In recent years, there has been a rapid advance in the field of radio detection of air showers induced by high-energy cosmic rays. Estimating the depth of shower maximum Xmax with improved accuracy is of great interest for the study of primary particle composition. One of the major systematic uncertainties in the Xmax-measurement arises from variations of the refractive index in the atmosphere. The refractive index varies with temperature, humidity and pressure, and the variations can be on the order of 10% for (n-1) at a given altitude. The effect of a varying refractive index on Xmax-measurements is evaluated using COREAS: a microscopic simulation of the radio emission from the individual particles in the cascade simulated with CORSIKA. We discuss the resulting offsets in Xmax for different frequency regimes, and compare them to a simple physical model. In typical circumstances, the offsets in Xmax range from 4 to 11 g/cm^2 for the 30 to 80 MHz frequency band. Therefore, for precision work it is required to include atmospheric data from the time and place of the air shower into the simulations. The aim is to implement this in the next version of CoREAS/CORSIKA using the Global Data Assimilation System(GDAS), a global atmospheric model based on meteorological measurements and numerical weather predictions. This can then be used to re-evaluate the air shower measurements done with the LOFAR radio telescope.