About 99.9% of the work I do takes place in an office or other indoor space. Occasionally, I get the opportunity to go out in the field for ground truthing. Earlier this summer, I had an opportunity to do this on the lower Columbia River for a project I am doing for the
Center for Natural Lands Management (CNLM).
 |
Our ride for the day |
In 2009, when the Washington CNLM office was still a part of the Nature Conservancy, I worked on a habitat analysis for the streaked horned lark (as I wrote about on the blog
here). The purpose of the current project is to update the analysis with 2011 data and increase the number of sites by a factor of three.
For the original study, I used image segmentation software to classify the vegetation from a combination of scanned and georeferenced photos (black and white, infrared, and some true color) as well as true color NAIP orthophotos. For the current study, I am using 2011 near-IR NAIP with ArcGIS to produce supervised and unsupervised vegetation classifications.
 |
The scale of the dredging/deposition operation is massive |
 |
The dredging barge in the channel |
I produced draft classifications for three islands and met my CNLM project manager in Cathlamet to get on an
Environ boat to tour the islands. Environ is a consultant to the
Port of Portland's dredging operations and we were able to tag along with them to gather our field data.
For the ground-truthing I used my Android phone in conjunction with an app called
Locus Pro. It worked really well for gathering GPS data with geo-tagged photos so I could take a point and a photo then load it into ArcGIS afterwards (as a KML) to assess the accuracy of the draft classification. It was extremely helpful to be standing on the islands while holding the draft map because I gained a much clearer sense both of the fine differences between classes and also the sometimes surprising differences between what a class looks like on an image and what it looks like on the ground. And we even saw a few larks!
 |
Moss/lichen vegetation |
 |
Formerly 'Larky' habitat invaded by Scotch broom |
 |
Taking a point |
 |
Taking another point (in a very scenic location!) |
It was a long day, but extremely productive, and I spent the drive back to Seattle thinking of ways to integrate more field work into my GIS projects!
No comments:
Post a Comment