As an earth system scientist and ecologist I model ecosystem processes, such as vegetation growth or drought / disturbance resistance, using various (retrospective) proxy measurements in a model data fusion approach. Throughout my career I worked interdisciplinary borrowing heavily from fields, outside of ecology, such as image vision processing (computer science), remote sensing and engineering to assist in either field measurements and/or model driven analysis.
Most of my work is aimed at predicting rhythmical changes in vegetation growth, i.e. vegetation phenology, using cheap imaging sensors (PhenoCams). The rhythmical timing of vegetation phenology is intimately linked to yearly variations in weather and long-term changes in climate. Not only does the seasonal development of vegetation serve as an important climate change indicator, the seasonal plant carbon uptake also provides an important feedback to the Earth system.
African tropical forest drought
I’m interested in understanding the influence of drought on the productivity of African tropical forests. Here, I rely on retrospective analysis of wood core stable isotopes and historical phenological and meteorological observations to reconstruct and model past vegetation responses to year-to-year variability in precipitation. A large part of this research is supported through the Jungle Rhythms crowd-sourcing project. A newly funded project, COBECORE, will provide matching climate data throughout the Congo basin supporting further spatial analysis.
I collaborate with the International Food Policy Research Institute (IFPRI) where I designed the protocols for a cellphone camera based methodology to track crop growth and disturbance in the fields of smallholder farmers. This research supports index-based micro-insurance, to alleviate income variability for these farmers.
Open source development and outreach
A large part of my operational code and hardware is open source as I’m a avid supporter of the open source movement. Although this does not constitute formal research, it is my opinion that contributing to the open source research community is key in ensuring reproducibility and fast research development.
Of equal importance is my scientific outreach. My personal blog communicates my research to a broad audience as do some of my popular science projects. These projects include a cheap raspberry pi (<$50) PhenoCam for classrooms (or researchers on a budget) and the Virtual Forest project which uses the latest VR technology to transport people to Harvard Forest using a live 360 image stream (reaching ~ 20,000 people in < 3 days and featured on Motherboard.tv).