After all these updates to the model I discussed above, I used it to constraint the Titan's haze particles. I will now discuss the results and the procedure that I followed briefly.
The constraints for the size of the aerosol particles mostly come from the forward scattering measurements. The solar aureola (SA) camera direclty measures the forward scattering part of the phase functions in the DISR model for the SA channels. The rest of the phase function obtained from the parameterization model except for the adjustments required by the observations, such as backscattering peak. Therefore, constraints from the forward scattering angles should be more accurate as this portion is not affected from any change in the parameterization model. I still used the backscattering portion of the DISR phase functions when comparing to model outputs, because the best fitting phase functions in the DISR model are obtained by comparing the measured radiances to the radiative transfer model outputs. Also, especially in the red channel the difference in the both version of the model is not significant. However, this time I allowed more flexibility in the acceptable error margin comparing to the forward angles. Consider adding comparison of old and new plot.
Constraints from the phase functions
I did not use the DISR phase functions given in \citep{Tomasko_2008} for the red channel, because of the possibly too strong backward peak artificially added to phase functions for longer wavelengths, at and below 80km altitude. Here in figure 3 some phase functions with or without enhanced backward peak are shown calculated by the T-Matrix model. Although in this figure monomer numbers of the aggregates, N, are smaller than best fitting aggregate models to the observations, the backward peak is not a strong function of N, but rather depend strongly on size parameter, α, and the refractive index. The backward peak is enhanced as the size parameter and real part of the refractive index, nr, increases and as the imaginary part decreases. Although the constraints from single scattering albedo suggest that there is little absorption in Red channel, the size parameter is small in the Red channel, most likely less than 0.35. As shown in figure 3 b. a strong backward peak is not observed in aggregates with 1000 monomers.
Moreover, \citet{de_Bergh_2012} found that DISR phase functions produce too strong intentisities at the some NIR wavelengths comparing to VIMS observations, and removing the backward peak reduces the positive bias from 40% to 10% as shown in fig 11 ( here in fig 5). Thus, unlike \citet{Tomasko_2008} they adopted a single aerosol phase function throughout the atmosphere without a backward peak, to be able to match the VIMS observations. More recently \citet{Doose_2016} modified the original aerosol model in the \citep{Tomasko_2008} by analzying DISR imagery including some measurements which were not used before. Similar to \citet{de_Bergh_2012}, they found that backward peak in the phase functions below 80 km is too strong. As a result they adopted a single phase function for each wavelenght thourughout the atmosphere which is a combination of 85% of the upper-level which does not include backward peak and 15% of the lower level phase function. The new phase functions fit the observations better than the original two-layer phase functions, and has a moderate backward peak which better fits the expectations.
I, therefore, updated the
Since the
The figure 1 shows the best fitting models to the SA observations. These models