The output didn't calculate the whole area

Hi,

I ran the Habitat Quality model but The output didn't calculate the whole area. I checked everything but I don't know what's wrong with this. I failed to upload my inputs here. If you can help me please leave me with your email address. I will send you my inputs. Thank you so much!

Comments

  • swolnyswolny Member, NatCap Staff
    Hi @uncxlk -

    Send your inputs to swolny at stanford.edu and I'll check it out.

    ~ Stacie

  • Thank you!
  • swolnyswolny Member, NatCap Staff
    Hi @uncxlk -

    The problem is that your threat raster luodi_c.tif does not cover your whole land use/land cover map. You can see that the boundary of the output quality_c.tif is the same as the boundary of luodi_c.tif. I tried removing "luodi" from threats.csv and the model produced results that cover the whole LULC map. So change luodi_c.tif to cover the whole LULC map and try again.

    ~ Stacie

  • Thank you so much for helping me figure out the problem. But I get all the tif files from two LULC map (one for current and one for future) with the same settings. Why does the boundary of luodi_c is different from else? Could you please help me find out the reason?
  • swolnyswolny Member, NatCap Staff
    I don't know why the boundary of luodi_c is different, that will depend on the method you used to create it from the LULC. Go back and re-create it, making sure that the result has the same extent (or larger extent) as the LULC.

    ~ Stacie

  • I re-creat it again but it makes no difference. I’m so sorry but I don’t know where I can change the extent when I creat it? Could you please give me some instructions? Thank you!
  • swolnyswolny Member, NatCap Staff
    edited February 26
    I don't know exactly how you are trying to create the raster, so it's hard to give instructions on what to do differently.

    One thing that might help is setting the Extent for your output raster to be the same as the LULC. You can do this in ArcGIS through Environments -> Processing Extent. I don't know QGIS as well, but many of the tools have a place to set the Extent of the output raster. 

    Another thing to note from the User Guide Data Needs section is that the threat rasters should have a value of 0 where there is no threat. Your threat rasters seem to have NoData where there is no threat, and that also makes it harder to see the full extent of the raster.

    ~ Stacie

    Post edited by swolny on
  • Thank you! I did all the procedures again and finally I got it! Thank you so much!
  • Hi Stacie, thank you so much for your patience!


    When I want to decide the half-saturation value, I find that values of deg_sum_c&f are from -2.97747e+038 to 0 while I can see some pixels degraded obviously. Is that normal? How should I evaluate the half-saturation value? 


    Thank you so much!

  • swolnyswolny Member, NatCap Staff
    Those values for deg_sum_c don't sound correct - they should be between 0 and 1, and then you can use the guidance given in the User Guide for choosing a half-saturation value based on the degradation score. First, which version of InVEST are you using? If you haven't, please try with the latest version (3.6.0) and let us know if it's still happening. If you are using 3.6.0, I'm not quite sure what the problem could be, so could you please send me your latest data and I'll check it out?

    ~ Stacie

  • Hi Stacie, I have sent you the data via email. Thank you so much!
  • swolnyswolny Member, NatCap Staff
    Thank you for sending your data, @uncxlk.

    @jdouglass, I've run the data on 3.6.0 (and 3.5.0 for comparison), looked at the results and I'm baffled. All of the intermediate files have reasonable values, as does quality_c.tif, yet deg_sum_c.tif has only very extreme negative values, as @uncxlk says above, not at all 0-1. How can quality_c.tif have reasonable values but deg_sum_c.tif not? I'll send you the input data to check it out.

    ~ Stacie
  • uncxlkuncxlk Member
    edited March 2
    Hi Stacie, Jdouglass told me that it was because that my threat rasters have pixel values of {0, 1, nodata}. So I'm trying to convert the threat rasters to have pixel values of {0, 1}. But after I ran the model I encounter a new problem. It says that "in get_raster-value_infoValueError could not open I:\input\u6700\u65b0\roads_c as a gdal.OF_RASTER".

    Actually I don't know how to convert a line data(roads) into a raster with pixel values of {0,1}. So I did it like this. First, I convert it to a raster with all pixel values of 1. So I got a raster of roads with pixel values of 1. Then I used the raster calculator in Arcgis 10.4 to evaluate all other nodata areas' pixel values into 0 within the processing extent. 

    After that I got the above problem. And the attachment is my logfile. I will send you my latest inputs via email.

    Thank you so much!
    Post edited by uncxlk on
  • jdouglassjdouglass Administrator, NatCap Staff
    I mentioned this over email, but I'll mention it here for posterity.

    There are a couple things at issue here.

    The 'ValueError could not open ...' error is because your threats folder has a folder in it called roads_c as well as a raster file called roads_c.tif.  The first is selected by the model, but it isn't a raster.  This isn't a bug in the model as there are valid formats where a folder called roads_c could represent a valid raster dataset.

    The second issue relates to the pixel values of the threat rasters.  As mentioned in other threads, threat rasters may only have pixel values of 0 or 1.  In your case, you have pixel values of 0 and 1, but your nodata value is set to 0.  As a result, your pixel values are interpreted as nodata and 1.  To resolve this, you'll need to remove the nodata value.
Sign In or Register to comment.