Using iCESM for Moisture Source Attribution and Precipitation Recycling Analysis

[hadi]

Dear iCESM Development Team and Community,


I am a researcher working on atmospheric moisture transport and precipitation recycling. I am interested in using iCESM to quantify moisture source contributions—specifically, determining the fraction of precipitation over the South US that originates from evaporation in the other remote sources.


From the literature (e.g., Nusbaumer & Noone, 2018; Dyer et al., 2017), I understand that iCESM includes water tagging capabilities that can track moisture from user-defined source regions through the hydrological cycle. However, I would appreciate clarification on the following:

1. Moisture source attribution: Can iCESM's water tagging framework explicitly quantify precipitation recycling ratios (i.e., the fraction of precipitation in a target region originating from a specific source region)?
2. Regional configuration: Is it feasible to define custom tagged regions such as the Gulf of Mexico, and track the subsequent contribution of that moisture to precipitation over continental areas like Texas?
3. Model version: Would iCESM1.2 or iCESM1.3 be more appropriate for this type of regional moisture source analysis?
4. Computational requirements: What is the typical computational cost for running water tagging simulations, and are there recommended best practices for defining the number and spatial extent of tagged source regions?

I would be grateful for any guidance, documentation references, or examples of similar applications. If there are existing datasets or published simulations that address moisture sources over the south-central United States, I would also appreciate pointers to those resources.


Thank you in advance for your time and assistance.


Best regards,

 

[nusbaume]

Hi Hadi,

Apologies for the slow reply! In response to your questions:

1. Yes, the water tagging framework can explicitly calculate precipitation recycling, down to the scale of a grid cell. This can be done by creating a water tag for the region of interest, and then calculating the fraction of precipitation that falls over that region that comes from that region's tag (with that ratio being equivalent to the recycling ratio).

2. Absolutely! You can find a source code example of regional tags, along with instructions for how to set them up in iCESM1.2, online here:

iCESM1.2/instructions at master · NCAR/iCESM1.2

Isotope-enabled CESM1.2. Contribute to NCAR/iCESM1.2 development by creating an account on GitHub.

github.com

Then you can just calculate the fraction of precipitation or water vapor from that tag relative to the total preciptation/vapor for any other region in the globe.

3. Either iCESM1.2 or iCESM1.3 should be adequate for water tagging experiments. The two papers you mentioned used iCESM1.2, in case that helps.

4. Formal computational cost estimates were never calculated, but in general a new water tag makes the model simulation ~10% slower. So adding ten water tags will make the model simulation approximately 2x more expensive.

Anyways, I hope that helps, and please let me know if you have any other questions or concerns.

Thanks, and have a great day!

Jesse

 

[hadi]

Thank you so much, Jesse. I am trying to run iCESM 1.3 for 1979-2014. Is there any SST input for iCESM from ERA5?
I could not find it here:
/glade/campaign/cesm/cesmdata/inputdata/atm/cam/sst

My WRF SST is using ERA5 for SST, so that is why I would like to have the same setup for SST.

Bests

 

[hadi]

Hi Jesse,

I ran iCESM 1.3 and in the very first hour, I could south america tagged precipitation. I mean, moisture from NPAC could travel to South America and precipitate in only 1 hour of simulation?

Attached is my user_nl_cam. running from 1979-2014.


Also, if you know any template case on derecho I can use, that would be helpful

Bests,
Hadi

 

[nusbaume]

Hi Hadi,

Depending on how you configure the model all water tracers start off with the same value as regular water, and then rain out over time except for in locations directly influenced by the tagged region. In other words the water tracers need time to "spin-up" to the correct value.

Given this, I generally ignore any results for at least the first few months when doing a "startup" run. If however the results still looks suspicious after a few months then it could certainly mean that something is wrong with the water tag's setup.

Anyways, I hope that helps, and have a great day!

Jesse

 

[hadi]

Hi Jesse, and thank you for your response.

we expect that it takes time for moisture tagged from NPAC to travel and spread to other remote regions, so we do not expect to see NPAC moisture within the first simulation hour in remote regions such as the Southern Oceans or East Asia.

Why do all water tracers start off with the same value as regular water, but not zero?


Can you help me understand how these initializations of tracers were done?

Bests,
Hadi