Hello, listserv!

I'm trying to do a stream network analysis, and I was hoping you fine folks could give me some pointers. Here's the problem:

We (myself and DEEP inland fisheries) are trying to quantify the effect of stream culverts on fish travel. We have fish sample data, more for the culverts, point shapefiles for each, and the high-res NHD stream geodatabase (with geometric network). What we want to be able to do is trace up the stream(s) a certain distance from the culvert, collect any fish sample data, trace down the stream(s) the same distance and collect the fish sample data, then get a table with the culvert data, fish sample data, and the distance with the upstream/downstream spatial relationships. This part is key. We also want the ability to stop the trace at a culvert.

I've attached a PDF with some figures to help show what I mean. The first page shows the trace we're trying to do, and the second shows the result we want. The third page shows a scenario where we want to stop the trace at a culvert. The fourth shows a situation where we need to trace along multiple streams. (For this one, not only would we want our table to show that the samples are upstream, but also that they're on separate tributaries). The fifth page shows why we don't want a simple distance analysis; it needs to be along the streams. The sixth shows an example of being able to limit the distance.

The big thing is that this process needs to be automated.  We have data for a decent chunk of the state, and it won't be practical to manually go through each culvert. The output needs to be something that can be dumped into a statistics program with minimal data processing.

I've attached copies of the culvert and fish sample point shapefiles. NHD database can be found here [http://nhd.usgs.gov/data.html]. Any pointers would be greatly appreciated! Please contact me for more info or anything else.

Greg