search

cmip6dr / CMIP6_DataRequest_VariableDefinitions

Definitions of variables in the CMIP6 Data Request
7 stars 0 forks source link

possibly incorrect cell_methods and time coordinate #386

Closed taylor13 closed 5 years ago

taylor13 commented 5 years ago

@martinjuckes I noticed that in the CMOR table 6hrLev the only variable with frequency = "6hr" and time coordinate = "time" is bs550aer. All other variables in that table have frequency = "6hrPt" and time coordinate = "time1".

I suspect that different variables reported on model levels need to have the same time coordinate (and frequency attribute) if CMOR is to successfully write them. Could you check with those requesting this variable and see if we can request bs550aer at frequency "6hrPt". (I would think they would want all variables in this table be sampled in the same way.) Otherwise we might have to put bs550aer in its own separate table: "6hrLevPt".

martinjuckes commented 5 years ago

OK .. I'll fix that (change bs550aer to 6hrPt and time-point).

sylvaine commented 5 years ago

It seems that the associated structure of variable bs550aer (which has uid 8fecd6da-267c-11e7-8933-ac72891c3257) still has a wrong title attribute:

In [1]: dq = dreq.loadDreq()

In [2]: bs550aer = dq.inx.uid['8fecd6da-267c-11e7-8933-ac72891c3257']

In [2]: dq.inx.uid[bs550aer.stid].title
Out[2]: 'Temporal mean, Global field (single level) [XY-na] {:lambda550nm scatter180} [am-tm]'

while its tmid is correct:

In [3]: dq.inx.uid[dq.inx.uid[bs550aer.stid].tmid].title
Out[3]: 'Instantaneous value (i.e. synoptic or time-step value)'