Publications of the Department of Geology D4Helsinki 2005

A long terrestrial sequence in Lantian – a window into the late Neogene

palaeoenvironments of northern China

Anu Kaakinen

Academic dissertation

To be presented with the permission of the Faculty of Science of the University of Helsinki, for public criticism in Auditorium,

Arppeanum, Snellmaninkatu 3, University of Helsinki, on June 21st, 2005,at 12 o’clock noon

Supervised by:Professor Juha Pekka LunkkaDepartment of GeosciencesUniversity of Oulu

Professor Mikael ForteliusDepartment of GeologyUniversity of Helsinki

Reviewed by:Docent Philip L. GibbardGodwin Institute for Quaternary ResearchDepartment of GeographyUniversity of CambridgeEngland

Professor Matti RäsänenDepartment of GeologyUniversity of Turku

Opponent:Professor Jef VandenbergheDepartment of Quaternary Geology and GeomorphologyVrije Universiteit, AmsterdamThe Netherlands

ISSN 1795-3499ISBN 952-10-2156-X (paperback)ISBN 952-10-2157-8 (PDF)http://ethesis.helsinki.fi/

Helsinki 2005Yliopistopaino

PhD-thesis No. 183 of the Department of Geology, University of Helsinki

Front cover: Exposure showing floodplain deposits from the late Miocene Bahe Formation

Abstract

A conformable, fossiliferous sedimentary sequence in the Lantian area is a unique terrestrial record of the northern Chinese Cenozoic Era. The region occurs within the Weihe Graben basin in the southernmost Loess Plateau. This work is focused on the Bahe and Lantian formations that comprise the late Neogene part of the clastic infill of the Weihe Graben. The basal Bahe Formation is ca. 280 m thick in the study area and spans from about 11 Ma to 6.8 Ma. Its facies associations are interpreted to have been deposited in a relatively low-energy, braided and multichannelled fluvial system with broad floodplains. The Bahe Formation is overlain by a ca. 50 metres thick sequence of mainly aeolian ‘Red Clay’, termed the Lantian Formation in the study area. It is composed of strongly rubified, silty clay to clayey silt beds that frequently contain carbonate-rich horizons, typical features of the ‘Red Clay’ deposits underlying the Pleistocene loess-paleosol sequences in northern China. The succession indicates relatively stable climatic and tectonic conditions through-out the deposition of the lower part of the sequence. A major sedimentological change oc-curred at the formation boundary at ca. 6.8 Ma. The boundary is not always sharp, as stated in earlier studies, but often transitional with a fluvial imprint in the lower part of the ‘Red Clay’. A total of 52 fossil vertebrate localities were discovered at different levels in the sequence, including remains of large and small mammals. Taphonomic and sedimentologi-cal studies show that the richest fossil vertebrate accumulations occur in proximal flood-plain sediments, such as sheet or crevasse-splay sands and silts. Magnetostratigraphy of the Lantian sequence allows calibration of most the fossil localities to the Geomagnetic Polarity Time Scale. The stable carbon and oxygen isotopic compositions were measured from the soil carbonates. In the fluvial part of the sequence, the records show only slight variation in δ13C and δ18O values implying relatively stable and water-stressed conditions. A shift in the carbon and oxygen stable isotope curves occurs at the formation boundary, indicative of a transition to more humid conditions during deposition of the Lantian Formation. There is no evidence suggesting a significant C4 component in the vegetation at any time. The envi-ronmental interpretation of the sedimentological and isotopic records is consistent with that derived from the vertebrate fossil faunas. In conclusion, coeval with the change in depositional regime from fluvial to mainly aeolian at the Bahe – Lantian Formation boundary, the surrounding area appears to have be-come somewhat more humid, with a predominantly open, possibly shrubby and woodland mosaic being succeeded by forested habitats. This transition from dry and open towards more a humid and forested landscape is opposite to the global trend. This environmental

Anu Kaakinen: A long terrestrial sequence in Lantian – a window into the late Neogene palaeoenvironments of northern China, University of Helsinki, 2005, 49pp., University of Helsinki, Publications of the Department of Geology D4, ISSN 1795-3499, ISBN 952-10-2156-X, ISBN 952-10-2157-8 (pdf-version).

change has been related to the intensification of the rain-bearing East Asian summer monsoon at about 7–8 million years ago, resulting in the transition to more humid and closed habitats. A comparison of the Lantian sequence with other records preserved in the Loess Plateau region reveals that, whilst the overall patterns of climatic variations in the Neogene throughout North China have been similar, significant local variations occurred. Relatively continuous and fossiliferous late Miocene sequences are rare in North China. Therefore, interpretations of the sediments, isotopes and vertebrate fauna from the Lantian sequence provide a unique window into the evolution of late Neogene environments in China.

1. Introduction 7 1.1. Long terrestrial sequences in Eurasia 7 1.2. Neogene terrestrial sequences in China 8 1.3. Lantian 9 1.4. Description of the project 10 1.5. Aims of the present study 102. General Setting 12 2.1. The Loess Plateau 12 2.2. Qinling Mountains 13 2.3. Lantian region 14 2.4. Geological background 15 2.4.1. Tectonic setting 15 2.4.2. General Stratigraphy 153. Material and methods 18 3.1. Sedimentology 18 3.2. Isotopes 18 3.3. Magnetostratigraphy 19 3.4. Taphonomy 204. Results 21 4.1 Sedimentology 21 4.1.1. Bahe Formation 21 4.1.2. Lantian Formation 21 4.2 Magnetostratigraphy 24 4.3 Fossil localities and taphonomy 26 4.3.1. Weathering analysis 26 4.4. Stable carbon and oxygen isotope record 275. Discussion 29 5.1. Completeness of the sedimentary record in Lantian sequence 29 5.2 Distribution of fossil vertebrate localities 30 5.3. Age of fossil vertebrate localities 32 5.4. Synthesis on palaeoenvironments and palaeoclimates 346. Conclusions 397. Acknowledgements 408. References 42Papers I-V

Contents

5

List of publications

I Zhang, Z., Gentry, A., Kaakinen, A., Liu, L., Lunkka, J.P., Qiu, Z., Sen, S., Scott, R.S., Werdelin, L., Zheng, S. & Fortelius, M. (2002): Land mammal faunal sequence of the late Miocene of China: new evidence from Lantian, Shaanxi Province. Vertebrata PalAsiatica 40(3), 165-176.

II Kaakinen, A. & Lunkka, J.P. (2003): Sedimentation of the Late Miocene Bahe Formation and its implications for stable environments adjacent to Qinling Mountains in Shaanxi, China. Journal of Asian Earth Sciences 22, 67-78.

III Andersson, K. & Kaakinen, A. (2004): Floodplain processes in the shaping of fossil bone assemblages: An example from the Late Miocene, Bahe Formation, Lantian, China. GFF 126, 279-287.

IV Kaakinen, A., Gose, W.A., Sen, S. & Lunkka, J.P. (manuscript): Magnetostratigraphy of the late Neogene terrestrial sequence in Lantian (Shaanxi Province, China).

V Kaakinen, A., Sonninen, E. & Lunkka, J.P. (submitted): Stable isotope record in paleosol carbonates from the Chinese Loess Plateau: implications for late Neogene paleoclimate and paleovegetation. Palaeogeography, Palaeoclimatology, Palaeoecology.

The author’s contribution to the articles

I AK conducted fieldwork jointly with JPL and was responsible for writing the sedimento-logical part of the manuscript.II AK performed fieldwork together with JPL, organised and analysed observations for the publication, prepared the manuscript.III AK was responsible for the sedimentological part of the study and wrote the article jointly with KA.IV AK participated in sampling and designed some of the sampling; carried out part of the laboratory analysis and interpretation. WG and SS analysed the data. AK prepared the manuscript.V AK carried out the sampling and interpretation of the data and was responsible for the article writing. ES analysed the isotopes.

6

7

1. Introduction

The Neogene Period (ca. 24–1.8 Ma) rep-resents an important episode in the Earth’s history, characterised by a succession of profound changes in both the terrestrial and marine realms that led to the modern con-figuration of climates and environments. The early Miocene global warmth culmi-nated in the mid-Miocene climatic optimum 17–15 Ma that marked the warmest period in the Neogene with high-latitude tem-peratures even 6ºC higher than at present (Flower, 1999). The climatic optimum was followed by a gradual cooling, associated with the change from an ephemeral to a permanent Antarctic ice sheet by 10 Ma ago (Flower and Kennett, 1995; Zachos et al., 2001) and cooling of Antarctic deep waters. These changes resulted in steepening of me-ridional thermal gradients, strengthening of boundaries between climatic zones and to the increased aridification of the mid-lati-tudes (Flower and Kennett, 1994). During the late Miocene (11–5 Ma), the global trend of steady climatic deterioration, drying and increased seasonality continued (Janis, 1993). Important tectonic changes also took place during the late Miocene. Among these changes the intensified uplift of the Himala-yas and Tibetan Plateau at 11 Ma as a result of the collision of India with Asia (e.g. Har-rison et al., 1993) generating or intensifying the Asian Monsoon circulation (Prell and Kutzbach, 1992; Rea, 1992) is particularly well-known. Prominent changes in terrestri-al ecosystems took place in the late Miocene as more open woodlands began to replace the forests (Potts and Behrensmeyer 1992), accompanied by the expansion of grasslands and savannas from 7–8 Ma onwards (Quade et al., 1995; Cerling et al., 1997). Among the most important bioevents in the late Mio-cene is the dispersal of grazing, three-toed

equid ‘Hipparion’ from North America into Eurasia and Africa (e.g. Garcés et al., 1997). In the Mediterranean basin, the Messinian salinity crisis – the desiccation and subse-quent catastrophic flooding of the Mediter-ranean Sea – occurred in the latest Miocene (about 6–5.3 Ma) (Krijgsman et al., 1999; Duggen et al., 2003). The early Pliocene subtle warm-ing was followed by Northern Hemisphere glaciation and global cooling in the late Pliocene and Pleistocene (Maslin et al., 1998; Zachos et al., 2001).

1.1. Long terrestrial sequences in EurasiaContinental stratigraphic sequences are generally viewed as incomplete records of the evolution of palaeoenvironments and palaeoclimates through time. The quality of proxy data in terrestrial sequences is often hampered by tectonic movements, discon-tinuous sedimentation, scarcity of fossils or sequences of variable fossil content. In contrast, many marine sequences provide records that are fossiliferous thoughout, hav-ing continuous and undisturbed sedimenta-tion. On the other hand, terrestrial records have a potential for higher temporal and spatial resolution while marine sequences have finite resolution due to low sedimenta-tion rates. In the absence of available radio-metric ages, terrestrial sequences can be dat-ed using magnetostratigraphy, applying the Geomagnetic Polarity Time Scale (GPTS) that is based on marine magnetic anomalies (Cande and Kent, 1995). Magnetostratigra-phy can provide a detailed and long-term chronological framework for the deposits, independent of biostratigraphy. Thus it is a useful tool for correlation in the continental domain and also allows correlation of ter-restrial sequences to marine records.

8

There are a few fossiliferous con-tinental sedimentary sequences in Eurasia, especially in the Indo-Pakistan and the Mediterranean regions, that provide long records of Neogene fossil faunas, paleoen-vironments and depositional histories. The most extensive and probably best-studied continental Neogene sequence in terms of geology and paleontology is the Siwalik Group (‘Siwaliks’) (Badgley and Behrens-meyer, 1995), exposed along the southern foothills of the Himalayans across most of northern Pakistan and India. The Siwalik Group consists of deposits of predominantly alluvial and fluvial environments from the middle Miocene throughout the Pliocene and contains an exceptionally continuous fossil record (e.g. Johnson et al., 1985; Badgley, 1986; Flynn et al., 1990; Willis, 1993a, 1993b; Willis and Behrensmeyer, 1994; Khan et al., 1997; Zaleha, 1997a, 1997b; Barry et al., 2002). In Western Eu-rope, the best stratigraphic information on continental mammal-bearing successions is available from Spain, where much re-search has been undertaken during past dec-ades especially on mammal biochronology and high resolution stratigraphy in several continental basins including the Vallesian, Aragonian and Turolian type areas (e.g. Calvo et al., 1993; Garcés et al., 1996; Kri-jgsman, 1996; van Dam, 1997; Opdyke et al., 1997; Daams et al., 1998, 1999a, 1999b; Abdul Aziz, 2001). In Greece, late Miocene continental successions are known from the classical localities of Pikermi and Samos Is-land (e.g. Solounias, 1981; Weidman et al., 1984). The stratigraphical record of the Si-nap Formation of central Turkey covers the middle and late Miocene (Kappelman et al., 1996; Fortelius et al., 2003).

1.2. Neogene terrestrial se-quences in ChinaNeogene continental deposits are widespread in Northern China and they frequently yield vertebrate fossil remains. Documentation of Chinese late Neogene vertebrate fossils dates back to the early twentieth century, when Schlosser (1903) described mammal fossils, mostly bought as “dragon bones” from chemist shops. During the 1920s and 30s, extensive field campaigns resulted in an increasing number of Chinese late Neo-gene vertebrate fossils being described and large collections produced (e.g. Andersson, 1923; Zdansky, 1923; Bohlin, 1926, 1935; Teilhard de Chardin and Young, 1931). From then onwards, many late Neogene mammal localities have become known in China, including the late Miocene fossilifer-ous sites, for example, from Tsaidam Basin (Qinghai Province), Amuwusu and Ertemte (Inner Mongolia) and Baode (Shanxi Prov-ince). Pliocene localities are known in sev-eral areas, for example, in the Yushe Basin (Qiu and Qiu, 1995). Although the study of Chinese Neogene faunas has challenged paleontologists for over a century, system-atic research into Neogene stratigraphy did not start before the late 1970s (Qiu and Qiu, 1995). This is partly because a significant part of the vertebrate fossil record comes from surface finds or from exposures where the degree of stratigraphical resolution is poor. Occasionally the work is also limited by a lack of knowledge on the exact position of the bone beds and by a notorious lack of dateable volcanic rocks. The stratigraphic information began to increase in the 1990s with the magnetostratigraphical correlation of key sections to the GPTS. However, mag-netostratigraphical control of fossiliferous late Neogene sections has been very limited and has been primarily concentrated on the

9

Yushe basin sections (Tedford et al., 1991; Flynn et al., 1995; 1997). Some other im-portant studies have recently been published from the Linxia Basin (Fang et al., 2003).

1.3. LantianThe first investigations in Lantian (Fig. 1) were conducted in the late 1950s (Liu et al., 1960). In the following decade, Chi-

nese paleontologists and geologists started to explore systematically the Lantian region (comprising Lantian, Lintong and Weinan counties) and carried out intensive exca-vations and geological mapping of the se-quences. The field campaign resulted in the discovery of various mammal localities, one of which yielded mandible of Homo erec-tus, as well as a geological map of Cenozoic

Lishan

Weihe

Bahe

Lantian

Qinling Shan

Figure 1. Detailed Landsat image of the Lantian region.

10 km

N

10

deposits in the Lantian area (Jia et al., 1966; Chow, 1978; Liu et al., 1978). Subsequently there followed a period of low research ac-tivity during which the studies were mainly concentrated on Pleistocene loess deposits in the area. Some geological and tectonic studies, mainly related to the investigations on revealing the uplift history of the Qin-ling Mountains and evolution of the Weihe Graben, were also extended into the Lantian region. Among these is the work by Bellier et al. (1988, 1991) that shed light on the subsidence history of the Weihe Graben in the Lantian region. During the 1990s, the magnetostratigraphical approach improved the chronology of the ‘Red Clay’ sequence in Lantian (Zheng et al., 1992; Yue, 1995; Sun et al., 1997), but studies were not ex-tended to the underlying fluvial deposits. It was not until the project reported here that detailed investigations on vertebrate palae-ontology and geology were resumed in the Lantian Neogene sequence. Since the pre-liminary results (Zhang Z.Q. et al., 1999; Paper I), project researchers have published numerous papers based on their findings on small mammals (Qiu et al., 2003; Qiu et al., 2004a, 2004b), bovids (Zhang, 2003; Chen and Zhang, 2004), sedimentology (Paper II) and taphonomy (Paper III).

1.4. Description of the projectThis thesis is a contribution to a Finnish – Chinese Lantian field research programme that was included in two subsequent, joint international and multidisciplinary projects entitled “Neogene Land Mammal Faunas of Eurasia: A problem-orientated field cam-paign in Anatolia and China (1996 – 1998), and “Physical and biotic changes in the con-tinental environments of Eurasia during the Neogene (1999 – 2001)”, both funded by the Academy of Finland. The fieldwork in Lan-tian was conducted during the period 1997–

2001. In the initial part, the field project be-gan as a bilateral collaboration between the University of Helsinki and the Institute of Vertebrate Paleontology and Paleoanthro-pology (IVPP) of the Academia Sinica (Bei-jing). However, soon afterwards it included international specialists from several other partner universities and institutions, includ-ing Muséum National d’Histoire Naturelle in Paris; The Natural History Museum in London; the Swedish Museum of Natural History; the University of Cambridge; the University of Texas at Austin and the Uni-versity of Utah. The aim of the Lantian Project was for the first time to provide a de-tailed documentation of the physical and bio-tic changes in a local section from the late Miocene of eastern Asia by applying a mod-ern multidisciplinary approach. A secondary objective was to compile a database of Chi-nese fossil mammal localities and faunas in order to analyse the evolution of mammal faunas and environments of Eurasia on a continental scale using the NOW (Neogene of the Old World) database and standards (Fortelius et al., 1996; http://www.helsinki.fi/science/now/)

1.5. Aims of the present studyThe main purpose of the work reported here is to create a high-resolution stratigraphical framework for the Bahe and Lantian forma-tions and to reconstruct the depositional and environmental history of the area through time. The main tools employed were con-ventional sedimentological methods, to-gether with paleomagnetic and geochemical techniques, combined with evidence from the fossil mammal assemblages recovered from the area. The first focus is on presenting a detailed geological analysis of the Bahe For-mation and on placing it in the context of the

11

broader tectonic and climatic settings. This analysis includes the investigation of lateral and vertical facies changes and interpreted primary depositional environment, gross sedimentation rates based on paleomagnetic dating and analysis of the palaeo-Bahe flu-vial system. It also examines evolution of the system during the late Miocene, as well as stable carbon- and oxygen-isotope evi-dence for environmental change in the area. The second focus is the placing of fossil occurrences in the sedimentary se-quence and correlation of fossil-find locali-ties to the GPTS (Cande and Kent, 1995). The occurrences of fossil concentrations in different stratigraphical horizons are ana-lysed. The aim is also to analyse the sedi-mentary processes involved in the origin of the fossil assemblages, and to assess their

similarities and differences in preservation, as well as the depositional processes that operated at selected localities. The third objective of this study is to describe and discuss the nature and tim-ing of environmental change that is mirrored in the shift of depositional regime from the Bahe to the Lantian formations. The specific issues addressed include discussion of the depositional continuity within the Lantian sequence, in general, and at the formation boundary, in particular. Finally, this work aims to summa-rise and compare the palaeoenvironmental interpretations derived from the different proxy evidence and to distinguish local changes from those of continental-scale during the late Neogene, including the uplift of the Tibetan Plateau and the development of the Asian monsoonal circulation.

12

2. General setting

2.1. The Loess PlateauThe Loess Plateau of north-central China lies in the middle reaches of the Yellow Riv-er (Huang He) (Fig. 1 in Paper V), rimmed by the Tibetan Plateau and the Helan Moun-tains to the southwest and west, the Taihang Mountains to the east, the Gobi desert in the north and the Qinling Mountains to the south. Averaging ca. 1000 m above sea level (Zhao, 1994) in elevation and cover-ing ca. 350,000 km2, it includes the bulk of the loess deposits of China (Liu, 1985) and contains the world’s greatest accumulation of loess. Administratively, the main body of the Loess Plateau occurs within the Shaanxi, Shanxi and Gansu provinces. The region is extensively covered by a mantle of Pleistocene-age intercalated loess-paleosol deposits. The loess deposits can be subdivided into the Early Pleistocene Wucheng Formation, Middle Pleistocene Lishi Formation and the Late Pleistocene Malan Formation. The Holocene soil at the top is often referred to as the Black Loam, and is locally interbedded or overlain by redeposited loess (e.g. Liu, 1985; Kukla, 1987; Kukla and An, 1989). The thickness of the loess-paleosol deposits in the central and southern Loess Plateau typically exceed 150 m and include over 30 interbedded loess – fossil soil units (Kukla, 1987; Kukla and An, 1989; Rutter et al., 1991). These loess and paleosol units are identified on the ba-sis of their colour, texture and structure, as well as using their magnetic and geochemi-cal properties. The loess units are of yellow colour, they are massive in structure and are rich in CaCO3. The grain-size composi-tion is dominated by the silt fraction. The brownish or reddish soils, in turn, are rich in clay, depleted of CaCO3, show pedogenic features and higher magnetic susceptibility

values than the surrounding loess. The loess deposits have been intensively studied using a variety of different geological, sedimento-logical, geochemical and dating techniques. The loess-paleosol deposits in the Loess Plateau are usually explained as recording alterations of the Asian monsoonal climate system from the late Pliocene throughout the entire Pleistocene (e.g. An et al., 1990, 1991; Ding et al., 1992, 1995; An, 2000). Earlier studies have shown that aeolian dust transport is closely associated with the Asian winter monsoon (e.g. Liu, 1985). There is a consensus that the loess units mainly accumulated during glacial events, with relatively cold and dry environments and a strengthened north-westerly winter monsoon; whereas the soil-forming periods correspond to reduced loess accumulation in moderately temperate and humid condi-tions during intervals with a strengthened summer monsoon (e.g. Liu, 1985; Kukla, 1987; Kukla et al., 1988; An et al., 1991; Derbyshire et al., 1995; Ding et al., 1995; Xiao et al., 1995). The Loess Plateau is subject to high soil erosion rates and has been dissected into a network of gullies and hills. Three main types of the landforms are often recognised: plateau landforms (yuan), elongated ridges (liang) and rounded hills (mao) (Liu, 1985). The yuan-type landforms (e.g. Bailuyuan Plateau in the Lantian region) develop in large and stable areas, and these landforms contain the most continuous records of aeo-lian accumulation. In places where loess cover is thin, the pre-existing landscape influences the present loessic landforms. In many places, terraced farmlands occupy the slopes of loess ridges or mounds. The present climate in the Loess Plateau is highly seasonal, being charac-terised by alterations of the summer and winter monsoons. In summer, differential

13

heating between the Asian continent and the oceans produces low atmospheric pressure over the continent and high pressures over the subtropical Pacific Ocean and the Indian Ocean. This atmospheric pressure gradient brings the influx of rain-bearing warm air from the oceans, known as the East Asian summer monsoon and the Indian summer monsoon (Domrös and Peng, 1988; Zhao et al., 1990). The dividing line between the East Asian (Pacific) and Indian monsoons lies from 105° to 110°E. During winter, the conditions are reversed. The Siberian–Mon-golian anticyclone results in the flow of cold and dry winter monsoon from north and west. Most of the annual rainfall in the Loess Plateau is concentrated in the months of July, August and September. Mean annu-al temperature varies from 14°C in the south to only 4°C in the north along a gradient of increasing latitude. Annual mean precipita-tion decreases from 650 mm in the southeast to less than 200 mm in the northwest. The Loess Plateau therefore covers the semi-hu-mid, semi-arid and arid climatic zones. Coinciding with the distribution of precipitation, the vegetation zones change from deciduous broad-leaved forest in the southeast through forest steppe and finally to steppe in the northwestern Loess Plateau area. However, nearly all the natural vegeta-tion has been removed by human activity during the past millennia. Nowadays most of the even ground on the Loess Plateau and along the river valleys is intensely cultivat-ed. The majority of the modern riv-ers in the Loess Plateau drain to the Yellow River. The Weihe river is the largest tribu-tary and merges with the middle reaches of the Yellow River at Huayin, where the Yellow River turns eastwards. Because of the high erodibility of the loess, most riv-ers in the Loess Plateau are heavily laden

with reworked loess sediment. The Yellow River has the highest suspended-sediment load in the world, with the total sediment discharge being 1.6 x 109 tonnes/yr (Yang et al., 2000).

2.2. Qinling MountainsThe east-west trending Qinling mountain belt, south of Lantian, forms a sharp physi-cal and environmental barrier that separates Northern and Southern China, and marks the southern boundary of the Loess Plateau. The mountain belt has an average elevation of 2000–3000 m a.s.l. and extends for over 1000 km (Mattauer et al., 1985) across Cen-tral China. The highest peak is Mount Taibai (3767 m a.s.l.). The southern slopes of the Qinling Shan are rather gentle, whereas the northern slopes have a steep east–west trending mountain front. The mountain range separates the subtropical humid cli-matic zone, with evergreen forest habitats in the south, from warm-temperate zone and deciduous, broad-leaved forests on the north side of the mountains (Zhao et al., 1990). The differentiation between the northern and southern slopes occurred in the Ceno-zoic, especially during the Neogene, most probably as a result of the tectonic uplift of the Qinling mountain range (Chen et al., 2001). The mountain belt is also a dividing crest between the middle branches of the Yellow and Yangtse rivers.

The Qinling orogen was formed by the collision of the North China Block (Sino-Korean Platform) and the South Chi-na Block (Yangtse Platform) (e.g. Mattauer et al., 1985; Yin and Nie, 1993; Hacker et al., 1996; Zhai et al., 1998). The bulk of the Qinling orogen is considered to be composed of metasedimentary rocks of amphibolite, eclogite and granulite grade (Hacker et al., 1996). The orogen structurally comprises a northern Palaeozoic belt and a southern

14

Mesozoic belt, both being cut by several east–west, left-lateral, strike-slip faults (e.g. Mattauer et al., 1985). The nature and tim-ing of the collision between the two blocks has long been an issue of debate (reviewed in Hacker et al., 1996; Meng and Zhang, 1999, 2000). On the basis of tectonic stud-ies, Meng and Zhang (1999, 2000) demon-strated that the orogen was built up through a two-stage collision: the southern edge of the North China Block and South Qinling orogen accreted during the middle Palaeo-zoic, whilst the final integration between the two blocks occurred in the Late Triassic. The present-day topography of the Qinling Mountains is mainly the result of Triassic – Cenozoic uplift following intracontinental collision (e.g. Meng and Zhang, 2000).

2.3. Lantian regionThe Lantian region in the southernmost Loess Plateau lies within the southeastern Weihe Graben basin, bounded to the south by the steep mountain front of the Qinling Shan. The general climate pattern of Lantian is characterised by a clear seasonal variabil-ity and monsoonal precipitation pattern. The mean annual temperature is 13.4°C, with mean monthly temperatures ranging from a low of –0.5°C in January to a high of 26.3°C in July. At Xi’an, ca. 40 km from Lantian town, the mean annual rainfall was about 575 mm over the period 1961–1990 (World Climatological Normals, 1996). Most of the rain (60%) falls between July and October. The moist air is mainly brought from the South China Sea by the East Asian sum-mer monsoon although, according to Wang et al. (1997) and Wang and Follmer (1998), Lantian lies on the border of the East Asian and Indian monsoon circulations and the two summer monsoon circulations overlap in this area on a regular basis. The present climatic regime supports vegetation that

belongs to the deciduous broad-leaved for-est zone but nowadays the land is entirely cultivated, wheat and maize being the main crops.

The Bahe river drains the Lantian region with its headwaters in the Qinling Mountains and merges downstream into a major trunk river, the Weihe (Fig. 1). The Bahe river is ca. 100 km long and converges with the Wangyuhe river close to Lantian itself. The Bahe river has developed a se-quence of terrace surfaces and a wide flood-plain on the right (northern) river bank. On the left (southern) bank, the river flows at the foot of the eroded northern flank of the Bailuyuan Plateau exposing a sequence of late Neogene deposits. According to Wang et al. (1966), the asymmetrical topogra-phy of the river valley was caused by a gradual lateral shift of the river course as-sociated with the Lishan uplift during the Pleistocene. Four to five fluvial terraces and underlying sequences have been rec-ognised (Wang et al., 1966; Bellier et al., 1991). Zhang et al. (1995) dated the lower and higher terrace deposits to the Holocene and Early Pleistocene, respectively. The ter-race deposits have been interpreted as being remnants of climatic fluctuations between interglacial and glacial conditions, fluvial conglomerates representing humid condi-tions and overlain by loess deposits of gla-cial times (Zhang et al., 1995). The conclu-sion that the Bahe river was initiated in the Early Pleistocene was also earlier stated by Jia et al. (1966), Wang et al. (1966) and Xie et al. (1966).

The present-day discharge of the Bahe river is characterised by high season-ality and annual variability. However, the current river dynamics are highly modified by human activity. The river banks have been terraced in many places and, during past decades, there has been extensive ex-

15

ploitation of gravel resources by local resi-dents. The sediments that underlie the ter-race surfaces indicate that the Bahe river had a higher discharge and steeper gradient in the Middle Pleistocene than today (Xie et al., 1966).

The Bahe river valley separates two areas of higher relief, the Bailuyuan Plateau and the Henglingyuan Plateau (Fig. 1 in Pa-per II). The ca. 25-km long and 10-km wide Bailuyuan Plateau rises to nearly over 790 m a.s.l. and lies generally 300–320 m above the Bahe river. The northwest-trending Hen-glingyuan Plateau, north of the Bahe river, reaches elevations of 900 to 1000 m a.s.l., bordered by the Lishan Mountain and the Weihe river. Small rivers on the Henglingy-uan Plateau drain southwestwards into the Bahe river. The highest peak in the Lantian region is the Lishan Mountain (1300 m), an uplifted block in the Weihe Graben, located north-east of Lantian. 2.4. Geological background2.4.1. Tectonic settingSince the beginning of the Cenozoic Era, the Sino-Korean Platform underwent wide-spread continental rifting that resulted in the development of several extensional basins. Their formation is often attributed to sub-duction events on the western margin of the Pacific Plate (e.g. Chen and Dickinson, 1986) or related to the collision between the Indian and Eurasian plates (e.g. Tapponnier and Molnar, 1977), or to the combined ef-fect of these two (Grimmer et al., 2002; Rat-schbacher et al., 2003). The Weihe Graben forms part of the Weihe – Shaanxi rift system that runs along the eastern and southern pe-riphery of the Ordos block on the Sino-Ko-rean platform. The Weihe Graben is about 100 km wide and extends for over 400 km in east north-east to west south-west direction in the southern segment of the rift system. It

is assumed to be a pull-apart graben related to the strike-slip faults along the northern border of the Qinling Mountains. Field ob-servations, combined with satellite-imagery analysis, have established three successive extensional regimes (and related subsid-ence periods) in the Weihe Graben (Bellier et al., 1988, 1991; Zhang et al., 1995, 1998, 1999). According to these studies, subsid-ence in the graben started in the middle Eocene – early Oligocene and was related to west north-west – east south-east exten-sion. During the middle Neogene (mid-dle-late Miocene), the tension was directed north-east – south-westwards. The change to the late Pliocene – Pleistocene north-west – south-westwards orientated extension was established between 9 and 2.4 Ma. Seismic reflection profiles and boreholes reveal that the sedimentary fills of the Weihe Graben are wedge-shaped, and dip south-eastwards towards the bounding Qinling frontal faults (e.g. Wang, 1987; Bel-lier et al., 1988) (Fig. 2.). Within the graben, the southwards-tilted fault block of Lishan forms the main body of the Lantian region (Zhang et al., 1978).

2.4.2. General stratigraphyThe Weihe Graben basin is infilled by sedi-ments assigned to the Honghe, Bailuyuan, Lengshuigou, Koujiacun, Bahe and Lantian formations, each bounded by unconformi-ties. These breaks most probably represent tectonic events related to different phases of subsidence of the graben. Sedimentation continued throughout the Cenozoic, except for during the late Oligocene – early Miocene when subsidence in the graben temporarily ceased (Bellier et al., 1988). The strata are generally blanketed by Pleistocene loess de-posits. The existing stratigraphical scheme for the Lantian region was originally estab-lished on the basis of results obtained from

16

field observations undertaken during the 1960s (Jia et al., 1966; Zhang et al., 1978). Since then this stratigraphy has served as the basis for stratigraphical correlation in the basin. Subsequent studies based on verte-brate paleontology and magnetostratigraphy (e.g. Qiu and Qiu, 1995; Zheng et al., 1992; Sun et al., 1997) have, in part, refined the scheme, mainly in the Neogene part of the succession. The lithostratigraphical units consist primarily of interbedded conglom-erates, sandstones and siltstones reflecting depositional systems of alluvial fans, fluvial channels, floodplains and lakes.

The Honghe Formation (260 m) is the basal Cenozoic unit in the Lantian re-gion, representing the initial sedimentation in the subsiding graben basin. The basal conglomerate of the Honghe Formation on-laps over Precambrian crystalline basement rocks. Two fossiliferous units of assumed early(?) Late-Eocene age have been report-ed from the lower and upper members of the formation.

The Honghe Formation is overlain by a clastic succession of late Late Eocene to Early Oligocene age. This age assign-ment for the Bailuyuan Formation was con-strained on the basis of fossil vertebrates found in the lower and upper members of the formation. The formation outcrops

around the Lishan Mountain, except for the type section of the upper member that is de-fined from Maoxicun on the opposite side of the Bahe river. The composite thickness of the sequence is ca. 400 m, as estimated from the type sections of the members of the Bailuyuan Formation.

The overlying Lengshuigou Forma-tion is named after the gully from which the type section was described. The exposures are mainly distributed around the Lishan Mountain, and reach maximum thicknesses of ca. 80 m. The Lengshuigou Formation rests on a basal angular unconformity that represents a hiatus of several million years (cf. Zhang et al., 1978).

The distribution of the subsequent Kojiacun Formation is similar to that of the underlying Lengshuigou Formation. How-ever, the most complete section has been de-scribed at Koujiacun village, next to Maoxi-cun, on the southern bank of the Bahe river. The Lengshuigou and Kojiacun formations were defined by Zhang et al. (1978) as en-compassing the Middle and Late Miocene based on fossil vertebrate evidence. Qiu and Qiu (1995) refined the earlier interpretation and determined the age of the two forma-tions as Middle Miocene. They also show that the paleontological evidence contra-dicts the original correlation since, on the

WeiheLi Shan

Qinling Shan

PuchengBeishan

0 4 8 12 Km

-6000

-4000

-2000

0

2000

4000

SSW NNE

+

+

+

++

++

++

++

++

++

++

+

+

+

P1

N1+2

QQ

N2

N1

Study area

metr

es

Ar+Pt

Pz

P1

Figure 2. Simplified cross-section of the Weihe Graben east of Xi’an. Modified after Wang (1987). Ar+Pt – Archaean and Proterozoic; Pz – Palaeozoic; Q – Quaternary; N1 – Pliocene; N2 – Miocene; P1 – Eocene.

17

basis of more developed fossil assemblages, the Lengshuigou Formation should postdate the Koujiacun Formation.

The overlying Bahe Formation is the main focus of this thesis. It is widely observed in the Lantian region from the western foot of the Lishan Mountain to the northern flank of the Bailuyuan Plateau and eastwards in the most river valleys. The most extensive outcrops of the formation are identified on the northern flank of the Bailuyuan Plateau where the formation is reported to attain its maximum thickness of 350 m (Zhang et al., 1978). The stratotype of the formation has not been defined as such. The lower member is present within the confines of the villages of Koujiacun and Maoxicun where it rests disconformably on the Koujiacun Formation. The upper mem-ber is best exposed in Shuijiazui (the Main Section in this study). According to Jia et al. (1966) and Zhang et al. (1978), the lower member comprises whitish-grey sandstones and red-brown siltstones. The upper mem-ber, by contrast, is characterised by red and yellowish siltstones, sandy siltstones, with interbeds of sandy conglomerates and sand-stones. The Bahe Formation shows varia-tions in its lithological character across the stratigraphical sections. This variation re-flects changes in the regional palaeotopogra-phy and structural control. Based on its con-tained vertebrate fossil faunal assemblage, the Bahe Formation has been previously er-roneously assigned to the early Pliocene (Jia et al., 1966; Zhang et al., 1978) but this age was later revised to late Miocene by Li et al. (1984), Qiu (1990), Qiu and Qiu (1995), Paper I and Paper IV.

The Lantian Formation, consisting of a basal conglomerate of varying thick-ness, and an upper unit of red-brown silt-stones, overlies the Bahe Formation with a regional angular unconformity. The sedi-ments in the upper fine-grained unit of the formation are similar to other ‘Red Clay’ deposits in the Chinese Loess Plateau re-gion, being mottled red-brown clayey silts and having abundant nodular carbonate and Fe-Mn-coatings (e.g. Ding et al., 1999; Han et al., 2002). Various studies of grain size, geochemistry and pedology have shown that the ‘Red Clay’ deposits in Northern China are mainly aeolian in origin (e.g. Ding et al., 1998; Guo et al., 2001; Lu et al., 2001), and the deposits are thought to record the evolution of the Asian atmospheric circula-tion and related uplift of the Tibetan Plateau (An et al., 2001; Guo et al., 2002). The basal conglomerate of the Lantian Formation be-comes thicker and coarser grained towards the Qinling Mountain front, while the red clays and siltstones overlying the conglom-erate become progressively thinner. The thickest and most complete exposures of the Lantian Formation are found on the Bailuy-uan Plateau in the Shuijiazui where the for-mation may reach thicknesses of over 60 m (Jia et al., 1966), with a basal conglomer-ate of ca. 2 m thick (author’s observations). East of the town of Lantian, the formation thins to ca. 10 m, including ca. 5 m of the basal conglomerate (Jia et al., 1966). The Lantian Formation was originally postulat-ed to be of late Pliocene-age but more recent studies using magneto- and biostratigraphy have shown the base of the formation is of latest late Miocene age (cf. Sun et al., 1997; Paper IV).

18

3. Material and methods

The best exposures of the Bahe and Lantian formations outcrop in vertically rather con-tinuous sections along the Bahe River on the northern slope of the Bailuyuan Plateau (Fig. 3a). The strata are gently tilted and gener-ally young from west to east. In addition, the lateral extent varies from tens to hundreds of metres but is often covered locally by the vegetation of cultivated farmlands.

3.1. SedimentologyStratigraphical exposures of the Bahe and Lantian formations were measured using a Jacob’s staff and Abney level at 30 sites totalling over 1000 m of logged section. The studied sections were established on the basis of field investigation and most productive fossil-find localities. The main localities provide profiles exceeding 130 m in height, whereas the smallest logged outcrops were only a few metres high. The spacing between each section ranges from tens of metres to three kilometres. Detailed lithological sections were constructed from the studied sections (Fig. 4 A and B in Paper II). The sections were studied using standard sedimentological techniques (cf. Miall, 1996). Lithofacies were determined and recorded on a bed-by-bed scale. Vari-ous facies were identified based on grain size, physical sedimentary structures, bed-ding contacts, bioturbation, paleosol forma-tion, paleocurrent pattern and geometry of units. In some places, the lateral continuity of the exposure was sufficient to allow for observation of lateral facies changes and large-scale bed geometries to provide addi-tional information for the interpretation of depositional environments. However, verti-cal facies relationships were more common than lateral ones.

Paleocurrent directions were stud-ied throughout the sections mainly by measuring dip azimuths of the foresets of planar cross-stratified units and trough-axis orientations of trough-cross strata. In addi-tion, paleoflow was determined by measur-ing imbricated pebbles and cobbles within small outcrop areas of individual litho-facies units. At each outcrop a minimum of 50 clasts were selected. In total, over 360 paleocurrent indicators were measured at 25 localities. All the data obtained in the field were corrected for tectonic tilt, using the stereographic projection method described by Tucker (2003). Maximum clast size was deter-mined by measuring the a-axis of the ten largest clasts within a given bed, and meas-urements were made from most conglomer-ate beds throughout individual sections. For the petrological composition, ten medium to coarse-grained samples from the Bahe Formation were selected for thin-section analysis. Five hundred points were counted from each thin section. The 0.125–0.250 mm size fraction was analysed. These sedimentological techniques allow the differentiation of the major depo-sitional environments preserved in the Bahe and Lantian formations, the identification of smaller-scale environments (such as channel, bar, crevasse splay, floodplain and pond), and the use of these distinctions to make interpretations on the variability of the depositional processes throughout the study area.

3.2. IsotopesPedogenic carbonate was collected for iso-tope analyses from both the Bahe and Lan-tian formations in several stratigraphical exposures, but the sampling was mainly concentrated on the Main Section and Liu-jiapo Section. A total of 114 samples were

19

taken. Laboratory inspections using an X-ray diffractometer (XRD) (n=20) ensured that calcite was the dominant component in the nodules, and thin-section studies (n=10) have revealed the dominance of homog-enous micritic calcite (generally <5μm). They indicated that the carbonates had un-dergone minimal recrystallization. The outermost surface of the car-bonate nodules was etched away using hy-drochloric acid (10%). Carbonate samples were dried and ground. The fine ground pedogenic carbonate samples were then treated with phosphoric acid (1.91 g/cm3) at 25°C to generate carbon dioxide. The evolved CO2 gas was analysed for carbon and oxygen isotopes using the Finnigan MAT Delta E isotopes ratio mass spectrom-eter in the Dating Laboratory, University of Helsinki. The results are reported in the conventional δ notation (where δ = [(Rsa/Rstd) – 1] x 1000‰. Rsa and Rstd are the ra-tios of 13C/12C and 18O/16O in the samples and standard, respectively.) The standard for both oxygen and carbon in carbonate is V-PDB. Parallel measurements were taken to check the repeatability of the measure-ments. The carbon and oxygen isotope ratios in pedogenic carbonates mirror the prevailing climatic and environmental con-ditions. Carbon isotope ratios in paleosol carbonates are related to the proportions of C4 and C3 plants growing at the site. Con-sequently, the 13C/12C variations in paleosol carbonates in a stratigraphic sequence re-flect past changes in C4 and C3 plant abun-dances in biomass through time (e.g. Cer-ling, 1984, 1991; Quade et al., 1989). The oxygen isotope ratio is more complex since it is a function of several climatic factors such as temperature, evaporation, amount of rainfall and sources of precipitation (e.g. Cerling, 1984; Quade et al., 1995).

3.3. MagnetostratigraphySeven stratigraphical sections (Liujiapo, Locality 42, Pyramid, Main, Locality 12, Locality 12/19, Locality 19) form the basis for the paleomagnetic investigations in this study. Orientated samples from claystones, siltstones and sandy siltstones were taken using a portable petroleum-driven rock drill and a compass. Normally, two or three cores were collected per site. In a few cases, where drilling was not possible, for example in sandy beds, block samples were taken. Loose and disturbed material was removed. A smooth surface was created at the outcrop and then the orientation was measured using a Brunton compass. The blocks were cut or drilled into several specimens in the labora-tory. The collection for paleomagnetic measurement was done in two stages. First, in 1998, a general reversal pattern for the study area was determined by collecting samples from the Main Section, Pyramid Section and Huijiaxincun Sections. They were then analysed at the University of Tex-as at Austin. The samples were subjected to alternating field demagnetisation using a Schonstedt AF demagnetiser GSD-1. A few supplementary samples were analysed at the paleomagnetic laboratory of the Geo-logical Survey of Finland, using a three-axis SQUID magnetometer with an automatic AF-demagnetiser. Two years later, the Lo-cality 42 and Liujiapo sections were sam-pled. In addition, based on the results of magnetic-polarity determinations from the first sampling excercise, two intervals in the Main Section were sampled in detail and also extended in the lowermost part of the section. These samples were measured at the Laboratoire de Paléomagnétisme et Géodynamique, Institut de Physique du Globe de Paris, France. The samples were

20

subjected to thermal stepwise demagnetisa-tion and their magnetisation measured with a 2G three-axes cryogenic magnetometer in an antimagnetic room. The final magnetic zonation is based upon 470 stratigraphic levels in the Bahe and Lantian Formations. A typical site spacing of one site per 1.2 m was achieved for the Bahe Formation, and 0.4 m for the Lantian Formation, depending on the suit-ability of the rock and outcrop availability. The spacing was greater in the Huijixincun sections were the sediments were generally coarser. There are no radiometrically date-able rocks in the Lantian sequence. There-fore, magnetostratigraphy is the principle means by which the important Lantian ver-tebrate fossil localities could be linked to an absolute time scale. It also provides a solid basis for discussion of the deposition history in the southeastern Weihe Graben basin, as well as the paleoenvironmental evolution of the area.

3.4. TaphonomyMost of the larger vertebrate fossil speci-mens, and some of the smaller ones were discovered by excavation and prospecting the surface of naturally eroded outcrops by the project members. In total, 55 fossil localities were discovered, the majority in stratigraphic superposition, 52 of which oc-cur within the Bahe and Lantian formations (44 and 8 localities, respectively). Twenty-six of the fossil localities included mamma-lian material of Hipparion fauna (cf. Kurten, 1952). In addition to mammalian remains, the localities yielded fish, amphibians, rep-tiles, freshwarer molluscs and charophytes (Qiu et al., 2003). The total number of cata-

logued large mammal specimens recovered is 1667. Nine localities (6, 12, 30, 31, 34, 42, 44, 45, 49) were excavated to some ex-tent; however, specimens were not consist-ently collected and catalogued according to their occurrence in the field. Lithological data (and a sketch map) were recorded from most sites and, when possible, the fossil localities were placed in the stratigraphi-cal framework. Detailed sedimentological logging of fossil locality was conducted at ten sites. Screen-washing bulk samples pro-vided large collections of small mammals, such as rodents and insectivores. Over 20 sites were screen-washed, and most of these localities also have remains of large spe-cies. The entire collection is housed as the Lantian collection at the Institute of Verte-brate Paleontology and Paleoanthropology (IVPP), Beijing, China. An intergrated study of the weath-ering status and sedimentology was carried out from three localities. These sites were chosen because they are fossiliferous and well represent the typical lithological con-text of fossil sites in the Lantian sequence. The specimens included in the study repre-sent a random sample. Studying the weath-ering and abrasion status of the skeletal element enlightens the preburial history of the bone assemblage (e.g. Behrensmeyer, 1978), whereas determination of deposi-tional environment of the fossil locality is required in order to interpret the processes that formed a fossil assemblage. Weather-ing reflects mainly the time of exposure on a land surface, while abrasion is an indica-tor of distance travelled and intensity and/or time of interaction with moving sediment (e.g. Behrensmeyer, 1982, 1991).

21

4. Results

4.1. SedimentologyMost strata dip in a south-southeast – south-southwest direction, varying from 8º in the Main Section to 14º in the lower part of the Liujiapo Section. Although tectonic dis-placements were observed, the strata are relatively undisturbed.

4.1.1. Bahe FormationOn the basis of facies and facies association analysis, the Bahe Formation shows basic sedimentary characteristics that could be classed as it having been deposited within a fluvial system (papers I and II). Evolution of this sequence is represented by the alter-nation of coarse-grained channel-fills with overbank deposits. These sequences repre-sent a variety of proximal to distal flood-plain depositional settings.

The channel fills observed in the Bahe Formation are coarse-grained and show an almost complete lack of lateral ac-cretion features (Fig. 3b). There are a few well-confined channel forms, but overall the channel margins show mainly low dips and often form broad sheets in the outcrops. This suggests that these channel fills can be attributed to frequent shifting of a river channel in the basin, indicative of a channel with relatively low sinuosity.

The Bahe Formation is character-ised by the dominance of overbank strata in the sediment succession. The overbank sediments result from deposition by bed-load and suspended load on the floodplain outside the main channels resulting from overtopping of the banks during river floods or avulsions. Massive, occasionally sandy, fine-grained deposits with characteristically indications of palaeosol formation, form the bulk of the floodplain sediments and of the

studied sediment sequence in general (Fig. 3c). Most of these overbank fines are inter-preted to represent well-drained floodplain facies assemblage, the key diagnostic crite-ria being their red-brown colour, the com-mon occurrence of carbonate nodules, and the absence of coal. The poorly-drained floodplain assemblage is often interlayered with the well-drained floodplain deposits, but it is less common. This assemblage has a distinguishing greyish-green colour and commonly contains rhizoconcretions. Apart from these fine-grained units, the continuum of overbank deposits also includes crevasse splay and sheet-flow sands, organised in thin, commonly less than 1 m-thick, fining-upward sequences. The floodplain depos-its also comprise a few lake marl deposits that form resistant ridges providing reliable guide horizons for correlation.

Nodular calcite formation is com-mon in the overbank deposits and points towards the existence of a semi-arid type of climate. The abundance of pedogenical overprinting in the sediments in general indicates subaerial exposure and relatively slow sedimentation rates in the sequence.

The exposed sediments of the Bahe Formation do not indicate any distinct changes in the fluvial regime through time. There is, however, a large-scale upward fin-ing trend in the Bahe Formation. Given that there are no marked temporal differences in the palaeosol character, this upward-fining trend has been related to tectonic or auto-cyclic dynamics of the fluvial system rather than to climatic factors.

4.1.2. Lantian FormationLithologically, the Lantian Formation, also termed informally the ‘Red Clay’ is char-acterised by a monotonous succession of ca. 50 m of structureless, red and reddish brown silty clay and clayey silt with as-

22

Bahe Fm

Lantian FmLoess

Gm, Sp

Fm+c

Figure 3. (a) General view to the northern slope of the Bailuyuan Plateau with the modern Bahe river in the foreground. (b) A typical upward-fin-ing fluvial sequence in the Bahe Formation. (c) Repetitive alteration of the fine-grained flood-plain deposits with palaeosol horizons and inter-calated minor sand deposits in the lower part of the Locality 42 Section.

(a)

(b)

(c)

23

(d)

(e)

Figure 3. (cont.) (d) The Liujiapo Section showing the Bahe – Lantian Formation boundary. The boundary horizon is indicated with an arrow. (e) Detail of the carbonate nodule horizon in the upper part of the Lantian Formation.

Bahe Fm

Lantian Fm

→

24

sociated carbonate nodules. The reddish brown clay and siltstones are rich in ferro-manganese coatings and carbonate nodules. Slickensides, rhitzoliths or green mottling were not detected, features that are typical of the floodplain facies in the underlying Bahe Formation. The carbonate nodules ap-pear scattered throughout the formation, but also as broad zones and complexes in which white nodules are profuse. The upper and lower contacts of these nodule-rich horizons are gradational. The lower boundary of the Lan-tian Formation is defined by a conglomerate stratum that reaches an average thickness of 1–2 m and a maximum of 3 m (Fig. 3d). The facies architecture of the so-called bound-ary conglomerates is channelised or tabu-lar and the unit often exhibits plane- and trough-cross stratification or is horizontally interbedded with sandstones. Overlying the conglomerate, fine-grained sandstones and granule lenses, interbedded with red-brown silts and clays, are present in minor abun-dances. Sand bodies above the conglomer-ate are, in general, tabular with thin massive beds that are often calcareous. From the upper part of the se-quence there are fewer detailed stratigraphi-cal or sedimentological remarks, mostly as a consequence of the uniform succession. A striking feature in the uppermost Lantian Formation is the existence of very hard, whitish, nodule-rich layer, up to 2 m thick, close to the top of the ‘Red Clay’ exposures (Fig. 3e). This thick horizon is, in practice, the only nodule bed in the Lantian Forma-tion that is straightforwardly traceable to the other sections in the study area. The princi-pal changes present in the Lantian Forma-tion are reduction in the grain size, with no sand or gravel lenses appearing in the up-per part of the formation, and variations in CaCO3 content. The formation is thoroughly

calcareous with CaCO3 content varying be-tween ca. 10–80%. Although this content displays a wide scattering of values, there is an obvious trend of increasing values to-wards the top of the formation. The ‘Red Clays’ have been inter-preted as being aeolian in origin (e.g. Sun et al., 1997, 1998; Ding et al., 1998; Guo et al., 2001; Lu et al., 2001) but field observa-tions show that lower part of the ‘Red Clay’ sequence in Lantian is at least partly of flu-vial origin. However, the fluvial imprint is restricted to the lowermost part of the Lan-tian Formation. The cessation of major flu-vial channels indicates that the palaeo-Bahe was deflected from this area during deposi-tion of the bulk of the Lantian Formation.

4.2. MagnetostratigraphyRock magnetic studies show that sediments from the Bahe and Lantian formations are reliable recorders of the geomagnetic field polarity. The resulting magnetostratigraphi-cal columns contain a relatively high number of reversals that have been correlated to the GPTS using the distinctive sequence and pattern of polarity reversals, supported by the information derived from the bio- and lithostratigraphy. The sequences studied are long, and with the exception of the Huijiax-incun sections, allow independent correla-tions to the GPTS. In general, the correla-tion of the magnetic polarity pattern to the GPTS is virtually straightforward. The dis-tinctive long intervals of normal polarity in the Main and Locality 42 sections were used as the principal anchor points for the GPTS correlation and were correlated, respec-tively, with chrons C5n.2n and C4n.2n. The remaining magnetozones were correlated to the GPTS based on the assumption that most chrons were detected as a result of the high sampling density. As a consequence, the preferred correlation constrains the sampled

25

sedimentary sequence to the period from ca. 11 Ma to 3 Ma. According to the correlation es-tablished in the study, the Bahe Formation was laid down between 11 and 6.8 Ma. This age span is in good agreement with the bios-tratigraphical control, which suggests a late Miocene age. According to a convention adopted by earlier scholars, the stratigraphi-cal boundary between the two formations is interpreted to coincide with a specific con-glomerate stratum. Interpolation of the ac-cumulation rate yields a basal age of 6.8 Ma for the Lantian Formation (i.e. to the base of the so-called boundary conglomerate), which is consistent with the earlier studies conducted in the area but is younger than the base of the ‘Red Clay’ deposits in other areas of the Loess Plateau.

The age control allows calculation of the accumulation for each polarity zone, i.e. for well-defined time intervals. The de-rived average sedimentation rates for the Bahe Formation are fairly uniform at the different localities (Fig. 4) being between 4 and 6 cm/ka. Figure 4 also demonstrates that although the average sedimentation rates remain relatively constant, there are differ-ences in the accumulation rates within the individual sections. Such short-term vari-ations are typical in fluvial sequences and may result, for example, from fluctuations in erosion and transport, or from deposi-tional hiatuses. The irregularity of the ac-cumulation curve between the chrons may also result from inconsistencies in sampling density and from interpolating the chron boundaries between the sampling horizons.

Figure 4. Plot of stratigraphical thicknesses with absolute time scale. The polarity column is from GPTS of Cande and Kent (1995); black denotes normal polarity and white denotes reversed polarity intervals.

0

20

40

60

80

100

120

140

Time (Ma)

Sedi

men

tthi

ckne

ss(m

)

234567891011

Main Section

Liujiapo Section

Locality 42 Section

Pyramid Section

Locality 12

Locality 19

Locality 12/19

Lantian FmBahe Fm

26

The Bahe–Lantian formation bound-ary marks a prominent change in accumula-tion rates. The Lantian Formation displays lower sedimentation rates of, on average, 1.38 cm/ka. The stable sedimentation rates im-ply a relatively constant tectonic regime during the deposition of the Bahe Forma-tion.

4.3. Fossil localities and tapho-nomyOnly a few sites, out of 49, were considered to be productive. The six richest fossil lo-calities (6, 12, 30, 31, 42, 44) account for most of all fossil material (Fig. 5). Of the most productive fossil vertebrate localities, two are in the lowermost part of the Lantian Formation (Locality 42 and 44). Except for these, very few fossil vertebrate specimens were found from the Lantian Formation and none from the upper half of the formation, despite extensive prospecting. Most locali-ties comprise a single sedimentary stratum

as a source of the fossil bone material but sometimes they comprise multiple strata, such as in Locality 6, and do not always coincide with lithological changes. It is evi-dent from Figure 6 that most fossil localities in the Bahe and Lantian formations occur in fine-grained sediments (clays and silts). Coarser-grained sediments (sandstones and conglomerates) account for one third of the sites. Eight sites were documented in a unit showing clear indications of palaeosol for-mation (root traces, carbonate nodules). The host sediment in these units may be clay, silt or sand.

4.3.1. Weathering analysisThe combined sedimentological and weath-ering evidence from Locality 31 conclusive-ly indicates a crevasse splay setting. Fossils exhibit almost no indications of abrasion due to transport nor were they exposed to weathering or scavenging prior to deposi-tion. The skeletal elements are articulated and aligned along a single foreset bed. In

Figure 5. Number of fossil specimens collected in individual fossil-find localities. Note a logarithmic scale on vertical axis.

1

10

100

10001 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55

Fossil locality

No

offo

ssil

spec

imen

sco

llect

ed

27

consequence, it seems likely that animals or carcasses were introduced to the floodplain by a sudden avulsion and covered within a short time (Paper III). The sedimentological evidence indicates that Locality 6 represents suc-cessive sheet flood deposits showing clear indications of palaeosol formation. The fos-sils are generally found disarticulated, heav-ily weathered and often abraded. There are also indications of carnivore activity and trampling. Equally with the sedimentologi-cal features, there are two modes of fossil accumulations: the scattered and weathered bones in the calcareous horizons are inter-preted as representing autochtonous attri-tional mortality on the soil surface whereas the concentrations of bones are mostly as-sociated with sheetflood events that both sorted and transported skeletal remains.

The depositional setting in Local-ity 42 can be interpreted to be one of a mass movement. As a consequence, the fossil as-semblage comprises a collection of elements at various stages of degradation that may de-rive from catastrophic mortality in a debris flow, reworked remains from older deposits as well as from the floodplain surface, both locally derived and transported.

4.4. Stable carbon and oxygen isotope recordBoth δ13C and δ18O show variation within the sampled time period 10–2.7 Ma (Paper V). The distinct trends in the record permit the separation of the profile into two dif-ferent intervals (Fig. 2 in Paper V), which correspond to the Bahe and Lantian forma-tions. The average values through the Bahe Formation are –8.71 ± 0.20‰ and –11.06 ± 0.26‰ for δ13C and δ18O, respectively. In the Lantian Formation, the δ13C values fall in the range –10.76‰ to –7.27‰ and δ18O between –8.00‰ and –11.36‰. The stable pattern of the isotope record in the Bahe Formation suggests that no marked change in local vegetation or climate took place during ca. 10–6.8 Ma. All δ13C values are below the threshold value –8‰ for pure C3 vegetation. Most measurements are, how-ever, in the upper C3 plant range, suggesting that water stress conditions prevailed in the area.

In general, the transition between the formations appears to be characterised in the isotope record by a considerable scatter in values within and between the sampled stratigraphical levels. Another feature is the decrease in average carbon isotope and in-crease in oxygen isotope values in the Lan-tian Formation compared to those from the Bahe Formation. The up-section decrease in δ13C values is interpreted as resulting from increased water supply although it may also reflect cooler growing season tempera-tures. Interpretation of this shift in oxygen isotope ratios at the formation boundary is not straightforward. The relatively enriched δ18O values may seem to be in conflict with the more depleted δ13C values in the Lantian Formation. However, considering the inter-preted increase in summer rain from the car-bon isotope records, it is likely that the en-

Figure 6. Fossil localities by lithology from the Bahe and Lantian formations.

28

riched oxygen isotope values were mainly caused by a shift to a seasonal pattern with increased summer precipitation.

The broad scatter of δ18O values is explained as probably produced by over-printing resulting from low sedimentation rates. Other sources of scatter may include changes in the water source or fluctuations in the intensity of summer monsoon rain-fall. In addition, the highly variable oxygen isotope ratios may be attributed to natural variability of meteoric water between the summer and winter monsoons, or may re-flect seasonal differences in infiltration of precipitation to the soil column.

The carbon isotope data suggests that C4 biomass was absent in Lantian in the late Neogene. This contrasts with the obser-vations made in Lingtai in the Chinese Loess Plateau where an expansion of C4 plants is documented at ca. 4 Ma ago. Evidence from other areas outside China, such as Pakistan,

Nepal, Africa and the Americas, also indi-cates a late Miocene emergence of C4 grass-es in the ecosystems. The nearly complete lack of evidence of C4 biomass in Lantian is explained in terms of a cool rainy season favouring the growth of C3 plants over C4 plants. The Lantian δ13C record suggests that despite the lowered CO2 concentra-tions in the atmosphere of the Miocene that favour C4 plants over C3 plants, the expan-sion of C4 grasslands to the Chinese Loess Plateau was largely controlled by regional climatic factors.

It is notable that the isotope data display a strong positive correlation in the Lantian Formation but not in the Bahe For-mation. This relationship indicates that dur-ing the deposition of the ‘Red Clay’ a strong environmental control, probably climate, af-fects both isotope ratios whereas in the Bahe Formation, isotope ratios reflect more the response to specific local conditions.

29

5. Discussion

5.1. Completeness of the sedi-mentary record in Lantian se-quenceThe magnetic polarity succession for the Lantian sequence is nearly complete, which is indicated by the fact that almost all the known magnetic reversals have been re-corded (Paper V). However, the accumula-tion history cannot be considered complete in terms of the lithological nor chronologi-cal record preserved in the sediments. It is well established that rates and styles of dep-osition in a sedimentary basin exhibit large spatial and temporal variability and that a stratigraphical section at a particular site is a net product of deposition, erosion and hia-tuses of varying length and magnitude (e.g. Sadler, 1981; Beer, 1990; McRae, 1990a, 1990b). Fluvial systems are especially dy-namic, and the sediment accumulation rates within the system vary considerably from large channels to floodbasin ponds and to virtual non-deposition where mature pal-aeosols develop. For example, in their con-ceptual model for sediment accumulation in fluvial sequences, Friend et al. (1989) used the following minimum estimates reported from modern environments; 0.1 mm/yr for sandstone/conglomerate units, 0.001 mm/yr for siltstone/mudstone units and 0.0005 mm/yr for mature palaeosols. The rates should be considered as rough approximations, and may vary by several orders of magnitude (e.g. Sadler, 1981).

Facies evidence, considered alone, gives indications of the completeness and time restored in sedimentary units. The Bahe Formation strata are characterised by relatively fine-grained deposition on the floodplain with pulses of sheet flows and common palaeosol development. Sheet flows are typically short-lived (on the scale

of hours), while extensive pedogenesis is commonly associated with time spans on the order of 102–104 yr (Bown and Kraus, 1981, 1987; Retallack, 1984, 1986). Togeth-er with the nature of deposition, erosion, representing ‘missing time’ in the sequence, is an important factor affecting the com-pleteness of strata. In the Bahe Formation, the sand units originated from unconfined sheet flows. Since they generally show no scour in their base, these sheet floods were not highly erosional during their emplace-ment, although the flow power may have been high. The pervasive channel deposits in turn, can be associated with erosional re-gimes in the strata. In the Bahe Formation, more erosion is recorded in the lowermost part of the studied sequence, and in the Hui-jiaxincun area.

The accumulation histories of dif-ferent stratigraphical sequences and differ-ent time spans are compared in Figure 4 where the short-term variation in deposi-tion rates within the individual sections, as well as their lateral variability in deposition rates, can be distinguished. The lateral vari-ability in accumulation rates is apparent, for example, in the Hujiaxincun area (Locality 12, 12/19 and 19 sections). In this area, the variability can also be seen as lateral facies changes in the field. However, when examin-ing the overall long-term accumulation rates in the Bahe Formation, the pattern becomes stable. Hence, although the deposition has been discontinuous at any one site and has been characterised by individual sedimenta-tion events, over the whole region, in gen-eral, it has been stable or ‘continuously epi-sodic’ sensu McRae (1990b).

In the Lantian Formation, the sedi-mentation rates are much lower, especially within the uppermost sampled interval be-tween ca. 4.3–3 Ma, which is coeval with the thick, nodular caliche horizon in the upper-

30

most part of the ‘Red Clay’ sequence. Thick nodular palaeosols are commonly taken to represent depositional hiatuses or periods of very low sedimentation rates (e.g. Retal-lack, 1984; Kraus and Bown, 1986). How-ever, the variability of accumulation rates over the individual magnetozones is less prominent compared to the underlying Bahe Formation. The lateral facies distribution seems to remain relatively uniform without distinct erosion surfaces except at the lower boundary. Given the considerations above, the Lantian Formation at the site studied seems to offer a more complete and continu-ous stratigraphical record than the underly-ing Bahe Formation. However, considering the exceptionally low sedimentation rates in the Lantian Formation, the time resolution associated with the Bahe Formation may be regarded as exceeding that present in the preserved Lantian Formation.

In many stratigraphical succes-sions, there is an inverse trend of decreasing deposition rates with increasing time, inter-preted as resulting from post-depositional compaction and a greater opportunity for se-quences to contain more and longer hiatuses (Sadler, 1981). This trend is not observed within the Lantian sequence.

Long-term sediment accumulation histories in depositional basins can be con-sidered as being controlled by a combina-tion of depositional style and extra-basinal factors such as tectonism and climate. The sediment accumulation histories together with the facies character reflect that there were no marked changes in these dominant controlling factors during the deposition of the Bahe Formation. This further suggests that there was a long-term balance between sediment supply and basin accommodation space attributable to constant subsidence and tectonic uplift of the basin margin.

The contrast in the sedimentation

rates between the Bahe and Lantian forma-tions is related to the different sedimentation processes in their respective depositional environments (i.e. predominantly fluvial and aeolian, respectively), combined with the tectonic and climatic factors. The nature of the formation boundary will be discussed in chapter 5.4.

5.2. Distribution of fossil verte-brate localitiesVertebrate fossil sites in the Bahe and Lan-tian formations can be classified by their dis-tribution according to lithology, depositional environment, number of sites and specimens per time unit and lithostratigraphical unit. As mentioned earlier, the majority of large mammal fossils were collected from fine-grained sediments, minor sandstone de-posits, or from palaeosol units formed on fine-grained or sand beds. These sediments and soils are interpreted as representing floodplains. Only a few sporadic skeletal re-mains were found in major channel-lag and cannel-fill deposits (cf. Papers I, II, III). The greater frequency of floodplain assemblag-es is expected because floodplains are the most prevalent environment in the Lantian sequence. This is not a general rule: for in-stance, in the Palaeocene lower Fort Union Formation of Wyoming, the most produc-tive fossiliferous facies is not the dominant one (Badgley et al., 1995). On the whole, the vertebrate fossils seem to occur princi-pally in all major recognisable depositional contexts on the palaeo-Bahe floodplains in-cluding crevasse splays, ephemeral ponds, well-drained land surfaces, topographic depressions and flash floods. The only de-posits in the sequence that lack vertebrate fossils are the lake marl deposits. Four sites that were associated with the ‘Red Clay’ deposits yielded only a few specimens. At first this could indicate

31

that there was a collection bias in favour of the Bahe Formation sediments, but this was not the case as much effort was equally put into prospecting the Lantian Formation. The scarcity or nearly total lack of vertebrate fossil remains from the ‘Red Clay’ may rather reflect low deposition rates, in which the bones become destroyed before burial. Moreover, a more humid environment in the Lantian Formation, as suggested by isotope data and fossil assemblages, may also ex-plain the poor fossil preservation. It is note-worthy that the Lantian Formation Locality 42 accumulation bears close similarities to the late Neogene Pikermi and Samos red bed localities in Greece. Equally with Local-ity 42, these classical Greek localities yield generally well preserved vertebrate fauna from lens shaped, densely packed bone accumulations, interpreted to have been accumulated in local depressions (Solou-nias, 1981). In general, the origin of fossil vertebrate accumulations in the ‘Red Clay’ deposits of northern China has remained un-clear. Limited information is available from Zdansky’s classic paper (1923). This early work acknowledges the similar lithological character of the Baode ‘Hipparion Clay’ and Pikermi deposits, and describes most fossil vertebrates at Baode as occurring in later-ally restricted ‘nests’. Further comparison of the Samos and Pikermi sequences with those from Lantian and other ‘Red Clay’ sites could contribute significantly to an un-derstanding the origin of fossil localities in the ‘Red Clay’ deposits of Northern China in general. Although pre-burial history of the fossil assemblages was studied only at three localities, some general assump-tions on the nature of fossil accumulations in the Lantian sequences can be proposed. The fine-grained units were deposited un-der relatively low velocity flow conditions

suggesting that fossil assemblages in these deposits were subjected to minimal trans-port prior to burial (Fig. 6). The skeletal ele-ments recovered from the sand deposits are likely to have been transported for some dis-tance and partly sorted by the current flow. However, since the amount of fossil remains found from coarse-grained channel fills or lags is negligible, it is hypothesised that the fossil assemblages in the Lantian sequence represent their principal living habitats, i.e. they are autochtonous or parautochtonous fossil assemblages (sensu Behrensmeyer and Hook, 1992). On the basis of integrated work on weathering status and sedimentol-ogy conducted at three localities (Paper III), combined with lithological information col-lected from the remaining fossil sites (Paper I), it also seems to be highly probable that the bone concentrations are a product of both attritional mortality and sedimentation processes as well as, although to a lesser ex-tent, of biological agencies like carnivorous mammals. It is well recognised that the amount of time preserved at the fossil site varies and the time represented in the fossil assemblage may differ from that represented by the depositional event (e.g. Behrensmey-er, 1982). This is consistent with the obser-vations made in the Lantian sequence. For example, the fossil assemblage from Local-ity 31 results from a mortality in a crevasse splay event, and therefore, the sedimentary event and fossil assemblage can be consid-ered both instantaneous and contemporane-ous. In contrast, the sediments at Locality 6 represent multiple short-lived depositional events, whereas the bulk of fossils is not directly related to the sedimentary process but considered to result from continuous attritional accumulation on floodplain sur-faces. The fossil assemblage from Locality 42, in turn, combines catastrophic mortality

32

in a mass-flow event with reworked skeletal elements from the land surface and older de-posits (Paper III). It is seems feasible to as-sume that most fossil mammal assemblages from the Lantian sequence are time-aver-aged. For example, Behrensmeyer (1982) suggested that a time period of 102–104 years is required to produce an attritional bone accumulation in an overbank setting.

5.3. Age of fossil vertebrate lo-calitiesMagnetic polarity stratigraphy allows age determination for fossil localities. Mag-netozones have a certain thickness in a lo-cal section, along with a time span that is derived from the correlation of the local magnetostratigraphy to the GPTS. Each magnetozone is bounded by the termination and onset of subsequent and latter magne-tozones of known ages. By convention, the transition between the different polarities is placed at the midpoint between those sites. Consequently, when a fossil-bearing unit occurs within a magnetozone with known upper and lower limits, the age of the fos-sil bed can be interpolated between the time lines (e.g. Locality 12). Similarly, the age of localities that are associated with top-most or bottommost magnetozones can be extrapolated by assuming a constant rate of sedimentation during the deposition of the subsequent or superjacent magnetozone (or magnetozones) (e.g. Locality 6). In total, 30 localities of 52 are given ‘precise’ ages based on the correla-tion of local sections to the GPTS (Table 1). Estimates for the ages of the localities vary in accuracy. The age determinations given for the localities that occur within the sam-pled sections are securely based, as are the age estimations for the localities that can be physically traced to the palaeomagneti-cally dated sections by using stratigraphical

markers like palaeosols or lake marls. How-ever, sites traced laterally over long distanc-es, such as 35, 38 and 43, have some strati-graphical uncertainty. The same applies to stratigraphical levels of fossil localities that are measured vertically over a long distance relative to a marker horizon. For a few local-ities, ‘exact’ ages are not achieved, but their positions have been bracketed by uppermost

Table 1. Ages for selected fossil mammalian lo-calities

Locality Estimated age (Ma)

3 8.775 8.036 8.037 7.9410 8.0311 8.0312 9.9013 9.2814 9.5615 9.5116 9.9718 9.9019 9.9520 9.9021 9.8927 > 6.828 < 6.830 8.1831 8.8832 8.1833 8.0334 8.1135 9.1638 8.7142 6.8 – 6.043 8.7844 6.8 – 6.046 9.9147 < 6.848 9.8150 10.2151 8.2452 7.6753 < 8.0

33

or lowermost age boundaries (e.g. localities 42 and 44). It is not possible to provide ages for all fossil sites in this way: Some locali-ties are surface finds (e.g. Locality 4), the stratigraphical position of a locality in a local section may be unclear (e.g. Locality 45), or, the correlation of a local section to the section with a magnetostratigraphic con-trol can be ambiguous (e.g. Locality 49). The oldest known find site is Local-ity 50, at the base of the Main Section, that falls within the lower portion of subchron C5n.2n and has an absolute age estimate of about 10.21 Ma. This site has yielded small mammalian material from a red-brown sandy siltstone that have also yielded Hip-parion material. The youngest fossil occur-rence in the Bahe Formation is at Locality 52, recorded in the Locality 42 Section at the ca. 68 m level within the normal interval of C4n.2n, its numeral age estimate is 7.67 Ma. In general, the Bahe Formation fauna derives from deposits that span ca. 10.2 to 7.7 Ma (Table 1). Fossil occurrences

are not uniformly distributed throughout the sequence, and the fossiliferous levels are separated by nonfossiliferous strata of vary-ing thickness of sediments (cf. Paper I). Fig-ures 7 and 8 illustrate the unequal distribu-tion of fossil localities and number of speci-mens per unit time. According to Figure 7, the majority of the fossil-bearing units in the Bahe Formation occur between 10.2 and 8.7 Ma and between 8.2 and 7.7 Ma. From Fig-ure 8 it is evident that the bulk of the Bahe Formation fauna is essentially clustered to age intervals around 9.9 Ma, 8.8 – 8.9 Ma and 8 – 8.2 Ma. The uneven distribution of fossil localities is a typical feature in terres-trial sequences and is also found, for exam-ple, in the Siwaliks (e.g. Flynn et al., 1990; Barry et al., 2002). The variation may de-rive from the episodic nature of sedimenta-tion, hiatuses and erosion in fluvial systems. However, it may also reflect real differences in environmental preferences of the original vertebrate communities, or changes in con-ditions of sediment transport and fossil pres-ervation (e.g. Cassiliano, 1997). Consider-

Figure 7. Histogram showing the distribution of fossil localities in the Bahe Formation per unit time.

0

1

2

3

4

5

6

7

6,577,588,599,51010,511Age (Ma)

No

offo

ssil

loca

litie

s

34

ing the observation that there is little change in the fluvial style or accumulation rates within the Bahe Formation, it is assumed that the difference in fossil-site abundance does not reflect real environmental changes to conditions more or less favourable for terrestrial mammalian communities, or the preservational potential of individual litho-facies. Some bias in the fossil recovery in Lantian sequence may result from irregular outcrop availability. The upper part of the Bahe Formation (i.e. ca. 8–7 Ma) is poorly exposed, whereas the sediments deposited between ca. 9.7–8 Ma are extensively avail-able along the wide outcrop belt in the study area. From this it follows that the effort used in prospecting the middle part of the Bahe Formation was much greater. The lower part of the Bahe Formation (i.e. ca. 11–10 Ma) was recovered only from the Main Section. Here, the rarity of fossils is probably not a consequence solely of the collecting bias but may also result from greater abundance of coarse-grained lithologies unfavourable for fossil preservation.

In all, 8 localities were discovered in the Lantian Formation. Of these, four sites, including the productive sites 42 and 44, and nearly all specimens were unearthed from the fluvial deposits of the lower-most Lantian Formation. From a lithologi-cal comparison with the Liujiapo section, which has magnetostratographical control, it is evident that the fossil-sites should be placed between ca. 6.8 and 6.0 Ma.

5.4. Synthesis on palaeoenvi-ronments and palaeoclimatesThis section endeavours to combine differ-ent lines of evidence and using the informa-tion obtained during this study, to answer to questions such as: what was the contem-poraneous climate during deposition of the Bahe Formation? What was the vegetation like? Did the environments change at the Bahe – Lantian formation boundary? The sedimentological evidence from the Bahe Formation portion of the Lan-tian sequence strongly suggests a climate that was relatively dry. The evidence for

Figure 8. Histogram showing the distribution of fossil specimens collected in the Bahe Formation per unit time.

0

100

200

300

400

500

600

6,577,588,599,51010,511Age (Ma)

No

offo

ssil

spec

imen

sco

llect

ed

35

‘semi-aridity’ includes nodular calcite (Pa-per II), commonly attributed to semi-arid to arid climates where evapotranspiration ex-ceeds precipitation (Goudie, 1983; Wright and Tucker, 1991; Retallack, 2001). The common occurrence of sheet floods inter-preted from tabular sandstone deposits (Pa-per II and Paper III), suggests rainfall was probably rather low and intermittent caus-ing periodic flood events in an otherwise low discharge setting. A constantly humid environment would favour the development of peat and other highly carbonaceous sedi-ments that are absent from the Bahe Forma-tion. On the other hand, the absence of evap-orates and desiccation cracks from the Bahe Formation sediments argues against fully arid conditions. Indications of relatively dry climates in the Bahe Formation also come from carbon isotope compositions that in-dicate plants seem to have been adapted to cope with a shortage of water (Paper V). Furthermore, the studies of molar hypsod-onty (tooth height) (Fortelius et al., 2002; Mikael Fortelius, personal communication) and the associated mammalian fauna (Paper I) support the palaeoclimatic interpretation made from both the isotope compositions and sedimentological observations. Coupled with the recognition that the Bahe Formation represents a relatively dry climate, the fossil assemblages also pro-vide a picture of ‘semi-open habitats (Paper I). Since there are no indications of C4 grass-lands in the Bahe Formation (Paper V), it can be argued that climates predominantly favoured herbaceous and woody C3 plant assemblages, probably with locally higher concentrations of trees. Attempts were made to undertake palynological investigations from these sequences, but no fossil pollens and spores were obtained (Wei-Ming Wang, written communication). Some, although rare, pollen and spore assemblages were

recovered from the Bahe Formation during earlier investigations (Anonymous, 1966; Zhang et al., 1978). Their pollen assemblage from Shuijiazui (i.e. the ‘Main Section’) was characterised by a dominance of herbaceous over arboreal taxa, an interpretation not in-consistent with the deductions made here. The dental mesowear pattern analysis from fossil herbivores may provide some insight into the feeding preferences and dominant vegetation of past environments (Fortelius and Solounias, 2000). The mesowear pat-tern from rhinoceroses from the Lantian se-quence reveal that all species appear to have been browsers or mixed feeders (Paper I). The preliminary results from other herbiv-ores indicate that none of the fossil species show the mesowear pattern of true grazers (Mikael Fortelius, personal communica-tion). This therefore excludes the dominance of grasslands, but closely corresponds to the interpretation based on isotopic evidence. Combining the evidence from sedimentology, isotope compositions and vertebrate fossil assemblages leads to the in-terpretation that the environment remained stable throughout deposition of the forma-tion (Papers I, II, IV, V). Limited informa-tion is currently available for the seasonality of climate. However, there are several lines of evidence indicating that the Bahe Forma-tion palaeoenvironments were subject to alternating wet and dry periods: 1) the oc-currence of flash-flood deposits indicates a peak rainfall pattern. 2) Results from X-ray diffraction analysis (n=11) from levels with-in the Bahe Formation reveal that smectite is present in the palaeosols (Christopher Jeans, written communication); pedogenic smec-tite principally develops in climates subject to seasonal precipitation (Chamley, 1989; Gibson et al., 2000). 3) In addition, the com-mon occurrence of blocky ped fabrics and slickensides in the floodplain deposits can

36

be attributed to periodic wetting and drying cycles (cf. Retallack, 2001). Pedogenic fea-tures, when considered by themselves, do not indicate the length of periodicity of the wetting and drying. However, considering that the winter monsoon pattern was already established by the early Miocene (Guo et al., 2002), the alterations can be interpreted to have occurred seasonally. The lithology of the Lantian se-quence points to a gradual but dramatic change of the sedimentological features from the Bahe to the Lantian formations. Field observations reveal that the sedimen-tological pattern changes gradually from mainly fluvial to predominantly aeolian ‘Red Clay’ deposits with continuously re-peated nodule-rich horizons (Paper I; Paper IV). The Bahe – Lantian formation bound-ary culminates in the deposition of bodies of conglomerates in the latest Miocene, suggesting that this represents response to a significant penecontemporaneous uplift of the Qinling Mountains (Zhang et al., 1978). This resulted in increasing accumulation and median grain sizes especially adjacent to the mountain front. It seems likely that this event was also concomitant with the change in the direction of extensional tectonism in the Weihe Graben that has been dated to 6 ±3 Ma (Bellier et al., 1988, 1991). These geological arguments thus suggest that the apparent lithological difference between the formations could merely mirror a tectonic-controlled lateral change of river course. As a result, the aeolian component that might have earlier been disguised by the fluvial ac-tivity would have become visible as fluvial action diminished (cf. Paper IV). However, there are several reasons for proposing that the differences between the formations are also a result of climate change. The strong-est evidence for a climate change at the for-mation boundary comes from that recorded

in the isotope geochemistry and vertebrate assemblages; a change in fossil mammal hypsodonty is also significant. By virtue of more depleted δ13C and enriched δ18O val-ues in the Lantian Formation, the climatic conditions are interpreted to have changed towards wetter with respect to the Bahe Formation (Paper V). Both fossil assem-blages and hypsodonty provide compara-ble evidence for a change in environmental pattern from drier and open to more humid and forested situations (Paper I; Fortelius et al., 2002). These conditions could have enhanced ‘wet deposition’ of dust and trap-ping by dense vegetation (cf. Pye, 1987). An increase in humidity was most probably triggered by the onset and/or intensification of the Asian monsoon circulation through its effect on summer precipitation (cf. An et al., 2001; Paper I; Paper V). Since the Bahe Formation contains some elements indica-tive of seasonality, the changes at the forma-tion boundary may rather reflect a change in the seasonality pattern. It is also worth noting that the evidence for forested and humid habitats in the latest late Miocene in Lantian contrasts to the climate models sug-gesting progressively increasing aridity and opposite to the trend observed in western Eurasia and North America (cf. Janis, 1993; Fortelius et al., 2002; Paper I). The proxy data presented here is scanty for the uppermost Bahe Formation and hence, the detailed timing of the envi-ronmental change in the Lantian sequence remains to be documented. While lithologi-cal change across the formation boundary appears to have been gradational, the ac-cumulation rate curve points to an abrupt change. Although there is a lack of isotope data and vertebrate faunal remains from the topmost Bahe Formation, a distinct change of isotope values is found in the lowermost Lantian Formation (Paper V), associated

37

with a ‘post-turnover fauna’ (Paper I). All these observations lead to the interpretation that the environmental change preceded the formation boundary by an unknown but probably not very long time interval. From the above a series of conclu-sions follow. Sediment variation across the Bahe – Lantian Formation boundary reflects tectonically controlled processes that were caused by the uplift of the Qinling Moun-tains and associated relative basin subsid-ence. These basin-scale events ultimately controlled the position and hydrology of the palaeo-Bahe River within the basin and, consequently, the nature of its deposits. Su-perimposed on the tectonic event is the cli-matic change that stems from the strength-ening of the Asian monsoon circulation. The regional tectonic events and climatic chang-es represented in the Lantian sequence can be considered as ultimately related. This can be judged by the fact that there is a causal link between the uplift of the Himalayas – Tibetan Plateau, intensification of the Asian monsoon system (Kutzbach et al., 1993) and dust accumulation in central China (An et al., 2001), and on the other hand, between the growth of the Tibetan Plateau and the extensional tectonics in North China (cf. Peltzer et al., 1985; Bellier et al., 1988). The ‘Red Clay’ deposits are litho-logically similar over broad geographical areas in the Chinese Loess Plateau. This lithological homogeneity implies wide-spread similarity in the climatic processes controlling their deposition. The same holds for the palaeontological record; the fossil vertebrate assemblages derived from the Lantian Formation are similar to those from the ‘Red Clay’ sites in Baode (north-ern Shanxi Province) and Fugu (northern Shaanxi Province) (Paper I). In addition, there are evident parallels, for instance, be-tween grain-size patterns of the ‘Red Clay’

over a large area (Ding et al., 1998). How-ever, there is a marked difference in the changing vegetational pattern. The carbon-isotope studies indicate practically a pure C3 biomass at Lantian (Paper V) and in the Linxian basin (north-eastern edge of Tibetan Plateau), prior to ca. 2.5 Ma (Wang, 2004). Whereas the Lingtai sequence indicates a C4 grassland expansion at ca. 4 Ma (Ding and Yang, 2000). At Baode, the carbon isotope record from fossil tooth enamel from several ‘Red Clay’ localities indicates C4 vegetation in palaeodiets, interpreted as representing a period of transition between a C3 and C4 world (Benjamin H. Passey, personal com-munication). Together this all suggests sig-nificant local differences in environmental evolution within the Chinese Loess Plateau region. The Miocene–Pliocene transition at Lantian lies within the ‘Red Clay’ (Lan-tian Formation). The available proxy data does not show any distinguishable differ-ences at the boundary zone. Although the palaeontological evidence from Lantian does not extend younger than 6 Ma, the other fossil sites demonstrate that there is no evident faunal turnover in the northern Chi-nese mammalian fauna recorded at this time (Zhang and Liu, in press). However, through their analysis of grain-size distribution from the Xifeng ‘Red Clay’, Qiang et al. (2001) and Guo et al. (2004) have demonstrated that aeolian dust accumulation decreased from the preceding values at ca. 5 Ma. A similar pattern has been observed in aeolian mass accumulation rates in North Pacific Ocean sediments (Rea et al., 1998). Since aeolian dust accumulation primarily reflects aridity in the source areas (Rea et al., 1998), it is concluded that the contemporaneous decrease in deposition rates in both records is caused by a shift towards less dry con-ditions of the Asian interior at the Miocene

38

–Pliocene transition (Guo et al., 2004). Taken together, the Bahe and Lan-tian formations exposed at the marging of the Bailuyuan Plateau offer a view of the late Neogene palaeoenvironmental evolu-tion in the Weihe Graben basin. Although this picture is spatially limited, it provides

new insights into the late Neogene palaeoen-vironmental evolution of the Loess Plateau and contributes information that adds to a more complete understanding of the diver-sity of habitats in North China during the late Neogene.

39

6. Conclusions

• Sedimentological evidence indicates that the Bahe Formation represents a shallow braided to multichannelled river with adjacent laterally extensive floodplain areas. The fluvial deposits are characterised by a predominance of aggradation and lack of lateral accre-tion.

• A detailed chronology, based on mag-netostratigraphy, has been developed. The strata studied record about 8 Ma of sedimentation (ca. 11–3 Ma) in the subsiding Weihe Graben basin. The Bahe – Lantian formation boundary is dated at 6.8 Ma.

• The lithological character and isotope compositions, combined with informa-tion from mammalian fossil assem-blages, suggest that the Bahe Forma-tion environments remained stable.

• The steady sedimentation rates imply a relatively constant tectonic regime dur-ing deposition of the Bahe Formation. There is a prominent change in deposi-tion rates at the Bahe – Lantian forma-tion boundary. Contrary to previous in-terpretations, this formation boundary does not represent a basin-wide hiatus.

• The depositional environment changed from predominantly fluvial in the Bahe Formation to mainly aeolian in the Lantian Formation. This formation boundary apparently coincides with a considerable change in the vegetation-al and climatic pattern from relatively dry climates and open environments to more humid and forested ones. This is opposite to the previously observed global trend.

• Vertebrate fossils are locally abundant in the Bahe Formation and lowermost part of the overlying Lantian Forma-tion. The vertebrate remains occur primarily in floodplain deposits. The process of origin of fossil assemblages is shown to be a complex interaction of hydraulic processes, attritional mortal-ity and biological agencies.

• Most localities can be tied to the Geo-magnetic Polarity Time Scale (GPTS). The fossil-find localities are not uni-formly distributed. The faunal record is best preserved between 9.9–7.7 Ma and between 6.8–6.0 Ma.

• There is virtually no indication of a C4-plant contribution to the soil carbon-ate.

• The Bahe Formation record indicates at least some seasonal precipitation, al-ternating with seasonally dry periods.

40

7. Acknowledgements

This study was carried out at the Depart-ment of Geology, University of Helsinki. The study was part of the Finnish – Chinese Lantian field project and was mainly con-ducted in the Graduate School of Environ-mental Geology. During the process of this work I have received help and support of many people. I wish to express my sincere gratitude to all those who made this study possible. I owe my special thanks to Juha Pekka Lunkka who was elemental to my engagement into the Lantian project dur-ing the first field seasons. His experience in field geology and critical questioning accompanied me over the years in one way or another and also significantly influenced and improved my own scientific work. In addition, the inspiring discussions with him enhanced my understanding of sedimentol-ogy. I greatly appreciate his contribution to my work. I have been fortunate to work with Mikael Fortelius. He has created an inter-national and multidisciplinary environment for doing this study and provided me with opportunities to participate research projects under his guidance. During all the stages of this work both his knowledge and profound scientific expertise have been of invaluable help. I am much indebted to Veli-Pekka Salonen. His faith in the significance of this work was of great support. He also com-mented upon many of the papers. Espe-cially, I want to thank him for the patience and support he showed when the work was prolonged, and for showing a genuine in-terest in the progress of graduate students, regardless of whether they study under his supervision or not.

The official pre-examiners Philip Gibbard and Matti Räsänen are gratefully acknowledged for their constructive criti-cism and valuable comments on the manu-script, and for the flexible and rapid review-ing process. I had the joy to work with a number of people during the several field campaigns in Lantian: Mikael Fortelius, Jussi Kenk-kilä, Kari Lintulaakso, Juha Pekka Lunkka, Lena Selänne (the University of Helsinki); Cui Ning, Liu Liping, Qiu Zhuding, Zhang Zhaoqun, Zheng Shaohua, Zhou Wei, Shou Huaquan, Du Wenhua (Institute of Verte-brate Paleontology and Paleoanthropol-ogy, Beijing); Xavier Filoreau, Vincent Pernegre, Şevket Şen (the Natural History Museum, Paris); Wulf Gose, Mulugeta Fes-eha, Robert Scott (the University of Texas at Austin); Ki Andersson, Maria Björnsdot-ter, Lars Werdelin, (the Swedish Museum of Natural History); Linda Ayliffe (the Univer-sity of Utah); Philip Gibbard (the University of Cambridge) and Alan Gentry (the Natural History Museum, London). I wish to ex-press my warm thanks to all fellow mem-bers in the project for help, discussions and company. I am particularly grateful to Liu Liping and Zhang Zhaoqun for their valu-able contribution and friendship, and Lena Selänne for sharing many field experiences and concerns in scientific and non-scientific issues. All co-authors in papers I–V are warmly thanked for sharing their knowl-edge and skills with me. I very much appre-ciate their input and their high professional expertise. Numerous people, including those mentioned above, have offered their advice, encouragement, and time for discussions on various Lantian-related topics. I would like to extend my gratitude to Hayfaa Abdul Aziz, Steve Boreham, Christopher Jeans,

41

Martin Konert, Lu Huayu, Benjamin H. Passey, Lauri Pesonen, Nicholas Shackle-ton, Jef Vandenberghe, Wang Wei-Ming and Zhou Liping. Anne Forss, Jutta Forssell, Jussi Kenkkilä, Martti Lehtinen and Kristina Löyttyjärvi are acknowledged for their valu-able help in various laboratory analyses. I have greatly enjoyed the free, easy and friendly atmosphere in the Depart-ment of Geology. Friends and colleagues in the department, past and present, have made my working pleasant and provided their help in various aspects. Especially, I wish to thank Seija Kultti who has shared a lot of both pain and pleasure over the years. Many ideas developed and grew in discussions with her. Most of the writing work for this

thesis was done at the Tvärminne Zoologi-cal Station and Lammi Biological Station of the University of Helsinki. Personnel of the stations are thanked for creating inspiring and welcoming atmosphere to work in. I am thankful to all of my friends and relatives outside the university who have helped me to understand the dimen-sions of life. My entire family and family-in-law have supported me in all respects. I could not have gone through this without the support and care of my parents Leena and Antti and sisters Inka and Tuula. Final-ly, I want to thank Otto for making the last few years easier. The financial support from the Graduate School of Environmental Geology and Finnish Cultural Foundation is grate-fully acknowledged.

42

8. ReferencesAbdul Aziz, H., 2001. Astronomical forcing in

continental sediments. An integrated strati-graphic study of Miocene deposits from the Catalayud and Teruel basins, NE Spain. Geologica Ultraiectina 207, 191 p.

An, Z.S., 2000. The history and variability of the East Asian paleomonsoon climate. Quater-nary Science Reviews 19, 171-187.

An, Z., Liu, T., Lu, Y., Porter, S.C., Kukla, G., Wu, X., Hua, Y., 1990. The long-term pale-omonsoon variation recorded by the loess-paleosol sequence in Central China. Quater-nary International 7/8, 91-95.

An, Z., Kukla, G. J., Porter, S. C., Xiao, J., 1991. Magnetic susceptibility evidence of monsoon variation on the Loess Plateau of Central China during the last 130,000 years. Quaternary Research 36, 29-36.

An, Z.S., Kutzbach, J.E., Prell, W.L., Porter, S.C., 2001. Evolution of Asian monsoons and phased uplift of the Himalaya – Tibetan since late Miocene times. Nature 411, 62-66.

Andersson, J.G., 1923. Essays on the Cenozoic of northern China. Memoirs of the Geologi-cal Survey of China, Series A 3, 1-152.

Anonymous, 1966. The Institute of Botany, Chi-nese Academy of Sciences & Pollen and Spore Division of Geological Institute, Min-istry of Geology. The Paleobotanic study of the Lantian area, Shaanxi. In: Monograph of the Lantian Cainozoic. IVPP, Science Press, Beijing, pp. 158-181 (in Chinese).

Badgley, C., 1986. Taphonomy of mammalian fossil remains from Siwalik rocks of Paki-stan. Paleobiology 12, 119-142.

Badgley, C., Behrensmeyer, A.K., 1995. Two long geological records of continental eco-systems. Palaeogeography, Palaeoclimatol-ogy, Palaeoecology 115, 1-11.

Badgley, C., Bartels, W.S., Morgan, M.E., Beh-rensmeyer, A.K., Raza, S.M., 1995. Tapho-nomy of vertebrate assemblages from the Paleogene of northwestern Wyoming and the Neogene of northern Pakistan. Palaeoge-ography, Palaeoclimatology, Palaeoecology 115, 157-180.

Barry, J.C., Morgan, M.E., Flynn, L.J., Pilbeam, D., Behrensmeyer, A.K., Raza, S.M., Khan,

I.A., Badgley, C., Hicks, J., Kelley, J., 2002. Faunal and environmental change in the late Miocene Siwaliks of northern Pakistan. Paleobiology Memoirs. Memoir 3, Supple-ment to 28(3), 1-72.

Beer, J.A., 1990. Steady sedimentation and litho-logic completeness, Bermejo Basin, Argen-tina. Journal of Geology 98, 501-517.

Behrensmeyer, A.K., 1978. Taphonomic and ecologic information from bone weathering. Paleobiology 4, 150-162.

Behrensmeyer, A.K., 1982. Time resolution in fluvial vertebrate assemblages. Paleobiol-ogy 8, 211-227.

Behrensmeyer, A.K., 1991. Terrestrial vertebrate accumulations. In: Allison, P.A., Briggs, D.E.G. (Eds.), Taphonomy – releasing the Data Locked in the Fossil Record. Plenum Press, New York, pp. 291-335.

Behrensmeyer, A.K., Hook, R.W., 1992. Paleoen-vironmental contexts and taphonomic modes. In A.K. Behrensmeyer, J.D. Damuth, W.A. DiMichele, R. Potts, H.-D. Sues, S.L. Winf (Eds.): Terretrial Ecosystems through time. The University of Chicago Press, Chi-cago and London, pp. 15-136.

Bellier, O., Mercier, J.L., Vergély, P., Long, C.X., Ning, C.Z., 1988. Evolution sédimentaire et tectonique du graben cénozoïque de la Wei He (province du Shaanxi, Chine du Nord). Bulletin de la Société Géolologique de Fran-ce IV (6), 979-994.

Bellier, O., Vergély, P., Mercier, J.L., Ning, C.Z., Deng, N.G., Yi, M.C., Long, C.X., 1991. Analyse tectonique et sédimentaire dans les monts Li Shan (Province du Shaanxi – Chi-ne du Nord); datation des regimes tectoni-ques extensifs dans le garben de la Wei He.). Bulletin de la Société Géologique de France 162, 101-112.

Bohlin, B., 1926. Die Familie Giraffidae mit be-sonderer Berucksichtigung der fossilen For-men aus China. Paleontologia Sinica Ser. C 4(1), 1-178.

Bohlin, B., 1935. Cavicornier der Hipparion-Fauna Nord-Chinas. Paleontologia Sinica, Ser. C 9(4), 1-166.

Bown, T.M., Kraus, M.J., 1981. Lower Eocene alluvial paleosols (Willwood Formation, northwest Wyoming, USA) and their signi-ficance for paleoecology, paleoclimatology

43

and basin analysis. Palaeogeography, Pal-aeoclimatology, Palaeoecology 34, 31-56.

Bown, T.M., Kraus, M.J., 1987. Integration of channel and floodplain suites, I: develop-ment of sequence and lateral relations of alluvial paleosols. Journal of Sedimentary Petrology 57, 587-601.

Calvo, J.P., Daams, R., Morales, J., López-Mar-tínez, N., Agustí, J., Anadón, P., Armenter-os, I., Cabrera, L., Civis, J., Corrochano, A., Díaz-Molina, M., Elizaga, M., Hoyos, M., Martín-Suarez, E., Martínez, J., Mois-senet, E., Muñoz, A., Pérez-García, A., Pé-rez-Conzález, A., Portero, J.M., Robles, F., Santisteban, C., Torres, T., van der Meulen, A.J., Vera, J.A., Mein, P., 1993. Up-to-date Spanish continental Neogene synthesis and paleoclimatic interpretation. Revista de la Sociedad Geológica de España 6, 29-40.

Cande, S.C., Kent, D.V., 1995. Revised calibra-tion of the geomagnetic polarity timescale for the Late Cretaceous and Cenozoic: Jour-nal of Geophysical Research 100, B4, 6093-6095.

Cassiliano, M., 1997. Taphonomy of mammalian fossils across the Blancan – Irvingtonian boundary: Palm Spring Formation, Anza – Borrego desert of southern California. Palaeogeography, Palaeoclimatology, Pal-aeoecology 129, 81-108.

Cerling, T.E., 1984. The stable isotopic composi-tion of soil carbonate and its relationship to climate. Earth and Planetary Science Letters 71, 229-240.

Cerling, T.E., 1991. Carbon dioxide in the atmos-phere: evidence from Cenozoic and Meso-zoic paleosols. American Journal of Science 291, 377-400.

Cerling, T.E., Harris, J.M., MacFadden, B.J., Leakey, M.G., Quade, J., Eisenmann, V., Ehleringer, J.R., 1997. Global vegeta-tion change through the Miocene/Pliocene boundary. Nature 389, 153-158.

Chamley, H., 1989. Clay Sedimentology. Spring-er-Verlag, New York, 623 p.

Chen, B.-L., Yin, G.-M., Li, W.-L., Guo, L., Li, L., Lu, Y.-C., Guo, S.-L., 2001. Determina-tion of tectonic uplift rates of Qinling moun-tains in central China by fission tracks. Ra-diation Measurements 34, 405-408.

Chen, G.-F., Zhang, Z.-Q., 2004. Lantiantragus

gen. nov. (Urmiatheriinae, Bovidae, Artio-dactyla) from the Bahe Formation, Lantian, China. Vertabrata PalAsiatica 42(3), 205-215.

Chen, Q., Dickinson, W.R., 1986. Contrasting nature of petroliferous Mesozoic – Ceno-zoic basins in eastern and western China. American Association of Petroleum Geolo-gists Bulletin 70, 263-275.

Chow, M., 1978. Tertiary mammalian fauna of the Lantian district. Professional Papers on Stratigraphy and Paleontology 7, 98-108 (in Chinese).

Daams, R., Alcalá, L., Álvarez Sierra, M.A., Azanza, B., van Dam, J., van der Meulen, A.J., Morales, J., Nieto, M., Pelaéz-Cam-pomanes, P., Soria, D., 1998. A stratigraphi-cal framework for Miocene (MN4-MN13) continental sediments of Central Spain. Comptes Rendus de l’Académie des scien-ces de Paris, Ser II 327, 625-631.

Daams, R., van der Meulen, A.J., Álvarez Sierra, M.A., Pelaéz-Campomanes, P., Calvo, J.P., Alonso Zarza, A.M., Krijgsman, W., 1999a. Stratigraphy and sedimentology of the Aragonian (Early to Middle Miocene) in its type area (North-Central Spain). Newslet-ters on Stratigraphy 37, 103-139.

Daams, R., van der Meulen, A.J., Pelaéz-Cam-pomanes, P., Álvarez-Sierra, M.A. 1999b. Trends in rodent assemblages from the Aragonian (early-middle Miocene) of the Catalayud-Daroca Basin, Aragon, Spain. In: Agusti, J., Rook, L., Andrews, P. (Eds.), Hominoid Evolution and Climatic Change in Europe, vol. 1. The evolution of Neogene Terrestrial Ecosystems in Europe. Cam-bridge University Press, Cambridge, pp. 127-139.

Dam, J.A. van, 1997. The small mammals from the upper Miocene of the Teruel-Alfambra region (Spain): paleobiology and paleocli-matic reconstructions. Geologica Ultraiec-tina 156, 204 p.

Derbyshire, E., Kemp, R., Meng, X., 1995. Vari-ations in loess and palaeosol properties as indicators of palaeoclimatic gradients across the Loess Plateau of north China. Quater-nary Science Reviews 14, 681-697.

Ding, Z.L., Yang, S.L., 2000. C3/C4 vegetation over the last 7.0 Myr in the Chinese Loess

44

Plateau: evidence from pedogenic carbonate δ13C. Palaeogeography, Palaeoclimatology, Palaeoecology 160, 291-299.

Ding, Z., Rutter, N., Han, J., Liu, T., 1992. A coupled environmental system formed at about 2.5 Ma in East Asia. Palaeogeogra-phy, Palaeoclimatology, Palaeoecology 94, 223-242.

Ding, Z.L., Liu, T.S., Rutter, N.W., Yu, Z.W., Guo, Z.T., Zhu, R.X., 1995. Ice-volume forcing of east Asian winter monsoon varia-tions in the past 800,000 years. Quaternary Research 44, 149-159.

Ding, Z.L., Sun, J.M., Liu, T.S., Zhu, R.X., Yang, S.L., Guo, B., 1998. Wind-blown origin of the Pliocene red clay formation in the cen-tral Loess Plateau, China. Earth and Plan-etary Science Letters 161, 135-143.

Ding, Z.L., Xiong, S.F., Sun, J.M., Yang, S.L., Gu, Z.Y., Lu, T.S., 1999. Pedostratigraphy and paleomagnetism of a ~7.0 Ma eolian loess – red clay sequence at Lingtai, Loess Plateau, north-central China and the impli-cations for paleomonsoon evolution. Palae-ogeography, Palaeoclimatology, Palaeoecol-ogy 152, 49-66.

Domrös, M., Peng, G.B., 1988. The Climate of China. Springer-Verlag, Berlin, 360 p.

Duggen, S., Hoernie, K., van den Bogaard, P., Rüpke, L., Morgan, J.P., 2003. Deep roots of the Messinian salinity crisis. Nature 422, 602-606.

Fang, X., Garzione C., Van der Voo, R., Li, J., Fan, M., 2003. Flexural subsidence by 29 Ma on the NE edge of Tibet from the mag-netostratigraphy of Linxia Basin, China. Earth and Planetary Science Letters 210, 545-560.

Flower, B.P. 1999. Warming without CO2? Na-ture 399, 313-314.

Flower, B.P., Kennett, J.P., 1994. The middle Mi-ocene climatic transition: East Antarctic Ice sheet development, deep ocean circulation and global carbon cycling. Palaeogeogra-phy, Palaeoclimatology, Palaeoecology 108, 537-555.

Flower, B.P., Kennett, J.P., 1995. Middle Mi-ocene deepwater paleoceanography in the southwest Pacific: relations with East Ant-arctic Ice Sheet development. Paleoceanog-raphy 10, 1095-1112.

Flynn, L.J., Pilbeam, D., Jacobs, L.L., Barry, J.C., Behrensmeyer, A.K., Kappelman, J.W., 1990. The Siwaliks of Pakistan: time and faunas in a Miocene terrestrial setting. Journal of Geology 98, 589-604.

Flynn, L.J., Qiu, Z.X., Opdyke, N.D., Tedford, R.H., 1995. Ages of key fossil assemblag-es in the late Neogene terrestrial record of northern China. In: Berggren, W.A., Kent, D.V., Aubry, M.-P., Hardenbol, J. (Eds.), Geochronology, Time Scales and Global Stratigraphic Correlation. SEPM Special Publication 54, 365-373.

Flynn, L.J., Wu, W.Y., Dawson, W.R., 1997. Dat-ing vertebrate microfaunas in the late Neo-gene record of Northern China. Palaeogeog-raphy, Palaeoclimatology, Palaeoecology, 133, 227-242.

Fortelius, M., Werdelin, L., Andrews, P., Bernor, R.L., Gentry, A., Humphrey, L., Mittmann, H.-W., Viranta, S., 1996. Provinciality, di-versity, turnover, and paleoecology in land mammal faunas of the later Miocene of Western Eurasia. In: Bernor, R.L., Fahlbus-ch, V., Mittmann, H.-W. (Eds.), The Evo-lution of Western Eurasian Neogene Land Mammal Faunas. Columbia University Press, New York, pp. 414-448.

Fortelius, M., Solounias, N., 2000. Functional characterization of ungulate molars using the abrasion-attrition wear gradient: a new method for reconstructing paleodiets. Amer-ican Museum Novitates 3301, 1-36.

Fortelius, M., Eronen, J., Jernvall., J., Liu, L., Pushkina, D., Rinne, J., Tesakov, A., Vislo-bokova, I., Zhang, Z., Zhou, L., 2002. Fossil mammals resolve regional patterns of Eura-sian climate change over 20 million years. Evolutionary Ecology Research 4, 1005-1016.

Fortelius, M., Kappelman, J., Sen, S., Bernor, R. (Eds.), 2003. Geology and Paleontology of the Miocene Sinap Formation, Turkey. New York, Columbia University Press, 409 p.

Friend, P.F., Johnson, N.M., McRae, L.E., 1989. Time-level plots and accumulation rates of sediement sequences. Geological Magazine 126, 491-498.

Garcés, M., Agustí, J., Cabrera, L., Parés, J.M., 1996. Magnetostratigraphy of the Vallesian (late Miocene) in the Vallès-Penedès Basin

45

(northeast Spain). Earth and Planetary Sci-ence Letters 142, 381-396.

Garcés, M., Cabrera, L., Agustí, J., Parés, J.M., 1997. Old World first appearance datum of “Hipparion” horses: Late Miocene large-mammal dispersal and global events. Geol-ogy 25, 19-22.

Gibson, T.G., Bybell, L.M., Mason, D.B., 2000. Stratigraphic and climatic implications of clay mineral changes around the Paleocene/Eocene boundary of the northeastern US margin. Sedimentary Geology 134, 65-92.

Goudie, A.S., 1983. Calcrete. In: Goudie, A.S., Pye, K. (Eds.), Chemical Sediments and Geomorphology. Academic Press, London, pp. 93-131.

Grimmer, J.C., Jonckheere, R., Enkelmann, E., Ratschbacher, L., Hacker, B.R., Blythe, A.E., Wagner, G.A., Wu, Q., Liu, S., Dong, S., 2002. Cretaceous–Cenozoic history of the southern Tan-Lu fault zone: apatite fis-sion-track and structural constraints from the Dabie Shan (eastern China). Tectono-physics 359, 225-253.

Guo, Z.T., Peng, S.Z., Hao, Q.Z., Biscaye, P.E., Liu, T.S., 2001. Origin of the Miocene – Pliocene Red-Earth Formation at Xifeng in Northern China and implications for pal-aeoenvironments. Palaeogeography, Palaeo-climatology, Palaeoecology 170, 11-26.

Guo, Z.T., Ruddiman, W.F., Hao, Q.Z., Wu, H.B., Qiao, Y.S., Zhu, R.X., Peng, S.Z., Wei, J.J., Yuan, B.Y., Liu, T.S., 2002. Onset of Asian desertification by 22 Myr ago inferred from loess deposits in China. Nature 416, 159-163.

Guo, Z.T., Peng, S.Z., Hao, Q.Z., Biscaye, P.E., An, Z.S., Liu, T.S., 2004. Late Miocene – Pliocene development of Asian aridifica-tion as recorded in the Red-Earth Formation in northern China. Global and Planetary Change 41, 135-145.

Hacker, B.R., Wang, X., Eide, E.A., Ratsch-bacher, L., 1996. The Qinling-Dabie ultra-high-pressure collisional orogen. In: An, Y, Harrison, T.M. (Eds), The Tectonics of Asia. Cambridge University Press, Cambridge, pp. 345-370.

Han, J., Fyfe, W.S, Gu, Z., 2002. Assessment of the palaeoclimate during 3.0-2.6 Ma regis-tered by transition of Red Clay to loess –

palaeosol sequence in central North China. Palaeogeography, Palaeoclimatology, Palae-oecology 185, 355-368.

Harrison, T.M., Copeland, P., Hall, S.A., Quade, J., Burner, S., Ojha, T.P., Kidd, W.S.F., 1993. Isotopic preservation of Himalayan/Tibetan uplift, denudation, and climatic histories of two molasses deposits. Journal of Geology 101, 159-177.

Janis, C.M., 1993. Tertiary mammal evolution in the context of changing climates, vegeta-tion, and tectonic events. Annual Review of Ecology & Systematics 24, 467-500.

Jia, L., Zhang, Y., Huang, W., Tang, Y., Ji, H., You, Y., Ding, S., Huang, X., 1966. Neogene of Lantian, Shaanxi Province. In: Mono-graph of the Lantian Cainozoic. IVPP, Sci-ence Press, Beijing, pp. 1-31 (in Chinese).

Johnson, N.M., Stix, J., Tauxe, L., Cerveny, P.F., Tahirkheli, R.A.K., 1985. Paleomagnetic chronology, fluvial processes and tectonic implications of the Siwalik deposits near Chinji Village, Pakistan. Journal of Geology 93, 27-40.

Kappelman, J., Sen, S., Fortelius, M., Duncan, A., Alpagut, B., Crabaugh, J., Gentry, A., Lunkka, J.P., McDowell, F., Solounias, N., Viranta, S., Werdelin, L., 1996. Chronology and biostratigraphy of the Miocene Sinap Formation of Central Turkey. In: Bernor, R., Fahlbusch, V., Mittman, H.W. (Eds.), The Evolution of Western Eurasian Neo-gene Mammal Faunas. Columbia University Press, New York, pp.78-95.

Khan, I.A., Bridge, J.S., Kappelman, J., Wilson, R., 1997. Evolution of Miocene fluvial envi-ronments, eastern Potwar Plateau, northern Pakistan. Sedimentology 44, 221-251.

Kraus, M.J., Bown, T.M., 1986. Paleosols and time resolution in alluvial stratigraphy. In: Wright, V.P. (Ed.), Paleosols, their recogni-tion and interpretation. Princeton University Press, pp. 180-207.

Krijgsman, W., 1996. Miocene magnetostratigra-phy and cyclostratigraphy in the Mediterra-nean: extension of the astronomical polarity time scale. Geologica Ultraiectina 141, 207 p.

Krijgsman, W., Hilgen, F.J., Raffi, I., Sierron, F.J., Wilson, D.S., 1999. Chronology, caus-es and progression of the Messinian salinity

46

crisis. Nature 400, 652-655.Kukla, G., 1987. Loess stratigraphy in Central

China. Quaternary Science Reviews 6, 191-219.

Kukla, G., An, Z., 1989. Loess stratigraphy in central China. Palaeogeography, Palaeocli-matology, Palaeoecology 72, 203-225.

Kukla, G., Heller, F., Liu, X.M., Xu, T.C., Liu. T.S., An, Z.S., 1988. Pleistocene climates in China dated by magnetic susceptibility. Ge-ology 16, 811-814.

Kurten, B., 1952. The Chinese Hipparion fauna. Societas Scientarum Fennica. Commenta-tiones Biologicae XIII 4, 82 p.

Kutzbach, J.E., Prell, W.L., Ruddiman, W.F., 1993. Sensitivity of Eurasian climate to sur-face uplift of the Tibetan Plateau. Journal of Geology 101, 177-190.

Li, C.K., Wu, W.Y., Qiu, Z.D., 1984. Chinese Neogene: subdivision and correlation. Ver-tebrata PalAsiatica 22(3), 163-178 (in Chi-nese with English summary).

Liu, T.S., 1985. Loess and the environment. Chi-na Ocean Press, Beijing, 251 p.

Liu, T.S., Ding, M.L., Gao, F.Q., 1960. Cenozoic stratigraphic sections between Xi’an and Lantian. Scientia Geologica Sinica 4, 199-208 (in Chinese).

Liu, T.S., Li, C.K., Zhai, R.J., 1978. Pliocene vertebrates of Lantian, Shensi. Professional Papers on Stratigraphy and Paleontology 7, 149-200 (in Chinese).

Lu, H., Vandenberghe, J., An, Z., 2001. Aeolian origin and palaeoclimatic implications of ‘Red Clay’ (north China) as evidenced by grain-size distribution. Journal of Quater-nary Science 16, 89-97.

Maslin, M.A., Li, X.S., Loutre, M.F., Berger, A., 1998. The contribution of orbital forcing to the progressive intensification of Northern Hemisphere Glaciation. Quaternary Science Reviews 17, 411-426.

Mattauer, M., Matte, Ph, Malavieille, J., Tappon-nier, P., Maluski, H., Qin, X.Z., Lun, L.Y., Qin, T.Y., 1985. Tectonics of the Qinling Belt: bild-up and evolution of eastern Asia. Nature 317, 496-500.

McRae, L.E., 1990a. Paleomagnetic isochrones, unsteadiness and non-uniformity of sedi-mentation in Miocene fluvial strata of the Siwalik Group, northern Pakistan. Journal

of Geology 98, 433-456.McRae, L.E., 1990b. Paleomagnetic isochrones,

unsteadiness, and uniformity of sedimenta-tion in Miocene intermontane basin sedi-ments at Salla, eastern Andean cordillera, Bolivia. Journal of Geology 98, 479-500.

Meng, Q.-R., Zhang, G.-W., 1999. Timing of col-lision of the North and South China blocks: Controversy andreconciliation. Geology 27, 123-126.

Meng, Q.-R., Zhang, G.-W., 2000. Geologic framework and tectonic evolution of the Qinling orogen, central China. Tectono-physics 323, 183-196.

Miall, A.D. 1996. The Geology of Fluvial De-posits. Springer-Verlag, Berlin, 582 p.

Opdyke, N., Mein, P., Lindsay, E., Pérez-González, A., Moissenet, E., Norton, V.L., 1997. Continental deposits, magnetostratig-raphy, and vertebrate paleontology, late Neogene of Eastern Spain. Palaeogeogra-phy, Palaeoclimatology, Palaeoecology 133, 129-148.

Peltzer, G., Tapponnier, P., Zhang, Z., Xu, Z.Q., 1985. Neogene and Quaternary faulting in and along the Qinling Shan. Nature 317, 500-505.

Potts, R., Behrensmeyer, A.K., 1992. Late Ceno-zoic terrestrial ecosystems. In: Behrensmey-er, A.K., Damuth, J.D., DiMichele, W.A., Potts, R., Sues, H.-D., Wing, S.L. (Eds.), Terrestrial ecosystems through time: evo-lutionary paleoecology of terrestrial plants and animals. University of Chicago Press, Chicago, pp. 419-519

Prell, W.L., Kutzbach, J.E., 1992. Sensitivity of the Indian monsoon to forcing parameters and implications for its evolution. Nature 360, 647-652.

Pye, K., 1987. Aeolian Dust and Dust Deposits. Academic Press, London, 334 p.

Qiang, X.K., Li, Z.X., Powell, C.McA., Zheng, H.B., 2001. Magnetostratigraphic record of the Late Miocene onset of the East Asian monsoon, and Pliocene uplift of northern Tibet. Earth and Planetary Science Letters 187, 83-93.

Qiu, Z.X., 1990. The Chinese Neogene mamma-lian biochronology – its correlation with the European Neogene mammalian zonation. In: Lindsay, E.H. (Ed.), European Mammal

47

Chronology. Plenum Press, New York, pp. 527-556.

Qiu, Z.X., Qiu, Z.D., 1995. Chronological se-quence and subdivision of Chinese Neogene mammalian faunas. Palaeogeography, Pal-aeoclimatology, Palaeoecology 116, 41-70.

Qiu, Z.D., Zheng, S.H., Sen S., Zhang, Z.Q., 2003. Late Miocene micromammals from the Bahe Formation, Lantian, China. In: Re-umer, J.W.F., Wessels, W. (Eds.), Distribu-tion and Migration of Tertiary Mammals in Eurasia. Deinsea 10, 443-453

Qiu, Z.D., Zheng, S.H., Zhang, Z.Q., 2004a. Mu-rids from the late Miocene Bahe formation, Lantian, Shaanxi. Vertebrata PalAsiatica 42(1), 67-76.

Qiu, Z.D., Zheng, S.H., Zhang, Z.Q., 2004b. Gerbillids from the late Miocene Bahe for-mation, Lantian, Shaanxi. Vertebrata PalA-siatica 42(2), 193-204.

Quade, J., Cerling, T.E., Bowman, J.R., 1989. Systematic variations in the carbon and oxygen isotopic composition of pedogenic carbonate along elevation transects in the southern Great Basin, United States. The Geological Society of America Bulletin 101, 464-475.

Quade, J., Cater, J.M.L., Ojha, T.P., Adam, J., Harrison, T.M., 1995. Late Miocene envi-ronmental change in Nepal and the northern Indian subcontinent: stable isotopic evi-dence from paleosols. The Geological Soci-ety of America Bulletin 107, 1381-1397.

Ratschbacher, L., Hacker, B.R., Calvert, A., Webb, L.E., Grimmer, J.C., McWilliams, M.O., Ireland, T., Dong, S., Hu, J., 2003. Tectonic of the Qinling (Central China): tec-tonostratigraphy, geochronology, and defor-mation history. Tectonophysics 366, 1-53.

Rea, D.K., 1992. Delivery of Himalayan sedi-ment to the northern Indian Ocean and its relation to global climate, sea level, uplift, and seawater strontium. American Geo-physical Union Geophysical Monograph 70, 387-402.

Rea, D.K., Snoeckx, H., Joseph, L.H., 1998. Late Cenozoic eolian deposition in the North Pa-cific: Asian drying, Tibetan uplift, and cool-ing of the northern hemisphere. Paleocea-nography 13, 215-224.

Retallack, G., 1984. Completeness of the rock

and fossil record: some estimates using fos-sil soils. Paleobiology 10, 59-78.

Retallack, G., 1986. Fossil soils as grounds for interpreting long-term controls on ancient rivers. Journal of Sedimentary Petrology 56, 1-18.

Retallack, G.J., 2001. Soils of the Past: An In-troduction to Paleopedology. Blackwell Sci-ence, Oxford, 404 p.

Rutter, N.W., Ding, Z.L., Evans, M.E., Liu, T.S., 1991. Baoji-type pedostratigraphic section, Loess Plateau, north-central China. Quater-nary Science Reviews 10, 1-22.

Sadler, P.M., 1981. Sediment accumulation rates and the completeness of stratigraphic sec-tions. Journal of Geology 89, 569-584.

Schlosser, M., 1903. Die fossilen Saugethiere Chinas nebst einer Odontographie der re-centen Antilopen. Abh. Bayer Akad. Wiss. 22, 1-221.

Solounias, N., 1981. Mammalian fossils of Sa-mos and Pikermi. Part 2. Resurrection of a classic Turolian fauna. Annals of Carnegie Museum 50, 231-270.

Sun, D., Liu, D., Chen, M., An, Z., Shaw, J., 1997. Magnetostratigraphy and palaeocli-mate of Red Clay sequences from Chinese Loess Plateau. Science in China (Series D) 40(4), 337-343.

Sun, D.H., Shaw, J., An, Z.S., Chen, M.Y., Yue, L.P., 1998. Magnetostratigraphy and paleo-climatic interpretation of a continuous 7.2 Ma late Cenozoic eolian sediments from Chinese Loess Plateau. Geophysical Re-search Letters 25, 85-88.

Tapponnier, P., Molnar, P., 1977. Active faulting and tectonics in China. Journal of Geophysi-cal Research 82, 2905-2930.

Tedford, R.H., Flynn, L.J., Qiu Z., Opdyke, N.D., Downs, W.R., 1991. Yushe Basin, China: Paleomagnetically calibrated mammalian biostratigraphic standard for the late Neo-gene of eastern Asia. Journal of Vertebrate Paleontology 11, 519-526.

Teilhard de Chardin, P., Young, C.C., 1931. Fos-sils from the Late Cenozoic of Northern China. Paleontologia Sinica, Ser. C 9(1), 1-67.

Tucker, M.E., 2003. Sedimentary Rocks in the Field. Wiley, 234 p.

Wang, H., Follmer, L.R., 1998. Proxy of mon-

48

soon seasonality in carbon isotopes from paleosols of the southern Chinese Loess Plateau. Geology 26, 987-990.

Wang, H., Ambrose, S.H., Liu, C.-L.J., Follmer, L.R., 1997. Paleosol stable isotope evidence for early hominid occupation of east Asian temperate environments. Quaternary Re-search 48, 228-238.

Wang, J.-M., 1987. The Fenwei rift and its re-cent periodic activity. Tectonophysics 133, 257-275.

Wang, N., Chao, J., Qian, Z., 1966. Topographic structures and history of the Bahe River. In: Monograph of the Lantian Cainozoic. IVPP, Science Press, Beijing, pp. 75-88 (in Chi-nese).

Wang, Y., 2004. A 25-Ma record of paleodiet and environmental change from carbon and ox-ygen isotopes in mammalian tooth enamel and paleosols from NE margin of the Tibet-an Plateau. Geological Society of America, Abstracts with Programs 36(5), p. 38.

Weidmann, M., Solounias, N., Drake, R.E., Cur-tis, G.H., 1984. Neogene stratigraphy of the Eastern Basin, Samos Island, Greece. Geo-bios 17, 477-490.

Willis, B., 1993a. Ancient river systems in the Himalayan foredeep, Chinji Village area, northern Pakistan. Sedimentary Geology 88, 1-76.

Willis, B., 1993b. Evolution of Miocene fluvial systems in the Himalayan foredeep through a two kilometre-thick succession in northern Pakistan. Sedimentary Geology 88, 77-121.

Willis, B., Behrensmeyer, A.K., 1994. Archi-tecture of Miocene overbank deposits in northern Pakistan. Journal of Sedimentary Research B64, 60-67.

World Climatological Normals, 1996. World Meteorological Organization, Climatologi-cal Normals (CLINO) for the period 1961-1990. WMO Region II: Asia. Secretariat of the World Meteorological Organization, Geneva.

Wright, V.P., Tucker, M.E., 1991. Calcretes: an introduction. In: Wright, V.P., Tucker, M.E. (Eds.), Calcretes. International Association of Sedimentologists, Reprint Series 2, pp.1-22.

Xiao, J., Porter, S.C., An, Z.S., Kumai, H., Yoshikawa, S., 1995. Grain size of quartz

as an indicator of winter monsoon strength on the loess plateau of central China during the last 130,000 yr. Quaternary Research 43, 22-29.

Xie, Y., Ding, H., Li, C., Jin, Z., 1966. Sedi-mentary environments reconstruction of the Lantian area. In: Monograph of the Lantian Cainozoic, IVPP, Science Press, Beijing, pp. 89-123.

Yang, D., Yu, G., Xie, Y., Zhan, D., Li, Z., 2000. Sedimentary records of large floods from the middle reaches of the Yellow River, China. Geomorphology 33, 73-88.

Yin, A., Nie, S., 1993. An indentation model for the North and South China collision and the development of the Tan-Lu and Honam fault systems, eastern China. Tectonics 12, 801-813.

Yue, L.P., 1995. Paleomagnetic polarity bound-ary were recorded in Chinese loess and Red Clay, and geological significance. Acta Geo-physica Sinica 38(3), 3120-327 (in Chinese with English abstract).

Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, and ab-errations in global climate 65 Ma to present. Science 292, 686-693.

Zaleha, M.J., 1997a. Fluvial and lacustrine pal-aeoenvironments of the Miocene Siwalik Group, Khaur area, northern Pakistan. Sedi-mentology 44, 349-368.

Zaleha, M.J., 1997b. Intra- and extrabasinal con-trols on fluvial deposition in the Miocene Indo-Gangetic foreland basin, northern Pa-kistan. Sedimentology 44, 369-390.

Zdansky, O., 1923. Fundorte der Hipparion-Fauna um Pao-Te-Hsien in N.W. Shansi. Bulletin of the Geological Survey, China 5, 69-82.

Zhai, X., Day, H.W., Hacker B.R., You, Z., 1998. Paleozoic metamorphism in the Qinling orogen, Tongbai Mountains, central China. Geology 26, 371-374.

Zhang, Y., Huang, W., Tang, Y., Ji, H.X., Long, Y., Tong, Y., Ding, S., Huang, X., Zheng, J., 1978. Cainozoic of the Lantian area, Shaanxi. Memoirs of Institute of Verte-brate Paleontology and Paleoanthropology, Academia Sinica, Special Publication A, 14. Science Press, 1-64 p. (in Chinese).

Zhang, Y.Q., Vergély, P., Mercier, J., 1995. Ac-

49

tive faulting in and along the Qinling Range (China) inferred from SPOT imagery analy-sis and extrusion tectonics of south China. Tectonophysics 243, 69-95.

Zhang, Y.Q., Mercier, J.L., Vergély, P., 1998. Extension in the graben systems around the Ordos (China), and its contribution to the extrusion tectonics of south China with respect to Gobi–Mongolia. Tectonophysics 285, 41-75.

Zhang, Y., Vergély, P., Mercier, J., Wang, Y., Zhang, Y., Huang, D., 1999. Kinematic his-tory and changes in the tectonic stress re-gime during the Cenozoic along the Qinling and southern Tanlu fault zone. Acta Geol. Sin. 73 (3), 264-274.

Zhang, Z.-Q., 2003. A new species of Shaanx-ispira (Bovidae, Artiodactyla, Mammalia) from the Bahe Formation, Lantian, China. Vertebrata PalAsiatica 41(3), 230-239.

Zhang, Z.Q., Zheng, S.H., Liu, L.P., Fortelius, M., Lunkka, J.P., Kaakinen, A., Selanne, L., 1999. Biostratigraphic study on the Bahe Formation. In: Proceedings VII Annual Meeting Chinese Society of Vertebrate Pale-ontology, China Ocean Press, pp. 55-60.

Zhang, Z.Q., Liu, L.P. (in press). The Late Neo-gene mammal biochronology in the Loess Plateau, China. Annales de Paleontologie.

Zhao, J., Zheng, G.M., Wang, H.D., Xu, J.L., 1990. The Natural History of China. Collins, London, 224 p.

Zhao, S., 1994. Geography of China. Environ-ments Resources Population and Develop-ment. Wiley, New York, 332 p.

Zheng, H., An, Z., Shaw, J., 1992. New contribu-tions to Chinese Plio-Pleistocene magneto-stratigraphy. Physics of the Earth and Plan-etary Interiors 70, 146-153.