Developing an Optimal Spatial Predictive Model for Seabed Sand Content Using Machine Learning, Geostatistics, and Their Hybrid Methods

Li Li,Huang Huang,Siwabessy Siwabessy,Nichol Nichol

doi:10.3390/geosciences9040180

Abstract

Seabed sediment predictions at regional and national scales in Australia are mainly based on bathymetry-related variables due to the lack of backscatter-derived data. In this study, we applied random forests (RFs), hybrid methods of RF and geostatistics, and generalized boosted regression modelling (GBM), to seabed sand content point data and acoustic multibeam data and their derived variables, to develop an accurate model to predict seabed sand content at a local scale. We also addressed relevant issues with variable selection. It was found that: (1) backscatter-related variables are more important than bathymetry-related variables for sand predictive modelling; (2) the inclusion of highly correlated predictors can improve predictive accuracy; (3) the rank orders of averaged variable importance (AVI) and accuracy contribution change with input predictors for RF and are not necessarily matched; (4) a knowledge-informed AVI method (KIAVI2) is recommended for RF; (5) the hybrid methods and their averaging can significantly improve predictive accuracy and are recommended; (6) relationships between sand and predictors are non-linear; and (7) variable selection methods for GBM need further study. Accuracy-improved predictions of sand content are generated at high resolution, which provide important baseline information for environmental management and conservation.

Highlights

Seabed mapping and characterization utilize datasets that describe the physical form and composition of seabed features that, when integrated, contribute important baseline information to support evidence-based environmental management [1,2,3,4,5,6]
We tested the performance of relative variable influence (RVI) for the generalized boosted regression modelling (GBM) model, but we found no improvement in predictive accuracy when compared to the full model
The hybrid methods applied in this study significantly improved predictive accuracy in comparison with other methods including random forests (RFs) and GBM (Table 7)

Summary

Introduction

Seabed mapping and characterization utilize datasets that describe the physical form and composition of seabed features that, when integrated, contribute important baseline information to support evidence-based environmental management [1,2,3,4,5,6]. Predictions of seabed sediments have been generated [7,8,9,10] at regional and larger scales for the Australian margin (e.g., seabed sand content at 1000 m resolution [11] and 250 m resolution for the northwest region [12]). These predictions were based largely on bathymetry-derived variables (e.g., slope, relief) and other spatial measures (e.g., latitude, longitude) due to data availability at these scales [7,8,9,10]. In contrast to the limited availability of predictive variables at regional and larger scales, with the increased application of high-resolution acoustic multibeam technologies to seabed mapping, usually more predictive variables are available at Geosciences 2019, 9, 180; doi:10.3390/geosciences9040180 www.mdpi.com/journal/geosciences

Methods

Results

Discussion

Conclusion