Operational Oil Spill Modelling: From Science to Engineering Applications in the Presence of Uncertainty - Mathematical Modelling and Numerical Simulation of Oil Pollution Problems - page 112

Fig. 5.6 HyposPy visualization of forecast tracks, continued from Fig. 5.5 . Forecast tracks in the
second day (0000-1200) are moving consistently closer to the beach. Visualization using Google
Earth
HyosPy development was motivated by a need for integrative tools that are exten-
sible and flexible so that new models and new data sources can be readily imple-
mented. HyosPy is formulated with two module levels: high-level modules controls
the overall logic and the Application Program Interface (API), whereas lower-level
modules process specific tasks, e.g. converting data from particular hydrodynamics
model to a common NetCDF format for oil spill models. Adding new models or data
is straightforward as the input/output for each module is designed without down-
ward/upward restrictions. Using the Google Maps/Earth visualization tools provide
portability, with results displayable over the web on any Java-enabled browser for
all supported terminal platforms (e.g., laptop, tablet, and smart phone) and operating
systems (Windows, Mac OSX, Linux, and Android) [ 18 ].
5.8 Evaluation of Uncertainty
Evaluating uncertainty with hindcast models and field-deployed drifter experi-
ments [ 35 ] can provide insight into setting up both the “best” parameters and upper
and lower bounds for a multi-model operational system (e.g. as in Sect. 5.6 ). For an
oil spill model, a typical IC uncertainty is in the oil spill shape. Typical BC prob-
lems include the appropriate effect of wind drag on the oil and effective diffusivity
of the oil. Quantification methodologies for these issues are discussed and demon-
strated below. These methodologies were designed to provide rapid multiple oil spill