Fire Data Sources
Many sources of weather, terrain, and fuel data for initialization, assimilation, and verification are freely available through public sites, many in the form of files viewable in Google Earth. Fire ignition tim/location data and imagery of fire extent at later times for specific incidents is still generally in the form of custom datasets where research or suppression teams were available to collect it.
Fuel Maps
- LandFire http://landfire.cr.usgs.gov
10-hr Dead Fuel Moisture
- Remote Automated Weather Station (RAWS) data http://raws.wrh.noaa.gov/roman/ discontinued
- MesoPy access to http://api.mesowest.net/v2/
Live Fuel Moisture
- http://wfas.net based on NVDI data
Fire Ignition Time and Location Data
- Custom
Fire Perimeters
- Geospatial Multi-Agency Coordination (GEOMAC) http://www.geomac.gov/
- Incident Information System http://inciweb.nwcg.gov/
Fire Detection
- Firemapper https://fsapps.nwcg.gov/afm/
- Latest MODIS from Forest Service http://activefiremaps.fs.fed.us/data/kml/conus_latest_AFM_bundle.kml
- Modis raw data ftp://152.61.133.5/MOLT/MOD14.005 Tera, ftp://152.61.133.5/MOLA/MOD14.005 Aqua satellite
HDF File Information
HDF stands for Hierarchical Data Format, there are two current standards; HDF4 (.hdf extension) and HDF5 (.h5 extension). All of our sources for active fire data and geolocation data use HDF4, though this could change in the future.
Reading HDFs
Matlab uses 'hdfread' to read a dataset into an array, requires filename and the path to the relevant data inside the hdf file. Here's an example call:
fire_mask = hdfread('MOD14.A2016266.1505.006.2016266215413.hdf', '/fire mask')
Where the filename is 'MOD14.A2016266.1505.006.2016266215413.hdf' (more on the naming convention later), and the path inside the hdf to the firemask data is '/fire mask'
More on matlab hdf here https://www.mathworks.com/help/matlab/ref/hdfread.html
Python information: http://hdfeos.org/software/pyhdf.php
Useful links:
- HDF4 User Guide: https://support.hdfgroup.org/release4/doc/UsrGuide_html/UG_PDF.pdf
- Software useful for browsing the data structure, finding data paths, etc.: http://www.giss.nasa.gov/tools/panoply/
- A way to grab only parts of an HDF file from a server? : http://hdfeos.org/software/hdf5_handler.php (looks to require server side support, need to check if our sources support it)
MODIS Data:
MOD14 Files:
Level 2 active fire data (750(?)m resolution) from the MODIS instrument on the Terra satellite (Aqua satellite also has a MODIS instrument, that fire data is under MYD14 with the same properties). Contains 4 fields important for our purposes:
- fire_mask: Array containing sensor data for each pixel. Classes:
- 0 missing input data
- 1 not processed (obsolete)
- 2 not processed (obsolete)
- 3 non-fire water
- 4 cloud
- 5 non-fire land
- 6 unknown
- 7 fire (low confidence)
- 8 fire (nominal confidence)
- 9 fire (high confidence)
- FP_latitude: Variable length vector (the number of pixels identified as firepixels) containing latitude of all firepixels in the mask.
- FP_longitude: Variable length vector (the number of pixels identified as firepixels) containing longitude of all firepixels in the mask.
- FP_confidence: Variable length vector (the number of pixels identified as firepixels) containing confidence of fire identification for each firepixel in the mask.
These can be found at: ftp://ladsweb.nascom.nasa.gov/allData/6/MOD14/
Example Filename: 'MOD14.A2016266.0615.006.2016266201759.hdf' ~200KB - 800KB
[Dataset name].A[Year][DayofYear].[24hrTime].[Collection number].[collection time?].hdf
(The files here are delivered in 5 minute intervals, and are organized by year, then by Julian Day Number (n corresponds to the nth day of the year).)
Geolocation metadata can be found at ftp://ladsweb.nascom.nasa.gov/geoMeta/6/TERRA
Download the file for the corresponding date and check to see if a granule you are considering lies within certain bounds. No analog for VIIRS files is currently known.
MOD03 Files:
Level 1 geolocation data from the MODIS instrument on the Terra satellite (Aqua satellite also has a MODIS instrument, that geolocation data is under MYD03 with the same properties). Contains 2 fields important for our purposes:
- Latitude: Array (same size as fire_mask) containing the latitude of each pixel in the firemask
- Longitude: Array (same size as fire_mask) containing the longitude of each pixel
These can be found at: ftp://ladsweb.nascom.nasa.gov/allData/6/MOD03/
Example Filename: 'MOD03.A2016266.2210.006.2016267050406.hdf' ~28MB - 32MB
[Dataset name].A[Year][DayofYear].[24hrTime].[Collection number].[collection time?].hdf
(The files here are delivered in 5 minute intervals, and are organized by year, then by Julian Day Number (n corresponds to the nth day of the year).)
VIIRS Data:
VNP14IMG_NRT files:
Level 2 active fire data (375m resolution) from the VIIRS instrument on the Suomi-NPP Satellite. Contains 4 fields important for our purposes:
- fire_mask: Array containing sensor data for each pixel. Classes (Note: these are slightly different than the MOD14 fire mask classes):
- 0 not-processed (non-zero QF)
- 1 bowtie
- 2 glint
- 3 water
- 4 clouds
- 5 clear land
- 6 unclassified fire pixel
- 7 low confidence fire pixel
- 8 nominal confidence fire pixel
- 9 high confidence fire pixel
- FP_latitude: Variable length vector (the number of pixels identified as firepixels) containing latitude of all firepixels in the mask.
- FP_longitude: Variable length vector (the number of pixels identified as firepixels) containing longitude of all firepixels in the mask.
- FP_confidence: Variable length vector (the number of pixels identified as firepixels) containing confidence of fire identification for each firepixel in the mask.
These can be found at: ftp://nrt3.modaps.eosdis.nasa.gov/allData/5000/VNP14IMG_NRT/
They are held behind a password wall, registration is easy at https://urs.earthdata.nasa.gov/users/new
Example Filename: 'VNP14IMG_NRT.A2016323.0748.001.hdf' ~ 0.5MB - 5MB
[Dataset name].A[Year][DayofYear].[24hrTime].[Collection number].hdf
(These files are delivered in 6 minute intervals, and are organized by year, then by Julian Day Number (n corresponds to the nth day of the year).)
NPP_IMFTS_L1 files:
Level 1 geolocation files (375m resolution) from the VIIRS instrument on the Suomi-NPP Satellite. Contains 2 fields important for our purposes:
- Latitude: Array (same size as fire_mask) containing the latitude of each pixel in the firemask
- Longitude: Array (same size as fire_mask) containing the longitude of each pixel
These can be found at: ftp://ladsweb.nascom.nasa.gov/allData/5000/NPP_IMFTS_L1/
No password this time.
Example Filename: 'NPP_IMFTS_L1.A2016323.0930.001.2016323192201.hdf' ~ 1GB (yikes!)
[Dataset name].A[Year][DayofYear].[24hrTime].[Collection number].[collection time?].hdf
(The files here are delivered in 6 minute intervals, and are organized by year, then by Julian Day Number (n corresponds to the nth day of the year).)
Algorithms, investigations, etc
This section will contain a discussion of the generation and structure of our files to be used with the wrfxpy
Tools and websites to help locate the data...?
To be continued...
Fire Location and Intensity Data
- Satellite fire detection from FireMapper by the NASA Moderate-Resolution Imaging Spectroradiometer (MODIS) and VIIRS
- VIIRS products