Crop Production - Units in Map

judypdxjudypdx Member
edited September 7 in Terrestrial Models
Could someone confirm the units of measure for the map and CSV output for the Crop Production model?  In the written instructions, there is a reference to the Monfreda paper, which states units in that paper are in metric tonnes per hectare (t/ha).  However, in the "Interpreting Results" section there is only a reference to "production rate per pixel".  Is the production rate also in t/ha?  I've copied and pasted the text from the instructions below.  There is no mention of units in the CSV file either.  My numbers for a wheat study in Kansas, US do not correspond to wheat production reports of about 3.25 Mg/ha (48 BU/acre).   Output is in "0.025" to "0.063" per pixel, so it can't still be t/ha (or Mg/ha).  Thanks so much.  Judy

  1. Result Table (CSV)
    This is an aggregating table that lists all the crops modeled in the run,
    the area covered, percentile or modeled production, observed production,
    and nutrient information for each mode. It is the primary output of the
    model.
  2. Crop
    Production Rasters
    For each crop
    modeled, there will be a raster corresponding to that crop name, the type
    of modeling (observed, percentile, or regression). The result is the production rate per pixel.
Post edited by judypdx on

Comments

  • RichRich Administrator, NatCap Staff
    Hi Judy, true, the output is in "tons/pixel" but I can see how that's confusing. If you look at a particular pixel, the value in that pixel is the tons of production in that pixel. So if you like, the units can be "tons". Sound okay?
  • judypdxjudypdx Member
    edited September 15

    Hi Rich,

    Thanks so much for the reply back.  I was also thinking the units were in  tons (metric, so Mg) per pixel as well, but
    when I calculated back to tons per hectare, the yield for wheat seemed quite a
    bit off from the averages for the region I was studying.  The average yield for wheat in Kansas was 48 BU/ac in 2012 (3.25
    Mg/ha) according to the regional data I found online. The mean value from the model
    is 0.056 tons/pixel.  When I convert that
    to tonnes/ha (each pixel is 30 x 30 m), I end up with 0.62 tonnes/ha.

    I've attached a
    pdf of the Crop Production model's output, which was done for Sumner County,
    Kansas (2012 data).  I filtered the raster map for wheat pixels only.  Do you think that might have caused this
    problem?  The raster is too big to upload, but it came from CropScape (link below) for Sumner County, Kansas:

    https://nassgeodata.gmu.edu/CropScape/

    Kind regards,

    Judy

    P.S.  Would you
    also be the contact person about questions regarding the Carbon Storage and
    Sequestration model?


    Post edited by judypdx on
  • RichRich Administrator, NatCap Staff
    Hi Judy, the contact scientist for this and the carbon models is Becky Chaplin-Kramer bchaplin@stanford.edu, but she's traveling internationally at the moment. I don't think she'll be back until next week.

    For the low wheat yield, I'm only going to guess about this, but our model has data for 117 different crops. Is it possible that "wheat" in your model could be a combination of a couple of different crops we model? You can take a look at our observed yield rasters in the sample data at  [invest-directory]data\CropProduction\model_data\observed_yield. I'm sorry if that's naive, I'm very far from being a crop scientist.

    Also this may be naive too, but aren't there percentile outputs from that model? When you're looking at 0.62 tonnes/ha, could that be from a lower percentile? It otherwise seems almost on the same order.


  • Hi Rich,  Sorry for taking so long to reply back.  I couldn't track down the observed yield rasters you mention.  It's possible that wheat for the dataset the crop production model provides is a combination, and that the average is quite a bit lower, but it still seems pretty far off.  I used the observed data and not the percentile outputs, but thanks for double checking.

    I noticed that the crop production model is no longer on the website for the 3.5.0 rollout, so perhaps there was an issue and it's been taken down.  I will contact Becky Chaplin-Kramer this coming week.  Thanks so much for her contact information.

    Appreciate your quick responses, and again, apologies for not doing the same!

    Judy
  • RichRich Administrator, NatCap Staff
    Okay I'm sorry I couldn't help, but FYI for you and others that might see this, the Crop Model is still part of the invest suite and part of the InVEST user's guide. Could you clarify what you mean when it's no longer on the website? 
  • Hi Rich, I was looking at the following link and didn't see it:

    But in doing a google search for "Invest Crop production model," I was brought to the 3.5.0 version: http://data.naturalcapitalproject.org/nightly-build/invest-users-guide/html/crop_production.html

    So it is there, just not listed with the other models on the InVEST/InVEST Models link.

    Thank you for pointing that out.
Sign In or Register to comment.