H2OSOI goes beyond WATSAT

[ekluzek]

** < Dominik.Rains > - 2016-04-12 07:47:23 -0600**
Bugzilla Id: 2306
Bugzilla CC: dlawren, rfisher, swensosc,
Bugzilla Attachment: watsat-max_sm.png - WATSAT - H2OSOI (areas masked where H2OSOI stays below saturation)

Created attachment 61
WATSAT - H2OSOI (areas masked where H2OSOI stays below saturation)

We are observing in our regional CLM 4.5 model setup, using a custom surface dataset and forcings, that over our simulation period H2OSOI often surpasses WATSAT significantly (~8%) in areas with high organic matter and precipitation.

We assume that, independently of our inputs, soil moisture should be constrained by WATSAT at all times.

The modeled catchment is the Murrumbidgee basin in Australia.

[ekluzek]

Erik Kluzek < erik > - 2016-04-13 14:13:56 -0600

Hi Dominik

Can you check which soil layer the max(h2osoi) ends up being? The top layer can exceed WATSAT because of ponding. And it looks like ponding is handled differently for urban versus other land-unit types. So you might check what the land-unit types are as well.

Thanks for reporting the issue!

[ekluzek]

** < Dominik.Rains > - 2016-04-15 03:58:30 -0600**

Hi Erik,

many thanks for the input.

it is indeed the topmost layer, no urban areas in that specific area.
I was not aware of the ponding. I will also check for the other layers.

As a side question: Can you suggest the best way of computing residual water content for each CLM pixel?

Best regards,
Dominik

[ekluzek]

Erik Kluzek < erik > - 2016-04-18 12:01:55 -0600

Dominik

I checked with Dr. Swenson our local hydrology expert. We don't calculate residual water content, anywhere so you'll have to figure it out based on it's definition.

He also showed me that pond max is set to zero in clm45 for cesm1_2_1 (so likely cesm1_2_2 as well) for clm4_5 (although it's non zero for urban). So if you are getting WATSAT>H2OSOI for non-urban points in the first layer, that may indicate somethings wrong still. Can you tell us more about what version of the model you are using?

[ekluzek]

** < Dominik.Rains > - 2016-04-19 05:50:07 -0600**

Dear Erik,

the version is cesm1_2_2 / clm4_5.
All input data is custom including the forcings (era-interim).

Are there any basic checks we can perform to exclude a mistake on our side?
My assumption is that regardless of the input H2OSOI should be constrained but maybe this is not always the case.

A colleague also pointed me towards the paper 'http://journals.ametsoc.org/doi/abs/10.1175/2008JHM1011.1' but we are not sure if this is related.

[ekluzek]

Looking at the latest code, there are certainly points where h2osoi is limited by watsat. But, I can't say that it's limited everywhere it should be.

[billsacks]

Here is the attachment from bugzilla.

[henrique-f]

Hi all,
I'm running ctsm5.2.005 and found a similar issue with H2OSOI.
I am getting very high H2OSOI values at 26 cm depth, reaching about 0.88 m3/m3. The bulk density provided in the surfdata input is 1 g/cm3, and the particle density provided in the param file is 2700 kg/m3. Based on that, the expected saturation water content would be about 0.63 m3/m3...
I am about to dive in the code to try to figure out what is going on, but wondered if you guys already have an idea of where the problem could be.
Thank you,
Henrique

[olyson]

The default calculation for the saturation water content or porosity (the CLM history field WATSAT) is described in the technical note here:

https://escomp.github.io/ctsm-docs/versions/master/html/tech_note/Hydrology/CLM50_Tech_Note_Hydrology.html#soil-water

See equation 2.7.49.
I think the calculation in the code is done in subroutine SoilStateInitTimeConst in biogeophys/SoilStateInitTimeConstMod.F90. You could print out some variables to see what is happening.