首页 | 期刊简介 | 编委会 | 投稿指南 | 常用下载 | 联系我们 | 期刊订阅 | In English
引用本文:刘璐璐,赵亮,蔺诗颖,冯建龙.基于MaxEnt和GARP的阿蒙森海域南极磷虾(EUPHAUSIA SUPERBA)的分布区预测.海洋与湖沼,2023,54(2):399-411.
【打印本页】   【HTML】   【下载PDF全文】   查看/发表评论  下载PDF阅读器  关闭
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 389次   下载 774 本文二维码信息
码上扫一扫!
分享到: 微信 更多
基于MaxEnt和GARP的阿蒙森海域南极磷虾(EUPHAUSIA SUPERBA)的分布区预测
刘璐璐, 赵亮, 蔺诗颖, 冯建龙
天津科技大学 海洋与环境学院 天津 300457
摘要:
南极磷虾是南大洋生态系统的关键物种,在南极碳汇过程中起到重要作用,近年来受到越来越多的关注。针对位于南大洋太平洋扇区的阿蒙森海域,运用最大熵模型(maximum entropy modeling,MaxEnt)和预设规则的遗传算法(genetic algorithm for rule-set production,GARP)两种生态位模型,基于已采集的南极磷虾分布点的数据,对其在阿蒙森海域的潜在分布区进行了预测和分析,并采用受试者工作特征曲线(receiver operating characteristic curve,ROC)下的面积(area under curve,AUC)和真实技巧统计法(true skill statistic,TSS)对模型结果进行评估。结果表明:MaxEnt模型中的高适生区刻画细致,GARP模型预测的高适生区分布范围更广。为克服单个模型的不确定性得到更佳结果,将两个模型的预测结果进行集合。集合后的结果模拟精度显著提高(AUC为0.946,TSS为0.78),达到了极好的预测效果。磷虾的高适生区集中分布在65°~73°S,占总面积的6.2%,中适生区占总面积的5.7%。海冰、平均海平面气压最小值和纬向流速最大值是MaxEnt中贡献最高的3个变量,3个变量贡献达81.3%。相较于MaxEnt模型,GARP模型中各个变量遗漏误差相对较平均。研究表明,集合的结果能够提高物种分布预测的准确性,阿蒙森海域南极磷虾的分布预测结果可以为磷虾保护、利用提供科学参考。
关键词:  南极磷虾  最大熵模型(maximum entropy modeling,MaxEnt)  预设规则的遗传算法(genetic algorithm for rule-set production,GARP)  阿蒙森海域
DOI:10.11693/hyhz20201100320
分类号:
基金项目:青岛海洋科学与技术试点国家实验室山东省专项经费,2022QNLM030002-1号;国家海洋局极地考察办公室项目(南极海域对气候变化的影响和响应),RFSOCC2020-2022-No.18号。
附件
PREDICTED DISTRIBUTION OF ANTARCTIC KRILL (EUPHAUSIA SUPERBA) IN THE AMUNDSEN SEA USING MAXENT AND GARP
LIU Lu-Lu, ZHAO Liang, LIN Shi-Ying, FENG Jian-Long
College of Marine and Environmental Sciences, Tianjin University of Science and Technology, Tianjin 300457, China
Abstract:
Antarctic krill Euphausia superba is a key species in the Southern Ocean ecosystem, and plays an important role in carbon sequestration in the Antarctic, which has gained more and more attention in recent years. The potential distribution of krill in the Amundsen Sea in the Pacific sector of the Southern Ocean was predicted and analyzed using MaxEnt and GARP niche models in this study. Occurrence points and environmental variables were used to simulate the distribution. The ROC (receiver operating characteristic) curve was used to evaluate the model's performance. Results reveal that the high suitability areas predicted by the MaxEnt model were more detailed, whereas those of the GARP model predicted a wider spread of high suitability areas. To overcome the uncertainty of a single model and obtain better outcomes, the prediction results of the two models were combined. The ensemble prediction's simulation accuracy was significantly enhanced[AUC (area under curve) is 0.946 and TSS(true skill statistic) is 0.78], and excellent prediction results were obtained. The high suitable areas of krill were concentrated in 65°S and 73°S, accounting for 6.2% of the entire area, and the moderate suitable areas accounted for 5.7% of the entire area. The three variables with the highest contribution of 81.3% in MaxEnt were ice, the minimum value of mean sea level pressure, and the velocity of eastward latitudinal flow. Compared with those of the MaxEnt model, the omission error of each variable in the GARP model was relatively even. Therefore, using ensemble prediction to estimate species distribution could enhance the accuracy. The prediction results of Antarctic krill distribution provide scientific data for krill conservation and utilization in the Amundsen Sea.
Key words:  Antarctic krill  MaxEnt (maximum entropy modeling)  GARP (genetic algorithm for rule-set production)  Amundsen Sea
版权所有 海洋与湖沼 Oceanologia et Limnlolgia Sinica Copyright©2008 All Rights Reserved
主管单位:中国科协技术协会 主办单位:中国海洋湖沼学会
地址:青岛市海军路88号  邮编:266400  电话:0532-82898753  E-mail:ols@qdio.ac.cn
技术支持:北京勤云科技发展有限公司