Central Data Archive for STARDEX (D14 and D17)

2004-10-27, Version 0.4.1


The central data archive provides easy access to observations, reanalysis, model and downscaled data for STARDEX partners and end users. It includes seasonal time series of the STARDEX indices and daily data for precipitation, minimum and maximum temperature, and circulation indices. The data will be stored in NetCDF files conforming to the Climate and Forecast (CF) metadata conventions and made available through a OPeNAP/DODS server.

This document describes the organization of the data archive and the conventions used to store the data in the NetCDF files.

NetCDF homepage: http://my.unidata.ucar.edu/content/software/netcdf/
CF conventions: http://www.cgd.ucar.edu/cms/eaton/cf-metadata/

Directory Hierarchy


<data-type> data or indices
<region> STARDEX region (alps, europe, germany, greece, iberia, italy, uk)

All directory names should consist of lowercase letters only.

File Naming Convention

A generic filename has the following form:

component description
<variable> variable name
<region> region and location of the stations
<type> type of data: station (st), gridded (gp), region (reg), time series (ts)
<institution> Specifies who produced the data.
<source> The method of production of the data (e.g. downscaling method).
<exp> Description of the experiment (e.g. driving GCM, scenario).

The file names should consist of lowercase letters only. Consult appendix A for further details and some example file names.

File Contents

In this section we define the metadata, the coordinate systems, and the variables for the different types of data to be stored. The types of data include: station observations, upscaled observations, downscaled data, model data on the one hand and preciptation, minimum and maximum temperature and the STARDEX indices on the other hand.


The metadata should contain all relevant information about the dataset.

name contents mandatory
Conventions CF-1.0 x
title A short description of what is in the dataset. x
source The method of production of the data (e.g. downscaling method). x
experiment Description of the experiment: Driving GCM, which scenario.
institution Specifies where the data was produced. x
comment Additional information about the data or methods used to produce it.
history Provides an audit trail for modifications to the original data.

The most important metadata related to downscaling are source and experiment. source should contain a concise description of the downscaling method, whereas experiment specifies the data source used for the predictors. Consult appendix B for further details.

Coordinate Systemes

The coordinate system should provide an accurate description of the location of the data in space and time. Three different coordinate systems are indroduced: one for station data, one for gridded data on a regular lat/lon grid, and one for gridded data on a rotated grid. The table below lists the variables required to describe this coordinate systems.

type of data coordinate variables auxillary coordinate variables grid mapping variable
station data station, time lon, lat, hgt, station_id, station_name -
lat/lon grid lon, lat, hgt, time - -
rotated grid rlon, rlat, hgt, time lon2d, lat2d rotated_pole

For ensembles an additional dimensions can be added to the above coordinate systems, e.g.

type of data coordinate variables auxillary coordinate variables grid mapping variable
station data station, time, member lon, lat, hgt, station_id, station_name -


Here we define the data variable names and their standard attributes. See appendix C for a complete variable list, including all the STARDEX indices.

variable long_name units
pre precipitation mm d-1
tmin minimum temperature degC
tmax maximum temperature degC

The long_name should be, despite its name, a short name for the variable which can be used for labeling plots.

File names


filename description
pre.al-fic.st.eth.obs.nc FIC observations at the 10 common alpine stations
pre.al-fic.st.eth.dloci.ha3p-a2a.nc downscaled FIC OBS from HADAM3P, Scenario A2a
pre.al-ch.st.eth.obs.nc Swiss precipitation stations (high-resolution dataset)
pre.al-alra.gp.eth.obs.nc Alpine Reanalysis Observations
pre.al-fic.gp.eth.alra-obs.nc Observed gridpoint precipitation at the FIC stations
pre.al-nit.reg.adgb.obs.nc Observed regional precipitation series for northern Italy
pre.gr.st.auth.uobs.nc Upscaled observed greek station data
pre.gr-west.reg.auth.obs.nc observed precipitation for the western Greek region
ct500.gr.ts.auth.ncep.nc 500hPa circulation types for Greece from NCEP
ctthick.gr.ts.auth.ncep.nc 1000-500hPa thickness circulation types for Greece from NCEP
pre.gr.st.auth.ct.ncep.nc DS precip for Greek stations using NCEP circulation types
pind.al-fic.st.eth.obs.nc precipitation indices for the 10 alpine FIC stations
pind.al-fic.st.eth.dloci.ha3p-a2a.nc precipitation indices for downscaled station data
tind.al-fic.st.uibe.obs.nc temperature indices for the 10 alpine FIC stations

Variable Name

abbreviation description
pre precipitation
tmin minimum temperature
tmax maximum temperature
pind precipitation indices
tind temperature indices
ct500 500hPa circulation types
ctthick 1000-500hPa thickness circulation types


abbreviation description
al Alps
eu Europe
ge Germany
gr Greece
ib Iberia
it Italy
uk United Kingdom
al-fic the common FIC stations for the Alps
al-ch the high-resolution Swiss station dataset
al-alra the Alpine Reanalysis
it-er Emilia-Romagna
gr-west western greek region
gr-east eastern greek region

Type of Data

abbreviation description
st station data
gp gridpoint data
reg regionally averaged data
ts time series


abbreviation description
adgb University of Bologna, Italy
arpa Servizio Meteorologico Regionale, ARPA-SMR Emilia-Romagna, Italy
auth University of Thessaloniki, Greece
cnrs Centre National de la Recherche Scientifique, France
dmi Danish Meteorological Institute, Denmark
eth Swiss Federal Institute of Technology, Switzerland
fic Fundación para la Investigación del Clima, Spain
fts Fachhochschule Stuttgart - Hochschule für Technik, Germany
kcl King's College London, UK
unibe University of Berne, Switzerland
ustutt Institut für Wasserbau, Germany
uea University of East Anglia, UK


abbreviation description institution
obs observations
ncep NCEP reanalysis
ha3p-cta HADAM3P, control, member a (see also appendix A.7)
anal2 two-step analogue method FIC
ann artificial neural network AUTH
ann-gamma ANN using hybrid Bernoulli/Gamma data misfit term UEA
ann-gammamc as ann-gamma but with Monte Carlo simulation UEA
ann-irbf individual radial basis fn ANN KCL
ann-mlp multi layer perceptron ANN KCL
ann-rbf radial basis fn ANN KCL
ann-sse ANN using sum-of-squares data misfit term UEA
cca canonical correlation analysis ARPA, AUTH, UEA, UNIBE
cr conditional resampling KCL
cwg conditional weather generator DMI
dloci dynamical local rescaling of GCM pre intensity ETH
hyper4 random sampling within 4-dim hyperspace ADGB
loc local rescaling of GCM precipitation ETH
loci local rescaling of GCM precipitation intensity ETH
mar multivariate auto-regressive model USTUTT
mlr multiple linear regression ARPA, USTUTT
mlr-ct MLR using circulation types AUTH
ppci potential precipitation circulation index CNRS

Note that the abbreviations for the downscaling methods do not include the institution, as the institution is a standard part of the file name.


abbreviation description
ha3p-cta HADAM3P, Control, Member a
ha3p-a2b HADAM3P, Scenario A2, Member b
ech4-a2a ECHAM4, Scenario A2, Member a
ncep NCEP Reanalysis


Two Examples

name example
Conventions CF-1.0
title STARDEX indices of precipitation at the 10 FIC stations in the Alps
source FIC stations (OBS)
experiment HADAM3P-A2a
institution ETH
history 2004-06-16: generated by STARDEX indices software version 3.3.1

name example
Conventions CF-1.0
title Precipitation on the Alpine grid
source loci-eth (DS)
experiment HADAM3P-CTa
institution ETH


Precipitation at the 10 FIC stations in the Alps
STARDEX indices of precipitation at the 10 FIC stations in the Alps


source description
FIC stations (OBS) observations from the FIC stations
CHRM (MOD) CHRM model data
NCEP (RA) NCEP reanalysis data
dloci-eth (DS) downscaled data, method dloci from ETH
cca-auth (DS) downscaled data, using CCA developed by AUTH
ann-gamma-uea (DS) downscaled data, using a neural network developed by UEA

Note that the source metadata should include the type of data in parenthesis, i.e. observations (OBS), model data (MOD), reanalysis (RA), or downscaled (DS).

Note also that for downscaled data/indices the source metadata consists of two parts, the downscaling method and the institution that developed the method. See appendix A.6 for the abbreviations to be used for the downscaling method and appendix A.5 for the abbreviations to be used for the institution.


experiment description
ERA15 ERA15 reanalysis
NCEP NCEP reanalyis
HADAM3P-CTa HADAM3P, control, member a
HADAM3P-A2b HADAM3P, scenario A2, member b
ECHAM4-CTa ECHAM4, control, member a


See Appendix A.5.

List of NetCDF variables

Data Variables

variable long_name units
pre precipitation mm d-1
tmin minimum temperature degC
tmax maximum temperature degC
ct500 500hPa circulation types

STARDEX Precipitation Indices

variable long_name units
pav mean mm d-1
pfre frequency 1
pint intensity mm d-1
pqNN NN% quantile mm d-1
pfNN fraction of total above NN% quantile 1
pn10mm no events $>$ 10mm 1
pnlNN no events $>$ clim NN% quantile 1
pfl90 fraction of total above clim 90% quantile 1
pxNd maximum N-day total mm
ppww mean wet-day persistence 1
ppdd mean dry-day persistence 1
ppcr correlation of spell lengths 1
pxcdd max no consecutive dry days day
pxcwd max no consecutive wet days day
pwsav mean wet-spell length day
pwsmed median wet-spell length day
pwssdv stddev wet-spell length day
pdsav mean dry-spell length day
pdsmed median dry-spell length day
pdssdv stddev dry-spell length day

STARDEX Temperature Indices

variable long_name units
tnav mean Tmin degC
tnq10 Tmin 10% quantile degC
tnq90 Tmin 90% quantile degC
tnf10 fraction of days Tmin $<$ 10% quantile 1
tnf90 fraction of days Tmin $>$ 90% quantile 1
tnfd number of frost days 1
tncwd cold wave duration day
tnfsl frost season length day
txav mean Tmax degC
txq10 Tmax 10% quantile degC
txq90 Tmax 90% quantile degC
txf10 fraction of days Tmax $<$ 10% quantile 1
txf90 fraction of days Tmax $>$ 90% quantile 1
txice number of ice days 1
txhwd heat wave duration day
txhw90 90% quantile heat wave duration day
trav mean diurnal temperature range degC
trq10 temperature range 10% quantile degC
trq90 temperature range 90% quantile degC
tiaetr intra-annual extreme temperature range degC
tgdd growing degree days degC day
tgsl growing season length day

next_inactive up previous