Soil depth index subtleties (general) and its application in the groundwater calculation (specific)?

LeonLeon Member

So was having a bit of a fiddle with my code (I've moved company since my old username @LonBar, and have forgotten the password linked to that email :S) and then started questioning the logic in soil depth... looking at your step-by-step instructions I've noted two things:

1) '...where 0 represents the highest soil depth, and 1 represents the lowest...'
  •  does that mean 1 represents the shallowest soil depth data or the deepest soil depth?
    • is currently configured to create a soil depth index with 1 as the deepest soil depth just by normalising input soil data (which in my case is magnitude of depth); so I'm possibly missing the subtlety of setting an index with (1 - <soil depth normalised with 1 meaning shallowest>)
    • with apologies for what is a semantic nit-pick - I just wanted to be sure! 
2) Reading the RIOS manual, it seems that deeper soil depths contribute to the upslope source calculations in the same way for the objectives it appears in. But where I see erosion/phosphate/nitrogen makes sense with this schema (shallower soil depth ⇒ less soil ⇒ less potential flow to the river network), the explanation for groundwater seems to indicate that a deeper soil depth allows greater potential for water to be taken from the surface (shallower soil depth ⇒ lower capacity of soil layer to hold water ⇒ more potential flow to the river network).
  • Have I misinterpreted something here? 
    • possibly captured by the <1 - SD_{1=shallow}> I've neglected above?
    • possibly upslope source has a different interpretation for the groundwater objective?


  • swolnyswolny Member, NatCap Staff
    Sorry for the slow response, @Leon. You've made me go back and look at the doc and my code for the first time in a while. :-) It's great to know that you're still on the case. 

    For question 1, I think the confusion may be because the step-by-step doc wording isn't as clear as it could be. Here's the entirety of that paragraph, "Soil Depth must be normalized into a raster dataset where each cell is assigned an index value between 0 and 1, where 0 represents the highest soil depth, and 1 represents the lowest. Determine the maximum value of the Soil Depth raster (or max value for Latin America).  Use Raster Calculator to divide the raster by the maximum value (resulting raster = normalized).  Use Raster Calculator again to subtract the resulting values from 1 (1 – normalized).  The resulting raster is the Soil Depth Index"

    So there are two steps involved with creating the Soil Depth Index. First, doing a simple normalization of the soil depth raster (which would produce 0 for shallowest soils, 1 for deepest), then taking (1-normalized), which would produce 0 for deepest soils, 1 for shallowest. This is then used in the Upslope Source calculation, where it serves as an indicator of the state of the area above each pixel. If the upslope area is in poor condition (including having shallow soils), then it's more important to do management activities on that pixel to help mitigate the upslope condition. We'll see this again with the next question.

    For question 2, I'm seeing that we don't have the groundwater equations in the step-by-step doc at all - sorry about that! 
    You're right that in the pre-processor, the soil depth index is treated the same for all objectives. I did not come up with these equations (and it has been a while since thinking about this), so am speculating here (and will ask @adrianvogl, the mind behind the theory to chime in), but I suspect that it has to do with the fact that here we're thinking about what's happening upslope of each pixel. If it's the case that the area upslope of a pixel is less conducive to groundwater recharge (because the soils are shallow), then it's even more important to do restoration/conservation on that pixel, to help water on that flow path infiltrate instead of running off. This same kind of theory holds for other objectives - if the area upslope of a pixel is in bad shape, it's more important to do activities on that pixel than if the area upslope was in good shape.

    Note that the pre-processor is only calculating the Upslope Source, Downslope Retention and Riparian inputs to RIOS. It is also the case that soil depth is considered separately as a RIOS input, with its own set of weights and directionality, separate from what is in the pre-processor calculations. Looking at the equations in the RIOS User guide for groundwater, the soil depth-specific input to the ranking equation normalizes the soil depth input but does not take (1-normalized) like the pre-processor does. So this is saying that, for each pixel, deeper soils should be prioritized for both doing protection and management activities, because improvements there are more likely to actually contribute to groundwater recharge.

    Does this help, or muddy the waters? @adrianvogl, please correct any mistakes I've made or provide clarification.

    ~ Stacie

Sign In or Register to comment.