1 / 12

IOP Algorithm Workshop @ OOXIX Jeremy Werdell NASA Ocean Biology Processing Group 7 Oct 2008

IOP Algorithm Workshop @ OOXIX Jeremy Werdell NASA Ocean Biology Processing Group 7 Oct 2008. attendees: Antoine Mangin (ACRI) Odile Hembise Fanton d’Andon (ACRI) Bryan Franz (NASA) Paula Bontempi (NASA) Catherine Brown (LOV) Samantha Lavender (U. Plymouth)

evelina
Download Presentation

IOP Algorithm Workshop @ OOXIX Jeremy Werdell NASA Ocean Biology Processing Group 7 Oct 2008

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. IOP Algorithm Workshop @ OOXIX Jeremy Werdell NASA Ocean Biology Processing Group 7 Oct 2008 IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

  2. attendees: Antoine Mangin (ACRI) Odile Hembise Fanton d’Andon (ACRI) Bryan Franz (NASA) Paula Bontempi (NASA) Catherine Brown (LOV) Samantha Lavender (U. Plymouth) Emmanuel Boss (U. Maine) Sean Bailey (NASA) Gene Feldman (NASA) Stephane Maritorena (UCSB) Hubert Loisel (U. Littoral) Takafumi Hirata (PML) Jeremy Werdell (NASA) Tim Moore (NURC) Jill Schwarz (NIWA) Tim Smyth (PML) Mark Dowell (JRC) Vittorio Brando (CSIRO) Mike Behrenfeld (OSU) Yannick Huot (LOV) ZhongPing Lee (MSU) unable to attend: Andre Morel (LOV), Paul Lyon (NRL) IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

  3. algorithm “shoot out” workshop motivation & goal: achieve community consensus on an effective algorithmic approach for producing global-scale, remotely sensed SAA IOP products what we attempted to do: extend the IOCCG SAA survey by evaluating application of SAA algorithms to satellite radiometry reviewing & consolidating SAA construction desirable features: combination of accuracy and geographic coverage flexible, multi-sensor implementation computational efficiency to support operational environment open source software and accompanying LUTs associated SAA uncertainties SAA = semi-analytical algorithm IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

  4. pre-workshop achievements (Mar - Sep) - dialog & discussion air-sea transmission, Rrs  rrs(0-) calculation of Rrs (bandpass correction, f/Q) temperature & salinity dependence of aw & bbw spectral data products to be considered (adg, bb, etc.) evaluation metrics & SAA failure conditions inversion methods & linearization issues calculation of uncertainties SAA product validation & sensitivity analyses strategies to produce level-3 products http://oceancolor.gsfc.nasa.gov/forum/oceancolor/board_show.pl?bid=24 IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

  5. pre-workshop achievements (Mar - Sep) - analyses in situ-to-in situ & satellite-to-in situ match-ups global (level-3) comparisons spatial coverage (level-2) comparisions sensitivities to parameterization & noisy input sensitivity to inversion method level-2 vs. level-3 inversion http://oceancolor.gsfc.nasa.gov/MEETINGS/OOXIX/IOP/analyses.html IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

  6. total a and bb are sums of coefficients for all components in seawater each coefficient expressed as product of magnitude and spectral shape construction (& deconstruction) of an SAA … satellite provides Rrs() a () and bb () are desired products IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

  7. correlated Bulk Inversion: * no predefined shapes * piece-wise solution: bbp(), then a(), via (empirical) Kd () via RTE * ex: LS01 1 3 Spectral Optimization: * define shape functions for (e.g.) bbp(), adg(), aph() * solution via L-M, matrix inversion, etc. * ex: RP95, HL96, GSM 2 uncorrelated Spectral Deconvolution: * partially define shape functions for bbp(), adg() * piece-wise solution: bbp(), then a(), then adg() + aph() * ex: QAA, PML, NIWA construction (& deconstruction) of an SAA … satellite provides Rrs() a () and bb () are desired products IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

  8. consensus to refine spectral optimization to initiate process … our STARTING point: * dynamic bbp retrieval * dynamic aph spectral model * IOP-based f/Q tables * Raman scattering * fluorescence * T/S dependence on aw & bbw * optical water class parameterization * uncertainties & propagation of error metrics defined to evaluate progress Spectral Optimization: * define shape functions for (e.g.) bbp(), adg(), aph() * optimization via L-M IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

  9. discussion of uncertainties & their calculation: Wang et al. 2005 GlobColour Lee et al. 2008 (OOXIX personal communication) uncertainties associated with: * input Rrs * models & shape functions IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

  10. specify sensor wavelengths to fit e.g., 412,443,490,510,555 e.g., 412,490,555 select aph form and set params tabulated: , ap*() gaussian: ,  select adg form and set params exponential: , S select bbp form and set params power law: ,  power law: ,  via Hoge & Lyon power law: ,  via QAA select Rrs[0-] to bb/(a+bb) quadratic: g1, g2 f/Q: (tbd) specify inversion method Levenburg-Marquart Amoeba (downhill simplex) Lower-Upper Decomposition Singular-Value Decomposition specify output products a (), aph (), adg (), bb (), bbp () = any sensor wavelength(s) Ca (given ap* at )  (dynamic model params) internal flags generalized IOP model (GIOP) in l2gen IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

  11. summary: consensus was reached on the way forward NASA will implement the GIOP w/i next 3-6 months & begin producing global time-series of IOPs for all missions for which we’re responsible the group will continue our dialog, review results of data processing, & make recommendations for improvements NASA will reintroduce refinements & reprocess the data once we have agreement that products are as good as (currently) possible, full mission reprocessing(s) will be initiated all code will be available via SeaDAS NASA will implement code for optical water class mapping & evaluate how to implement this with class-based SAA parameterization IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

  12. http://oceancolor.gsfc.nasa.gov/MEETINGS/OOXIX/IOP http://oceancolor.gsfc.nasa.gov/forum/oceancolor/forum_show.pl IOP Algorithm Workshop @ Ocean Optics XIX, 7 Oct 2008, PJW NASA/SSAI

More Related