Comparison of deterministic, probabilistic and machine learning-based methods for bathymetric surface modeling

The evolution of water resources management is closely related to technological and scientific advances in the hydrographic field. For this, it is necessary to map the submerged bottom, especially with equipment such as echo sounders offering greater data collection precision. However, the resulting...

Full description

Saved in:
Bibliographic Details
Published in:Modeling earth systems and environment 2025-02, Vol.11 (1), p.6
Main Authors: de Andrade, Laura Coelho, Silva, Arthur Amaral e, Veloso, Gustavo Viera, Filho, Elpídio Inácio Fernandes, Ferreira, Italo Oliveira
Format: Article
Language:English
Subjects:
Bathymetric surveys
Chemistry and Earth Sciences
Computer Science
Data collection
Depth
Earth and Environmental Science
Earth Sciences
Earth System Sciences
Echo sounding
Ecosystems
Environment
Learning algorithms
Machine learning
Math. Appl. in Environmental Science
Mathematical Applications in the Physical Sciences
Original Article
Parameters
Physics
Repeaters
Statistics for Engineering
Water resources
Water resources management
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The evolution of water resources management is closely related to technological and scientific advances in the hydrographic field. For this, it is necessary to map the submerged bottom, especially with equipment such as echo sounders offering greater data collection precision. However, the resulting product of a bathymetric survey consists of a sample of spaced points without a continuous representation of the study area. Thus, interpolators such as Universal Kriging and IDW are commonly used to predict depth in unsampled locations. In this sense, this work aimed to evaluate different interpolators, such as Machine Learning (ML), IDW, and Fixed Rank Kriging (FRK), in the creation of Digital Depth Models (DDMs). From the results, it can be inferred that the best interpolator was the one that presented a better value for the performance parameters (RMSE, MAE, R 2 , and LCCC), as well as a reliable DDM with the one used as a reference, without changing its form. The maximum and minimum depth values are significant. In addition, despite not having the best values of performance parameters, FRK proved to be a potential interpolator for representing bathymetric surfaces. Graphical abstract
ISSN:2363-6203
2363-6211
DOI:10.1007/s40808-024-02189-8