东北长白山泥炭地硅藻−水位埋深转换函数

Diatom-based water-table depth transfer function from peatlands in the Changbai Mountains,Northeast China

本研究基于2016、2019和2021年夏季在东北长白山地区7处泥炭地采集的160个表层样品和环境因子数据,结合表层样品硅藻分析,构建了硅藻属种与环境因子数据库,该数据库包括64个硅藻属种和11个环境变量。优势硅藻属种组合沿水位埋深梯度(0~66 cm)呈现明显分异,偏典范对应分析结果表明,水位埋深单独解释了硅藻组合数据总方差的8.4%,远大于其他环境因子的单独解释份额。基于此,选取加权平均模型和最大似然估计模型构建硅藻−水位转换函数。不同模型分析结果显示,典型回归耐受值降权加权平均模型构建的硅藻−水位转换函数表现最佳,剔除残差大于20%水位梯度(13.2 cm)的样品后进行留一法交叉检验,该模型的推导值与实测值相关系数最大和推导误差最小。本文构建的硅藻−水位转换函数为定量反演长白山地区泥炭地过去水位变化提供了新的有效方法。

Peatlands,as unique limno-terrestrial ecosystems,play an important role in water storage,biodiversity conservation and global carbon cycling.Water table is a key environmental variable controlling the biogeochemical process in peatlands,and hence it is of great significance to reveal water-table changes for further understanding the evolution of peatlands.Based on one hundred and sixty samples collected from seven peatlands in the Changbai Mountains,Northeast China,in the summer of 2016,2019 and 2021,this study established a dataset consisting of sixty-four diatom species and eleven environmental variables.The diatom assemblages registered clear variations along the water-table gradient from 0 to 66 cm.Partial CCA revealed that Depth to the Water Table(DWT)independently explained 8.4%of the total variance in diatom composition,which was much higher than the sole effect of other environmental factors.Diatom-based water-table transfer function was developed using the Weighted Averaging model and the Maximum Likelihood model,respectively.Our results showed that tolerance downweighted Weighted Averaging model with classical deshrinking(WA.cla.tol)had the best performance.After removing the samples with residuals>20%of total measured water table range(13.2 cm),the WA.cla.tol model yielded the highest coefficient of determination(R2)between diatom-inferred DWT and observed DWT,and the lowest Root Mean Square Error of Prediction(RMSEP)in leave-one-out cross validation.The diatom-based water-table transfer function built in this study provides a new and effective method for quantitative reconstruction of water-table changes in peatlands of the Changbai Mountains.