I received an AGU Outstanding Student Paper Award for my presentation at the 2017 AGU Fall Meeting in New Orleans, Louisiana. My talk was titled 'Tracking fine-scale seasonal evolution of surface water extent in Central Alaska and the Canadian Shield' and focused on my current research on understanding fine-scale seasonal changes in surface water across the North American Arctic using Planet CubeSat imagery. I created the above animation for this talk to illustrate seasonal declines in surface water extent in the Yukon Flats, Alaska, as viewed with CubeSat imagery.
Our paper presenting a first demonstration of using Planet Labs CubeSats to track fine-scale changes in surface water extent has recently been published in Remote Sensing. CubeSats, tiny, satellites the size of a loaf of bread, represent a major potential advance in remote sensing due to their ability to image the entire planet at 3-6m resolution every day. However, due to the cheap sensors used, they have significant quality limitations and therefore remain generally unused within the hydrology research community. In this paper we present a new method for detecting surface water extent from variable quality CubeSat imagery and validate our water extraction method using concurrent WorldView imagery. We additionally apply our method over a small 625 km2 area of the Yukon Flats, Alaska and analyze the hydrologic connectivity of the floodplain.
- Surface water area can be reliably extracted from Planet imagery with an RMSE of <11%.
- Observed changes in lake extent in the Yukon Flats in summer 2016 are spatially heterogeneous and reflect a complex interplay of flow paths, underlying geology and permafrost presence.
- While cross-sensor consistency, automated cloud masking and geolocation problems still need to be resolved, Planet CubeSat imagery represents a valuable new resource for hydrologic remote sensing.
I also had the opportunity to give a talk about my CubeSat research at the 2017 AGU conference in December and speak at the Planet Town Hall session, a summary of which can be found here.
My masters research on rapid supraglacial lake drainage events in West Greenland has just been published in Journal of Geophysical Research - Earth Surface.
- We present a new method for detecting rapid drainage events and find that drainage detection is strongly dependent on frequency of observation
- Observation bias correction yields a predicted rapid drainage percentage of 36-45%, which is 20-30% higher than non bias-corrected estimates
- Inland lakes at high elevations (>1600m) drain as frequently as lakes at low elevations
This summer I had the opportunity to spend 6 weeks conducting fieldwork outside of Kangerlussuaq, Greenland with a team from UCLA and Rutgers University. Our work this summer was the second field campaign of a NASA-funded three-year grant to assess drainage efficiency of supraglacial rivers on the Greenland ice sheet. This summer’s research had two main parts. First, six members of the group conducted a helicopter-supported 10-day ice camp in the ablation zone of the ice sheet where research primarily focused on measuring the discharge of a large supraglacial river. Our group also maintained a 2.5 month-long camp at the ice sheet margin where we studied the seasonal evolution of the surface hydrology of the sheet margin.
I primarily worked at the camp at the ice sheet margin, and each day we hiked 3-4 miles each way to our various study sites on the ice sheet, measuring supraglacial stream discharge and conducting experiments to assess the variability of the water content of the ice. The work was very demanding both physically and mentally, but the scenery was absolutely incredible. The turquoise blue water of supraglacial streams and steep walls of ice canyons never ceased to amaze me. We often joked we had the prettiest study site in Greenland!
For more information about our work this summer, check out the photos below, this piece in the New Yorker by Elizabeth Kolbert, our NASA blog, a video about our work made by climate science videographer Peter Sinclair.