1 / 13

Inverse modeling

Inverse modeling. Semi-analytical algorithms : sensitivity analysis of the values of the backscattering spectral dependency n and the a_nap&cdom slope S on two stations Ma é va DORON, Ocean Optics 2004. To begin with. Roesler and Perry ’95 model of inverse modeling

erek
Download Presentation

Inverse modeling

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. Inverse modeling Semi-analytical algorithms : sensitivity analysis of the values of the backscattering spectral dependency n and the a_nap&cdom slope S on two stations Maéva DORON, Ocean Optics 2004

  2. To begin with • Roesler and Perry ’95 model of inverse modeling • 3 basis vectors in input: • spectral slope of the absorption of NAP and CDOM S (DL); • spectral slope of backscattering n in nm-1; • spectral absorption of chlorophyll (normalized). • 3 parameters in output : • chl in mg.m-3; • a_cdom&nap (440nm) in m-1; • bb(440nm) in m-1. • Data from the second cruise: one station off the DMC dock and one station near the ocean.

  3. Goal • First idea • Comparison of the data retrieved by the GSM model and the R&P95 model (hyperspectral & multispectral at 412, 442, 490, 510 and 555 nm); • Sensitivity to S and n to the retrieved values • Comparison of the sensitivities for the hyperspectral and the multispectral model

  4. Original values Station A – Off of DMC Dock

  5. Original valuesStation B – SW of Pemaquid Pt.

  6. Sensitivity • One spectral dependency n ,instead of 0 for large particles and 1 for small particles. • Varying the n – values : from 0 to 2. • Change the values of S : from 0.005 to 0.04 nm-1 : Realistic range of Scdom = 0.007–0.013nm-1 Realistic range of Snap = 0.016–0.022nm-1

  7. Sensitivity for the n(bb) slope - hyperspectral

  8. Sensitivity for the S(cdom&nap) slope - hyperspectral

  9. Sensitivity for the n(bb) dependency - multispectral

  10. Sensitivity for the S(cdom&nap) slope - multispectral

  11. Sensitivity for the S(cdom&nap) slope – multispectral- different scales

  12. Discussion • This was a first qualitative approach. • The results are encouraging because for realistic values of S are not erratic. • Any idea about how to test the robustness of the algorithm with multispectral data?

More Related