Machine Learning Solutions to Regional Surface Ocean δ18O‐Salinity Relationships for Paleoclimatic Reconstruction

IF 3.2 2区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Paleoceanography and Paleoclimatology Pub Date : 2023-08-21 DOI:10.1029/2023PA004612

N. K. Murray, A. R. Muñoz, J. L. Conroy

{"title":"Machine Learning Solutions to Regional Surface Ocean δ18O‐Salinity Relationships for Paleoclimatic Reconstruction","authors":"N. K. Murray, A. R. Muñoz, J. L. Conroy","doi":"10.1029/2023PA004612","DOIUrl":null,"url":null,"abstract":"Stable isotope‐based reconstructions of past ocean salinity and hydroclimate depend on accurate, regionally constrained relationships between the stable oxygen isotopic composition of seawater (δ18Osw) and salinity in the surface ocean. An increasing number of δ18Osw observations suggest greater spatial variability in this relationship than previously considered, highlighting the need to reassess these relationships on a global scale. Here, we use available, paired δ18Osw and salinity data (N = 11,119) to create global interpolations of each variable. We then use a self‐organizing map, a specialized form of machine learning, to define 19 regions with unique δ18Osw‐salinity relationships in the surface (<50 m) ocean. Inclusion of atmospheric moisture‐related variables and oceanic tracer data in additional self‐organizing map experiments indicates global surface δ18Osw‐salinity spatial patterns are strongly forced by the atmosphere, as the SOM spatial output is highly similar despite no overlapping input data. Our approach is a useful update to the previously delimited regions, and highlights the utility of neural network pattern extraction in spatiotemporally sparse data sets.","PeriodicalId":54239,"journal":{"name":"Paleoceanography and Paleoclimatology","volume":null,"pages":null},"PeriodicalIF":3.2000,"publicationDate":"2023-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Paleoceanography and Paleoclimatology","FirstCategoryId":"89","ListUrlMain":"https://doi.org/10.1029/2023PA004612","RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"GEOSCIENCES, MULTIDISCIPLINARY","Score":null,"Total":0}

引用次数: 0

Abstract

Stable isotope‐based reconstructions of past ocean salinity and hydroclimate depend on accurate, regionally constrained relationships between the stable oxygen isotopic composition of seawater (δ18Osw) and salinity in the surface ocean. An increasing number of δ18Osw observations suggest greater spatial variability in this relationship than previously considered, highlighting the need to reassess these relationships on a global scale. Here, we use available, paired δ18Osw and salinity data (N = 11,119) to create global interpolations of each variable. We then use a self‐organizing map, a specialized form of machine learning, to define 19 regions with unique δ18Osw‐salinity relationships in the surface (<50 m) ocean. Inclusion of atmospheric moisture‐related variables and oceanic tracer data in additional self‐organizing map experiments indicates global surface δ18Osw‐salinity spatial patterns are strongly forced by the atmosphere, as the SOM spatial output is highly similar despite no overlapping input data. Our approach is a useful update to the previously delimited regions, and highlights the utility of neural network pattern extraction in spatiotemporally sparse data sets.

查看原文本刊更多论文

用于古气候重建的区域表层海洋δ18O-盐度关系的机器学习解决方案

基于稳定同位素的过去海洋盐度和水文气候重建取决于海水的稳定氧同位素组成（δ18Osw）和表层海洋盐度之间的准确、区域约束关系。越来越多的δ18Osw观测表明，这种关系的空间变异性比以前考虑的更大，这突出了在全球范围内重新评估这些关系的必要性。在这里，我们使用可用的、成对的δ18Osw和盐度数据（N=11119）来创建每个变量的全局插值。然后，我们使用自组织地图（一种专门的机器学习形式）来定义表层（<50m）海洋中具有独特δ18Osw-盐度关系的19个区域。在额外的自组织地图实验中纳入大气湿度相关变量和海洋示踪剂数据表明，全球表面δ18Osw盐度空间模式受到大气的强烈影响，因为尽管没有重叠的输入数据，但SOM空间输出高度相似。我们的方法是对先前定界区域的有用更新，并强调了神经网络模式提取在时空稀疏数据集中的实用性。

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

Paleoceanography and Paleoclimatology Earth and Planetary Sciences-Atmospheric Science

CiteScore

6.20

自引率

11.40%

发文量

107

期刊介绍： Paleoceanography and Paleoclimatology (PALO) publishes papers dealing with records of past environments, biota and climate. Understanding of the Earth system as it was in the past requires the employment of a wide range of approaches including marine and lacustrine sedimentology and speleothems; ice sheet formation and flow; stable isotope, trace element, and organic geochemistry; paleontology and molecular paleontology; evolutionary processes; mineralization in organisms; understanding tree-ring formation; seismic stratigraphy; physical, chemical, and biological oceanography; geochemical, climate and earth system modeling, and many others. The scope of this journal is regional to global, rather than local, and includes studies of any geologic age (Precambrian to Quaternary, including modern analogs). Within this framework, papers on the following topics are to be included: chronology, stratigraphy (where relevant to correlation of paleoceanographic events), paleoreconstructions, paleoceanographic modeling, paleocirculation (deep, intermediate, and shallow), paleoclimatology (e.g., paleowinds and cryosphere history), global sediment and geochemical cycles, anoxia, sea level changes and effects, relations between biotic evolution and paleoceanography, biotic crises, paleobiology (e.g., ecology of “microfossils” used in paleoceanography), techniques and approaches in paleoceanographic inferences, and modern paleoceanographic analogs, and quantitative and integrative analysis of coupled ocean-atmosphere-biosphere processes. Paleoceanographic and Paleoclimate studies enable us to use the past in order to gain information on possible future climatic and biotic developments: the past is the key to the future, just as much and maybe more than the present is the key to the past.