Snow depth transition to bare sea ice - Icepack

[simondriscoll]

Hi,

in every version of Icepack I have downloaded/looked at (including the most recent) the parameters for melt ponds are given as:

&ponds_nml

hp1 = 0.01
hs0 = 0.
hs1 = 0.03
dpscale = 1.e-3
frzpnd = 'hlid'
rfracmin = 0.15
rfracmax = 1.
pndaspect = 0.8

hs0 is the "snow depth of transition to bare sea ice in m". In the documentation it refers to hs0 default value as 0.03, which contradicts that in every Icepack release I've looked at over a few years. I assume there is a typo somewhere therefore. Also the documentation refers to dpscale as 1.0, whereas the model releases all have 1e-3.

Secondly, could I be informed of what "snow depth of transition to bare sea ice in m" exactly means? (and apologies for my confusion) Is it the depth of snow where, if it calculated as lesser than this, the model simply ignores remaining snow, discards it, and treats it as bare sea ice?

Thanks!

 

[eclare@lanl_gov]

Hi Simon,

I'm sorry you seem to be finding the codes so frustrating to use. They are admittedly complex and there is not one "right" answer or set of parameters that suits all of our users. CICE and Icepack are research codes and must be configured differently depending on the situation. There's no way to turn those many decision options into a piece of software that acts like a black box, i.e. insert parameter A and answer B comes out. Users are required to dig in to both the documentation and the code to fully understand what it does.

The melt pond schemes are explained in detail in this paper, which is referenced in the documentation:

E. C. Hunke, D. A. Hebert, O. Lecomte (2013). Level-ice melt ponds in the Los Alamos Sea Ice Model, CICE. Ocean Modelling, 71, 26–42. doi: 10.1016/j.ocemod.2012.11.008.

The documentation is not always entirely consistent with the code and scripts, since those evolve over time and we have many test configurations with different namelist values, but you can learn how the parameterizations work (including things like the snow transition) from the peer-reviewed literature, and you can check the code and scripts themselves to see/confirm exactly what they are doing. If you find something in the online documentation which is wrong, compared with the code, then please do bring it to our attention. But note that differences between namelist values and hard-coded values used for initialization are not necessarily an indication that there's a mistake.

A "default" value is somewhat relative in the sense that it might be the initial value hardwired in the code, which can be changed using the namelist input, or it can be the standard namelist value, or it can be the tested and/or recommended value. This code has been around for nearly 3 decades and there have been a lot of changes over the years that can result in inconsistencies with older texts. In any case, parameters like hs0 can be changed in the namelist and so you can choose to use 0.03 or 0 or whatever you like, run the code and see what happens. hs0 is the snow depth below which there is a transition (of the form 1/(1+hs0) if I remember correctly) in the property it's being applied to -- I think this is done in a couple of different places in the code, but only for some parameterization choices. hs1 is similar, and note that it has the value 0.03 in the namelist above, and so for the parameterizations set in this namelist, hs1 is active rather than hs0. We do use hs0 for the melt pond scheme(s) but it predates that -- it has been used for snow-influenced radiation calculations for decades, and it is mentioned in the older literature cited in the CICE or Icepack documentation associated with the radiation schemes.

To reiterate, I recommend reading the online documentation for guidance in using the code, the peer-reviewed literature for in-depth explanations of how and why the physical parameterizations work, and the code itself for understanding the details.

Best
e

 

[eclare@lanl_gov]

Also, I've recommended in the past that you get in touch with Danny Feltham's group at the University of Reading. I saw them last week and asked whether they had heard from you, and they said no. They are very familiar with the code, especially the melt pond parameterizations. Working with a larger research group on complex problems like sea ice is much more productive than working by yourself! Who are you working with, or who supports your work? What is the application?

 

[simondriscoll]

Dear Elizabeth,

apologies. Whilst I know parameter values would change over time, my assumption was that the model documentation for the Icepack release would be in keeping with that release. E.g. the values in that version/pdf release of Icepack version 1.2.1 would correspond to the pdf documentation https://buildmedia.readthedocs.org/...pack/icepack1.2.1/cice-consortium-icepack.pdf "Icepack1.2.1" That's all.

I am a PhD student at the University of Reading, Danny Feltham is on my advisory board, and reads my work as part of this.

"If you find something in the online documentation which is wrong, compared with the code, then please do bring it to our attention."

This is what I believed I was doing here.

"But note that differences between namelist values and hard-coded values used for initialization are not necessarily an indication that there's a mistake."

Noted thank you.

Also I was instructed by you the best way for any Icepack queries was to post "questions on the forum so that the Consortium's liaisons and others in the community can help out." So I believe I was just as you say flagging up a possible inconsistency in model documentation/and model release values. :) Apologies if this is no longer the correct method.

"hs1 is similar, and note that it has the value 0.03 in the namelist above, and so for the parameterizations set in this namelist, hs1 is active rather than hs0. We do use hs0 for the melt pond scheme(s) but it predates that -- it has been used for snow-influenced radiation calculations for decades, and it is mentioned in the older literature cited in the CICE or Icepack documentation associated with the radiation schemes."

Thanks. I thought this is why it is called the "Snow depth of transition to bare sea ice", whilst hs1 is the snow depth of transition to pond ice. I assumed that hs0 and hs1 were to do with a transition of snow, depending on whether that was snow over bare ice or a melt pond, respectively, which made sense.

Best regards,

Simon

 

[eclare@lanl_gov]

I have added an Icepack issue to check the documentation for consistency. This is the right place to post such questions. Thank you.