利用植物估计气候是定量重建过去气候参数的主要途径之一。共存似然估计正态分布法(P-CRACLE)是一种基于生物分类学的气候定量重建方法, 它利用全球范围内的植物分布数据来获取植物的耐受区间, 并假定植物分布的气候变量沿正态分布, 以此计算植物组合最有可能共存的气候区间。该方法基于全球现代植物样地较好地验证了该方法的可靠性, 但在获取植物分布数据以及气候数据的处理方面仍存在不足。为更深层次验证该方法的可靠性, 本文挑选了 26 个研究点, 基于全球生物多样性信息设施(GBIF)数据库获取植物的全球分布资料, 并使用两种分辨率(0.5 弧分、2.5 弧分)的气象网格数据(WorldClim 2.1)获取分布点的气候参数, 引入传统的相互气候范围方法(MCR)与 P-CRACLE 同时计算研究点的气候共存区间。结果显示, 使用不同分辨率气象网格数据的结果无明显差别, MCR 和 P-CRACLE 获取年平均温度的平均分辨率分别为 8.3 ℃、1.7 ℃, 年平均降水的平均分辨率为 1120 mm、280 mm, P-CRACLE 获取的气候区间分辨率远比 MCR 更精确。随后, 基于每一种植物分类群的气候参数到达研究点气候观测值的百分位, 选取百分位 10%–90%的植物分类群再次进行 P-CRACLE 的共存分析, 并对比了前人的共存结果, 重建的准确性显著提高, 这为未来提高古气候定量重建的准确性提供了方向。本研究认为, 利用 P-CRACLE 进行新生代以来的古气候定量重建具有非常好的应用前景。

Using plants as a proxy to estimate climate is one of the main approaches to quantitatively reconstructing past climate parameters. The parametric function of Climate Reconstruction Analysis using Coexistence Likelihood Estimation (P-CRACLE) is a quantitative climate reconstruction method based on biological taxonomy. P-CRACLE uses worldwide plant distribution data to obtain corresponding climate parameters. This method assumes that the occurrence of a species is normally distributed along the variables of the climate to calculate the climate intervals where plants coexist. Based on 165 modern floras of the world, the reliability of this method is well verified, but there are still controversies over the acquisition and processing of plant distribution and climate data. To further verify the reliability of this method, we selected 26 research sites of modern floras, obtained plant distribution data from the Global Biodiversity Information Facility (GBIF) database and the climate parameters by using the meteorological grid data (WorldClim 2.1) with two different resolutions (0.5 arcminute and 2.5 arcminute). We apply the traditional mutual climate range (MCR) and P-CRACLE to jointly calculate the climate parameters for all research sites. The results show no apparent difference when using different resolutions of climate grid data. Annual average temperatures obtained using MCR and P-CRACLE are 8.3 ℃ and 1.7 ℃, respectively, and the average resolutions of annual average precipitation are 1120 mm and 280 mm, respectively. Therefore, the resolution of climate intervals obtained by P-CRACLE is far more accurate than that by MCR. Then, based on the percentile of each plant taxon’s given climate parameter to the observed value of the research site, the plant taxon with a percentile higher than 10% and lower than 90% was selected to re-analyze the coexistence of P-CRACLE. Comparison with the coexistence results of previous studies shows that the accuracy of the reconstruction is significantly improved, providing a direction for increased accuracy of quantitative paleoclimate reconstruction in the future. This study suggests that P-CRACLE can be applied to the quantitative reconstruction of paleoclimate since the Cenozoic with very high resolution and reliability.

国家自然科学基金(42161144012, 41772181, 42030505, 41888101)和国家“万人计划”青年拔尖人才(2019000041)联合资助