Magill, C.R.; Horspool, N.A.; Pastor-Paz, J., Moratalla, J.M., Mooney, P. 2023 Inter-operability of RiskScape™ and the Oasis loss modelling platform. Lower Hutt, NZ: GNS Science. GNS Science report 2023/21. 31 p.; doi: 10.21420/FDTF-WT42
Abstract
The Institute of Geological & Nuclear Sciences (GNS Science), the National Institute of Water & Atmospheric Research (NIWA) and Toka Tū Ake EQC see the benefits of risk modelling and risk science being accessible across a larger number of platforms. Oasis is used widely within the international insurance and re-insurance industries. If models developed for RiskScape™ were able to be deployed in Oasis, this may increase uptake by these sectors. The key difference between the two software engines is that RiskScape™ has the abilityto combine hazard and exposure based on geospatial joins, whereas Oasis (standard implementation) utilises tabular data to combine hazard, exposure and vulnerability. Geospatial functionality may be included in an Oasis wrapper implementation by linking exposure files to GIS boundaries. RiskScape™ represents vulnerability as functions, whereas Oasis represents vulnerability as tabular data linked to hazard and exposure using keys. In both standard and wrapper Oasis implementation, RiskScape™ models will need to be modified to be inter-operable. For simple single-hazard models, standard implementation may be applicable; however, additional development would be required to convert hazard and vulnerability information to linked tables. If this process was to be done repeatedly, for example, because of model upgrades, it is recommended that this process be considered at the time of model development with conversion processes automated. For more complex models – multi-hazard, time-varying and high-spatial resolution – Oasis wrapper implementation may be more advantageous. This would also allow some of the geospatial and multi-hazard functionality of RiskScape™ to be carried across, as well as for consistency to be retained between model outputs. With wrapper implementation, less additional development is anticipated, although models will need to be compatible with Open Data Standard exposure inputs and have geospatial information provided numerically rather than spatially. A priority development step for RiskScape™ may be to accept Open Exposure Data files, with an option to output loss results as Open Results Data. We agree with Oasis that a proof-of-concept exercise to test inter-operability would be advantageous (auths)
