 Welcome back to my YouTube channel. In today's video I'm going to do a quick review of the FAPDEM, which is the Copernicus DEM with forest and buildings removed. And I'm going to compare it with SRTM in terms of performance for stream and catchment delineation. This website you can download the data. It comes in zip files. And I'm going to download the DEM for the RUR catchment. And here you can find all the metadata. You can see from the license that you can use this for non-commercial use. Click on the URL to download the zip file and then extract from the zip file the one degree DEM tiles that you need. So I have here four SRTM tiles and four FAPDEM tiles. And I'm going to mosaic them into one roster that covers the study area for which we are going to do the test. So here first for the SRTM, create a virtual layer and I call it SRTM Mosaic. I keep the data in the separate groups for SRTM and FAPDEM so we can easily compare them. Do the same for the FAPDEM. And let's call this one FAPDEM Mosaic. Now let's compare these DEMs by subtracting them with the roster calculator. I'm going to do SRTM minus FAPDEM and I create an on-the-fly roster layer which I call Difference SRTM FAPDEM. This is the result. Let's drag it to the top and let's style it to interpret the result. And here we see that there are some big patches here where there are big differences, both positive and negative. But also other areas have differences which become more visible if I stretch the colors to those areas. And let's add an OpenStreetMap layer to see where those differences occur. And we would expect that to occur in urban areas or in forested areas. I'm going to multiply, use some blending and there we can see that indeed the differences are mostly in the forested areas. Let's switch off the blending. Let's install the Profile tool to check in profiles what the difference is between the two DEMs. Let's do that here for those mines. I'm going to add the layers, give them different colors for the lines. Let's draw cross-section and you can see the difference. Well, this difference can also be caused by the different DEM products that are the original ones which might have a different timestamp. So these mines they move and they are not exactly at the same location in both DEMs. That's also what you can see when there are some shifts. Until now we have used the raw downloaded DEMs, but for further analysis we need to project it. I'm also going to clip it to a bounding box to make it a bit smaller. So here I'm going to export both DEMs to UTM Zone 32 North, the study area. And take boundary from the bounding box and fit to square pixels of 30 meters. Before we continue with the analysis let's have a closer look at this difference raster again to see what value ranges are there. And we can do that with plotting a histogram. And here you see that the peak is at zero, but it's slightly skewed, which we can expect because the fab DEM is corrected for urban and forests, which makes it generally in some places a bit lower. And here we see the range, minimum and maximum value of the difference. So there are quite large differences. Now for the delineation of streams and catchments we're going to use the PC Raster Tools. I'm going to import the DEMs to the PC Raster Format. In other videos I explained how to use the PC Raster Tools plugin to install it. So now both DEMs are in the PC Raster Format. The next step is to calculate for both the flow direction or the local drain direction map with the LDD Create tool. Let's calculate one for the fab DEM and another one for the SRTM DEM. Let's have a look at the results. And we see that the SRTM has much more noise. Well, the fab DEM is smoother. And let's use these flow direction maps then for the delineation of streams and the catchment. So to derive the rivers I can use this flow direction map and the stream order tool. So let's calculate the strahler orders for the fab DEM and do the same for SRTM. Now with the stream orders we can calculate the rivers by using a threshold for strahler. So I'm going to use the spatial tool to create a raster with only value 8. That I'm going to use as a threshold. So later I will calculate all the strahler orders larger equal than 8. Call it ordinal 8. And then I can use this in a condition. Use the comparison operators and as an input I'll use the fab strahler larger equal than 8. And we save it to fab river. And I do the same for the SRTM. So SRTM strahler larger equal than 8. Change the name to SRTM river. And there's the result. Put it in the correct groups. Let's compare the results. So I style it with palleted unique values. Remove the zeros. So we'll have in blue the SRTM river. And then for the fab DEM. Palleted unique values. And remove the zeros and that's in red. And let's compare it with a reference layer from OpenStreetMap. So move this one to the bottom. And there's OpenStreetMap. And now we can compare. So let's zoom in and see if these rivers match with each other and with the background OpenStreetMap. And we see here in the natural area that they're very close together. The blue and the red pixels. Here the red pixels follow maybe better the sinuosity of the meanders there. Well, the blue pixels are a bit straight. But here's something strange. There's a gap. The line is not continuous there. It could be that there's a different Stryla order there that has been filtered output. That would be a bit strange. It looks like that it generates another subcatchment here. And did not really fill that part to have a continuous catchment there. We'll see what effect of that is later when we calculate the catchment. Let's follow it further downstream. And let's see what it does in these urban areas. And we see there that the red line is much closer to the channel in this part. Seems to pick up some more streams that don't really look like that they are there. But here it flows nicely the stream that we can see on OpenStreetMap. Here we can also see that the red line is better than the blue line. Let's have a look closer to the outlet. We see here that the FAPDEM catches nicely the MoseRiver. But it's disconnected there. So let's calculate the catchment of a pixel that is a river in both datasets. Copy the coordinate. Add value 1. Save this comma-separated file. And then we use this to calculate the catchment. But first we need to import this to the pcraster format. Using column file to pcraster map. We use a nominal here. We could also use Boolean or Ordinal. And let's call it Outlet. It's the same for both, so we don't need to distinguish between FAP and SRTM. And we're going to use this outlet for calculating the catchment. So with the catchment tool, use the FAPDEM LDD and the outlet to calculate the catchment for the FAPDEM. And the same for SRTM. Let's zoom to the layer and we can compare them. And there we see that the northern part is quite similar. But the southern part is missing in the FAPDEM. So these discontinuities have an effect that it divides the root catchment into two subcatchments, basically. And we need some further research why this is happening. It looks like the other subcatchment drains to that mine. Let's change the colors to better compare them. Change SRTM to blue. And the FAPDEM catchment to red. And here we see the differences. So in this video we evaluated the FAPDEM and compared it to SRTM for hydrological applications. And I think from this analysis I can conclude that FAPDEM delineates the streams better than SRTM. But for the catchment we need to check it for different circumstances. I think here the mine has a large impact on the delineation of the catchment and divides it in subcatchments. I hope you've enjoyed this video. Please subscribe to the channel if you want to receive updates and looking forward to see you again next time.