Data and codes for vegetation-based fire risk forecast
Data files
Jan 23, 2025 version files 9.95 GB
-
Data_and_codes.zip
9.95 GB
-
README.md
28.39 KB
Abstract
This dataset intends to provide the clarity for the readers on using vegetation dynamics in a fire risk forecast framework. The dataset relates to Vegetation Optical Depth (a merged output from three satellite retreived products), Normalized Difference Vegetation Index (publicly accessible data product from Terra - Moderate Resolution Imaging Spectroradiometer as MOD13C1 Version6.1), and Forest Fire Danger Index (purchased from the Australian Bureau of Meteorology). Additional data on Koppen climate classification (publicly accessible on the Australian Bureau of Meteorology, (BoM, http://www.bom.gov.au/)) are provided to demonstrate the application aspects of the study. The study focuses on a fire risk forecast framework and the major processing steps of the given datasets are also provided.
README: Dataset title: Data and codes for vegetation-based fire risk forecast
Access on Dryad (DOI link: https://doi.org/10.5061/dryad.0gb5mkmbv)
Folder: 1. Data
This folder contains the raw data: Vegetation Optical Depth (VOD), Normalized Difference Vegetation Index (NDVI), Forest Fire Danger Index (FFDI), and Koppen climate classification. Below is the summary of the folder/file structure and a detailed description of their contents.
Folder & file structure summary:
1a. *VOD* > *VOD_netcdf* > 2002 > daily .nc files for the year (**vod_10km_20020101.nc** to **vod_10km_20020131.nc**)
> 2003 > daily .nc files for the year (**vod_10km_20030101.nc** to **vod_10km_20030131.nc**)
> 2004 > daily .nc files for the year (**vod_10km_20040101.nc** to **vod_10km_20040131.nc**)
> 2005 > daily .nc files for the year (**vod_10km_20050101.nc** to **vod_10km_20050131.nc**)
> .... > ....
> .... > ....
> 2019 > daily .nc files for the year (**vod_10km_20190101.nc** to **vod_10km_20190131.nc**)
1b. *NDVI* > *Sample_NDVI* > 16-day .hdf files for 16-day NDVI acquisition days in 2002
(**MOD13C1.A2002193.061.2020077195453.hdf** to **MOD13C1.A2002353.061.2020084110532.hdf**)
1c. *FFDI* > *Sample_FFDI* > daily .nc files for 16 days in 2002 from 2002.06.27 to 2002.07.12
(**FFDI20020627.nc** to **FFDI20020712.nc**)
1d. *Koppen climate classification* > *AUS_2021_AUST_SHP_GDA94*
> *STE_2021_AUST_SHP_GDA94*
> *kpngrp*
Contents in these folders are described below.
Descriptions:
1a. VOD (Vegetation Optical Depth) \ Data
- Source: derived by merging the Land Parameter Retrieval Model-based 0.1° VOD (Liu et al., 2011; Liu et al., 2015; Meesters et al., 2005; Owe et al., 2008; Owe et al., 2001; Santi, 2010) (from three satellites with passive microwave instruments, i.e., the Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) aboard the Aqua satellite, the microwave imager from the Tropical Rainfall Measuring Mission (TRMM), and the Advanced Microwave Scanning Radiometer 2 (AMSR2) onboard the Global Change Observation Mission-Water (GCOM-W1) satellite. [These details are referenced in the manuscript]
- Resolution: 0.1 degrees, daily
- Spatial extent: Australia-mainland [North bounding latitude 10.0°S, South bounding latitude 40.0°S, West bounding longitude 110.0°E, East bounding longitude 160.0°E]
- Temporal extent: 01 July 2002 to 31 December 2019
- File arrangement in subfolders: VOD_netcdf > 18 subfolders named by year from 2002 to 2019 > each year's subfolder contains daily files per year, i.e. no. of files per each year's subfolder = no. of days of that year.
- File count: 366 files per year × 4 leap years + 365 files per year × 14 non-leap years = 6574 files
- File format: Network Common Data Form (NetCDF) >> file extension .nc
- File naming convention: AAA_BBCC_YYYYMMDD.nc (e.g. vod_10km_20020101.nc) AAA: 'vod' stands for data product name 'vegetation optical depth' BB: '10' which is the numerical value of the spatial resolution denoted in 2 digits CC: 'km' which stands for 'kilometers', the unit of the spatial resolution\ YYYY: year of the acquisition date in 4 digits MM: month of the acquisition date in 2 digits DD: calendar day of the month of the acquisition date in 2 digits .nc: standard file extension for the NetCDF file
- Variable name: 'vod_daily'
- Units: none
- Datatype: double
- Spatial data matrix (2D) dimensions: Y dimension = 350 rows, X dimension = 500 columns
- Remarks: VOD data represent vegetation moisture content and density
1b. NDVI (Normalized Difference Vegetation Index) \ Data
- Source: Terra MODIS (Moderate Resolution Imaging Spectroradiometer) (MOD13C1-v061 product)
- Resolution: 0.05 degrees, 16-day composites
- Spatial extent: Global (study used data for Australia-mainland) [North bounding latitude 10.025°S, South bounding latitude 39.975°S, West bounding longitude 110.025°E, East bounding longitude 159.975°E]
- Temporal extent: 12 July 2002 to 19 December 2019
- Access: Publicly available at https://lpdaac.usgs.gov/products/mod13c1v061/
- Note: This folder provides a sample of raw NDVI data used in the study from 2002.07.12 to 2002.12.19
- File arrangement in subfolders: Sample_NDVI > 11 data files from 2002.07.12 to 2002.12.19 (each file is a 16-day composite of NDVI).
- File count: 11 files
- Format: Hierarchical Data Format (HDF) >> file extension .hdf
- File naming convention: AAABBCC.XYYYYDDD.EEE.yyyydddtttttt.hdf (e.g. MOD13C1.A2002193.061.2020077195453.hdf) AAA: 'MOD' indicates the MODIS sensor onboard the Terra satellite (if it is onboard the Aqua satellite, AAA would be 'MYD') BB: '13' refers to the product type, in this case 'vegetation indices' CC: 'C1' indicates product's spatial resolution, in this case, climate modeling grid (CMG) with 0.05-degree spatial resolution X: 'A' stands for 'Acquisition date' of the data YYYY: year of the acquisition date in 4 digits DDD: Julian day of the year of the acquisition date in 3 digits (i.e. July 12, 2002 is 193rd day, so DDD = 193) EEE: '061' is the processing version (i.e. version 6.1, a reprocessed data version with improved algorithms and corrections) yyyydddtttttt:timestamp representing when the file was created (yyyy for year in 4 digits, ddd for Julian day of the yar in 3 digits, tttttt for time of processing in 6 digits in HHMMSS format (19:54:53 = 7:54 PM UTC)) .hdf: standard file extension for the HDF file
- Variable name: 'CMG 0.05 Deg 16 days NDVI' >> this is the variable used in the study
- Units: 'NDVI' (Normalized Difference Vegetation Index)
- Datatype: int16
- Valid range: [-2000, 10000]
- Scale factor: 10000 (scaled NDVI value in 'double' datatype = grid NDVI value in 'double' datatype / scale factor)
- Spatial data matrix (2D) dimensions: Y dimension = 3600 rows, X dimension = 7200 columns
- Remarks: NDVI data measures vegetation greenness, an indicator of plant health and biomass
- Other variable names, units, datatypes: these are irrelevant to the study but provided for user's information 'CMG 0.05 Deg 16 days EVI' >> units 'EVI'(Enhanced Vegetation Index) >> datatype 'int16' 'CMG 0.05 Deg 16 days VI Quality' >> units 'EVI' (Enhanced Vegetation Index) >> datatype 'uint16' 'CMG 0.05 Deg 16 days red reflectance' >> units 'bit field' >> datatype 'int16' 'CMG 0.05 Deg 16 days NIR reflectance' >> units 'reflectance' >> datatype 'int16' 'CMG 0.05 Deg 16 days blue reflectance' >> units 'reflectance' >> datatype 'int16' 'CMG 0.05 Deg 16 days MIR reflectance' >> units 'reflectance' >> datatype 'int16' 'CMG 0.05 Deg 16 days Avg sun zen angle' >> units 'degrees' >> datatype 'int16' 'CMG 0.05 Deg 16 days NDVI std dev' >> units 'NDVI'(Normalized Difference Vegetation Index) >> datatype 'int16' 'CMG 0.05 Deg 16 days EVI std dev' >> units 'EVI'(Enhanced Vegetation Index) >> datatype 'int16' 'CMG 0.05 Deg 16 days #1km pix used' >> units 'Pixels' >> datatype 'uint8' 'CMG 0.05 Deg 16 days #1km pix +-30deg VZ' >> units 'Pixels' >> datatype 'uint8' 'CMG 0.05 Deg 16 days pixel reliability' >> units 'rank' >> datatype 'int8'
1c. FFDI (Forest Fire Danger Index) \ Data
- Source: Australian Bureau of Meteorology (BoM)
- Resolution: 0.05 degrees, daily
- Spatial extent: Australia-mainland North bounding latitude 10.00°S, South bounding latitude 45.50°S, West bounding longitude 112.00°E, East bounding longitude 156.25°E
- Temporal extent: 01 July 2002 to 31 December 2019
- Access: Data are not publicly available. They were purchased from BoM. For access, please request data at https://reg.bom.gov.au/climate/data-
- Note: This folder provides a sample of raw FFDI data used in the study from 2002.06.27 to 2002.07.12
- File arrangement in subfolders: Sample_FFDI > 16 data files from 2002.06.27 to 2002.07.12 (each file contains FFDI data for the given day).
- File count: 16 files
- Format: Network Common Data Form (NetCDF) >> file extension .nc
- File naming convention: AAAAYYYYMMDD.nc (e.g. FFDI20020627.nc) AAAA:'FFDI' stands for data product name 'Forest Fire Danger Index' YYYY: year of the acquisition date in 4 digits MM: month of the acquisition date in 2 digits DD: calendar day of the month of the acquisition date in 2 digits .nc: standard file extension for the NetCDF file
- Variable name: 'FFDI'
- Units: 'FFDI' (Forest Fire Danger Index)
- Datatype: int16
- Spatial data matrix (2D) dimensions: Y dimension = 886 rows, X dimension = 691 columns
- Remarks: FFDI quantifies fire risk based on McArthur forest fire danger rating system that uses weather conditions and surrogate variables of fuel moisture [details are given in the manuscript]
- Other variable names, description, units, datatypes, array length: these variables provide additional details for processing the 'FFDI' variable 'time' >> time dimension >> units 'days' >> datatype 'double' >> length = 1 'lat' >> latitude >> units 'degrees' >> datatype 'single' >> length = 691 'lon' >> longitude >> units 'degrees' >> datatype 'single' >> length = 886
1d. Köppen climate classification \ Data
- Note: this folder contains the 3 sub folders kpngrp, AUS_2021_AUST_SHP_GDA94, STE_2021_AUST_SHP_GDA94 each of their content is described below ----------------------------------------
### kpngrp
- Source: Australian Bureau of Meteorology (BoM)
- Note: direct download from BoM at http://www.bom.gov.au/climate/maps/averages/climate-classification/?maptype=kpngrp
- Files & contents: > kpngrp.txt - Koppen climate classification indices for Australia spatial domain, provided by BoM > readme.txt - 'README' file for kpngrp.txt, provided by BoM
- Remarks: The Köppen Climate Classification indices provides climate zone boundaries based on temperature and precipitation.
- Access: Köppen - major classes grid data are publicly available at http://www.bom.gov.au/climate/maps/averages/climate-classification/?maptype=kpngrp
### AUS_2021_AUST_SHP_GDA94
- Source: Australian Bureau of Statistics (ABS)/Digital boundary files/Australian Statistical Geography Standard (ASGS)-Edition 3
- Note: direct download from ABS at https://www.abs.gov.au/statistics/standards/australian-statistical-geography-standard-asgs-edition-3/jul2021-jun2026/access-and-downloads/digital-boundary-files
- Files & contents: > AUS_2021_AUST_GDA94.shp - main shape file with Australian administration boundary data (geometric features) > AUS_2021_AUST_GDA94.dbf - attribute data associated with each geometric feature > AUS_2021_AUST_GDA94.prj - projection file with the coordinate system & projection information for the shapefile (e.g., GDA94 in this case, which is the Geocentric Datum of Australia 1994) > AUS_2021_AUST_GDA94.shx - shape index file with index data to locate & access features in the shape file. > AUS_2021_AUST_GDA94.xml - metadata file with metadata about the shapefile (data source/creation date/authorship/description of attributes/other technical details)
- Remarks: .shp file can be opened & analyzed by open-source geospatial software (QGIS) or programming languages (Python/R/MATLAB) .prj & .xml files are optional but highly recommended for proper spatial alignment and documentation
- Access: publicly available at ABS at https://www.abs.gov.au/statistics/standards/australian-statistical-geography-standard-asgs-edition-3/jul2021-jun2026/access-and-downloads/digital-boundary-files
### STE_2021_AUST_SHP_GDA94
- Source: Australian Bureau of Statistics (ABS)/Digital boundary files/Australian Statistical Geography Standard (ASGS)-Edition 3
- Note: direct download from ABS at https://www.abs.gov.au/statistics/standards/australian-statistical-geography-standard-asgs-edition-3/jul2021-jun2026/access-and-downloads/digital-boundary-files
- Files & contents: > STE_2021_AUST_GDA94.shp - main shape file with Australian states boundary data (geometric features) > STE_2021_AUST_GDA94.dbf - attribute data associated with each geometric feature > STE_2021_AUST_GDA94.prj - projection file with the coordinate system & projection information for the shapefile (e.g., GDA94 in this case, which is the Geocentric Datum of Australia 1994) > STE_2021_AUST_GDA94.shx - shape index file with index data to locate & access features in the shape file. > STE_2021_AUST_GDA94.xml - metadata file with metadata about the shapefile (data source/creation date/authorship/description of attributes/other technical details)
- Remarks: .shp file can be opened & analyzed by open-source geospatial software (QGIS) or programming languages (Python/R/MATLAB) .prj & .xml files are optional but highly recommended for proper spatial alignment and documentation
- Access: publicly available at ABS at https://www.abs.gov.au/statistics/standards/australian-statistical-geography-standard-asgs-edition-3/jul2021-jun2026/access-and-downloads/digital-boundary-files
Sharing/access information
This is a summary of links to publicly available data and sources from the which certain data was purchased.
Links to publicly accessible locations of the data:
- NASA official website for MODIS > [https://lpdaac.usgs.gov/products/mod13c1v061/] for NDVI
- Australian Bureau of Statistics > [https://www.abs.gov.au/statistics/standards/australian-statistical-geography-standard-asgs-edition-3/jul2021-jun2026/access-and-downloads/digital-boundary-files] for Australian administration boundary data & Australian states boundary data
- Australian Bureau of Meteorology > [http://www.bom.gov.au/climate/maps/averages/climate-classification/?maptype=kpngrp] for Koppen major climate classification grid data in Australian spatial domain
The FFDI data used in the study were purchased from the Australian Bureau of Meteorology (BoM) under authors' self-service data request. BoM does not make these data publicly available.
> Access to FFDI data
If the readers wish to access the FFDI data, please follow the procedure outlined below.
Request self-service data at https://reg.bom.gov.au/climate/data-services/data-requests.shtml
Provide following specifications in the data request.
* Category: Climate & historical weather information
* State or territory: readers can specify the state of interest. If they require data for the entire continent, opt for "No specific part of Australia".
* Contact details
* Message details: Area of interest (e.g. Australia); type of data (e.g. FFDI); time period (e.g. 01/01/2002 to 31/12/2019); frequency of data (e.g. daily); proposed use of data (e.g. research).
> Meta data
Gridded fire weather climatology meta data is available at http://www.bom.gov.au/climate/averages/climatology/gridded-data- info/metadata/md_ave_ffdi.shtml
> FFDI processing
For users' clarity, a set of sample raw FFDI data is provided in 1. Data >> 1c. FFDI >> Sample_FFDI as described in # Descriptions section above.
Free/open software to read the data
.n' files: can be opened and read using Panoply, ncview, Integrated Data Viewer (IDV), Python (with netCDF4 or xarray), QGIS (with NetCDF Plugin)
.hdf files: can be opened and read using Panoply, HDFview, Integrated Data Viewer (IDV), PyHDF (Python Library)
.shp files: can be opened and read using QGIS, ArcGIS (does not support advanced editing)
Code/Software (Folder: 2. Scripts)
We have submitted the scripts used for our study's analysis in the 2. Scripts folder as detailed below.
All the scripts are MATLAB scripts ('.m' files) which were run using MATLAB version '9.14.0.2489007 (R2023a) Update 6'
All the input data files are MATLAB-processed data files ('.mat' files)
All the output data files are MATLAB-processed data files ('.mat' files) & MATLAB-processed figures ('.fig' files)
Folder & file structure summary:
- Scripts*
01. Study area
input files- koppen_data_300_500_au.mat,AUS_2021_AUST_GDA94.shp,STE_2021_AUST_GDA94.shp
script- script1_koppen_major_climate_classification_map.m
output files- koppen_data_300_500_au.mat,AUS_2021_AUST_GDA94.shp,STE_2021_AUST_GDA94.shp
02. Preprocessing
> sample1_ndvi
input files- sample_ndvi_au_16day_0dot1.mat
script- sample_script_1_preprocess_ndvi.m
output files- samplefig1.fig
> sample1_vod
input files- sample1_vod_au_16day_0dot1_at20040524.mat,sample2_vod_au_16day_0dot1_at20040524.mat,
timeline_16day_2002to2019.mat
script- sample_script_1_preprocess_vod.m
output files- samplefig1.fig,samplefig2.fig
> sample1_ffdi
input files- sample_ndvi_au_16day_0dot1.mat,sample_ffdi_au_daily_0dot1_20020627to20020712,
timeline_16day_2002to2019.mat
script- sample_script_1_preprocess_ffdi.m
output files- samplefig1.fig,samplefig2.fig
03. Preliminary testing
> Crosscorrelation
> xcf_analysis_ndviffdi
input files- ffdi_input_signals_16day_0dot1deg.mat,contpixelsnf0dot1.mat,
ndvi_input_signals_16day_0dot1deg.mat,coast.mat,waypoints_OZ.mat,worldlo.mat
script- script1_xcf_ndvi_ffdi.m
output files- xcfmatrix_ndviffdi.mat,xcfresults_ndviffdi.mat,xcfmatrix_ndviffdi.fig
> xcf_analysis_vodffdi
input files- ffdi_input_signals_16day_0dot1deg.mat,contpixelsvf0dot1.mat,
vod_input_signals_16day_0dot1deg.mat,coast.mat,waypoints_OZ.mat,worldlo.mat
script- script1_xcf_vod_ffdi.m
output files- xcfmatrix_vodffdi.mat,xcfresults_vodffdi.mat,xcfmatrix_vodffdi.fig
> Hurst
input: continentalpixelids_0dot1.mat,vod_16day_0dot1deg_2002to2019.mat*
script: script1_hurstmatrix_vod.m
output:hurstmatrix_vod.mat,hurstmatrix_vod.fig
--
input: continentalpixelids_0dot1.mat,ndvi_16day_0dot1deg_2002to2019.mat*
script: script1_hurstmatrix_ndvi.m
output:hurstmatrix_ndvi.mat,hurstmatrix_ndvi.fig
--
input: continentalpixelids_0dot1.mat,ffdi_16day_0dot1deg_2002to2019.mat*
script: script1_hurstmatrix_ffdi.m
output:hurstmatrix_ffdi.mat,hurstmatrix_ffdi.fig
04. FFDI segment extraction
input: continentalpixelids_0dot1.mat,ffdi_16day_0dot1deg_2002to2019.mat,
**timeline_16day_2002to2019.mat*
script: script1_segment_extraction.m
output:upswing_timeranges.mat.mat
--
input: continentalpixelids_0dot1.mat,upswing_timeranges.mat,ffdi_16day_0dot1deg_2002to2019.mat,
**vod_16day_0dot1deg_2002to2019.mat,timeline_16day_2002to2019.mat*
script: script2_segment_data.m
output:vod_ffdi_segmentdata.mat
05. Models
> a. Model estimation
input: continentalpixelids_0dot1.mat,vod_ffdi_segmentdata.mat* script: script1_model_estimation.m
output:substruct_nonempty_segmentdata.mat,model_information.mat, models_var_ar.mat,lag_distribution_vodffdi_segmentbased.mat --
input: lag_distribution_vodffdi_segmentbased.mat script:script2_fig5_lag_distribution.m
output:fig5_lag_distribution_vodffdi.fig
> gct_sample
> ndvi_ffdi
input: ffdi_16day_0dot1deg_2002to2019.mat,ndvi_AU_0dot1deg_16day_2002to2019.mat, continentalpixelids_0dot1.mat
script:script1_gct_ndviffdi.m
output:gctspatialmatrix_ndviffdi.mat,gctresults_ndviffdi.mat,gc_spatialmap_ndviffdi.fig
> vod_ffdi
input: ffdi_16day_0dot1deg_2002to2019.mat,vod_AU_0dot1deg_16day_2002to2019.mat, continentalpixelids_0dot1.mat
script:script1_gct_vodffdi.m
output:gctspatialmatrix_vodffdi.mat,gctresults_vodffdi.mat,gc_spatialmap_vodffdi.fig
> b. Calibration&validation
> 1. calibration_outputs
input: model_information.mat,models_var_ar.mat,vod_ffdi_segmentdata.mat, substruct_nonempty_segmentdata.mat,ffdi_16day_0dot1deg_2002to2019.mat,
vod_16day_0dot1deg_2002to2019.mat,continentalpixelids_0dot1.mat,
timeline_16day_2002to2020.mat,upswing_time_segments.mat
script:script1_calibration_ffdivodmodel.m
output:leastsquarefit_input_data.mat,model_functions.mat,leastsquarefit_output_info.mat
[in output_data_model_var_ffdivod subfolder]
--
input: model_information.mat,models_var_ar.mat,vod_ffdi_segmentdata.mat, substruct_nonempty_segmentdata.mat,ffdi_16day_0dot1deg_2002to2019.mat,
vod_16day_0dot1deg_2002to2019.mat,continentalpixelids_0dot1.mat,
timeline_16day_2002to2020.mat,upswing_time_segments.mat script:script2_calibration_ffdi_model.m
output:leastsquarefit_input_data.mat,model_functions.mat,leastsquarefit_output_info.mat
[in output_data_model_ar_ffdi subfolder]
--
input: leastsquarefit_input_data.mat,leastsquarefit_output_info.mat
[from respective subfolders output_data_model_var_ffdivod & output_data_model_ar_ffdi] script:script3_calibration_metrics.m
output:calibration_metrics.mat,
[auxiliary .mat files in subfolders output_data_model_var_ffdivod & output_data_model_ar_ffdi]
> 2. validation_outputs
input: 'upswing_testing.mat,ffdi_16day_0dot1deg_2002to2019.mat,continentalpixelids_0dot1.mat
script:script1_evaluate_var_ffdivod.m
output:validation_outputs_ffdivod.mat
[auxiliary .mat files in subfolder output_data_model_var_ffdivod]
--
input: 'upswing_testing.mat,ffdi_16day_0dot1deg_2002to2019.mat,continentalpixelids_0dot1.mat script:script2_evaluate_ar_ffdi.m
output:validation_outputs_ffdi.mat,
[auxiliary .mat files in subfolder output_data_model_ar_ffdi]
> *3. metric plots*
input: **calibration_correlation_obsvssim_24leads_var&ar.mat**,
**calibration_avgNSE_avgRMSE_24leads_var&ar.mat**,
**validation_correlation_obsvssim_24leads_var&ar.mat**,
**validation_avgNSE_avgRMSE_24leads_var&ar.mat**
script:**script1_metric_plots.m**
output:**samplefig1_nse_calibration.fig**,**samplefig2_rmse_calibration.fig**,
**samplefig3_correlation_calibration.fig**,**samplefig4_nse_validation.fig**,
**samplefig5_rmse_validation.fig**,**samplefig6_correlation_validation.fig**
> *4. nse_spatial_analysis*
input: **continentalpixelids_0dot1.mat**,**fcast_var.mat**,**fcast_ar.mat**,**gc_pixels.mat**, **pixels_with_novodnoffdi.mat**,**allnanpixels.mat**
script:**script1_nse_spatial_classification.m**
output:**nsevarmatrix.mat**,**nsearmatrix.mat**,**nse_class_matrix.mat**,
**figs3a_nse_var_segmentbased.fig**,**figs3b_nse_ar_segmentbased.fig**,
**fig7_nse_classification.fig**
> *c. AIC_comparison*
input: **aicvalues_var_ar_allpixels.mat**,**aicvalues_var_ar_gcpixels.mat*,
**continentalpixelids_0dot1.mat**,**nansquarevod.mat'**
script:**script1_aic_spatial_maps.m**
output:**fig6(a-1)_aic_ffdivod_spatialmap.fig**,**fig6(a-2)_aic_ffdivod_gcpixels_spatialmap.fig**,
**fig6(b-1)_aic_ffdi_spatialmap.fig**,**fig6(b-2)_aic_ffdi_gcpixels_spatialmap.fig**
> *d. Sample forecasts
input: **continentalpixelids_0dot1.mat**;**ffdi_16day_0dot1deg_2002to2019.mat**, **validation_outputs_ffdivod.mat**,**simulated_data_ffdi_validation.mat**,
**simulated_data_ffdivod_validation.mat**,**timeline_16day_2002to2020.mat**
script:**example_script1_model_predictions.m**
output:**example1_fig9inmanuscript.fig**,**example2_figS6(1)inmanuscript.fig**,
**example3_figS6(2)inmanuscript.fig**,**example4_figS6(3)inmanuscript.fig**,
**example5_figS6(4)inmanuscript.fig**
> *06. Check for climate zones*
input: **nsevarmatrix.mat**,*nse_class_matrix.mat**,**koppen_climate_classes_au_300_500.mat**,
**continentalpixelids_0dot1.mat**,**coast.mat**,**waypoints_OZ.mat**,**worldlo.mat**
script: **script1_application_to_koppen_classes_au.m**
output:**positive_nse_matrix.mat**,**samplefig1_spatial_map_12monthnse_vodffdi.fig**, **tally_nse_kopppen.mat**,**koppen_indices.mat**,**nse_per_koppen.mat**,**nse_per_majkoppen.mat**
**samplefig2_boxplot_majorkoppen_vodffdi.fig**
General notes:
- In above summary, supplemental datasets were processed in advance to facilitate the main processing steps & reduce computational time. These auxiliary datasets serve as shared inputs across multiple sub folders. >> continentalpixelids_0dot1.mat (all the land pixels with reference to VOD/NDVI/FFDI spatial maps, this avoids the methods being unnecessarily executed in ocean pixels) > > contpixelsnf_0dot1.mat (all the land pixels extracted from 'continentalpixelids_0dot1.mat' for NDVI/FFDI) > > contpixelsvf_0dot1.mat (all the land pixels extracted from 'continentalpixelids_0dot1.mat' for VOD/FFDI)\ > > timeline_16day_2002to2020.mat (16-day time steps from 12-07-2002 to 01-01-2020 as indexed from NDVI 16-day acquisition days (MOD13C1v061) to facilitate time series indexing) > > nansquarevod.mat (continental pixels where vod is absent were pre-identified to index them equally during spatial map comparison)
- In above summary, the geospatial data used for creating spatial visualizations are > > coast.mat (coastline data in the form of latitude and longitude points that can be used to overlay coastlines on maps)\ > > waypoints_OZ.mat (geospatial waypoints specific to Australia) > > worldlo.mat (world boundary data including coastlines/country borders, source: U.S. National Imagery and Mapping Agency's (NIMA))
- When applying certain phases of the methodology across the entire spatial domain, it becomes computationally intensive. In such instances, we have also included example scripts demonstrating pixel-based execution to facilitate the reader's understanding.
---
# Contact
For questions or regarding further details, please contact:
Dinuka Kankanige ([d.kankanige@unsw.edu.au](mailto:d.kankanige@unsw.edu.au))
Methods
Dataset relates to Vegetation Optical Depth (a merged output from three satellite retreived products), Normalized Difference Vegetation Index (publicly accessible data product from Terra - Moderate Resolution Imaging Spectroradiometer as MOD13C1 Version6.1), and Forest Fire Danger Index (purchased from the Australian Bureau of Meteorology). They were spatially and temporally resampled from their raw resolutions, and processed for the analysis inetended in the study (fire risk forecast modeling). All the processing was conducted in MATLAB 9.14.0.2489007 (R2023a) Update 6.