Improving prediction for potential spawning areas from a two-step perspective: A comparison of multi-model approaches for sparse egg distribution

IF 2.1 4区 地球科学 Q2 MARINE & FRESHWATER BIOLOGY
Zunlei Liu , Yan Jin , Linlin Yang , Xingwei Yuan , Liping Yan , Yi Zhang , Hui Zhang , Min Xu , Xiaojing Song , Jianhua Tang , Yongdong Zhou , Fen Hu , Jiahua Cheng
{"title":"Improving prediction for potential spawning areas from a two-step perspective: A comparison of multi-model approaches for sparse egg distribution","authors":"Zunlei Liu ,&nbsp;Yan Jin ,&nbsp;Linlin Yang ,&nbsp;Xingwei Yuan ,&nbsp;Liping Yan ,&nbsp;Yi Zhang ,&nbsp;Hui Zhang ,&nbsp;Min Xu ,&nbsp;Xiaojing Song ,&nbsp;Jianhua Tang ,&nbsp;Yongdong Zhou ,&nbsp;Fen Hu ,&nbsp;Jiahua Cheng","doi":"10.1016/j.seares.2023.102460","DOIUrl":null,"url":null,"abstract":"<div><p>Recently, much work has been put into improving species distribution models, especially for systematic conservation planning for important ecosystems. Understanding fish spawning sites requires a thorough examination of ichthyoplankton. These examinations usually produce sparse counts contaminated by inaccurate detection, making it impossible to directly infer the abundance or occurrence from observational data, which could lead to inaccurate model predictions. A flexible modeling framework for estimating and inference about the abundance of eggs with ensemble models that include the presence/absence and abundance components is described in this study. The generalized linear model, generalized additive model, integrated nested Laplace approximations, and random forest habitat modeling approaches were compared within this framework to those currently being used for fish conservation planning at regional scales. Additionally, the distribution of suitable habitats for small yellow croaker (<em>Larimichthys polyactis</em>) spawning stocks were mapped based on the ensemble model. Furthermore, the promotion ability of ensemble models with different weighting methods was evaluated. The outcomes demonstrated that machine learning algorithms performed better than statistical models, and the geometric weighted ensemble model further increased prediction accuracy. However, there was no significant difference compared to the optimal individual model (<em>p</em> &gt; 0.05). The predicted distributions of the four models can be divided into two groups. The central sea of Jiangsu was recognized as the most suitable area by the GAM with a fixed effect for each year and INLA models, while the GLM was similar to the RF with spatial effect (RF-LL) and demonstrated Haizhou Bay as the most suitable area. The ensemble model discovered several areas of highly suitable habitat that dominated areas in the two groups of models and revealed many finer-scale patterns in the egg distribution. According to the ensemble model, although 5.37% of the area could be suitable habitat, only 0.12% was highly suitable. It is suggested that examining small yellow croaker spawning aggregation areas would benefit from using an ensemble modeling approach to identify and prioritize conservation areas.</p></div>","PeriodicalId":50056,"journal":{"name":"Journal of Sea Research","volume":null,"pages":null},"PeriodicalIF":2.1000,"publicationDate":"2023-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S1385110123001296/pdfft?md5=591658e7fc800afc054c2e6fa922513c&pid=1-s2.0-S1385110123001296-main.pdf","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Sea Research","FirstCategoryId":"89","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1385110123001296","RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"MARINE & FRESHWATER BIOLOGY","Score":null,"Total":0}
引用次数: 0

Abstract

Recently, much work has been put into improving species distribution models, especially for systematic conservation planning for important ecosystems. Understanding fish spawning sites requires a thorough examination of ichthyoplankton. These examinations usually produce sparse counts contaminated by inaccurate detection, making it impossible to directly infer the abundance or occurrence from observational data, which could lead to inaccurate model predictions. A flexible modeling framework for estimating and inference about the abundance of eggs with ensemble models that include the presence/absence and abundance components is described in this study. The generalized linear model, generalized additive model, integrated nested Laplace approximations, and random forest habitat modeling approaches were compared within this framework to those currently being used for fish conservation planning at regional scales. Additionally, the distribution of suitable habitats for small yellow croaker (Larimichthys polyactis) spawning stocks were mapped based on the ensemble model. Furthermore, the promotion ability of ensemble models with different weighting methods was evaluated. The outcomes demonstrated that machine learning algorithms performed better than statistical models, and the geometric weighted ensemble model further increased prediction accuracy. However, there was no significant difference compared to the optimal individual model (p > 0.05). The predicted distributions of the four models can be divided into two groups. The central sea of Jiangsu was recognized as the most suitable area by the GAM with a fixed effect for each year and INLA models, while the GLM was similar to the RF with spatial effect (RF-LL) and demonstrated Haizhou Bay as the most suitable area. The ensemble model discovered several areas of highly suitable habitat that dominated areas in the two groups of models and revealed many finer-scale patterns in the egg distribution. According to the ensemble model, although 5.37% of the area could be suitable habitat, only 0.12% was highly suitable. It is suggested that examining small yellow croaker spawning aggregation areas would benefit from using an ensemble modeling approach to identify and prioritize conservation areas.

从两步的角度改进对潜在产卵区域的预测:稀疏卵分布的多模型方法的比较
近年来,人们对物种分布模型的改进,特别是对重要生态系统的系统保护规划进行了大量的研究。了解鱼类的产卵地点需要对浮游鱼进行彻底的检查。这些检查通常产生的稀疏计数受到不准确检测的污染,使得不可能直接从观测数据推断丰度或发生,这可能导致不准确的模型预测。本研究描述了一种灵活的建模框架,用于使用包括存在/不存在和丰度成分的集成模型估计和推断卵子的丰度。在此框架内,将广义线性模型、广义加性模型、集成嵌套拉普拉斯近似和随机森林栖息地建模方法与目前用于区域尺度鱼类保护规划的方法进行了比较。此外,利用集合模型绘制了小黄鱼产卵群适宜生境分布图。此外,还对不同加权方法对集成模型的提升能力进行了评价。结果表明,机器学习算法优于统计模型,几何加权集成模型进一步提高了预测精度。但与最优个体模型相比,差异无统计学意义(p > 0.05)。四种模型的预测分布可分为两组。具有固定效应的GAM模型和INLA模型均将江苏中部海域确定为最适宜区域,而具有空间效应的GLM模型与RF模型相似,均将海州湾确定为最适宜区域。集合模型发现了在两组模型中占主导地位的几个高度适宜栖息地区域,并揭示了许多更精细尺度的卵分布模式。综上模型显示,虽然5.37%的面积可为适宜生境,但高度适宜生境的面积仅为0.12%。建议使用集成建模方法来确定和优先考虑小黄鱼产卵聚集区域,这将有利于研究小黄鱼产卵聚集区域。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Journal of Sea Research
Journal of Sea Research 地学-海洋学
CiteScore
3.20
自引率
5.00%
发文量
86
审稿时长
6-12 weeks
期刊介绍: The Journal of Sea Research is an international and multidisciplinary periodical on marine research, with an emphasis on the functioning of marine ecosystems in coastal and shelf seas, including intertidal, estuarine and brackish environments. As several subdisciplines add to this aim, manuscripts are welcome from the fields of marine biology, marine chemistry, marine sedimentology and physical oceanography, provided they add to the understanding of ecosystem processes.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信