This forum is shutting down! Please post new discussions at

Habitat Quality Threat Raster

Hi Guys, perhaps this question might relate to just basic GIS operation. But I am really confused what does it mean in the Habitat Quality Model ? And How to do that? 

LULC Raster: It says that  "The LULC raster should include the area of interest, as well as a buffer of the width of the greatest maximum threat distance.

Threat Raster: "These threat maps should cover the area of interest, as well as a buffer of the width of the greatest maximum threat distance"

I have attached both my LU map and a raster Map as images. They don't overlap in some sections. How to address this. and how do I buffer my raster in what ways. Wouldn't buffer give me unwanted data in my raster? 

1366 x 768 - 198K
1366 x 768 - 164K


  • Also; I get this error while running habitat quality model 

    08/21/2017 12:00:59  osgeo.gdal         ERROR    [errno 1] GTiffDataset::Crystalize:No space for tile arrays
    08/21/2017 12:00:59  osgeo.gdal         ERROR    [errno 1] MissingRequired:TIFF directory is missing required "TileOffsets" field
    08/21/2017 12:01:00  osgeo.gdal         ERROR    [errno 1] TIFFWriteEncodedTile:No space for tile arrays
  • swolnyswolny Member, NatCap Staff
    Hi @rosean -

    The recommendation for these rasters to also include "a buffer of the width of the greatest maximum threat distance" is intended to make sure that you are modeling all of the threats that are likely to impact your area of interest. For example, there might be a road that lies 100m outside of your LULC map, but if it has a MAX_DIST of, say, 1km, that road can impact your area of interest. If you do not include that road in the analysis (with a buffer), you're not representing all of the threats to the habitat, so will overestimate the habitat quality/underestimate the habitat degradation.

    Of course, to represent these, you need to have LULC and threat data that covers a wider area than your analysis area. If you do not, then you'll need to note this limitation in your results. If you do have this extra data, then you can create a buffer around your current LULC map (the one that represents your actual area of interest) and use that buffer to re-clip the threats and LULC data to put into the model.

    If there are places where the threats and LULC don't overlap, I'm actually not sure how the model handles it. In other models, if there is NoData in one layer, the result will be NoData. But in this model it might just not have the threats applied - I'll have to let the software team respond to that, as well as the error you're getting.

    ~ Stacie
  • RichRich Administrator, NatCap Staff
    That "TIFFWriteEncodedTile:No space for tile arrays" is indicative of running out of disk space.  I don't know how big your data are, but can you check to see that you have enough space on your working drive?  The model will make a bunch of intermediate rasters depending on your combination of inputs.  
  • Many thanks @swolny and @Rich for the prompt response. This is one of the reasons I like to use InVEST; you guys are super supportive. 

    I now understand the buffer thing and will use it that way. About the "TiffWriteEncodedTile: No space for tile arrays"; It turned about that among the many threat rasters that I was using; one of them didn't have a proper projection system :)  I matched the projection and now have my results for the habitat model. This might be useful for others to note as well. 
  • swolnyswolny Member, NatCap Staff
    Hey Rich, how does Habitat Quality handle places where threats don't overlap the LULC? Does it produce NoData in those areas in the results? Or is there data in the results, it just doesn't have any threats applied?

    ~ Stacie
  • RichRich Administrator, NatCap Staff
    edited August 2017
    Hi Stacie, the threat vectors are trimmed to the extent of the landcover raster.  So you won't see anything in the case where threats don't overlap the LULC, 
    Post edited by Rich on
  • RichRich Administrator, NatCap Staff
    Hi Stacie, we emailed on the side, and what I said above is not true.  The threat maps should cover the area of interest, as well as a buffer of the width of the greatest maximum threat distance if there are threats outside the landcover map extents that may otherwise affect it.  The final result will be clipped and rescaled to be the same shape and cell size of the landcover map.
Sign In or Register to comment.