2020-09-24

Agenda Item 1:

CRTM 2.4 Release Update: Technical items associated with the release

Discussion:

Ben: [shows slides]. In the chat I will put the link to the 2.4 Release issue.

https://github.com/JCSDA/CRTM_dev/issues/67

https://app.zenhub.com/workspaces/crtm-development-5aaf935412f8e82ae4ed50d4/reports/release?release=5f6ba3d34093734cf5b5d4c6

Release 2.4 is really sort of a preliminary release leading up to 3.0, so it does not contain any of the polarized features, but it contains all code optimizations and bugfixes. An alpha version is available on Github and Emily Liu has agreed to test it. There are some new coefficient files, I don’t know if the folks from STAR want to share some new ones. One of the things that I have listed here is the netCDF interface for cloud- and aerosol tables. It’s not clear whether we’ll make it to the transmittance coefficients in netCDF4. We will need to have comprehensive tests for REL-2.4 so we can coordinate our advertisement with the JEDI release. Using 2.4 with JEDI is optional and I am making it so that 2.4 will not break anything. That won’t be the thing with version 3.0.

I merged the code, got the netCDF support. We are working with new scattering tables. Scott Sieron, this is what you brought from Penn State and we will merge that in the release. If you are willing to do that you can review the release.

Scott: I am not sure what to do.

Ben: I understand, I had a talk with Daryl and I am happy to help you.

We are implementing new tests, which is a challenge. We need to bring them in the same format as JEDI. My idea is to work the next week on tests that cover the CRTM from end to end. The TL and AD tests might be a challenge. The last time sink is the updated documentation. That will be an update from 2.1.3, which is the last user guide version. So that’s where we are at, a little bit slow on things. But there will be no more development outside the develop branch in the CRTM_dev repository. Sometime next week there will be a public CRTM repository.

Any questions?

Kevin: I didn’t see the CMAQ work from Yingtao?

Ben: Yingtao made a pull request last night. I’ll have Cheng review that.

Kevin: I’ll have Yingtao follow up with you on any new sensors that we have.

Ben: One thing we are doing at the JCSDA is keeping in touch with the UFO team to have better coordination with the Ops folks. Comments? Just FYI, I promised you a fast CRTM and this is it with OpenMP.

Hui-Ya: We need to change the way CRTM is called?

Ben: Yes, that’s a question how we need to send the profiles in. We can send in a few profiles at a time.

Hui-Ya: What’s the max number you can send in?

Ben: I’m not sure. I have tested it for a million.

Hui-Ya: So potentially we can send all in?

Ben: There is a certain overhead.

Hui-Ya: That sounds good.

Greg Thompson

3:15 PM

@Hui-Ya Can I double-check with you how UPP handles the hydrometeor radiative effective size for different Microphysics schemes? I coded that up in WRF - but I want to ensure how UPP handles it before going to CRTM.

Result:

A list of the new features of REL-2.4 is available on Github.

Tasks:

JEDI-compatible ctests need to be added to the CRTM.

Responsible People:

Ben

Deadline:

October 2020

Agenda Item 2:

Brief Science Byte on melting layer modeling (Johnson)

Discussion:

Ben: This has been something that has been on my radar for a while. Right now, the CRTM is handling no particle melting. So, I’ll just share my slides again.

Hopefully you can see them.

Is a melting simulation even necessary?  Impact on passive observations and radar observations, what’s going to be the potential impact? There are a number of technical challenges.

Here you can see a schematic diagram. The short dash is the freezing line and the short dash is sort of the rain/freezing separation.

What we would be looking to do in the CRTM is trying to model these transitions from a radiative perspective. This is a graph from Galligani that shows brightness temperatures. We also have the rain rate. The bright band is the melting layer. Well within that melting layer you can see strong brightness temperature differences. So, it is clear that there is some effect going on here from a radiative perspective.

What you can see here is some impact on polarization difference over the ocean. This is suggesting that it is fairly important to characterize the radiative impact of a melting layer both in terms of radiance and polarization. If you take it to the extreme, this is from a paper I wrote, if you have a melting particle you can have a strong BT change over a short melting range. When I was doing the CSAM radar simulations in the CRTM you can see the bright band from enhanced reflectivity in the data but not in the CRTM.

Greg: @Ben - I plan to return to this mtg, after a brief check-in with the code sprint JEDI daily brief.

Ben: Modelling the particle morphology is another piece in this. Here are some of my simulations. So, a lot of the work has already been done. One technical challenge is that the CRTM is not good enough for this kind of study. We would need some sort of vertical slicing to resolve the melting line. What’s the particle morphology and PSDs? These are open questions. I just wanted to introduce this topic. Any questions?

Maybe Greg? – He’s gone.

My question to everybody – what do we have from the model domain as a description for melting? -

Hui-Ya: GFS is increasing vertical resolution.

Ben: Do you know by how much?

Cory: There are 127 layers.

Ben: Do you happen to have an estimate of the vertical height?

Cory: No.

Ben: More or less than 500m?

Cory: Certainly less than half a kilometer.

Ben: That’s what we need as resolution.

Ben: I also wanted to know your opinion on the scientific value because I know that the RTTOV guys thought about doing this.

Hui-Ya: Yes, certainly.

Ben: Hui-Ya, have you tested UPP with the higher vertical resolution?

Hui-Ya: Yes, but none of us have time to test the impact.

Ben: Looks like Greg is back. Do you have any microphysical melting questions?

Greg: No, I left while you were presenting.

Ben: So, the question is, do you think this is even possible to get a mapping from model to CRTM?

Greg: I do believe that is possible. There is 2D distrometer data and basically for every degree of wet-bulb temperature you get a significant degree of melting. Since melted snow is way easier to see than graupel and these stratiform regions have been flown a lot by aircraft I think this is a very good first-order approximation we’re persueing.

Ben: I think on thermal radiance there is a significant impact.

Greg: On one of your slides I saw that you have a high latitude which has a high melting point.

Ben: You often see terrain intersecting melting layers. The weather impact is also important.

Cory Martin - NOAA Affiliate

3:38 PM

500 mb is ~ 250m per layer; decreases towards sfc; above 100 mb, quickly increases

Ben: So, I think this would probably be 1 or 2 years out. It’s something from my own research.

Result:

-       Melting layers have a significant radiative effect, both in terms radiance and polarization.

-       Resolving the melting layer is a challenge, but it seems possible and it could provide some scientific value.

Tasks:

-        

Responsible People:

Ben

Deadline:

~ 2 years

Agenda Item 3:

Coefficient Generation Update

Discussion:

Patrick: The apodization shell script is finished and now I am working on the shell script that combines the transmittances for the regression. The VIIRS coefficients for the comparison with STAR are finished.

Ben: Did you make the comparison?

Patrick: No, I only computed the coefficients.

Ben: Remind me, so it was AVHRR and VIIRS for the comparison?

Patrick: No, only VIIRS.

Result:

-       VIIRS coefficients have been computed.

-       The apodization shell script is finished.

Tasks:

-       Create VIIRS comparison plots.

-       Write transmittance concatenation shell script.

Responsible People:

Patrick

Deadline:

October 2020

Agenda Item 4:

Aerosol Update

Discussion:

Cheng: I’m looking at Yingtao’s branch now. We have two very different aerosol structures now. There need to be some modifications for check_crtm.fpp and I am testing that on my local branch. My concern is that merging Yingtao’s changes will break check_crtm.fpp.

Yingtao: Cheng, what’s the program for testing the CRTM?

Cheng: The aerosol profiles are tested in check_crtm.fpp.

Yingtao: The aerosol types are hard-coded in the

Sarah Lu: The differences is in the aerosol species?

Yingtao: The difference is in the aerosol name.

Ben: There definitely needs to be more planning for the aerosol implementation. There has to be some legacy support.

Result:

-       The new REL-2.4 netCDF branch has been merged in the CRTM.

Tasks:

-       Find a way to include Yingtao’s CMAQ work in the CRTM

Responsible People:

Cheng Dang

Deadline:

October 2020

Agenda Item 4:

General questions and comments

Discussion:

Ben: Any updates on research?

Hui-Ya: We’ll aim for simulated GOES implementation <unintelligible>

Ben: It’s kind of a technical thing.

Greg Thompson

3:45 PM

so does HRRR at NOAA-GSD/GSL

Ben: So, it sounds like it’s more of a technical issue to make that Awebs conversion.

Hui-Ya: We were supposed to go to Aweb two years ago.

Greg: I don’t do HRRR myself, but it’s done at NOAA.

Hui-Ya: Would you be willing the microphysics scheme in the GFDL?

Greg: Sure.

Ben: How are things with the DTC?

Hui-Ya: UPP will be part of UFS. UPP carries a lot of libraries up to the previous release. DTC comes with the 10 libraries that UPP needs.

Ben: So, your concern is primarily the operational implementation.

Ben: There will be an upcoming quarterly review meeting for all contributors, just as a side note.

Ming do you have any CSEM updates?

Ming: We already set up the CSEM repository. <unintelligible>. In this first release we have all the CRTM surfaces and we will include an atlas. For the MW land model, we will include <unintelligible> and some new models based on machine learning. I think technically it’s ready. We need to work together with you whether we need to modify the CRTM. We need to take advantage of the new data structures in the CSEM. For the retrieval we had some issues. <unintelligible> I’m working with Kevin to get the CSEM on to a public Github repository.

Hui-Ya Chuang - NOAA Federal

3:58 PM

@Greg, I just emailed you UPP Github site. Thanks.

Ming: Any adjustments and improvements you need are no problem for us.

Ben: Let’s sit down before December and make a work plan.

Ming: <unintelligible>

Ben: We will be starting earlier with the AOP this year.

Ming: <unintelligible>

Ben: Any other items? One last comment, Jake Liu gave a JEDI talk today about an RTTOV C++ interface. As soon as that’s available there are some things to do to allow one-to-one comparisons between RTTOV and CRTM in a scattering setting.

Ming: <unintelligible>

Ben: Ok, I’ll set up a meeting.

Result:

N/A

Tasks:

N/A

Responsible People:

N/A

Deadline:

N/A