Page History

Introduction

Here we document the ERA5-Land dataset that, in its consolidated version, covers the period from January 1950 to 2-3 months before the present. In addition, the ERA5-Land-T version delivers non-checked close to Near-Real-Time (NRT) daily updates. ERA5-Land-T is synchronized with the close to NRT daily updates provided by the ERA5 climate reanalysis (ERA5T).

Introduction

Here we document the ERA5-Land dataset that, in its consolidated version, covers the period from January 1950 to 2-3 months before the present. In addition, the ERA5-Land-T version delivers non-checked close to Near-Real-Time (NRT) daily updates. ERA5-Land-T is synchronized with the close to NRT daily updates provided by the ERA5 climate reanalysis (ERA5T).

ERA5-Land is a replay of the land component of the ERA5 climate ERA5-Land is a replay of the land component of the ERA5 climate reanalysis, forced by meteorological fields from ERA5. Note that ERA5-Land always uses forcing fields based on the final release of ERA5 (i.e., expver=0001). ERA5-Land comes with a series of improvements making it more accurate for all types of land applications. In particular, ERA5-Land runs at enhanced resolution (9 km vs 31 km in ERA5). The temporal frequency of the output is hourly and the fields are masked for all oceans, making them lighter to handle. Click this link here for comparison of the ERA5-Land features against other ECMWF reanalyses.

...

Convention used in MARS: the date and time of the data is specified with three MARS keywords, 'date', 'time' and 'step'. For parameters labelled as analyses (see list of parameters), step=0 hours so that date and time specify the analysis time. All forecasts start at 00UTC (time=00 hours), and for parameters labelled as forecasts (see list of parameters), date specifies the forecast start day and step specifies the number of hours since the start of the forecast, with a maximum of step=24 hours. The combination of date, time and forecast step defines the validity date/time. For analyses, the validity date/time is equal to the analysis date/time.

...

For sub-daily data for the HRES (stream=oper) the parameters labelled as analyses (type=an) are available hourly. The once daily short forecasts, run from 00 UTC, also provide data hourly, with steps from 01 to 24. The uncertainty is currently provided by ERA5 EDA fields, which are available every 3 hours for the surface fields.

Data update frequency

Initial release data, i.e. data with just a few days behind real time, is called ERA5-Land-T. Both for the CDS and MARS, daily updates for ERA5-Land-T are available about 5 days behind real time and monthly mean updates are available about 5 days after the end of the month.

...

• Synoptic monthly means, for each particular time and forecast step (stream=mnth) - in the CDS, referred to as "monthly averaged by hour of day".
• Monthly means of daily means, for the month as a whole (stream=moda) - in the CDS, referred to as "monthly averaged". These monthly means are created from the hourly data for the HRES.

Monthly means for:

• parameters labelled as analyses or instantaneous forecasts are created from data with a valid time in the month, between 00 and 23 UTC on each day of the month.
• accumulations are created from data with a forecast period falling within the month. Monthly means of daily means for accumulations are created from the last forecast step (24) of the forecasts for each day of the month.

...

Parameters described as "instantaneous" in Table 2 are valid at the specified time.

Note that that in the tables below, "an" and "fc" is just a label used for convention to archive the data in MARS. 

...

Table 1: surface parameters: invariants (in time)

Table 2: stream=oper/mnth/moda, levtype=sfc: surface parameters: instantaneous

...

...

1. 
   

   Expand
   title Actual and potential evapotranspiration
   Actual evapotranspiration in the ERA5-Land datasets is called "Total Evaporation" (param ID 182) and is the sum of the following four evaporation components: Evaporation from bare soil Evaporation from open water surfaces excluding oceans Evaporation from the top of canopy Evaporation from vegetation transpiration For the ERA5-Land datasets, actual evapotranspirationand it's four components can be downloaded from the C3S Climate Data Store (CDS) under the category heading "Evaporation and Runoff". For details about the computation of actual evapotranspiration, please see Chapter 8 of Part IV : Physical processes, of the IFS documentation: ERA5-Land IFS cycle 45r1 The potential evapotranspiration in the ERA5-Land CDS dataset is given by the parameter potential evaporation (pev).  Pev data can be downloaded from the CDS under the category heading "Evaporation and Runoff", in the "Download data" tab for the ERA5-Land datasets. Note The definitions of potential and reference evapotranspiration may vary according to the scientific application and can have the same definition in some cases. Users should therefore ensure that the definition of this parameter is suitable for their application. Please note that to give more flexibilty to users,  PEV in ERA5 and ERA5Land are not computed in the same way: ERA5: the definition of PEV in ERA5 is computed for agricultural land as if it is well watered (no soil moisture stress)  and assuming that the atmosphere is not affected by this artificial surface condition.  ERA5Land: the definition of PEV in ERA5Land is computed as an open water evaporation (Pan evaporation) and assuming that the atmosphere is not affected by this artificial surface condition. Note Please note that based on ERA5-Land atmospheric forcing, other independent (offline) methods such us "Priesley-Taylor1 (1972) , Schmidt2 (1915) or de Bruin3 (2000)" can also be used to estimate Potential evapotranspiration. 1PRIESTLEY, C. H. B., & TAYLOR, R. J. (1972). On the Assessment of Surface Heat Flux and Evaporation Using Large-Scale Parameters, Monthly Weather Review, 100(2), 81-92. Retrieved Aug 27, 2021, from https://journals.ametsoc.org/view/journals/mwre/100/2/1520-0493_1972_100_0081_otaosh_2_3_co_2.xml  2Schmidt, W., 1915: Strahlung und Verdunstung an freien Wasserflächen; ein Beitrag zum Wärmehaushalt des Weltmeers und zum Wasserhaushalt der Erde (Radiation and evaporation over open water surfaces; a contribution to the heat budget of the world ocean and to the water budget of the earth). Ann. Hydro. Maritimen Meteor., 43, 111–124, 169–178. 3de Bruin, H. A. R., , and Stricker J. N. M. , 2000: Evaporation of grass under non-restricted soil moisture conditions. Hydrol. Sci. J., 45, 391–406, doi:10.1080/02626660009492337.

   
   

2. 
   

   Expand
   title How to use lake-related fields
   Independently whether a model grid point is over a lake or not, the IFS computes lake variables all over the globe, at each grid-box. This is to ease output field aggregation at diverse model resolutions and to have a warm start of the model with shorter spin-up time if lake cover is upgraded, i.e.,  it is still a decent lake initial condition if lake location are updated or a new lake is added operationally. Lake depths (input parameter for our lake parametrization) are specified for each grid-box either with in-situ values or with a default 25 m value; over ocean we use ocean bathymetry. Worth to mention that the later default values will be changed soon (extra information in this HESS reference). The computed  lake variable values are not taken into account in the total grid-box flux calculations if lake is not present in the grid-box. The lake fields provided in ERA5-Land can be used in combination with the lake location. The latter in the model is determined by lake cover field (parameter name CL, in fraction: 0 - grid-box has no lakes, 1 - grid-box is fully covered with lake/s). Lake depths are presented in the field DL (in meters). The ECMWF model also contains an ice module, a snow module and a bottom sediments module. The present setup of the IFS is running with no bottom sediment and snow modules (snow accumulation over ice is not allowed and snow parameters are used only for albedo purposes). In the implementation in the IFS lake ice can be fractional within a grid-box with inland water (10 cm of ice means 100 % of a grid-box or tile is covered with ice; 0 cm of ice means 100 % of the grid-box is covered by water; in between a linear interpolation is applied) (Manrique-Sunen et al., 2013). At present, the water balance equation is not included for lakes and the lake depth and surface area are kept constant in time (IFS Documentation, 2017, chapter 8 and 11 ). Lake parametrization also requires the lake fraction CL, lake depth DL (preferably bathymetry), and lake initial conditions. DL is the most important external parameter that uses the lake parametrization.

...

1. the lake parametrization.

   
   

Known issues

1. ERA5-Land forecast parameters are missing for the validity time of 1st January 1950 00 UTC.

2. 
   

   Expand
   title Uncertainty fields
   As it was done for ERA5, the original plan for ERA5-Land was to provide an estimate of the uncertainty fields based on a dedicated 10-member ensemble run. The latter generated an ensemble of forcing fields that should, in principle, reproduce the space of uncertainty for the land surface fields. The first experiments demonstrated that the spread of the ensemble was clearly under dispersive, i.e. the uncertainty was unrealistically low. A reason for this is the low spread shown by the ensemble of ERA5 forcing fields. As a result of these experiments we took the decision of not providing the uncertainty fields of ERA5-Land. The opposite would have assigned, for instance, unrealistically high confidence to ERA5-land fields in a data assimilation experiment.  Our recommendation is, for the time being, to use the uncertainty estimate of the corresponding ERA5 field, which should provide a second order approximation to the estimate of the real uncertainty. Future experiments will also perturbe, among other, key land surface model parameters, therefore providing a more realistic spread of the ERA5-Land ensemble surface fields.

   
   

3. 
   

   Expand
   title Swapped value of the components of the total evapotranspiration
   Three components of the total evapotranspiration have values swapped as follows: - variable "Evaporation from bare soil" (mars parameter code 228101 (evabs)) has the values corresponding to the "Evaporation from vegetation transpiration" (mars parameter 228103 (evavt)), - variable "Evaporation from open water surfaces excluding oceans (mars parameter code 228102 (evaow)) has the values corresponding to the "Evaporation from bare soil" (mars parameter code 228101 (evabs)), - variable "Evaporation from vegetation transpiration" (mars parameter code 228103 (evavt)) has the values corresponding to the "Evaporation from open water surfaces excluding oceans" (mars parameter code 228102 (evaow)).

   
   

4. 
   

   From 1950-1978 ERA5-Land was forced by the preliminary ERA5 back extension which has a sub-optimal representation for a number of tropical cyclones.

   The over-estimation of a number of tropical cyclones for this period affects some products over the oceans in the vicinity of tropical cyclone tracks. Over land much smaller impact is expected, and therefore, the effect on the ERA5-Land product from 1950-1978 is more limited.

   This is supported by the figure below that plots, for each location, the minimum pressure from the ERA5 forcing (top panels) and the maximum daily accumulated total precipitation for ERA5-Land (lower panels) for the (preliminary) back extension (left panels) and  for the period from the late 1970s to 2010 inclusive (right panels). Note that these show the most extreme situations,  i.e., the absolute extremes in the about 30-year periods that were considered in each plot. Less extreme statistics, like 99, 95 (etc.) percentiles or mean distributions will show a much smaller impact of tropical cyclones.

   From these panels it is seen that for the forcing from ERA5 (top panels), in general, local minimum pressure is similar between 1950-1978 and 1979-2010. There are of course sampling differences between the two, each about 30-year, periods. Large differences that are likely related to  anomalously strong tropical cyclones are very localized, such as for some areas over North Australia, East Madagascar, Philippines and Northeast China. Note again that these affect a few cases only in the 29-year dataset.

   The effect on the ERA5-Land precipitation is shown in the lower panels. Even for these extremes it is difficult to pin-point locations that could be affected by anomalous tropical cyclones.

    

   Caption: locally minimum of 6-hourly pressure for ERA5 forcing data (top panels) and maximum daily total precipitation (lower panels) over the indicated period of time for the back extension (left panels) and for later periods (right panels). Numbers represent the averages for the locally extreme values over the indicated areas.

   Expand
   title Low values of snow cover and snow depth on the eastern side of the Antarctic ice sheet
   Low values of snow cover and snow depth were found on the eastern side of the Antarctic ice sheet, as shown in Fig. 1. The issue is due to the application of an old glacier mask to the Antarctica, which excludes the patch shown in the figure as glacier. Inaccuracies in the glacier mask are due to errors in satellite measurements datasets. While, due to the lower horizontal resolution, in ERA5 this ice sheet part is a sea point, in ERA5-Land the area is categorised as land without an initial ice mass. Since the initialization doesn't consider a glacier there (estimated at a constant 10 m of snow water equivalent), the low amount of precipitation along with potential excess of sublimation makes them to obtain unrealistic low numbers there. Image Removed  Fig 1: ERA-Land Snow depth (m of water equivalent) on 01-01-2015 eastern side of the Antarctic ice sheet.
   Expand
   title Limited impact from sub-optimal tropical cyclones in the forcing from the ERA5 preliminary dataset for 1950-1978.
   Image Removed	Image Removed
   Image Removed	Image Removed

   Expand
   title Definition of Potential Evaporation (PEV) modified
   Note that on 18-11-2021 we modified the definition of Potential Evaporation (PEV) provided in the CDS catalogue entry for both, the hourly and the monthly fields. The reason is that until this date the definition of PEV was similar to that of the same field provided by ERA5, whereas in reality they are computed differently:    PEV in ERA5 is computed for agricultural land as if it is well watered (no soil moisture stress)  and assuming that the atmosphere is not affected by this artificial surface condition.   PEV in ERA5-Land is computed as an open water evaporation (Pan evaporation) and assuming that the atmosphere is not affected by this artificial surface condition.

   Expand

   title	Incorrect values of monthly averaged reanalysis of accumulated fields from September 2022 to February 2024

   An issue with the ERA5-Land monthly averaged data from 1950 to present dataset was identified in January 2024 on both CDS discs and MARS tapes. It affects the following part of the dataset:

   Product: Monthly averaged reanalysis

   Time period: from September 2022 to February 2024 

   Area: Global

   Variables: all accumulated variables (please refer to Table 3 in ERA5-Land documentation)

   Incorrect data for period September 2022 to February 2024 is no longer available to download using the interactive CDS Download data form over the web, but remains accessible for the time being via CDS API or direct MARS access. 

   Incorrect data from September 2022 to February 2024 will not be replaced on CDS discs nor MARS tapes. Instead users can recover the same (and correct) data using the product monthly averaged reanalysis by hour of day at 00:00.

   A fix was implemented for ERA5-Land monthly averaged data stored on CDS discs in time for the release of March 2024 monthly averaged data (i.e. from March 2024 onwards).

   low numbers there. Image Added  Fig 1: ERA-Land Snow depth (m of water equivalent) on 01-01-2015 eastern side of the Antarctic ice sheet.

   
   

5. 
   

   Expand
   title Limited impact from sub-optimal tropical cyclones in the forcing from the ERA5 preliminary dataset for 1950-1978.
   From 1950-1978 ERA5-Land was forced by the preliminary ERA5 back extension which has a sub-optimal representation for a number of tropical cyclones. The over-estimation of a number of tropical cyclones for this period affects some products over the oceans in the vicinity of tropical cyclone tracks. Over land much smaller impact is expected, and therefore, the effect on the ERA5-Land product from 1950-1978 is more limited. This is supported by the figure below that plots, for each location, the minimum pressure from the ERA5 forcing (top panels) and the maximum daily accumulated total precipitation for ERA5-Land (lower panels) for the (preliminary) back extension (left panels) and  for the period from the late 1970s to 2010 inclusive (right panels). Note that these show the most extreme situations,  i.e., the absolute extremes in the about 30-year periods that were considered in each plot. Less extreme statistics, like 99, 95 (etc.) percentiles or mean distributions will show a much smaller impact of tropical cyclones. From these panels it is seen that for the forcing from ERA5 (top panels), in general, local minimum pressure is similar between 1950-1978 and 1979-2010. There are of course sampling differences between the two, each about 30-year, periods. Large differences that are likely related to  anomalously strong tropical cyclones are very localized, such as for some areas over North Australia, East Madagascar, Philippines and Northeast China. Note again that these affect a few cases only in the 29-year dataset. The effect on the ERA5-Land precipitation is shown in the lower panels. Even for these extremes it is difficult to pin-point locations that could be affected by anomalous tropical cyclones.   Image Added Image Added Image Added Image Added Caption: locally minimum of 6-hourly pressure for ERA5 forcing data (top panels) and maximum daily total precipitation (lower panels) over the indicated period of time for the back extension (left panels) and for later periods (right panels). Numbers represent the averages for the locally extreme values over the indicated areas.

   
   

6. 
   

   Expand
   title Definition of Potential Evaporation (PEV) modified
   Note that on 18-11-2021 we modified the definition of Potential Evaporation (PEV) provided in the CDS catalogue entry for both, the hourly and the monthly fields. The reason is that until this date the definition of PEV was similar to that of the same field provided by ERA5, whereas in reality they are computed differently:    PEV in ERA5 is computed for agricultural land as if it is well watered (no soil moisture stress)  and assuming that the atmosphere is not affected by this artificial surface condition.   PEV in ERA5-Land is computed as an open water evaporation (Pan evaporation) and assuming that the atmosphere is not affected by this artificial surface condition.

   
   

7. Expand
   title Incorrect values of monthly averaged reanalysis of accumulated fields from September 2022 to February 2024
   An issue with the ERA5-Land monthly averaged data from 1950 to present dataset was identified in January 2024 on both CDS discs and MARS tapes. It affects the following part of the dataset: Product: Monthly averaged reanalysis Time period: from September 2022 to February 2024  Area: Global Variables: all accumulated variables (please refer to Table 3 in ERA5-Land documentation) Incorrect data for period September 2022 to February 2024 is no longer available to download using the interactive CDS Download data form over the web, but remains accessible for the time being via CDS API or direct MARS access.  Incorrect data from September 2022 to February 2024 will not be replaced on CDS discs nor MARS tapes. Instead users can recover the same (and correct) data using the product monthly averaged reanalysis by hour of day at 00:00. A fix was implemented for ERA5-Land monthly averaged data stored on CDS discs in time for the release of March 2024 monthly averaged data (i.e. from March 2024 onwards).

Resolved issues

1. ERA5-Land-T rewritten 1st - 6th January 2025: At the turn of 2025, an unfortunate technical issue with the production of ERA5 resulted in a large part of satellite observations not being used. This led to small but systematic differences, impacting mainly humidity in the tropics, and in lesser extent other ERA5 parameters as well from the 1st of January 2025 onwards. The issue was resolved and ERA5T and ERA5-Land-T were rewritten with the correct data from 12 UTC 1^st January to 12 UTC 6th January 2025. The consolidated ERA5 and ERA5-Land will contain these corrected values.

How to cite the ERA5-Land dataset

...

• cite the CDS catalogue entry;
• provide clear and visible attribution to the Copernicus programme and attribute each data product used;

Step 1: Check the licence to use Copernicus Products for attribution/reference clause.

...