K parameter units and the Harmonized World Soil Database

gaviotingaviotin Member
edited August 2013 in Pre/Post-Processing
Hi everybody!
I am working with InVest to select priority areas for some environmental services. Under the Sediment Retention Model I was trying to understand the meaning of the soil erodibility factor K, and I found myself doing a unit analysis to make sure that the final required units for K (Mg.h.Mj-1.mm-1) are the result of every input parameter units. I then realized two problems:

1) to obtain Mg in the final value of K I should use the organic matter (OM) as Mg.ha-1 . The harmonized world soil database HWSD (which is recommended in the InVest manual) only provides with % of organic carbon. It means that first I must transform organic carbon to organic matter by using a factor correction of 2 g OM/g carbon (as suggested by Brown, 1982 for tropical forests) which in that case I need the soil density values to obtain carbon grams from the % of organic carbon. Even if the units were already stated in one of the constant parameters in the formula, I still have to do the conversion from organic carbon to organic matter, and is not as simple as just multiplying by 2 (because of the units). Has anybody had this problem??

2) The HWSD only provides with % of sand whitout making any difference between the very fine sand and the rest of sand classes. Therefore, the m parameter will be overestimated when using the % of sand directly in the first part of the m formula (silt (%) + very fine sand (%)). As I couldn´t find an indirect way to know which proportion of sand makes this very fine sand I will leave out the "very fine sand" value from the calculation to avoid increasing the error. What do the experts think about this??

Thanks in advance for your suggestions!

Claudia Véliz


  • Regarding point 1 I just found two soil density values in the HWSD database: The reference bulk density and the bulk density. As far as I´ve read it seems that I should use the reference bulk density for the tropical forests (in areas wihtout antropogenic effect) am I right or maybe I just missunderstood what I've found in some papers?
  • PerrinePerrine Moderator, NatCap Staff

    As pointed out by Duc, if you are working in Europe the soil database may be useful, with direct values of erodibility: http://eusoils.jrc.ec.europa.eu/library/themes/erosion/Erodibility/
    The SOTER is also an option to look at.

    If you use the HWSD, here are some thoughts on your two points:
    1) This old post contains a tutorial with useful information on how to use the HWSD. 
    In particular, you'll see that Stacie's recommending a factor of 1.72 to convert OC to OM.

    2) In the same tutorial, it is suggested to use the soil texture classes (e.g. "loamy sand") and the table provided in the User's guide (copied from here). The equation in the user guide uses slightly different soil classes (including the very fine sand), which are not readily available in the HWSD. So if you don't have other sources you may introduce an error by using the sand instead of fine sand.
    Note: For further information on USDA soil classes, please have a look at this document.

  • here is an interesting update, I don't know how many of you are using this recent released data base from the ISRIC but for us it worked pretty well.

     These are 1km resolution 3D modelled grids at different depths and with confidence interval values for Carbon, pH, Sand, silt , clay and bulk density. They have been very useful for our study area given that we recalculated K parameter and the obtained values are in accordance with the ones field based obtained from Ochoa (20013) see figures in paper:


    hope it is useful

  • PerrinePerrine Moderator, NatCap Staff
    Thanks Claudia.
    Soilgrids is a very useful resource indeed. We have added it to the User's guide and it will appear shortly in the online version.

Sign In or Register to comment.