Heron Island Satellite Imagery Classified Benthic Data and Coral Halo Analyses and Models

From "Seeing halos: Spatial and consumer-resource constraints to landscapes of fear"
Published in American Naturalist

Access this dataset and code on Zenodo (10.5281/zenodo.8335657)

Authors: 
Theresa W. Ong*[1], Lisa C. McManus[2], Vitor V. Vasconcelos[3,4], Luojun Yang[5], Chenyang Su[1]


[1]Department of Environmental Studies and Graduate Program in Ecology, Evolution, Environment and Society, Dartmouth College
[2]Hawaiʻi Institute of Marine Biology, University of Hawaiʻi at Manoa
[3]Computational Science Lab, Informatics Institute, University of Amsterdam, Netherlands
[4]Institute for Advanced Study, University of Amsterdam, Netherlands
[5]Department of Ecology and Evolutionary Biology, Princeton University

*Corresponding author: theresa.w.ong@dartmouth.edu, p:1-603-646-4008, f:1-603-646-1682,  6182 Steele Hall, Hanover, NH 03755, USA




This study examines the relationship between grazing halos, consumer-resource interactions, and the spatial distribution of sheltering habitat for prey exhibiting anti-predator preferences to graze near shelter using a combination of models and satellite data of coral reefs. 

This dataset includes satellite imagery data from Heron Island, Australia downloaded from Google Earth Pro in 2023-4, with image from Maxar Technologies dated 2016 clipped to the shallow lagoon layer from the Allen Coral Atlas shape file classified into benthic categories: corals, algae and sand using a combination of unsupervised machine learning spectral classification and manually identification of region to undergo further processing with segmented mean shift applied to better detect objects. We also include manual classifications for two panels used to test accuracy of automated classifications. 

AUTHOR CONTRIBUTIONS REDACTED


## Description of the data and file structure


1. coralclass.txt 

This is the classified Heron Island data. The data is in the form of a rectangle with
ncols         4422
nrows         2362
xllcorner     390070.51518494 (x lower left corner in UTM)
yllcorner     7404070.4008209 (y lower left corner in UTM)
cellsize      1.6612890475868 m (Rastered size of each classified pixel)
NODATA_value  -9999

 	"1" = "Coral" 
	"2" = "Sand"
	"3" = "Algae" 

Positions in matrix correspond to spatial coordinates in pixels of the above cellsize. The xll/yll provide georeferenced location data using Datum/Projection: WGS 1984 UTM Zone 56S


## Sharing/Access information

Data was derived from the following sources:
 - Allen Coral Atlas for shallow lagoon layer from its geomorphic map:

Kennedy, E.; Roelfsema, C.; Lyons, M.; Kovacs, E.; Borrego-Acevedo, R.; Roe, M.; Phinn, S.; Larsen, K.; Murray, N.; Yuwono, D., et al. (2020) Reef Cover: a coral reef classification to guide global habitat mapping from remote sensing. bioRxiv 2020, doi:10.1101/2020.09.10.292243

 - Coral halo sizes for patch reefs at Heron Island over time in Supplementary Information of:

Madin, E. M. P., Precoda, K., Roelfsema, C. M., and Suan, A. (2022). Global Conservation potential
in coral reef halos: Consistency over space, time, and ecosystems worldwide. The American
Naturalist, 200(6):857–871.

available for download on Zenodo (https://doi.org/10.5281/zenodo.6426257; Madin 2022).

** Note that this dataset is not reproduced here since it is already published but appears as "halodataHx.csv" in our R script.

 - Google Earth Pro imagery is available from Google Earth: 
https://earth.google.com/web/search/heron+island/@-23.4425153,151.95594246,-2.88745373a,17263.4191697d,35y,0.00000001h,0t,0r/data=CigiJgokCUgRwMjsgjpAEUcRwMjsgjrAGUu8YpO1_ENAIdkmKIj6WFHA

The imagery is dated 7/3/2016 for classifications and 6/2024, 7/3/2016, 10/29/2011,295
8/2/2006, 4/25/2001 for temporal analysis


Accessed with: Google Earth Pro

7.3.6.9345 (64-bit)
Build Date

Thursday, December 29, 2022 10:32:21 PM UTC
Renderer

OpenGL
Operating System

Mac OS X (10.15.7)
Graphics Driver

Intel(R) Iris(TM) Plus Graphics 655 (2.1 INTEL-14.7.28)
Maximum Texture Size

16384×16384
Available Video Memory

2047 MB
Server

kh.google.com


2. Tclassfications_A.csv

This file includes a matrix of 28X29 pixels in the upper half of the satellite image that was manually classified by lead author. Here, cells marked 1=algae and all other cells are marked 0.

3. Tclassfications_AL.csv

This file includes a matrix of 28X29 pixels in the lower half of the satellite image that was manually classified by lead author. Here, cells marked 1=algae and all other cells are marked 0.

4. Tclassfications_C.csv

This file includes a matrix of 28X29 pixels in the upper half of the satellite image that was manually classified by lead author. Here, cells marked 1=coral and all other cells are marked 0.

5. Tclassfications_CL.csv

This file includes a matrix of 28X29 pixels in the lower half of the satellite image that was manually classified by lead author. Here, cells marked 1=coral and all other cells are marked 0.

6. Tclassfications_S.csv

This file includes a matrix of 28X29 pixels in the upper half of the satellite image that was manually classified by lead author. Here, cells marked 1=sand and all other cells are marked 0.

7. Tclassfications_SL.csv

This file includes a matrix of 28X29 pixels in the lower half of the satellite image that was manually classified by lead author. Here, cells marked 1=sand and all other cells are marked 0.

----------------------------

The next set of files were manually classified by another individual who is a non-expert in corals, and their classifications compared to expert classifications to assess ambiguity for manual classifications.


8. AL_Uclassfications_A.csv

This file includes a matrix of 28X29 pixels in the upper half of the satellite image that was manually classified by a volunteer with non-expert knowledge of the coral system. Here, cells marked 1=algae and all other cells are marked 0.

9. AL_Lclassfications_A.csv

This file includes a matrix of 28X29 pixels in the lower half of the satellite image that was manually classified by a volunteer with non-expert knowledge of the coral system. Here, cells marked 1=algae and all other cells are marked 0.

10. AL_Uclassfications_C.csv

This file includes a matrix of 28X29 pixels in the upper half of the satellite image that was manually classified by a volunteer with non-expert knowledge of the coral system. Here, cells marked 1=coral and all other cells are marked 0.

11. AL_Lclassfications_C.csv

This file includes a matrix of 28X29 pixels in the lower half of the satellite image that was manually classified by a volunteer with non-expert knowledge of the coral system. Here, cells marked 1=coral and all other cells are marked 0.

12. AL_Uclassfications_S.csv

This file includes a matrix of 28X29 pixels in the upper half of the satellite image that was manually classified by a volunteer with non-expert knowledge of the coral system. Here, cells marked 1=sand and all other cells are marked 0.

13. AL_Lclassfications_S.csv

This file includes a matrix of 28X29 pixels in the lower half of the satellite image that was manually classified by a volunteer with non-expert knowledge of the coral system. Here, cells marked 1=sand and all other cells are marked 0.


14. halos_yes_no.csv

This file includes scoring of selected plots representing C= Clustered and D= Dispersed coral patch reefs in "Treatment" column, the replicate number in the "Rep" column, code to indicate unique plots in "ID" column, year of historical aerial imagery under "Year" column and the score for isolated halos where 1= present and 0= absent in the "Halos" column.
 

## Code/Software

Included are 2 notebooks to analyze data and produce figures:

1. Coralhalos_spatialdata_revised.Rmd

This is a R Markdown file used to analyze classified spatial data and produce Figs. (1, 3a-b, 5b, 6c, B2, B3) in manuscript. Also included is Coralhalos_spatialdata_revised.html, the knitted output file.

Markdown file was created with RStudio (Version 2023.06.0+421 (2023.06.0+421) and R (R  version 4.2.3 (2023-03-15) GUI 1.79 High Sierra build (8160))

File requires the following packages/libraries:

spatstat 3.0-2, sp 1.5-1, sf 1.0-14 for spatial analyses and plots
bbmle 1.0.25 for maximum likelihood estimation
caret 6.0-94 for accuracy tests
stringr 1.5.0, ggpubr 0.5.0, tidyverse  1.3.2 for data wrangling
ggplot2 3.4.4 for data visualization

2. Coralhalos_geo_CRmodels_revisedFINAL.nb

This is a Mathematica notebook for geometric model and consumer-resource model and all associated figures from manuscript (Figs. 2c-d, 4, 5a, C1). 

Created using Wolfram Mathematica Version 13.0

Optional: Check settings in Mathematica to enable parallel processing to speed up the analysis. 


All other figures or figure components in manuscript were created using Adobe Illustrator 2023 v. 27.4.0 and Adobe Photoshop 23.5.5.
