Estimating the depth of shower maximum $\mathrm{X_{max}}$ with high precision
is of great interest for the study of primary particle composition. One of the systematic
uncertainties in reconstructing $\mathrm{X_{max}}$ from the radio emission of air
showers is the limited knowledge of the atmospheric parameters like humidity,
pressure, temperature and the index-of refraction. Using the Global Data Assimilation System (GDAS), a global atmospheric model,
we have implemented time-dependent realistic atmospheric profiles in the air shower simulation codes CORSIKA and the radio plug-in CoREAS.
This program is now available within CORSIKA and flexible to be adapted for different air shower experiments.
We have analyzed the LOFAR cosmic ray data with dedicated simulations for each detected air shower with event specific GDAS atmospheres
and investigated the effects of pressure and humidity on the reconstructed $\mathrm{X_{max}}$.
This study shows that for bulk of the events, where the ground pressure is close to US standard atmosphere values,
there is a small systematic shift in $\mathrm{X_{max}}$ that is less than 2 $\mathrm{g/cm^{2}}$ and for very low
pressure values the shift is up to 15 $\mathrm{g/cm^{2}}$