Increasing human activities during the past 2,100 years in southwest China inferred from a fossil pollen record

Besides natural climate variations, human activities can also cause significant changes in regional vegetation composition, especially during the late Holocene period. Southwest China has experienced a long history of human activities in terms of deforestation and agricultural development. Here we present a continuous fossil pollen record covering the past 2,100 years from Beihai Wetland in western Yunnan Province, for a better understanding of regional vegetation history and potential human activities. The pollen results indicate that the forest components show decreasing trends, including Alnus dominated deciduous broadleaved forest and evergreen broadleaved forest represented by evergreen Quercus. On the other hand, herbs such as Poaceae and Artemisia expanded along with the decline of forest, especially during the period after AD 1120. The results of ordination analysis further confirm such a pattern in which the first PCA component (PC1, capturing 42.2% of total variance) represents an increasing trend, which is subsequently interpreted as an index of human activities. Our findings show broadly identical patterns to previously published sedimentary records from Yunnan as well as historical documents. The human activities increased following the expansion in regional population, in the form of deforestation and regional agricultural development. They may also have caused the expansion of herbs, especially Poaceae, which replaced forest. Therefore, as indicated by the fossil pollen record from the Beihai Wetland, human activities in southwest China increased gradually during the past 2,100 years and more rapidly after AD 1120. Results of sediment influx and grain size analyses showed generally consistent patterns with the signs of summer monsoons captured by stable oxygen isotope records from stalagmites, indicating limited influence from human activities within the wetland system compared to the regional landscape.