View in EDS

A novel approach for improving the spatiotemporal distribution modeling of marine benthic species by coupling a new GIS procedure with machine learning.

Saved in:
Bibliographic Details
Title: A novel approach for improving the spatiotemporal distribution modeling of marine benthic species by coupling a new GIS procedure with machine learning.
Authors: Benavides Martínez, Iván. F.1,2 (AUTHOR) ibenavidesm@unal.edu.co, Rueda, Mario3 (AUTHOR), Ortíz Ferrin, Omar Olimpo4 (AUTHOR), Díaz-Ochoa, Javier A.5 (AUTHOR), Castillo-Vargasmachuca, Sergio6 (AUTHOR), Selvaraj, John Josephraj1,7 (AUTHOR)
Source: Deep-Sea Research Part I, Oceanographic Research Papers. Jan2024, Vol. 203, pN.PAG-N.PAG. 1p.
Subjects: Oceanographic maps, Whiteleg shrimp, Machine learning, Species distribution, Ocean bottom
Abstract: In this study we developed and validated a new integrative method for constructing and assessing Species Distribution Models (SDMs) for marine benthic species across spatial and temporal scales. This methodology synthesizes four dimensions-latitude, longitude, depth and time-into a unified predictive output, achieved by integrating a novel GIS tool, Bathymetric Projection (BP) with machine learning algorithms. BP can extract multi-dimensional data from remote sensing and satellite-derived datasets across the oceanic water column, including latitude, longitude, depth, and time. This data is then mapped onto the ocean floor using a reference bathymetric layer. The resulting seafloor layer is a predictive variable within the SDMs for benthic species. The model output quantifies the probability of species occurrence, or habitat suitability, as determined by environmental conditions specific to the ocean floor across the four dimensions. We validated this approach using two benthic species from the Colombian Pacific Ocean: the White shrimp (Litopenaeus occidentalis) and the Coliflor shrimp (Solenocera agassizii). The model's performance ranged from 87% to 98%, as indicated by 30-fold cross-validation on test datasets. Depth ranges and monthly seasonal patterns were accurately predicted when compared to independent literature data, confirming the model's capability to generate precise spatiotemporal distributions for benthic species. This research underscores the necessity of incorporating depth and time into SDMs, given their critical interactions with latitude and longitude in determining species distribution. The methodology offers a valuable tool for informing decision-making processes in fisheries and environmental management by enhancing the predictive accuracy of spatiotemporal distributions for benthic species, thereby facilitating more holistic interpretations of habitat suitability. • Bathymetric Projection generate spatiotemporal seafloor predictors to improve the realism of benthic species distribution models. • It was validated with two benthic species resulting in high accuracy, predicting both depth ranges and monthly seasonal patterns. • We emphasize the importance of including depth and time in benthic species distribution models to improve management decisions. [ABSTRACT FROM AUTHOR]
Copyright of Deep-Sea Research Part I, Oceanographic Research Papers is the property of Pergamon Press - An Imprint of Elsevier Science and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
Description
Abstract:In this study we developed and validated a new integrative method for constructing and assessing Species Distribution Models (SDMs) for marine benthic species across spatial and temporal scales. This methodology synthesizes four dimensions-latitude, longitude, depth and time-into a unified predictive output, achieved by integrating a novel GIS tool, Bathymetric Projection (BP) with machine learning algorithms. BP can extract multi-dimensional data from remote sensing and satellite-derived datasets across the oceanic water column, including latitude, longitude, depth, and time. This data is then mapped onto the ocean floor using a reference bathymetric layer. The resulting seafloor layer is a predictive variable within the SDMs for benthic species. The model output quantifies the probability of species occurrence, or habitat suitability, as determined by environmental conditions specific to the ocean floor across the four dimensions. We validated this approach using two benthic species from the Colombian Pacific Ocean: the White shrimp (Litopenaeus occidentalis) and the Coliflor shrimp (Solenocera agassizii). The model's performance ranged from 87% to 98%, as indicated by 30-fold cross-validation on test datasets. Depth ranges and monthly seasonal patterns were accurately predicted when compared to independent literature data, confirming the model's capability to generate precise spatiotemporal distributions for benthic species. This research underscores the necessity of incorporating depth and time into SDMs, given their critical interactions with latitude and longitude in determining species distribution. The methodology offers a valuable tool for informing decision-making processes in fisheries and environmental management by enhancing the predictive accuracy of spatiotemporal distributions for benthic species, thereby facilitating more holistic interpretations of habitat suitability. • Bathymetric Projection generate spatiotemporal seafloor predictors to improve the realism of benthic species distribution models. • It was validated with two benthic species resulting in high accuracy, predicting both depth ranges and monthly seasonal patterns. • We emphasize the importance of including depth and time in benthic species distribution models to improve management decisions. [ABSTRACT FROM AUTHOR]
ISSN:09670637
DOI:10.1016/j.dsr.2023.104222