Macrobenthic Species Distribution Modeling Based on a Binary Logistic Regression:A Case Study of Freshwater Restoration Wetlands in the Yellow River Delta of China
YANG Wei,LI Xiaoxiao,LI Ming,SUN Tao
DOI: https://doi.org/10.5846/stxb201608261739
2017-01-01
Abstract:Studies on changes in the macrobenthic species distribution could help to understand the ecological restoration efficiency of wetlands and elucidate the ecological succession process.Based on the results of field samplings and species identification from 2014-2015 in the Yiqianer national nature reserves in the Yellow River Delta wetlands of China,in which freshwater release has been implemented since 2010,we selected three characteristic,dominant macrobenthic species,namely,Chironomidae species,Sinocorophium sinensis,and Alitta succinea,corresponding to freshwater species,euryhaline species and brine species,respectively.We selected salinity,pH,water moisture,total organic carbon (TOC),and median particle size of sediment as environmental variables.Based on the binary logistic regression approach,in which a binary variable (0/1) indicates whether the target species were found at the sampling sites,80 groups of the field sampling data were used to calibrate the parameters of the distribution possibility model for the three species,whereas the other 20 groups were used to verify the simulated accuracy of models.Based on the simulated distribution possibility of the target species at the sampling sites,we spatialized their occurrence possibility using inverse distance weighted interpolation and characterized their changes on a regional scale.We found that the models for Chironomidae species and A.succinea corresponded well with the survey results and their precision rates reached 84.9％ and 77.9％,respectively.In contrast,the accuracy of the simulated occurrence probability of S.sinensis reached only 26.7％,which may be because of the mobility traits of S.sinensis and small sampling data,and because the selected environmental variables could not cover the main limited survival factors for S.sinensis.The distribution possibility for A.succinea showed no significant difference between before and after freshwater release.In general,the distribution possibilities of A.succinea in the study region were lower than 0.55.Furthermore,the areas with distribution probabilities between 0.4 and 0.55 accounted for about 28.2％ of the study region and were concentrated in the intertidal zone.The main high distribution possibility area for Chironomidae species was the freshwater restoration area with the implementation of freshwater release.The area with a distribution possibility ＞ 0.5 for Chironomidae has increased 9.9-to 10.8-fold after the ecological restoration.Owing to the low accuracy of the simulated distribution possibility,we did not spatialize its occurrence possibility of S.sinensis on a regional scale.The steep increase in the occurrence possibility of Chironomidae species after freshwater release indicated that the restoration wetlands increasingly showed obvious freshwater characteristics,consistent with the significant reed growth and increase in the freshwater habitat patches observed during field investigation.Although the freshwater habitat succession was unpredictable owing to significant fluctuations with respect to release periods,we should take essential measures,such as continuous freshwater release,long-term monitoring system,quality improvement of released freshwater,and cost reduction and high frequency of freshwater release,to maintain the long and difficult transition.