|
APOLLO (acronym for
Automatic Procedure to mOdeL voLcanic ash dispersiOn)
is a platform-independent procedure designed to facilitate the
execution and subsequent interpretation of volcanic ash
transport and fallout models. The APOLLO procedure is built on
a series of open-source programs that perform different tasks,
generate input data needed by models and do simple
postprocessing. Several fallout models can be
included. However, the user is not constrained to use just the
"default" models but can, alternatively, add other model(s)
with minor modifications on the source codes. To this purpose,
APOLLO contains a library (named LibApollo) that acts
as an interface between programs/models and input data files.
Data from files and databases generated by different programs
included in the APOLLO procedure can be read directly through
simple LibApollo routine calls without having a
detailed knowledge of the file/database format. The APOLLO
procedure generates all the data needed by models, including a
terrain and a meteorological database, the definition of the
source term, and the granulometric distribution. A
meteorological database for a particular area contains
short-term predictions, typically up to few days, for
meteorological variables (e.g. wind field, temperature,
turbulence related variables, etc.) defined at the nodes of a
3D structured grid. The meteorological database(s) is(are)
absolutely independent from models and can be updated
automatically, typically every day as new meteorological
prognostics are available. A run can start automatically after
the construction of a meteorological database or at any user
defined time (a run is mainly a scenario; it may content
several simulations from different fallout models). Whenever a
fallout model runs it simply reads the required meteorological
(and, if necessary terrain) data from the database as well as
the files that define the source and the granulometric data.
Clearly, the kind of data to read varies from model to model
(a model is not constrained to use the entire contents of the
database). For example, if a model assumes that the wind field
is horizontally uniform it is sufficient to use a selected
value from each vertical layer of the database, for instance
the average or a manually specified profile. The gathering of
data from a database is, consequently, a model dependent step
and must be implemented ad hoc for each particular
model. After a run, the last step of the procedure is to
postprocess the outcomes of models in order to draw maps with
pre-defined physical quantities. All models can share the
same postprocess treatment, so that if two or more different
models output the same quantity (e.g. ground deposit
thickness) their respective maps are directly comparable.
|
