History of modifications

Related Documents

Acronyms

About the dataset

This dataset is based on the provisional data from ECMWF's ERA5 (referred to as ERA5T) and ERA-Interim (no longer produced) reanalyses and therefore is subject to change should a significant production problem be found to have occurred. The release of this dataset usually happens a few days after the last observations of each month have been made.
This dataset contains the following products calculated from the ERA5 and ERA-Interim data:

• monthly means
• monthly anomalies with respect to 1981-2010 climatologies
• monthly anomalies with respect to 1991-2020 climatologies for ERA5 only
• 12 months running mean anomalies with respect to 1981-2010 climatologies
• 12 months running mean anomalies with respect to 1991-2090 climatologies for ERA5 only
• monthly climatologies based on 1981-2010
• monthly climatologies based on 1991-2020 for ERA5 only

for the following ECVs:

• Surface air temperature
• Total precipitation
• Volumetric soil moisture for the uppermost layer of soil 0-7 cm (the percentage of water per unit volume)
• Surface air relative humidity
• Sea-ice cover

Anomalies are calculated relative to one of two climatological averaging periods, 1981-2010 or (for ERA5 only) 1991-2020. The anomaly for a particular variable and month is the difference between the value of the variable for that month and its climatology, which is the average value of the variable during the reference period for that same month of the year. The 12-month running-mean anomalies are calculated such that anomalies are averaged over a 12-month period, from the selected month minus 11 months to the selected month.

Some parameters have been adjusted to compensate for various production issues, as indicated in Table 1. The adjustments described below refer to adjustments to the monthly mean fields. These fields are then used to calculate the respective climatological fields and anomalies. A detailed evaluation of ERA5 surface temperature and humidity is provided by Simmons et al. (2021).

Table 1: Main variables and note on adjustments 

ERA5

Surface air temperature

This product is based on the monthly mean ERA5 dataset.
Surface air temperatures have been adjusted for the period 1979-2013 to compensate for an inadvertent failure in production to utilize observationally-based analyses of the water temperatures of the Great Lakes. Observationally-based analyses were used in ECMWF's earlier ERA-Interim reanalyses. The monthly average surface air temperatures over the Great Lakes from ERA5 were thus adjusted by the 1981-2010 average of the differences between the ERA-Interim (adjusted as below) and ERA5 temperatures for the month in question. The adjustment is applied only over the Great Lakes; elsewhere the monthly average temperatures and corresponding climatological fields are derived entirely from ERA5 analyses.

In v01 of the data set fields with the wrong processing were made available. V02 rectifies this, see more here.

Surface air relative humidity

Relative humidity is calculated from the archived sub-daily temperature, dewpoint temperature and surface pressure fields, and averaged for each month in the same way as for other fields.

Values of relative humidity over land and sea are taken from the analyses.

0-7cm volumetric soil moisture

0-7cm volumetric soil moisture is derived from the ERA5 "Volumetric soil water layer 1" monthly mean field. For all ERA5 soil moisture calculations a mask is applied for regions designated to have permanent ice cover (including Antarctica and much of Greenland) or no vegetation, or that otherwise have a climatological mean annual total precipitation rate of less than 0.3 mm day^-1. The soil moisture values are set to undefined over these regions. Over other areas values are identical to the original field.

ERA-Interim

Surface air temperature

This product is based on the monthly mean ERA-Interim dataset.
Values over sea are taken from the background forecast model, not the analyses, to avoid a detrimental effect of analysing biased air-temperature observations from ships. Values over ice-free sea prior to 2002 are further adjusted by subtracting 0.1°C to account for a change in bias that arose from changing the source of sea-surface temperature analysis. Grid cells designated as sea are defined by the fractional land sea mask. The adjustment is described in detail in Simmons, et al. (2017).

Surface air relative humidity

Similar to ERA5, relative humidity is calculated from the archived sub-daily temperature, dewpoint temperature and surface pressure fields, and averaged for each month in the same way as for other fields.
Values of relative humidity over sea are taken from the background forecast model, not the analyses, for consistency with what is done for the temperature.

0-7cm volumetric soil moisture

0-7cm volumetric soil moisture is derived from the ERA-Interim "Volumetric soil water layer 1" monthly mean field. Similar to ERA5, for all ERA-Interim soil moisture calculations a mask is applied for regions designated to have permanent ice cover (including Antarctica and much of Greenland) or no vegetation, or that otherwise have a climatological mean annual total precipitation rate of less than 0.3 mm day^-1. The soil moisture values are set to undefined over these regions. Over other areas values are identical to the original field.

About reanalysis

Atmospheric reanalysis combines a weather model with observational data from satellites and ground sensors to build a complete long-term record of Earth's climate. Reanalysis datasets are used for climate monitoring and studies, meteorological research and have become widely used in a variety of sectoral applications. They also support numerical weather prediction by providing a reference against which a quality of forecasts can be checked.

ERA5

ERA5 is a global atmospheric reanalysis currently covering a period from 1979 onwards. It will eventually cover a period from 1950 to near real time. ERA5 was produced using 4D-Var data assimilation in Cycle 41R2 of ECMWF's Integrated Forecasting System. The native ERA5 dataset includes one 31 km high resolution realization (reduced Gaussian grid N320), with a temporal resolution of one hour. However, data in this dataset are available on a 0.25° regular latitude/longitude grid with monthly resolution.
ERA5 replaces ECMWF's previous atmospheric reanalysis ERA-Interim and has several improvements compared to ERA-Interim:

• use of Cycle 41R2 brings benefits from 10 years of improvements in numerical weather analysis and prediction;
• use of improved historical observations, boundary conditions and external forcing;
• finer spatial and temporal resolutions.

Surface air temperature

ERA5 surface air temperature is defined on the whole global domain and over all surfaces. Over land, values of surface air temperature are determined from observational records for regions where plentiful observations of surface air temperature were made. Elsewhere, the background forecast model plays a stronger role, helping values of surface air temperature to be derived from other types of observation, such as sea-surface temperatures and winds. Satellite data on the extent of sea-ice cover are important in winter, as surface air temperatures tend to be much warmer over open sea than over ice. Observations of conditions higher in the atmosphere provide some additional information.

Surface air temperature is adjusted as described in Table 1.

Sea-ice cover

ERA5 does not analyse the sea ice directly, instead it makes use of the fractional sea- ice cover or concentration produced elsewhere with additional steps to ensure consistency across the whole time period. The EUMETSAT OSI-SAF Climate data Record version 1.2 product (OSI-409a) was used in the sea ice analysis until August 2007. Later this product was replaced by the operational OSI-SAF product, which is a part of the OSTIA product, to determine ERA5 sea-ice cover values. The ERA5 data are defined only for model grid points that are designated as sea or lakes.

Sea-ice cover is not adjusted compared to the original ERA5 values.

Total precipitation

The ERA5 total precipitation values come from a sequence of 12-hour background forecasts. These forecasts owe their skill to many types of data, of which a set of rain- affected microwave radiances assimilated over sea was used in significant numbers from 1992 onwards. From 2009, a precipitation product over the USA, based on radar and rain gauge data, was used.
Monthly mean total precipitation data is in units of "m", effectively accumulated over a day, thereby giving "m day^-1". For consistency with the ECMWF parameter database, the units in the GRIB files are "m".

No adjustment is applied to the original ERA5 values.

0-7cm volumetric soil moisture

Soil-moisture values are for the uppermost 7 cm of soil, as modelled by the Hydrology Tiled ECMWF Scheme for Surface Exchanges over Land (H-TESSEL). H-TESSEL uses a variety of soil texture classes with specific properties and takes into account surface runoff. Soil moisture analyses are constrained by the mismatches between 2m temperature and humidity observations and corresponding background forecasts, as well as by the soil moisture pseudo-observations derived from ASCAT and corresponding background forecasts, in snow-free regions. The constraint is particularly effective where the density of synoptic observations is high.

Soil moisture is adjusted as described in Table 1.

Surface air relative humidity
Values of the relative humidity of surface air are determined directly from observational records for regions where plentiful observations of surface air humidity were made.
The GRIB code for the surface air relative humidity in the GRIB files provided is that of the relative humidity "157".

Surface air relative humidity is not adjusted.

ERA-Interim

ERA-Interim is a reanalysis for the period from 1979 until the end of August 2019. It combines information from meteorological observations with background information from a forecast model, using 4D-Var data assimilation in Cycle 31R2 of the IFS. The atmospheric observing system underwent several improvements leading up to 1979.

Surface air temperature

ERA-Interim surface air temperature values are derived using the same scheme as for ERA5.

Surface air temperature is adjusted as described in Table 1.

Sea-ice cover

ERA-Interim does not analyse sea ice observations directly. Instead, it incorporates analyses of fractional sea-ice cover (or concentration) produced elsewhere. The analyses used for dates prior to 2002 were those developed collaboratively for use in the earlier ERA-40 reanalysis. This was followed for several years by the operational analyses of the US National Centers for Environmental Prediction (NCEP). For dates from February 2009 onwards, the Operational Sea Surface Temperature and Sea Ice Analysis (OSTIA) product was used to determine the sea-ice cover. The OSTIA product in turn uses a global sea-ice analysis produced by the EUMETSAT Ocean and Sea Ice Satellite Application Facility (OSI-SAF). OSTIA is also used in the ECMWF's operational forecasting. The OSTIA product includes data on the ice cover of inland seas and large lakes. These data are carried over into what for convenience is termed simply the ERA- Interim sea-ice cover. The processing of the product undertaken by ECMWF is described in the documentation of the forecasting system. Only analysed values of ice cover greater than 20% are accepted in the processing of this product. Values lower than this nevertheless occur in the monthly mean ERA-Interim data products.

Sea-ice cover is not adjusted compared to the original ERA-Interim values.

Precipitation

Similar to ERA5, values of total precipitation come from a sequence of 12-hour background forecasts. These forecasts owe their skill to many types of data, in particular to the rain-affected microwave radiances assimilated over sea, which were used in significant numbers from 1992 onwards. Observations from rain gauges are not assimilated, but provide largely independent datasets for evaluation.
Monthly mean total precipitation data is in units of "m", effectively accumulated over a day, thereby giving "m day^-1". For consistency with the ECMWF parameter database, the units in the GRIB files are "m".

Total precipitation is not adjusted compared to the original ERA-Interim values.

0-7cm volumetric soil moisture

Soil-moisture values are for the uppermost 7cm of soil, as modelled by the Tiled ECMWF Scheme for Surface Exchanges over Land ( TESSEL ). TESSEL uses a single soil texture class with the same infiltration capacity for all grid points. This means that the soil moisture anomalies of deeper layers are quite strongly correlated with those of the top 7cm. In addition, the correlation between different soil layers can be amplified by the smoothing effect of monthly averaging. Direct observations of soil moisture are not assimilated; analyses are constrained instead by the mismatches between synoptic temperature and humidity observations and corresponding background forecasts, in snow-free regions when meteorological conditions are appropriate. The constraint is particularly effective over Europe and North America, where the density of synoptic observations is high.

Soil moisture is adjusted as described in Table 1.

Surface air relative humidity

Over land, values of the relative humidity of surface air are determined directly from observational records for regions where plentiful observations of surface air humidity were made. Elsewhere, the background forecast model plays a stronger role, enabling values of surface relative humidity to be derived less directly from other types of assimilated observation.
The GRIB code for the surface air relative humidity in the GRIB files provided is that of the relative humidity "157".

Surface air relative humidity is adjusted as described in Table 1.

Differences between published versions

The table below summarizes GRIB codes and grids for each variable in the dataset. Also given are current file version numbers and differences between previous version numbers.

Table 2: Information about GRIB codes, version numbers and differences between version numbers.

Known issues

• Update from v01 to v02 of ERA5 surface air temperature:

In v01, the intended adjustment (see page 3) was applied globally, rather than restricted to the Great Lakes, affecting mean and anomaly fields from 1979-2013. Climatologies in v01 were adjusted only over the Great Lakes. Fields for 2014 onwards are not affected. In v02 the following has been done:

1. 1. reprocessing of all fields (mean, anomalies, climatologies) with the intended adjustment only over the Great Lakes (based on the ERA5 lake fraction)
   2. regridding from the native grid to the regular grid done with an updated interpolation package compared to v01.

• Update from v01 to v02 of ERA5, from v02 to v03 of ERA-Interim 12-month running means:

A minor bug affected v01 of ERA5 and v02 ERA-Interim 12-month running averages. To remove this bug reprocessing was done to create v02 ERA5 and v03 of ERA-Interim 12-month running means.

• Introduction of additional reference period:

In March 2021, the dataset was extended by introducing an additional reference period for the calculation of ERA5 climatological and anomaly fields. There is no change to the underlying calculation and as such the version numbers remain as before.

• ERA5T issue over Central Asia Sep-Dec 2021 and the C3S Climate Bulletin

Summary

"Essential climate variables for assessment of climate variability from 1979 to present" constitutes the basis for the C3S monthly Climate Bulletin. For the months of September, October, November and December 2021, erroneously large values of soil moisture, and to a lesser extent, of 2m temperature and of humidity became noticeable in the timely ERA5 product (ERA5T) for central Asia. A full description of the issue and how it has been addressed can be found here.

After correction, the timely published "Essential climate variables for assessment of climate variability from 1979 to present" have been replaced for each of the affected months with a newly post-processed version based on the corrected, consolidated ERA5 (rather than ERA5T, which bore the erroneous values). Graphics and time-series data of the monthly climate bulletin have been accordingly updated and replaced for the affected months in the respective bulletin. The table of the bottom of the page outlines which months have been updated to date. 

Example of differences 

October hydrological variables for Europe

The outdated and new maps of hydrological variables for Europe for October highlight the main difference between the dataset containing the erroneous values and the new one based on the consolidated ERA5. The issue is very apparent in Central Asia in the soil moisture maps and, to some extent, for the other variables too. 

Figure 1: Anomalies in precipitation, the relative humidity of surface air, the volumetric moisture content of the top 7 cm of soil and surface air temperature for October 2021 with respect to October averages for the period 1991-2020. The darker grey shading denotes where soil moisture is not shown due to ice cover or climatologically low precipitation. Data source: Left ERA5T (erroneous), Right: ERA5 consolidated (new & corrected) Credit: Copernicus Climate Change Service/ECMWF.

The resulting difference for the same variables in regions other than Central Asia, in general for other variables and for aggregated statistics covered in the Climate Bulletin is smaller than their respective typical error, and as such there is no further impact of the issue on the information given in the Bulletins. For full traceability and transparency, the original data and graphics based on ERA5T are made available in the zip-files in the table below. 

Access to new and erroneous monthly summaries 

Even though the issue mainly affects Central Asia, for the months September to December 2021 all “Essential climate variables for assessment of climate variability from 1979 to present" data entries in the CDS as well as all related graphics and data-files displayed in the Climate Bulletin (temperature, hydrological variables, sea ice) have been replaced. The replacement of graphics and data had however no impact on the text of the bulletin, as the issue was known at the time of writing, and the figures given in the text do not differ between the two versions of the dataset. 

The table below outlines which months have been replaced to date and where to find the original erroneous and the new corrected versions of all data and graphics.

Table 3: Original erroneous and the new corrected versions of data and graphics.

• Incomplete retrieval of ERA5 during DHS move affecting RH used in the Climate bulletin 

The “Essential climate variables for assessment of climate variability from 1979 to present" dataset constitutes the basis for the C3S monthly Climate Bulletin.  

In the frame of a major hardware change at ECMWF, and for a limited period, the services usually provided by the Data Handling System, from which this dataset is retrieved, were distributed across different systems. For this, a bespoke retrieval script was developed, within which an error led to the incomplete retrieval of surface air relative humidity data for the month of September 2022. Instead of covering the period 1-30 September, the retrieved files only contained the data for the period 07-30 September. This one-off problem, had an impact on all data and information derived from those, namely the data for September and October 2022. Upon publication of data for November 2022, the problem was identified and the data for the month of September 2022 correctly retrieved. As of mid-December 2022, the published "Essential climate variables for assessment of climate variability from 1979 to present" data have thus been replaced on the CDS for each of the affected months with a new version.   

 In turn, related maps, time-series and data of the monthly climate bulletin have been accordingly updated and replaced for the affected months in the respective bulletin. The table of the bottom of the page outlines which months have been updated to date.  

Table 4: Original erroneous and the new corrected versions of data and graphics. 

References

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., Bid-lot, 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. Roy. Meteor. Soc., 137, 553–597, https://doi.org/10.1002/qj.828 .

Simmons, A., H. Hersbach, J. Muñoz-Sabater, J. Nicolas, F. Vamborg, P. Berrisford, P. de Rosnay, K. Willettand J. Woollen. Low frequency variability and trends in surface air temperature and humidity from ERA5 and other datasets. 2021, ECMWF Technical Memoranda, 811, doi: 10.21957/ly5vbtbfd.

Simmons, A. J., Berrisford, P., Dee, D. P., Hersbach, H., Hirahara, S., and Thépaut, J.- N., 2017: A reassessment of temperature variations and trends from global reanalyses and monthly surface climatological datasets, Q. J. Roy. Meteor. Soc., 143, 101–119, https://doi.org/10.1002/qj.2949 .

Simmons, A., K. Willett, P. Jones, P. Thorne, and D. Dee, 2010: Low‐frequency variations in surface atmospheric humidity, temperature, and precipitation: Inferences from reanalyses and monthly gridded observational data sets. J. Geophys. Res., 115, D01110.

Philippe Lopez 2013: Experimental 4D-Var Assimilation of SYNOP Rain Gauge Data at ECMWF, https://doi.org/10.1175/MWR-D-12-00024.1

Related articles

Related articles appear here based on the labels you select. Click to edit the macro and add or change labels.