Machine-learning based reconstructions of primary and secondary climate variables from North American and European fossil pollen data
Abstract We test several quantitative algorithms as palaeoclimate reconstruction tools for North American and European fossil pollen data, using both classical methods and newer machine-learning approaches based on regression tree ensembles and artificial neural networks. We focus on the reconstruct...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2019-11-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-019-52293-4 |
| _version_ | 1818345876446248960 |
|---|---|
| author | J. Sakari Salonen Mikko Korpela John W. Williams Miska Luoto |
| author_facet | J. Sakari Salonen Mikko Korpela John W. Williams Miska Luoto |
| author_sort | J. Sakari Salonen |
| collection | DOAJ |
| description | Abstract We test several quantitative algorithms as palaeoclimate reconstruction tools for North American and European fossil pollen data, using both classical methods and newer machine-learning approaches based on regression tree ensembles and artificial neural networks. We focus on the reconstruction of secondary climate variables (here, January temperature and annual water balance), as their comparatively small ecological influence compared to the primary variable (July temperature) presents special challenges to palaeo-reconstructions. We test the pollen–climate models using a novel and comprehensive cross-validation approach, running a series of h-block cross-validations using h values of 100–1500 km. Our study illustrates major benefits of this variable h-block cross-validation scheme, as the effect of spatial autocorrelation is minimized, while the cross-validations with increasing h values can reveal instabilities in the calibration model and approximate challenges faced in palaeo-reconstructions with poor modern analogues. We achieve well-performing calibration models for both primary and secondary climate variables, with boosted regression trees providing the overall most robust performance, while the palaeoclimate reconstructions from fossil datasets show major independent features for the primary and secondary variables. Our results suggest that with careful variable selection and consideration of ecological processes, robust reconstruction of both primary and secondary climate variables is possible. |
| first_indexed | 2024-12-13T17:09:20Z |
| format | Article |
| id | doaj.art-b8548e10953e4fd095bb0a622c57d84b |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-12-13T17:09:20Z |
| publishDate | 2019-11-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-b8548e10953e4fd095bb0a622c57d84b2022-12-21T23:37:37ZengNature PortfolioScientific Reports2045-23222019-11-019111310.1038/s41598-019-52293-4Machine-learning based reconstructions of primary and secondary climate variables from North American and European fossil pollen dataJ. Sakari Salonen0Mikko Korpela1John W. Williams2Miska Luoto3Department of Geosciences and Geography, University of HelsinkiDepartment of Geosciences and Geography, University of HelsinkiDepartment of Geography and Center for Climatic Research, University of Wisconsin–MadisonDepartment of Geosciences and Geography, University of HelsinkiAbstract We test several quantitative algorithms as palaeoclimate reconstruction tools for North American and European fossil pollen data, using both classical methods and newer machine-learning approaches based on regression tree ensembles and artificial neural networks. We focus on the reconstruction of secondary climate variables (here, January temperature and annual water balance), as their comparatively small ecological influence compared to the primary variable (July temperature) presents special challenges to palaeo-reconstructions. We test the pollen–climate models using a novel and comprehensive cross-validation approach, running a series of h-block cross-validations using h values of 100–1500 km. Our study illustrates major benefits of this variable h-block cross-validation scheme, as the effect of spatial autocorrelation is minimized, while the cross-validations with increasing h values can reveal instabilities in the calibration model and approximate challenges faced in palaeo-reconstructions with poor modern analogues. We achieve well-performing calibration models for both primary and secondary climate variables, with boosted regression trees providing the overall most robust performance, while the palaeoclimate reconstructions from fossil datasets show major independent features for the primary and secondary variables. Our results suggest that with careful variable selection and consideration of ecological processes, robust reconstruction of both primary and secondary climate variables is possible.https://doi.org/10.1038/s41598-019-52293-4 |
| spellingShingle | J. Sakari Salonen Mikko Korpela John W. Williams Miska Luoto Machine-learning based reconstructions of primary and secondary climate variables from North American and European fossil pollen data Scientific Reports |
| title | Machine-learning based reconstructions of primary and secondary climate variables from North American and European fossil pollen data |
| title_full | Machine-learning based reconstructions of primary and secondary climate variables from North American and European fossil pollen data |
| title_fullStr | Machine-learning based reconstructions of primary and secondary climate variables from North American and European fossil pollen data |
| title_full_unstemmed | Machine-learning based reconstructions of primary and secondary climate variables from North American and European fossil pollen data |
| title_short | Machine-learning based reconstructions of primary and secondary climate variables from North American and European fossil pollen data |
| title_sort | machine learning based reconstructions of primary and secondary climate variables from north american and european fossil pollen data |
| url | https://doi.org/10.1038/s41598-019-52293-4 |
| work_keys_str_mv | AT jsakarisalonen machinelearningbasedreconstructionsofprimaryandsecondaryclimatevariablesfromnorthamericanandeuropeanfossilpollendata AT mikkokorpela machinelearningbasedreconstructionsofprimaryandsecondaryclimatevariablesfromnorthamericanandeuropeanfossilpollendata AT johnwwilliams machinelearningbasedreconstructionsofprimaryandsecondaryclimatevariablesfromnorthamericanandeuropeanfossilpollendata AT miskaluoto machinelearningbasedreconstructionsofprimaryandsecondaryclimatevariablesfromnorthamericanandeuropeanfossilpollendata |