IOP inversion from shallow waters

1
IOP inversion from shallow waters
Peng Wang
2
Rrs f(a, bb)
Lu(0-) upwelling radiance Ed(0) downwelling
irradiance
3
PBR approach

• Basis vectors
• absorption
• af(?) af(?0) Sfamicro(?0)(1-Sf)apico(?0)
• adg(?) adg(?0) exp(-S(?-?0))
• backscattering
• bbp(?) bbp(?0) (?/?0)-Y
• Radiance Reflectance equation 40010650
• Rrs 0.0949( bb/(bba)) 0.0794 (bb/(bba))2
• Linear regression method
• Data set simulated by hydrolight (Rrs, a and bb)

4
Some results IOP comparison (a and bb)
Total Absorption Comparison
average of relative difference 11.9
5
total backscattering comparison
average of relative difference 14.2
6
Now story changed
7
Rrs from shallow waters
Strange rrs
Matlab complain
No solutions !!!
8
Ed
Lu
Ludp
LuB
Rrs Lu/Ed (Ludp LuB)/Ed Ludp/Ed
LuB/Ed Rrsdp RrsB
downwelling irradiance, upwelling radiance from
water column and bottom
9
simple idea, hard application fortunately

• Semi-Analytical Hyperspectral Model of Lee, et
  al, 1998

basically rrs rrsdp1-exp(-2KH)
rrsBexp(-2KH) rrsdp(1-A0exp-(1/cos?w)D0(1D1
u)0.5aH) A1?exp-(1/cos?w)D0(1D1u)0.5
aH. rrs subsurface remote-sensing reflectance,
sr-1 rrsdpsubsurface remote-sensing reflectance
for deep waters, sr-1 rrsB subsurface
remote-sensing reflectance for the bottom, sr-1
K diffuse attenuation, m-1 H bottom depth,
m ?wsubsurface solar zenith angle, rad u bb/(a
bb) a attenuation coefficient (abb), m-1
? bottom albedo A0,1 D0,1 D0,1 from Lee et
al, 1998

10
After subtracting the bottom influence, we get
Now matlab smiled and we got solutions !!!
11
Coefficient of Variance(express sample
variability relative to the mean of the sample)

• Total
• absorption
• Total
• Backscattering

12
IOP inversion results from shallow waters

• Total
• Absorption

Total Backscattering
13
Conclusions

• Bottom reflectance has a huge impact on the
  remote sensing reflectance
• Current semi-analytic algorithm can be
  successfully applied to invert IOPs after bottom
  correction
• PBR approach can find strange rrs which is caused
  by the environment or bad measurements?

14
Acknowledgements