uniomni
commented
13 years ago TsuDAT relies on the availability of a Probabilistic Tsunami Hazard Assessment (PTHA) which provides sea-ward boundary conditions for ANUGA with associated probabilities. As the PTHA for Indonesia won't be ready until late this year, we have identified some logical steps that will make the process incremental and also allow us to demonstrate the vision to stakeholders. We therefore propose the following path
- Upload tsunami inundation map for Monado projected to WGS84 geographic. This data was precomputed using ANUGA and manually reprojected for ease of use with Risiko.
- Get freely available population density dataset from CIESIN: http://sedac.ciesin.columbia.edu/gpw/global.jsp and bundle it up with the risiko_demo_data
- Write a plugin that will tag the cells where population is exposed to more than 1m of water and count the total. The plugin should return a caption to this effect (following the procedure outlined in the HKV flood study and to be done according to ticket #88) This three steps would form the first demonstration
Next step is then to it all over, but with the native inundation map from ANUGA which is projected to UTM. If this turns out to cause problems (as is stipulated in ticket #40), then this will be a priority for this task to reproduce the issue in unit tests (both within Risiko and for a stand alone GeoServer) and work with the community to get it fixed. If, on the other hand, things go smoothly, then we have another nice unit test verifying this capability and can move forward assured that projections will be handled.
The next step is then to upload the tsunami inundation map to Tsudat and setup a demonstration that although the calculation is not done with Tsudat the resulting layers can indeed be used in Risiko. This would form a nice center piece for a second demonstration to stakeholders to communicate where we are going.
Meanwhile, we are planning to also do a tsunami inundation study for Padang (still manual without TsuDAT) and this layer would then form the next study, namely tsunami impact not just to population data as with Manado, but also buildings and this would be a second plugin implementing a simple tagging of buildings according to x < 1m, 1m <= x < 3m and x >= 3m. Each class would have a percentage of collapse or associated with it which would be shown on the resulting map. Again, there would probably be a caption providing some global statistics (e.g. total loss for this event is 215 billion IDR). When this is working, we have a third demonstration and by then also confidence that TsuDAT and Risiko can interoperate so that we are ready to roll when the PTHA becomes available!
jj0hns0n
nickhorspool
ingenieroariel
AIFDR is currently supporting a program to provide TsuDAT with the required data for Indonesia. In preparation for this, and to build off the momentum that has been generated with BNPB through the Australian training workshop in June 2011, we need to be able to demonstrate how a tsunami impact analysis can be done in Risk in a Box using:
1) Implementation and testing of new tsunami impact plugins:
2) (This is for the TSUDAT end): The keywords should match those of the plugin, so that’ll be (for the moment): category: hazard subcategory: tsunami
3) Display the resultant damage data reliably and meaningfully (https://github.com/AIFDR/riab/issues/82). There is currently no design for this, but in the case of the building damage function above, how do we determine what field to colour by and what do we put in the legend. It could e.g. be done using keywords too. Associated with that is the need to provide a dynamically generated legend (https://github.com/AIFDR/riab/issues/35).
4) Allow caption keyword to be displayed in text box e.g. under the selection combo (was issue #77 now https://github.com/AIFDR/riab/issues/88). It is almost there and would along with styling, legends and feature selection make Risiko useful for real life work. That way, the tsunami risk plugin could output things like “Total estimated loss: $xxxx”. See the HKV flood study plugin for an example (https://github.com/AIFDR/riab/blob/master/impact/plugins/flood/HKV_flood_study.py).
5) Take this opportunity to make sure that data projected to e.g. UTM (as is the output from ANUGA) can be ingested and used in Risiko (https://github.com/AIFDR/riab/issues/40)
6) Make use of the feature selection ability to query calculated results (and indeed input data). See https://github.com/AIFDR/riab/issues/38