Sign in

Joseph Kasprzyk

@josephrobertbob.bsky.social
196 followers 43 following 9 posts

Associate Professor and Associate Chair, Civil Environmental and Architectural Engineering, University of Colorado Boulder. Posts about water resources planning and management, decision making under deep uncertainty, motorsports, chess, and other things!

PostsRepliesMedia
Joseph Kasprzyk @josephrobertbob.bsky.social · 13/12/2024
#AGU24 This afternoon see me at poster 1316: a multi objective test problem in Borg-RiverWare, with Edie Zagona, David Neumann, and Patrick Lynn from CADSWES
010
Joseph Kasprzyk @josephrobertbob.bsky.social · 12/12/2024
I recently told the true #AGU24 story about how I was worried about carrying the poster I made myself on my Frontier flight. My colleagues were shocked about this mundane behavior and it gave me the idea to start a #NoEgoHydrologist account. @mgooseff.bsky.social do you approve?
320
Joseph Kasprzyk @josephrobertbob.bsky.social · 12/12/2024
The Post-2026 Web Tool uses DMDU concepts to help anyone explore tradeoffs in alternative #ColoradoRiver management plans, made by Virga Labs in partnership with Reclamation and @cuboulder.bsky.social CADSWES research www.crbpost2026dmdu.org
crbpost2026dmdu.org
Home | DMDU
This is a flexible, powerful platform designed to allow Colorado River Basin stakeholders and the public to explore operational strategies for Lake Powell and Lake Mead as part of the Post-2026 National Environmental Policy Act (NEPA) Process. This tool is intended to support early-stage exploration of operational strategies that may eventually be incorporated into the Post-2026 NEPA alternatives. For more information on this NEPA process, see Colorado River Post-2026 Operations.
000
Joseph Kasprzyk @josephrobertbob.bsky.social · 12/12/2024
In case you missed my #AGU24 Monday poster with @aditibhaskar.bsky.social and Jorge Santiago Ramírez Núñez, here it is! Using CNN image classification to predict flow presence and absence at West Stroh Gulch.
Deep Learning Flow Detection in Non-Perennial Streams. CNN does a good job in an initial test, but work needs to be done in matching the training protocol (every image one at a time) with the use case (finding evidence of flow over a long period of time).
111
Joseph Kasprzyk @josephrobertbob.bsky.social · 12/12/2024
Hello World
030