Speaker: Venkat Lakshmi, John L. Newcomb Professor of Engineering, University of Virginia
In order to study land surface hydrology, we need to use a multitude of tools, namely, modeling, observations, and their synergism. After multiple decades of hydrological modeling, we still have major challenges. However, we have novel observations and mathematical methods that are now available and can be harnessed to achieve progress. These include earth observations that are available at global scales and at high spatial resolutions and with frequent
temporal repeat. Artificial Intelligence and Machine Learning (AI/ML) can be used (specifically Transfer Learning) to determine streamflow in un-gaged or poorly gaged watersheds.
In this talk, Lakshmi will focus on four major questions and provide examples for each of the
questions. These examples will highlight both advances and limitations for each issue.
(i) Do earth observations compare well with in-situ counterparts?
(ii) How important is rainfall in hydrological modeling?
(iii) Can we get a higher spatial resolution of Earth observations?
(iv) How do we study un-gaged/poorly gaged watersheds?