Hello! I'm studying similiar things for the mountain area. I found for myself next things.

First of all, ASTER is really raw DEM with big amount of artefacts and errors especially at the north. I decided not to use it at all. even despite the fact that it is 30m.

ALOS is much more better, but there is a holes sometimes or black areas. Somewhere there can be really big. I've attached an example with this NoData areas. You should be accurate with it.

The best from your list is SRTM of course. Even despite it 90m resolution it is really good. I suggest to use already prepared DEMs (hydrologically condicted) from HYDROSHEDS. Less then one month ago they've got a update and it is great.

On the other hand you can use ALOS DEM not HYDROSHEDS if your watershed is really small (less then 100 km2). However don't forget to prepare it. Filter and fill sinks in SAGA for example.

For really small watersheds (around 20 km2) I found for myself that the digitazing of topomaps is the best way.

http://hydrosheds.org/

Hello! 

You can read some details about SRTM elevation data here: https://lta.cr.usgs.gov/SRTM1Arc

I tend to use USGS Earth Explorer to access and download DEMs with spatial resolution 30 m. 

https://lta.cr.usgs.gov/SRTM1Arc

https://earthexplorer.usgs.gov/

For small catchments I strongly suggest flying a drone for dem restitution, if possible. In arid regions, lidar data, if available, is really convenient. 90m may be well too small depending on what do you consider "small watershed". I tend not to use corrected dems because it limits your knowledge of what is adequate to "clean" given your specific issue to be addressed. Check USGS Earth Explorer as Petko suggests. When nothing better is at hand almost anything is worth testing.

Thanks all for your very kind answers.

In fact; my watershed is about 150 Km2. I've downloaded ASTER GDEM (30m), SRTM (30 m) and Alos Palsar DEM (12.5 m). I found that the aster GDEM have more coarse comparing to others, and I decided to not use it. The SRTM and Alos Polsar DEM seems to be Identical. but I observed that, in the same area, the Alos Polsar DEMs give higher altitude value than the SRTM DEM. any explanation for this fact.

Miguel Llorente, It's not possible to use Drone in my region and Lidar data are not available.

All the DEM that you mentioned are not fine resolution, you can go for hydraulic modeling to be useful for planning, won't able to obtain site specific detailed flood scenario. Of course, 12.5 m resolution Alos DEM is better. Even it you get higher elevation, I suppose it won't be problem because the relative change in value is important .

Hi,

The difference in elevation between the two DEMs can be explained by the different interpolation methods that were used to create the DEMs. But it should not be a big difference though. For your study type (hydrological application) I would use a DEMs that was created by Topo-to-raster interpolation. 

For the few SRTM and ALOS Palsar DEM data sets that I have recently worked with it appears as Ali stated that they are pretty much identical.  I have not investigated the source of the 12.5m Palsar DEMs but am thinking that perhaps they are generated by 'up sampling' the 30m SRTM data.  Keep in mind that the original SRTM data are at a 60m resolution and the 30m version is derived from those data.  Therefore, perhaps both the 30m SRTM and 12.5m Palsar data sets are simple up sampling of the 60m SRTM data set.  If so then what should be said about the 30m and 12.5m data sets is that they have 30m and 12.5m 'pixels' not resolution?  In my analyses I did shaded relief (hill shade) and 1st derivatives of both data sets and the amount of high frequency information is close to identical; you would expect the 12.5m data to have more / better high frequency information (that is, better detail).

Pat

Dear all, I suggest you to read this one paper about comparing different DEM's. Shortly authors concluded that in very mountainius areas ASTER v2 is really good.

Article Comparison of free high resolution digital elevation data se...

Dear all, in this link a description of Alos Palsar product. the document affirm, if I don't misunderstand, what Pat S. Chavez, Jr stated. The Alos Palsar DEMs are SRTM DEMs of 30 m resolution up-sampled to a resolution of 12.5 m. Except for the US.

Regards.

https://media.asf.alaska.edu/uploads/RTC/rtc_product_guide_v1.2.pdf