Finding datasets

Revision as of 00:04, 30 April 2019 by P.petrelli (talk | contribs)

Finding climate data at NCI

There are a lot of climate data resources which are available at NCI. Only some of them have been listed in the NCI data catalogue and even then there are a few tips that can help you working out if the dataset you are looking for is already available or not.

 

Step 1 - check the NCI data catalogue

NCI uses a geonetwork catalogue to list its collections, this covers most of the bigger data collections hosted by NCI. To find a dataset you can run a text search or select from a list of available attributes.

Free text search

Geonetwork free text search will look for an exact match in the record title and description. For example if I am looking for "precipitation" datasets I need to type the entire word, if I type only "precip" it will return only this dataset:

"Proportion of days per month with precip > 0.2mm: ..."

Likewise typing "precipitation" will return 62 datasets including any record that has the exact word in their title or description.

Selection by topic and other attributes

NCI created some categories dividing the datasets based on "topic", they are shown on the catalogue main page.  Once you click on one of them you will see a panel on the left side of the screen showing all the available attributes you can use to select datasets. The same is true if you run a text search, you could then refine your selection using this panel. Probably the most useful attributes are the keywords, they are based on keywords set by the data manager themselves or on Field Of Research (FOR) standards.

Unfortunately the selection panel can be a bit obscure, for example often the FOR codes appears as a sequence of numbers rather then their definition for example "0401" rather then "atmospheric science".  This is because the dataset descriptions and attributes are provided by the owners of the data, so there is a big variance in what is included and how keywords are chosen or interpreted.

Another limitation is that sometimes there is a record for a data collection but not for the single datasets included in it. Some of the data collections are clear as CMIP6 contains only CMIP6 data but atmospheric re-analysis or even the ARCCSS own data collection are more heterogenous and if a child record is not present for each dataset in the collection, it can be hard to get an idea of what is actually available.

NCI is working on an improved and more user-friendly interface and is also working with all the data managers to improve the quality of the records. Feel free to send them feedback or tell us if you prefer and we will pass it on.

Step 2 - check the CMS wiki

The CMS wiki lists all the datasets we download and manage for our researchers. Some of these are also listed in geonetwork, mostly are not because for example we have downloaded only a small subset.

 

Step 3 - ask the helpdesk:  cws_help@nci.org.au 

 

If you still can’t locate the datasets you were looking for or you find the dataset but the description wasn’t sufficient to determine that it covers your needs then feel free to e-mail us on the helpdesk.

As both geonetwork and the wiki are potentially incomplete or some of the records might be out of date, it is always a good idea to double check with us. We also might know about other data which is on raijin but not necessarily listed or make enquire on your behalf to our partners to help you locate a specific product.

 

Step4 - request us to download the data

If you still couldn’t find what you were looking for we can help you downloading the data. When we receive a request to download data, we quickly check the storage and time required for the task. Unless these are “enormous” we usually download the data for you and put in a shared environment where others can also access it. If the dataset require a lot of storage or time to download or ongoing maintenance we might need to check with the infrastructure committee before going ahead.

It is rare that we have to say no to someone and we don’t do it without a fairly strong reason because we prefer download and manage updates to the data and to have it shared in a central location where other can access it too. We also want to avoid that precious disk space that should be used for analysis by all your group gets bogged with data.

 

Datasets hosted on raijin and managed by the ARCCSS

 

Other datasets hosted on raijin

  • CMIP5- Coupled Model Intercomparison Project Phase 5 data on raijin
  • CMIP6- Coupled Model Intercomparison Project Phase 6 data on raijin

 

ARCCSS datasets and software published on Research Data Australia (RDA)

The Centre of Excellence for Climate System Science has started publishing its datasets on Research Data Australia (RDA), the Australian National Data Service (ANDS) metadata repository. The first datasets to be published were from the Climate Model Downscaling Data for Impacts Research (CliMDDIR), then the ACCESS CMIP5 simulations.

  • ACCESS- CMIP5 simulations
  • CliMDDIR- Climate Model Downscaling Data for Impacts Research
  • ARCCSS collection- ARCCSS datasets on the NCI Data Catalogue
  • C20C+ ACCESS- Atmospheric ACCESS1.3 historical all forcing model output for the Climate of the 20th Century Plus (C20C+) Detection and Attribution sub-project
  • MarineHeatWaves- Marine heatwaves detection code

 

External data resources