Parameterization of the light absorption by components of sea water - - PowerPoint PPT Presentation

Parameterization of the light absorption by components of sea water in the Black sea coastal zone

E.V. Dmitriev, T.Y. Churilova, M. Chami, G. Khomenko, G.A. Berseneva, O.V. Martynov, E.B. Shybanov, M. E-G. Lee, G.K. Korotaev

Light absorption in coastal sea waters

) ( ) ( ) ( ) ( ) ( λ λ λ λ λ

φ NAP CDOM w

a a a a a + + + =

Four-component model: 1) Pure sea water - w 2) phytoplankton - φ 3) colored dissolved organic matter - CDOM 4) nonalgal particles - NAP

) (

) ( ) (

λ φ φ

φ

λ λ

E

Chl A a = ) ( ) ( ) ( λ λ λ

φ NAP p

a a a − =

Parameterizations of the spectral absorption for different sea water components

by phytoplankton by nonalgal particles by colored dissolved

• rganic matter

Measuring of phytoplankton absorption coefficient )) ( (

) ( ˆ ) (

r NAP

S r NAP NAP

e a a

λ λ

λ λ

− −

=

)) ( (

) ( ˆ ) (

r CDOM

S r CDOM CDOM

e a a

λ λ

λ λ

− −

=

The oceanographic platform near Katsiveli

Coordinates 44°23’N and 33°59’E

Observations were made in summer 2002 (from July 27 until August 15)

The platform located 600 meters from the southern Crimea coast in the Black Sea The deck of the platform is 12 meters high from the sea level The bottom depth of the sea increases from 28 to 33 meters in the offshore direction

4 8 12 16 Wind speed (m s-1) Secchi disk depth (m) wind speed secchi disk depth 26.07 30.07 03.08 07.08 11.08 15.08 Date : Day.Month.2002

Date: Day.Month.2002

• 250
• 200
• 150
• 100
• 50

Depth z (m)

Temperature (°C) 26.07 28.07 30.07 01.08 03.08 05.08 07.08 09.08 11.08 13.08 15.08

9 11 13 15 17 19 21 23 25

5 10 15 20 25

Variability of the wind speed and Secchi disk depth near the platform during the experiment. Concentration of phytoplankton pigments in the Black Sea as derived by SeaWiFS satellite sensor (top) on July 26 2002 and (bottom) on August 11 2002. Temporal variation of the vertical profile of the temperature along the experiment.

360 400 440 480 520 560

Wavelength, nm

20 40 60 80 100

Relative contribution (%) CDOM NAP Phytoplankton pigment

The absorption budget reveales that the Crimea coastal waters clearly fall in the case II water type with a yellow substance dominated regime.

360 400 440 480 520 560

Wavelength, nm

20 40 60 80 100

Relative contribution (%) CDOM NAP Phytoplankton pigment

360 400 440 480 520 560

Wavelength, nm

20 40 60 80 100

Relative contribution (%) CDOM NAP Phytoplankton pigment

360 400 440 480 520 560

Wavelength, nm

20 40 60 80 100

Relative contribution (%) CDOM NAP Phytoplankton pigment

Relative contribution of phytoplankton, nonalgal particles and CDOM absorption to the total absorption

August 2 August 3 August 10 August 14

, ) ( ) (

) (λ

λ λ

ph

E ph ph

Chl A a =

1 ) ( *

) ( ) (

−

=

λ

λ λ

ph

E ph ph

Chl A a

Parameterization of the light absorption by particles and phytoplankton [ Bricaud et al, JGR, 1998]

0.1 1 10

Tchla, mg/m3

0.001 0.01 0.1 1

aph(443), m-1 aph(443)=0.0514*Tchla0.566, r2=0.58

Calibration for "CASE I" waters [ Bricaud et al, JGR, 1998] Calibration for "CASE II" waters

• f the Black Sea coastal zone

Parameterization of the specific absorption [ Bricaud et al, JGR, 1998]

, ) ( ) (

) (λ

λ λ

ph

E ph ph

Chl A a =

, ) ln (ln ) ln )(ln ) ( ln ) ( (ln ) ( ), ln ) ( ) ( ln exp( ) (

2

Chl Chl Chl Chl a a E Chl E a A

ph ph ph ph ph ph

− − − = − = λ λ λ λ λ λ

Spectral values of numerical coefficients Aph - top and Eph -bottom defining the parameterization of light absorption by phytoplankton as a function of the sum of chlorophyll a and phaeopigment concentration

Parameterization of light absorption by phytoplankton

where Tests of normality of the distribution of aph(λ). The upper bar - Jarque-Bera test, the middle bar - Lilliefors test, the lower bar - hybrid test. The significance level 0.05.

Bricaud et al 1998 Our fitting

Parameterization of absorption by phytoplankton for different values

• f Tchla concentrations

The spectral distribution

• f the determinate and random errors
• f the absorption parameterization

Spectral distributions of the specific absorption by phytoplankton in different waters [Babin et al, JGR, 2003]

Variations of the spectral absorption at fixed value of Tchla concentration.

Thin grey curves (5 in total) are spectral absorption measured for different water at the chlorophyll concentration 0.65 mg/m3.

The correlation coefficients between the distributions

• f spectral absorption at

fixed value of Tchla

Parameterization of light absorption by nonalgal particles (NAP) and color dissolved organic matter (CDOM)

The exponential fit of the NAP absorption by different methods The root-mean-square approximation errors of DLM (solid grey curve) and NLSM (solid black curve).

August 14, depth 16 m

The example corresponds to the largest error

• f nonlinear least-square fitting method

) (

) ( ˆ ) (

r NAP

S r NAP NAP

e a a

λ λ

λ λ

− −

=

) (

) ( ˆ ) (

r CDOM

S r CDOM CDOM

e a a

λ λ

λ λ

− −

=

The general form (nm) is a reference wavelength

r

λ

Fitting in the spectral region from 380 to 730 nm, excluding the intervals 400-480 and 620-710 nm Fitting in the spectral region from 350 to 700 nm

Scatter plots of - the slope parameter estimated by DLM

• squares and by NLSM - stars, with

respect to the NAP absorption at the wavelength 443 nm. The black solid and dash curves signify the corresponded mean values of . Histograms of the exponential slope parameter of parameterization

• f the NAP absorption spectrum

NAP

S

NAP

S

NAP

S

DLM, 118 samples NLSM, 118 samples

The fitting is done for the spectral region from 380 to 730 nm, excluding the intervals 400-480 and 620-710 nm Average slopes presented in [Babin et al 2003] for different coastal waters range from 0.0116 up to 0.0130

NAP

S

NAP

S

NAP

S

2σ=0.0024 2σ=0.002

Parameterization derived in [Babin et al, JGR, 2003] Parameterization obtained from our measurements in the Black Sea 2002

) 443 ( ) 0024 . 0104 . (

) 443 ( ) (

− ± −

=

λ

λ e a a

NAP NAP

The exponential fit of the CDOM absorption using NLSM

) 443 ( 0179 .

) 443 ( ) (

− −

=

λ

λ e a a

CDOM CDOM The spectral curves are fitted by exponential function in the spectral region from 350 to 500 nm.

) 443 ( 0123 .

) 443 ( 75 . ) (

− −

=

λ

λ e a a

NAP NAP ) 443 ( 0176 .

) 443 ( ) (

− −

=

λ

λ e a a

CDOM CDOM

The presented results show that the Crimea coastal waters fall in the case II water type with a CDOM dominated regime and clearly have their own optical particularities.

Conclusions

• We demonstrated that the parameterization of phytoplankton absorption obtained

from our measurements is significantly different from the published one in [Bricaud et al. 1998] and have proofed that this cannot be explained by poor statistic. Our fitting of Eph reveals a strong minimum at the wavelength 580 nm, which can not be observed in the parameterization of Bricaud at al.

• The absorption by NAP and CDOM can be parameterized well using an

exponential regression. We compared two different techniques of estimating slope parameter: the data linearization method (DLM) and the nonlinear least-square method (NLSM). It is shown that the NLSM is more preferable. The mean value of SNAP obtained by NLSM, is amounted to 0.0104±0.0024, that is less than all average slopes represented in [Babin et al., 2003] for different coastal waters. We would like to underline that the difference between mean values of slope parameters for DLM and NLSM is comparable with the range of the average slopes for different waters.

• 2 of 4 our measurements of CDOM spectral absorption are significantly go out the

uncertainty intervals for the parameterization constructed in [Babin et al., 2003]. Thus it should be calculated more exact employing additional measurements.