- This event has passed.
Colloquium: Colin Gleason, Univ. of Massachusetts, Amherst
November 15, 2018 @ 2:00 pm - 3:30 pmFree
Join us on Thursday, November 15, for a special Thursday KU Geography & Atmospheric Science colloquium, as we welcome Colin Gleason, Civil and Environmental Engineering, University of Massachusetts, Amherst. He will present:
“Combining big-data remote sensing, AMHG, and river routing to estimate daily discharge over an entire river network: a SWOT template”
2:00 pm: Meet and greet in 205 Lindley Hall.
2:30 p.m.: Colloquium in 317 Lindley Hall.
Refreshments will be provided. All are welcome to attend.
Remote sensing is gaining attention as a viable path toward estimating discharge in ungauged basins, and the Mass conserved Flow Law Inversion (McFLI) approach has recently been successfully demonstrated for truly ungauged basins. However, this technique suffers from two major drawbacks: 1) Existing satellites lead to temporally and spatially sparse discharge estimates, and 2) McFLI algorithms consider all reaches as independent hydrologic entities, making river networks subject to unrealistic flow discontinuities. This talk proposes a solution to both problems. We establish a 58-reach topological river network over the Missouri Basin, and then treat the Missouri as an ungauged basin: we use 30 USGS gauges exclusively for validation and invoke no other in situ data. We first route three globally available runoff products through the network without calibration to produce daily discharges in all channels to form a base case that reflects current logic in ungauged discharge estimation. We then estimate discharge via the Bayesian AMHG-Manning (BAM) McFLI algorithm from ~1M width measurements derived from 12,000 Landsat scenes in Google Earth Engine. Routing runoff alone produces hydrographs with less than 50% nRMSE at just 3/30 gauges, and including McFLI flows improves discharge nRMSE at 21/30 gauges. We also discuss alternate error metrics and their utility. Although this approach can be improved, as detailed herein, it is truly applicable globally because it is computationally efficient and uses no in situ data. The results presented here can serve as a template for global remote sensing of discharge in ungauged basins.