ERA-20C reanalysis data was used to reconstruct the weather situation during a historical flash flood event in the Jona river basement (Canton of Zürich, Switzerland) back in 1939 (see report). Vertical temperature and moisture profiles are crucial in the search for relevant convective storm processes. Thus, for a periode between 1980 and 2010 radiosonde observations from Payerne, Switzerland were compared with reanalysis data. ERA-20C tends to overestimate relative humdity in the lower and upper troposphere. This yields to overestimation of precipitable water and other parameters in a thermodynamic profile. These findings have been taken into account within mentioned flash flood study.

## Data

The IGRA-2 radiosounding data set is provided by the NOAA. No further quality checks are applied to the data. Obvious errornous data records have been removed, though. Changes in sensors and techniques might affect the integrity of the data set. ERA-20C renalysis data is available from the ECMWF. Profiles from the reanalysis data set are extracted at 47°N / 7°E, coordinates which are slightly to the North and East of the sounding launch place (Payerne, Switzerland). Long term comparison have been performed on a subset by month (May through September only) and time (12Z observations only).

## Calculation of Relative Humidity

Relative humidity $$RH_w$$ is the ratio in per cent of the observed vapour pressure $$e$$ to the saturation vapour pressure $$e_s$$ with respect to water at the same temperature and pressure (WMO CIMO Guide 4.A.15): $RH_w = 100\left({ e \over e_s}\right)_{T, p}$

$$RH_w$$ is also related to the mixing ratio: $RH_w = 100 {r \over r_w} \cdot {0.62198 + r_w \over 0.62198 + r}$ where $$r_w$$ is the saturation mixing ratio at the pressure and temperature of the moist air (WMO CIMO Guide 4.A.16). Specific humidity $$q$$ can be approximated by the mixing ratio $$r$$ if $$r << 1$$:

$q = {r \over r + 1} \approx r$

Finally $$RH_w$$ is calculated from $$q$$ as follows:

$RH_w = 100 {q \over q_w} \cdot {0.62198 + q_w \over 0.62198 + q}$

where $$q_w$$ is

$q_w = 0.62198 {e_s \over p}$

Saturation vapour pressure $$e_s$$ is calculated with the formula from Goff-Gratch (1965).

## Results

• Meridional wind is underestimated by ERA-20C in the lower and middle troposphere (see figure 1)
• Relative humidity taken as-is and calculated from specific humidity, respectivly, differ significantly. Possibly, this is caused by a different calculation of the vapour pressure (over ice/water). See figure 2.
• Temperature tends to be underestimated by ERA-20C in middle and upper troposphere (see figure 3)
Figure 1: Aggregated atmospheric profiles from Payerne radiosoundings 1980 - 2010 at 1200Z and respective ERA-20C reanalysis data. Height (vertical axis in log-p scale) is plotted against observed values. Zonal and meridional wind components from reanalysis (blue) and radiosoundings (magenta). Linetype stands for mean (solid) and percentiles (dotted, dashed). Source: NOAA IGRA-2, ECMWF ERA-20C
Figure 2: Same as figure 1, but showing relative humitidy (as-is and derived from specific humidity
Figure 3: Same as figure 1, but showing temperature and dew point