Soil Texture and colour

[nickh07@yahoo_com_au]

Hi,

I understand that soil texture and colour is specified in seperate input datasets but does each PFT have a set texture and colour (I would guess not since they need an input file each)? And if so is this automatically determined when I run the script to generate the surface datasets for soil color, PFT's, LAI etc.? I am wandering how to implement a (semi) accurate representation of soil colour/texture for a palaeo study.

Regards,

Nick

 

[slevis]

As you suspected, the pft data and the soil texture and soil color data are unrelated. The code points to the appropriate input files when you run the script that generates a surface dataset. We refer to these input files as "raw datasets" and, if remember right, the soil texture file contains data at 5 minute resolution. The units are called "mapunits" and correspond to soil profiles in the igbp soil dataset. For you to change soil textures and colors it may be easier to make changes directly to your surface dataset, where texture is given in the more intuitive units of %sand/silt/clay.

Sam

 

[nickh07@yahoo_com_au]

Hi Sam, the file mksrf_soitex.10level.nc has %clay and %sand for all ten layers, (no silt though), we are using a reduced grid (2x2 deg resolution). Is this (mksrf_soitex.10level.nc) the 'surface dataset' you're referring to or something else?

Regards,

Nick

 

[slevis]

I think I see the source of our misunderstanding:
-You are at the stage where you run a script to create what I call "raw datasets."
-I was referring to the next step where you run the CLM3 for the first time and fsurdat is blank in the namelist. At this stage the raw datasets are used to create what I call a "surface dataset." After that first time, the CLM uses the surface dataset as an input (fsurdat in the namelist).

So, mksrf_soitex.10level.nc is a "raw dataset" in clm lingo. It contains "soil mapunits" and corresponding %clay and %sand. The silt equals 100 minus %clay minus %sand.

The "surface dataset" you will create by running the CLM with fsurdat=' '. This dataset shows %clay and %sand in 3-D instead of showing "soil mapunits."

Sorry if I confused you earlier. By the way, if you're still confused, I recommend reading next the clm3 documentation.

Sam

 

[MrIgnition]

I think I see the source of our misunderstanding:
-You are at the stage where you run a script to create what I call "raw datasets."
-I was referring to the next step where you run the CLM3 for the first time and fsurdat is blank in the namelist. At this stage the raw datasets are used to create what I call a "surface dataset." After that first time, the CLM uses the surface dataset as an input (fsurdat in the namelist).

So, mksrf_soitex.10level.nc is a "raw dataset" in clm lingo. It contains "soil mapunits" and corresponding %clay and %sand. The silt equals 100 minus %clay minus %sand.

The "surface dataset" you will create by running the CLM with fsurdat=' '. This dataset shows %clay and %sand in 3-D instead of showing "soil mapunits."

Sorry if I confused you earlier. By the way, if you're still confused, I recommend reading next the clm3 documentation.

Sam

Thank you so much Sam for this. Although this thread will be 3 years old next week, I need some help with this now.
I am grateful for the mention of "raw dataset". That is the stage I am also. I intend to run CLM5 over a large domain at a very fine resolution meaning I have to ensure that my input files are also at high resolution (or close to that). I am beginning with soiltexture and i have sought appropriate sand, clay and silt dataset from soilgrids.
I have learnt from the above discussion how to generate silt (100 - (sand + clay)), but I like to know;

CLM uses 10 layers but soilgrids has 6. What do you recommend I do?

Thanks in anticipation.