. .F u ‘ m .. f‘fj‘ . a? . (I'm‘ Ii“ , u‘lt‘_ _ h. . -?.'.; I” A ow... ‘— ~ . -. .‘ _~,' '4! V“ .?~r .‘l‘cm‘f '5: L}? L0 “qr 1:41;. “(A 3% . , EQIHZJE 4“,“...4! Hui" fin” A". "' ~‘ $342135. "3‘7: .4 "‘ "WPQ'AQ .' "'99: I -, IQTFfi , : ‘K’ ‘x 1 r ~. z." .x' ‘2». . '%~Lv 5y... . —‘ Vi ‘ tfifxéfl" .1» _..:v (fa-‘3. ‘lfi. .7 J :- I " 7 ' O‘ . 1 -. ‘31“. ' .u .\ "'5sz ' H‘.‘ : 1*. 'H' .- . .. . ., "fin'; l- ' :A.’ “a. ‘ “' . 5' u l.‘ ,o- F 4‘, , I .m- ‘bnfi‘é: :I“.':‘(LI;':‘«~I.., '1 ‘ " thy“ y—v —— WW —u - 1 u-v ., . MW"; '|'%,".I¥a '3 | V .'i ’IJI 15’::I:‘L “‘43:; .“ I .I- -"lrI' v :I “1:1! '..I...AI . v.5;l, “at" ' ,.\*'.L'). h | . . . 1‘ .. .r- . . ' 1,” ...I A I “fix ”'f’LH‘l l"'.'I: M . ‘. I . . .I ”J: H" 1 II; “I. Jq|.;l.ijl |.::.’LL1;£'.'L;L u! I: I \_L AFAIK é {lg-’2” 11' 5‘13": AA 311293 010632 llllll \lllllllllllllllll I L: This is to certify that the dissertation entitled PATTERNS OF DISTURBANCE AND VEGETATION CHANGE IN THE MIOCENE SUCCOR CREEK FLORA OF OREGON- l DAHO presented by LORETTA S l MMONDS SATCHELL has been accepted towards fulfillment of the requirements for PH.D. Botany 8 Plant Pathology degree in Ralph E. Taggart Major professor Date 8" I 0-83 MSU is an Affirman've Action/Equal Opportunity Inuirution 0'12771 MSU LlBRARlES _:—_ RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. i’ga..!3‘{v ,1 2 l2; I “Jae PATTERNS OF DISTURBANCE AND VEGETATION CHANGE IN THE MIOCENE SUCCOR CREEK FLORA OF OREGON-IDAHO By Loretta Simmonds Satchell A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Botany and Plant Pathology 1983 ABSTRACT PATTERNS OF DISTURBANCE AND VEGETATION CHANGE IN THE MIOCENE SUCCOR CREEK FLORA OF OREGON-IDAHO By Loretta Simmonds Satchell Pollen analysis of 600 m of section in the Miocene Sucker Creek Formation shows three patterns of vegetation change that can be related to two kinds of disturbance. 1. A continuous pollen sequence through 70 m shows a pattern of repeated forest disturbance and secondary succession. Forest assemblages of mixed broadleaved evergreen, broadleaved deciduous, and conifer elements shift abruptly to an elm dominated broadleaved deciduous assemblage, apparently in response to cyclic disturbance that caused widespread damage, particularly to the evergreen elements. Possible disturbances included climate cycles that periodically killed vulnerable trees, or catastrophes such as fire, hurricane, or flood, that destroyed areas of forest. On such cleared sites deciduous species with early successional roles gained rapid dominance, then gradually yielded to mixed broadleaved-conifer dominance. In the time represented by this 70 m section, the forest did not attain steady state. 2. Discontinuous pollen sequences through the remaining section preserve a record of episodic disturbance by volcanism and of primary plant succession on an ash covered landscape. A pioneer herb-shrub complex is succeeded by an alder dominated broadleaved deciduous scrub. Alder quite probably improved the nitrogen content of the mineral ash for the other woody plants that followed, as it does today. Loretta Simmonds Satchell 3. An aquatic vegetation succession is identified which developed concurrently with the terrestrial volcanic ash succession. A submerged, Open water assemblage dominated by algae and aquatic fungi is succeeded by a Glyptostrobus swamp forest in which the aquatics were excluded. Volcanism apparently led to the blockage of drainage systems, creating shallow lakes with opportunities for aquatic and swamp plants. The vegetation was constantly changing. The continuous pollen sequences provide a record of forest disturbances which occurred with a frequency that precluded the attainment of a steady state forest composition. The discontinuous pollen sequences provide a record of volcanism that destroyed the vegetation with a frequency that rarely allowed the development of forest. In both settings of disturbance, broadleaved deciduous trees dominated early successional stands. ACKNOWLEDGMENTS I am deeply grateful to R. E. Taggart, N. E. Good, and P. J. Murphy of the Department of Botany and Plant Pathology, and to A. T. Cross and R. L. Anstey of the Department of Geology for their committee guidance, interest, advice, and manuscript critiques; to A. T. Cross and R. E. Taggart for guidance in every aspect of the study; to W. C. Elsik of Exxon Company, U.S.A. for assistance with identification of some palynomorphs and comments on the manuscript; to R. M. Ritland and J. H. Ritland for field assistance; and to G. K. Satchell for assistance and companionship in the field and laboratory. I thank the Department of Botany and Plant Pathology for financial support in the form of teaching and curatorial assistantships, and the Department of Geology for a grant in support for the 1978 field season. ii TABLE OF CONTENTS Page LIST OF TABLES .................................................. v LIST OF FIGURES ................................................. vi LIST OF PLATES .................................................. vii GEOGRAPHIC TERMINOLOGY .......................................... viii INTRODUCTION .................................................... 1 Previous Studies of the Succor Creek Flora ................. 3 The Present Study .......................................... 4 METHODS ......................................................... 6 Field and Laboratory Methods ............................... 6 Description of the Collecting Locality ................ 6 Measuring and Collecting Methods ...................... 6 Sample Preparation for Study of Palynomorphs .......... 8 Analytical Methods ......................................... 9 Methods of Morphological Study and Taxonomic Treatment of Palynomorphs ........................... 9 Methods of Quantitative Analysis ...................... 10 STRATIGRAPHY .................................................... 14 Sucker Creek Formation ..................................... 14 Stratigraphy .......................................... 14 Structural Setting .................................... 15 Age ................................................... 16 Devils Gate Section ........................................ 17 Structural Setting .................................... 17 Lithology ............................................. 22 Descriptions of Units ................................. 22 Stratigraphic Interpretation .......................... 27 Stratigraphic Relationships of Sucker Creek Sections ....... 36 Lithologic Comparison of Upper Devils Gate Units and Type Section .................................... 38 Summary ............................................... 40 Page THE DEVILS GATE FLORA ........................................... 41 Additions to the Succor Creek Flora ........................ 41 Additions to the Palynoflora ............... ' ........... 41 Additions to the Macroflora ........................... 45 Contribution to Floristic Analysis .................... 46 Interpretation of Inconsistencies of macrofossil and microfossil records ................................. 48 Vegetation Reconstruction .................................. 55 Summary of Characteristics of the Succor Creek Flora.. 55 Floristic and Climatic Comparisons: Mixed Temperate Forests of East Asia ................................ 57 Climate Requirements of East Asian Mixed Forests ...... 61 Floristic and Climatic Comparisons: Mixed Temperate Forests of North America ............................ 66 Floristic and Climatic Comparisons: Miocene Mixed Forests of Oregon and Idaho ......................... 69 Summary ........................... * .................... 71 STRATIGRAPHIC PALYNOLOGY ........................................ 73 Description of the Continuous Sequence Unit I .............. 73 General Patterns ...................................... 74 Dynamics of Individual Taxa ........................... 77 Description of the Devils Gate Units 11 to IV .............. 81 Continuous Sequence (Unit IV) ......................... 81 Disjunct Samples (Units II to IV) ..................... 84 Patterns of Occurrence of other Palynomorph Groups.... 86 Patterns of Disturbance and Vegetation Change .............. 87 Shift from Forest to Herb Assemblage .................. 89 Development of Taxodiaceae ............................ 93 Development of an Aquatic Element ..................... 96 Deciduous-Mixed Forest Dynamics ....................... 98 Characteristics of the Disturbance .................... 99 Natural Disturbances ....................................... 113 Climate ............................................... 114 Catastrophic Disturbances ............................. 119 Pattern and Process in a Disturbed Forest ............. 123 Conclusions ........................................... 126 SUMMARY AND CONCLUSIONS ......................................... 128 The Devils Gate Study ...................................... 128 General Conclusions ........................................ 133 LITERATURE CITED ................................................ 137 APPENDIX ........................................................ 147 iv Table LIST OF TABLES Families and genera recognized in the Devils Gate section ............................................... Temperate mixed forests of Taiwan and Szechuan (China) ............................................... Composition of pollen assemblages from selected levels in the Devils Gate section ..................... Comparison of the composition of the mixed pollen assemblages of Unit I and Unit IV ..................... Some ecological characteristics of deciduous genera that dominate the post-disturbance intervals .......... Page 42 58 78 85 106 Figure 10 11 12 13 14 15 LIST OF FIGURES Page Locations of sampled sections in the Sucker Creek area .................................................... 7 Summary of units measured and sampled at the Devils Gate locality ........................................... 18 Approximate positions of measured and sampled units at the Devils Gate locality ............................. 21 Stratigraphic relationships and present positions of Units I and II .......................................... 29 Alternate interpretations of the Devils Gate section.... 31 Stratigraphic comparison of Unit III and Units IV-V ..... 34 Stratigraphic and paleontologic comparison of the Type section and Devils Gate section units IV-V .............. 39 Comparison of pollen and leaf abundances in Miocene leaf beds of Oregon-Idaho .................................... 50 Topographic and thermal fields affecting distribution of temperate mixed forests of East Asia .................... 63 Pollen diagram from Unit I .............................. 76 Pollen diagram from discontinuous pollen sequences in Units II to IV .......................................... 83 Summary pollen diagrams for the Unit IV Glyptostrobus sequence ................................................ 95 Recalculated summary pollen diagrams from Unit I ........ 101 Summary pollen diagrams of Unit I showing trend in recovery of dominant evergreens (conifers and oaks) following disturbance ................................... 104 Summary of disturbance and vegetation change in the Devils Gate section ..................................... 132 vi LIST OF PLATES Plate Page 1 View of the Devils Gate locality showing locations of sampled units ...................................... 20 2 View of Devils Gate Units III, IV and V ............... 33 3 Additions to the Succor Creek palynoflora ............. 153 vii GEOGRAPHIC TERMINOLOGY The Miocene flora and fauna and the creek for which they were named were spelled either "Succor" or "Sucker" prior to 1960 when the name "Sucker Creek" was officially adopted by the State of Oregon. Kittleman et al. (1965) used "Sucker" to name the formation exposed in the area, and Graham (1965) to publish a revision of the fossil flora. In 1966 the State Geographic Names Board reversed its earlier official ruling, changing it to the "Succor" form. "Sucker" must be retained for the formation, has been used for the fossil flora (Taggart 1971, 1973; Taggart and Cross 1974), and continues to be used for the fossil fauna (Shotwell 1968, Tedford et al. in preparation), but "Succor" has recently become the preferred spelling for the name of the fossil flora (Niklas and Gianassi 1977 and later, Taggart and Cross 1980 and later) and will be used here for consistency. viii INTRODUCTION Late Tertiary paleobotanical studies in the Pacific Northwest have documented broad trends in the modernization of the vegetation of the region. Evolutionary models of vegetation change and stratigraphic and climatic interpretations of the fossil floras have mostly been based on the assumption of long term climatic control of vegetation and thus on the assumption that the plant fossil assemblage preserved at any point in time is one of steady state, climatically controlled, climax vegetation. Studies of Quaternary and recent vegetation history show ubiquitous patterns of disturbance and of continual vegetation change on the time scale of plant succession that have led to the conclusion that long term stability may be the exception and episodic disturbance the rule in the history of vegetation (e g., Brubaker 1975, Davis and Webb 1975, Miles 1979, White 1979, Davis 1981). Extrapolated to pre-Quaternary forest vegetation, this concept has implications for models of evolution of Tertiary vegetation and for the biostratigraphic and climatic interpretations of Tertiary fossil plant assemblages. A test of the concept of intrinsic disturbance in the pre- Quaternary history of vegetation requires the preservation of a strati- graphically continuous plant record in a sedimentary setting that allows resoTution of vegetation change on a time scale of plant succession. 2 In the late Tertiary volcanic terrains of the Pacific Northwest pollen and spores provide the most abundant, continuous, and widespread plant fossil record. Many of the Neogene volcanic sections in the region were deposited very rapidly and are barren of plant fossils. In occasional sections, fine volcaniclastics were reworked into lake beds at a rate that allowed the incorporation of a more or less continuous pollen record with resolution of vegetation change on a time scale of plant succession. In such sequences in the Miocene Sucker Creek Formation of eastern Oregon and western Idaho, stratigraphically controlled palynologic studies have shown a record at disturbance and plant succession in the Succor Creek flora (papers by Taggart and Cross). The present study involves the stratigraphic palynology of 600 m of volcaniclastic section in the Sucker Creek Formation. The results provide new insight into the patterns of both cyclic and episodic disturbances of the Succor Creek flora and into the processes of post-disturbance vegetation succession. These findings contribute not only to ecological theory and models of vegetation evolution but also to biostratigraphy and the reconstruction of paleoclimate. Ecological and evolutionary assumptions form the basis of the application of fossil plant assemblages in biostratigraphy and in the reconstruction of paleoclimate. Thus, insight into the ecological significance of plant assemblages and the history of disturbance enhances effective paleoclimatic and biostratigraphic interpretation of late Tertiary plant assemblages in the Pacific Northwest. 3 Previous Studies of the Succor Creek Flora The Succor Creek flora is one of the best studied in the region, including many contributions to classical macrofossil taxonomy and more recent intensive stratigraphic palynological studies. Historically, the Succor Creek flora is a complex of isolated florules collected from volcanic strata which outcrop in the Oregon-Idaho border area. Contributions to the macroflora include Knowlton (1898), Berry (1932), Brooks (1935), Arnold (1936a, 1936b, 1937), Smith (1938, 1939, 1940), and Chaney and Axelrod (1959). Graham (1963, 1965) undertook a thorough systematic revision of the flora based on approximately 2500 specimens from 15 localities. Included in the study was an examination of the leaf-bearing matrix for the occurrence of pollen and spores. This pioneering palynological study added an important dimension to our understanding of the flora by showing the presence of a significant conifer element in the pollen flora. Pigga was shown to be the most abundant conifer pollen type, averaging 20% of total pollen and spores. Able; and Isuga were also found to be consistently present. Shah (1968; Smiley, Shah, and Jones 1975) extended the regional record of the flora with collections from leaf-bearing localities in the Weiser area of southwestern Idaho assigned to the Sucker Creek Formation. Taxonomic studies have generally considered the individual florules to represent a single contemporaneous flora consisting of a diverse deciduous forest complex growing under warm temperate conditions with moderate, evenly distributed rainfall. An approach to study of the taxonomic affinities of some elements of the flora has recently been undertaken. Using well preserved material it has been possible to compare cell ultrastructure and 4 chemistry of Miocene Succor Creek and living leaf genera (Gianassi and Niklas 1977, Niklas and Gianassi 1978, Niklas et al. 1978). Paleoecological studies of the Succor Creek flora have been undertaken over the past decade by Taggart and Cross (Taggart 1971, 1973; Taggart and Cross 1974, 1980; Taggart et al. 1982; Cross and Taggart 19831). The approach has involved continuous sampling of stratigraphically sequential sections for palynological study and collection of macrofossils from each horizon containing leaves, seeds, or wood. These studies have suggested patterns of vegetation dynamics of the flora in response to short term climatic oscillations, and to volcanic hydrologic, and fire disturbance. The Present Study The updated taxonomic treatment and stratigraphically-controlled palynologic perspective of the Succor Creek flora contributes signifi- cantly to the interpretative framework of the present study. The site for this study was selected because it provides a stratigraphic and topographic extension of previous studies in the formation. The measured and sampled section exceeds 600 m and is thus by far the thickest studied to date from the Sucker Creek Formation. The thickness of the section enhances the possibility of obtaining detailed data on the patterns of vegetation change and disturbance. The site is 10 km north of the type locality, previously the most northerly section documented by palynology, and thus provides a significant extension ¥ 1Many references will be made to these studies. Where cited collectively, they will be referred to as "papers by Taggart and Cross" for brevity. 5 of the study area towards the Snake River. The site is here referred to as the "Devils Gate locality". The study of the Devils Gate locality had the following objectives: 1. Describe the depositional units exposed at the Devils Gate locality and interpret their relative stratigraphic positions. 2. Identify the micro- and macrofossils preserved at the locality and compare the paleoflora with modern floras. 3. Document the stratigraphic palynology of the Devils Gate sequence. 4. Interpret the patterns and processes of vegetation change and the incidence and nature of disturbance, and derive insights for models of evolution of Miocene vegetation, and for biostratigraphic and paleo- climatic interpretation of Miocene plant assemblages. METHODS Field and Laboratory Methods Description of the Collecting Locality The Devils Gate section (Figure 1) is exposed on both sides of the Succor Creek Road, approximately 15 km south of the junction with Oregon State Highway 201 (Idaho 19). It is 10 km north of the type locality and 8 km north of the campground in the scenic Succor Creek Gorge. The location is in the Owyhee Ridge Quadrangle, Malheur County, Oregon (Sec. 31 51/2 and Sec. 32 31/2 R.46 E. and T.23 S.). It is referred to as the Devils Gate locality in this paper for the nearest named landmark, a prominent gap in the overlying Jump Creek Rhyolite situated about 1.2 km northwest of the site. Measuring and Collecting Methods Field work was conducted in the late summers of 1977 and 1978. The surface of most of the exposed sequence was covered with weathered bentonitic ash, the thickness varying with the slope. Continuous trenches were excavated to expose relatively less weathered rock for comprehensive sampling of the section. Each lithologic unit was Sampled. Where lithologic change was imperceptible through extensive thicknesses, the unit was sampled in approximately 1.6 m increments. In three portions of the section particularly deep weathering precluded Oregon idoho Devils Gate ‘. Devils Gate 9; Section ‘ Malheur Co. Oregon Owyhee Co. Idaho -<< 9 Shortcut Section \ / .’ .. Rockville School . \ Rockville Section .‘ o’ \ ‘ / :., Whiskey Créek{$ Section / \ ,J I /”' Sheoville. Figure 1. Locations of sampled sections in the Succor Creek area (after Taggart and Cross 1980). 8 the excavation of continuous sampling trenches. In these sequences holes were dug to somewhat less weathered rock. Samples were taken and a lithologic description attempted at measured intervals. Two hundred and fifty samples were collected from 630 m of section for palynological analysis. In the course of measuring and sampling the stratigraphic sequence, stumps were found at two horizons and two leaf beds were exposed and collections made. Fossil bone fragments were observed at seven horizons which were traced along the outcrop in a search for additional faunal remains. Sample Preparation for Study of Palynomorphs Initial trials were conducted to develop a standardized processing scheme which could be used routinely, with little modification. The aim was to develop a technique that would effectively macerate the range of volcanic rock types encountered in the sections while providing as uniform a processing schedule as possible. Cold HF treatment (after Taggart 1971) was employed to avoid possible destruction of palynomorphs caused by high temperature in the reaction of concentrated HF with these highly siliceous volcaniclastics. After HF treatment most samples required repeated washes in hot HCl to remove fluorosilicates. Many productive samples needed no further treatment except staining. A few samples containing excessive quantities of fine organic detritus were treated briefly with bleach prior to staining. Details of Treatment. 1. A 20 or 30 9 sample was crushed until the largest fragments were less than 10 mm. 9 2. A fragment was tested with HCl and the few samples containing carbonates received a preliminary treatment in 10% HCl for 4 to 8 hours. These were subsequently washed until all trace of the acid was removed. 3. A plastic beaker containing the crushed sample was placed in an ice water bath and crushed ice was added to the dry sample. Reagent grade HF (49%) was added very slowly and intermittently to keep the temperature of the reacting mixture cold. The sample was left to stand in HF for about 24 hours with frequent stirring for the first 8 to 12 hours and was then allowed to settle overnight. 4. The sample was centrifuged and washed repeatedly in hot, dilute HCl (10%) until all the fluorosilicate gel was removed. Need for subsequent treatment was determined by microscopic examination of the sample. 5. Fine, organic detritus was reduced by treatment with commercial bleach (5.25% NaOCl) until a color change was observed. This was followed by several washings. 6. Residues were stained with 1% safranin and stored in H E.C. (hydroxyethyl cellulose WP-O9, Union Carbide Corp ). 7. Four to six slides were prepared from each productive residue using H.S.R. (Harleco synthetic resin) as a mounting medium. Analytical Methods Methods of Morphological Study and Taxonomic Treatment of Palynomorphs Most palynomorphs were assigned to living genera or families on the basis of comparison with modern reference material and published atlases. 10 Several atlases of north temperate airborne pollen with good generic descriptions and illustrations provided a useful supplement to the reference slide collections for this study: Regional atlases - Martin and Drew 1969; Richard 1970 a,b; Huang 1972; Adams and Morton 1972, 1974; McAndrews et al. 1973; Lieux 1980 a,b, Lieux and Godfrey 1982. Systematic treatments - Ferguson 1977 (Cornaceae); Helmich 1963 (Aggr); Sivak 1973 (Isuga), 1975 (bisaccates), 1976 (Cathaya); Stone and Broome 1975 (Juglandaceae). Two or more distinct morphotypes of some genera were found and each was assigned informal identifying codes for counting purposes. Some pollen types could not be identified. In part, these unknowns are probably due to the limits of scope of the search for modern equiva- lents. Further reference study and more detailed morphological study including scanning electron microscopy should resolve some of these taxonomic problems. Other unknowns probably represent the pollen of extinct taxa. All unidentifiable pollen types were found to be quantitatively insignificant in the subsequent pollen analytical study. Photographs were taken of all pollen and spore types to document the range of variation of each type for constant reference during pollen counting. Succor Creek palynomorphs have been adequately treated systematically (Graham 1965, Taggart 1971, 1973); only the few new Devils Gate forms are described and illustrated here (Appendix and Plate 3). Methods of Quantitative Analysis Sampling and Counting. Counting was carried out along regular traverses of the slides at a magnification of 950x. Traverses were 11 positioned evenly over the slide, and counts were divided between several slides. Five hundred palynomorphs of vascular plants were counted unless this was not possible due to the low productivity of a sample. Algal bodies and fungal spores were tallied outside of the basic pollen sum during the counting procedure. All the interpretations in a study of this sort obviously depend on accurate and consistent pollen identification. Care was therefore taken to assign to each pollen type a generic name, where an earlier assignment had been made, and also to a specific photograph. This allowed subsequent reassignments as more grains were examined and as observations and interpretations became more critical with experience. Where the essential morphological features of a grain were obscured it was assigned to one of several "indeterminable" classes indicating various levels of uncertainty. Morphologically determinable types which could not be assigned to a modern taxonomic group were referred to coded categories for counting purposes. Presentation of Pollen Spectra. The basic pollen sum was the total number of higher plant spore and pollen grains counted in a sample. The relative abundance of each pollen type was calculated as a percentage of the basic pollen sum. Pollen and spores of aquatic plants, algae, and fungi were excluded from this sum because their production is local and more or less independent of the factors influencing the sum of terrestrial plants. Their abundances were calculated as percentages of the basic sum plus their own sum, following the practice of Quaternary palynologists (Birks and Birks 1980, p. 168). 12 Relative abundances of groups of taxa were also calculated to elaborate broad trends through the section. Relative abundances were used to plot the pollen diagrams. Interpretation of Relative Pollen Diagrams. The paleoecological interpretation of pollen diagrams is complicated by differential processes of production, dispersal, deposition, and preservation of pollen. These processes have been the subject of intensive study by Quaternary paleoecologists. A thorough recent review has summarized the significance of some of the principal factors (Birks and Birks 1980). Paleoecological reconstructions in this study are made within the accepted constraints of this body of information. Through a single sequence of lake sediments, pollen assemblages are more or less controlled for many of the variables of pollen production and sedimentation so that major shifts in pollen floras through the sequence can be interpreted with some confidence as shifts in vegetation distribution or composition. Studies from lakes attest to the validity of pollen records for the purpose of reconstructing dynamics of vege- tation through time (Janssen 1966, 1967, Davis et al. 1971, Webb 1974, Brubaker 1975). In a lake sequence, a major potentially distorting influence must be considered, that of local pollen input from lake margin vegetation which can change through time under local hydrologic control independently from the regional vegetation. This potential problem in the Devils Gate setting will be evaluated before interpre- tations are made of the pollen dynamics of continuous pollen sequences. The significance of differences in pollen assemblages from different basins in space and time is much more uncertain than 13 interpretation of changes through time in a single depositional setting, and comparisons made with other Succor Creek sections will be in very general terms. STRATIGRAPHY Sucker Creek Formation The Sucker Creek Formation is in the Owyhee Upland physiographic region of the southern half of eastern Oregon. To the west and south are the northernmost extension of the Basin and Range and the High Lava Plains regions and to the east is the Snake River Plain, Idaho. The Owyhee Upland forms a broad plateau deeply dissected by many streams that drain toward the Snake River. Stratigraphy Surface Section in Southeastern Oregon. The Sucker Creek Formation, described by Kittleman et al. (1965), includes volcaniclastic sediments exposed along the general course of Succor Creek, in the vicinity of the Oregon-Idaho border. Kittleman described the type section of 180 m which was extended to about 200 m by Taggart and Cross (1980). The base of the unit is not exposed but Kittleman et al. (1965) estimated a total thickness of about 490 m. The formation lies unconformably beneath the Miocene Owyhee Basalt (equivalent of the Upper Yakima Member of the Columbia River Basalt) in the region of the Owyhee Reservoir and beneath the Pliocene Jump Creek Rhyolite (Kittleman et al. 1965) (equivalent of the Idavada Volcanics) to the east of the reservoir in the area under study. 14 15 Subsurface Extension into Southwestern Idaho. Clastic Miocene beds in southwestern Idaho have been considered to represent the Sucker Creek Formation (Shah 1968, Smiley et al. 1975). Outcrops are limited and are separated from Sucker Creek area at the surface by younger rocks in westernmost Idaho. Some authors, while considering the Idaho outcrops equivalent in age and stratigraphic position, mapped them separately as the Payette Formation (Newton and Corcoran 1963). While the outcrops in Idaho are limited, there is evidence from recent wells of an extensive subsurface distribution of the Sucker Creek Formation. Deep test holes in Owyhee County, Idaho (Warner 1977) indicate that Miocene volcani- clastic sediments (assigned to the Sucker Creek Formation) thicken markedly in a short distance north-northeast of the studied outcrops to an estimated 2000 m beneath the western end of the Snake River Plain. Structural Setting Miocene. The thickness of the Sucker Creek Formation requires a structural setting that provides for the enormous volume of volcani- clastic sediment and the rapid subsidence of the basin to accommodate up to 2000 m of sediment. The basin axis is interpreted to occur along the line of a rift zone (the Idaho rift) that was developing as a rapidly downwarping graben basin during Sucker Creek time (Warner 1976, 1977). The studied outcrop sections are placed along the southern flank of the basin with the Devils Gate section the nearest of the studied sections to the axis of the basin. Isopach projections indicate that the total thickness of the formation in the Devils Gate area is between 1000 and 2000 m. Alternate interpretations are reviewed by Cross and Taggart (1983). 16 Pliocene-Pleistocene. In the area of thickest development the Sucker Creek sediments are now deeply buried. The top of the formation appears to plunge to a maximum depth of about 3400 m below sea level along the line of the rift, not more than 50 km from where they outcrop at an elevation of 900 m above sea level. Such displacement indicates substantial post-Miocene downwarping or block faulting along the Plio-Pleistocene rift. In the area of downwarping the formation is overlain by late Miocene rhyolites and Pliocene to recent sediments (Warner 1977). Age Radiometric dates have not been successful in clearly establishing the age of the Sucker Creek Formation. An age of about 16 m.y. is tentatively assigned to the sediments, which is consistent with a Barstovian mammal age (Shotwell 1968) and places the formation in the uppermost Early Miocene on the Berggren (1972) Cenozoic time scale. The radiometric age assignment is based on dates from clastics within the formation and from the overlying basalt. A vitric tuff from the Type section taken 119 m above the base (Kittleman, personal com- munication to A. T. Cross), is dated at 15.4 m.y. (sanidine) and 18.5 m.y. (glass shard) (Kittleman in Laursen and Hammond 1974). A basalt, thought to outcrop in the vicinity of the Whiskey Creek section, is dated at 16.7 m.y. (K/Ar) (Evernden and James 1964, and written communication to A. T. Cross). Collecting site and stratigraphic relationships of the basalt are unknown. The overlying Owyhee basalt is dated at 14.4 m.y. based on an average of 4 flows (Bottomly and York 17 1976), 13.1 to 13.9 m.y. on the basis of 16 successive lava flows (Watkins and Baksi 1974), and 24.6, 15.2, and 14.4 m.y. (Kittleman in Laursen and Hammond 1974). Radiometric dates and macrofossil data indicate an older age for some outcrops east of the Whiskey Creek section. K/Ar dates between 36 m.y. and 22 m.y. and Sr/Rb dates with a mean of 22 m.y. are reported for leaf-bearing tuffs that also have distinctive floral components (Niklas and Gianassi 1978). Preliminary macrofossil data indicate an older age (possibly Oligo-Miocene) for exposures in the Coal Mine Basin now under study by Taggart and Cross. Devils Gate Section Structural Setting The Devils Gate section is exposed in a 7 unit series of en eschelon fault blocks (Figure 2, Figure 3, Plate 1). The faults between the blocks extend more or less northwest-southeast. The blocks dip generally from southwest to northwest and the altitude of the strata vary from almost horizontal to about 20°. Internally, the blocks may be warped so that the dip varies across a single bedding plane. The prominent cliff (Unit V, Plate 1), for example, is essentially horizontal on the south end and dips 11° NNW at the north end, indicating that the block has dropped and rotated to the north. Internally, the blocks may also show minor faults that displace a part <>f the unit as on the south end of Unit V, for example. When erosion (accurs along such a fault line, small valleys may develop and displaced lJoerons of a unit may become isolated. These small-scale structural 18 Figure 2. Summary of measured and sampled units at the Devils Gate locality. Number of samples collected in each unit, positions of productive samples, bone fragments, leaf and stump horizons and some lithologically distinctive beds are shown (modified after Taggart et al. 1982, fig. 1, p. 539.) 19 Meters Uni t/SamPleS = Pollen 1 Coll. OBone flmmjmm Jump Creek 7 Rhyolite ///// /A C l d - . .2213: VII, 4(0) Wflz 600 _ VI /30(0) gm 500 _ v /72(2) - i ----- ° 2:." Iv/25(9) ; 400 _ 7// Me“ W /% 300 _ 200 _ III/51(5) _ ‘ £21 100 _ ‘- White 11 /32(4) ; gimme “”mmfieim' 1/35(31) g y o 2 fl " W % 20 .mcwyzxm mg» mecow ucm compomm mcu mamo mproxcm xmmco aszw asp .pmoz mcmxoou .mpwcz um_qsmm co mcowpaoop m:P3o;m apppmqu wumw m_w>mo mcp we zww> .H mum—m . é... . .. nos... . , {a , . 1.: Whmyjrzn , . . .. ”Helix 4, .. _. . M7. . .«r a»: ... .. #9.... , . .355 , . z....... . .27.... w . . ,. . 4......» . .gmu.ar....&.r..... . Ls .. JDIIH‘. 44.1001. . h . .. r0... 4.11...." ., .... .. .c.ao... ...... a . Wows. . ...... .. «“thpr ”Hwy... .4 v - OI. . v .. , , mow... tripe??? , .L ...: . . . . 21 .Aooo.em\flv mans ownamcmoaou Xm>c=m .Pomw .m.:--mpmcmcum:o Havoc ucmxm>mcw .Nm .umm new mpmcmcumao mmuwm mmczgo .Hm .umm co acmemmcmpcm .xumpmoop mama mpw>mo may we mama: empaEmm ncm cocammos mo meowummoa mumewxogaa< .m mczmwu a.ma — P P b FL 38353 :23: Jr! 73.x-.. 12% 2' .....K 33. got... .53... SJ 338 3 J 2 I 22 features of the section make it difficult to measure and establish relative positions and stratigraphic correlations of the 7 units. Lithology Sediments of the Devils Gate section include massive yellowish-gray to drab-olive volcanic sandstones which often contain pumiceous frag- ments and glass shards; fine, drab-olive to brown tuffs, buff platy tuffs, white opaline shales, occasional thinly-bedded paludal lignitic shales, and calcareous shales. Both fluvial and airfall volcanic sediments are present. Depositional environments recognized at the Devils Gate site include fluvial, lacustrine, and occasionally paludal. The more massively bedded, coarse, light-colored, volcanics are relatively resistant to erosion and form bare, low cliffs. Darker, fine, tuffs are poorly indurated and form low, deeply weathered, slopes. The weathered bentonitic clays swell when wet and shrink when dry to give a characteristic granular ("popcorn") surface to the outcrops. Descriptions of Units Seven exposed units with a combined thickness of 629 meters have been measured, described, and sampled (Figure 2). Intervening concealed section adds a minimum of 106 meters to the section making a total measured thickness of about 735 meters, perhaps 183 meters less, if a duplicated section is demonstrated as will be discussed later. The units lie on an approximately east-west transect along 1.75 km of northwest trending faulted terrain. The base of the measured section lies to the east. The upper contact is with the overlying Pliocene Jump Creek Rhyolite at the western edge of the area under consideration 23 (Figure 3 and Plate 1). The sampled units are exposed primarily on east to northeast facing slopes with beds dipping generally northwest to southwest. The measured units are numbered from the base of the section, in the order of their inferred stratigraphic positions. Generalized descriptions of the sampled units will precede a stratigraphic interpretation of the 7 units. Unit_1. The base of Unit I is concealed. The exposed sequence of 74 m faces northeast. It consists of alternating series of dark (brownish-gray and grayish-brown) weathered claystones and siltstones and light (pinkish-gray to pale yellowish-brown) indurated platy claystones and siltstones that weather to an even lighter color on the surface. At about 12 m above the base is a macrofossil zone yielding leaves, seeds, twigs, and fish scales. Throughout several meters of light colored platy beds above the main leaf zone occasional Glyptostrobus fragments are found. Permineralized wood fragments are found weathering out at two levels, about 24 m and 39 m above the base. The uppermost bed of this unit is a distinctive light gray, silvery, poorly consolidated, granular ash (6 m) containing purplish-tinted plant fragments, mostly twigs. Unit I is distinguished in the Devils Gate section by the alternation of deeply-weathering, dark beds and resistant, indurated, platy, light-colored beds. The sequence is mostly fine-grained and is extraordinarily productive of pollen. It appears to have been deposited in an uninterrupted lacustrine setting. Unit II. The base of Unit II is marked by a massive, light gray, silvery, poorly consolidated ash with lenticular inclusions rich in 24 purplish twig fragments. The bed is apparently indistinguishable from the uppermost bed of Unit I. The measured block is exposed on a southwest- facing slope and is internally faulted with 24 m of repeated section recognized by marker beds. The unit consists of 52 m of non-repeated section. Above the gray ash about 30 m of distinctive white and light- gray opaline shales occur. Within the evenly and thinly-bedded opaline series are found some thin dark earthy sandstone layers. The upper 14 m consists of yellowish-gray to pale olive volcanic sandstones and siltstones which become increasingly pumiceous towards the top. A few bone fragments were recovered in the upper third of the unit. Unit II is distinguished in the Devils Gate section by the basal ash and the white opaline series. This sequence can be traced to the northwest across the Succor Creek road where they are useful in relating Units I and II to the younger units in the section. The basal ash probably represents an airfall. The evenly bedded opaline series appears to be a lake deposit and the overlying volcanic sandstones are mostly fluvial. Only 4 of 32 pollen samples were productive. These occur at irregular intervals through the unit. Unit III. The base of Unit III is marked by a north trending, steeply dipping, fault plane that juxtaposes Unit III against the white opaline series of Unit II to the south of the measured face described above. The lowest 37 m of Unit III was measured from the fault line in a southwesterly direction across a saddle to a stump horizon. The attempt was made to trace the stump bed along its strike (mostly concealed but inferred horizontal extension) northward across the small valley, using intermittent wood fragments weathered out on the surface 25 to mark its position. In the north end of the valley the (inferred equivalent) wood-bearing horizon was taken as a starting point to continue measuring the section in a west southwesterly direction across a deeply weathered, low slope for 65 m stratigraphically, and up a steeper, less weathered, rock face for another 85 m. The total measured section was 184 m thick. The low slope consists of deeply weathered, banded, olive to olive-brown to yellowish-brown, sandstones and siltstones, some pumiceous, and most poorly indurated. The cliff face comprises a series of light yellowish-gray thickly bedded, volcanic sandstones and siltstones, some flecked with angular fragments of yellow pumice or dark glass shards, and several prominent ledge-forming beds. Bone fragments weather out along the strike of a horizon near the cliff base. The upper series consists of drab yellowish-gray and light olive-gray sandstones and siltstones, often pumiceous. The section is concealed by deep weathering and vegetation to the top of the hill. Apart from some swamp and lacustrine beds in the lower third, most of the unit is probably fluvial. Unit III yielded 5 productive pollen samples from the lower 40 m of section at the stump zone and above. Unit IV. Unit IV is informally named the "Glyptostrobus unit" for the Glyptostrobus stump and leaf beds near its base. It is separated from Unit III by an uncertain amount of covered section on both sides of the main road. Unit IV consists of dark, organic siltstones and sandstones, largely concealed by deep weathering on very low slopes. The unit (as estimated) is 55 m thick from the base of the exposed beds below the stump horizon to the base of the green ash series of Unit V. 26 Unit_y. The unit forms a broad cliff that faces approximately east and is clearly seen from the Succor Creek Road. Thickness, measured from the base of the green ash series, to the covered section at the top of the cliff is 116 m. Above the green series (10m), Unit V was measured up the cliff face through a sequence of light-colored, mostly yellowish-gray, volcanic sandstones including two ledge forming horizons (22 m); series of mostly coarse volcanic sandstones often with pumice fragments in colors of dusky yellow and light olive-gray (47 m); and an upper light colored series with persistent ledge forming horizons (37 m). Fossil bone fragments weather out at several levels in the lower third and at one level in the upper third. Only 2 samples of 72, from a zone at about 74 m from the base, yielded pollen. Most of the sediments were apparently rapidly deposited by fluvial processes and are little altered, coarse volcaniclastics. Unit VI. Unit VI can be seen looking west from the top of Unit V. It is a down-faulted block and the amount of concealed section at the base separating the exposed sequence from the top of Unit V cannot be determined. The outcrop consists of 61 m of volcanic sandstones and siltstones with several ledge forming pumiceous sandstones. It has no distinguishing lithologic features. The top is covered. All 30 pollen samples were barren. Unit VII. The unit was measured from the covered section in the saddle above Unit VI up to the Jump Creek Rhyolite that unconformably overlies the Succor Creek Formation in the area and forms the prominent ridge across the skyline (Plates 1 and 2). Unit VII is comprised of 27 88 m of mostly yellowish-brown, coarse volcanic sandstones and siltstones. The four pollen samples were barren. Stratigraphic Interpretation Relative stratigraphic positions are reasonably clear in the several blocks of measured section and, for the most part, sampling gaps occur where strata have been buried by down-faulting or where strata are concealed by deep weathering on slopes with low relief. The amount of missing section is uncertain. The major stratigraphic problem is the possibility of repeated section in Units III to V. In the following discussion I will describe the stratigraphic uncertainties in detail and present an alternative Devils Gate section. However, it should be men- tioned here, that because of the pattern of occurrence of samples with significant pollen content these stratigraphic uncertainties appear to be of little consequence in the interpretation of vegetation dynamics through Devils Gate time. Relationship of Unit I and Unit II. In the field, the outcrops of Unit I and Unit II are adjacent with a small valley separating them. The two blocks dip in different directions and beds cannot be traced across the small valley. A stratigraphic relationship of the two adjacent blocks can be reasonably inferred lithologically. The lowest beds in Unit II are lithologically distinctive, including a light gray or silver gray ash bed and overlying white opaline shale series. Such a sequence of beds is found at the top of Unit 1, indicating that Unit II is stratigraphically above it. If this is correct, the valley between the 2 blocks may mark the position of a fault line along which the two 28 blocks were vertically displaced, laterally separated, slightly rotated, and eroded, to bring them into adjacent positions with differently dipping beds. (Figure 4). Repeated Section on Unit II. Where a minor fault has occurred on the face of a tilted block, the faulted portion may have slumped resulting in a repeated sequence of beds. Such a situation is found in Unit II where 28 m of section is repeated. In this case, marker beds, interpreted as repetitive, have been used to make the stratigraphic correlation. Stratigraphic Position of Unit III. The relative position of Unit III is the most uncertain. The exposed block is bracketed by a major fault plane below and by concealed section above. At the base, a fault plane, which extends approximately northward above the opaline shale beds, juxtaposes Unit III against the eroded (?) surface of the opaline beds of Unit II. A concealed section and possible unrecognized faults occur between the top of Unit III and the Glyptostrobus stump zone at the base of Unit IV. It has not been possible to trace Unit III to the northwest across the Succor Creek Road. Two independent sets of measurements of the section and lithologic comparisons of Unit III with Units IV and V suggest that Unit III may be repeated in part in Units IV and V in exposures to the north of the road. Although Unit III has not been located west of the road, the lower Devils Gate section (Units I and II) can be related to the upper part of the section (Units IV to VII) by the white opaline series (of Unit II) with its basal silver ash bed. This pair of distinctive 29 A. INFERRED DEPOSITIONAL SEQUENCE Unit 11 SILVER ASH Unit I B. POST DEFORMATION SECTION Figure 4. Stratigraphic relationships and present positions of Units I and II. 3O lithologic units can be traced continuously along the outcrop across the road into the valley below the Glyptostrobus stump zone. This correlation allows a second measurement of the interval between the silver ash bed (base of Unit II) and the Glyptostrobus stump bed (base of Unit IV) at a location several hundred meters north of the section just described (Figure 5). The brunton-pace traverse measurement of 177 m for the section to the north indicates a consistently thinner interval than the measurement of Unit III to the south which is 324 m. This difference of 183 m between the top of the silver ash bed up through the opaline beds to the base of the Glyptostrobus stump layer is too great to be explained by sedimentary deposition. In these measurements, consideration must be allowed for uncertainty concerning the presence, location and magnitude of concealed faults, changes of dip of strata, and the possibility of repeated section in concealed strata. However, the magnitude of this discrepancy (147 m) and its approximation to the thickness of Unit III (183 m), require an alternate interpretation of the section such as in Figure 5C where Unit III is excluded and is considered to be a duplicate of a large portion of Units IV and V. This interpretation of repeated section based on comparison of the two measured sections is further supported by a lithologic comparison of Unit III with Units IV and V. Several similarities are found. The two sequences (Figure 6 and Plate 2) have a basal horizon of silicified stumps or logs that outcrop in drab-olive and brown banded, deeply weathered, poorly exposed valley sediments. At about 65 m above the stump zones the two sequences are comprised of prominent cliffs with indurated light-colored, yellowish-gray, coarse, pumiceous sandstones. Both have distinctive 31 Figure 5. Alternate interpretations of the Devils Gate section. A. Total measured section Units I to VII. B. North section between silver ash bed and stump zone. C. Interpreted section with Unit III equivalent to Units IV-V. meters 600 - Green Ash I 400 - 200 " white poxcelanite beds silver ash bed 32 Jump Creek Rhyol l te ¢: )\ Gl- tostrobus Beg III ------d \ ,— Ilfl VII VI ///// Green Ash .... —-—q Glyptostrobus Bed \S \\\\\\\\ r—-J Jump Creek Rhyolite III white porcelanite beds silver ash bed W Concealed Section 33 A.m:owpumm owcamcmvbmcum com o wcsmwm mmmv .mmumF ucm mums «con co mcompmqu ozu mpmowucw cmpcmu Law: mzoccm wmcoN asapm mcp to mcowpwuor 03p wmeTUcw pcmwc pm mzocc< .> new .>H .HHH mute: mama m_w>mo co zmc> , e . n . . JOIw ._ v . ” .....vw. J i.v ,u . .... (a... .N mom—a 34 cw mmogu op mcwuconmmccou mums cmxcms op L040; mzocc< .>->H mews: ucm HHH awe: co comwcmasou uwgamcmwumcpm 0:2 953. J\ 3:058» A022. 9.7.0500: c2300.. ‘09:: f3 333—0230....» 0230 6223.23: 030.. o:..:oc-cc._u o. one-g0 \SS \ HN~ awe: .N aba_a 0:3 3:3 ®< 3:03.03 A022. 9.20.3.0: 3300.. 009:... ...-o quu-pficflul—ofll 0ac~ou .uoco_ou-ugo_. nap-teétpu 3 09.28 0090. «c058: 0000. 00900. l 25. \‘I :005 0098: \\ > a >~ muwcz .o mczmwu 1 0m 11 00H 1 nfia mcmumz 35 ledges at about 96 meters. Bone beds are present in this part of each sequence. However, two distinctive features of Units IV-V have not been found in Unit III. One is a Glyptostrobus dominated leaf bed found above the stump zone in Unit IV which may be due to a slight difference in swamp environment between the two locations. The other is a conspicuous lithologic feature well exposed at the prominent Unit V cliff base at about 55 m. Bright green ash beds occur at this level which, as will be discussed later, may have correlative value even beyond the Devils Gate locality. They have not been found in Unit III but since the supposed equivalent horizon is deeply weathered and extensively covered and has been sampled only at intervals where some exposure by digging was successful, there is some possibility of their occurrence in the unit. One green-colored sample of somewhat comparable appearance and lithology was collected from Unit 111 at 32.4 m. The palynology of the two units has not produced sufficient evidence to resolve this correlation question. In both sections, the lower beds sampled were deeply weathered and samples were taken only at intervals where less weathering and vegetation cover permitted excavation to sufficient depth to obtain satisfactory samples. Thus, comparable or continuous sets of pollen samples were not obtained and, as demonstrated elsewhere in this paper, pollen assemblages may change markedly in very short vertical intervals through a stratigraphic unit. Nevertheless, two features of the pollen profiles compare favorably. First, the general composition of the mixed forest pollen assemblages is similar in both sets of samples. Second, the composition of both sets of samples are distinctly different from those in the underlying Unit I, 36 particularly in having high abundances of Abies, Tsuga, and Alnus, and in the absence of Lithocarpus. Third, the sequence of pollen assemblages is comparable: Unit III (m) Pollen Assemblage Unit IV (m) 39.9 herb-shrub 32.4 28.8 alder-dominated 13.5 woody 12.9 herb-shrub 10.2 0.6 mixed forest (abundant 1.8 to 6.9 montane conifers) O herb-shrub ' O The consequence of correlating Unit III with Units IV-V is to reduce the total Devils Gate composite section to approximately 445 m (Figure 5C). The following sections on stratigraphic palynology will discuss the attributes of Unit III along with the treatment of the other units. This approach is adopted for the sake of completeness but it should be kept in mind that the evidence for regarding Unit III as a repeated section is persuasive. Stratigraphic Relationships of Sucker Creek Sections The four previously published sections (Figure 1) are the Whiskey Creek ("Valley" section of Taggart 1971, Taggart & Cross 1974), Rockville, Shortcut, and Type section (type locality of Kittleman et al. 1965). Their stratigraphic relationships are discussed by Taggart and Cross (1980). On the basis of lithology they fall into three separate depositional settings: Whiskey Creek, Rockville-Shortcut, and Type 37 section. Stratigraphic palynological interpretation indicates that the sections have some overlapping time relationships. This is based in part on similarities in the pollen spectra, particularly the montane conifer element interpreted in a setting of inferred relative topographic positions of the sections with respect to the axis of the basin. A palynologic comparison of the Devils Gate section with previously studied sections is included in later discussion. However, at this point it is appropriate to describe an apparent lithologic similarity between the upper Devils Gate section (Units IV and V) and the Type section that indicates a possible correlation of the two. Despite the difficulties of correlating Sucker Creek outcrops in the area, the lithologic characteristics of color, texture, and induration of some of these sedimentary rocks are so striking that they allow recognition of unique combinations and sequences which permit correlation with some confidence. Taggart (I971) correlated the Rockville-Shortcut sections across 12.8 km of terrain. This lithologic correlation was later supported by palynologic interpretation of vegetation shifts. The Type section and the Devils Gate section are 10 km apart, a correlative distance that is not unreasonable by comparison with the Rockville-Shortcut correlation. The similarity of the two cliff-forming blocks is apparent from a marker bed of unique lithology, correlative sequential combinations of several distinctive rock units, and their positions relative to the marker bed. 38 Lithologic Comparison of Upper Devils Gate Units and Type Section On the basis of lithology (Figure 7), each section can be divided into two parts, separated by the distinctive green ash series that is not recognized at any other level in the two sequences or in any other Succor Creek outcrops studied. In the Type section it is conspicuous at or near the base of the lower cliff face where it weathers to a "hoodoo" type of feature. In the Devils Gate section, the green series occurs at the cliff base but has not been eroded to such conspicuous geomorphic character. Below the "hoodoo" green beds, dark, relatively fine, thinly-bedded sediments with several prominent lignitic shale zones are present. A stump or wood zone occurs about 50 m below the "hoodoo'I green beds near the base of each exposed unit. These sediments form low, deeply- weathering slopes (Type section) and valley floor (Unit IV Devils Gate section) and yield some productive pollen samples. Above the green beds, each section is well exposed on relatively steep cliffs. Each section is comprised of vari-colored, coarse, volcanic sandstones alternating with light gray volcanic sandstones of angular, pumiceous fragments, and glass shards which generally form resistant ledges. The two cliff sequences (125 m) are barren of palynomorphs. Both sections contain fossil bone fragments. Type section mammalian material has been studied by Shotwell (1968). In summary, this lithologic comparison is largely made on the basis of the equivalent positions of thick, light colored cliffs with resistant ledges and bone beds, and the equivalent positions of fine, dark, deeply weathered ashes, lignites, and wood zones. In both 39 you.“ Pollen Saloles Samples & \1 122.12.... Ibne ___.,$amles dominated 1===pby ell pollen : UNIT. V l 1 F: 1—b lone lone ———————————————— .===1I 1===ib 5000 Done W 9".“ _______________ Hoodoos Unit IV 11mm sh E ——————————— Loverlignite : lignite :- 2“ — Z _____ “00¢ _ 5m zone I ----- _ l \i ® A. & Type Upper . Devils Gate Section Section Figure 7. Stratigraphic and paleontologic comparison of the Type section and Devils Gate section Units IV-V. 40 sections the positions of these beds are established relative to the green beds. Additional support for placing such correlative value on the green beds comes from subsurface wells. Warner (1977) describes a marker bed used in subsurface correlation as a green chloritic ash, the "Green Hornet Ash". This bed occurs above his second marker described as a white porcelanite, the "Snowbird Shale" which fits the description of the white opaline shales of the Devils Gate Unit II. In the "average" subsurface section the green and white beds are separated by approxi- mately 150 m, and in the Devils Gate section they are separated by approximately 200 m. Summary This section has presented descriptions of the seven measured units that comprise the sampled Devils Gate section and a stratigraphic interpretation that places Units I and II, and IV to VII in stratigraphic sequence with some uncertainties concerning the amount of concealed section. Unit III is considered to be a repeated section of the main portion of Units IV and V. A correlation of Units IV-V with the Type section is postulated which, if true, would place most of the productive pollen samples at Devils Gate below the Type section. The stratigraphic palynology of the Devils Gate section may therefore represent a valuable extension, in time as well as space, of the study of Succor Creek vegetation dynamics. THE DEVILS GATE FLORA This section reports additions to the Succor Creek flora, analyzes some biases in preservation in an attempt to determine the relative importance of some genera in the Succor Creek vegetation, and proposes an emphasis for the reconstruction of the vegetation and climate. The purpose of this floristic and climatic analysis is to provide a setting for interpretation of the stratigraphic changes in pollen assemblages through the Devils Gate section. Additions to the Succor Creek Flora Additions to the Palynoflora The taxonomic study of the Devils Gate palynomorphs (Table 1) added very few pollen genera to the well-studied Succor Creek flora. New pollen records are Keteleeria, Pseudotsuga, Fraxinus, and Platanus. All have been previously reported as macrofossils. The other pollen genera were quite abundant and their floristic and ecological significance will be discussed in due course. Zygospores of an aquatic alga, Ovoidites (Zygnemataceae), spores of an aquatic fungus (Chytridaceae), and several acritarchs of unknown botanical affinity are also reported here. They occur together and in only a few samples at Devils Gate but when present are generally very abundant. Descriptions of new forms are contained in the Appendix. 41 42 Table 1. Families and genera recognized in the Devils Gate section. S-seed, C-cone, W-wood, L-leaf. MACRO- MICRO- FOSSIL FOSSIL FUNGI Phycomyceta Chytridaceae + Imperfecti + ACRITARCHS Undifferentiated Psophosphaera Micrhystridium Sigmapollis + + + + ALGAE Chlorophyta ‘ Zygnemataceae Ovoidites + Chrysophyta Botryococcaceae Botryococcus + LYCOPSIDA Lycopodiaceae Lycopodium + SPHENOPSIDA Equisetaceae Eguisetum St, C + PTEROPSIDA Polypodiaceae + Osmundaceae + GYMNOSPERMS Cupressaceae + Thuja? S Pinaceae Keteleeria L? Picea L? + + + + + + Podocarpaceae Podocarpus Taxaceae Taxodiaceae + Glyptostrobus L, C, W Ephedraceae Ephedra + + + 43 Table 1. (continued) ANGIOSPERMS MACRO- FOSSIL Aceraceae Acer L, S Aquifoliaceae Ilex Berberidaceae Mahonia L Betulaceae Alnus L Betula Carpinus-Ostrya Corylus Caprifoliaceae Chenopodiaceae- Amaranthaceae Compositae Artemisia Eleagnaceae Shepherdia Fagaceae Castanea Lithocarpus Fagus Quercus Hamamelidaceae Liquidambar Juglandaceae Carya Juglans Pterocarya L, S Lauraceae Persea L Malvaceae Meliaceae Cedrela S Nyssaceae Nyssa Oleaceae Fraxinus S Onagraceae Platanaceae Platanus L Rosaceae Crataegus L [— l_' MICRO- FOSSIL + + + + + + + + + + + + + + + ... Table 1. (concluded) Salicaeae Salix Tiliaceae Tilia Ulmaceae Celtis Ulmus Gramineae Potamogetonaceae Potamogeton Typhaceae Typha 44 MACRO- FOSSIL MICRO- FOSSIL 45 Additions to the Macroflora Collections were made from two leaf beds. One bed is a light colored tuff that contains sparse but well preserved leaves and seeds of diverse conifers and angiosperms. The other is a dark, lignitic, bentonitic clay bed that has preserved numerous Glyptostrobus leaf shoots. Unit I Mixed Leaf Bed. The leaf bed is found at about 12 m above the base of Unit I (Figure 2 and Plate 1) in a pinkish-gray to pale yellowish-brown tuff. The collection is notable for three features: 1. Additions to the Succor Creek record for the Juglandaceae. Here specimens of Pterocarya (leaf and seed) were found, the only such fossils reported in extant Succor Creek collections. Pterocarya pollen has been reported by Graham (1965) and Taggart (1971). A leaf specimen was cited by Chaney and Axelrod (1959) but it could not later be located (Graham 1965). The leaf collected from Unit I is a fine specimen with seven large, attached leaflets. 2. The assemblage is mixed, with five conifer genera, four broadleaved evergreen genera (Quercus, Mahonia, Cedrela, and Persea), and seven broadleaved deciduous genera (Table 1). 3. Some genera are represented by more than one species. The most diverse genus is Age: with at least 5 species differentiated. Macrofossil genera are represented by pollen in the same sediment, with only two exceptions, Cedrela and Persea. These are both tropical to warm temperate, evergreen trees, with insect pollinated flowers whose pollen is typically not released and incorporated in the pollen rain. 46 Furthermore, even when deposited, the pollen of the Lauraceae (including Persea), deteriorates rapidly and is generally not preserved. Unit IV Glyptostrobus Bed. This leaf bed occurs in association with a Glyptostrobus stump zone. It contains innumerable Glyptostrobus twigs, both sterile and fertile, cones and stems of Eguisetum, and broadleaved angiosperm leaves which are poorly preserved and unidentifiable and are overlain by Glyptostrobus leaves. Leaves are preserved in a volcanic ash mudstone which deteriorates upon exposure. The bed is noteworthy because it is the richest Glyptostrobus leaf bed found to date in the Sucker Creek Formation. These sediments also contain the highest percentage of taxodiaceous pollen found in the Devils Gate section (45%). It is apparently the record of a Glyptostrobus swamp forest. Graham's collections of 2500 specimens included only 50 of Glyptostrobus representing 2% of the flora. His most productive locality yielded only 10%, leading to the conclusion that swamps were a rare development in the Succor Creek setting (Graham 1965). Contribution to Floristic Analysis The few new plants recorded here complement previous records but provide little new floristic insight. However, with the addition of this study to the record of the Succor Creek flora, a pattern of co-occurrence of macrofossils and pollen in leaf beds and a pattern of pollen dynamics through time in the Devils Gate section can be to suggest an emphasis for reconstruction of the Succor Creek vegetation. 47 In the Succor Creek flora (72 gymnosperm and angiosperm genera and families) 28 taxa (39%) are preserved as both macrofossils and microfossils, 25(36%) as macrofossils alone, and 19 (25%) as pollen alone (data from Graham 1965, Taggart and Cross 1980, and this study). The two sets of data are thus complementary and reconstructions of vegetation have been strongly aided by considering both macrofossils and microfossils (Graham 1965, papers by Taggart and Cross). The flora includes diverse genera of three major physiognomic subdivisions of forest vegetation based on leaf characteristics: coniferous, broadleaved deciduous, and broadleaved evergreen. A comparison of the relative abundances of the physiognomic elements represented by both macrofossils and microfossils in the same beds in Succor Creek localities (Graham 1965) shows that broadleaved taxa, important as macrofossils. are underrepresented by pollen. Broadleaved elements constitute 94% of the total leaf flora and 60% of the pollen flora. Particularly notable is the relative representation of oak leaves (48%) and oak pollen (7%). Acer, Populus, Salix, and Platanus are also consistently more abundant in macrofossil assemblages. Other broadleaved taxa are more abundant in pollen assemblages from the leaf beds including Alnus, Carya, Celtis, Fagus, Fraxinus, Liquidambar, and Ulmus. As a group the conifers are much more strongly represented by pollen (26%) than macrofossils (6%). In particular spruce leaves and seeds are very rare or absent in the leaf beds where pollen assemblages average 20% spruce. 48 Interpretation of Inconsistencies of Macrofossil and Microfossil Records Discrepancy between pollen and leaf representation of a taxon is to be expected because of a large number of obvious factors that differ- entially bias the record of the two classes of fossils. Factors include differences in production, time and modes of dispersal, modes of tran- sport, and durability of the two classes of fossils. The significance of the discrepant recovery of pollen and leaf fossils is, however, sometimes difficult to assess and interpretations of relative popu- lations in the original vegetation are largely a matter of judgment and intuition. Studies of leaf taphonomy are few (Birks 1973, Spicer 1981) and not particularly applicable to the Succor Creek leaf flora for either leaf types or the depositional setting. In general, studies show that leaf material is transported relatively short distances and indicate that macrofossil assemblages are thus biased strongly towards representation of local vegetation. By contrast, many pollen assemblages are representative of the regional vegetation and their significance on a regional scale is rather well understood on the basis of many studies of temperate forests that show overall good correspondence between a pollen assemblage and the vegetation of a region (Birks and Birks 1980, Davis and Webb 1975, Webb and McAndrews 1976). In the Succor Creek flora, two genera with discrepant records, the broadleaved evergreen oaks (represented by leaves) and the conifer spruce (represented by pollen), particularly merit evaluation because of their dominance in the Succor Creek fossil record, their dominance of two of the vegetation elements, and apparent ecological disparity. 49 Previous studies have emphasized the significance of spruce (papers by Taggart and Cross) in the regional vegetation. The following discussion will attempt to support the position that both the oaks and spruce had important roles in the regional vegetation and that zonation was indistinct. The lines of support will include modern studies of dispersal and deposition of the pollen; comparison with the structure of modern, temperate forests in regions where both evergreen oaks and spruces are prominent; and an analysis of the pattern of occurrence of forest elements in more or less contemporaneous Miocene floras of the Pacific Northwest. Interpretation of the Role of Evergreen Oaks in the Flora. In the Succor Creek leaf beds that yield pollen as well as leaves, evergreen oak leaves are much more abundant than oak pollen (Figure 8). Yet, North American oaks are known as good pollen producers that tend to be overrepresented in modern pollen rain relative to their importance in the vegetation (e.g., Davis and Goodlet 1960, Webb 1974). On this basis an interpretation of the discrepancy is that the oaks were unimportant regionally, that they dominate the leaf assemblages only because they grew on the immediate margin of the basins, and thus that their relative abundance in pollen spectra reflects dilution of a local riparian element in the regional pollen rain. Two factors can be used to argue for more than a limited riparian role. First, the presumed living equivalents of three of the leaf species (9. consimilis, Q. dayana, Q. hannibali) are not necessarily restricted to riparian or lake margin habitats (Graham 1965) and some have very wide altitudinal ranges. The fourth species, Quercus simulata, has uncertain affinities. Second, 50 Quercus Ulmus-Zelkova Pinus Alnus Age: Mascall Gray 1958 Blue Mtns Gray 1958 Stinking Hater Gray 1958 Succor Creek Graham 1965 RELATIVE ABUNDANCE pollen ||||| leaves %%%% Figure 8. Comparison of pollen and leaf abundances from some Oregon Miocene leaf beds. 51 three or all four oak leaf species occur at most localities in fairly stable proportions in association with a diverse assemblage of other broadleaved evergreen, broadleaved deciduous and coniferous taxa. The present day setting in which three or four species of evergreen oaks are found together with diverse broadleaved evergreen, broadleaved deciduous and coniferous taxa is in broad regional mixed forests of East Asia (Wang 1961). An alternative interpretation of the fossil discrepancy, in light of the taxonomic diversity of the Succor Creek oaks and the ecological diversity of their living equivalents, is that they were a regional component of the vegetation that was being underrepresented by pollen in the regional pollen rain. Despite their usual overrepresentation in the North American pollen rain, there is indication from modern studies of pollen underrepresentation in some circumstances. Pollen is underrepresentative of oak trees in some vegetation types. While overrepresentation is reported for eastern North American studies generally (e 9., Davis and Goodlet 1960, Webb 1974) and for some European studies (e g., Wijmstra 1978), other European studies show oak and vegetation to be about equally represented (Janssen 1966, Faegri and Iversen 1975). A study from central Taiwan in a mixed vegetation that includes evergreen oaks shows that the pollen was more abundant than oak in that vegetation by a ratio of 1:1.7 (Tsukada 1967). The Taiwan study may be particularly relevant to the problem under consideration because the vegetation and pollen rain is similarly of mixed evergreen oaks, broadleaved deciduous genera, and conifers. It appears that oak is either not necessarily an overproducer or that other factors are 52 operating in some settings to affect its relative abundance in the pollen sum. There are no studies of the deposition and preservation bias of these entire or serrate margined, coriaceous, evergreen oak leaves in a mixed vegetation. It might be expected that in this mixed flora, these leaves would have distinct advantages in their availability for transport and burial by winter storm, in their transportability, and in burial. Thus one might expect that the fossil leaf/pollen discrepancy would be greater than the tree/pollen discrepancy. The question of the role of oaks in the flora is not resolved by these observations but they leave open the possibility that disparate representation of oak pollen and leaves in the flora does not necessarily mean a limited local occurrence for the four oak species in a mixed fossil assemblage. The possibility of a regional role will be further supported with discussion of modern East Asian vegetation that have several evergreen oaks in important regional roles in mixed forests. Interpretation of the Role of Spruce in the Flora. In this case the discrepancy in the fossil record is the inverse of that for the oaks. Spruce macrofossils are rare in Succor Creek leaf beds. Only a few seeds have been found whereas pollen is consistently abundant (averaging 20%) in the pollen assemblages of the same leaf-bearing beds (Graham 1965). Spruce pollen represents more than 50% of samples in other horizons in the formation and the conclusion has been reached that these pollen abundances reflect an important role for spruce in the regional vegetation removed from the immediate basins (papers by Taggart 53 and Cross). This interpretation is based on modern pollen studies that show Spruce pollen abundance to be a fair representation of spruce trees in the vegetation. Studies of spruce pollen in both moss and lake samples show a general pattern of relatively poor dispersal capacity and some show underrepresentation of the vegetation at local as well as regional levels (Janssen 1966, 1967, Webb and McAndrews 1976. Hicks 1977, Prentice 1978). Webb and McAndrews (1976) found the occurrence of pollen to be generally a good indicator of the regional distribution of spruce in the vegetation in eastern and central North America. But when they compared relative abundances of pollen and trees they found the "relative abundance of spruce trees to be somewhat greater than that of pollen; this fact indicates the general underrepresentation of spruce in the pollen record". Values exceed 20% only in the boreal forest and tundra. Caution may be indicated in comparing relative abundances of spruce from these modern pollen studies with those in Succor Creek samples because of the substantial difference in the representation of pine pollen. Pine is an associate in modern spruce dominated forests. In pollen spectra from spruce forests pine is the dominant pollen type with the consequence that all other relative abundances (including spruce) are relatively lower. By contrast, pine pollen frequencies are relatively low in the Succor Creek assemblages that show high spruce pollen. Pine was probably not abundant in the Succor Creek forests and consequently less a factor in the distortion of other pollen percentages than it is in the modern pollen spectra. Nevertheless, 20% spruce 54 pollen in evergreen oak leaf dominated sediments (Graham 1965) strongly indicate that spruce was an important regional component. Other conifers are similarly disparately represented by pollen and macrofossils in the Succor Creek flora. In the Devils Gate section hemlock pollen abundance peaks at 10% and fir pollen abundance at 8%. In modern sediments the pollen of hemlock and fir are even more strongly underrepresentative of their regional contribution to the vegetation than is spruce. Maximum values of fir pollen (20%, mean 1.6%) are found in small pockets in the boreal forest of North America (Webb and McAndrews 1976). Hemlock pollen abundance is highest in pollen spectra from the mixed conifer hardwood forests of eastern North America (maximum abundance 15%, mean 2.4%). These comparisons also suggest important roles for hemlock and fir. In the mixed forest vegetation of central Taiwan, Picea, Abies, and Isuga dominate stands 1600 m above an evergreen oak dominated site where lake sediments were sampled for pollen. In this setting total pollen abundance for the three conifers did not exceed 5% (Tsukada 1967). Such evidence lends strong support to the placement of these conifers close to the Succor Creek basins as a regional dominant where their combined contribution may exceed 50% of the total pollen (papers by Taggart and Cross, this study). The evidence from modern pollen studies supports the interpretation that spruce was a major component of the Succor Creek vegetation at least somewhat removed from the margin of the basins (based on the paucity of fossil seeds in the leaf beds) but not restricted to limited areas of high elevation. Substantial populations of hemlock and fir 55 also occurred regionally. Similarly, they could not have been confined to distant sites of high elevation. Vegetation Reconstruction Summary of Characteristics of the Succor Creek Flora 1. The preceding discussion focused on an interpretation of the relative abundance in the Succor Creek flora of two genera that have disparate but important representation in leaf or pollen assemblages and argued for regional roles for spruce and (presumably evergreen) oaks. 2. The oaks were represented by four species of coriaceous, entire, or sparsely serrate-margined leaves. Three or four oak leaf species occur together at most macrofossil collecting localities and apparently were closely associated in life. They are thought to have been evergreen on the basis of the habit of the living equivalents (Graham 1965). However, it should be noted that uncertainties surround this designation. The systematic treatment of entire-margined and irregularly-serrate fossil leaves of three of the genera of Fagaceae is very difficult. Uncertainties surround even the distinction of entire-margined species of Quercus from the genera Lithocarpus and Castanopsis (Graham 1965). Furthermore, the large suite of modern oaks with entire to serrate-margined leaves exhibit a range of habit from evergreen to semi-evergreen to deciduous. In the diverse Mexican oak flora, for example, entire-margined and serrate-leaved types are more commonly deciduous than evergreen. 56 3. Other evergreen broadleaved trees, shrubs, ground-covers, and vines are represented as macrofossils including: Castanopsis, Lithocarpus, Cedrela, Oreopanax, Arbutus, Mahonia (3 species), Hiraea, Ilex, and Anoda. 4. Diverse deciduous genera occur. Some genera are represented by several species that often occur in the same deposit: Aggr (6 species), Betula (3 species), Platanus (2 species), Populus (3 species), Ulmus (4 species). 5. Diverse conifer macrofossil remains have been found, though their occurrence is mostly erratic, with the exception of Glyptostrobus. Conifer genera are: Glyptostrobus, Taxodium, Thuja, Cephalotaxus, Picea, Keteleeria, and Pinus. Diverse conifer pollen is found, some types in great abundance: Abies, Cedrus, Keteleeria, Picea, Pinus, Tsuga, Pseudotsuga, Taxodiaceae-Cupressaceae-Taxaceae, and Podocarpus. Glyptostrobus apparently developed locally pure stands in swamp conditions and also occurred in mixed stands on drained slopes. These characteristics will be compared to modern East Asian temperate forests to suggest a Succor Creek vegetation of mixed physiognomic and floristic diversity where evergreen oaks, diverse broadleaved evergreen and deciduous trees and shrubs, and diverse conifers grew either in overlapping stands or in a blurred mosaic of stands, with an increasingly strong representation of conifers in the regional vegetation at higher elevations. This emphasis differs from Graham (1965) in suggesting the lack of clear zonation of the conifers and an association of broadleaved evergreen and deciduous elements of the flora. It differs from Taggart and Cross (papers) in recognizing the broadleaved evergreen component of 57 the flora and in proposing a significant role for that element at lower elevations in a mixed vegetation. The regional importance of conifers in the Succor Creek vegetation has been cogently stressed (Taggart and Cross 1980, Cross and Taggart 1983). In the following discussion comparisons will be made with temperate East Asian mixed forests and climate. Then comparisons will be made with some North American forests that show strong floristic similarity to elements of the Succor Creek flora but do not show the association of elements. They provide a test for the proposed reconstruction of the Succor Creek environment by identifying the environmental parameters that causes one element or another of the mixed vegetation to be eliminated. Finally, the discussion will summarize regional trends in the association of broadleaved evergreen and conifer elements of the vegetation of some well studied, more or less contemporaneous Miocene floras in Oregon and Idaho. Floristic and Climatic Comparisons: Mixed Temperate Forests of East Asia The striking feature of temperate East Asian forests is that many associations can be characterized as "mixed", with broadleaved evergreen, broadleaved deciduous, and conifer elements. The floristic composition of each element is largely shared by adjoining forest types and it is only the proportional representation of each element and its position in the forest mosaic that changes (Table 2). Wang's treatise (1961) is used here as the primary source of data on the composition of the mixed forests of China. 553 .wmmkaooomm .cmcmummu .mmmooc=:4 .mzucmam comcmco>0 3 432 .84 xucoaocucnu< .o_=umm mamumoo:u_4.xmo .cacm>0 .mmwmm .m0_n< .wmzmh msocmw III'|'I"'I|"I|'|'| mzo=u_umc-:00cmc0>0 umxwz Amuowa .mown< omcmuumom+v mzoacwomc-:00coc0>0 u0x_z acou A_omp acazv z<3IQMNm cam: .Am:_;uv consumNm cc: cmzwmh we mummcow umx_s mumcmasmh ma:_a .omcmmrwwmw .mammmcmzu .maxmu .o_m:mmw .mzxcp .mxmccQH .mwvwmam maoao_00u .wzucmmo cmmcmc0>0 uax_z wmucmzm cmmgmco>0 m0_umam msoakuwn wwuwm .A.ammv omaww umx_z mmwumnm maoan_umv chcg IIII'II I all .I, A.am~v aaa_a .ma_n< maoc00_:ou A_om_ acazv Acute saxcazov z0v nmxwz A.:owam>0_0 new: 0:0 -cogeoo cmcficoo mcwmmmcu:_v xcoumcmvca :mmcmc0>0 twmuxumo cmmcmc0>0 .mmuucmmamamcc «mmmmW -ouJMWM .cchammwcczu .WJMMH whHamN mammw sMMQMm .m_umm omoaozu A.:w50u mflmomxum:Ec:uv umxwz mzcwa .mmmwm .cmo_a .muwafi maoc0»_:ou Aaom_ acazv N “.mcaz _aca=auv z<3c0 .mam -cmoo:u_4 .mwmmmcwumuu fiucmcwso: comcm0>0 any u0x_z m=:.mccu mam—ma: .m>oxpmN .m:e_= 1 .maocmmm cowcmcm>0 .mmmmuouammm .m_E:: rmcwccmu «wm:cmxommammm ~.:wson mHMmmNummsmguv u0x_z . Amzcwnp mmo_m .mmwn< . mam» mao:0»_:ou . Amoa— oumxampv H A.m:az Pmcucmuv z<3~u ou>mo one cw mmcmco :o_umummm> ucw mocmncaumwu do xcm553m mt<3m hmu¢Ou Ilv mhz _ 3 § 5 g : .mH menace mwz<=u zc~h no zmuhhzm omaaahmna muzo uo>oopuooco a >uoaoz>4¢m wuz . ‘11}; ...u 5 .\ I u '21 20 19