Transcript
Journal of the Geological Society , London , Vol. 162 , 2005, pp. 653–659. Printed in Great Britain.653 Abnormal pollen grains: an outcome of deteriorating atmospheric conditionsaround the Permian–Triassic boundary C. B. FOSTER 1 & S. A. AFONIN 21 Geoscience Australia, P.O. Box 378, Canberra, A.C.T. 2601, Australia (e-mail:
[email protected]) 2 Palaeontological Institute, 117647, Profsoyuznaya 123, Moscow, Russia Abstract: Records of abnormal gymnosperm pollen morphotypes from coeval localities in Russia and NWChina provide independent and unique evidence of deteriorating atmospheric conditions at the close of thePermian. Although at similar palaeolatitude, the Vologda region of European Russia and the Junggar Basin of Xinjiang Province, China, were thousands of kilometres apart and on different tectonic plates. The co-occurrence of the megaspore Otynisporites eotriassicus Fuglewicz 1977, and its restricted distributionelsewhere in marine sections with either Late Permian ammonoids or the basal Triassic conodont marker Hindeodus parvus , confirms the age range, and correlation of the Russian and NW Chinese sections. In threeChinese sections, rapid changes in carbon isotopic ( 13 C org ) composition of between 4 and 10‰ occur over the intervals showing abnormal pollen. These patterns reflect bioclastic facies contributions responding tolocal environmental changes. They are nevertheless coincident with global perturbations in the atmosphere,caused by extensive volcanic activity around the Permian–Triassic boundary. Keywords: Permian, Triassic, pollen, carbon isotopes, palaeoclimate. The boundary between the Permian and Triassic Systems ismarked by a world-wide extinction of marine fauna, with lessthan 10% of the fauna surviving into the Mesozoic (Erwin1993). It has been long debated if large-scale changes in theterrestrial realm were contemporaneous with those in the sea;claims to the contrary emphasized different evolution rates. Eventhe turnovers of palaeophytic flora have been considered asdiachronous across the globe (Meyen 1973; Traverse 1988).The problem is exacerbated by severe difficulties in recogniz-ing the Permian–Triassic (P–T) boundary in non-marine sec-tions, away from the Global Boundary Stratotype Section and Point (GSSP) for the base of the Triassic, at Bed 27c in theMeishan Section, Changxing County, Zhejiang Province, China(Jin et al . 1997; see Fig. 2).The GSSP is defined by the first occurrence of the conodontspecies Hindeodus parvus in the evolutionary lineage Hindeoduslatidentatus – Hindeodus parvus – Isarcicella isarcica . For conti-nental deposits, such as those studied here, a number of proxiesfor the P–T boundary have been used: shifts in carbon isotopiccomposition (Sephton et al . 2002); the first appearance datum(FAD) of both land animals and plants, including the amphibian Lystrosaurus , and lycopod spores assigned to Aratrisporites .Erwin et al . (2002) have summarized possible causal mechan-isms.The hypothesis of synchronous extinction within the plantworld, in keeping with a destruction, death and decay scenario,was supported by the putative ‘spike’ of saprophytic fungalremains at many P–T boundary sections (Visscher et al . 1996).This theory has been refuted, as the ‘fungal’ remains are of algalorigin, and their stratigraphic occurrence is not limited to thelatest Permian–Early Triassic (Foster et al . 1997; Afonin et al .2001; Foster et al . 2002). Other criteria are now needed to effectcorrelation using plant microfossils, and our research is directed to establishing correlation with the GSSP in southern China.We report findings from our continuing study of plant micro-fossil assemblages from Nedubrovo, Vologda region, Russia(Krassilov et al . 1999 a ; Afonin 2000) and the DalongkouAnticline and Lucaogou sections, Junggar Basin, Xinjiang Pro-vince, NW China (Metcalfe et al . 2001; Figs 1 and 3). TheChinese sample set comprised 27 samples from a 198 m meas-ured section from the north limb of Dalongkou Anticline, 12samples from a 210 m section of the south limb, and 21 samplesfrom a 120 m section at Lucaogou. Samples span the Lower Triassic Jiucaiyuan Formation and underlying Guodikeng and Wutonggou formations. Sections from all areas are of non-marine srcin. Of critical importance to our argument is that theassemblages from the Chinese and Russian localities are coeval.Plant microfossils provide that evidence. Correlation: Dalongkou and Nedubrovo In the Dalongkou section, spore-pollen floras are recovered fromthe siltstone- and claystone-dominated Guodikeng Formation and from the coarse sandstones of the overlying lower TriassicJiucaiyuan Formation (Figs 2 and 3). The boundary betweenthese formations is marked by a discontinuously developed, thinconglomerate that immediately overlies the Guodikeng clays-tones. Although mappable in the field, the lithological boundarywas not considered by previous workers to mark the P–T boundary (Yang et al . 1986; Ouyang & Norris 1999). Thoseworkers placed the P–T boundary between 30 and 50 m belowthe basal sandstones of the Jiucaiyuan Formation, in the upper Guodikeng Formation. The criteria to define the P–T boundaryare biological, and relate to changes in the palynoflora; theyinclude the appearance of lycopod spores, including Aratrispor-ites spp., a widely used marker for the Early Triassic (see Foster & Archbold 2001).Whereas the previous studies were based on relatively fewsamples (12 and three productive samples, respectively, Yang et al . 1986; Ouyang & Norris 1999), our sample set is morecomprehensive (39 samples from continuously measured sectionsin both the north and south limbs of the Dalongkou Anticline,and 21 samples from the Lucaogou section) and therefore provides better details of the stratigraphic distribution of taxa. Specimens of Aratrisporites spp. were not identified in our study,and we suggest that the previously reported specimens should bere-examined to ascertain and confirm that they belong to thisgenus. We have identified over 100 species of spores and pollenfrom the Chinese sections, and many of the taxa that from thetwo previous studies were apparently restricted stratigraphicallywere recovered from samples spanning almost the entire Guodi-keng Formation.Based on occurrences of at least 27 shared taxa, there is astrong correlation between palynofloras from the Chinese sec-tions and those reported from Nedubrovo (Afonin 2000). Keytaxa include: gymnosperm pollen Platysaccus queenslandi deJersey 1962, Klausipollenites schaubergeri (Potonie´ & Klaus)Jansonius 1962, Lunatisporites pellucidus (Goubin) Helby ex deJersey 1972, Lunatisporites transversundatus (Jansonius) Fischer 1979, Lueckisporites virkkiae (Potonie´ & Klaus) Clarke 1965, Decussatisporites mulstrigatus Hou & Wang 1986, Scutasporites sp. cf. S. unicus Klaus 1963, Ephedripites permasensis Yaroshen-ko 1997, Vesicaspora sp. A; lycopod and fern spores ‘ Rewanis- pora ’ sp. of Yaroshenko (Yaroshenko et al . 1991), Leptolepidites jonkeri (Jansonius) Yaroshenko & Golubeva 1991, Limatulaspor-ites fossulatus (Balme) Helby & Foster 1979; and algal remains Reduviasporonites chalastus (Foster) Elsik 1999. The appearanceof the distinctive megaspore Otynisporites eotriassicus Fuglewicz1977 in both the Chinese and Russian palynofloras is alsosignificant. O. eotriassicus was first described from poorly dated marine-influenced deposits in Poland (Fuglewicz 1977), where it is the Fig. 2. Stratigraphic distribution of Otynisporites eotriassicus Fuglewicz 1977 (shading and symbol) in key sections and proposed correlation with theGSSP for the Permian–Triassic boundary. Radiometric ages from Mundil et al . (2001). It should be noted that each section is drawn to scale. Datasources: 1, Looy (2000); 2, Kozur (1998); 3, this study (see also Fig. 3 for details); 4, P–T boundary placement based on vertebrate evidence (see Afonin2000; Lozovsky et al . 2001). J, Jiucaiyuan Formation; Ast, Astashikhiskaya Formation; V, Vyatsky Formation. Fig. 1. Late Permian reconstruction showing localities of Russian(Nedubrovo) and Chinese (Dalongkou) sections yielding abnormal pollen, and localities where the megaspore Otynisporites eotriassicus Fuglewicz 1977 has been recorded. 1, East Greenland; 2, Nedubrovo; 3,Poland; 4, Italy; 5, Dalongkou.C. B. FOSTER & S. A. AFONIN654 eponymous species of a zone of the Lower Bundsandstein(Fuglewicz 1980). More importantly, it occurs in close associa-tion with the P–T boundary marker H. parvus in marine sectionsin both Italy and Greenland (Fig. 2). In Italy, O. eotriassicus occurs within the uppermost Permian Tesero Member, and belowthe FAD of H. parvus (Kozur 1998). In Greenland, O. eotriassi-cus was reported from 10 samples from the uppermost PermianWordie Creek Formation, Jameson Land, 2 m below the FAD of H. parvus (Looy 2000; Looy et al . 2001). However, unlike therecord in Italy, the species is still present in the lowest Triassic,4 m above the occurrence of H. parvus . Recently, Looy (pers.comm.), has extended the range of O. eotriassicus into theuppermost 15 cm of the Schuchert Dal Formation (9.35 m abovethe base of the Oksedal Member), which, on faunas recovered from lower in the Oksedal Member, is dated to the Late Permian(Twitchett et al . 2001). The lack of fauna with O. eotriassicus atthis locality is problematic, because, as noted by Utting et al .(2004), the uppermost part (50 cm) of the Schuchert DalFormation at other localities in Jameson Land has been consid-ered of Early Triassic age, based on occurrences of certainlycopod spore taxa but not including Aratrisporites spp. That agedetermination has not been widely accepted, and has beenchallenged by Utting et al . (2004). Despite this new finding, weconcur with Lozovsky et al . (2001) and Krassilov (2003) that theFAD of O. eotriassicus is an important marker for the latestPermian–earliest Triassic. Some of the shared taxa have arelatively long time range in the Late Permian, but it is the co-occurrence of these taxa that is significant and, on this evidence,we conclude that the Nedubrovo and upper lower Guodikengassemblages are coeval (Fig. 2). Abnormal fossil pollen In both the Russian and Chinese assemblages there are signifi-cant numbers of abnormal morphotypes of gymnosperm pollen.For the type of pollen being considered here, normal grains are bisaccate; that is, with two air sacs, which aid dispersal as partof the fertilization process. A single abnormal pollen grain hasone, three, or four or more sacci (Fig. 4). As discussed below,species of living gymnosperms show morphologically identical pollen abnormalities. Apart from the number of sacci not beingequal to two, the abnormal pollen in the Chinese and Russianassemblages can be assigned confidently, using other morpholo-gical criteria, to different genera, including Klausipollenites , Alisporites , Scutasporites , Protohaploxypinus and Lunatisporites (Fig. 4). We note, however, that the abnormalities illustrated herewere not recognized in the previous studies, and that using saccinumber as a primary taxonomic criterion, grains with four sacciwere assigned to the genus Bascanisporites Balme & Hennelly1955 (see Yang et al . 1986, plate 26). This had the effect of increasing the apparent diversity of the assemblages, and, in thisexample, extending the geographical range of the genus. Thereare, of course, genuine polysaccate grains, such as Bascanispor-ites , and they too can demonstrate a range of morphologicalvariation, although bisaccate forms have yet to be illustrated.However, as demonstrated by this study, morphological variationneeds to be assessed in the context of the entire assemblage, and multiple taxonomic criteria used for species discrimination.Abnormalities among pollen assigned to Klausipollenites schaubergeri and Alisporites sp. are most common. In assem- blages from Nedubrovo, 5–6% of K. schaubergeri pollen areabnormal; similarly, 3–4% of this species from the upper lower Guodikeng Formation of the Dalongkou Anticline north limb areabnormal. Grains assigned to Alisporites sp. show 4% abnormal-ities in assemblages from both the north (sample 3373) and south(sample 3400) limbs of the Dalongkou Anticline (Fig. 3). Modern abnormal pollen: natural variation andresponse to external stress Natural variation Pollen abnormalities occur in both living gymnosperms and flowering plants. Causes of abnormalities are either naturalintraspecific variation (Pant & Bhatnagar 1973; Lindstro¨m et al .1997; Krassilov et al . 1999 b ; Pozhidaev 2000, 2001) or responses to environmental stress on the parent plant (see below):the resultant morphotypes are the same. To discriminate betweenresponse to external stresses and natural variation, we have used a measure of the percentages of abnormal grains: . 3% repre-sents a response to external stress. Our benchmark of 3% is based on observations of natural variation in living gymnos- perms, and on studies of pollen from single glossopterid sporangia from the Late Permian of Antarctica (Lindstro¨m et al .1997).In a study of pollen from a white spruce, Picea glauca (Moench) Voss, Wilson (1963) found that in 5000 grains 96.9%were normal bisaccate, 0.6% had reduced sacci, 1.4% weretrisaccate, and 1.1% monosaccate. Wilson commented that a previous study by Lakhanpal & Nair (1956) of Picea smithiana Fig. 3. Isotopic curves, abnormal pollen, and distribution of Otynisporites in Chinese sections. Placement of P–T boundary is under review by the current authors; above according to Ouyang & Norris(1999). Samples 3373 and 3440 are labelled.PERMIAN–TRIASSIC POLLEN 655 Fig. 4. Normal and abnormal pollen grains of Klausipollenites schaubergeri and Alisporites sp. The specimens that are identified by collection number 640 are housed at Geoscience Australia; those with collection number 4748 are housed at the Palaeontological Institute, Moscow. Sample and slidenumber as shown (6403378-8), and the specimen is located by England finder (H35). All scale bars represent 25 m. ( a )– ( f ) Pollen grains of Alisporites sp.: ( a )– ( c ) north limb, Dalongkou Anticline, China; ( d )– ( f ) south limb, Dalongkou Anticline, China. ( a ) Normal bisaccate (two detached saccatestructures) pollen grain, 6403373-8; H35. ( b ) Pollen grain showing lateral detachment of saccus exoexine to form a third saccate structure, 6403373-8;L45/2. ( c ) Trisaccate pollen, with three detached exoexinal saccate structures, 6403373-6; T19/1. ( d ) Trisaccate pollen grain; proximal equatorialdetachment of sacci suggests a monosaccate condition, 6403400-11; H19. ( e ) Trisaccate pollen grain, 6403400-12; N24/3. ( f ) Trisaccate pollen grain,6403400-2; T18/3. ( g )– ( l ) Pollen grains of K. schaubergeri : ( g )– ( i ) Nedubrovo, Russia; ( j )– ( l ) Dalongkou, China. ( g ) Normal pollen grain with twodetached exoexinal saccate structures, 4748/27; T51/1. ( h ) Trisaccate pollen grain with three detached exoexinal saccate structures, 4748/x; Z63. ( i )Tetrasaccate pollen grain detached exoexinal structures, 4748/36; J30. ( j ) Trisaccate pollen grain, 6403381-2R; V6. ( k ) Trisaccate pollen grain, 6403381-11; Y11/1. ( l ) Tetrasaccate pollen grain, 6403381-11; Y24/2.C. B. FOSTER & S. A. AFONIN656 (Wall.) Bioss. showed that 97% of grains were normal, and thatif he (Wilson) had calculated percentages according to their methods, 98.9% of grains from P. glauca would be classed asnormal and 1.1% as monosaccate. In a study of 42000 pollengrains from Pinus roxburghii Sarg., Srivastava (1961) noted that0.18% were abnormal, and of the abnormal grains 16% weretetrasaccate, 48% were trisaccate, and 36% disaccate. Plaxina(1969) noted that not more than 2% of bisaccate grains of Abies siberica Ledeb. showed a variation in size.Although lacking quantitative data, Vishnu-Mittre (1957) illu-strated abnormal gymnosperm pollen from India, and there arevery close morphological similarities to specimens in the presentassemblages. Mehra & Dogra (1965) described abnormal saccategrains from a single tree of Abies pindrow (Royal) Spach, and closely comparable morphotypes were found in this study (com- pare Fig. 4e and Mehra & Dogra 1965, fig.1).Lindstro¨m et al . (1997) analysed taeniate saccate pollen fromintact sporangia, assigned to Arberiella sp. cf. A. africana Pant& Nautiyal, recovered from the Permian Amery Group in thePrince Charles Mountains, East Antarctica. They reported thatfrom one sporangium, yielding 2656 pollen grains, only sixgrains (0.2%) were monosaccate, and one grain (0.03%) wastrisaccate; the remainder of grains were bisaccate. Although theglossopterid sporangia are found in Late Permian sediments, theyare older (palynological Stage 5) than the abnormal pollenassemblages from Russia or China. The study by Lindstro¨m et al . (1997) provides a measure of natural variation amongstPermian gymnospermous taxa. Response to modern external stress Wilson (1965) reported abnormal forms of pollen of Pinus flexilis James from ‘stunted trees and somewhat layered trees atthe tree line . . . in the Rocky Mountains of Colorado’. Thealtitude at the site is c . 3350 m, with tree heights of about 1.2 m;the habit was exposed to strong winds and severe frosts. Pollenof P. flexilis is normally bisaccate, and pollen from trees of 6–9 m height growing at altitudes of about 1800 m were ‘as freeof teratological forms as is possible’, and were used as areference, or normal, population (Wilson 1965). From a count of 5000 grains from a stunted and stressed tree, 12% of the pollenwas abnormal, and 94% of those grains belonged to onemorphotype, with a distally detached hemisphere monosaccus;the remainder belonged to five other morphotypes. Some of theseabnormal forms show close similarities to the monosaccate formsof Alisporites sp. found in our study.Pollen wall abnormality is common in modern plants and isused as an indicator of adverse external environmental condi-tions, caused either by atmospheric pollution or increased levelsof UV-B radiation or a combination of them. Pollution mayinclude: acid rain, from sulphur dioxide; aerosol emissions;increased carbon dioxide levels; toxins from industrial manufac-ture; heavy metal ions; and radiation. In Russia and Slovakia, for example, the increased percentages ( . 3%) of abnormal pollenfrom extant gymnosperms and angiosperms is used as a proxy tomeasure and monitor air pollution (Shkarlet 1972; Kormut’a´k et al . 1994; Ostrolu´cka et al . 1995; Kormut’a´k 1996; Dzyuba1998; Dzyuba et al . 1999, 2001; Dzyuba & Tarasevich 2001).The effect of radiation on pollen morphology resulting fromthe accident at the Chernobyl atomic power station, near Kiev,was shown by Sirenko (2001). Pollen abnormalities included gymnosperm grains ( Pinus ) with three or four sacci (instead of two), the absence of germinal apertures, wall thickenings, and reduction in size (among pollen of Alnus and Betula ). The abilityof such pollen to germinate, and effect fertilization, was reduced by between 20 and 100% within a 30 km zone around theChernobyl atomic power station (Sirenko 2001). It is importantto note that non-viable, non-germinating pollen grains of recent plants do not all display structural changes in wall structure(O. F. Dzyuba, pers. comm.), and so the percentage of abnormalfossil pollen morphotypes may underestimate the lack of viabili-ty of any given species.In discussing abnormal pollen of extant Pinus flexilis , Wilson(1965) concluded that ‘Teratology in pollen grains, can and doesin many cases, result in sterility which may have more far-reaching effects on the geological history of plants than is nowappreciated’. In the ensuing 40 years, few papers have discussed the significance or presence of abnormal pollen from thegeological record (see Visscher et al . 2004), and yet the extensiveevidence of volcanism at the P–T boundary (Campbell et al .1992), and its consequent effects on climate, suggests that suchrecords should be common. Discussion Using our benchmark of . 3% pollen abnormality as an indicator of environmental stress, we conclude that gymnospermous plantsfrom the Russian and Chinese sections were adversely affected by global atmospheric conditions. This is because the sectionsare essentially coeval and they are geographically separated,occurring on different palaeogeographical plates and thousandsof kilometres apart. Abnormality is not an outcome of geographi-cally localized environmental conditions.We concur with Krassilov (2003) and Visscher et al . (2004,and references therein) that a severe ecological crisis occurred onland at the close of the Permian. Records of intense volcanicactivity around the P–T boundary, with basalts covering over 1.5 3 10 6 km 2 in Siberia, and in South China an ash horizon of 10 cm thickness covering more than 1 3 10 6 km 2 (Yang et al .1995) suggest catastrophic atmospheric and climatic conditionsat this time. Stratospheric clouds from modern volcanic eruptionscontain, inter alia , sulphate aerosols that produce acid rain,enhance reflection of solar radiation, and cause significant strato-spheric ozone destruction, increasing UV-B radiation that causes plant mutagenesis (Barbera & Self 1978; Rampino et al . 1979;Teramura 1980; Tevini 1993; Manning & Tiedemann 1995;Briffa et al . 1998; Torabinejad et al . 1998; Robock 2000;Shumilov et al . 2000; Van Ulden & Van Dorland 2000).Field studies of the Chinese localities reported here have notyet proved the presence of volcanic ash beds, although clays are being still analysed (see Metcalfe et al . 2001), but there isevidence of volcanism in northern European Russia, with twogenerations of smectite (from altered volcanic ash) in sedimentsfrom Nedubrovo (Eroshev-Shak et al . 2002).Visscher et al . (2004) provided evidence of mutagenesisamong herbaceous lycopsids that they considered resulted fromenhanced UV radiation during the end of the Permian. The present paper provides the first evidence of mutagenesis from thegymnosperm flora.As a response to a global atmospheric phenomenon, thereshould be, prima facie , many records of abnormal pollen fromother localities at the P–T boundary. That there is not reflects both the incompleteness of the sedimentary record preserved (for example, the global stratotype for the P–T boundary is acondensed marine section with very poorly preserved and low-diversity palynological assemblages recovered (Ouyang & Utting1990)) and, perhaps more importantly, the taxonomy used byresearchers. As noted above, tetrasaccate specimens from the PERMIAN–TRIASSIC POLLEN 657