Comparison between a forward simulation with CHIMERE and various observations¶
Here we run a forward simulation of CHIMERE including the comparison to observation data.
Please note here some instructions about how pyCIF deals with paths.
So far, there is only one possibility for
obsvect
. For the first simulation, there is no previous file to be read but the file generated by this simulation is dumped to be used in later simulations.##################################################################### ##################################################################### # Observation vector and observation uncertainties if needed # Also projects information from the observation to the model space obsvect: plugin: name: standard version: std dump: TrueIn case all the
monitor.nc
files to use for comparison are ready and arranged in a directory into sub-folders by type (e.g. “OMI”, “IASI”) then sub-folders by species (e.g. “CH4”, “N2O”), then it is possible to simply indicate here the root directory.##################################################################### ##################################################################### # Observation vector and observation uncertainties if needed # Also projects information from the observation to the model space obsvect: plugin: name: standard version: std dump: True dir_obsvect: directory_with_perfectly_ordained_and_formatted_monitor_netcdf_files/If the branching is empty or not right, or the files are not named “monitor.nc”, then the information will not be used. Therefore, in most cases, information for measurements is provided in
datavect
, in the dedicated components (see j V).The requirements for the standard obsvect are
model
(e) anddatavect
(j)
If the simulation is to be compared to observations, a component dedicated to observations must be specified. The CIF can use various types of observations, either in pre-formatted files or in raw data files.
For satellite data, the plugin
satellites
has been designed for conveniency and includes various transformations. It is then easy to specify the monitor formatted files to use with the minimum information on how to transform the model’s output into equivalents of the data, e.g. total or partial columns. The names of theparameters
must be the same as the names in the output of the model i.e. the names of the species in OUPUT_SPECIES. The format of the file name provided makes it possible to divide a large amount of data into smaller monitor.nc files, e.g. per month or even per smaller chunks for satellite data.satellites: parameters: S2: dir: directory_where_monitor_files_for_S2_are located file: monitor_file_only_for_a_period_of_7day_%Y%m%d.nc formula: 3 pressure: Pa product: level chosenlev: 5 cropstrato: False correct_pthick: TrueWARNING errors may happen when no observation is available during the simulation XXCHECK AND CORRECT?XX
Other sets of transformations can also be specified by the user in
obsoperators
with atransform_pipe
XXX more tutos later on this, useful for transformation which are not ready-made in the codeXXXXXX
XXXXXXX faire tuto ajout nouvelle lecture d’obs?XXXX