My formal background is in Computational Science and Remote Sensing, and my research focuses on the development and application of computational algorithms for the analysis of spatio-temporal remote sensing, numerical modeling and social media “Big Data” related to environmental hazards and renewable energy. My research falls under the general description of CyberScience and can be summarized as a tripartite of tools, data and topics.
Tools: GeoInformatics is the framework that provides the geospatial analysis tools, primarily based on spatio-temporal data mining, machine learning and artificial intelligence.
Data: Remote sensing, numerical simulations, and social media are the primary source of ‘big data’ used in the analysis.
Topic: Environmental hazards and renewable energy are the challenging problems of high societal impact addressed.
Currently, I work on three projects, and I am constantly looking for venues for new collaboration and students.
1. The fusion of remote sensing, numerical model and social media data during emergencies. Remote sensing is the de-facto standard in observing the Earth and its environment during emergencies, but gaps in the data are inevitable due to sensor limitations of atmospheric opacity.
The goal is to ‘fill the gaps’ in remote sensing observations using social media data, other non-authoritative sources, and numerical models. This project is currently being funded by DOT and by ONR.
2. Using GeoInformatics to optimize numerical model forecasts for renewable energy. Given a single deterministic future forecast, and a history of past forecasts and associated observations, the goal is to build a probabilistic forecast that captures the risk of over- or under-producing electricity.
This project is also partially supported by DoE and Xcel Energy through NCAR, and my code is being used operationally by Excel Energy.
3. The source characterization of unknown and potentially toxic pollutants using remote sensing, numerical models and ground sensor measurements. I reconstructed the non-steady release rate for the radioactive leak at the Fukushima nuclear power plant. The methodology I developed uses an evolutionary (genetic) algorithm guided by machine learning rule induction. This work is currently being supported by ONR.
Other irrelevant interests: My big passion is sailing. I am a volunteer instructor at the US Naval Academy (USNA), in the Offshore Sailing Training Squadron (OSTS) program. I race on the Chesapeake bay about once a week, and I have done several offshore passages in the Atlantic Ocean and in the Mediterranean Sea.
- remote sensing
- environmental hazards
- social media