Gravity wave driven seasonal variability of temperature differences between ECMWF IFS and lidar measurements at 54S in the lee of the Southern Andes

In November 2017, the DLR Institute of Atmospheric Physics started running the ground-based Compact Rayleigh Autonomous Lidar (CORAL) at the southern tip of South America in Rio Grande that is located at the east coast of Argentina in the lee of the Andes. We used this independent (i.e., not assimilated in the ECMWF IFS) and high-resolution lidar data of the year 2018 and some individual months in 2019 and 2020 to investigate middle atmosphere temperature deviations in IFS analyses and short-term forecasts at higher mid-latitudes in the southern hemisphere (54 S).

 

We found a generally good agreement between IFS and CORAL temperature data below 45 km altitude and the calculated monthly mean temperature deviations are smaller than +/-2 K.  The temperature deviations are more variable in time and the sign of the monthly mean deviations varies throughout the year above 45 km altitude. There, the largest positive differences (+2 K), i.e. IFS temperatures were too warm, are found for May 2018. The largest negative differences (-10 K), i.e. IFS temperatures were too cold, are found for August 2018.  The standard deviation of the temperature differences is significantly larger (up to 15 K) and increases with altitude in the winter half year (April to September 2018) compared to the summer half year. The better agreement of IFS temperature with ground-based lidar measurements in the summer months previously reported in literature for the northern hemisphere also manifests for the southern hemisphere and more recent cycles of the IFS. The largest temperature differences above 45 km altitude in the winter half year are due to gravity waves (GWs) and it was found that amplitude and phase deviations are equally important at the location of Rio Grande. In general, the IFS underestimates GW potential energy density in the middle atmosphere, especially within the sponge layer. Monthly mean GW potential energy density at 45-60 km altitude gets up to four times larger when the sponge is removed but is still less than 50 % of the amount of GW potential energy density found in the CORAL data.