FORECASTING DRIFT
519
tional forecast model data are required, and uncertainty assessment is
an essential part of the forecast information returned.
4.1
Geophysical forcing data access
An important task for the operational implementation is accessing the
best possible forcing data at a given time and location. This can be a
complicated task. In a SAR case, for example, the LKP may be many
hours, even days old. At met.no, forcing data sets covering the last
7 days are maintained for rapid retrieval to meet such an eventuality.
Furthermore, there may be several candidate forecast models, with different horizontal extent and resolution, capable of supplying the same
type of forcing data (Figure 4). The choice of model data set to use for a
drift forecast will in principle depend on the location and the presumed
forecast accuracy of the data. However, in practical implementation, the
choice is limited to models that are considered “officially operational”
in the sense of established quality and robustness (e.g., supported by
automatic backup systems, computer redundancy, archiving, etc.). In
a typical national service, there will be a small number (1-2) of operational models for weather, ocean and wave forecasting, together with
several pre-operational models being tested in the daily routine with the
aim of replacing or supplementing the existing operational models. At
met.no, the drift services currently obtain their atmosphere and wave
forcing data from a selection of operational models, including met.no,
ECMWF and UK Met Office, while ocean data are obtained from one
operational model at met.no. The default is the met.no operational
models (cf. Figure 4). In the event of total failure to obtain model
forecast data, an operator may enter uniform values of wind, wave and
current manually.
Recent developments in global ocean modeling and, not least, data
exchange capability (e.g., the European Mersea project) are making it
feasible to access adequate ocean forcing data from other operational
forecasting centers. Thus, there is potentially a wide range of alternative data sets available. The met.no drift forecast service is being
extended to allow selective access to a fuller range of forcing data sets,
from local, high resolution in-house models to global data sets obtained
from external sources.
The challenge of this approach is devising methods to determine which
sets are best for a given emergency situation. For
ex
archived (e.g., for post mortem reruns), as well as make the necessary
agreements on formats, data product requirements and delivery
data
ternal data sets, one must ensure that they are reliably available and
forcing
519
tional forecast model data are required, and uncertainty assessment is
an essential part of the forecast information returned.
4.1
Geophysical forcing data access
An important task for the operational implementation is accessing the
best possible forcing data at a given time and location. This can be a
complicated task. In a SAR case, for example, the LKP may be many
hours, even days old. At met.no, forcing data sets covering the last
7 days are maintained for rapid retrieval to meet such an eventuality.
Furthermore, there may be several candidate forecast models, with different horizontal extent and resolution, capable of supplying the same
type of forcing data (Figure 4). The choice of model data set to use for a
drift forecast will in principle depend on the location and the presumed
forecast accuracy of the data. However, in practical implementation, the
choice is limited to models that are considered “officially operational”
in the sense of established quality and robustness (e.g., supported by
automatic backup systems, computer redundancy, archiving, etc.). In
a typical national service, there will be a small number (1-2) of operational models for weather, ocean and wave forecasting, together with
several pre-operational models being tested in the daily routine with the
aim of replacing or supplementing the existing operational models. At
met.no, the drift services currently obtain their atmosphere and wave
forcing data from a selection of operational models, including met.no,
ECMWF and UK Met Office, while ocean data are obtained from one
operational model at met.no. The default is the met.no operational
models (cf. Figure 4). In the event of total failure to obtain model
forecast data, an operator may enter uniform values of wind, wave and
current manually.
Recent developments in global ocean modeling and, not least, data
exchange capability (e.g., the European Mersea project) are making it
feasible to access adequate ocean forcing data from other operational
forecasting centers. Thus, there is potentially a wide range of alternative data sets available. The met.no drift forecast service is being
extended to allow selective access to a fuller range of forcing data sets,
from local, high resolution in-house models to global data sets obtained
from external sources.
The challenge of this approach is devising methods to determine which
sets are best for a given emergency situation. For
ex
archived (e.g., for post mortem reruns), as well as make the necessary
agreements on formats, data product requirements and delivery
data
ternal data sets, one must ensure that they are reliably available and
forcing