FOSS4G North America Conference 2013 Preliminary Program

carbon on snow can alter the energy equation and affect overall melt rates. Products generated
by the system are generally available less than 24 hours after original upstream data acquisition
by the satellite, and are distributed to water resource managers and other principle decision
makers on a daily basis.
The software infrastructure behind the Snow Data System is overwhelmingly comprised of open
source software. From the core data management components that utilize Apache OODT, to
data persistence and transformation via Postgres, PostGIS, and GDAL, to front­end interaction
and visualization tools that leverage tools like GeoServer and Leaflet, the Snow Data System is a
model of full­stack commitment to leveraging open source software. This talk will discuss the
way the strengths of these open source software products were combined to deliver a powerful
near­real­time data processing infrastructure and the accompanying challenges, lessons, and
rewards of having chosen the open­source route.
Selling Open Source to the Census Bureau
Dan Little, Excensus LLC
The twitter verse is crowded with complaints of closed data from the U.S. Census Bureau. We
have been working with them for over 10 years and have a team legitimately concerned about
data accessibility. We have pushed for the inclusion of a MapServer, PostGIS, Python, and
C­based stack for major applications. We have been very successful using, deploying, and
gaining acceptance for our solutions because they work. This has been an evolutionary process
and we've met basic road blocks along the way. This presentation will provide some insight to
that process and describe ways we were able to offer proof to our clients that allowed them to
take the "risk on open source."
Landscape Scale LiDAR ­ based Forest Structure Analysis
Doug Newcomb, U.S Fish and Wildlife Service
In an effort to tie landscape scale forest structure data to localized bird species preferences,
25.5 billion muliple return LiDAR points of varying densities collected for floodplain mapping were
converted to seven 3.3 billion point LAS files and normalized to a 6.096m ( 20ft ) statewide
elevation grid. The normalized files were then simultaneously analyzed for different forest
canopy metrics using GRASS 7.0 r.in.lidar to a series of 753661000 cell, 18.288m (60ft)
resolution grids and compared to polygons of nest locations (25m buffer) of different bird
species.
17 output grids were separately analyzed using a 17 dimension cluster analysis algorithm on the
Oak Ridge National Laboratory Titan Supercomputer to 50 and 100 clusters. Output of the
cluster analysis was used to identify areas of inconsistent LiDAR data collection and clusters of
species habitat preference.
Deploying Open Source Software for Shared OGC Service Hosting at the National