Abstract. Possible uncertainties of lidar measurements of middle-atmospheric temperatures, measured with the novel airborne Rayleigh lidar system ALIMA, are investigated on the basis of data from the SouthTRAC-GW campaign in September 2019 and corresponding simulations of photon counts of the ALIMA system. We evaluate uncertainties due to the attenuation by Rayleigh extinction and ozone absorption, (signal-induced) photon noise, the photon background, and the nonlinearity of photon counting detectors. Ozone absorption induces an altitude-dependent cold bias in the retrieved temperatures of 2 K between 25 km to 55 km. Rayleigh extinction introduces a similar uncertainty of 2 K below 25 km that can be decreased by a suitable correction. Photon noise can introduce uncertainties of ±25 K at high altitudes (above 70 km) for high temporal resolutions (1 min), but on average the photon noise influences the temperature by only 1 K to 2 K at 70 km and decreases downwards. Uncertainties related to the photon background and the nonlinearity of the detectors, with a dead time correction applied, play a minor role in the temperature uncertainty. The analysis of the photon background in the ALIMA measurements of six research flights of the SouthTRAC-GW campaign proves the assumption of a constant photon background with altitude as well as the Poisson distribution of the photon counts. The airborne operation of ALIMA is advantageous as the high flight altitudes reduce the Rayleigh extinction by up to 17 % and thus result in higher signal levels compared to a ground-based operation. Overall, our analysis reveals that temperatures can be retrieved from ALIMA measurements with a remaining uncertainty of ≤ 1 K if all known biases are corrected.