Data file with densities of acid, mineral and salt componets vs. depth
scaled for the South Pole, is here.
To generate scattering coefficients using this data (depth dependence)
for the light with the wavelength 532 nm run
dust inputs -L=532 -n -ice -v -zm < table_data > depthTo generate scattering coefficients for some specific depth and a set of wavelengths, run this script:
foreach i ( 190 195 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600 620 640 660 680 700 720 740 760 780 800 )To list all options, run "dust -h", and to see the description of the data sets and outputs, run "dust -usage". Different debug codes do different things. Debug option -d64, however, shows the pre-integrated values of scattering coefficients, particle radius being equal to its average value, for comparison with the "pure Mie theory", or for the applications that do not require integration over the radii distribution.echo Lamda=$i
dust inputs -L=$i -n -ice -v -zm < fixed_depth >> wave_depend
We plotted absorption coefficient a, scattering coefficient b, and g = <cos(theta)> in the region of visible wavelengths (300 nm - 600 nm) for generic particles with refractive index n = 1.55 + 0.005i, and then fitted them with the power law function in the region from 350 nm to 400 nm. The first two plots are are made for a single radius distribution of sizes (no distribution, the only radius present is r_0, or, in other words, sigma=1+0). Here you will find plots of the exponents of the power law. And here is the same plot assuming that the refractive index of the ice stays constant. You can clearly see that the deviation of the exponents from their expected values -1, -4, and -2 at small radii is due to the refractive index wavelength dependence. In the next two plots size distribution sigma is assumed to be 2.03. This plot is for wavelength dependent n, and this one is for constant n.
This file shows wavelength dependence of the absorption coefficient a, scattering coefficient b, and average g = <cos(theta)> in the large range of wavelengths, assuming constant refractive indices for 4 main components: mineral, salt, acid, and soot. These components are taken with their average representative concentrations in the SP ice. In the limit of small wavelengths (geometrical optics) we see that all plots for the scattering coefficient approach some constant value. This value is equal to the sum of difraction and reflection. Difraction per particle is pi*r^2, and reflection plus absorption is another pi*r^2. If the imaginary part of the refractive index is zero, absorption is zero, and, therefore, the total scattering per particle is 2*pi*r^2. In the other case, when the imaginary part of the refractive index is non-zero, absorption can be anything from pi*r^2 to 0, the more opaque the particle is, the bigger the absorption. Therefore, in the case of big absorption we get only pi*r^2 for the scattering coefficient. In any case, the scattering coefficient should lie in the region between pi*r^2 and 2*pi*r^2 times the particle concentration. The table shows these values together with the asymptotic values taken from the plot. Since all curves come to their asymptotic values from above, those curves with zero imaginary refractive index have asymptotic values outside the specified region, even though they are pretty close to it. Here is a plot that shows total a (with ice contribution) and effective b ( b*(1-g) ).
This file shows a plot of a_ice, b_eff and g vs. wavelength in the range from 190 nm to 800 nm for some average component concentrations in the depth range from 1550 to 1900 m.
This
file shows dependence of concentrations of three dust components with
depth, and this
one is produced for a smaller depth span, to only cover the location
of the where three peaks. Depths were obtained by some scaling of the original
Vostok data depths to SP. Because some developments took place since these
plots were produced, depth scaling function has changed significantly,
so that depths, given on these plots are off from their correct values.
I used them, however, to get some average values of peaks
and valleys
by fitting them with gaussians and polynomials. Since third peak in every
plot failed to produce an acceptable fit, i took the maximum value for
this peak from the plot instead of from the fit. The following plots show
the most recent data we have for the concentration of the mineral component
of dust in the antarctic ice: all
depths, only
three peaks. This plot
is my attempt to fit peaks and valleys in this data with polynomials. Since the first peak in the Dome Fuji data
failed to produce an acceptable fit, we used an average of the 5 points near the top, to approximate the peak value.
The table below summarizes the results of this paragraph,
and takes the depths from the locations of the dust bands found from the
AMANDA callibration data. Three peaks are named 1, 3 and 5, and two valleys
are called 2 and 4.
component \ # |
1
|
2
|
3
|
4
|
5
|
depth, m |
1580
|
1680
|
1750
|
1825
|
1880
|
acid, ng/g |
290
|
190
|
280
|
210
|
210
|
mineral, ng/g |
870
|
320
|
500
|
130
|
490
|
salt, ng/g |
420
|
210
|
330
|
240
|
350
|
This plot shows wavelength dependence of absorption and scattering coefficients for all 5 points. And this one shows the depth dependence of these coefficients for 3 wavelengths. An assumption was made that all or some of the acid has gone from being mainly represented by droplets to being concentrated at the boundaries of the ice crystals, where the contribution of it to the light scattering is at least an order of magnitude less than from that predicted by the Mie theory for spherical particles. The next two plots show the wavelength dependence of absorption and scattering coefficients for 5 depths, and the depth dependence of these coefficients for 3 wavelengths, with the assumption that acid component doesn't contribute.
Some relevant plots: