Regional historical changes for REF #3926
Conversation
…on_clouds_cycles.yml
…/ESMValTool into benchmarking_maps4monitoring
|
Could you please check if observational data is already available in obs4MIPs? #3932 (comment) For example, for ERA-5 obs4MIPs provides the following variables: 'psl', 'ta', 'tas', 'ua', 'uas', 'va', 'vas', 'zg' and GPCP-SG version 2.3 is available as "GPCP-V2.3". |
Do we have them on Levante @hb326 ? The only one I find in Tier1 is: |
Dear @bouweandela, I changed GPCP-SG to obs4MIPs now, but unfortunately, I don't manage to get the ERA5 obs4MIPs data: On Friday I first got a timeout, then it changed to this and is still the same now: |
|
Unfortunately, the only server providing that data is offline.. |
It is running again, I try to get them, but I won't manage today anymore (and it is only tas, psl and uas from the variables I use and not all files are there, pr and hus seem to be missing completely, for them I don't get a wget scripe) |
And now it is again not possible, to reach the server at all :( |
|
Dear @bouweandela I updated ERA(-)5 psl, tas, and ua to obs4MIPs now. The download was rather slow. pr and hus I did not find at all, therefore I did not change the recipes for them (the change would be the same as for the other variables if data would be available: instead of ) |
| <<: *var_default_abs | ||
| reference_dataset: GPCP-SG | ||
| additional_datasets: | ||
| - {dataset: GPCP-SG, project: obs4MIPs, type: atmos, version: 2.3, tier: 1, |
There was a problem hiding this comment.
| - {dataset: GPCP-SG, project: obs4MIPs, type: atmos, version: 2.3, tier: 1, | |
| - {dataset: GPCP-V2.3, project: obs4MIPs, tier: 1, |
The name of the obs4MIPs dataset on ESGF is GPCP-V2.3. type is not a required facet for obs4MIPs data. The only available version on ESGF is v20180519, but the version does not need to be specified because ESMValCore will automatically use the latest available version.
There was a problem hiding this comment.
The name is GPCP-SG if I download the data from ESGF (only the [THREDDS Catalog] worked for me (and on Levante))?
There was a problem hiding this comment.
You are right that GPCP-SG is used in the filename, but what matters is the value for the source_id, as that is what we call dataset in ESMValCore (see here for the mapping from 'our' facet names to ESGF names). Using source_id=GPCP-V2.3 in an ESGF search returns a file:
{
"responseHeader":{
"status":0,
"QTime":5,
"params":{
"df":"text",
"q.alt":"*:*",
"indent":"true",
"echoParams":"all",
"fl":"*,score",
"start":"0",
"fq":["type:File",
"project:\"obs4MIPs\"",
"source_id:\"GPCP-V2.3\""],
"sort":"id asc",
"rows":"500",
"q":"*:*",
"shards":"localhost/solr/files,localhost:8985/solr/files,localhost:8986/solr/files,localhost:8987/solr/files,localhost:8988/solr/files,localhost:8989/solr/files,localhost:8990/solr/files",
"tie":"0.01",
"facet.limit":"-1",
"qf":"text",
"facet.method":"enum",
"facet.mincount":"1",
"wt":"json",
"facet":"true",
"facet.sort":"lex"}},
"response":{"numFound":1,"start":0,"maxScore":1.0,"docs":[
{
"id":"obs4MIPs.NASA-GSFC.GPCP.atmos.mon.v20180519.pr_GPCP-SG_L3_v2.3_197901-201710.nc|dpesgf03.nccs.nasa.gov",
"version":"1",
"cf_standard_name":["precipitation_flux"],
"checksum":["4dd4678b79ef139446c8406da5aae4fed210abb2f2160ef95f6988bf83e4525b"],
"checksum_type":["SHA256"],
"data_node":"dpesgf03.nccs.nasa.gov",
"dataset_id":"obs4MIPs.NASA-GSFC.GPCP-V2.3.atmos.mon.v20180519|dpesgf03.nccs.nasa.gov",
"dataset_id_template_":["%(project)s.%(institute)s.%(source_id)s.%(realm)s.%(time_frequency)s"],
"index_node":"esgf-node.llnl.gov",
"instance_id":"obs4MIPs.NASA-GSFC.GPCP.atmos.mon.v20180519.pr_GPCP-SG_L3_v2.3_197901-201710.nc",
"institute":["NASA-GSFC"],
"latest":true,
"master_id":"obs4MIPs.NASA-GSFC.GPCP.atmos.mon.pr_GPCP-SG_L3_v2.3_197901-201710.nc",
"model":["Obs-GPCP"],
"product":["observations"],
"project":["obs4MIPs"],
"realm":["atmos"],
"replica":false,
"size":19348352,
"source_id":["GPCP-V2.3"],
"time_frequency":["mon"],
"timestamp":"2018-02-17T17:17:54Z",
"title":"pr_GPCP-SG_L3_v2.3_197901-201710.nc",
"tracking_id":["4070c751-6c2d-440f-a4d7-5b325fb98990"],
"type":"File",
"url":["https://dpesgf03.nccs.nasa.gov/thredds/fileServer/obs4MIPs/observations/NASA-GSFC/Obs-GPCP/GPCP/V2.3/atmos/pr/pr_GPCP-SG_L3_v2.3_197901-201710.nc|application/netcdf|HTTPServer"],
"variable":["pr"],
"variable_long_name":["Precipitation"],
"variable_units":["kg m-2 s-1"],
"_version_":1662343940370595840,
"retracted":false,
"_timestamp":"2020-03-27T18:45:20.965Z",
"score":1.0}]
},
"facet_counts":{
"facet_queries":{},
"facet_fields":{},
"facet_ranges":{},
"facet_intervals":{},
"facet_heatmaps":{}}}while a search using source_id=GPCP-SG in an ESGF search returns no files:
{
"responseHeader":{
"status":0,
"QTime":8,
"params":{
"df":"text",
"q.alt":"*:*",
"indent":"true",
"echoParams":"all",
"fl":"*,score",
"start":"0",
"fq":["type:File",
"project:\"obs4MIPs\"",
"source_id:\"GPCP-SG\""],
"sort":"id asc",
"rows":"500",
"q":"*:*",
"shards":"localhost/solr/files,localhost:8985/solr/files,localhost:8986/solr/files,localhost:8987/solr/files,localhost:8988/solr/files,localhost:8989/solr/files,localhost:8990/solr/files",
"tie":"0.01",
"facet.limit":"-1",
"qf":"text",
"facet.method":"enum",
"facet.mincount":"1",
"wt":"json",
"facet":"true",
"facet.sort":"lex"}},
"response":{"numFound":0,"start":0,"maxScore":0.0,"docs":[]
},
"facet_counts":{
"facet_queries":{},
"facet_fields":{},
"facet_ranges":{},
"facet_intervals":{},
"facet_heatmaps":{}}}There was a problem hiding this comment.
But when I write "dataset: GPCP-V2.3" in the recipe it does not run (maybe it would, if I'd allow download?) but it does with dataset: GPCP-SG (because on Levante, these data are available)?
There was a problem hiding this comment.
Could you try running it with --search-esgf=always?
There was a problem hiding this comment.
Thanks for the answer @bouweandela ! I tried again with search_esgf: when_missing in config, then it works with "GPCP-V2.3". But it does not work any more, when I set it to search_esgf: never to run the next recipe, which needs the same data (or even to rerun the same recipe again). Then it still does not find the data? The reason seem to be, that it does not create a folder. It puts the data under:
(esmvaltool) [b380216@levante1 ESMValTool]$ ls -lhrt /work/bd1083/b380216/extraobsv2/Tier1/
total 19M
drwxr-sr-x 2 b380216 bd0854 20K Jun 15 12:28 ERA-5
-rwx------ 1 b380216 bd0854 19M Jun 25 15:39 pr_GPCP-SG_L3_v2.3_197901-201710.nc
When I move the data by hand into a sub-folder called "GPCP-V2.3" it works, so I change it in these recipes now to GPCP-V2.3 now.
But I wonder if we need to change something for the downloader in cases where file name and the source_id disagree
and/or if we then should rename?
/work/bd0854/DATA/ESMValTool2/OBS/Tier1/GPCP-SG
to
/work/bd0854/DATA/ESMValTool2/OBS/Tier1/GPCP-V2.3
?
But this would cause other recipes, which currently use "GPCP-SG" not to work anymore on Levante?
These recipes are:
- esmvaltool/recipes/recipe_model_benchmarking_timeseries_pr_e.yml
- esmvaltool/recipes/examples/recipe_check_obs.yml
- esmvaltool/recipes/ipccwg1ar5ch9/recipe_flato13ipcc_figure_96.yml
- esmvaltool/recipes/ipccwg1ar6ch3/recipe_ipccwg1ar6ch3_atmosphere.yml
- esmvaltool/recipes/model_evaluation/recipe_model_benchmarking_boxplots.yml
- esmvaltool/recipes/model_evaluation/recipe_model_benchmarking_maps.yml
- esmvaltool/recipes/model_evaluation/recipe_model_evaluation_basics.yml
- esmvaltool/recipes/model_evaluation/recipe_model_evaluation_clouds_clim.yml
- esmvaltool/recipes/model_evaluation/recipe_model_evaluation_precip_zonal.yml
-esmvaltool/recipes/ref/recipe_ref_scatterplot.yml
There is also still the issue, that I cannot find pr and hus as cmorized ERA5 data?
There was a problem hiding this comment.
But it does not work any more, when I set it to
search_esgf: neverto run the next recipe, which needs the same data (or even to rerun the same recipe again).
You need to set the rootpath and DRS so it matches the directory tree in your config-user.yml file. For example, mine looks like this for the obs4MIPs project on Levante:
rootpath:
obs4MIPs:
/work/bd0854/DATA/ESMValTool2/OBS: default
/work/bd0854/DATA/ESMValTool2/download: ESGFI wonder if we need to change something for the downloader
The downloader uses DRS ESGF, for obs4MIPs that is https://github.com/ESMValGroup/ESMValCore/blob/4d8fcbfb331e0745db58321cc41fe7de88c82393/esmvalcore/config-developer.yml#L110-L119
to store files. As you can see, the only requirement on the filename is that it starts with the short_name for that DRS setting, so what comes after it should't matter.
There was a problem hiding this comment.
There is no issue with the file name, but with the name of the directory, which is usually the data set name:
It downloads the data to:
Tier{tier}/
but expects: 'Tier{tier}/{dataset}' for finding the data, if they are not downloaded in the same run.
I'd expect that most users don't change config_developer (because it is part of ESMValCore) so if I use dataset: GPCP-V2.3 it expects this as directory name and won't find the data if they are directly under Tier{tier}/ or in /work/bd0854/DATA/ESMValTool2/OBS/Tier1/GPCP-SG ?
Co-authored-by: Bouwe Andela <[email protected]>
Co-authored-by: Bouwe Andela <[email protected]>
Co-authored-by: Bouwe Andela <[email protected]>
Co-authored-by: Bouwe Andela <[email protected]>
Co-authored-by: Bouwe Andela <[email protected]>
|
Thanks everyone! I noticed that the data is not written out for |
|
@bouweandela thanks for merging! There seem to be generally no data written with the plots for seaborn_diag.py (I tried the standard recipe recipe_seaborn.yml). Is writing data generally expected for all recipes in ESMValTool or is this only/mainly an issue for the REF? I probably should open a new issue for this then? |
Yes, see our contribution guidelines. For the REF it is even more important because most users of REF output will not be capable of running the recipe themselves. |
Description
Time series, trends, and evaluation plots for regional (CORDEX) Historical changes in climate variables for REF
Based on existing monitoring and benchmarking recipes:
and
from Benchmarking recipes (Lauer et al.)
ref/recipe_model_benchmarking_boxplots_region_trend.yml needs some code changes in monitor/multi_datasets.py all others only add settings to the recipes.
AR6 regions (from shape files) used.
Before you get started
Checklist
It is the responsibility of the author to make sure the pull request is ready to review. The icons indicate whether the item will be subject to the 🛠 Technical or 🧪 Scientific review.
New or updated recipe/diagnostic
To help with the number of pull requests: