Technological advances have created an unprecedented availability of inexpensive sensors able to stream environmental data in real-time for which we still seek appropriate data management technology that can keep up with this onslaught of sampling in previously unavailable spatial and temporal density.

In previous work we have shown that DSEs can be extended to generate smooth representations of continuous spatio-temporal fields sampled by up to 250K sensors on-the-fly in near real-time, creating a new representation every second.

In this paper we have investigated a spatio-temporal stream operator framework that efficiently executes predicate operators over spatio-temporal fields. We introduced a definition of predicates over dynamic fields, analyzed requirements for stream query evaluation and presented several pipelined stream based query operators algorithms. The work is based on the assumption that it is more efficient to find ‘seeds’ of regions that are part of the predicate result and expand them into the complete predicate result regions instead of interpolating the entire continuous phenomenon first, and filtering all cells based on the predicate condition. We investigated different seed expansion algorithms (Breadth First and Scanline region growing, and tile expansion) as well as exploring the impact of using the knowledge of the previous window query result. Our analysis and performance results show that both region growing algorithms perform best for all data set sizes and characteristics; tile-based approaches are efficient for tiles sizes 4×4 and 8×8. History-aware tile expansion performs better if the phenomenon changes slowly (as expected). Future work will include investigating adaptive query evaluation using the different algorithms based on the changing phenomenon characteristics.