Improving knowledge of the ozone global distributions in the mesosphere–lower thermosphere (MLT) is a crucial step in understanding the behaviour of the middle atmosphere. However, the concentration of ozone under sunlit conditions in the MLT is often so low that its measurement requires instruments with very high sensitivity. Fortunately, the bright oxygen airglow can serve as a proxy to retrieve the daytime ozone density indirectly, due to the strong connection to ozone photolysis in the Hartley band. The OSIRIS IR imager (hereafter, IRI), one of the instruments on the Odin satellite, routinely measures the oxygen infrared atmospheric band (IRA band) at 1.27 µm. In this paper, we will primarily focus on the detailed description of the steps done for retrieving the calibrated IRA band limb radiance (with <10 % random error), the volume emission rate of O₂ (a¹Δ_g) (with <25 % random error) and finally the ozone number density (with <20 % random error). This retrieval technique is applied to a 1-year sample from the IRI dataset. The resulting product is a new ozone dataset with very tight along-track sampling distance (<20 km). The feasibility of the retrieval technique is demonstrated by a comparison of coincident ozone measurements from other instruments aboard the same spacecraft, as well as zonal mean and monthly average comparisons between Odin-OSIRIS (both spectrograph and IRI), Odin-SMR and Envisat-MIPAS. We find that IRI appears to have a positive bias of up to 25 % below 75 km, and up to 50 % in some regions above. We attribute these differences to uncertainty in the IRI calibration as well as uncertainties in the photochemical constants. However, the IRI ozone dataset is consistent with the compared dataset in terms of the overall atmospheric distribution of ozone between 50 and 100 km. If the origin of the bias can be identified before processing the entire dataset, this will be corrected and noted in the dataset description. The retrieval technique described in this paper can be further applied to all the measurements made throughout the 19 year mission, leading to a new, long-term high-resolution ozone dataset in the middle atmosphere.

中文翻译：

---

使用了O的OSIRIS观测白天mesospheric臭氧的检索₂（一个¹ Δ_克）发射

增进对中层低层热层（MLT）中臭氧全球分布的了解是了解中层大气行为的关键一步。但是，MLT中阳光照射下的臭氧浓度通常很低，以至于其测量需要灵敏度很高的仪器。幸运的是，由于与Hartley波段中的臭氧光解作用密切相关，因此明亮的氧气气辉可以间接替代白天的臭氧密度。奥丁卫星上的仪器之一OSIRIS红外成像仪（以下称IRI）通常在1.27 µm处测量氧红外大气波段（IRA波段）  。在本文中，我们将主要侧重于获取校准的IRA波段肢体辐射（<10  ％随机误差），的体积排放率Ô ₂ （一个¹ Δ_克）（具有<25  ％的随机误差）和最后的臭氧数量密度（具有<20  ％的随机误差）。该检索技术适用于IRI数据集中的1年样本。生成的产品是一个新的臭氧数据集，沿轨采样距离非常短（<20  km）。通过比较同一航天器上其他仪器的同时进行的臭氧测量，以及Odin-OSIRIS（光谱仪和IRI），Odin-SMR和Envisat-MIPAS的区域平均值和月平均值比较，证明了这种检索技术的可行性。 。我们发现IRI在75公里以下似乎有25％的正偏差，而在以上某些地区则有50％的正偏差 。我们将这些差异归因于IRI校准的不确定性以及光化学常数的不确定性。但是，IRI臭氧数据集在50至100 km之间的臭氧总体大气分布方面与比较的数据集一致 。如果可以在处理整个数据集之前识别出偏差的来源，则将对此进行更正并在数据集说明中注明。本文所述的检索技术可以进一步应用于整个19年任务中进行的所有测量，从而在中层大气中产生一个新的长期高分辨率臭氧数据集。