AVHRR Data Flow at CCRS Presented by Gunar Fedosejevs

1
AVHRR Data Flow at CCRSPresented by Gunar
Fedosejevs

• AVHRR Data Flow at the Canada Centre for Remote
  Sensing (CCRS) includes satellite data reception,
  pre-processing, archiving, and enhanced data
  processing for derived products in support of
  modelling and global climate research.

2
AVHRR Data Flow at CCRS

• CCRS as the official Canadian remote sensing
  centre has two satellite receiving stations
  
• The Prince Albert Satellite Station (PASS)
• The Gatineau Satellite Station (GSS)

3
AVHRR Data Flow at CCRS

• PASS has been receiving direct readout High
  Resolution Picture Transmission (HRPT) data from
  the Advanced Very High Resolution Radiometer
  (AVHRR) instrument onboard the National Oceanic
  and Atmospheric Administration (NOAA) series of
  satellites since 1973 (http//www.ccrs.nrcan.gc.ca
  /ccrs/data/satsens/sats/noaa_e.html).
• PASS has been receiving HRPT data from NOAA 16
  polar-orbiting satellite for ascending (daytime)
  orbits since April 1, 2001 and descending
  (night-time) orbits shortly thereafter for most
  of North America.

4
AVHRR Data Flow at CCRS

• NOAA 14 has been drifting into a later orbit
  where solar contamination has become a problem
  for thermal calibration.
  
• The high solar zenith angles from the later
  orbits also result in unwanted increased
  bi-directional reflectance distribution function
  (BRDF) effects in channels 1 and 2.
• CCRS has ceased to acquire NOAA 14 (ascending
  orbit) data on an operational basis as of March
  31, 2001.
  
• NOAA 14 AVHRR data are still archived by USA to
  NOAA SAA.
  
• Users may request acquisition of data by PASS for
  any operational AVHRR sensor through the CCRS
  order desk.
  

5
AVHRR Data Flow at CCRS

• Raw Products
• Satellite Acquisition Services Order Desk
• Canada Centre for Remote Sensing
• 588 Booth Street
• Ottawa, Ontario
• CANADA
• K1A 0Y7
• Tel [phone]
• Fax [phone]
• Email orderdesk_at_ccrs.nrcan.gc.ca

6
AVHRR Data Flow at CCRS

• NOAA 1516 contain the new AVHRR/3 instrument
  with six spectral channels. Channel 1 (visible),
  2 (NIR), 3A (SWIR), 3B (MIR), 4 (TIR), 5 (TIR)
  have centre wavelengths of 0.65, 0.85, 1.6, 3.6,
  10.5, 11.5 um, respectively.
• Satellite data transcription systems were
  designed to handle 5 channels of the AVHRR/2
  onboard NOAA 14 and earlier missions.
  
• Thus, NOAA 1516 are programmed to acquire
  channel 3A during the day (ascending orbit) and
  channel 3B during the night (descending orbit).
  

7
AVHRR Data Flow at CCRS

• For descending orbits passing over the Canadian
  arctic, this often results in both channel 3A and
  3B data being acquired.
  
• Channel 3A can be used for cloud/snow/smoke
  discrimination.
  
• Technical specifications for NOAA 16 AVHRR/3 can
  be found in the NOAA KLM User's Guide
  (http//www2.ncdc.noaa.gov/docs/klm/index.htm).
  

8
AVHRR Data Flow at CCRS

• Data Reception at Mont Joli
• NOAA 14 AVHRR HRPT data from ascending daytime
  orbits were acquired in 1999 and 2000 over
  Canada's East Coast by the DFO Maurice Lamontagne
  Institute at Mont Joli for CCRS under a
  Memorandum of Understanding (MOU).
• Reception of NOAA 16 AVHRR HRPT daytime data at
  Mont Joli commenced on April 5, 2001 under the
  same MOU.
  
• DFO provides near real-time delivery of AVHRR
  data as raw and L1B products via ftp to PASS.
  
• PASS software pre-processes DFO-received data to
  make them compatible with PASS-received data.
  

9
AVHRR Data Flow at CCRS

• NATAS Data Processing
• HRPT data received at PASS and Mont Joli are
  processed on the NOAA AVHRR Transcription and
  Archive System (NATAS) system at PASS
  (http//ceocat.ccrs.nrcan.gc.ca/client_acc/guides/
  avhrr/ch4.html).
• NATAS raw data products are generated in the
  Committee on Earth Observations
  Satellites/Landsat Ground Station Operators
  Working Group (CEOS/LGSOWG) standard format
  (refer to STD-TM 90-678F document on Standard Raw
  AVHRR CCT Image Format Specifications).
• NATAS products are archived on 5-gigabyte 8-mm
  Exabyte cassettes kept at PASS.

10
AVHRR Data Flow at CCRS

• CEOCat Data Catalogue
• All NATAS raw data products can be viewed on the
  Canadian Earth Observation Catalogue (CEOCat)
  (http//ceocat.ccrs.nrcan.gc.ca/cgi-bin/client_acc
  /ceocate/holdings.phtml).
• NATAS-generated digital quicklook browse (JPEG)
  images and Catalogue Update Files (CUFs) are
  transferred on a daily basis via ftp to a CCRS
  VAX server at 588 Booth St.
• A batch job initiated daily from the CEOCat
  server at 615 Booth St. retrieves the data from
  the VAX server.
  
• Channels 2, 1 and 1 are displayed as an RGB false
  colour composite for orbits received from April 1
  to October 31 or channel 4 as a B/W image for
  orbits received in the winter.

11
AVHRR Data Flow at CCRS

• CCRS ftp transfer
• Since 2000, CEOS-formatted NATAS raw data .zip
  files are transferred by ftp from PASS over the
  internet to a secure CCRS AVHRR ftp site for
  privileged user access.
• The .zip files for NATAS data are labelled
  NOAA_16_yyyymmdd_hhmm_nnnnn where nnnnn is the
  orbit number. The GMT time (hhmm) is for the
  first scan line when the scene is oriented with
  north at the top.

12
AVHRR Data Flow at CCRS

• CCRS ftp transfer (continued)
• For the descending orbits the jpegs are labeled
  as 5-minute segments starting with the start
  orbit time. For the ascending orbit data, the
  jpegs are labeled as 5-minute segments but
  starting from the orbit acquisition start time
  and going forwards (north along the orbit) even
  though the jpegs are oriented north at the top of
  the image. The time in the jpeg filename applies
  to the south end of the images.
• The time in the .zip filename is the acquisition
  stop time even though the top scan lines in the
  jpeg may be blank fill.

13
AVHRR Data Flow at CCRS

• EMS Data Holdings
• All NATAS data received on the CCRS ftp server
  are archived on exabyte or DLTs in the
  Environmental Monitoring Section (EMS) the
  Applications Division at CCRS
• EMS Data Holdings are applied to various research
  studies using GeoComp-n (http//www.ccrs.nrcan.gc.
  ca/ccrs/rd/ana/geocomp/geocomp_e.html), ABC3 and
  other data processing packages in the context of
  global climate change and vegetation process
  modelling

14
AVHRR Data Flow at CCRS

• MRSC GeoComp-n Data Processing
• The Manitoba Remote Sensing Centre (MRSC)
  retrieves NATAS data from the CCRS ftp site to be
  processed on the operational system for Geocoding
  and Compositing of AVHRR data (GeoComp-n).
• GeoComp-n creates near real time composite
  products of Canada at 1-km resolution in Lambert
  Conic Conformal (LCC) projection.
  
• Canada-wide daily and 10-day composite products
  containing selected data layers are generated by
  GeoComp-n for CCRS.
  
• Alternate composite periods, geographic coverage
  and data layer selections are available to MRSC
  subscribers.
  
• JPEG browse images (using channel 12 TOA
  radiance and NDVI) and CUFs are sent from MRSC
  to CEOCat.

15
AVHRR Data Flow at CCRS

• GeoComp-n Basic Composite Product Layers
• TOA radiance and brightness temperature
• TOA reflectance, surface reflectance,
  BRDF-corrected surface reflectance and associated
  NDVI
  
• Satellite and sun zenith/azimuth angles
• Quality control (QC) mask
• Input scene mask
• Pixel count mask
• Relative date
• Residual geometric error mask

16
AVHRR Data Flow at CCRS

• GeoComp-n Advanced Composite Product Layers
• Pixel contamination (CECANT) mask
• Land Surface temperature
• Leaf area index (LAI)
• Instantaneous FPAR
• Daily Mean FPAR
• Instantaneous APAR Daily Total APAR
• Composite Mean APAR
• Fire (hot spot) mask
• PAR albedo

17
AVHRR Data Flow at CCRS

• Geocoded and Composite Products
• Manitoba Remote Sensing Centre
• Roy Dixon
• 1007 Century Street
• Winnipeg, Manitoba
• CANADA
• R3H 0W4
• Tel [phone]
• Fax [phone]
• Email rdixon_at_nr.gov.mb.ca
• WWW http//www.gov.mb.ca/conservation/geomatics/r
  emote_sensing/index.html

18
AVHRR Data Flow at CCRS

• MRSC Data Transfer to CCRS
• MRSC GeoComp-n products generated for CCRS are
  sent on exabyte periodically to CCRS.
  
• In 2000 this data set included virtually all
  product layers for the daily and 10-day
  composites for NOAA 14.
  
• In 2001/2002 only a subset of product layers
  were requested for the daily and 10-day
  composites.
  
• The 10-day composite radiance data are
  reprocessed in the

19
AVHRR Data Flow at CCRS

• ABC3V2 Data Processing
• At the end of a growing season (April 1- October
  31), all 10-day composite data are staged on the
  ABC3V2 server for processing in the Atmosphere,
  Bi-directional and Contamination Corrections of
  CCRS (ABC3) software written in Visual C by
  Rasim Latifovic,
• http//www.ccrs.nrcan.gc.ca/ccrs/rd/apps/landcov/
  corr/emcorr_e.html.
  
• ABC3V2 applies recalibration and other
  radiometric enhancements. Details can be found in
  ABC3V2.doc on CCRS anonymous ftp site
  ftp//ftp.ccrs.nrcan.gc.ca/ftp/ad/EMS/ABC3Document
  ation/.
• The 1-km resolution output products (.pix files
  with a file size of 5700 pixels by 4800 lines)
  are available in LCC projection for model input
  and derived product generation.

20
AVHRR Data Flow at CCRS

• ABC3V2 Data on GeoGratis
• A subset of ABC3V2 product layers from 1993 to
  1998 were converted from LCC to geographic
  (latitude/longitude) projection and archived
  along with browse and thumbnail images in
  GeoGratis at URL http//www.geogratis.gc.ca/frames
  .html.
• These data sets may be searched for on the
  GeoConnections Discovery Portal at URL
  http//ceonet.ccrs.nrcan.gc.ca/.
  
• A download of selected dates and product layers
  of 10-day composite data from GeoGratis is
  bundled with README.doc and ABC3V2.doc.

21
AVHRR Data Flow at CCRS

• ABC3V2 Data on GeoGratis (continued)
• The browse images are 1/10th in size with an RGB
  assignment of BRDF-corrected NDVI (using only
  values from 0.9-2.0), ch2 (using only values from
  0-0.65) and ch1 (using only values from 0-0.35),
  respectively.
• The thumbnails are 100 pixels by 42 lines in size
  before reprojection.
  
• Retrieval of data sets is based on UNIX tools.
• If retrieving data files on PC using Winzip, the
  user must uncheck the "TAR file smart CR/LF
  conversion" box under Winzip/Options/Configuration
  s/Miscellaneous.

22
AVHRR Data Flow at CCRS

• EMS Fire Processing at CCRS
• Batch/perl script/EASI scripts retrieve the NATAS
  data and processes them in GeoComp-n for fire
  mapping research by the Fire M3 scientists
  (http//fms.nofc.cfs.nrcan.gc.ca/FireM3/).
• While there are additional benefits of using
  channel 3A data to the fire modelling, the
  current hot spot mapping algorithm implemented
  for NOAA 11 and 14 in GeoComp-n was designed to
  use channel 3B data.
• A major research effort is ongoing to perform hot
  spot and smoke detection in NOAA 16 daytime data
  using channel 3A and hot spot detection in
  night-time data using the reliable channel 3B
  data. Channel 3A data can contain a thermal
  emissive component. Depending on the size and
  temperature of the fire, pixels can be darker
  (low reflectivity of burned vegetation) or
  brighter (hot fire) than for background
  vegetation.
• Clock corrections for improved accuracy in
  geocoding are available for NOAA 11, 12, 14, 15
  and 16 from NOAA/NESDIS (http//www.osdpd.noaa.gov
  /PSB/NAVIGATION/navpage.html).

23
AVHRR Data Flow at CCRS

• EMS Fire Processing at CCRS (continued)
• A mask of false hot spots in night-time
  composites can be created from permanent thermal
  sources such as smoke stacks in industrial
  plants. A similar mask can be built for daytime
  composites where bright surfaces (e.g. White
  Sands, New Mexico) are mistaken for smoke in
  channel 3A.
• Another source of false hot spots is noisy
  pixels/lines in the raw data. These occur in the
  NOAA data received by PASS because of data
  downlink errors when the satellite is low on the
  horizon. PASS employs a reception mask of zero
  degrees elevation that can cause bad pixels or
  calibration sample data. The DFO data acquired at
  Mont Joli with a mask of 5 degrees contain no bad
  pixels. These bad pixels must be manually removed
  as bad line replacement is not envoked for the
  fire processing in GeoComp-n so as not to miss
  any hot spots.

24
AVHRR Data Flow at CCRS

• GeoComp-n Changes for NOAA 15/16
• GeoComp-n thresholds based on raw counts from the
  NOAA 15 and NOAA 16 dual gain AVHRR/3 sensor may
  need to be revised for bad pixel detection, cloud
  threshold and GCP radiometric threshold and for
  any products using raw counts for channels 1 and
  2 or computed NDVI.
• GeoComp-n products for NOAA 15 use post-launch
  time-dependent PWL calibration coefficients based
  on Tahnk and Coakley ice data.
  
• GeoComp-n products for NOAA 16 use post-launch
  time-dependent PWL calibration coefficients based
  on Heidinger et al. calibration data.
  
• PWL calibration coefficients are available on
  CCRS calval web page http//www.ccrs.nrcan.gc.ca/c
  crs/rd/ana/calval/calhome_e.html.
  
• A new set of BRDF correction coefficients must be
  created from a full season of data.
  
• New product algorithm coefficients specific to
  NOAA 16 need to be derived because of spectral
  band differences between NOAA satellites.
  

25
AVHRR Data Flow at CCRS

• AVHRR SST Product
• While Geocomp-n and ABC3V2 can generate land
  surface temperature (LST) products CCRS doesn't
  have an operational sea surface temperature (SST)
  product. PCI does provide limited support for the
  SST product in XPACE computed from NOAA 9, 12 and
  14 data. The PCI model NOAA_SST.mod was built at
  CCRS to compute SST for NOAA 9, 11, 12 and 14 and
  will run in PCI V6.3.

26
AVHRR Data Flow at CCRS

• NATAS Historical Data
• CCRS plans are to retrieve 1993-1998 NATAS data
  from the PASS archive to process through
  Geocomp-n and ABC3V2.
  
• PASS is recovering historical NOAA data (from HDDT and creating a NATAS exabyte archive.
  The NATAS data are sent to the CCRS AVHRR ftp
  site as zip files. The browse images and CUFs are
  sent to CEOCat.
• NOAA data from 1990-1993 with almost complete
  seasonal coverage are now in CEOCat and the EMS
  Data Holdings.
  
• In support of the 1-km global mapping initiative
  by the USGS EROS Data Center in Sioux Falls,
  Dakota, year-round acquisition of NOAA 11 and 14
  data occurred from April 1992 until October 1997
  notwithstanding the data gap between the demise
  of NOAA 11 in September 1994 and the start-up of
  NOAA 14 in March 1995 (?).
• The data going back to 1983 are much sparser but
  all 10-bit historical data will be archived in
  CEOCat in time.

27
AVHRR Data Flow at CCRS

• AVHRR Space Counts
• Instead of assuming that the calibration offsets
  (based on pre-launch values for NOAA 15 and 16)
  as specified in the GeoComp-n radiometric
  calibration file are the correct ones the space
  counts or offset can be read directly from the
  raw image files generated by NATAS.
• GeoComp-n doesn't retrieve space counts for the
  non-thermal channels for inclusion in the PCIDSK
  metadata.
  
• A Perl script was created to extract the space
  counts from the suffix (calibration sample) data
  in the NATAS image file and write them to a text
  file. The space count records were imported into
  an Excel file to view the time-dependent
  behaviour.

28
AVHRR Data Flow at CCRS

• AVHRR Solar Contamination
• As NOAA satellites drift into a later orbit, the
  sun impinges on the AVHRR instrument for part of
  the orbit, which results in thermal fluxes in the
  blackbody cavity that affect the PRT readings of
  the blackbody temperature. The rapidly changing
  thermal calibration coefficients can result in
  brightness temperature errors on the order of
  several degrees Kelvin.
• Trishchenko, A.P., Z. Li. 2001. A method for the
  correction of AVHRR onboard IR calibration in the
  event of short term radiative contamination.
  IJRS, vol. 22, no. 17, pp. 3619-3624.
• Trishchenko, A.P. 2002. Removing unwanted
  fluctuations in the AVHRR thermal calibration
  data using robust techniques. Journal of
  Atmospheric and Oceanic Technology (submitted).
• Trishchenko, A.P., G. Fedosejevs, Z. Li and J.
  Cihlar. 2002. Trends and uncertainties in thermal
  calibration of AVHRR/NOAA-9 to 16. Journal of
  Geophysical Research Atmospheres (submitted).

29
AVHRR Data Flow at CCRS

• NDVI Spectral Conversion
• Alex Trichtchenko wrote a journal article
  describing the spectral dependency of NDVI
  between NOAA satellites, which included empirical
  fits between NOAA 14 TOA NDVI values and those
  from other NOAA satellites. Models that were
  created for NOAA 14 need to apply a spectral
  conversion to NOAA 16 data. The spectral
  conversion is target reflectance dependent.
• Reference
• Trishchenko, A.P., J. Cihlar, Z. Li. 2002.
  Effects of spectral response function on surface
  reflectance and NDVI measured with moderate
  resolution satellite sensors. Remote Sensing of
  Environment , vol. 81, pp. 1-18.

30
AVHRR Data Flow at CCRS
31
AVHRR Data Flow at CCRS
32
AVHRR Data Flow at CCRS
33
Maximal detectable temperature in ch.3B
34
Maximal detectable temperatures for NOAA-14
(HRPT scenes from PASS)

35
Conclusions

• CCRS provides complete support for NOAA AVHRR
  data to users
  
• Web access to quick look images
• COFUR access to raw data
• Near real-time access to composite data by
  subscription
  
• Web access to free historical enhanced data

36
AVHRR Data Flow at CCRS

• References
• Adair, M., J. Cihlar, B. Park, G. Fedosejevs, A.
  Erickson, R. Keeping, D. Stanley, and P.
  Hurlburt. 2001. GeoComp - n, an advanced system
  for generating products from coarse and medium
  resolution optical satellite data. Part 1 System
  characterization. Canadian Journal of Remote
  Sensing, vol. 28, no. 1, pp. 1-20.
• Cihlar, J., J. Chen, Z. Li, G. Fedosejevs, M.
  Adair, W. Park, R. Fraser, A. Trishchenko, B.
  Guindon, and D. Stanley. 2001. GeoComp-n, an
  advanced system for the processing of coarse and
  medium resolution satellite data. Part 2
  biophysical products for the northern ecosystem.
  Canadian Journal of Remote Sensing, vol. 28, no.
  1, pp. 21-44.
• Cihlar, J., Latifovic, R., Chen, J., Trishchenko,
  A., Du, Y., Fedosejevs, G., and Guindon, B. 2002.
  Systematic corrections of AVHRR image composites
  for temporal studies. Remote Sensing of
  Environment (in press).

37
Contact

• Gunar Fedosejevs
• Data Acquisition Division
• Canada Centre for Remote Sensing
• 588 Booth Street
• Ottawa, Ontario, Canada K1A 0Y7
• E-mail Gunar.fedosejevs_at_ccrs.nrcan.gc.ca
• URL http//www.ccrs.nrcan.gc.ca/ccrs/