Slide 5 of 29
- The “hard-edged” people: land ownership records, utilities management: telecomms., transport planning, (military ?).
- The process-oriented studies: dispersion, flow, integration with models which currently only use GIS as a post-hoc results publishing step: need tools and HCI to aid input data management too.
These are both “single truth” groups: the Temporal GIS at all times contains our current best guess at the truth for the temporal dataset. Legal evidence systems might want to compare alternative versions of past events, to discover conflicts and co- evidence, e.g. an oil-spill in fog with ship’s pilot, radar operator, shore and other ship-based observers.
Multiple future worlds of policy implications are also multi-truth too.
- Temporal geo-statistics: I don’t know enough about this to comment.
- Temporal image interpretation (1) clearly vital as past observations and interpretations will aid later interpretation, e.g. forest detected -> clear-felling. (2) “low-level” algorithms of “ 2+1D pixel manipulation: I don’t know enough to comment, but I would guess that MATLAB and Mathematica are better tools.
There is a temporal analogue to generalisation, e.g. generating max.flooded are by day from original hourly data. Cyclic events (migration, crops) also need special treatment.
