Working with observed data in GOTM

In the context of GOTM, the term `observations' should be understood in a broad sense: it may refer to data either measured in nature or generated artificially. The inclusion of such data into GOTM can serve different purposes. Examples are time-series of external pressure-gradients, which can be used to drive the model, or observed profiles of the temperature to which model results can be relaxed.

Two different types of `observations' are considered so far in GOTM: time series of scalar data and time series of profile data. The first type is used to introduce, for example, sea surface elevations into the model. The latter is used to include, for example, temperature or velocity fields.

All specifications concerning the `observations' are done via the namelist file obs.inp. Each of type of variable has its own namelist in obs.inp. Common for all namelists is a member with the suffix _method, used to specify the action performed to generate or acquire the variable, respectively. Observations can be, for example, read-in from files or computed according to an analytical expression. Some types of observations (e.g. turbulent dissipation rates) are not used directly during the calculations in GOTM. but can be conveniently interpolated to the numerical grid to allow for an easy comparison of measured data and model results.

For all types of observations, one _method is always `from file'. All input-files are in ASCII with a very straight-forward format. The necessary interpolation in space is performed as an integral part of the general reading routines. Temporal interpolation is performed as part of the specific reading routines, e.g. get_s_profile.F90.

Karsten Bolding 2012-01-24