CSR MAGIC Satellite Remote Sensing Products

1
CSR MAGIC Satellite Remote Sensing Products

• Real Time Answers to Real Time Questions

2
Real Time Answers to Real Time Questions

• Why does CSR operate a ground station?
• Increased availability of data supports
  innovations
• Familiarity with the data improves understanding
• Constant stream of data supports evolution of
  applications
• Access to data not otherwise available e.g more
  AVHRR
• Some satellite data not available from other
  sources, e.g. FY1-D and Oceansat
• More timely access to data (i.e. We need it now!)

3
Pass Summary for 2005/04/13

• 39 Satellite passes in view of antennas
• 5/4 Day/Night AVHRR Austin Views
• 1/1 Day/Night MVISR Austin Views
• 2/2 Day/Night MODIS Austin Views
• 1 Daytime OCM Austin View

4
S/L Band Products

• AVHRR
• Day bands 0.63µm(1), .91µm(2) and inverse
  11µm(4)
• Night inverse of 3.7µm(3), 11µm(4) and 12µm(5)
• NDVI
• FY-1
• Day 0.63µm(1), 0.55µm(9) and 0.45µm(7)
• Night Same as AVHRR inverse 3.7µm(3), 11µm(4)
  and 12µm(5)

5
AVHRR DAY/NIGHT Samples
6
MVISR DAY/NIGHT Samples
7
X-Band Products

• MODIS
• Fire Mask Product
• NDVI Product
• 3-band Image TCEQ Subsets
• Modis Aerosol Product
• NAIP Subsets
• Ocean Color Products
• OCM
• NAIP Subsets
• Ocean Color Products

8
MODIS Institutional Algorithms

• A Standard Foundation
• Initial experience with MODIS began with products
  downloaded from NASA EDG
• Allowed previously developed tools to be applied
• Allows DB data and EDG data to be interchanged
• Generated Base Products
• Geolocation File MOD03
• L1b MOD02
• Cloud Mask MOD35
• MOD04 Aerosol
• MOD14_RR Fire Mask

9
MODIS Fire Mask

• Updated algorithm from Rapid Response Team
  ver.4.3.2
• Tabulated Fire Detections
• Mapped Fire Locations
• Validation Concerns for Modelling

10
Sample Firemask Image
11
MODIS NDVI Product

• Generated from Surface Corrected Reflectances
• Coded NDVI and EVI equations into emathp
• Generate Masked and Unmasked products
• Building Weekly Composites

12
NDVI EVI samples
13
NDVI Daily vs. Weekly Comparison
14
3-Band Subsets

• True Color TCEQ Subsets
• Based on Creating Reprojected True Color MODIS
  Images A Tutorial by Liam Gumley and Jacques
  Descloitres
• psuedo Surface Corrected Reflectance
• Reprojection using UTCSR legacy tools
• Band sharpening based on Red band
• 16-bit to 8-bit mapping using piecewise linear
  LUTs
• Overlays and Rendering with Terascan
• Thermal Scenes
• Selected bands nearest AVHRR bands
• Inverted histogram equalized enhancements

15
L1B to Surface Corrected Comparison
16
Unsharpened to Sharpened Comparison
17
Piecewise Linear LUTs
18
Equalized vs. Piecewise LUT
19
TCEQ 250m Subsets
20
250m Austin/San Antonio Subset
21
250m Corpus Christi Subset
22
MODIS Thermal
23
MODIS Aerosol Product

• Modis Aerosol Products
• Image product
• Text tabular product for CAMS sites
• Issues with elevated values

24
True Color and Aerosol Image
25
Aerosol Image with CAMS Stations
26
AOT Product Issues
27
Future Interests

• L-Band products
• Adapt Surface Corrected Reflectance algorithm to
  MVISR
• Coriolis?
• X-Band products
• Improve fire inventory capability
• Reconstructed cloud free NDVI
• Departure-from-normal NDVI
• Elevated AOT product, other AOT enhancements
• Forecast AOT trajectories

28

• NAIP Flight Weather
• and
• Hurricane Tracker

29
NAIP Background

• NAIP National Agriculture Imaging Program
• UTCSR provides near-real-time (NRT) satellite
  image data for supplementing the decision making
  of data collection.
• Full-resolution images are delivered to the
  website within five minutes after acquired and
  processed by TeraScan ingestion software.

30
NAIP Details

• Using JPEG to compress the full-resolution
  true-color image for faster transmission over the
  Internet.
• Each JPEG is associated with a World File for
  defining the geolocation used in GIS software.
• Uncalibrated MODIS and other image data are
  delivered in five minutes after acquisition.
• Update with the DAAC calibrated MODIS L1B image
  when available for better image quality.

31
HurricaneTracker Background

• Meteorological satellites monitor the movement of
  hurricanes constantly but with coarse resolution.
  (ex GOES 1km)
• With NRT high resolution data, the emergency
  management authority can localize the state of
  hurricanes more precisely.

32
HurricaneTracker Details

• Receive the predicted and historical trajectories
  of the hurricane from National Hurricane Center
  (NHC).
• Determine the region of interest based on
  prediction from NHC when telemetry is acquired.
• Subset and ingest the telemetry, and generate
  true-color image.
• Deliver to the web interface to report current
  status of the hurricane.

33
Hurricane Ivan, 2004
34
HurricaneTracker Results

• Using a subset of the whole telemetry centered at
  the hurricane, the size of data and image has
  been more manageable for rapid response.
• The highest resolution of the image is 250m from
  MODIS which has been a reliable source for
  emergency management.
• Always guarantee the hurricane center is captured.

35
Ocean Color Products

• Using SeaDAS

36
Facts about SeaDAS

• SeaDAS is a data analysis system from Goddards
  Space Flight Center (GSFC).
• With the experiences from SeaWiFS research,
  GSFCs OceanColor team becomes a main source of
  science support for MODIS ocean products,
  especially Aqua.
• SeaDAS can be user-friendly by exploiting IDL
  GUI.
• SeaDAS is a single package for multiple sensors
  to process ocean color products.

37
Using SeaDAS with TeraScan

• SeaDAS supports MODIS, SeaWiFS, CZCS, OCTS, MOS,
  OSMI, and AVHRR.
• TeraScan systems is flexible to use command line
  interface for scripting.
• UTCSR glues SeaDAS packages as an
  alternative/comparison to the ocean products
  generated from TeraScan.
• SeaDAS has been successfully installed on Xserve
  G5 cluster for testing its portability.

38
MODIS Ocean Color Products

• 305 Level 2 products are available from SeaDAS.
• Operational ocean products
• Chlorophyll concentration-a using OC4 algorithm
• Diffuse attenuation coeff. At 490 nm (K_490)
• Aerosol optical depth at 865 nm (tau_865)
• Epsilon (eps_78)
• Aerosol angstrom coeff at 510 nm
• Normalized water-leaving radiances.
• Level 2 processing flags.
• SeaDAS 4.7 for MODIS processing has some banding
  issue.

39
Shape Mask

• Importing Shapefiles to TeraScan

40
Challenge of Inland Water

• Algorithms are complicated and empirical.
• Need calibration on site.
• Require higher accuracy of image registration.
• Need detailed land/sea mask to remove unwanted
  data.

41
Shape Mask

• Shapefile ESRIs vector data format used in
  ArcView. A common format used by GIS community.
• Many accurate shapefiles around the world are
  available.
• Convert regular shape file to TeraScans coast
  lines and land/sea mask.
• Result in better geolocation if the quality of
  the shapefile is accurate.
• Improve the quality of navigation in TeraScan for
  local area.
• Suitable for local/small area mapping.

42
Comparison
Aqua MODIS 2005-04-06 2005 UTM DCW - Pink Texas
Reservoirs - Orange
43
DCW - Pink Texas Reservoirs - Orange
44
END
45
EXTRA Slides
46
Data Repository

• Developed because of need
• No central DAAC for OCM or FY-1
• Improve data reliability
• Simple FTP based deployment
• Uses standard software tools
• Sets of client and server shell scripts developed
• Allow interactive or script based interactions
• 1-week Cache with deep archive support
• Ongoing Development
• Data merging to reduce our storage requirement
• Individual frame requests
• More user friendly interface

47
Map of CAMS stations Tabular subset
48
Map of CAMS stations
49
AOT Product Issue Zoom Louisiana