Using mod_landm_coslat for the whole grid

[mborreggine@g_harvard_edu]

Hello all,

I'm doing an LGM run of CESM1 using component set B1850 and following the paleoclimate simulation guide from NCAR. I'm currently working on creating a new initialization file for CAM (bnd_topo).

The last step before getting bnd_topo is to run "mod_landm_coslat.ncl". This is all in the paleoToolkit. As I understand it, mod_landm_coslat is used to edit your topography "if you have extended land over ocean points" (and presumably vice versa). The issue I'm having is that I'm in a sea level lab that models a very specific kind of paleotopography, and our model tends to be different globally (e.g., coastlines around the entire world would be different from the usual USGS base topography, rather than just the usual changes to Hudson Bay, Bering Strait, Australia, Indonesia, etc.).

I'm uncertain how to use mod_landm_coslat to correctly reclassify land and ocean cells to their correct category (land or ocean) as they are in my topography. It seems that areas where topography is reclassified is outlined (e.g., Hudson Bay, Bering Strait) in the code and a new land mask is created and applied. But theoretically I would need to do this for the entire world since my topography is so different from the USGS base topography. Is it ok to run mod_landm_coslat.ncl on the entire grid?

Thanks!

 

[shields@ucar_edu]

I am not sure what version of the toolkit you are using, but the variable "LANDM" is applied to the cloud physics and radiation to define a land fraction ramped at the coastlines that will eventually be zero over the ocean. This should be done over the entire globe.

 

[mborreggine@g_harvard_edu]

I am not sure what version of the toolkit you are using, but the variable "LANDM" is applied to the cloud physics and radiation to define a land fraction ramped at the coastlines that will eventually be zero over the ocean. This should be done over the entire globe.

Thank you for this response. I'm using the most recent paleotoolkit from github. I was confused because in the code (pictured here - mod_landm_coslat.ncl from paleotoolkit/cesm1/atm/definesurf/public/cam4), spaces that would usually have land extended over them / taken away (e.g., Bering strait, Hudson Bay, etc.) are outlined specifically - I thought I was meant to outline the lat and lon of every space I've extended land over / taken land away from, like so.

 

[nanr]

mod_landm_coslat.ncl is meant to be used for 'Near Modern' paleo simulations where small changes are made to the modern coastline, where the code operates over small sections of the grid by extending the landm_coslat variable over new land. For example, adding land over Hudson Bay, or closing the Bering Strait. This code is not intended to be used for 'Deep Time' simulations where the coastlines are completely different from modern.

Note that LANDM_COSLAT was used in cam versions CAM3/CAM4 as a scale factor for offshore aerosols, and is no longer used in CAM5.

 

[mborreggine@g_harvard_edu]

mod_landm_coslat.ncl is meant to be used for 'Near Modern' paleo simulations where small changes are made to the modern coastline, where the code operates over small sections of the grid by extending the landm_coslat variable over new land. For example, adding land over Hudson Bay, or closing the Bering Strait. This code is not intended to be used for 'Deep Time' simulations where the coastlines are completely different from modern.

Note that LANDM_COSLAT was used in cam versions CAM3/CAM4 as a scale factor for offshore aerosols, and is no longer used in CAM5.

Thanks for this response - this gets to the crux of my problem. I am doing Near-Modern runs, so generally topography is quite similar to today. However, my lab uses state-of-the-art sea-level modelling that results in non-uniform changes across the globe - of course with the coastlines being affected the most. If I plan to use a topographic construction in CESM where virtually all coastlines are different from today (but the topography is still not as different as it could be, for example, in a deep time run), do I need to outline every coastline in mod_landm_coslat?

 

[nanr]

In answer to your question, I think you could run this script over the entire grid, and it would achieve the result that you want: it would match the landm_coslat to the landmask.

Just curious: What version of the model are you using?

 

[mborreggine@g_harvard_edu]

In answer to your question, I think you could run this script over the entire grid, and it would achieve the result that you want: it would match the landm_coslat to the landmask.

Just curious: What version of the model are you using?

Thank you so much for this answer - very helpful. I'm using CESM1.1.1, which is the version my department utilizes through the university's supercomputer.