# Satellite imagery pre-processing and feature extraction for the mapping of coastal ecosystems using Google Earth Engine: A workflow for practitioners

## TL;DR

This paper provides a step-by-step workflow for using Google Earth Engine to preprocess and extract features from satellite images for mapping coastal ecosystems.

## Contribution

The paper introduces a general, adaptable workflow for satellite image pre-processing and feature extraction in Google Earth Engine tailored for coastal ecosystem mapping.

## Key findings

- The workflow produces processed satellite images suitable for machine learning classification of coastal ecosystems.
- The method empowers local managers, especially in low-resource settings, to monitor their coastal areas effectively.

## Abstract

The use of Google Earth Engine (GEE) is increasingly common in geospatial analysis of satellite images for various environmental management purposes due to its easy accessibility and capabilities to support complex pre-processing and mining of geographic data. In the context of coastal management, GEE provides opportunities for cost-efficient mapping of coastal habitats and their ecosystem service potentials. Understanding the extent of coastal habitats and the spatial and temporal variabilities of their ecosystem services can be useful for management and intervention purposes. GEE is well-suited for this due to its user-friendliness, particularly for non-experts of programming languages, such as area managers and other practitioners. However, there is no specific methodological guideline for the pre-processing and feature extraction of satellite images in GEE that can be readily adopted by these practitioners. This study develops general methodological steps to perform those processes that can be adapted to different management needs. Highlights of this study:•Steps detailed in this method paper will produce processed satellite images readily applicable for machine learning to classify coastal ecosystems.•The development of this adaptable workflow can benefit and empower local area managers, particularly in low-resource settings, to conduct monitoring of their area.

Steps detailed in this method paper will produce processed satellite images readily applicable for machine learning to classify coastal ecosystems.

The development of this adaptable workflow can benefit and empower local area managers, particularly in low-resource settings, to conduct monitoring of their area.

Image, graphical abstract