1 / 39

GOES-R AWG Imagery Team: Cloud and Moisture Imagery August 27, 2009

GOES-R AWG Imagery Team: Cloud and Moisture Imagery August 27, 2009. Presented By: Tim Schmit NOAA/NESDIS/STAR. Outline. Runoff / Selection Environment Description of Selected Algorithm Validation of Selected Algorithm Summary of IV & V Feedback. Project Objectives (1).

conwayj
Download Presentation

GOES-R AWG Imagery Team: Cloud and Moisture Imagery August 27, 2009

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. GOES-R AWG Imagery Team: Cloud and Moisture Imagery August 27, 2009 Presented By: Tim Schmit NOAA/NESDIS/STAR

  2. Outline • Runoff / Selection Environment • Description of Selected Algorithm • Validation of Selected Algorithm • Summary of IV & V Feedback

  3. Project Objectives (1) • Develop algorithms and deliver Algorithm Theoretical Basis Documents (ATBDs) to the GOES-R Ground Segment Project (GSP) • Cloud and Moisture Imagery for ABI • 54 F&PS Requirements • 16 ABI channels multiplied by 3 coverage areas • Multi-band spectral products (16 channels at 2 km resolution) by 3 coverage areas in NetCDF4 format • Multi-band spectral products (16 channels at 2 km resolution) by 3 coverage areas in McIDAS format

  4. Cloud and Moisture Imagery MRD Requirement • Cloud and Moisture Imagery reports digital maps of clouds, moisture, and atmospheric windows through which land and water are observed, by reporting radiance measurements converted first to brightness temperature and then digital counts from 0-255 from all of the bands sensing clouds and moisture from an imaging instrument. Infrared imagery bands are often chosen either along spectral absorption features including those of water vapor bands or CO2 and in regions with no absorption that permit observations of the surface. Visible bands are also chosen to sense the surface and the low lying cloud and fog interfering with observations of the surface. Low light imagery in the visible band is also included. Cloud and moisture imagery provides input to other algorithms producing other environmental products (same as CONUS product except this version provides mesoscale coverage).

  5. Requirements – Cloud and Moisture Imagery C – CONUS FD – FullDisk M - Mesoscale

  6. Cloud and Moisture ImageryProduct Qualifiers C – CONUS FD – FullDisk M - Mesoscale

  7. Cloud and Moisture Imagery Product (CMIP) Algorithm Objectives • Provides state-of-the-art Imagery products over the GOES-R observation domain, in all ABI scanning modes • Single time step • Meets the GOES-R mission requirement specified for the Imagery products • Algorithm is based upon heritage techniques to minimize changes for current user base. • Computational efficiency

  8. Runoff / Selection Environment: Candidate Algorithm Selection Given this is a legacy-type product, no ‘run-off’ per se was generated. Following procedures consistent with current image generation. The understanding for current GOES processing steps were the guide: Heritage from current GOES Lessons learned from other imagery processing There are different methodologies for the Planck conversion from radiance to brightness temperature (for the infrared bands): Adopted the framework version 8 8 8 8

  9. Outline • Runoff / Selection Environment • Description of Selected Algorithm • Validation of Selected Algorithm • Summary of IV & V Feedback

  10. Description of Selected Algorithm: Overview • The purpose of the imagery team is two-fold: • Demonstrate how to convert from GRB scaled integers to other physical units, such as radiance, brightness temperatures and brightness values. • Build files that can be used for processing most all of the ABI products, such as clouds, soundings, etc. • Imagery is the one and only key product for GOES-R. • Algorithm Development Status • Code delivered in late 2008 and 2009 • Draft ATBD delivered in September 2008 • Proxy & simulated products tested: simulated ABI • Software has been integrated into AIT frame work • TRR in July 2009 (used simulated ABI data)

  11. Cloud and Moisture Imagery Algorithm • Algorithms steps: • Produce Scaling Coefficients • Define the conversion from reflectance/radiance to Scaled Integer (SI) for use in building the GRB data stream • Factors should be time invariant across satellites • Products Generated • Un-scale the ABI SI (Scaled Integers) data in the GRB to reflectance factor or radiance • Define the process for converting radiance to Brightness Temperature (BT) for IR bands. • Down-scale the higher-resolution bands for the required multi-band files. • Define the process for converting to brightness values (BV) for all bands. • BV would (at least) be 8-bit, but could support higher bit depths

  12. Cloud and Moisture Imagery Algorithm • Convert ABI scaled reflectance from the GOES ReBroadcast (GRB) data stream (L1B) into reflectance factor for bands 1-6; convert GRB scaled radiance (L1B) to spectral radiance and then to BT for bands 7-16. • Linear Stretch • Bi-Linear Stretch • assumed for IR bands • BT < 242K, BV = 418 –TBB; BT >= 242K, BV = 660 – (2*TBB) • Logarithmic Stretch • Power Law Stretch • Square-Root Stretch (RF of 0-1) • Square Stretch • Exponential Stretch • Generate CMIP in “classic” NetCDF (V4) and McIDAS AREAs. Convert CMIP into other formats, e.g. HDF. NetCDF is Climate and Forecasts conventions compliant. • Extract geolocation information from the input data stream

  13. Cloud and Moisture Imagery Algorithm • Bi-Linear Stretch (bands 7-16) • Advantages • Covers more temperature range than a 1K/1Count strech • Heritage (McIDAS, etc.) • Efficient • Algorithm is mature • Disadvantages • Not as straight forward as a linear stretch. • Square-Root Stretch (bands 1-6) • Advantages • Highlights important features that would have been too dark • Heritage (AWIPS, McIDAS, etc.) • Low risks • Efficient • Algorithm is mature • Disadvantages • Not as straight forward as a linear stretch. Note: Users can still employ their own enhancements to the data values.

  14. Reflective bands Emissive (IR) bands ABI Band Characteristics

  15. The Advanced Baseline Imager: ABI Current Spectral Coverage 16 bands 5 bands Spatial resolution 0.64 mm Visible 0.5 km Approx. 1 km Other Visible/near-IR 1.0 km n/a Bands (>2 mm) 2 km Approx. 4 km Spatial coverage Full disk 4 per hour Scheduled (3 hrly) CONUS 12 per hour ~4 per hour Mesoscale Every 30 sec n/a

  16. 0.47 m 0.64 m 0.86 m 1.38 m AWG Proxy ABI Simulations of Hurricane Katrina 1.61 m 2.26 m 3.9 m 6.19 m NOAA/NESDIS STAR and GOES-R Imagery Team 6.95 m 7.34 m 8.5 m 9.61 m 10.35 m 11.2 m 12.3 m 13.3 m

  17. Corresponding current Imager bands NOAA/NESDIS STAR

  18. “Information Volume” Improved attributes with the Future GOES Imagers “Information volume”

  19. Processing Outline For AREAs For Images “digitize” “Output” netCDF (w/ SI, un-scaling coefficients, Planck coeffs, etc.) CMIP “Input” netCDF (w/ SI, etc.) Original Proxy For All Products

  20. Down Scaling Done by Re-sampling • Building 2 km files from higher spatial resolution data • Several ABI bands (1,2,3 and 5) will need to be ‘down-scaled’ for the multi-band ‘2 km’ files. • It is assumed that the data for these multi-band files will be sub-sampled. • quicker • traceable, less ‘blurry’ • doesn’t create artificial values • If the scene has a bi-modal distribution • straight forward (due to resolutions nested in one another) • Advanced (quantitative) users, and users generating products, should use the full resolution (single band) files.

  21. Imagery Unit Description – Top Level 21 21

  22. Teff:effective temperature C1: 1.191066 x 10-5 [mW/(m2.sr.cm-4)] C2: 1.438833 (K/cm-1) v:central wave number of the band T: BT α,β: depend on band and detector Teff = C2 ν ln(C1v3/ Lλ + 1) T = βTeff + α Algorithm Description Radiometric Conversion • Algorithms • Radiometric conversion • Convert spectral radiance to BT (bands 7-16)

  23. Algorithm Description Radiometric Conversion A parallel method converts radiance (mW/(m2.sr.cm-1) to BT (K) or vice versa. fk1 = C1ν3 fk2 = C2ν bc1 = -α/β bc2 = 1/β Draft Table of Planck coefficients (ftp://ftp.ssec.wisc.edu/ABI/SRF/)

  24. Algorithm Description Radiometric Conversion Another implementation of this method converts radiance (mW/(m2.sr.cm-1) to BT (K). This computes FK1/2 internally, using the area-weighted central wavenumber. C1: 1.191066 x 10-5 [mW/(m2.sr.cm-4)] C2: 1.438833 (K/cm-1) Equation version: FK1 = C1 * WaveNumber ** 3 FK2 = C2 * WaveNumber Var_Tmpy = log(1.0 + FK1 / Radiance) Var_Tmpy = FK2 / Var_Tmpy Brt_Temp = ( Var_Tmpy - BC1) / BC2

  25. Assumptions for Imagery • Assumptions: • Reflective bands units: reflectance factor • Similar to MODIS/VIIRS/GOES uses • With mean Sun-Earth distance adjustment • ABI calibration is reflectance-based • Emissive bands units: mW/(m2 ster cm-1) • BT can be derived from the radiance values • Navigation information will be calculated from the projection information • From the “Fixed Grid” geo projection • Normalized Geostationary Projection • Each ‘swath’ is also in this ‘perfect’ projection

  26. Assumptions for Imagery • Include version number (say of the Planck coefficients, instrument number, etc.) in the data stream • Include Quality Control index • Zero count is a ‘missing’ value. • Assuming bit depth of 12 and 14 bits per pixels for the data • E-mail from L. Rokke on 7/25/2008... • “… to 12 bits for the 064 band, and 14 bits for all other bands …” • Assuming 8-bit display for BV calculations

  27. Assumptions for Imagery • Only one SRF (instrument Spectral Response Function) per band • On current GOES there is one SRF per detector • The maximum radiance value for scaling does not need to change from instrument to instrument. • Each output netCDF is CF compliant • http://cf-pcmdi.llnl.gov/ • Imagery team is not providing angles for every users, only Fixed Grid information.

  28. Outline • Runoff / Selection Environment • Description of Selected Algorithm • Validation of Selected Algorithm • Summary of IV & V Feedback

  29. Validation Imagery does not have traditional ‘truth’ datasets for comparison, such as radiosondes or aircraft data. In light of this, we have defined our own ‘truth’ data via high resolution NWP runs, coupled with advanced forward modeling. “digitize” “Output” netCDF (w/ SI, un-scaling coefficients, Planck coeffs, E_sun, etc.) CMIP “Input” netCDF (w/ SI, etc.) Original Proxy “off-line” “Validation” netCDF. Un-scale from SI, reflectance/radiance and Brightness Temperature (TBB) and Brightness Values (BV). Radiances.

  30. Simulated ABI bands in McIDAS area McIDAS-X

  31. Simulated ABI bands in NetCDF McIDAS-V

  32. Simulated ABI band 8 (6.2 um) proposed GOES-R sub-point location McIDAS-V

  33. Test Readiness Review: Imagery – Software Verification (bands 1-6) Difference of (AIT-UW) These small differences for the various parameters demonstrate that the conversion processes can be reproduced via the cloud and moisture imagery team processing.

  34. Test Readiness Review: Imagery – Software Verification (bands 7-16) Difference of (AIT-UW) As of July 10, 2009

  35. Outline • Runoff / Selection Environment • Description of Selected Algorithm • Validation of Selected Algorithm • Summary of IV & V Feedback

  36. Summary of IV & V Feedback • The imagery team was one of the last teams to form • Imagery Kickoff Meeting – 06/26/08 • The 80% imagery ATBD was scheduled to be completed in September 2009. • Posted on the STAR page on August 25th, 2009 • In the meantime, informal feedback will be addressed during this review.

  37. Summary of IV & V Feedback (A) • Feedback: Informally, discussions were received if enough “pixel” time information was going to be available. • Response: The imagery team can only pass on the time information in the GRB (GOES-R Re-broadcast) data stream. It is expected, similar to current GOES, that detailed time stamp information will be broadcast with each ‘swath’.

  38. Summary of IV & V Feedback (A) • Feedback:It is expected that the issue of scaling for the ABI bands 1-6 from radiances or reflectance factor will be raised. • Response: The imagery team recommends the GRB (GOES-R Re-broadcast) should be scaled from reflectance factor, although including additional information for users that need the data to be in radiance space. The rationale is for a more accurate product and less processing. A white paper on this issue is being prepared by the GOES-R calibration working group.

  39. Open Discussion • The time is now open for discussion

More Related