ArcGIS 10: Stream Modeling with LIDAR

3414
2
07-27-2011 02:10 PM
Labels (1)
PeterPierce
New Contributor
Greetings,
Once again, I'm fairly new to the hydrologic modeling scene, so any input is appreciated. I'm running ArcGIS 10 with Spatial Analyst (with the hydrology tool set) as well as ArcHydro now. The group I'm working with is trying to generate an Environmental Benefits Index that includes some erosion focused layers that focus on field scale analyses, including localized drainage pathways and proximity and a stream power index layer (flow accumulation and Slope). My thinking is that we may break down the streams further into perennial, intermittent, and ephemeral streams in order to effectively describe erosional conditions (giving different weights to each for the index).

However, intermittent and ephemeral stream layers are not something generally available, so it sounds like I'll have to generate them. I tried out both hydrology tool set and archydro 2.0, with the former generating no results and the later unsatisfactory. Looked up TauDEM, but it lacks GUI but not command line (http://hydrology.usu.edu/taudem/taudem5.0/index.html). As we're a non-profit group, we're limited to using 10, and I don't have any experience with python or command line, and another forum post indicated that there were some issues with python versus command line (is TauDEM even functional with python?).

Any advice on how I can solve this with the tools at hand or any suggestions on literature?
Tags (2)
0 Kudos
2 Replies
KevinMacLeod1
Occasional Contributor III
Greetings,
Once again, I'm fairly new to the hydrologic modeling scene, so any input is appreciated. I'm running ArcGIS 10 with Spatial Analyst (with the hydrology tool set) as well as ArcHydro now. The group I'm working with is trying to generate an Environmental Benefits Index that includes some erosion focused layers that focus on field scale analyses, including localized drainage pathways and proximity and a stream power index layer (flow accumulation and Slope). My thinking is that we may break down the streams further into perennial, intermittent, and ephemeral streams in order to effectively describe erosional conditions (giving different weights to each for the index).

However, intermittent and ephemeral stream layers are not something generally available, so it sounds like I'll have to generate them. I tried out both hydrology tool set and archydro 2.0, with the former generating no results and the later unsatisfactory. Looked up TauDEM, but it lacks GUI but not command line (http://hydrology.usu.edu/taudem/taudem5.0/index.html). As we're a non-profit group, we're limited to using 10, and I don't have any experience with python or command line, and another forum post indicated that there were some issues with python versus command line (is TauDEM even functional with python?).

Any advice on how I can solve this with the tools at hand or any suggestions on literature?


Hi PeterP, I am interested in this same topic.

To my knowledge, we can not get fine-scale accurate delineation of the three flow duration stream types (perennial/ephemeral/intermittent) without fieldwork. I am researching this for my thesis. I presumed you mean small streams versus rivers from your use of the word stream. This is because there is much natural variability of hydrology in the field, based on hydrogeology, land cover and use, etc, using just one variable like flow accumulation (drainage area) will usually not provide an accurate stream network. (I have seen this myself by checking such networks in the field).

Data from a few study watersheds is needed to create "training" data and reference data, to base and test a statistical model on. A statistical model predicts the probability per pixel as to how likely each pixel is one of the three categories. This is as opposed to hydrologic modeling that conserves mass and attempts to predict the exact flow volume and timing (ie, creates a hydrograph of points on the streams), which would likely be prohibitively complex (e.g. BASINS/HSPF, ArcSWAT, other engineering programs, GeoRAS/HMS, etc).

You can delineate the origins of each type in the field using a method like the NC Stream Identification Method to define the transitions between the three flow types. This is similar to wetland delineation. Then, you can model new sites, using variables including slope, curvature and flow accumulation, as well as other remotely sensed data. The problem is that this is a very fine scale analysis. You will need high-resolution lidar DEMs that are conditioned properly for hydro analyses. (For example a DEM with <5m resolution).

The leading research on this topic is here (North Carolina): http://portal.ncdenr.org/web/wq/swp/ws/pdu/hwssites   and also by USGS in MA here http://water.usgs.gov/osw/streamstats/massachusetts.html

Also I am experimenting with TauDEM. It has some very interesting possibilities, but I have not come to any conclusions yet. Tarboton's papers are excellent on explaining TauDEM and the hydro principles behind it.

Good luck, I hope this helps!  I'm glad others are looking into this topic! My hope is if the USGS launched a national-scale project similar to the one in North Carolina, for each ecoregion, we could increase accuracy of the "Blue lines" on USGS topos and begin to delineate channels with the three types of flow. This in turn would facilitate community planning by DOT and other large public and private entities, when considering buffer zones for each type of stream.

If you have time, let us know how your work turns out!

Thanks,

Kevin M
0 Kudos
JayChristensen
New Contributor
I've used TauDEM using the python code in Arc 10 for the past few months with LiDAR data. You can place the python scripts into the Arctoolbox just as you would with a tool or modelbuilder. I don't know if it has all the functionality of the complete TauDEM for 9.3. The help and explanations are either abbreviated or not there, but help can be found in the supporting documentation. I've primarily used it to take advantage of the parallel processing (very helpful with the larger LiDAR datasets) and the multiple flow direction capability of D-Infinity and haven't encountered issues thus far.  The area I have been working on is too flat to realistically use the stream delineation techniques included in TauDEM. Some good resources on terrain-processing for stream flow analysis can be found in Wilson (2012) Geomorphology 137:107-121 or a flow-analysis book chapter from Tarboton and Baker 2008 found on the TauDEM website. Recent efforts of headwater stream modeling include Sofia et al. (2011) Hydrology and Earth System Sciences 15:1387-1402, Jaeger et al. (2007) Environmental Management 40:775-786, and Heine et al. (2004) Annals of the Association of American Geographers 94:477-490.  Hope this helps, Jay
0 Kudos