Tb single frequency simulation from a radiosonde profile launched from 500m asl: Obs_height and add_obs_height_to_layer.

[Davide88]

Dear all,
I kindly ask support about obs_height setting in the case of single frequency Tb simulations starting from a RAOB profile. In particular, I am a little bit confused about how to set variables in the case of a Radiosonde launched at a site which is not at sea level, but, for example, at 500m asl.

Questions are:

1. The .lay file, at its third line, should report 833000 0. or 833000 500. as obs_height?
2. pam.nmlSet['add_obs_height_to_layer'] should be set as True or False?
3. Would it be necessary or in general correct to set pam.p['obs_height'][0, 0, :] = np.array([833000.0, 500])?

In attachment, you can find the first three lines of the .lay generated from the radiosonde profile. You can see underlined in red the site altitude, while the red row is pointed to the obs_height for which I have some doubts.
Thank you in advance for any help.
Best regards,
Davide Rossi.

[Davide88]

Hello,
just a clarification: the observation perspective is from ground station to TOA, thus not from space to ground.
Awaiting yours,
Davide R.

[mariomech]

Dear Davide,
on the 13th of October, I sent you the following message, but it seems you didn't get it.

1. If the profile starts at 500 m and you do not manually add a layer between 0 and 500 m, I assume it does not make any difference whether you choose 833000 0 or 833000 500.
2. That's up to you. What it actually does is add an extra layer at the height where you want your outputs. This makes sense if there is no close-by atmospheric layer, cause PAMTRA always takes the next available layer as output. For example, assume your atmospheric altitudes are [0, 1000, 5000, 10000]. If you want an output at 4000 m, PAMTRA will provide you with what it calculates for 5000 m, provided you do not set add_obs_height_to_layer to True. If you set it to True, PAMTRA adds a layer at 4000m by interpolating the atmospheric variables to that layer and provides the radiative transfer output at that layer. In most cases, there is little to no difference, so it is not True by default.
3. Necessary - no; correct - yes.

PAMTRA always outputs both perspectives.

Working .lay or .lev files are somehow deprecated and only kept for backward compatibility. But I understood what you mean. I created that files more than 15 years ago.

Best regards
Mario

[Davide88]

Dear Mario,
thank you for your reply. I didn't receive any message on October 13th actually, do you mean by email or here on github?

Well, let's consider the following profile:

   1 1 107 2
    2024  9 19  1300         107             1           1
  833000 0.
   46.8      23.5     459.00000      0.45253     -1.41927   289.88634    4.00000
                                                1.17072e+01  0.00000e+00  0.00000e+00  0.00000e+00  0.00000e+00  0.00000e+00
 459.00000  96750.10000    289.88634    57.00000 7.05711e-03 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00
 462.68140  96607.93777    289.84216    57.50928 7.06105e-03 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00
 487.98511  96321.31509    289.54590    58.19417 7.03246e-03 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00
[...]

If I run pamtra as follow:

    target_freq = 89.0
    freqs = np.array([target_freq])
    pam.p['obs_height'][0, 0, :] = np.array([833000.0, 459])
    pam.nmlSet['active'] = False
    pam.nmlSet['gas_mod'] = 'R98'  # ROSENKRANZ
    pam.nmlSet['freqs'] = freqs.tolist()

I get brightness temperature Tb= 62.276 K

Instead, if I run pamtra as this second following configuration

    target_freq = 89.0
    freqs = np.array([target_freq])
    # pam.p['obs_height'][0, 0, :] = np.array([833000.0, 459]) --> obs_height commented, the only difference with respect to the previous case
    pam.nmlSet['active'] = False
    pam.nmlSet['gas_mod'] = 'R98'  # ROSENKRANZ
    pam.nmlSet['freqs'] = freqs.tolist()

I get brightness temperature Tb= 71.345 K

I cannot understand which one is corretct, what do you suggest?

Would you advice to solve the problem using an updated procedure or using a different format?
Thank you in advance,
Davide Rossi.

[mariomech]

Can you please provide the full script and the input file. That makes it easier for me to run it myself.

[Davide88]

Example_obs_height_Pamtra.zip

Dear Mario,
thank you for your kind support.
You can find in attachment the example I am referring to. It is built using the "simple passive" example, but focusing on the 89GHz instead of the full spectrum.
The script compute both the Tb when obs_heigh is left as default and then when it is set to the first level red in the .lay profile: you will se the differences directly printed as output.

Awaiting yours,
Davide Rossi.

[mariomech]

Okay. PAMTRA internally works with layers and levels, which are the boundaries that define the layers. When reading in a .lay file, all the values are defined in layers. Internally, PAMTRA creates levels around the layers, where the number of levels equals the number of layers plus one. Take a look at pam.p['hgt_lev']. The lowest level is thereby the surface, which is specified in the fourth line of the .lay input file at the last position (4 m in that case). This means the difference in brightness temperature between obs_height 0 and 459 m is the temperature emitted by the layer in between.

Cheers

Mario

[Davide88]

Ok thank you, now it is more clear.
I totally neglected the 4m at the fourth line of the .lay file. I tought it was referred somehow to hydrometeors.
In fact, if I set 459m instead of 4m in the fourth line, I get the same value for both the obs_height setting, meaning 62.31K.

Thus, just to conclude (sorry to bother), the right setting to simulate the Tb of a ground based radiometer, with the ground at 459 m above sea level, is to:

1. put the 459 in the fourth line as the ground surface level, regardless of the obs_height setting.
2. put 833000 459 in the third line
   In this way, the profile lower level coincides with the observation height ground level, both in ASL.

Am I right?
Best regards,

[mariomech]

Yes, that would be the correct way, although you can set the obs_height anywhere below ground. If no atmosphere is defined, none will be included in the calculation. Just give it a try.

Best regards
Mario