This data set contains daily files of thermal-infrared brightness temperature (TB) statistics for each frame and each video recorded with the FLIR A315 camera during Polarstern cruise PS149 (Nicolaus 2026). The videos were recorded with 1 frame/second between 6 July and 28 August 2025. The TB statistics files (PS149_flir_tb_statistics_.nc) include the mean, standard deviation, minimum and maximum TB of each frame (image), the TB in the centre of the image as well as the count of TBs within TB bins for the full image. For enhanced spatial context, daily mp4 videos of minutely FLIR images (PS149_flir_tb_.mp4) were compiled including the time stamp of the image in UTC and a colour bar indicating the range of the brightness temperature. Note that not all video players are able to play the mp4 files as we have chosen a lossless video codec (libx264rbg, crf 0). We recommend using the VLC media player to play the mp4 files.We additionally provide a file to correct biases of the FLIR recordings (PS149_flir_tb_offset_correction.nc) using in-situ surface temperature measurements and reference infrared target temperatures in the footprint of the FLIR camera on the sea ice. We assumed an emissivity of the snow/surface scattering layer of 0.985 (0.996) before (after) fresh snow had fallen onto the sea ice. The emissivity assumptions are based on Thielke et al. (2022), Nalliet al. (2023) and Cox et al. (2023). The reference IR targets are aluminium plates coated with Nextel Suede Coating 3101, 7329 S139 tiefschwarz with an emissivity of 0.98. The contribution of the reflected atmospheric contribution is generally low because of the high emissivities but not negligible. We assumed reflected atmospheric TBs of 273.15 K in overcast, 205 K in clear-sky and 239.075 K in broken-sky conditions. The clear-sky value is taken from the subarctic summer values given in Smith and Tuomi (2008). The file also contains an uncertainty estimate of the bias correction as well as the in-situ temperatures and FLIR TBs used to determine the bias.