Fix 3 - Githubissues

Ouranosinc / PAVICS-landing

The landing page serving as the entrance to PAVICS

1 stars 0 forks source link

Fix 3 #9

Closed tlogan2000 closed 3 years ago

tlogan2000 commented 3 years ago

@huard Could you take a quick look at notebooks (data analysis / subsetting) so far to get any feedback?

tlogan2000 commented 3 years ago

@aulemahal comments welcome as well... overall goal is to provide a series of notebooks explaining analysis steps see #3.

First two are fairly general but next series (index calc, ensembles, visualization) will more or less outline steps to do a 'Portraits Clim' type calculation.

huard commented 3 years ago

I suggest we use siphon instead of threddsclient. Siphon is supported by unidata.
Should we use "we" or a more neutral form ?

huard commented 3 years ago

Note that there is no driving scenario attribute in the metadata, ie rcp45 can only be parsed from the file name. I suggest we add that info in the NcML layer.

tlogan2000 commented 3 years ago

Note that there is no driving scenario variable in the metadata.

ok. can add to the ncmls

tlogan2000 commented 3 years ago

Note that there is no driving scenario variable in the metadata.

ok. can add to the ncmls

Note I actually would like to get the spatially extended dataset of cb-oura-1.0 on the server before launch... I will need to transfer the files to thredds. Covers alberta-east and portions of northern USA. Covers most (all?) of man. hydro's basins

huard commented 3 years ago

I made suggestions for the first notebook. Will continue with others tomorrow. Let me know if I'm going in the right direction with this before I continue. I left all your compute cells at the end in case you don't like what I did.

tlogan2000 commented 3 years ago

I made suggestions for the first notebook. Will continue with others tomorrow. Let me know if I'm going in the right direction with this before I continue. I left all your compute cells at the end in case you don't like what I did.

ok thanks. I was doing some changes to siphon as well... merge might have ended up 'weird' ... I will work on notebook 2 a bit so plz wait until tomorrow

tlogan2000 commented 3 years ago

Note that there is no driving scenario attribute in the metadata, ie rcp45 can only be parsed from the file name. I suggest we add that info in the NcML layer.

PCIC data uses a combined driving_experiment_id: historical,rcp85 is this a reasonable approach for the bias-corrected sims that span historical and future periods?