Associate Professor Paul Siqueira of the Electrical and Computer Engineering Department has received a Charles Bullard Fellowship in Forest Research, worth more than $40,000 for the 2012-13 academic year, from Harvard University. The Bullard Fellowship program supports “advanced research and study by individuals who show promise of making an important contribution, either as scholars or administrators, to forestry and forest-related subjects from biology to earth sciences, economics, politics, administration, or law.” Professor Siqueira’s specialty is in the design, development, and use of remote sensing techniques for applications in terrestrial ecosystems. The base for Siqueira’s Bullard Fellowship research will be at the Harvard Forest in Petersham, Massachusetts.
Bullard Fellowships are generally awarded to individuals in mid-career who have established themselves in public service, in academia, or in the private sector. The fellowship is a highly competitive program which accepts only five to seven recipients per year. Judgments of the fellowship committee are based primarily on the quality of the applicant's professional accomplishments and academic record.
Before coming to UMass in 2005, Siqueira was at the NASA Jet Propulsion Laboratory in Pasadena, California, where he worked on the engineering of airborne and space-borne microwave remote sensing systems and their application to earth sciences. Among many important projects he worked on were the Shuttle Radar Topography Mission and the Japanese Aerospace Exploration Agency’s Global 2 Rainforest Mapping Project.
Siqueira, who will be on sabbatical leave from UMass Amherst during the upcoming academic year, will be doing his research at Harvard Forest on a probabilistic model of vegetation structure and biomass that would help to characterize forests on a global basis by creating a critical link between detailed ecological models and remote sensing data.
Harvard Forest is well-known in the scientific community as a test-site for the characterization of forest structure through remote sensing. Characterizing vegetation on a global basis is a complex task with many far-reaching benefits, including the study of animal and plant habitats, monitoring the impact of population, economic pressures, and climate change on the world’s forests, and making ecological and environmental projections.
The state-of-the-art technology for using remote sensing to characterize vegetation structure is primarily lidar and radar. While there has been a considerable degree of success in using these two technologies, the global application is severely limited, thus making it almost impossible to design reliable space-borne remote sensing systems capable of meeting our international needs for forest monitoring.
The missing component is a frame of reference based on an ecosystem’s inherent physical and environmental constraints, such as water resources, light availability, and soil type. Models that take such constraints into account do exist, ranging from regional climate-type models such as the ED2 model, to individual tree-based models, known as IBMs.
What has not been accomplished thus far, to a high degree of resolution, is the use of remote sensing and IBMs in a combined probabilistic model. Such a model would simulate possible trajectories of the forest environment and project what the remote sensing system would observe. A model such as this would serve as a key link between remote sensing data and the larger-scale models used for regional assessments and projections. (April 2012)