Does anyone use Digital Elevation Models (DEM) in GIS software to
determine elevation for designs of gravity fed systems?
Is GIS easier to use than abney levels or theodolites for partners that do not have such skills? Are there any other options that people are using to determine elevations/slopes change?
Within WaterAid, I am not aware of anybody using this. Even outside WaterAid, I doubt the usefulness, as most of the public available DEM have poor resolution (e.g. grid of 1x1km).
GPS are much more commonly used. The normal ones are notoriously inaccurate in the third dimension (height) but more sophisticated have barometers. If these are used properly, these can be sufficiently reliable for preliminary surveys.
However, to accurately calculate pipes needed, breakage tanks, bending of pipes etc., nothing beats a theodolite and if the investment is considerable, a theodolite should be used.
On another note: at least as important is to monitor the source over at least a full year (preferably much longer) to be certain that there is enough water also during the dry season. Failure due to sources running dry is more common than incorrect height measurements.
The DEM data we can get a hold of is a minimum of 25sqkm grids with data points at 5m intervals, so has significantly higher resolution than the 1x1km grids you proposed. The cost is significant, USD1250 per 25sqkm. If we were to reduce this and get data points every 20m, it would cost half the price. I'm wondering if the investment is worth it seeing our partners do not have the technical knowledge to use GIS software (although they don't have surveying skills either).
On another note, the East Timor CP has trialled an iPhone application called Theodolite. From all accounts it seems to be easier to use than an abney level and theodolite. Have any of the other UK CPs utilised such an app and what has their experience been with its accuracy?
Thanks for the advice on water flow monitoring, we are already implementing this in some communities. I have recommended that communities collect water flow data once a fortnight and anecdotally record if there was rain in the preceding days prior to the monitoring to account for quick response of the surface flow compared to constant rate of groundwater flow. I did not want to burden communities by doing too much 'volunteer' work in collecting such data (hence fortnightly monitoring) however do you think we need to increase the frequency of such monitoring?
I've used the SRTM data that John Jacobs mentioned above, the newer release, with about 30m intervals. Overall, the data was great, but you need to beware of some substantial offsets (10-20m), which mean you need to adjust elevations if you do field surveys. This was in DRC, elevations around 800m above sea level, in a fairly hilly area.
As a first approach, it provides a fantastic "preview". You can get the data into QGIS, and design a basic network with the EPAnet plugin, all for free. It is only useful over fairly big distances, otherwise, the noise in the data and the resolution will make your calculations meaningless. Also, you need someone fairly comfortable with GIS systems, as the process is not super easy, so this might be a show stopper for you.
If you're going to invest substantial money in building a network, I'd advise you add some sort of field survey, at which point you're back to abney levels or smartphone apps.
Some good answers here, and I have had some good luck developing pipeline profiles in GIS from DEMs. The precautions mentioned by others here are important, as DEMs often limit a level of precision with elevation and spatial data. The 5m grids you suggested sounds great, but of course you want to be sure of the elevation data accuracy and precision. As with any method of developing a pipeline profile, you want to be sure allow for the margin of error introduced by your elevation data (be it DEMs or abney level). That is, be sure not to get too close to your minimum and maximum elevation tolerances for your pipeline. As Arjen mentioned, theodolites are a reliable method for getting quite accurate profiles, but if you don't have access that kind of analysis, there are less accurate methods can be used with the right provisions.
There are, I believe quite a number of ways in ArcGIS to create a profile from a DEM. The way I have used is with the 3D analyst extension. I am not sure if you have access to the ArcGIS software and extensions but if you do, I can look for some specific instructions to send your way.
This thread is public, all members of KnowledgePoint can read this page.