The University of Adelaide
Browse
IMAGE
2007_reprojected.tif (5.1 MB)
IMAGE
2008_reprojected.tif (5.1 MB)
IMAGE
2009_reprojected.tif (5.1 MB)
IMAGE
2010_reprojected.tif (5.1 MB)
IMAGE
2011_reprojected.tif (5.1 MB)
IMAGE
2013_reprojected.tif (5.1 MB)
IMAGE
2014_reprojected.tif (5.1 MB)
IMAGE
2015_reprojected.tif (5.1 MB)
IMAGE
2016_reprojected.tif (5.09 MB)
IMAGE
2017_reprojected.tif (5.09 MB)
IMAGE
1988_reprojected.tif (5.09 MB)
IMAGE
1993_reprojected.tif (5.1 MB)
IMAGE
1994_reprojected.tif (5.1 MB)
IMAGE
1995_reprojected.tif (5.1 MB)
IMAGE
1996_reprojected.tif (5.1 MB)
IMAGE
1997_reprojected.tif (5.09 MB)
IMAGE
1999_reprojected.tif (5.1 MB)
IMAGE
2000_reprojected.tif (5.1 MB)
IMAGE
2001_reprojected.tif (5.1 MB)
IMAGE
2003_reprojected.tif (5.1 MB)
1/0
29 files

Benthic cover maps associated with paper "Landsat historical records reveal large-scale dynamics and enduring recovery of seagrasses in an impacted seascape", in Science of the Total Environment

journal contribution
posted on 2022-09-07, 01:46 authored by Milena Fernandes, Andrew HennessyAndrew Hennessy, Wallace LawWallace Law, Robert Daly, Samuel GaylardSamuel Gaylard, Megan LewisMegan Lewis, Kenneth ClarkeKenneth Clarke

This dataset includes benthic exposure annual maps as well as decadal benthic cover trajectory and persistence maps derived from summer Landsat imagery between 1988 and 2018 over 501 km2 off the coast of Adelaide, South Australia.


Image processing included spectral subsetting, density slicing (Landsat 5 and Landsat 7) or Support Vector Machine (SVM) supervised classification (Landsat 8) and temporal majority filtering. Training and accuracy assessment of the classification were based on video transect data.


The benthic exposure maps were classified into seagrass and sand. The trajectory maps identified an increase or decrease in seagrass cover for each pixel every 10 years, no change, or too variable to detect trends in time. The persistence maps showed the most frequent benthic cover for each pixel within each decade.


The analysis of annual changes highlighted major temporal shifts in the system, while the analysis of decadal products allowed interpretation of long-term spatial dynamics.


The associated paper published in Science of the Total Environment investigated the relationship between these temporal/spatial changes in benthic cover and land-based inputs, including breakpoints and timelags in system trajectory, and the identification of vulnerable seagrass meadows.


Annual benthic exposure file names use the following convention: "YYYY_reprojected.tif", where YYYY = the year in which the summer covered by that map started (e.g., for mapping covering the summer of 1988-1989 the year is recorded as '1988')

Data values for annual benthic exposure:

* Reef = NoData

* Seagrass = 1

* Sand = 0


Benthic cover trajectory file names use the following convention: "traj_XX-YY_reprojected.tif" where XX = the year of the summer at the start of the epoch, and YY = the year of the summer at the end of the epoch.

Data values for benthic cover trajectory:

* No clear trend = -1

* Sand - no change = 0

* Seagrass - no change = 1

* Seagrass - loss = 2

* Seagrass - gain = 3

* Reef = NoData


Benthic cover persistence file names use the following convention: "pers_XX-YY_reprojected.tif" where XX = the year of the summer at the start of the epoch, and YY = the year of the summer at the end of the epoch.

Data values for benthic cover persistence:

* Seagrass = 50

* Sand = 200

* Equal persistence (seagrass and sand) = 65,535

* Reef = NoData


Funding

Funded by SA Water Corporation

History