3,600 years of human adaptation to drought intensification on the southern Tibetan Plateau

The Tibetan Plateau (TP), known as Earth’s “Third Pole,” averages over 4000 m a.s.l. and presents cold temperatures, aridity, and highly variable weather, constraining resources and human habitation. Despite these challenges, Tibetans developed agro-pastoral systems several thousand years ago. Yet, how these survival strategies evolved—especially the transition toward a modern economy dominated by pastoralism with supplemental farming—and the role of climate remain unclear. Prior work indicates that the introduction of cold-tolerant barley and wheat enabled adaptation after ~3600 cal yr BP and supported permanent settlement mainly below 3500 m a.s.l., leaving a knowledge gap for higher elevations that comprise ~85% of the TP. Archaeological, zooarchaeological, paleoproteomic, lipid residue, and genetic studies provide pieces of the puzzle but remain spatially and temporally fragmented for high-elevation regions. Microfossils preserved in sediments offer continuous, high-resolution archives of agricultural and pastoral activity. This study analyzes pollen, coprophilous fungal spores, and charcoal from a precisely dated sediment core from Angrenjin Co on the southern TP to determine the history and drivers of high-altitude subsistence strategies (arable agriculture and pastoralism) over the past ~3600 years.

- Barley/wheat adoption after ~3600 cal yr BP facilitated human adaptation and permanent settlement in parts of the TP, mainly at lower elevations (<3500 m a.s.l.). - High-elevation archaeological records (e.g., Bangga, Khog Gzung, Qugong) indicate well-adapted agro-pastoral systems in the late Holocene, but details of pastoral evolution remain unclear. - Zooarchaeology, paleoproteomics, and lipid residues suggest pastoralism and dairying emerged ~3500–3000 years ago in higher regions; yak domestication dates to ~7300 years ago. However, evidence is scattered and sometimes inconsistent. - Sedimentary microfossils (e.g., Sporormiella-type spores, crop pollen) in other world regions (Andes, Nile Delta, northern China) have tracked shifts between herding and farming tied to climate variability (e.g., aridity, monsoon changes, ENSO), motivating similar approaches on the STP. - Existing TP records show increased grazing signals at various times during late Holocene, possibly linked to climatic deterioration and/or social factors, but continuous, high-resolution histories for the southern high-altitude TP have been lacking.

Study site: Angrenjin Co (29°17′–29°20′N, 87°9′–87°13′E; 4306 m a.s.l.) on the southern Tibetan Plateau is a closed-basin lake (watershed 194 km²) in a semi-arid alpine climate. Mean annual precipitation ~300 mm (~90% June–September); mean annual temperature ~7.1 °C (1977–2015). Surrounding vegetation is alpine steppe/meadow; local agriculture includes high-altitude barley, rapeseed, and wheat; livestock include yaks, sheep, and goats. Core recovery and chronology: A 142 cm gravity core (AR-02) was taken near the lake center at ~14 m water depth in October 2011. Lithology: gray silt (142–80 cm), interbedded black/gray silt (80–50 cm), black silt (50–0 cm). Eight radiocarbon ages on bulk organic matter (Beta Analytic) were calibrated with IntCal20 in OxCal 4.4.1. The upper 40 cm were dated with 210Pb (CRS model) and 137Cs at 1-cm resolution. Paleomagnetic secular variation (PSV) from Angrenjin Co was compared to an East Asia PSV stack and the Pfm9k.1a global geomagnetic model. Ages were integrated using Bacon age modeling. Independent 210Pb–137Cs and PSV validations indicate the core spans ~3600 years. Pollen, fungal spore, and charcoal analyses: 142 samples (0–142 cm) were processed using standard acid–alkali treatments. Lycopodium tablets (27,560 spores/tablet) were added for concentration calculations. Residues were sieved at 7 µm with ultrasonic agitation. Identification of fungal spores followed published plates/descriptions. Counts were made at 400× magnification using a Leica DM 750. Data were plotted and analyzed with Tilia 3.01 and stratigraphically constrained cluster analysis (CONISS). Targeted counts included Poaceae pollen grains >40 µm, coprophilous fungal spores (Sporormiella-type, Coniochaeta sp.), Glomus-type bodies/cells, Gelasinospora sp., and charcoal particles >100 µm. Stellera pollen was identified following prior work. Indicator group definitions: - Cereal indicators: Poaceae >40 µm (a robust proxy for arable activity) and the ratio of cereal-type Poaceae to total Poaceae. - Grazing indicators: Stellera abundance and coprophilous fungal spores (Sporormiella-type, Coniochaeta sp.) as proxies for large herbivore presence and grazing intensity. - Erosion indicator: Glomus-type (arbuscular mycorrhizal fungal remains) as an indirect proxy for soil erosion/disturbance. - Fire indicators: Gelasinospora sp. (fires stimulate germination/fruiting) and charcoal >100 µm (local fire intensity). Trend analysis: Generalized additive models (GAMs) with restricted maximum likelihood (REML) fits were applied to regional climate proxy time series (e.g., pollen PCA1, geochemical PCA1, reconstructed summer precipitation from Angrenjin Co; grain-size PCA1 from Paru Co; median grain size from Nam Co; reconstructed mean annual air temperature) to infer smoothed trends and uncertainties (mgcv package in R).

- Two main zones over the past 3600 years were identified by CONISS and indicator concentrations: • Zone 1 (3600–1800 cal yr BP): High cereal signal and low grazing signal. - Poaceae >40 µm: mean 54 grains/g (range 0–157 grains/g); high cereal-type Poaceae/total Poaceae ratio. - Grazing indicators low: Stellera mean 5 grains/g; Sporormiella-type mean 37 spores/g; Coniochaeta sp. mean 3 spores/g. - Erosion indicator present: Glomus-type mean 332 spores/g. - Fire indicators low: sporadic Gelasinospora; charcoal >100 µm mean 2 particles/g. • Zone 2 (1800 cal yr BP to present): Declining cereal signal and sharply increased grazing and fire signals. - Poaceae >40 µm declines to mean 32 grains/g; cereal-type/total Poaceae ratio low. - Grazing indicators increase: Stellera mean 8 grains/g; Sporormiella-type mean 157 spores/g; Coniochaeta sp. mean 18 spores/g. - Erosion indicator persists: Glomus-type mean 376 spores/g. - Fire indicators increase: frequent Gelasinospora; charcoal >100 µm mean 14 particles/g. - Subsistence interpretation: • 3600–1800 cal yr BP: Dominant arable agriculture supplemented by low-intensity pastoralism; peak arable intensity ~2000 cal yr BP with rapid decline after ~1800 cal yr BP. • 1800 cal yr BP–present: Intensified, dominant pastoralism with low-level cultivation; increased local fires likely linked to dung fuel use accompanying intensified herding. - Climatic linkage: • Stage 1 corresponds to relatively wet conditions on the southern TP (multiple proxies, GAMs). • After ~1800 cal yr BP, regional proxies indicate drying (weakened monsoon, higher evaporation) and warming (enhanced mean annual air temperature), coincident with the subsistence shift. • Drought likely reduced seasonal rainfall below barley’s minimum requirement (~378 mm) and increased evaporation, constraining cultivation and favoring pastoralism. - Archaeological concordance: Zooarchaeological records show low domesticate proportions during Stage 1 and high proportions during Stage 2 at high-altitude sites; archaeobotanical evidence aligns with an arable peak near ~2000 cal yr BP and decline thereafter.

The study reconstructs a continuous, high-resolution history of human subsistence strategies in a high-altitude region of the southern Tibetan Plateau, demonstrating a clear two-stage evolution. During 3600–1800 cal yr BP, strong cereal pollen signals and low coprophilous fungal spores indicate a system dominated by arable agriculture with supplemental herding. After ~1800 cal yr BP, cereal indicators decline while coprophilous spores and Stellera increase, showing intensified pastoralism, accompanied by increased local fire activity likely related to the long-standing practice of dung fuel use in fuel-scarce alpine environments. These transitions align temporally with regional climatic changes reconstructed from multiple proxies and summarized with GAMs: a shift from relatively wetter conditions to drier, warmer conditions, including weakened monsoon precipitation and increased evaporation. The inferred mechanism is that drying reduced seasonal rainfall below the threshold required for high-altitude barley cultivation and, together with higher evaporative losses, depressed soil moisture and crop yields. In response, human groups adapted by expanding pastoralism (economically more resilient under aridity) and reducing cultivation intensity. The archaeological record corroborates these patterns: domesticate proportions are low in Stage 1 and high in Stage 2 at high-elevation sites, while archaeobotanical data show an arable peak near ~2000 cal yr BP and a rapid decline thereafter. Overall, the findings indicate adaptive resilience of high-altitude populations on the STP to late Holocene drought intensification, with a strategic shift toward pastoralism that likely underpins aspects of the modern plateau economy.

Well-dated, high-resolution pollen, coprophilous fungal spore, and charcoal records from Angrenjin Co reveal a two-phase evolution of subsistence over the past ~3600 years in the high-altitude southern Tibetan Plateau. Between ~3600 and 1800 cal yr BP, subsistence was dominated by arable agriculture with relatively low-intensity pastoralism. After ~1800 cal yr BP, intensified pastoralism became dominant while arable agriculture decreased markedly. Regional paleoclimate proxies indicate that this transition coincided with intensified drought and rising temperatures, suggesting that increasing aridity drove the adaptive shift from cultivation toward herding. This transformation likely laid the foundation for the modern pastoralism-focused economy of the high-altitude TP.

- Archaeological evidence at high elevations remains spatially patchy; some inferences rely on integrating scattered site records, which may bias temporal continuity. - Direct evidence for dairy (milk proteins) in the southern TP is scarce, limiting precision on the chronology and extent of dairying. - Archaeological evidence of cultivation after ~1800 cal yr BP is sparse and could reflect both genuinely reduced farming and insufficient excavations. - Chronology relies partly on bulk-organic 14C ages (due to absence of macrofossils), which can be affected by old-carbon effects; this was mitigated by independent 210Pb–137Cs and PSV constraints but residual uncertainty remains. - Proxy–process relationships (e.g., pollen size thresholds, coprophilous fungi abundance) are indirect and can be influenced by multiple environmental and taphonomic factors; local differences among paleoclimate records introduce interpretative uncertainty.