A comparison of relative pollen productivityfrom forest steppe, typical steppe and desert steppe in Inner Mongolia

Relative pollen productivity (RPP) is the ratio of the absolute pollen productivity of a certain pollen taxon to the pollen reference taxon. It is an important parameter that can be used for quantitative reconstruction of paleovegetation at a landscape scale. Here we report a study on the estimation of the RPPs of five common pollen types of grassland vegetation from forest steppe, typical steppe and desert steppe in Hulunbuir, Xilinhot and Sonid Left Banner of Inner Mongolia. We used a 1 m×1 m quadrat to record plant species, rooted frequency and percentage cover at upwind to downwind direction at 0, 1, 2, 3.5, 5.5, 7.5, 9.5, 14.5 and 19.5 m distance from the trap. For the sites of 20–100 m distance from the trap, we investigated vegetation composition of the 1 m × 1 m quadrat with a 10-m interval. In the three study areas, Artemisia , Compositae, Gramineae, Chenopodiaceae and Cyperaceae were dominant plants. In the Hulunbuir forest steppe, Artemisia and Gramineae had the highest percentage. In the Xilinhot with typical steppe and the Sonid Left Banner with desert steppe, the percentage of Gramineae was the highest. We collected the pollen samples using modified Tauber traps from 2007 to 2013. The percentage of Chenopodiaceae in the Hulunbuir forest steppe was the lowest, but reached the highest in the Sonid Left Banner desert steppe, in contrast to those of Compositae, Gramineae and Cyperaceae. Our results show that the fall speed ( v g ) varies among the five pollen types with 0.032 m/s for Gramineae, 0.027 m/s for Cyperaceae and Chenopodiaceae, 0.021 m/s for Artemisia and 0.019 m/s for Compositae. The RPP Che. was found the highest and RPP Gra. the lowest in the three areas. When we used Artemisia as a reference taxon, the RPPs of the same pollen type are different in the three areas. For example, in the Hulunbuir forest steppe, the RPP Art . was 1.00, RPP Com . was 0.33, RPP Gra . was 0.44, RPP Che . was 4.78 and RPP Cyp. was 1.21. However, in the Xilinhot typical steppe, RPP Art. became 1.00, RPP Com . 4.72, RPP Gra . 0.01, RPP Che . 2.05 and RPP Cyp. 1.31×10 - 4 ; in the Sonid Left Banner desert steppe, RPP Art . was 1.00, RPP Com . 6.72, RPP Gra . 2.15, RPP Che . 36.25 and RPP Cyp . 0.41 respectively. The main reasons for the RPP differences might be the species dissimilarity in diverse vegetation types and the number of the sampling sites used in the models. The estimated relevant source areas of pollens (RSAPs) in all the three areas were less than 20 m. The coverage of different plants was reconstructed by the REVEALS model and presented a large difference among the three areas. In the Hulunbuir forest steppe, the Artemisia ’s coverage was 29.16%, Compositae18.66%, Gramineae 39.40%, Chenopodiaceae 3.53% and Cyperaceae 9.25%. In the Xilinhot typical steppe, the coverage of Artemisia reached 13.42%, Compositae 0.12%, Gramineae 61.69%, Chenopodiaceae 17.59% and Cyperaceae 7.17%. And in the Sonid Left Banner desert steppe, Artemisia reached a coverarge of 93.81%, Compositae 0.59%, Gramineae 1.22%, Chenopodiaceae 3.13% and Cyperaceae 1.25%. Compared with the actual vegetation coverage, the reconstructed coverage of Artemisia and Chenopodiaceae are higher, while that of Gramineae is far lower, which might be caused by the high background value of Artemisia .