Скачать или смотреть Creating synthetic observations using perfect_model_obs - Data Assimilation Research Testbed (DART)

DART requires an observation sequence file that specifies the type, time, location, value and error variance of observations in order to implement its data assimilation algorithms. The data in these observation sequence files often comes from real world observations, however, it’s also possible to create synthetic observations from model restart files. Three DART programs, create_obs_sequence, create_fixed_network_seq and perfect model obs can be used to create an observation sequence file consisting of synthetic observations. The first program, create_obs_sequence, is an interactive program requiring input from the user as it is run. create_obs_sequence generates an observation_sequence file that contains user specified data for the value and error variance of a hypothetical observation or it can be used to generate a set definition, in other words, an empty template typically named set_def.out that is used as input for the second program in this workflow: create_fixed_network_seq. This program takes set_def.out as input and creates an empty observation sequence file consisting of a set of repeated observations either repeated at a regular interval or at specified irregular times. This file is typically named obs_seq.in to signify that it is used as input for the third program, perfect_model_obs. The perfect_model_obs program creates an obs_seq.out file with observations corresponding to the times and locations specified in obs_seq.in and fills in the empty values by numerically integrating a model and using DART’s forward operators to create synthetic observations as the model integrates forward in time. Note that in these diagrams the obs_seq.out file is filled with observations that aren’t on the purple orange color scale that DATt’s schematics use for real world observations, instead it uses yellow blue color scale to signify that these are synthetic observations gathered from the model state.

This material is based upon work supported by the NSF National Center for Atmospheric Research, a major facility sponsored by the U.S. National Science Foundation and managed by the University Corporation for Atmospheric Research. Any opinions, findings and conclusions or recommendations expressed in this material do not necessarily reflect the views of the U.S. National Science Foundation.