1
Electronic Supplementary Material S8 to Miehe et al. “Persistence of Artemisia steppes in the Tangra Yumco basin, west-central Tibet (China) despite or in consequence of Holocene lake level changes?”
Palynological discussion of late Holocene climate change versus anthropogenic impact in the Tangra Yumco basin
(1) Methodological constraints
The climatic interpretation of the historical peat phase at Targo Xian remains vague because the set of grazing indicator pollen poses a methodological problem. In the Targo Xian record, these pollen types have their highest abundance both in the late Pleistocene wetlands (that were certainly not grazed by livestock) and in the modern swamp – a pattern found in pollen diagrams throughout the Tibetan highlands. The grazing indicators comprise a set of alpine species that invade montane vegetation types when grazing pressure is high. If this observation is ignored, the appearance of these pollen and spore types in pollen diagrams will be interpreted as a cold period (Shen 2003; Herzschuh et al. 2006, 2009 a,c; Shen et al. 2008; Herzschuh and Birks 2010). Shen (2003) has shown that the discrepancies between pollen spectra of modern vegetation types and those of palaeovegetation in central Tibet are weakest during the late glacial period, which is easily explained by the high proportion of alpine genera in the present set of human indicator pollen. To worsen the problem for the reconstruction of little-grazed palaeo-vegetation, strong wildlife concentrations seem to have similar effects on the vegetation as intensive livestock grazing. These phenomena strongly limit the transfer of modern pollen biomes (Prentice et al. 1996) to little-grazed palaeo-vegetation, as long as biome shifts are exclusively ascribed to climatic change.
(2) Vegetation development under strong human impact in the late Holocene at Targo Xian
At the end of the Tangra Yumco high stand (after 2.6 cal ka BP) proportions of charcoal and Glomus rise. Plantago appears at 2.6 cal ka BP together with strongly rising proportions of grazing indicators, especially Brassicaceae (Fig. 4: 8). A resident human population in the Tangra Yumco area is testified by the megalith assemblage at Sumbuk (30°30’ N / 86°36’ E). These assemblages usually accompany graves, dated between 2.84 cal ka BP and the 7th century AD (Bellezza 2008:91).
Highest charcoal and Glomus values are reached during the local lake phase at Targo Xian (180-140 cm) shortly before 1 cal ka BP. The increase in human-indicator pollen in PZ-3 (Fig. 4: 7, 8) suggests that human disturbance intensified, but this pattern may partly be ascribed to the region-wide cold period identified by Mischke and Zhang (2010) between 1.7 and 1.3 cal ka BP. The charcoal maxima might support the local oral history of a region-wide fire set by warriors of the Yarlung Dynasty in the 7th century AD, destroying the juniper forests and followed by a major drought (Nam Thak, pers. commun. 2003). The wetter period indicated in our core by the richness in fern spores and Artemisia pollen around 140 cm (~1 cal ka BP) and recognized region-wide after 1.5-1.3 cal ka BP (Shen et al. 2008; Demske et al. 2009; Mügler et al. 2010; Wrozyna et al. accepted), is in fact followed by the drought that led to the desiccation of the local lake at Targo Xian, resembling a phase of maximum aridity after 1.3 cal ka BP recognized at Bangong Co and Selin Co (Wei and Gasse 1999).
In any case the interpretation of pollen spectra in Tibet during this time segment should be done in the awareness of the expanding human activities during the Tibetan Empire (7th to 9th century AD).
The modern Cyperaceae bog (Fig. 4, PZ-4/5) has the whole set of human indicator pollen, including Cerealia (9) and fungal spores indicating dung (11), which shows a further intensification of land use, as follows.
PZ-4 gives clear evidence of agricultural activities, by the Cerealia record accompanied by Pteris-type and six pollen types which partly appear for the first time at 115 cm, indicating wastelands and cultivation (Fig. 4: 9). The combined first appearance of Apium-type, Asteraceae Subfam. Asteroideae, Boraginaceae, Sedum-type(incl. Rhodiola) and Trollius, the reappearance of Gentiana himalayensis-type, Leontopodium-type, Papilionaceae (first recorded together with Cercophora in PZ-1), Ranunculus acris-type and Vicia-type as well as the maxima of Asteraceae Subfam. Cichorioideae (Taraxacum, Youngia) and Saussurea, suggest the intensification of grazing around the same time (~0.9-~0.7 cal ka BP), even though spores of coprophilous fungi are not yet recorded.
The last major fire event is documented in our core at 95 cm (PZ-5A). It might be connected with the Mongol incursions in the early 13th century AD (Bellezza 2008). The peat itself does not show any carbonization. The interesting climatic information of this event is that the standing biomass of the zonal vegetation was still dense enough to support the spreading of fire. However, the drop in Cyperaceae pollen combined with a peak in Ephedra can be connected with drought or extensive fires (highest charcoal concentrations since PZ-1) and grazing. The peak in Botryococcus indicates more water in the ponds, and Chenopodiaceae show a relative minimum, which also speaks against severe drought conditions. More pollen of grazing indicators appear (Fig. 4: 9) or increase in proportion (Leontopodium-type, Parnassia). Obviously a phase of more intensive land use with respective erosion is documented at this time, but additionally, effects of the Little Ice Age cooling (0.6-0.1 cal ka BP, Mischke and Zhang 2010) might be reflected by these pollen assemblages.
The pollen data above 97 cm depth are a good example of the persisting dilemma in separating climatic and anthropo-zoogenic effects. The 7 cm thick sand layer in the peat, low pollen concentrations, and declines in Cyperaceae and Artemisia combined with a peak in Ephedra can all be taken as indications of drought and/or cold and the connected morphogenetic disturbances. However, Artemisia may be favoured or reduced by grazing, depending on the climate and the species present. The less favourable the climate, the more vulnerable are Artemisia species to grazing, and the longer they need to regenerate, especially the suffruticose ones (A. prattii (Pamp.) Ling & Y.R. Ling nowadays dominant, see Electronic Supplementary Material S2). Moreover, Artemisia species are less dominant in alpine steppes than in montane (“temperate”) steppes (Shen 2003; Herzschuh et al. 2009a). Thus a retreat in Artemisia may be due to desiccation, cooling or intensified grazing and/or burning.
Extra-regional palynomorphs show low arboreal pollen values and the first record of Pteridium aquilinum (fire successor like Pinus) around 0.7 cal ka BP. The highest concentrations of Pinus in PZ-4 (Electronic Supplementary Material S7), followed by the appearance of Quercus griffithii Hook. f. & Thoms. (secondary forests, village woodlots) are synchronous with the largest set of human indicator pollen at the study site. This suggests intensified human activities in the wider region including the Himalayan forests (Miehe et al. 2009; La Duo 2008).
Around 60 cm depth (c. 0.5 cal ka BP), further intensification of land use is indicated by the reappearance of the dung fungal spores (Cercophora-type) and of Euphorbia, the first records of Anaphalis and absolute maxima of Leontopodium-type, Potentilla-type and Sedum-type. High Cyperaceae values suggest that the swamp vegetation was well developed (no charcoal at 62 cm!). Without the analysis of fungal spores, and in view of the Little Ice Age as a probable climatic framework, part of these species may exclusively suggest climatic cooling; compare their presence in PZ-1 and Herzschuh et al. (2006) for respective conclusions. Gasse et al. (1996) and Van Campo et al. (1996) suggest a dry period around 0.7 cal ka BP in western Tibet; Kasper et al. (2012) at Nam Co around 0.9–0.5 cal ka BP, and Shen et al. (2008) at Co Ngoin at 0.7–0.3 cal ka BP. Except for the minimum in Botryococcus, which may also be due to cold, there are no indications of these droughts in our diagrams. Morrill et al. (2003) and Mischke and Zhang (2010) stressed that this regional climatic event had spatially very diverse effects. Lacking consideration of human impact in the interpretation of pollen diagrams may contribute to this seeming diversity.
Above 50 cm depth (~0.5 cal ka BP) PZ-5 shows strikingly low Cyperaceae values and a maximum of Glomus. Weak representation of Cyperaceae and the muddy and light-coloured substratum can be attributed to either inundation, drought, or strong grazing impact. Grazing impedes sedge flowering, and thus pollen production, which is common under the present grazing pressure. While Cyperaceae pollen decreases in the upper PZ-5, that of Potentilla-type (probably P. anserina L., colonizing muddy soil at the margin of the Kobresia schoenoides wetland) and spores of the dung indicators Cercophora and Sporormiella increase. This is easily explained with strongest grazing impact of livestock, which is additionally indicated by high values of Gentianahimalayensis-type, Parnassia and Thalictrum and by the appearance of Swertia (S. younghusbandii Burk.). These small rosette plants can establish only when the tall and dense Kobresia schoenoides is suppressed. The strong increase of Arcella can be explained by larger surfaces of moss and open substrate replacing Cyperaceae on hummocks. All of the mentioned species groups are at present only found in these water surplus sites, which supports the assumption that they replaced part of the Cyperaceae in consequence of over-grazing. This period had little charcoal influx. Obviously the vegetation cover was not dense enough to favour the spreading of fires, and the Cyperaceae bog had become a too valuable fodder resource to be set on fire. Both explanations would indicate present-day conditions. Discrepancies in the doubled upper 10 cm (PZ-5A from muddy surface, PZ-5B from Cyperaceae patch) are explainable with different degrees of compaction (uncertain chronology) and disturbance (more bioturbation due to trampling in PZ-5A).
Comparison with the pollen assemblages of the initial swamp (PZ-1) suggests a distinctively smaller impact of wildlife grazing on the swamp vegetation in the early Holocene.
References not cited in the main paper:
Demske D, Tarasov PE, Wünnemann B, Riedel F (2009) Late glacial and Holocene vegetation, Indian monsoon and westerly circulation in the Trans-Himalaya recorded in the lacustrine pollen sequence from Tso Kar, Ladakh, NW India. Palaeogeogr Palaeoclimatol Palaeoecol 279: 172–185
Herzschuh U, Winter K, Wünnemann B, Li S (2006) A general cooling trend on the central Tibetan Plateau throughout the Holocene recorded by the Lake Zigetang pollen spectra. Quat Int 154/155: 113–121
Herzschuh U, Kramer A, Mischke S, Zhang C (2009c) Quantitative climate and vegetation trends since the Late Glacial on the north-eastern Tibetan Plateau inferred from Koucha Lake pollen spectra. Quat Res 71: 162–171
Mügler I, Gleixner G, Günther F, Mäusbacher R, Daut G, Schütt B, Berking J, Schwalb A, Schwark L, Xu B, Yao T, Zhu L, Yi C (2010) A multi-proxy approach to reconstruct hydrological changes and Holocene climate development of Nam Co, Central Tibet. J Paleolimnol 43: 625–648
Prentice IC, Guiot J, Huntley B, Jolly D, Cheddadi R (1996) Reconstructing biomes from palaeoecological data: a general method and its applications to European pollen data at 0 and 6 ka. Climate Dynamics 12: 185–194