus contorta, Populus) and montane in outwash or kame terrace (J. E. Beget, personal taxa (, Tsuga mertensiana). At - 11 200 communication). in the region re-advanced in BP temperate lowland species began to replace mon- the early Holocene 8400-8300 BP (Beget 1984). No tane species so that by 10 000 BP open forests domi- bedrock is exposed within the drainage basin of Kirk. nated by Pseudotsuga and Alnus rubra had developed Surrounding mountains are composed of crystalline at most sites. These forests persisted until -6000 BP rocks, principally schists (McKee 1972). when modern closed forests of temperate conifers The lake lies within the Tsuga heterophylla Zone (Pseudotsuga, Thuja plicata, Tsuga heterophylla) de- (Franklin and Dyrness 1973), which regionally is dom- veloped. The climates inferred from this sequence are inated by Pseudotsuga menziesii (Douglas-), Tsuga cool-maritime conditions initially, gradually warming heterophylla (western hemlock), and Thuja plicata to a warmer and drier period relative to modern cli- (western red cedar). (Pacific silver fir), mate between 10 000 and 6000 BP, and then the de- A. grandis (grand fir), Picea sitchensis (Sitka ), velopment of the modern humid climate after 6000 Pinus contorta (lodgepole ), and P. monticola BP. (), are minor constituents. The most The role of fire in either maintaining plant com- common deciduous tree is Alnus rubra (red alder). Near munities or in initiating vegetation change has increas- the lake, Thuja plicata and Tsuga heterophylla are the ingly attracted the attention of Quaternary paleoecol- most common trees together with some Pseudotsuga ogists, especially since charcoal is abundant in lake menziesii. A peatland on the east side supports a stand sediments and easily quantified (Waddington 1969, of Pinus contorta with scattered Picea sitchensis and a Swain 1973, Cwynar 1978, Green 1981, Grimm 1983, shrub understory of Ledum groenlandicum (Labrador Lamb 1985, Winkler 1985a, b, Dunwiddie 1986). Some tea) and some Alnus sinuata (Sitka alder). A small stand researchers have recognized the potential significance of Tvpha latifolia (cattail) grows at the north end of of fire in the postglacial development of PNW forests the lake. (Hansen 1947, Heusser 1973a, Mathewes 1973, Leo- pold et al. 1982), but only Sugita and Tsukada (1982) METHODS and Dunwiddie (1986) have provided detailed studies I collected a 12.44-m core from peat on the west of charcoal from lake sediments, and Mathewes (1985) shore of the lake with a 5 cm diameter modified Liv- has stressed the need for direct evidence of past fire ingstone piston sampler (Wright 1967) and help. I re- regimes from sedimentary charcoal profiles for a full moved samples of 0.5 cm3 with a calibrated brass sam- understanding of postglacial vegetation change in the pler (Birks 1976) for pollen analysis (at 20-cm intervals PNW. My primary objective in this paper is to recon- above 9 m, and 5- or 10-cm intervals below 9 m) and struct the history of local vegetation around Kirk Lake, to determine percent mass-loss on ignition (LOI) at Washington in relation to disturbance by fire. I spe- 550°C. Tablets of Eucalyptus pollen or Lycopodium cifically address the following questions: (1) Are post- spores were added to pollen samples for determination glacial changes in vegetation associated with changing of pollen concentration and pollen accumulation rates fire regimes? (2) What was the nature of the vegetation (PARs). I prepared samples for pollen analysis using between 10 000 and 6000 BP and what was the relative standard methods (Faegri and Iversen 1975) and, in importance of change in climate vs. the disturbance addition, sieved basal inorganic sediments to remove regime in initiating the development and maintenance fine particles (Cwynar et al. 1979). I tallied a minimum of this vegetation, which has generally been interpreted sum of 300 upland pollen types (trees, shrubs, terres- as an open Pseudotsuga woodland possibly maintained trial herbs, pteridophyte spores, and unknowns) for by fire and resulting from the direct effects of a much each sample and used it to calculate pollen percentages. drier climate relative to modern? I argue that the vege- The Haploxylon/Diploxylon determination of pine tation was a closed forest, but more importantly, that pollen was based on a minimum of 25 grains per sam- a change in the fire regime and not the direct effects of ple. I zoned the pollen diagrams using constrained clus- climate change initiated and maintained this forest. ter analyses (Gordon and Birks 1972). Charcoal was analyzed by the method of Waddington (1969). Clark SITE DESCRIPTION (1984) has shown that charcoal counts derived from Kirk Lake is small (0.6 ha), relatively deep (7.6 m), pollen preparations are sensitive to the preparation and is on the western flank of the North Cascade Range techniques used; I treated all samples identically during at 190 m altitude (Fig. 1; 121°37' W, 48°14' N; Section sample preparation. 23 N/P, Township 32 N, Range 9 E). It lies just within For plant macrofossil analysis, I removed 5 cm long August 1987 FORESTAND FIRE HISTORYOF 793 core segments, measured their volumes (generally 60- r 80 cm3) by displacement in a graduated cylinder, wet- sieved each through screens of 1000-, 425-, and 106- Alm mesh (clayey or silty samples were disaggregated in a 5% sodium pyrophosphate solution), and then picked, identified, and enumerated all seeds, fruits, co- nifer needles, and other recognizable plant parts from the residues. Identifications were based on modern ref- erence material. I embedded in wax several needles from basal sediments and sectioned them for critical identifications. I included only samples with a minimum of 20 conifer needles in the preparation of the percentage diagram of conifer macrofossils (Dun- widdie 1987), with the exception of two samples from the early postglacial period; I amalgamated some sam- ples in order to raise the needle sum to 20.

STRATIGRAPHY Numerous coarse roots and wood fragments pre- vented the collection of the uppermost 1 m of sedi- ment. None of the core reacted with 10% HC1, indi- cating the absence of carbonates. The sediment FIG. 1. Location of Kirk Lake and the sites frequently mentioned in the text. stratigraphy was as follows (Munsell Color descriptions [Munsell Color 1975] are given in parentheses): Banded with some silt. Coarse Depth 1200-1244 sandy clay below sand at the very bottom (N 3/0). Inorganic. surface Below the ash couplet LOI was <6% (see Fig. 3), (cm) Description whereas above the ash LOI rose rapidly to 40-75% 100-360 Peat, well-decomposed at the bottom and throughout most of the core, increasing to a maximum coarser at the top (7.5 YR 3/4). of 94% in the upper peat. 360-715 Coarse gyttja grading to decomposed peat (5 YR 4/3). 715-1151 Fine detritus gyttja grading upward to CHRONOLOGY coarse gyttja. (Grading in color from 7.5 Seven radiocarbon dates (Table 1) were obtained YR 2/1 to 5 YR 2/3.) A volcanic ash oc- from the lower half of the core where most of the pollen curred between 775-771 (2.5 YR 8/1). stratigraphic events occur. The geographic distribution 1151-1184 Predominantly silt (N 4/0) with complex of Holocene tephra layers (Sarna-Wojcicki et al. 1983), banding of thin sand and clay layers and the pollen stratigraphy, and radiocarbon dates indicate some brownish, slightly organic layers. A that the tephra at 771-775 cm is from . couplet of volcanic ash layers occurred at Its age of ;6800 BP (S. Porter, personal communi- 1151-1153 cm; each layer was -2 mm cation) is used as another point on the age-depth curve. thick and formed a sharp contact with ad- Deposition times were calculated by linear interpola- jacent sediment. Slightly organic, brown- tion (Fig. 2). The two oldest dates (14 610 and 13 250 ish silt lay between the ash layers. BP) may not be reliable because the sediment dated has a low organic content (<6%), which may result in spuriously old dates (Olsson 1979). If these dates are TABLE1. Radiocarbondates from Kirk Lake, Washington. correct, the sedimentation rate would have been very slow, which seems unlikely because of the many small Sediment depth Radiocarbon age Laboratory layers of sand, which probably sedimented rapidly. An interval (BP + 1 number (cm) SD) hiatus below the ash couplet is an alternative to slow 865.0-870.0 8600 + 40 QL-1614 continuous sedimentation. In either case, reliable de- 960.0-965.0 9580 + 60 QL-1615 times cannot be based on the two oldest dates. 1060.0-1065.0 10510 + 60 QL-1616 position 1105.0-1110.0 11 180 + 60 QL-1617 Therefore, I extrapolated deposition times downward 1125.0-1130.0 11500 + 160 QL-1618 only to the top of the ash couplet at 11.51 m and 1153.5-1158.0 13 250 + 90 QL-1649 to the in sediment from fine to 1165.0-1173.0 14610 + 220 QL-1619 upward change type coarse detritus gyttja. 794 LES C. CWYNAR Ecology,Vol. 68, No. 4

Xp associated PARs (Fig. 4), charcoal accumulation rates 14 - Coarse t ti (CHARs; Fig. 4), and plant macrofossils (Figs. 5 and --detritus 14 Fine detritus g gyttja I Silt- gyttja 6) are summarized in Table 2. _ I , Although it is clear from the superposition of the CQ 12- pollen zone boundaries onto the macrofossil concen- 6.00 tration diagram (Fig. 5) that there is excellent agree- 14.89 co _ I ment between the pollen and macrofossil data, there O 10- ,-. 9.30 i are, nevertheless, several important local changes in LxJ /10.3 2 1 the vegetation that are not as evident in the pollen I 8 I 19 profiles as in the macrofossil record. In the mid-Ho- locene there is a shift from Diploxylon to Haploxylon pine pollen, although the overall abundance of pine A remains low. The macrofossil record, however, shows 6 7 8 9 10 11 12 that P. monticola became important in the local vege- DEPTH (m) tation when this shift occurred. Similarly, when Picea FIG. 2. Age-depth curve for Kirk Lake. · Mazamatephra pollen appears consistently in the late Holocene at val- (6800 BP), · radiocarbon dates used in the calculation of ues < 2% (a subtle change that occurs at other sites and deposition times in years per centimetre, dates that are not is in consideredreliable. generally disregarded vegetation reconstructions), leaves of Picea sitchensis recur indicating its local reap- pearance. RESULTS AND INTERPRE'TATION The percentage macrofossil diagram for conifers (Fig. Five local pollen assemblage zo nes are informally 6) indicates several interesting trends. Pseudotsuga was recognized on the basis of constrairled cluster analysis clearly the dominant conifer in the early Holocene, but (Fig. 3). The chief features of the,se zones and their Tsuga heterophylla was important locally despite its

KIRK LAKE, Washington Trees and Shrubs I

>s \C .c\ (0 b\ S \Cs? ,r' /^'_. / Co 5.5.y OS I a e 6 " E - *$o * Q(5 a?4 S~o'cj~osb~b·'~ ·~ive~'o Z 0 a Hrr,k4-rkkr-s k I iiii,HVr~mkr-n-,-r H HH H Pollen hhlF,-, hH 0 HHHc c [hllh 05.sse ' ' ' ' : emblage zones · '! : ! to .\ KL-5 . .

. 4- . . 1- . . . Pinus- I1. *: . . . Alnus rubro * I - . .

. . 3-

L I 4- KL-4

I) 5- Cupressaceae ii .E 1- .0 6-

7-

j~~I 14c ash : dates I 8- I (yr BP) .1 KL-3 - 8600 9- I : Alnus rubra- : Pteridium . -9580 .a) i! I U. 10 -

' ". .' , .:-10! 510 Glacier :I {Peak 1 - - I1' : : : . _ 11 180 * KL-2 Picea- couplet '*** | ' i *i i * 0m1 500 t-!' : . : : : -11 250 Alnus sinuafa C w :** : : : : . : . . . : -13 KL-1 Silt I ' :-rrr14610 : Pinus-Populus " a y -2 - i I 1 1 Ir-f i i' m rIr Tr r I I I,, r-,r rpp-p-r-m T M ppP Pl rrPpplpprprnp clay llrrrTrfll 0 60 .8 10 0 50

FIG.3. Pollen percentagediagram for Kirk Lake. August 1987 FORESTAND FIRE HISTORY OF NORTH CASCADES 795 low pollen percentages. Furthermore, Pseudotsuga de- the lake muds of Kirk Lake. Second, Pseudotsuga and clined in local abundance in the mid-Holocene as Tsu- other conifers have been selectively logged so that sed- ga heterophylla and Thuja plicata increased. The per- iments show a striking decrease of conifer pollen and centage diagram highlights these trends, but in general an increase of Alnus pollen, attributable to A. rubra, there is little difference between the concentration (Fig. which colonizes disturbed environments (Davis 1973, 5) and percentage (Fig. 6) macrofossil diagrams. When Heusser 1978a, c, Barnosky 1981, Sugita and Tsukada interpreting these diagrams, however, we should bear 1982, Dunwiddie 1987). Pollen spectra immediately in mind that nothing is known about the potential for below the settlement horizon better describe the pollen differential preservation of macrofossils and that de- rain of the natural forests of the region (Davis 1973, ciduous species were important (Populus and Alnus Heusser 1978d), but they are too few in number to rubra) early in the postglacial period. form an adequate basis for comparison. Kirk Lake is small, so most of the pollen it receives Basis for interpretation is derived from local or extra-local sources, i.e., from Paleoecologists now often use quantitative methods plants growing within several hundred metres of the to compare fossil pollen assemblages with pollen as- lake (Jacobson and Bradshaw 1981). Plant macrofossils semblages produced by modern vegetation so that are also derived from local source areas, but areas vegetation reconstructions may be refined (MacDonald smaller than those for pollen (Watts 1978, Birks 1980). and Ritchie 1986). This approach was not possible in The reconstructions possible from the record at Kirk this study for two reasons. First, pollen spectra from are therefore of the local vegetation, but because of the different types of deposits are not directly comparable lack of modern pollen and macrofossil samples for (e.g., Ritchie 1974, Birks 1977). Most samples of mod- comparison with the fossil assemblages, the reconstruc- ern pollen spectra from the PNW are derived from tions are necessarily qualitative and based on the abun- moss polsters (Heusser 1973b, 1978a, c, d) and there- dances of various taxa in the pollen and macrofossil fore cannot be compared with the fossil spectra from records. When interpreting the pollen record, I have

Herbs Aquatics and / Marsh Plants os P · ,. / ·/ s

Ilen VT-d-iH-HHF4-4,sV hFH~l? ~ H ~ HH -· Ihh H h, -·, ih rr. E I 'i-- n. ... hhh'hhhhhhll. ... Lih I;:.,!.1hh~-1hhhI 'i' :r

I` . I i I · I r i-

.-1

j~~~~~~~~~~ .~:·:::: I __ -

I r-mlrlm r rPP rrlrnrPrrrrrrrnrpnrTlmll i m nn""rrP1 i" 0 20 10 0 500 010 0 500 0.1 0.3 0.5 103 grains/cm3

FIG. 3. Continued. 796 LES C. CWYNAR Ecology, Vol. 68, No. 4

I

U0

l . 1 1 TI II . . I I I . l 7l I l I I 5 10 15 5 60 2.5 5.0 7.5 2 100 pm ·cmc-2-myr yr FIG. 4. Pollen and charcoal accumulation rate diagram.

/I Trees and Shrubs I I·""," is~ Q ~ FVV 9'$ O~ f~ <\9 9 \ ( QNC~,1~ I 1 6 r.6 \. I ,. I I I i .4~ I I I . I * =

_ I

E:

_. :p- I

__ *: ,- i · · · " ' ]' -- iE I ' I '-- - I N I I I SandyTInTIF' 510 50100 10 2 100 100 400 10 100 50 3 MacrofossilConcentration (macrofossils/ 100 cm )

FIG. 5. Concentration diagram of plant macrofossils from Kirk Lake. For the trees and shrubs, all macrofossils are leaves unless otherwise shown. For all other plants, they are fruits or seeds, unless otherwise indicated. The concentration data are plotted on a log scale. August 1987 FORESTAND FIRE HISTORY OF NORTH CASCADES 797 borne in mind that Pseudotsuga and Tsuga hetero- reconstructions phylla are generally under-represented and Alnus rubra Paleoecological over-represented in modern pollen spectra relative to Pinus-Populus zone (KL-1).--This zone is present their abundance in the vegetation (Heusser 1973b, at most sites in the PNW (Barnosky 1981) and is gen- 1978a, c, d). Dunwiddie (1985, 1987) has shown that erally interpreted as representing a pine woodland in the PNW, conifer needles can be identified to species dominated by Pinus contorta (confirmed at Kirk Lake and are abundant in the sediments of small lakes and by the predominance of Diploxylon-type pollen and ponds; their abundance in recent sediments is propor- macrofossils), Populus, and Tsuga mertensiana. The tional to the importance of the species in the local inorganic nature of the sediment indicates that the vegetation (within 30 m of the site). Therefore, al- vegetation cover was probably too sparse to stabilize though I have considered the pollen and macrofossil the fresh soil surfaces. The occurrence ofAlnus sinuata, records together, I have relied more on the specific which today grows in the Cascade Range on inorganic determinations and abundances of macrofossils to re- soils where forest canopies are open (Oliver et al. 1985), construct the local vegetation. The pollen record has also indicates that the forests were open. been primarily used to correlate the record at Kirk with Picea-Alnus sinuata zone (KL-2).--Similar pine- the records at other sites. dominated pollen assemblages from western Washing- Swain (1973) showed that past fires around a lake ton have generally been interpreted as representing could be identified by the presence of distinct charcoal open, woodland vegetation. This interpretation is par- peaks in the sediments of the lake. Changes in the ticularly appropriate for sites farther south in the gla- amount of charcoal accumulating in lake sediments ciated areas of southwestern Washington (e.g., Bar- (CHARs) may therefore be taken as an indication of nosky 1981, Tsukada and Sugita 1982) where PARs changes in the fire regime. However, it is not possible are comparatively low (5000-10 000 grains cm-2 yr- ), to determine whether the fire regime is changing with the sediment organic content is low, and open-ground respect to the frequency or magnitude of fires. taxa are common (Thalictrum, Epilobium, Saxifraga

Herbs 4 Trees and Shrubs¢ - /4 --:/-- -Floating and Submerged- · Emergent Aquatics E 0Aquatics,-.,*,,,S , _ ","^,4' 6x \ o ·o \\\O c 0& . Pollen o aI °,° a z-m °, Gb / assemblage zonesS

: Pinus - 2E-~~~~~~~~*....':~~~~~~~~~~~~~~..~. : *__\_ : ...L.~ :Alnus rubra

- 3- -

1= 4-: _ . l , - .- KL-4 . 5- r- . . PI Cupressaceae

6- : -- " r

7-: I; 8-. 14C dates KL-3 (yr B P) - 9-- -8600 Alnus rubra Pteridium ;. F i: 10-' cl -9580 : :: . r- .=:16 ' · ' · 11. , - · - · · · i_ -10 510 - , KL-2 Picea : _11 180 Alnus sinuata ------11 500 F o ' I r' , '' 1 r- , F r n m n m7 I m Ir- I -13 250 510 10'o 2r2 10 "14 610

FIG. 5. Continued. /

798 LES C. CWYNAR Ecology, Vol. 68, No. 4

E , 0 110 1 0 NSR I , , I I I I I 0 9^~') O1 - * 0) I I , H-1 I- I I I I I 1 I-

KL-5 2-

-.7 3- I

4-

5- KL-4 U 6-

7- 8- I 9- KL-3

10- I II- I-- KL-2 20' I I-- I i 0 60 20 60 60 20 20 100 100 100 %

FIG. 6. Percentage diagram of conifer-leaf macrofossils.

tricuspidata-type). At Kirk Lake the abundance of Al- somewhat open coniferous forest of Pinus contorta, nus sinuata indicates that the forest canopy must also Picea sitchensis, Tsuga mertensiana, and Abies lasio- have been partially open. The higher PARs at Kirk, carpa. The abundance of Picea sitchensis probably re- however, suggest that forests there were less open than flects its ability to colonize open habitats. At present, those in southwestern Washington. Because PARs are both Picea sitchensis and Pinus contorta colonize dunes a function not only of pollen productivity by vegetation along the PNW coast and invade prairie in south- but also of pollen transport to the lake by streams (Peck (Franklin and Dyrness 1973). 1973, Bonny 1976, Pennington 1979), other explana- Populus and Alnus rubra were important deciduous tions for these high PARs are possible, although they trees. A. sinuata was most prominent during this time, seem improbable. For example, erosion of pollenifer- probably growing in canopy gaps. Fire was a minor ous soil into the lake may have inflated the PARs some- influence on this vegetation. CHARs in this zone are what. However, it is unlikely that that is solely re- moderate relative to the rest of the record and are sponsible for the high PARs, because PARs remain consistent with dominance by P. contorta ssp. contorta, high at the end of the zone where the organic com- the coastal type that occurs in the PNW today. It differs ponent (LOI) of the sediment reaches 60% of sediment from ssp. latifolia, the inland or Rocky Mountain type, dry mass, in comparison with 6% at the beginning of in having open cones that shed their seeds seasonally the zone. Another possible explanation is that these (Critchfield 1980). This is generally regarded as an ad- high PARs result from changing patterns of sediment aptation to a less severe fire regime than that experi- deposition within the lake (Davis et al. 1984). Multiple enced by ssp. latifolia in the drier interior. cores would be needed to test this hypothesis rigor- Alnus rubra-Pteridium zone (KL-3). -This zone has ously, but it seems unlikely because the core was re- been variously interpreted as representing "structurally covered from the side of the basin, far from the pre- open forest, unlike the forest of today" (Heusser 1978b), sumed center of deposition where the water is 7.6 m unstable transitional open woodlands (Tsukada et al. deep today. 1981), woodland (Barosky 1981, Sugita and Tsukada The local vegetation around Kirk Lake during this 1982), or either open woodland or a forest mosaic (Leo- period was, therefore, a well-developed but nonetheless pold et al. 1982, Barnosky 1984). The interpretation August 1987 FOREST AND FIRE HISTORY OF NORTH CASCADES 799

TABLE 2. Summary charactersitics of pollen zones and associated plant macrofossils and charcoal.

Pollen zone KL-1 KL-2 KL-3 KL-4 KL-5 Informal name Pinus-Populus Picea-Alnus sinu- Alnus rubra- Cupressaceae Pinus-Alnus rubra- ata-type Pteridium type Depth (cm) 1180-1157.5 1157.5-1097.5 1097.5-772.5 772.5-275 275-100 Age (BP) >12000 12 000-11 030 11030-6830 6830-;2400 ;2400-900 Majorpalyno- Pinus, Populus, Pinus, Alnus sinu- Alnus rubra- Cupressaceae, Cupressaceae,Pi- morphs Alnus sinuata- ata-type, Picea, type, Pterid- Tsuga hetero- nus, Alnus rubra- type, Tsuga Alnus rubra-type ium phylla, Alnus type mertensiana rubra-type Total PARs (103 ... 30-65 15-30 15-25 grains.cm-2 yr-1) Majorterrestrial Pinus contorta Picea sitchensis, Pseudotsuga, Thujaplicata, Pinus contorta, plant macro- (rare) Abies (lasiocar- Tsuga hetero- Pseudotsuga, Ledum, Thuja fossils pa), Pinus contor- phylla Tsuga hetero- plicata, Tsuga ta, Alnus rubra, phylla, Pinus heterophylla and Alnus sinu- monticola,Pic- ata ea sitchensis CHARs (106 *** 2-4 < 1-8 < 1-2 Amm2cm-2. yr-1)

that this zone represents an open forest community is of Pseudotsuga menziesii, Thuja plicata, and Tsuga commonly based on two observations. First, Pteridium heterophylla (Davis 1973, Scott 1980), pollen spectra is a heliophytic fern in which sporulation decreases from sediment immediately below the settlement ho- with increasing shade (Conway 1957, Page 1976). Its rizon contain 5-60% Pseudotsuga pollen (Davis 1973, spores are abundant in this zone, implying the presence Leopold et al. 1982). Similarly, at Mineral and Hall of an open community. Secondly, the percentages for Lakes, presettlement Pseudotsuga percentages are - 15 Pseudotsuga pollen in this zone are maximal for the and 9%, respectively (recalculated from Figs. 4 and 5 postglacial period, and they exceed considerably per- in Sugita and Tsukada [1982] using the total pollen centages found in surface samples from modern closed and spores of upland plants as the pollen sum, as at forests in the Puget Lowland and adjacent areas (Heus- Kirk Lake). Dunwiddie (1987) found 20% Pseudotsuga ser 1973b, 1978a, c, d). Since trees generally flower pollen in the surface sediment of a pond on Mount most profusely when grown in open stands (Faegri and Rainier located within a dense old-growth stand of Iversen 1975), and the highest values for Pseudotsuga Pseudotsuga that formed 80% of the total basal area. pollen in surface samples come from Pseudotsuga The 10-17% Pseudotsuga pollen at Kirk Lake is there- woodlands in the rainshadow of the Olympic Moun- fore consistent with expected percentages for Pseudo- tains (Heusser 1978a), a woodland hypothesis was tsuga in closed forests. widely accepted. There are several difficulties with this Secondly, at sites where total PARs have been hypothesis. calculated for this zone, they are high: 20 x 103 First, the analogy with open, dry rainshadow com- grains-cm-2 yr-1 at Kirk Lake, 20-45 x 103 grains- munities is not compelling because Alnus rubra and cm-2.yr-1 at Davis Lake in the southern Puget Pteridium, both of which increase in this zone together Lowland (Barnosky 1981), 10 x 103 grains cm-2 yr-1 with Pseudotsuga, are infrequent in these communities at Mineral Lake, and 16 x 103 grains cm-2 yr-1 at (Franklin and Dymess 1973, Fonda and Berardi 1976). Hall Lake (Sugita and Tsukada 1982). There is no This analogy is also linked with the climatic interpre- change to a more-minerogenic sediment in this zone, tation that increased Pseudotsuga implies greater arid- so these high PARs do not result from increased ero- ity, but it seems paradoxical that pollen ofAlnus rubra, sion associated with an open forest. Presettlement PARs a species of moist habitats, increases in abundance for Pseudotsuga at Mineral and Hall Lakes (1200 and simultaneously with that of Pseudotsuga. Secondly, the 2500 grains cm-2 yr-1, respectively [Sugita and Tsu- low percentages of Pseudotsuga pollen in surface sam- kada 1982]), where closed forests dominated by Pseu- ples may be an artifact of selective logging of Pseu- dotsuga, Thuja plicata, and Tsuga heterophylla grew dotsuga coupled with the over-representation of Alnus before settlement, are similar to the mean values for rubra pollen. this zone at Kirk (2050 ± 920 grains cm-2 yr-1). Such Several lines of evidence suggest that the forest was a correspondence indicates that this zone represents a closed. First, at Lake Washington, where the natural closed forest. presettlement forest was dominated by closed stands Thirdly, as previously mentioned, Alnus sinuata does 800 LES C. CWYNAR Ecology,Vol. 68, No. 4

not grow in closed forests (Oliver et al. 1985). A. sinu- plicata, and Tsuga heterophylla trees predominate. The ata-type pollen declines as Pseudotsuga increases at coincident increase of Thuja and decline of Alnus (both the beginning of this zone, implying that the forest of which grow in moist habitats) in both the pollen and canopy became closed. macrofossil records suggest that Thuja plicata replaced How can the evidence for both closed and open for- Alnus rubra trees on the wettest sites least prone to ests be reconciled? Hansen (1948), Heusser (1973a, fire. Pinus monticola, a tree of well-drained sites, be- 1978a), Barosky (1981), Leopold et al. (1982), Ma- came a minor but consistent local component of the thewes (1973, 1984, 1985), and Tsukada et al. (1981) forest. argued that fire has been important for maintaining The greater particle size of the sediment in the upper open communities during this time. The charcoal pro- part of this zone and the occurrence of Ledum and file from Kirk Lake clearly shows increased charcoal Gaultheria shallon macrofossils indicate the encroach- accumulation rates in this zone. Similarily, at Hall Lake, ment of the growing peatland toward the coring site. 80 km southwest of Kirk Lake, high percentages and Both Pinus contorta and Picea sitchensis once again PARs of Pseudotsuga, Alnus, and Pteridium in the early grew locally. They were probably restricted to the de- Holocene are accompanied by postglacial maximum veloping peatland, as they are today at Kirk and other CHARs (Sugita and Tsukada 1982). Thus, disturbance similar sites in western Washington (Franklin and Dyr- by fire was more important during this time. ness 1973). Both Pseudotsuga and Alnus rubra are well adapted Pinus-Alnus rubra zone (KL-5).-The initiation of to survive and regenerate after fire (Fowells 1965), and peak deposition indicates the growth of the peatland they are the most important trees following disturbance over the coring site. The most extensive peatland in in the Tsuga heterophylla Zone (Munger 1940, Frank- the basin is on the east side of the lake opposite the lin and Dyrness 1973). In contrast, Tsuga heterophylla coring site. The increase in Diploxylon pine pollen and is especially sensitive to fire owing to its thin bark and of Pinus contorta macrofossils reflects the continued shallow roots, which are often exposed (Fowells 1965). colonization of the peat by pine as well as its greater Pteridium is also abundant during early successional proximity to the coring site. The increased amplitude stages following logging or fire, particularly during the in the abundance of pollen of most taxa, such as Alnus, 1st 5 yr after disturbance (McMinn 1951, Franklin and probably results from greater local heterogeneity of Dyrness 1973). fossil deposition on peat surfaces relative to lakes. It Repeated fires leave an irregular, patchy distribution is not clear, for example, whetherAlnus rubra or Pterid- of vegetation on the landscape in various stages of ium were actually more abundant at this time. The succession. A higher fire frequency would result in a vegetation changes of this period can be attributed to higher proportion of patches in relatively young succes- local peat growth and expansion, and the forests of sional stages, and thus a greater abundance of Alnus surrounding uplands may not have changed. rubra and Pteridium on the landscape. In addition, DISCUSSION pollen percentages of fire-sensitive Tsuga heterophylla declined early in this zone and remained at low values Pickett (1976) developed an evolutionary model of as the percentages of Pseudotsuga pollen increased. In succession in which he viewed landscapes as "mosaics the absence of disturbance, T. heterophylla invades of successional habitats, generated by random periodic stands of Pseudotsuga, replacing it in late seral stages. disturbance." In this model, vegetation on the scale of Because of the higher fire frequency during zone KL-3 landform units depends in part on the frequency of time, however, the development to later successional disturbance. Hansen (1947) argued that the postglacial stages was preempted more frequently so that Pseu- development of vegetation in the Puget Lowland was dotsuga was less often replaced by T. heterophylla, ex- influenced "regionally by fire and locally by soil con- cept probably at the wettest sites in the drainage where ditions, and to a limited extent by climate." Although fire frequency remained relatively low. the ultimate cause of postglacial vegetation change was I interpret this zone as representing a closed forest climate change, the charcoal record at Kirk indicates dominated by Pseudotsuga, Alnus rubra, and Tsuga that change in the fire regime was a proximate cause heterophylla as a mosaic of postfire seral communities for some of the postglacial changes in vegetation. Spe- of different ages. These forests were similar to modem cifically, the increase of Pseudotsuga, Alnus rubra, and closed forests of the region except that they lacked Pteridium coincident with a decline of Tsuga is readily Thuja, Tsuga heterophylla was less abundant, and the explained in terms of a shift to a relatively younger proportion of landscape patches in early successional mosaic of postfire successional communities as a con- stages was relatively high because of a higher frequency sequence of more frequent fires. The trend to warmer of disturbance by fire. and/or drier climate in the PNW during zone KL-3 Cupressaceae zone (KL-4).-The increased abun- time is well established (Heusser 1983, Barosky 1984), dances of Thuja plicata and Tsuga heterophylla pollen but if greater dryness itself were the proximate cause and macrofossils indicate the development of forests of these changes from zone KL-2 to KL-3, then the similar to those of today in which Pseudotsuga, Thuja increase of A. rubra, which grows in moist habitats, August 1987 FOREST AND FIRE HISTORY OF NORTH CASCADES 801 would be paradoxical. Rather, a small change to drier sediment of some Lake District lakes. Journal of Ecology 64:859-887. climate probably triggered a relatively large change in R. L. 1984. Effects on charcoal of the disturbance fire Clark, pollen preparation regime by increasing frequency. techniques. Pollen et Spores 26:559-576. Similarly, the later increased abundances of Tsuga het- Conway, E. 1957. Spore production in bracken (Pteridium erophylla and Thuja plicata trees in the vegetation and aquilinum [L.] Kuhn). Journal of Ecology 45:273-284. concomitant decreases of Pseudotsuga and A. rubra Critchfield, W. B. 1980. Genetics of lodgepole pine. Forest Service Research WO-37. may be attributed to less frequent fires. Paper Cwynar, L. C. 1978. Recent history of fire and The of the disturbance itself as vegetation importance regime from laminated sediment of Greenleaf Lake, Algonquin a factor influencing forest composition is apparent in Park, Ontario. Canadian Journal of Botany 56:10-21. the PNW where forests have recently changed drasti- Cwynar, L. C., E. Burden, and J. H. McAndrews. 1979. An method for cally in response to an altered disturbance regime dis- inexpensive sieving concentrating pollen and from sediments. Canadian Journal of sociated from climate has spores fine-grained change. Logging artificially Earth Sciences 16:1115-1120. elevated the frequency and severity of disturbance. The Davis, M. B. 1973. Pollen evidence of changing land use result is a mosaic of relatively young successional com- around the shores of Lake Washington. Northwest Science munities in which A. rubra and Pteridium are abun- 47:133-148. Davis, M. R. E. and J. Ford. 1984. Sediment dant. Recent pollen spectra contain abundant A. rubra- B., Moeller, focusing and pollen influx. Pages 261-293 in E. Y. Haworth and Pteridium Heusser type pollen spores (Davis 1973, and J. W. G. Lund, editors. Lake sediments and environ- 1978d), and they are remarkably similar in this respect mental history. University of Leicester Press, Leicester, En- to spectra from zone KL-3 (Davis 1973, Heusser 1978d). gland. Delcourt, H. R., P. A. Delcourt, and T. Webb III. 1983. ACKNOWLEDGMENTS Dynamic plant ecology: the spectrum of vegetational change in and time. Science Reviews 1:153-175. I thank W. A. Watts for the opportunity to do this research. space Quaternary P.W. 1985. Dichotomous to conifer Detailed reviews by C. Barnosky and E. Grimm greatly im- Dunwiddie, key foliage in the . Northwest Science proved an earlier version of this paper. P. W. Dunwiddie, G. 59:185-191. 1986. A record offorest on Mount M. MacDonald, R. W. Mathewes, J. C. Ritchie, S. Sugita, 6000-year history 67:58-68. and W. A. Watts made useful suggestions on the manuscript. Rainier, Washington. Ecology 1987. Macrofossil and of co- C. Barnosky, P. W. Dunwiddie, E. Grimm, and W. A. Watts pollen representation niferous trees in modern sediments from Ecol- helped to core the site. P. W. Dunwiddie helped to identify Washington. 68:1-11. some of the macrofossils. M. Stuiver kindly supplied the ra- ogy and Iversen. 1975. Textbook of diocarbon dates. A. Christmas provided clerical assistance. Faegri, K., J. pollen analysis. Third revised edition. New New USA. This work was funded by a National Science Foundation grant Hafner, York, York, R. and J. A. Bernardi. 1976. of Sucia to E. B. Leopold and an Operating Grant (U0242) to me from Fonda, W., Vegetation Island in Bulletin of the the Natural Sciences and Engineering Research Council of , Washington. Torrey Canada. Botanical Club 103:99-109. Fowells, H. A. 1965. Silvics of forest trees of the United States. United States Department of Agriculture Handbook 271. LITERATURE CITED Franklin, J. F., and C. T. Dyrness. 1973. Natural vegetation Agee, J. K. 1981. Fire effects on Pacific Northwest forests: of and Washington. United States Department of flora, fuels, and fauna. Pages 54-66 in The Northwest Fire Agriculture Forest Service General Technical Report Council 1981 Proceedings. Northwest Fire Council. Port- PNW-8. land, Oregon, USA. Gordon, A. D., and H. J. B. Birks. 1972. Numerical methods Baker, R. G. 1983. Holocene vegetational history of the in Quaternary paleoecology. New Phytologist 71:961-979. . 109-127 in H. E. Pages Wright, Jr., Green, D. G. 1981. Time series and postglacial forest ecol- Late environments of the United States. editor. Quaternary ogy. Quaternary Research 15:265-277. Volume 2. of Minnesota University Press, Minneapolis, Grimm, E. C. 1983. Chronology and dynamics of vegetation USA. Minnesota, change in the prairie-woodland region of southern Min- C. W. 1981. A record of late Barosky, Quaternary vege- nesota, USA. New Phytologist 93:311-350. tation from Davis southern Wash- Lake, Puget Lowland, B. and D. J. Easterbrook. 1974. Research 16:221-239. Hansen, S., Stratigraphy ington. Quaternary and of late sediments in the .1984. Late Pleistocene and Holocene envi- palynology Quaternary Puget early Sound of America Bulletin 85: ronmental of southwestern U.S.A. region. Geological Society history Washington State, 587-602. Canadian Journal of Earth Sciences 21:619-629. H. P. 1938. forest succession and cli- J. E. 1984. Tephrochronology of Late Wisconsin de- Hansen, Postglacial Beget, mate in the Sound 19:528-548. glaciation and Holocene glacier fluctuations near Glacier Puget region. Ecology . Peak, North Cascade Range, Washington. Quaternary Re- 1947. Postglacial forest succession, climate, and search 21:304-316. chronology in the Pacific Northwest. Transactions of the Birks, H. H. 1980. Plant macrofossils in Quaternary lake American Philosophical Society 37:1-130. sediments. Archiv fir Hydrobiologie (Ergebnisse der Lim- 1948. Climate versus fire and soil as factors in post- nologie) 15:1-60. glacial forest succession in the Puget Lowland of Washing- Birks, H. J. B. 1976. Late-Wisconsinan vegetational history ton. American Journal of Science 245:265-286. at Creek, central Minnesota. Ecological Monographs Hemstrom, M. A., and J. F. Franklin. 1982. Fire and other 46:395-429. disturbances of the forests in National Park. 1977. Modern pollen rain of the St. Elias Mountains, Quaternary Research 18:32-51. Yukon Territory. Canadian Journal of Botany 55:2367- Heusser, C. J. 1973a. Environmental sequence following 2382. the Fraser advance of the Juan de Fuca lobe, Washington. Bonny, A. P. 1976. Recruitment of pollen to the seston and Quarternary Research 3:284-306. 802 LES C. CWYNAR Ecology, Vol. 68, No. 4

1973b. Modern pollen spectra from Mount Rainier, catchment. Pages 43-60 in H. J. B. Birks and R. G. West, Washington. Northwest Science 47:1-8. editors. Quaternary plant ecology. Blackwell Scientific, Ox- 1978a. Modern pollen rain in the Puget Lowland ford, England. of Washington. Bulletin of the Torrey Botanical Club 105: Pennington, W. 1979. The origin of pollen in lake sediments: 296-305. an enclosed lake compared with one receiving inflow streams. 1978b. Palynology of Quaternary deposits of the New Phytologist 83:189-213. lower Bogachiel River area, , Washing- Pickett, S. T. A. 1976. Succession: an evolutionary inter- ton. Canadian Journal of Earth Sciences 15:1568-1578. pretation. American Naturalist 110:107-119. 1978c. Modern pollen rain of Washington. Cana- Pickett, S. T. A., and P. S. White, editors. 1985. The ecology dian Journal of Botany 56:1510-1517. of natural disturbance and patch dynamics. Academic Press, 1978d. Modern pollen spectra from . London, England. Bulletin of the Torrey Botanical Club 105:14-17. Ritchie, J. C. 1974. Modern pollen assemblages near the 1983. Vegetation history of the northwestern United arctic tree line, Mackenzie Delta region, Northwest Terri- States including Alaska. Pages 239-258 in S. C. Porter, tories. Canadian Journal of Botany 52:381-396. editor. Late Quaternary environments of the United States. Runkle, J. R. 1985. Disturbance regimes in temperate for- Volume 1. University of Minnesota Press, Minneapolis, ests. Pages 17-33 in S. T. A. Pickett and P. S. White, editors. Minnesota, USA. The ecology of natural disturbance and patch dynamics. Jacobson, G. L., Jr., and R. H. W. Bradshaw. 1981. The Academic Press, London, England. selection of sites for paleovegetational studies. Quaternary Sarna-Wojcicki, A. M., D. E. Champion, and J. O. Davis. Research 16:80-96. 1983. Holocene volcanism in the conterminous United Lamb, H. F. 1985. Palynological evidence for postglacial States and the role of silicic volcanic ash layers in corre- change in the position of tree limit in Labrador. Ecological lation of latest-Pleistocene and Holocene deposits. Pages Monographs 55:241-258. 52-77 in H. E. Wright, Jr., editor. Late Quaternary envi- Leopold, E. B., R. Nickmann, J. I. Hedges, and J. R. Ertel. ronments of the United States. Volume 2. University of 1982. Pollen and lignin records of late Quaternary vege- Minnesota Press, Minneapolis, Minnesota, USA. tation, Lake Washington. Science 218:1305-1307. Scott, D. R. M. 1980. The Pacific Northwest region. Pages Loucks, O. L. 1970. Evolution of diversity, efficiency, and 447-493 in J. W. Barrett, editor. Regional silviculture of community stability. American Zoologist 10:17-25. the United States. Second edition. John Wiley & Sons, New MacDonald, G. M., and J. C. Ritchie. 1986. Modern pollen York, New York, USA. spectra from the western interior of Canada and the inter- Shugart, H. H. 1984. A theory of forest dynamics: the eco- pretation of late Quaternary vegetation development. New logical implications of forest succession models. Springer- Phytologist 103:245-268. Verlag, New York, New York, USA. Mathewes, R. W. 1973. A palynological study of postglacial Sugita, S., and M. Tsukada. 1982. The vegetation history vegetation changes in the University Research Forest, in western . I. Mineral and Hall Lakes. Jap- southwestern . Canadian Journal of Bot- anese Journal of Ecology 32:499-515. any 11:2085-2103. Swain, A. M. 1973. A history of fire and vegetation in north- 1984. Evidence for warm and dry early Holocene eastern Minnesota as recorded in lake sediments. Quater- summers in southwestern British Columbia and adjacent nary Research 3:383-396. Washington. Abstracts of the Eighth Biennial Meeting of Tsukada, M., and S. Sugita. 1982. Late Quaternary dynam- the American Quaternary Association, University of Col- ics of pollen influx at Mineral Lake, Washington. Botanical orado, Boulder, Colorado, USA. Magazine (Tokyo) 95:401-418. 1985. Paleobotanical evidence for climate change Tsukada, M., S. Sugita, and D. M. Hibbert. 1981. Paleo- in southern British Columbia during late-glacial and Ho- ecology of the Pacific Northwest. I. Late Quaternary vege- locene time. Syllogeus 55:397-422. tation and climate. Verhandlungen der Internationale Ver- McKee, B. 1972. Cascadia: the geologic evolution of the einigung fur theoretische und angewandte Limnologie 21: Pacific Northwest. McGraw-Hill, New York, New York, 730-737. USA. Waddington, J. C. B. 1969. A stratigraphic record of the McMinn, R. G. 1951. The vegetation of a burn near Blaney pollen influx to a lake in the Big Woods of Minnesota. Lake, British Columbia. Ecology 32:135-140. Geological Society of America, Special Paper 123:263-282. Mooney, H. A., and M. Godron, editors. 1983. Disturbance Waitt, R. B., Jr., and R. M. Thorson. 1983. The Cordilleran and ecosystems. Springer-Verlag, New York, New York, Ice Sheet in Washington, , and Montana. Pages 53- USA. 70 in S. C. Porter, editor. Late Quaternary environments Munger, T. T. 1940. The cycle from Douglas-fir to hemlock. of the United States. Volume 1. University of Minnesota Ecology 21:451-459. Press, Minneapolis, Minnesota, USA. Munsell Color. 1975. Munsell soil color charts. Munsell Watts, W. A. 1978. Plant macrofossils and Quaternary pa- Color, MacBeth Division, Kollmorgen Corporation, Bal- leoecology. Pages 5367-5380 in D. Walker and J. C. Guppy, timore, Maryland, USA. editors. Biology and Quaternary environments. Australian Oliver, C. D., A. B. Adams, and R. J. Zasoski. 1985. Dis- Academy of Science, Canberra, Australia. turbance patterns and forest development in a recently de- White, P. S. 1979. Pattern, process, and natural disturbance glaciated valley in the northwestern Cascade Range of in vegetation. Botanical Review 45:229-299. Washington, U.S.A. Canadian Journal of Forest Research Winkler, M. G. 1985a. Charcoal analysis for paleoenviron- 15:221-232. mental interpretation: a chemical assay. Quaternary Re- Olsson, I. U. 1979. A warning against radiocarbon dating search 23:313-326. of samples containing little carbon. Boreas (Oslo) 8:203- 1985b. A 12,000-year history of vegetation and cli- 207. mate for Cape Cod, Massachusetts. Quaternary Research Page, C. N. 1976. The taxonomy and phytogeography of 23:301-312. bracken-a review. Botanical Journal of the Linnean So- Wright, H. E., Jr. 1967. A square-rod piston sampler for ciety 45:1-34. lake sediments. Journal of Sediment Petrology 37:975-976. Peck, R. M. 1973. Pollen budget studies in a small Yorkshire

Top View