User Tools

Site Tools


intercomparisons:c3_cumulus

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Next revision
Previous revision
intercomparisons:c3_cumulus [2016/11/16 16:28]
127.0.0.1 external edit
intercomparisons:c3_cumulus [2018/01/08 12:23] (current)
Line 3: Line 3:
 //Setup:// //Setup://
  
-  ​* Cumulus cloud field from LES model (same as used in I3RC project, $r_{eff}$ depending on LWC):+{{ :​intercomparisons:​phase_c:​c3_cumulus_cloud:​cumulus.png?​400 |Definition of cumulus cloudfield}} 
 + 
 +  ​* Cumulus cloud field from LES model (same as used in I3RC project, $r_{\rm eff}$ depending on LWC):
    ​{{:​intercomparisons:​phase_c:​cumulus.dat}},​ the format of the ascii file is as follows: ​    ​{{:​intercomparisons:​phase_c:​cumulus.dat}},​ the format of the ascii file is as follows: ​
       Nx Ny Nz flag       Nx Ny Nz flag
Line 69: Line 71:
   # case theta_0 z theta phi i_x i_y I Q U V Istd Qstd Ustd Vstd   # case theta_0 z theta phi i_x i_y I Q U V Istd Qstd Ustd Vstd
  
-==== Results for pure Rayleigh background atmospehre  ​====+==== Data of model results ​====
  
 +{{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​iprt_case_c3_mystic.dat.gz|Results of MYSTIC model.}}
 +
 +{{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​results_all_models_c3.tar.gz|Results of all models (MYSTIC, SHDOM, SPARTA, 3DMCPOL, MSCART).}}
 +
 +==== Results for pure Rayleigh background atmosphere ​ ====
 +
 +== MYSTIC results ==
 The following plots show the results obtained with MYSTIC. For all other models the patterns look the same.  The following plots show the results obtained with MYSTIC. For all other models the patterns look the same. 
  
 {{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​c3_mystic_noaer.png|C3 without aerosol, MYSTIC results}}  {{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​c3_mystic_noaer.png|C3 without aerosol, MYSTIC results}}
  
-** General **+** General ​remarks**
  
-  * Generally we find a very good agreement between all Monte Carlo models, only for a few cases they do not agree within 2 standard deviations.+  * Generally we find a very good agreement between all Monte Carlo models
   * The polarization patterns are the same in all models, also the pattern for V when variance reduction methods are applied   * The polarization patterns are the same in all models, also the pattern for V when variance reduction methods are applied
   * For SHDOM a quantitative comparison is difficult because it does not provide a standard deviation and systematic differences are expected (different treatment of cloud boundary. Smaller differences due to different solution method of VRTE).   * For SHDOM a quantitative comparison is difficult because it does not provide a standard deviation and systematic differences are expected (different treatment of cloud boundary. Smaller differences due to different solution method of VRTE).
Line 83: Line 92:
   * Circular polarization (Stokes component V) is very noisy in Monte Carlo results.   * Circular polarization (Stokes component V) is very noisy in Monte Carlo results.
  
-The following plots show the mean radiance (for each Stokes component), mean absolute difference with respect to MYSTIC (forward Monte Carlo), mean standard deviation, fraction of pixels which agree within 2 standard deviations (2σ), please note the non-linear scale of the colorbar.  ​ +== Statistics == 
-{{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​c3_statistics_1.png|Summary plots, statistics C3 without atmosphere}} +{{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​c3_statistics_rms_std_match_1_nstokes4.png|Summary plots, statistics C3 without atmosphere}} 
-{{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​c3_statistics_1.pdf|Summary plots, statistics ​C3 without ​atmosphere}}+Statistics of the Stokes vector results for scenario ​C3 (cumulus cloud 
 +field in molecular ​atmosphere).  The panels in the left column show 
 +the mean radiance $I_{\rm mean}$ (for Q, U, and V the mean of the 
 +absolute values) for all models and all 9 cases. The panels in the 
 +second column show the standard deviations $\sigma_{\rm rel}$. The 
 +third column shows the root mean square differences $\Delta_{\rm RMS}$ 
 +in per cent and the right column shows the match fractions $q$.
  
 +** Notes: **
   * Stokes component I (upper row): Obvious differences in mean radiances only for SHDOM. The mean standard deviation is almost the same for 3DMCPOL and SPARTA, both models use the same number of photons and no variance reduction. The standard deviation of MSCART is smallest, so for this case it seems that MSCART variance reduction works best. The match fraction shows that all Monte Carlo codes agree.   * Stokes component I (upper row): Obvious differences in mean radiances only for SHDOM. The mean standard deviation is almost the same for 3DMCPOL and SPARTA, both models use the same number of photons and no variance reduction. The standard deviation of MSCART is smallest, so for this case it seems that MSCART variance reduction works best. The match fraction shows that all Monte Carlo codes agree.
   * Stokes component Q (second row): Results are similar to I, however we find systematic differences for SPARTA (cases 2 and 3)    * Stokes component Q (second row): Results are similar to I, however we find systematic differences for SPARTA (cases 2 and 3) 
   * Stokes component U (third row): Also for U, there are systematic differences between SPARTA and other Monte Carlo models for cases 2 and 3.    * Stokes component U (third row): Also for U, there are systematic differences between SPARTA and other Monte Carlo models for cases 2 and 3. 
   * Stokes component V (bottom row): Mean radiance for V is dominated by noise, so here we can not compare the models quantitatively.   * Stokes component V (bottom row): Mean radiance for V is dominated by noise, so here we can not compare the models quantitatively.
-  ​* Generally the standard deviation for SPARTA is below 3DMCPOL because for SPARTA more photons were used (10e11 vs. 49e9). MYSTIC and MSCART used 49e9 photons and variance reduction techniques which decrease the standard deviation approximately by almost one order of magnitude.+  ​
    
 ==== Results with atmosphere ==== ==== Results with atmosphere ====
  
 +== MYSTIC results ==
 The following plots show the results obtained with MYSTIC. As for the cases without atmosphere the patterns look the same for all models. ​ The following plots show the results obtained with MYSTIC. As for the cases without atmosphere the patterns look the same for all models. ​
 {{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​c3_mystic_aer.png|C3 with aerosol, MYSTIC results}}  {{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​c3_mystic_aer.png|C3 with aerosol, MYSTIC results}}
  
-The following plots show the mean radiance (for each Stokes component), mean absolute difference with respect to MYSTIC (forward Monte Carlo), mean standard deviation, fraction of pixels which agree within 2 standard deviations (2σ), please note the non-linear scale of the colorbar. +== Statistics == 
-{{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​c3_statistics_2.png|Summary plots, statistics C3 with atmosphere}} + 
-{{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​c3_statistics_2.pdf|Summary plots, statistics ​C3 with atmosphere}} +{{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​c3_statistics_rms_std_match_2_nstokes4.png|Summary plots, statistics C3 with atmosphere}} 
-  * Stokes component I (upper row): Mean radiances agree well, largest differences for SHDOM. Match fraction shows systematic differences for SPARTA, again cases 2 and 3 as for case without aerosol. MSCART variance reduction ​works better ​than MYSTIC.+Statistics of the Stokes vector results for scenario ​C3 (cumulus cloud 
 +field and aerosol in molecular ​atmosphere).  The panels in the left 
 +column show the mean radiance $I_{\rm mean}$ (for Q, U, and V the mean 
 +of the absolute values) for all models and all 9 cases. The panels in 
 +the second column show the standard deviations $\sigma_{\rm rel}$. The 
 +third column shows the root mean square differences $\Delta_{\rm RMS}$ 
 +in per cent and the right column shows the match fractions $q$. 
 + 
 +** Notes: ** 
 +  * Stokes component I (upper row): Mean radiances agree well, largest differences for SHDOM. Match fraction shows systematic differences for SPARTA, again cases 2 and 3 as for case without aerosol. MSCART variance reduction ​more efficient ​than MYSTIC.
   * Stokes components Q and U: Similar results as for I.   * Stokes components Q and U: Similar results as for I.
   * Stokes component V: Results dominated by noise, no quantitative comparison possible. ​   * Stokes component V: Results dominated by noise, no quantitative comparison possible. ​
Line 122: Line 148:
 {{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​stats_sparta_rayleigh.pdf|SPARTA}} {{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​stats_sparta_rayleigh.pdf|SPARTA}}
  
-{{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​stats_mystic_bw_rayleigh.pdf|MYSTIC - backward MC}} 
  
 ** Stokes parameters I,Q,U,V ** ** Stokes parameters I,Q,U,V **
Line 168: Line 193:
   ​         ​      
 {{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​stats_sparta_aerosol.pdf|SPARTA}} {{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​stats_sparta_aerosol.pdf|SPARTA}}
- 
-{{:​intercomparisons:​phase_c:​c3_cumulus_cloud:​stats_mystic_bw_aerosol.pdf|MYSTIC - backward MC}} 
  
  
intercomparisons/c3_cumulus.1479313706.txt · Last modified: 2016/11/16 21:27 (external edit)