ERA-Interim production stopped on 31st August 2019
For the time being and until further notice, ERA-Interim (1st January 1979 to 31st August 2019) shall continue to be accessible through the ECMWF Web API. Keeping in mind that ERA-Interim is published with an offset of about three months from the dataset's reference date, ERA-Interim August 2019 data will be made available towards the end of October 2019.
Here we document the ERA-Interim dataset, which, covers the period from 1st January 1979 to 31st August 2019.
ERA-interim was produced using cycle 31r2 of the Integrated Forecast System (IFS - CY31R2) 2006, with 60 vertical levels, with the top level at 0.01 hPa. Atmospheric data are available on these levels and they are also interpolated to 37 pressure, 16 potential temperature and 1 potential vorticity
level(s). "Surface or single level" data are also available, containing 2D parameters such as precipitation, 2m temperature, top of atmosphere radiation and vertical integrals over the entire atmosphere.
Generally, the data are available at a sub-daily and monthly frequency and consist of analyses and 10-day forecasts, initialised twice daily at 00 and 12 UTC. Most analysed parameters are also available from the forecasts. There are a number of forecast parameters, e.g. mean rates and accumulations, that are not available from the analyses.
How to download ERA-Interim
The data are archived in the ECMWF data archive MARS and datasets are available through both the Web interface and the ECMWF WebAPI, which is the programmatic way of retrieving data from the archive.
The 4D-Var data assimilation uses 12 hour windows from 15 UTC to 03 UTC and 03 UTC to 15 UTC (the following day).
The model time step is 30 minutes.
Data organisation
The data can be accessed from MARS using the keywords class=ea and expver=0001. Subdivisions of the data are labelled using stream, type and levtype.
Stream:
oper: sub-daily products
wave: sub-daily ocean-waves products
mnth: synoptic monthly means
moda: monthly means of daily means
mdfa: monthly means of daily forecast accumulations
Type:
an: analyses
fc: forecasts
Levtype:
sfc: surface or single level
pl: pressure levels
pt: potential temperature levels
pv: potential vorticity levels
ml: model levels
In MARS: the date and time of the data is specified with three MARS keywords:
date
time
step
For analyses date and time, specify the analysis time and step equal to 0 hours. For forecasts date and time, choose the forecast start time and then step specifies the number of hours since that start time. The combination of date, time and forecast step defines the validity time. For analyses, the validity time is equal to the analysis time. Refer to ERA-Interim: 'time' and 'steps', and instantaneous, accumulated and min/max parameters for further details.
Spatial grid
The horizontal grid spacing of ERA-Interim atmospheric model and reanalysis system is around 80 km (reduced Gaussian grid N128) which became around 83km (
\( 0.75^{\circ} \)
) when interpolated to a regular lat/lon grid.
Depending on the parameter, the data are archived either as the full T255 spectral resolution and on the corresponding N128 reduced Gaussian grid, depending on their basic representation in the model. The coupled ocean-wave model data are produced and archived on a reduced
\( 1.0^{\circ}\times 1.0^{\circ} \)
latitude/longitude grid. For more information see ERA-Interim archive report Version 2.0, Section 2.
It is possible to specify the grid when downloading data. Available options are:
GRIB1 format (native format): native grid, different Gaussian grids, and regular lat/lon grids. Data will be interpolated to your chosen grid if different to the native one.
NetCDF format: NetCDF only supports regular lat/lon grids. Data is transformed to a regular lat/lon grid, interpolated to your selected resolution, and converted from the native GRIB1 format to NetCDF.
Longitudes range from 0 to 360, which is equivalent to -180 to +180 in Geographic coordinate systems.
Temporal frequency
Analyses of atmospheric fields on model levels, pressure levels, potential temperature and potential vorticity, are available every 6 hours at 00, 06, 12, and 18 UTC. Forecasts run twice at 00 and 12 UTC and provide 3 hours output for surface and pressure level parameters up to 24 hours, with decreasing frequency to 10 days.
Wave spectra
The ERA-Interim atmospheric model is coupled ocean-wave model resolving 30 wave frequencies and 24 wave directions at the nodes of its reduced
\( 1.0^{\circ}\times 1.0^{\circ} \)
latitude/longitude grid.
If you want to download the data in NetCDF format, please add the 'format' and 'grid' parameters:
#!/usr/bin/env python
from ecmwfapi import ECMWFDataServer
server = ECMWFDataServer()
server.retrieve({
......
"grid": "0.75/0.75", # Spatial resolution in degrees latitude/longitude
"format": "netcdf"
"target": "2d_spectra_201601.nc"
})
Decoding 2D wave spectra in GRIB
To decode wave spectra in GRIB format we recommend ecCodes. Wave spectra are encoded in a specific way that other tools might not decode correctly.
In GRIB, the parameter is called 2d wave spectra (single) because in GRIB, the data are stored as a single global field per each spectral bin (a given frequency and direction), but in NetCDF, the fields are nicely recombined to produce a 2d matrix representing the discretized spectra at each grid point.
The wave spectra are encoded in GRIB using a local table specific to ECMWF. Because of this, the conversion of the meta data containing the information about the frequencies and the directions are not properly converted from GRIB to NetCDF format. So rather than having the actual values of the frequencies and directions, values show index numbers (1,1) : first frequency, first direction, (1,2) first frequency, second direction, etc ....
For ERA, because there are a total of 24 directions, the direction increment is 15 degrees with the first direction given by half the increment, namely 7.5 degree, where direction 0. means going towards the north and 90 towards the east (Oceanographic convention), or more precisely, this should be expressed in gradient since the spectra are in m^2 /(Hz radian) The first frequency is 0.03453 Hz and the following ones are : f(n) = f(n-1)*1.1, n=2,30
Also note that it is NOT the spectral density that is encoded but rather log10 of it, so to recover the spectral density, expressed in m^2 /(radian Hz), one has to take the power 10 (10^) of the NON missing decoded values. Missing data are for all land points, but also, as part of the GRIB compression, all small values below a certain threshold have been discarded and so those missing spectral values are essentially 0. m^2 /(gradient Hz).
Decoding 2D wave spectra in NetCDF
The NetCDF wave spectra file will have the dimensions longitude, latitude, direction, frequency and time.
However, the direction and frequency bins are simply given as 1 to 24 and 1 to 30, respectively.
The direction bins start at 7.5 degree and increase by 15 degrees until 352.5, with 90 degree being towards the east (Oceanographic convention).
The frequency bins are non-linearly spaced. The first bin is 0.03453 Hz and the following bins are: f(n) = f(n-1)*1.1; n=2,30. The data provided is the log10 of spectra density. To obtain the spectral density one has to take to the power 10 (10 ** data). This will give the units 2D wave spectra as m**2 s radian**-1 . Very small values are discarded and set as missing values. These are essentially 0 m**2 s radian**-1.
This recoding can be done with the Python xarray package, for example:
import xarray as xr
import numpy as np
da = xr.open_dataarray('2d_spectra_201601.nc')
da = da.assign_coords(direction=np.arange(7.5, 352.5 + 15, 15))
da = da.assign_coords(frequency=np.full(30, 0.03453) * (1.1 ** np.arange(0, 30)))
da = 10 ** da
da = da.fillna(0)
da.to_netcdf(path='2d_spectra_201601_recoded.nc')
Units of 2D wave spectra
Once decoded, the units of 2D wave spectra are m2 s radian-1
Instantaneous and Accumulated parameters
Instantaneous parameters represent an average over the model time step (30min). Accumulated parameters are accumulated from the start of the forecast, ie. from 00 UTC or 12 UTC to the time step selected. All the analysed fields are instantaneous instead forecast data could be either instantaneous or accumulated, depending on the parameter. More detailed information on parameters are shown in Parameter listing.
Minimum/maximum since the previous post processing
In ERA-Interim there are some parameters named '...since previous post-processing', for example 'Maximum temperature at 2 metres since previous post-processing'. This represents the maximum temperature between the previous archived forecast 'Step' and the forecast 'Step'. For example, 'Maximum temperature at 2 metres since previous post-processing' with start time 00 UTC and Step=9, is the maximum 2m temperature in the 3-hour period between 06 UTC and 09 UTC.
Monthly means
ERA-interim sub-daily data are monthly averaged on data with valid times or accumulation periods that fall within the calendar month in question. The different monthly means are:
Synoptic Monthly Means (stream=mnth) are the monthly averages:
for each analysis time (at the four main synoptic hours - 00, 06, 12, and 18 UTC)
for each forecast start time (00 and 12 UTC) and step (3, 6, 9, 12 etc).
Monthly Means of Daily Means (stream=moda) are only available for analysis and instantaneous forecast data.
Monthly Means of Daily Forecast Accumulations (stream=mdfa) are similar to Monthly Means of Daily Means but are for accumulated fields (eg precipitation, radiation) and three step ranges are available.
Tables 1-6 below describe the surface and single level parameters (levtype=sfc), Table 7 describes wave parameters, Table 8 describes the monthly mean exceptions for surface and single level and wave parameters and Tables 9-13 describe upper air parameters on various levtypes. Information on all ECMWF parameters is available from the ECMWF parameter database.
Parameters described as "instantaneous" below, are valid at the specified time.
Instantaneous, invariant, surface and single level parameters table
count
name
an
fc
paramId
units
1
Low vegetation cover
x
27
(0-1)
2
High vegetation cover
x
28
(0-1)
3
Low vegetation type (table index - see Table 10.1 in IFS CY31R1 Documentation)
Computation of near-surface humidity and snow cover
Near-surface humidity
Near-surface humidity is not archived directly in ERA datasets, but the archive contains near-surface (2m from the surface) temperature (T), dew point temperature (Td), and surface pressure[1] (sp) from which you can calculate specific and relative humidity at 2m:
Specific humidity can be calculated over water and ice using equations 7.4 and 7.5 from Part IV, Physical processes section (Chapter 7, section 7.2.1b) in the documentation of the IFS for CY31R1. Use the 2m dew point temperature and surface pressure (which is approximately equal to the pressure at 2m) in these equations. The constants in 7.4 are to be found in Chapter 12 (of Part IV: Physical processes) and the parameters in 7.5 should be set for saturation over water because the dew point temperature is being used.
Relative humidity should be calculated: RH = 100 * es(Td)/es(T)
The relative humidity can be calculated with respect to saturation over water, ice or mixed phase by defining es(T) with respect to saturation over water, ice or mixed phase (water and ice). The usual practice is to define near-surface relative humidity with respect to saturation over water.
[1] Access to surface pressure varies by dataset. For example, for ERA-Interim surface pressure is available from the Web Interface and from the WebAPI, while for ERA-40 surface pressure is not available from the Web Interface, but only via the WebAPI.
Snow Cover
In the ECMWF model (IFS), snow is represented by an additional layer on top of the uppermost soil level. The whole grid box may not be covered in snow. The snow cover gives the fraction of the grid box that is covered in snow. The method for calculating snow cover depends on the particular version of the IFS and for ERA-Interim is computed directly using snow water equivalent (ie parameter SD (141.128)) as:
snow_cover (SC) = min(1, RW*SD/15 )
where RW is the density of water = 1000. The Physical depth of snow (SD) where there is snow cover is:
Please use this as the main scientific reference to ERA-Interim:
Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi, S., Andrae, U., Balmaseda, M. A., Balsamo, G., Bauer, P., Bechtold, P., Beljaars, A. C. M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A. J., Haimberger, L., Healy, S. B., Hersbach, H., Hólm, E. V., Isaksen, L., Kållberg, P., Köhler, M., Matricardi, M., McNally, A. P., Monge-Sanz, B. M., Morcrette, J.-J., Park, B.-K., Peubey, C., de Rosnay, P., Tavolato, C., Thépaut, J.-N. and Vitart, F. (2011), The ERA-Interim reanalysis: configuration and performance of the data assimilation system. Q.J.R. Meteorol. Soc., 137: 553–597. doi: 10.1002/qj.828
For a more technical documentation of the contents of the ERA-Interim dataset please use:
Berrisford, P., P. Kållberg, S. Kobayashi, D. Dee, S. Uppala, A. J. Simmons, P. Poli, and H. Sato, 2011: Atmospheric conservation properties in ERA-Interim. Q.J.R. Meteorol. Soc., 137: 1381–1399. doi: 10.1002/qj.864
Dee, D. P., and Coauthos, 2011: The ERA-Interim reanalysis: configuration and performance of the data assimilation system. Q.J.R. Meteorol. Soc., 137: 553–597. doi:10.1002/qj.828