2

I am on QGIS 2.18.2, trying to generate a map that shows all areas with slope steeper than 10% in a specific color (red in this case) and hides the rest. I followed most of the instructions in this post: Slope Analysis in QGIS but I still have some problems, unfortunately.

Here is what I did so far: I got the raw data from the client in the form of points, imported them into QGIS and, per the instructions from the post above, generated a DEM layer using the interpolation tool. The result had a few weird lines in it. I used IDW as an interpolation method, because when I used TIN it looked even worse. Next I did a slope analysis, which resulted in this picture:

slope analysis

My QGIS is in German. The slope layer is called "Neigung". The weird-looking areas around the edges are expected, that data is missing from the raw file (but that's ok, the only data I need is the area of a municipality in the middle of the area). The problem is that I don't really know what unit the output is in. It should be degrees, but I highly doubt that the steepest slope in the area is a mere 4.56 degrees.

For the slope analysis, I left the z-factor at 1. The points and the raster layers are all in Gauss-Kruger, so everything is in meters, if I understood it correctly. And per the above post, the factor should be one if the vertical and horizontal units are identical, which they are in my understanding (the data points give height in meters and are spaced 25m apart in each direction.

I don't know what exactly the slope tool generated. It should be degrees, but the range does not really fit the situation.

Improve this question
5
  • Welcome to GIS SE. If you take minute to go on the tour it'll help to know how to write a good question. You've asked several questions in one and doesn't tend to go down well with the community. As for the straight lines, those I assume look like artifacts retained from mosaic-ing several layers together to create the DEM
    – Knightshound
    Mar 23, 2017 at 8:29
  • Thanks for the comment, I'll edit it down to just one question. As for the mosaic suggestion, it all came in one portion, no mosaicing required.
    – cetequ
    Mar 23, 2017 at 8:36
  • Some of your questions can't be answered without having your data. Artifacts may come from the data. But you can evaluate if the slope is correct by taking heights from two points from your DEM and the distance between the points and using your calculator.
    – user30184
    Mar 23, 2017 at 8:50
  • 1
    If your aim is to visualize areas with slope<10% I would create the slope image directly to use percent as unit. I am not sure if QGIS is using this under the hood gdal.org/gdaldem.html but when you have your dem you can continue with "gdaldem slope" from the command line.
    – user30184
    Mar 23, 2017 at 8:55
  • 3
    Did you try to set the style classes yourself? The problem often is that qgis analyses only parts of the image to find the min and max values for the classification (which is good, as it would otherwise take ages). This might lead to an automatic classification that misses values for display or just shows weird classes. Set a classification on your own and just set two classes with values 10 and 90. Then all below 10 should be in one class and the rest in the second. Then have a look at the image and see if it makes sense.
    – Matte
    Mar 23, 2017 at 9:20

1 Answer 1

Reset to default
4

As it turns out, the comment by Matte hit the nail on the head. The slope dialogue generated values in degrees, but for some reason did not get proper min-max values. By manually setting the categories in the layer properties I was able to fix the problem and generate the map I needed.

The z-factor of one (1) was exactly right for height data in meters and Gauss-Kruger coordinates.

Improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.