ERA5 is a global reanalysis dataset with hourly weather data from 1979 until present in 31 km spatial resolution. Various atmospheric levels as well as soil information are included. It is updated regularly with a delay of 3 months.
ERA5 has been developed by ECMWF Copernicus Climate Change Service (C3S) since 2016 and available as open data. It is the successor of ERA-Interim.
A reanalysis dataset utilizes archived satellite observations and weather station measurements in combination with weather models to provide a globally consistent dataset with multiple atmospheric levels. For example, in meters above ground (2m, 10m, 100m etc.) or on different pressure levels (900 hPa, 850 hPa etc.).
|1979-01-01 until 2019-12-31
|ECMWF N320 reduced Gaussian Grid
|Global: 180°W to 180°E / 89.784°S to 89.784°N in WGS84 coordinates
For a more general description of the ERA5 reanalysis dataset, see here.
ERA5 is available via meteoblue as part of history+ to the general public and dataset APIs for business customers. Data are updated regularly and stored in meteoblue databases to provide easy access within seconds. meteoblue retrieves and stores data unmodified on the original ERA5 irregular reduced gaussian grid, to generate reproducible results, which are not influenced by the spatial interpolation scheme used to generate regular grids. We carefully handle data integrity to provide an independent and trustworthy data source.
With the use of meteoblue dataset APIs, it is also possible to downscale ERA5 to 1km or even more fine-grained spatial resolution. Although spatial interpolation and reprojection does not improve data quality, it is allows to combine different datasets in one common projection for further analysis.
ERA5 in meteoblue archives includes common variables that are used in scientific applications. In addition, meteoblue provides derived variables from the original ERA5 dataset. E.g. Wind speeds are calculated from wind U/V components or FAO reference evapotranspiration (ET₀, see below). A more detailed list of available variables and derivatives can be found below.
Currently no near-realtime updates are available. In the future, updates with a smaller delay of less than 2 weeks will be provided.
ERA5 Quality and limitations
Reanalysis datasets are based on measurements, but do not reflect measurements precisely. Gridded data divides the world evenly in small boxes
grid-cells. With a spatial resolution of 31 km, some measurement and observational data points reside in the same grid-cell, while the majority of grid-cells are not covered by any measurement station. A data assimilation process is used to combine observations and generate a global gridded dataset.
With a spatial resolution of 31 km and uncertainties in the data, local effects cannot be captured precisely. However, even for those locations you can analyze trends over the years, seasons, months, weeks or even day to day changes quite well, but be aware that the absolute values might be biased.
In comparison to weather models operated by meteoblue, ERA5 has a rather large delay until data is available. If your application relies on near-realtime updates or forecasts, we recommend using meteoblue NEMS datasets which offer data from 1985 up to 7 days ahead. So if you want to train your own AI or statistical model based on historical weather data and use it for forecasting, the meteoblue NEMS models have to be used, as ERA5 does not provide a forecast.
ERA5 is an excellent dataset for climate risk mitigation to minimize negative impacts and adapt to climate change. With a temporal extent starting in 1979, long term changes for agriculture, insurances, tourism and transportation can be estimated. With access to time-series data for any given location worldwide, history+ enables immediate data downloads for any location worldwide.
meteoblue provides an extensive set of commonly used weather variables of ERA5 and derived variables.
Many variables are provided on different atmospheric levels. Temperature is commonly used on 2 meters above ground and wind speeds on 10 m and 100 meters above ground. Selected variables are also available on different atmospheric pressure levels like 850 hPa.
ERA5 provides hourly values. Most weather variables are instantaneous in time: Temperature at 14:00 is the indicated temperature at 14:00. Other weather variables are aggregated over time. E.g. Precipitation at 14:00 is accumulated from 13:00 - 14:00. If not otherwise stated variables are instantaneous.
meteoblue history+ and APIs provide access to data for specified coordinates or locations. For the desired location, the most suitable ERA5 grid-cell with a 31 km extent will be selected. For complex terrain this might not be the closest grid-cell. If the user calls data for a location in a valley, the nearest grid-cell could be on a mountain with significantly higher elevation. The exact center-point coordinates of the selected ERA5 grid-cell will be shown in the downloadable XLSX or CSV files. Although the center-point coordinate might be a couple of kilometers apart of the point of interest, all weather data is still valid. Interpolating between grid-cells is highly discouraged.
Air temperature is an hourly instantaneous variable in ERA5. Typically, the
2 m above gnd level is used. With a spatial resolution of 31 km, this value might differ from local measurements:
- Within 31 km, temperature varies due to local effects like heat build-up, terrain, direct solar radiation on different surface albedos or vegetation. Please consider those effects for further analysis and interpretation.
- Locations of interest could be on different terrain elevations. An ERA5 grid-cell of 31 km has a defined elevation. meteoblue corrects temperature to match elevation temperature differences. This indicated with the level
2 m corrected.
Temperature in ERA5 is available for atmospheric pressure levels
700 hPa and
In contrast to the 2 meter above ground temperature, skin temperature is equivalent to the temperature measured directly above the surface. Depending on the vegetation and terrain type, skin temperatures could be significantly higher for example on sand. Surfaces that could lead to very high skin temperatures like concrete or tarmac are not modelled within reanalysis data. Skin temperature can be used to analyse dew.
Relative humidity and vapor pressure deficit
Relative humidity is calculated from 2 meter above ground air temperature and dew point temperature.
Vapor pressure deficit is important for plant water transport. It is not directly available in ERA5 but calculated using vapor pressure saturation according to A. Wexler.
Atmospheric air pressure is reduced to mean sea level as most commonly used for weather reports.
Precipitation and snow
Precipitation amount includes rain, convective precipitation as well as snow fall. It is accumulated over the past hour. Precipitation at 14:00 is the precipitation sum from 13:00 - 14:00. Especially convective precipitation in form of showers or very intense precipitation can be a local and short-term event and therefore not precisely covered in gridded reanalysis with a 31 km spatial resolution. Common quality indicators are probability of detection (POD) and Heidke Skill Score (HSS). More information about ERA5 precipitation quality can be found in meteoblue verification studies.
The total precipitation sums over a longer period tend to be relatively reliable. Analyzing pattern of droughts in frequency and duration works well.
Excessive precipitation leads to runoff. ERA5 runoff is calculated with the
HTESSEL land surface model and provides the amount of surface water, that is not absorbed by the soil within one hour. Soil water freezing and melting is considered. It does not provide any information about flood or river discharge.
Snowfall amount is based on the amount of precipitation that fell down as snow converted to centimeters. The conversion assumes that 1mm of water is equivalent to 7mm of snow. The accumulating snow depth is provided as an additional variable, that also account for compacting and melting. Snow depth and snow fall may not be perfectly consistent. Snow depth without snow fall can occur due to the combination of observational data and inconsistencies in the precipitation modeling.
Cloud cover, sunshine duration, cloud ice and cloud water
Clouds are expressed in percent with 0% = no clouds and 100% = fully covered. Clouds do not disperse evenly, but cluster and form patterns. Gridded datasets like ERA5 calculate the mean of the cloud cover over a 31 km grid-cell and within 1 hour.
Cloud layers are defined according to WMO as:
- low clouds: from the surface to 2 km (surface to 800 hPa)
- medium clouds: 2 to 6 km (800 to 450 hPa)
- high clouds: 6 to >13 km (450 to model top)
Total cloud cover combines all 3 cloud layers as a weighted average whereby high clouds are less significant to the total cloud cover.
Based on total cloud cover, sunshine duration in minutes is calculated. The number of minutes of daylight within one hour is multiplied by the total cloud cover fraction. A cloud fraction of 90% would indicate 6 minutes of sunshine within one hour. Sunshine duration as a sum over 1 year is an easily understandable way to compare different locations.
Total column cloud water is the total amount of water in clouds from bottom to top of the atmosphere excluding rain water droplets, which are much larger and heavier. It is expressed as liquid water in units of mm, which is equivalent to kg/m2.
Total column cloud ice. is the total amount of ice in clouds from bottom to top of the atmosphere. It is expressed as equivalent liquid water content in units of mm, which is equivalent to kg/m2. Snow, hail or other aggregated ice crystals are not included in this parameter.
The global horizontal irradiation (GHI) is offered as shortwave radiation in Watt per square meter W/m². It is the sum of direct and diffuse radiation on the horizontal plane. These values are not instantaneous, but backwards averaged over the preceding hour. A value at 14:00 is the average irradiation between 13:00 and 14:00. Therefore, these values cannot be compared directly with instantaneous measurements.
UV and longwave radiation are also given as backwards averages in W/m².
For daily radiation sums, selecting the unit
Joules in the interfaces will provide Mega Joules per day.
Long wave radiation
The long wave radiation in Watt per square meter W/m² is the infrared emmision from the sky and especially from the clouds. For daily radiation sums, selecting the unit
Joules in the interfaces will provide Mega Joules per day.
The UV radiation provided by ERA5 is the integration over all UV wavelengths (200 nm to 440 nm). It is not easily possible to compute the UV-Index based on this data as the UV is not available for different wavelengths. The erythemal UV index would need to be computed as an integration between 280 nm and 400 nm of the UV irradiance at ground level, weighted with the erythemal action spectrum.
Wind and gusts
Wind speeds in ERA5 are instantaneous values and available at the levels 10 meters, 100 meters, 850 hPa, 800 hPa, 700 hPa, 500 hPa and 250 hPa.
Wind direction is given in degrees from 0° (wind blowing from north), 90° (east wind), 180° (south wind) and 270° (west wind). If daily aggregations are called, the meteoblue APIs will provide the dominant wind direction. Instead of a simple mean wind direction, the dominant wind direction considers also the wind speeds at given directions.
Wind gusts are based on short-term wind speed turbulence in an hour. Gusts indicate the level of turbulence as such they could be lower than regular wind speeds.
Transpiration, evapotranspiration and potential evapotranspiration
Evapotranspiration is the sum of evaporation (evaporation over soils, lakes, seas) and transpiration (evaporation from plants). It depends on soil properties (e.g. moisture) and meteorological conditions (e.g. wind, temperature, radiation). Evapotranspiration is expressed in millimeter (mm) as a backwards sum analog to precipitation in ERA5.
Potential evapotranspiration is a theoretical value that indicates how much water would evaporate from a surface with unlimited water availability. Thus, it shows how much water could be evaporated with the total available energy. For instance, the potential evapotranspiration in the Sahara is very high, whereas the actual evaporation is almost zero, because there is no water available. Potential evapotranspiration is expressed in millimeter (mm).
While evapotranspiration and potential evapotranspiration are available in ERA5, FAO reference evapotranspiration (ET₀) is derived using hourly temperature, wind speed, humidity and shortwave radiation. See Reference Evapotranspiration (Eto)
Soil temperature and soil moisture
Soil temperature and moisture is available for the levels below surface
28-100 cm and
100-255 cm. Moisture is given as volumetric fraction from 0 to 1 describing the mean moisture within the specified level.
Soil moisture available to plant considers the ERA5 soil type classification with field water capacity and permanent wilting point.
Planetary Boundary Layer height (PBL height)
The PBL height indicates up to which height the influence of the surface is significant or up to which height pollutants emitted from the surface are mixed. The PBL can be as low as a few tens of metres, such as in cooling air at night, or as high as several kilometres on a hot sunny day. Air quality is significantly reduced in low PBL height situations, as pollutants emitted from the surface are mixed into a smaller volume of air.
Convective Available Potential Energy (CAPE)
CAPE is a measure of instability (or stability) of the atmosphere and can be used to assess the potential for the development of convection, which can lead to heavy rainfall, thunderstorms and other severe weather. In thunderstorms values of 1000 J/kg or more are often observed.
Geo Potential Height
The geopotential height approximates the actual height of a pressure surface above mean sea-level. In warm air, the geopotential height of e.g. the 800 hPa pressure level is at a much higher altitude than in cold air. The geopotential height also plays an important role in synoptic meteorology.
- 2018-06-01: First years of ERA5 available in the meteoblue databases
- 2019-04-01: Data starting from 1990
- 2019-11-01: Data starting from 1979
- 2020-03-01: Data now available until 2019-12-31
- 2020-03-27: ERA5 available in history+ with 40 years of data with over 50 weather variables