Currently, Table C2 change factors are computed on the fly by extracting values from the change factor rasters at the station locations. Charles has now precomputed these values (with corrections for locations with no raster value, e.g., at some coastal locations).
Current situation
There is one file per DV. These files are stored at /storage/data/projects/comp_support/dv-explorer/data/new_tables. Each of these files contains columns labelled as follows (in LTR order):
Location
Location (NBCC): for verifying table merge
Comparison: for verifying table merge
Prov
2020 Elev (m)
lon
lat
rlat
rlon
irlat
irlon
PCIC
NBCC 2015
CF values are computed from rasters using lat and lon to determine raster cell corresponding to each station. (This is where no-data errors can occur; Charles fixed this in his precomputations.)
These CSV files (old and new) affect only the content of the Table C-2 tab. Stations and values for display on Map tab are determined by a different set of files. This is probably not a good arrangement, but it makes the update simpler so we will not change it in this issue.
Updated files
There is one file per DV. A file contains columns labelled as follows (in LTR order):
Location: station name
prov
Longitude
Latitude
elevation (m)
<DV> (NBCC): NBCC 2015 historical value
<DV> (<units>): presumed PCIC revised historical value (confirm)
CF_<gw>C<units?>
where
<units> specifies the absolute units of the value, e.g., degC-day
<units?> is a string containing the absolute units of the change factor, (<units>), or the empty string if the change factor is specified in relative units
<gw> specifies the global warming measure: 0.5, 1.0, ..., 3.5.
Conversion
Straightforward, except for the column names. Options:
Rename the columns, replacing <DV> (NBCC) with NBCC, <DV> (<units>) with PCIC. This removes metadata (DV, units) from file; although it what we have now, it's not great to have files documented only by their file names.
Compute column names from config values (DV, units, gw); check that they exist in the file.
Summary
Currently, Table C2 change factors are computed on the fly by extracting values from the change factor rasters at the station locations. Charles has now precomputed these values (with corrections for locations with no raster value, e.g., at some coastal locations).
Current situation
There is one file per DV. These files are stored at
/storage/data/projects/comp_support/dv-explorer/data/new_tables. Each of these files contains columns labelled as follows (in LTR order):LocationLocation (NBCC): for verifying table mergeComparison: for verifying table mergeProv2020 Elev (m)lonlatrlatrlonirlatirlonPCICNBCC 2015CF values are computed from rasters using
latandlonto determine raster cell corresponding to each station. (This is where no-data errors can occur; Charles fixed this in his precomputations.)These CSV files (old and new) affect only the content of the Table C-2 tab. Stations and values for display on Map tab are determined by a different set of files. This is probably not a good arrangement, but it makes the update simpler so we will not change it in this issue.
Updated files
There is one file per DV. A file contains columns labelled as follows (in LTR order):
Location: station nameprovLongitudeLatitudeelevation (m)<DV> (NBCC): NBCC 2015 historical value<DV> (<units>): presumed PCIC revised historical value (confirm)CF_<gw>C<units?>where
<units>specifies the absolute units of the value, e.g.,degC-day<units?>is a string containing the absolute units of the change factor,(<units>), or the empty string if the change factor is specified in relative units<gw>specifies the global warming measure: 0.5, 1.0, ..., 3.5.Conversion
Straightforward, except for the column names. Options:
<DV> (NBCC)withNBCC,<DV> (<units>)withPCIC. This removes metadata (DV, units) from file; although it what we have now, it's not great to have files documented only by their file names.