ERA5 reanalysis

ERA5 is the fifth-generation global atmospheric reanalysis produced by the European Centre for Medium-Range Weather Forecasts (ECMWF). It provides hourly estimates of atmospheric, land, and oceanic climate variables from 1940 to the present at approximately 31 km horizontal resolution on 137 pressure levels (Hersbach et al., 2020).

Key characteristics

• Spatial resolution: ~31 km (TL639 spectral truncation) on a reduced Gaussian grid
• Temporal resolution: Hourly
• Vertical levels: 137 hybrid sigma-pressure levels, from the surface to 0.01 hPa (~80 km)
• Period: 1940 to present (near-real-time updates within 5 days)
• Data assimilation: 4D-Var (12-hour windows) assimilating satellite radiances, radiosondes, aircraft, surface stations, buoys, and scatterometer winds
• Model: IFS Cycle 41r2

Use as forcing data

ERA5 is one of the most widely used datasets for providing lateral boundary conditions in dynamical downscaling with regional climate models like WRF. It serves as the large-scale atmospheric reference state, which the regional model then refines at higher resolution.

Key advantages over earlier reanalyses (ERA-Interim, NCEP/NCAR):

• Higher spatial and temporal resolution
• Improved data assimilation and model physics
• Uncertainty estimates through a 10-member ensemble (ERA5 EDA)
• Better representation of tropical variability and the diurnal cycle

Applications in my research

ERA5 was used as forcing data in multiple research projects:

1. DARWIN project: ERA5 drove the WRF downscaling that produced the meso-scale climate analysis of the Galapagos Archipelago (Schmidt et al., 2025, Int. J. Climatol.). The downscaling resolved orographic Precipitation and Garua dynamics that ERA5 alone cannot capture at its ~31 km grid spacing.

2. Qaidam Basin: ERA5-derived boundary conditions forced the WRF simulations used to study Water balance sensitivity under Mid-Pliocene conditions (Wang, Schmidt et al., 2021, JGR-Atmospheres).

3. CER v2: ERA5 provided the large-scale forcing for three decades of high-resolution Precipitation analysis over Berlin-Brandenburg.

Limitations

Despite its quality, ERA5 has known limitations:

• Precipitation is a model-generated variable (not directly assimilated) and can have systematic biases, especially in complex terrain and for convective events
• The ~31 km resolution cannot represent small island topography like the Galapagos, necessitating Dynamical downscaling
• Observation density varies over time and space, affecting reanalysis quality in data-sparse regions and earlier decades

See also: Dynamical downscaling, Regional climate modeling