古植被定量重建与R值模型的发展——R值和ERV模型在兴隆山地区的应用

QUANTITATIVE RECONSTRUCTION OF PALEOVEGETATION AND DEVELOPMENT R-VALUE MODEL:AN APPLICATION OF R-VALUE AND ERV MODEL IN XINGLONG MOUNTAIN NATURAL PROTECTION REGION

基于兴隆山自然保护区及其周边地区30个样方的植被调查和表土花粉分析,运用R值计算方法和ERV模型估算了我国北方常见的几种乔木和草本植物花粉的代表性和相对花粉产量。R值显示蒿属花粉为超代表性,云杉属和沙棘属代表性略高,栎属和桦属代表性适中,禾本科和莎草科为低代表性。ERV模型估算结果显示云杉属、沙棘属和栎属的花粉产量较高,桦木属的花粉产量较低,而禾本科和莎草科的花粉产量很低。通过对结果的分析表明:云杉属、沙棘属、栎属、莎草科和禾本科根据R值计算得到的代表性结果和ERV模型估算得到的花粉产量结果有较好的一致性,但有很高背景花粉成分的桦属和蒿属的花粉产量较低,主要是由于R值没有考虑背景花粉成分对花粉百分比的影响。因此,在进行景观尺度的植被重建时一定要考虑背景花粉成分的影响,以确定当地是否有对应的母体植物大量分布。

On the basis of studies on 30 surface samples collected from Xinglong Mountain Natural Protection Region （35°38＇ -35°58＇N,103°50＇- 104°10＇E） and its surrounding area in the Gansu Province,this paper focused on the quantitative relationship between vegetation and pollen in arid area of Northwest China. All of the samples are moss, of which 11 samples were collected from montane shrub at an altitude of about 1800 -3000m,2 samples were collected from coniferous forest at an altitude of 2300 - 2500m,4 samples collected from mixed coniferous forest at an altitude of 2400 -2600m, 8 samples collected from deciduous broad-leaved forest at an altitude of 2200 - 2800m,3 samples collected from alpine shrub at an altitude of 3000 - 3500m and 2 samples from alpine meadows above an altitude of 3500m. The vegetation around the sampling localities was investigated within the area of 20- 20m2. Our result indicates that 18,766 pollen grains belong to 47 families and 35 genera were identified,including conifer mostly Pinus and Picea; deciduous broad-leaved mainly Quercus and Betula; shrubs and herbaceous are more abundant, which shrubs include Hippophae, Corylus/Ostryopsis, Rosaceae and Lonicera; herbaceous mainly Artemisia, Compositae, Gramineae and Cyperaceae. The representation and relative pollen productivity （RPP）of several common arboreal and herbaceous taxa in North China were calculated by R-value and Extended R-value model（ERV model）based on these pollen and vegetation data. The calculation results show that the R-values of Quercus, Picea,Betula, Hippophae,Artemisia, Gramineae and Cyperaceae are 0. 70 ,1. 79 ,1. 23 ,1. 78 , 7. 90 , 0. 30 and 0.03 ,respectively. The R-value reveals that Artemisia is over-represented, Picea and Hippophae have slightly high representation and Quercus and Betula have moderate representation,whereas Gramineae and Cyperaceae are under-represented. When the Quercus is used as reference taxon, the RPPs of Picea,Betula, Hippophae,Artemisia, Gramineae and Cyperaceae are 2.96,0. 16,1.67,9.08,0.01 and 0. 004, respectively based on the estimates of submode 1 ; while the RPPs are 4.27,0. 18,1.46,0. 32,0. 00002 and 0.02, respectively based on the estimates of submode 2. The results of ERV model show that Picea, Hippophea and Quercus have relatively high RPP,Betula have relatively low RPP,while Gramineae and Cyperaceae have very low RPP values. The results demonstrate that Picea,Hippophae, Quercus, Gramineae and Cyperaceae have good correlations between presentations calculated by R-values and RPPs estimated by ERV model. Previous modeling calculations consider impact of background pollen components to pollen percentage,leading to that Betula and Artemisia have low RPPs with high background pollen components. Our calculation results suggest that the impact of background pollen components have to be considered during the quantitative reconstruction of paleovegetation on landscape scale, in order to determine whether a large number of the parent plants distribute in local.