In conditional analysis of spatio-temporal data, either spatial snapshots are analysed using spatial tools (map algebra) and their results considered over time, or time series at locations are analysed using time series analysis tools, and later analysed spatially. For a complete, unconditional analysis we need a new set of tools that generalize the ideas of conditional analysis in space or time to tools that operate on space and time. This requires first an exploration and implementation of all methods that can be useful, new semantics for the function to be used and for all the possible features that can be found, as words like space-time interation or spatio-temporal change are too broad. Inspiration will be sought by the 4D fields analysed in meteorology, where symbols are used for gradients and expected movement. Other relevant terms, such as “change”, “growth”, “move” will be explored in terms of (i) how can they be identified, and (ii) how can they be visualized? This PhD reseach will come up with a complete set implemented of algebraic functions and visualisation primitives for spatio-temporal fields. Primary interest is directed to the analysis of synthesized data and on amazonian rain forest spectral and classified imagery.

Supevisors: Prof.dr. Edzer Pebesma, Prof.dr. Gilberto Camara (not confirmed; possibly reversed); Karine Reis Ferreira (Dr., by then)

this PhD could work at INPE or at IFGI, but spend some time at the other place as well.