Recently DEMs have been the most sort out solution for drainage and stream network studies. The recent availability of numerous global open source DEMs provides immense potential for drainage-based studies. Yet, the vertical and spatial resolution of the DEMs plays a critical role in the accuracy of the generated stream networks in addition to the algorithm used for flow direction and the accumulation threshold. Especially, while generating a stream network with Strahler stream order, how do we validate the extracted stream orders (validation using a topo derived stream network is valid - yet the toposheets would be small scale leading to missing out on first/second-order streams, which eventually would influence the number of streams order of a particular river basin). Is there any other way of validating DEM-derived stream orders?
I would also appreciate an opinion on how to decide on the number of stream orders for a particular river basin derived from DEMs. If toposheet is the only option which would be older and small scale, then how do we justify a higher stream order derived by using a high-resolution DEM - for eg., deriving an 8th order stream network instead of a 6h order in toposheet.
DEM approaches have made watershed delineation and creating of stream networks relatively easy. However, there are some drawbacks due to the fundamental nature of DEMs;
-DEM datasets are usually one-off creations and do not take into account, the topographic changes that would occur weeks, months, years, or decades later.
- The resolutions versus the scope of the area under consideration as you've rightly mentioned can be a challenge. A case in point is trying to use a 30m resolution DEM for the analysis of a 30 km2 area. There are bound to be generalizations or over/underestimations.
- If there are no straightforward means of directly validating the stream orders, one can take one step back and validate the DEM instead. A simple approach would be to take a GPS receiver to the study area and randomly take ground-truthing points. Some statistical measures like RMSE, correlation analysis, etc can then be used to determine the accuracy of the DEM and by extension the DEM-derived stream order
i am not sure if the resolution of 10 m (30m) and above is accurate enough to provide good networks to answer hydrologic related questions.
I was working with DEMs with a resolution of 1 m and derived the watersheds. Afterwards i had to verify it visually as well (with orthophotos).
The stream order is easily calculated with GIS products like QGis or ArcMap, both calculate very good results.
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India