Skip to content

Latest commit

 

History

History
128 lines (103 loc) · 4.42 KB

README.md

File metadata and controls

128 lines (103 loc) · 4.42 KB
Raw

AI for Shrubland Identification and Mapping

This repository contains the files required to generate the chapter “AI for Shrubland Identification and Mapping”, included in the book “AI for Earth Sciences”.

Chapter Introduction

This chapter walks through a procedure for predicting the prevalence of “shrubland” (defined here as low-statured vegetation between 1 and 5 meters in height) across a diverse region in New York State, patterned off the process used in Mahoney, Johnson, and Beier (2022). Due to the impacts of climate change and human land use patterns, these shrublands are becoming an increasingly important land cover type in the region, often representing an entirely novel ecosystem type. As a result of this novelty, these shrublands and the roles they play in the larger landscape (for instance, as habitats and as components of biogeochemical cycles) are poorly understood. Even identifying shrublands using remote sensing data, a potential way to monitor their development over time, is difficult given the relative rarity of shrublands in this region and their similar appearance to forest lands and wetlands in satellite imagery.

The chapter introduces step-by-step how to fit a feedforward neural network using the Keras module in the popular TensorFlow package and a subset of the data from Mahoney, Johnson, and Beier (2022). Due to the rarity of shrubland in this region of New York, the chapter focuses on the adjustments necessary when building models from data with imbalanced classes, and on how to interpret model performance metrics when fitting classification models for specific purposes. Chapter exercises prompt learners to investigate how different priorities for a model might result in notably different performance measures.

Repository Structure

- .
|
|-- figures/
  |
  |- diagram.png
  |- diagram.svg
  |- inset_map.png
  |- model_predictions.png
|
|-- filters/
  |
  |- author-info-blocks.lua
  |- scholarly-metadata.lua
|
|-- renv/
|
|-- scripts/
  |
  |- make_inset.R
  |- make_map.R
|
|-- .jupyter-cache/
  |
  |- *
|
|-- earth-science-ai-shrubland.qmd
|-- earth-science-ai-shrubland.Rproj
|-- paper.bib
|-- _quarto.yml
|-- README.qmd
|-- README.md
|-- renv.lock
|-- requirements.txt
|-- .gitignore

The diagrams figures/diagram.svg and figures/diagram.png was created by hand in InkScape. The other two figures are generated from the files in scripts/.

All files in filters/ are from the pandoc/lua-filters repository. These files generate the author information at the top of the document, and are run automatically as part of the quarto render process.

renv/ contains files related to the R virtual environment (“renv”) included with this project.

make_inset.R in scripts/ generates the figure at figures/inset_map.png. make_map.R generates model_predictions.png.

.jupyter-cache/ contains binary files used to “cache” outputs from the Python code in the manuscript, speeding up execution.

The manuscript itself, and all Python code used in the process of generating the manuscript, lives at earth-science-ai-shrubland.qmd. This is a Quarto document, which generates all of the relevant outputs (.tex, .ipynb, .pdf) through the command quarto render.

earth-science-ai-shrubland.Rproj is an RStudio project file, which makes it easy to open this directory in the RStudio IDE.

paper.bib is a BibTeX file used to generate the citations at the end of the manuscript.

_quarto.yml is a Quarto project file, which controls the rendering process for this directory.

README.qmd generates this file, which is README.md.

renv.lock lists all R packages (with versions) used in this project. requirements.txt lists all Python packages (with versions) used in this project.

.gitignore controls which files get included in this git repository.

.Rprofile enables renv when working in this directory.

References

Mahoney, Michael J, Lucas K Johnson, and Colin M Beier. 2022. “Classification and Mapping of Low-Statured ’Shrubland’ Cover Types in Post-Agricultural Landscapes of the US Northeast.” arXiv. https://doi.org/10.48550/ARXIV.2205.05047.