Strange results of SDR model

Hi,

I tried SDR model in my research site in Vietnam. The model run smoothly, but the results are very strange. For example, in the watershed_outputs: 
- Sed_retent is -1.465194e+022
- sed_export is - 1
- usle_tot is -1
other outputs, such as sed_export has values from -1 to 49; sed_retention_index varies from -1 to 47, and usle is from -1 to 64

My inputs were extracted from the following sources:
- dem: ASTER GDEM version 2
- erodibility (K): HWSD version 1.2, which falls between 0.41 and 0.76
- erosivity (R): estimated from worldclim current monthly precipitation by the formula developed by (Wischmeier and Smith, 1978). R ranges from 1597 to 5091
- c-factor: estimated from NDVI for tropical region suggested by(Durigon et al., 2014): C = (1-NDVI)/2, which ranges from 0.166 to 0.245
- p-factor: ranges from 0.15 to 0.2, because this area is mostly cover by forests (42%)

Do you think I made something wrong? Please assist me.

Many thanks,
Duc

Comments

  • DougDoug Administrator, NatCap Staff
    HI Duc,

    Do you think you could tell us what version of InVEST you are using?

    Thanks!
  • RichRich Administrator, NatCap Staff
    And if you are using an old version, can you try the development tip here: http://data.naturalcapitalproject.org/nightly-build/release_default/InVEST_dev38_3_1_1 [cb023e9383dc]_x86_Setup.exe

    There were some post 3.1.1 patches that might address issues like this.  If that's not the case can you dropbox me your data stack at richsharp@stanford.edu so I can take a look?
  • PerrinePerrine Moderator, NatCap Staff
    Also, you may want to double check the units for the erosivity. This won't fix the negative values but it will make sure that you have reasonable results when the model runs correctly.

    If you are using the equation with monthly data from Renard et al. 1994, the units will be in the international system, whereas the range of values for erodibility you provide seems to be in US customary units. More info in this thread.

    Best,
    Perrine

    Ref:
    Renard, K., & Freimund, J. (1994). Using monthly precipitation data to estimate the R-factor in the revised USLE. Journal of Hydrology, 157, 287–306.
  • Thank you all, 

    I double checked the units for the erosivity. The equation is: 

    (MJ mm ha-1 h-1 y-1) = sum(1.735*10 power (1.5*lg(Pi*Pi/P) - 0.08188) 

    Where Pi is monthly precipitation and P is annual precipitation.  

    I tried again with the latest version 3.1.1 and it fixed the negative values!!!

    One minor problem, there are a lots of no data pixels. Do you think that I need to fill dem before hand?

    Thanks,
    Duc

  • PerrinePerrine Moderator, NatCap Staff
    Hi Duc, 

    Good that you checked the erosivity. My point above was that the erodibility and erosivity units cancel each other out, to give a value in ton/ha/yr: cf. the user's guide
    So one needs to make sure that they are expressed in the same unit system (either US customary units, or international system).

    From the range you provided for erodibility values, it looks like they are not expressed in the international system (i.e. ton.ha.hr (ha.MJ.mm)^-1), so they are not matching the erosivity units.

    Best,
  • Thank you Perrine,

    Now I am clear. However, I am confusing with the unit of erodibility. I followed the user guide to calculate soil erodibility from soil texture class and topsoil organic carbon which guided by OMAFRA factsheet. The unit of K-erodibility given in the factsheet is tonnes/ha or tons/acre, I used tonnes/ha.

    Could you tell me how to convert the unit to the international system (i.e. ton.ha.hr(ha.MJ.mm)^-1)? 

    Best regards,
    Duc

  • PerrinePerrine Moderator, NatCap Staff
    Hi Duc, 

    Good point, the units in the OMAFRA factsheet are confusing: the authors made half of the units implicit for the R and K factor, under the rationale that they cancel each other out. They should mention the complete units, so you'd see that R is expressed in of ft.tonf.inch (acre.hr.yr)^-1 and K is in ton.acre.hr (acre.ft.tonf.inch)^-1 (US customary units).

    So... I reattach the Appendix 1 from the RUSLE guide (Renard et al. 1996) where you find the conversion factors, and where you can see that the units for R and K need to cancel each other out.

    In short, to convert K to the international system, you multiply by 0.1317.
    Cheers


    Ref:
    Renard, K., Foster, G., Weesies, G., McCool, D., & Yoder, D. (1996). Predicting Soil Erosion by Water: A Guide to Conservation Planning with the revised soil loss equation, Agriculture Handbook No. 703, 404 p.
  • Hi Perrine,

    It is great! Thank you very much.

    Best, Duc
This discussion has been closed.