[Esip-preserve] data reference vs. citation

Thu Mar 31 11:41:06 EDT 2016

Matt,

Thanks for forwarding and providing help for those of us who could not go past the pay wall. This is certainly encouraging, even though it is not quite what we are pushing for in terms of data citations. Certainly helps with reproducibility and trust.

Rama.

From: Esip-preserve [mailto:esip-preserve-bounces at lists.esipfed.org] On Behalf Of Matthew Mayernik via Esip-preserve
Sent: Wednesday, March 30, 2016 3:38 PM
To: esip-preserve at lists.esipfed.org
Subject: Re: [Esip-preserve] data reference vs. citation

Apologies to folks blocked by Nature's paywall. Here's are the two relevant sections of which I was referring. All of the numbered references in the "Data" section are to journal articles.

Data.

Summer temperature observations are daily maxima from the Global Historical Climatology Network-Daily database (GHCND; ref. 14 <http://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo2687.html#ref14> ). We rely on weather station observations of near-surface daily maximum air temperature14 <http://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo2687.html#ref14> , as opposed to reanalyses that infer near-surface conditions34 <http://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo2687.html#ref34>  without necessarily ingesting station data35 <http://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo2687.html#ref35> .

The analysis is confined to the 60 warmest days of summer based on the average climatology across US weather stations (24 June–22 August on a non-leap year, and 23 June–21 August on a leap year). These climatologically warmest days of summer are when further warming may be expected to have the greatest implications for health and crops. Records from individual weather stations are included only if they have at least 80% coverage during June, July and August for at least 80% of the years considered in the analysis. There are 1613 stations in the eastern US that fulfil these requirements for the 1982–2015 period, and 1092 for the longer 1950–2015 period. Changing the data coverage requirement to focus specifically on the 60 hottest days of summer has essentially no effect on the results, because >99.5% of the stations that pass the coverage requirement for June, July and August also pass the requirement for peak summer.

Daily SST data are from the NOAA OI SST2 data set36 <http://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo2687.html#ref36> , which is available at daily resolution beginning in September 1981. SST data are re-gridded from 1/4° to 1° spatial resolution. Daily geopotential height and wind fields beginning in 1979 are from the NCEP-DOE II reanalysis37 <http://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo2687.html#ref37>  at 2.5° resolution. Atmosphere–ocean fluxes from 1985 to 2009 are from the OAFlux project38 <http://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo2687.html#ref38>  at 1° resolution. Monthly SST data are from the 5° resolution HadSST3 data set21 <http://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo2687.html#ref21> , which uses only in-situ measurements from ships and buoys. All analyses are in terms of anomalies that are calculated by removing the first three annual harmonics, and then removing a linear trend.

Data and code availability.

All data used in the analysis are publicly available. NCEP-DOE II reanalysis fields for geopotential height and 10-m zonal and meridional winds are available at http://www.esrl.noaa.gov/psd/data/gridded/data.ncep.reanalysis2.html. Daily SST data from NOAA OI SST V2 are available at ftp://ftp.cdc.noaa.gov/Datasets/noaa.oisst.v2.highres. Monthly SST data from HadSST3, version 3.1.1.0 are available at http://www.metoffice.gov.uk/hadobs/hadsst3/data/download.html. Ocean–atmosphere heat flux data from the WHOI OAFlux Project are available at http://oaflux.whoi.edu/heatflux.html. Precipitation data from the CPC Unified Gauge-Based Analysis of Daily Precipitation are available at http://www.esrl.noaa.gov/psd/data/gridded/data.unified.daily.conus.html. Daily temperature data are from the Global Historical Climatology Network-Daily database, available from http://www1.ncdc.noaa.gov/pub/data/ghcn/daily.

Monthly values for the Pacific North America pattern, North Atlantic Oscillation, and El Niño–Southern Oscillation indices are from the Climate Prediction Center, and are available at http://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/norm.pna.monthly.b5001.current.ascii, http://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/norm.nao.monthly.b5001.current.ascii, and http://www.cpc.ncep.noaa.gov/data/indices/sstoi.indices, respectively. Monthly values for the Pacific Decadal Oscillation and Northern Annular Mode indices are from the Joint Institute for the Study of the Atmosphere and Ocean, and are available at http://jisao.washington.edu/pdo/PDO.latest and http://research.jisao.washington.edu/analyses0302/#data, respectively.

See https://github.com/karenamckinnon/PEP.git for code and formatted data that allow for reproduction of the results depicted in the main text figures.

On Wed, Mar 30, 2016 at 1:25 PM, Matthew Mayernik <mayernik at ucar.edu> wrote:

Hi all,

This recent Nature article is an interesting case for data citation. Toward the end of the paper, all of the datasets are described in the methods section, in a subsection called "Data", with citations given to published papers about the datasets. Below that, there is a "Data and Code availability" section, which gives URLs to all of the data & software resources relevant to the paper. They don't give any DOIs for the data/software, though I don't know if any of these resources have DOIs.

This is both a) a great case of being transparent in linking to the underlying data & code, and b) a great case of how data citations should not be done (at least according to our usual recommendations of putting them in the reference list instead of the text or acknowledgements, using DOIs, etc.).

Long-lead predictions of eastern United States hot days from Pacific sea surface temperatures
K. A. McKinnon, A. Rhines, M. P. Tingley & P. Huybers    
Nature Geoscience, (2016)
http://doi.org/10.1038/ngeo2687 

It seems like a good exhibit in how these things are evolving in a positive direction, if on a winding road,

Matt

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.deltaforce.net/pipermail/esip-preserve/attachments/20160331/48ce237c/attachment-0001.html>