Adding your observations to DART

First, you should understand that DART already supports a tremendous variety of observations. To fully support an observation means that the observation can be converted from its native format to the DART observation sequence format and that the observation forward operator is already implemented. Keep in mind that forward operators are not specific to any one model.

DART Observations and Available observation converter programs describe types of observations; their sources and methods for formatting them for use by DART. DART can use both real and “synthetic” observations. Synthetic observations are used in OSSEs and are useful for model interface development.

Real observations

The observation converters are in the observations/obs_converter directory, and are documented in Available observation converter programs.

The forward operators are functionally or logically grouped into Fortran modules in the observations/forward_operator directory. DART employs a ‘contractual’ style of programming in that the forward operator requests information from the model, and if the model cannot provide it, the forward operator may request different information in an attempt to collect the information needed to apply the operator. If the model cannot provide any of the required information, the forward operator fails, the DART QC for that observation is set to the appropriate value, and the program continues.

Synthetic Observations

Synthetic observations can be created manually using program create_obs_sequence or extracted from a numerical model using program perfect_model_obs.

When using perfect_model_obs the value of the synthetic observation is calculated from the model state at a chosen location using the appropriate forward operator H(x). The observation error variance is specified for later use and an observation error is added to the value; typically a random draw from a distribution sized according to the error variance.

When the observation is used to sample a model state directly, this is known as an identity observation. The observation error variance and observation error are specified in the same way as other synthetic observations, but the observation value is a copy of a chosen state variable. That is, the forward operator H(x) is the identity matrix (hence the name). The value of the observation is taken from the model state at a chosen grid point, which is the location of the observation.

Identity observations do not get listed as a type in the header of an observation sequence file, they are denoted in a given observation by a type of -x (negative x) where x is the index in the DART state vector that the observation corresponds to. This means that identity observations from one model can’t (usually) be used in an assimilation with another model.

The process for creating synthetic observations, and identity observations is outlined in Creating an obs_seq file of synthetic observations and Synthetic observations.