ERA5 versus ERA5-Land: is an 80–100 m altitude difference expected?

Hello,

I am comparing ERA5 and ERA5-Land surface pressure along a transect to understand the effect of the different model orographies.

I downloaded:

  • ERA5 hourly single-level data;
  • ERA5-Land hourly data;
  • ERA5 geopotential;
  • ERA5-Land invariant geopotential.

The transect is approximately:

  • longitude: −5.5°;
  • latitude: 35.5–36.25°;
  • time: 15 July 2020, 12:00 UTC;
  • fixed logger pressure: 950 hPa.

I convert geopotential \(\Phi\) to surface elevation as:

\[ z_{\mathrm{surface}} = \frac{\Phi}{g}, \]

with \(g=9.80665\ \mathrm{m\,s^{-2}}\).

I then estimate altitude from the fixed pressure using:

\[ z_{\mathrm{logger}} = z_{\mathrm{surface}}+ \frac{T}{L} \left[ \left( \frac{P_{\mathrm{logger}}}{P_{\mathrm{surface}}} \right)^{-R_dL/g} -1 \right], \]

where \(T\) is 2 m temperature, \(L=0.0065\ \mathrm{K\,m^{-1}}\), and \(R_d=287.05\ \mathrm{J\,kg^{-1}\,K^{-1}}\).

Over land, I obtain:

  • maximum difference in geopotential-derived surface elevation: approximately 106 m;
  • maximum difference in inferred altitude for the same fixed pressure: approximately 80 m;
  • maximum difference between the ERA5–ERA5-Land pressure difference and the pressure difference expected from the elevation difference: approximately 8.9 hPa.

ERA5-Land is unavailable over the water section of the transect, which explains the missing values there. The comparison above is based only on locations where ERA5-Land provides data.

My understanding is that ERA5-Land pressure is obtained by interpolating ERA5 atmospheric fields to the finer grid and applying an orographic/lapse-rate correction. The documentation describes temperature, pressure, and humidity corrections for the elevation difference between the ERA5 and ERA5-Land grids:

ERA5-Land documentation

Could you clarify whether differences of approximately 80–100 m in the inferred altitude are expected and physically normal in this situation?

In particular:

  1. Should ERA5 and ERA5-Land surface pressure, temperature, and geopotential reproduce approximately the same altitude after applying the hydrostatic equation?
  2. Is the pressure correction in ERA5-Land expected to satisfy the simple hydrostatic equation using the published 2 m temperature and geopotential?
  3. Could the difference be caused by the use of a daily environmental lapse rate, rather than the instantaneous 2 m temperature?
  4. Is it valid to use the ERA5-Land invariant geopotential together with ERA5-Land surface pressure and 2 m temperature for this calculation?
  5. Are there additional corrections or differences between the model geopotential and the pressure-correction geopotential that should be considered?

I can provide the exact NetCDF extracts, the Python download/calculation script, and the resulting transect table if useful.