This forum is shutting down! Please post new discussions at

Water yield gives out -1 in the AET column

Hello out there!
I ran the water yield model of InVEST and I am confused about the outputs. The results table only shows the value -1 for AET and for water yield_mn and the water yield volume values are way under 0. Do you have an idea where the problem could be? I attached the table and the log file. Thanks in advance!



  • swolnyswolny Member, NatCap Staff
    Hi Lisa -

    One interesting thing about the table values that are showing is that mean precip is always much lower than mean PET, which indicates that in these places there will be no water yield, since more water is being evapotranspired than is provided by rainfall (thus the negative values.) If you want to check this in a quick, coarse way, multiply your precipitation raster by 5 or something, so your precip is generally > PET and see how that affects the results. 

    ~ Stacie
  • DaveDave Member, Administrator, NatCap Staff
    Also it's worth updating to InVEST version 3.6.0 for this model. It got a significant update including some bugfixes. 

  • Hello together,
    thank you for your answers!

    I updated InVEST and reran the model with much higher precipitation values. The result is strange because now the PET values are much lower than with the last try an instead of -1 for AET and water yield I have 0 in every column. Log and table are attached.
  • swolnyswolny Member, NatCap Staff
    What do the per_pixel rasters (aet.tif, wyield.tif) look like? The information that is in the output shapefile is created by aggregating these two rasters within the input watersheds. 

    ~ Stacie
  • hillel_95hillel_95 Member
    edited February 23
    All the per pixel files have no plausible values (-3.40282e+038 to 3.40282e+038) and cannot be displayed...
    Post edited by hillel_95 on
  • swolnyswolny Member, NatCap Staff
    Hm. Can you send me your inputs and I'll test it out? swolny at

    ~ Stacie

  • I took a look at the intermediate files and found that my depth to root restricting layer depth file was somewhat broken. I rebuild it and averything worked fine :)
  • gis4sdgis4sd Member
    Dear all,
    I get a similar error, could you please help me to solve it out.
    Thank you very much.
    844 x 237 - 26K
  • DaveDave Member, Administrator, NatCap Staff
    Hi @gis4sd,

    Could you please also upload your logfile so that we can exactly which version of the model and all the parameters you used to get these results.

  • gis4sdgis4sd Member
    Hi Dave,

    I attached here the log file. Could you please take a look. 

    Thank you

  • gis4sdgis4sd Member

    Iam sorry, all attached files in my previous comments was run with my original data.

    I have just run the model with the resample data that I sent.  I attached all files in this comment, please take a look.
    Thank you very much
  • DaveDave Member, Administrator, NatCap Staff
    Sorry for the delay. I'm working on this now. I noticed you are running version 3.5.0, so I tried your data with version 3.6.0 and I'm getting an error there which may be related. I'll troubleshoot that and post more soon.
  • DaveDave Member, Administrator, NatCap Staff

    Good news! The error I was seeing with 3.6.0 has absolutely nothing to do with your data and everything to do with a *very* specific setup on my end. So, please try out version 3.6.0 with your data. My results show that the -1 problem is resolved in that version.
  • gis4sdgis4sd Member
    edited March 15
    Hi Dave,
    It is really a good new for me. Thank you very much for your help.
    I have a (stupid) question, could you please explain me: 
    I thought that Forest should be better than agriculture or grassland to keep water, is it true?
    - In the water yield model, if I change the type of land use, for example from agricultural land to forest land ((but keep all other parameters), the water yield output per pixel is decreased: because of increasing Kc (in many documents, I found that Kc of forest is higher than grassland).

    In my case study: in a such area, land was grassland or some thing like that in 2000. It changes to forest in 2015. Output wyield_.tif indicate that the area's water yield is decreased.
    So, is there any problems in my thinking? 

    Thank you so much.
    Post edited by gis4sd on
  • swolnyswolny Member, NatCap Staff
    Hi @gis4sd -

    Yes, it is true that forest usually uses more water than grassland. So we assign a higher value of Kc, which leads to more evapotranspiration. The water yield model calculates water yield as (precipitation - evapotranspiration), so if there is more evapotranspiration there will be less water yield. 

    Now, it may also be true in the real world that forests help water infiltrate, so more water may be released as baseflow later in the season. However, this simplified model does not consider baseflow separately. It assumes that any water that is not evapotranspired reaches the stream either by surface or subsurface flow within the year.

    ~ Stacie
  • gis4sdgis4sd Member
    Hi Swolny,
    Thank you very much for your responding. Your answer help me much.
    One again, I would like to express my deeply thank to you - Dave and Swolny
    Wish the best thing to NetCap team.

    All the best,
    Ngo Dang Tri
    From Hanoi, Vietnam 
  • Hi all, 
    I ran water yield scenario and got a very similar issue that a error of zero, I have attached the log file and input data here. Could you please help me with the problem? 
    Thank you very much.

  • swolnyswolny Member, NatCap Staff
    Hi @Songsong -

    Thanks for posting the logfile and data. The problem is that your inputs are not all in the same coordinate system. The watersheds and land use are in WGS_1984_Cylindrical_Equal_Area, while the other inputs are in World_Cylindrical_Equal_Area. Try reprojecting land use and watersheds and try again.

    ~ Stacie

Sign In or Register to comment.