Tropical Retrieval Simulations

 

These are the tropical parameters input:

Prob2IceCld=0.33 # probability of two ice cloud layers

MeanZtop1=14.7 # mean of gaussian upper cloud top height (km)

StdZtop1=2.0 # std. dev. gaussian upper cloud thickness (km)

MeanDelZ1=0.7 # mean of exponential dist. upper cloud thickness

MinZbot1=7.0 # minimum upper cloud base allowed (km)

MeanDelZgap=1.2 # mean of exponential dist. of cloud gap

MeanDelZ2=0.9 # mean of exponential dist. lower cloud thickness

TempLiqIce=(253 273) # temperature zone for ice/liquid boundaries (K)

ProbBndCld=0.37 # probability of boundary layer clouds with cirrus

MeanBndCldTopZ=0.7 # mean of exponential dist. of PBL cloud top height (km)

 

The errors presented below are only for bins containing more than 100 samples.

Experiment #1

This is a nadir view for IWP & Dme median error plotted against cloud base height and cloud top height for different combinations of submm and IR channels.

 

 

 

 

 

 

 

Experiment #2

 

This is 38 degree view from nadir for IWP & Dme median error plotted against cloud base height and cloud top height for different combinations of submm and IR channels.

 

 

 

 

 

 

 

 

Experiment #3

 

This is 55 degree view from nadir for IWP & Dme median error plotted against cloud base height and cloud top height for different combinations of submm and IR channels. It appears that the error is slightly less as the viewing angle increases from 0 to 38 to 55 degrees.

 

 

 

 

 

  

Experiment #4

 This is a nadir view, comparing using a particle mixture (mix4) for the retrieval database then picking a random shape (4non) for the observation or picking the mixture (mix4). This shows that randomly picking a shape does not increase the IWP error, and only affects the Dme error by a small amount.