[Ray Davis]

Hi Folks,
The Ohio River Forecast Center is working on a large HEC RAS model that covers the entire Ohio River mainstem (about 1,000 miles long,) and about 100 miles both upstream and downstream of the Ohio/Mississippi confluence. That's about 1,200 miles of cross-section data! Yikes! Considering that we've already processed a large amount of data, I'm wondering what's the most appropriate map projection to use for processing DEM data.

Our multi-agency working committee agreed to use US Albers Equal Area (USGS version) several months ago. We decided against using UTM because our project area covers three zones (15, 16, and 17.)

Last week, we consulted with some folks from the USACE Hydrologic Engineering Center (HEC) and they feel we ought to be using a conformal projection like UTM. They thought we'd run into river-length distortions.

I did a quick distance check, comparing two lat-long locations projected to UTM zone 16 and to Albers. One point was at Pittsburgh, the other near Paducah. There was a ~3.2 mile difference.

UTM Zone 16 distance: 507.5496 miles
Albers distance: 504.3068 miles

I then checked a number of points on the Ohio River, connecting maybe 20 dots from upstream to downstream. I got a difference of ~1.4 miles.

UTM Zone 16 distance: 662.7969 miles
Albers distance: 661.4064 miles

We've got about 1,000 cross-sections between the upstream and downstream ends of this stretch of the Ohio. My gut feeling is that any length errors would be very small between the cross-sections. Further, the total length error along the mainstem of the Ohio River would be very small.

I'm certainly considering going with the HEC's suggestion to use UTM 16. However, I'd sure hate to have to reprocess 15-20 GB's of DEM data. Help!

Thanks,

Ray