### Abstract:

Spectral beam attenuation coefficients, spectral volume scattering
function meaurements at 45°, 90°, and 130° particle size distributions
and chlorophyll a and phaeophytin pigment concentrations were measured during
May and September, 1977 in Monterey Bay, California. This data is examined
and statistical interrelationships between the parameters measured are
explored.
It was found that beam attenuation and suspended particulate volume
concentration are highly correlated with the slope of the particle size
distribution, low slopes corresponding to higher particle concentrations.
Correspondingly, the slope of the spectrum of the particulate beam attenuation
decreases with increasing particle concentration. This decrease is not as
much as theoretical values based on hyperbolic size distributions extending
to zero particle size would predict, but such theoretical values are based
on diameters which we cannot measure nor do we expect the hyperbolic
distribution to hold to such small sizes. Changes in the type of
particles present with changing particle concentration also affect the
spectral properties.
Variations in the spectral behavior with location both horizontally
and vertically were found. These variations were partially explained by
corresponding changes in particle size distributions and biological parameters.
The volume scattering functions behaved similarly to the beam attenuation
coefficients except that the particle size distributions were not able to
explain much additional variation in the scattering beyond that explained by
total attenuation and location parameters. Unmeasured parameters such as
particle shape may be more important for scattering than attenuation in
explaining spectral and angular differences between locations.
The particle size distributions and especially the larger particles
were more important than the pigment parameters in predicting certain ratios
of particulate spectral beam attenuation coefficients such as Cp(450)/Cp(500),
Cp(400)/Cp(650) and CP450/Cp(650). The ratio of pigments to suspended
volume was also significant in determining these ratios, and is believed
to be related to the effective index of refraction for the collection of
suspended particles. An exception was the parameter Cp(400)/Cp(450) which
was not predicted well by the slope of the size distribution. Instead, the
pigment-suspended volume ratio and total suspended volume were foremost in
importance followed by the concentration of the smallest measured particles
(1.75-2.5 um diameter) and the ratio of the small particles to medium-sized
particles (6.2-10 um). This particle ratio may extrapolate to the very,
small size particles better than the overall slope and thus be related to
Rayleigh (a-4) scattering which might affect Cp(400)/Cp(450) while the
large particle scattering dominates the other ratios.
Particle size distributions were best predicted by the sum of the
attenuation values and Cp(400)/Cp(650). While several other parameters
removed a statistically significant amount of additional variance, they
did not improve the models for practical purposes, Particles less than
10 um diameter had much smaller original variance (on a logarithmic scale)
than the larger particles and not much variance was removed by the attenuation
parameters. Better correlations would have resulted if an electronically
tripped rosette sampler had been attached to the transmissometer package
instead of relying on separate bottle costs.
The findings of this report were not even consistent between. May and
September and therefore it is highly unlikely that they could be applied to other regions. A better approach would be to confine the area of interest
both vertically and geographically, determine typical particle collections
on which to do laboratory experiments and then use matrix minimization
techniques to decompose measured optical spectra into the predetermined
possible components.