Skip to content

Latest commit

 

History

History
43 lines (30 loc) · 1.87 KB

README.md

File metadata and controls

43 lines (30 loc) · 1.87 KB

Project forked from John Burley and Chris Bradley's PaddocksTS project

Main changes thus far:

  • Bash files and Pre-segment code adapted to pull Sentinel-2 data from DEA datasets on AWS rather than on NCI
  • 01_pre-segment.py now can take either a geojson file or lat/lon/buffer
  • If using a geojson file, update the filepath to your geojson folder in the _main.sh file before running it and change the STUB name accordingly

Workflow to download new sentinel datasets:

  1. Open one of the _main.sh files and save it with the name of your new dataset

  2. Create a new geojson file with the area you want to download

  3. Name the geojson by the area being downloaded.

  4. Make a copy of one of the _main.sh files and save it with the name of your new dataset

  5. In the _main.sh files, change the STUB name to be the same as the filename and the geojson area

  6. Include a timestamp in the name, for example: geojson: ACT-NE-v1.geojson

One year: * stub: ACT-NE-v1_2021 * main.sh file: ACT-NE-v1_2021_main.s

Multiple years:
* stub: ACT-NE-v1_2020-2024
* main.sh file: ACT-NE-v1_2020-2_main.sh

PaddockTS notebooks have been adapted for ACT Veg mapping using various ACT datasets.

  • New Jupyter notebooks for processing this data are now in the act-environment repo (at least until I have time to organise them better)

Note for JupyterLab setup on GADI/ARE at NCI

When launching a new JupyterLab the following must be set:

  • Walltime: Whatever you are likely to use
  • Queue: expressbw
  • Compute Size: cpus=28 mem=252G <- this can be smaller if you aren't loading huge time-series pkl files
  • Project: xe2
  • Storage: gdata/xe2+scratch/xe2+gdata/v10+gdata/ub8
  • Module directories: /g/data/v10/public/modules/modulefiles/
  • Modules: ffmpeg/4.3.1 tensorflow/2.15.0 gdal/3.6.4