Abstract: Objective Columnar water vapor (CWV) is an important parameter for characterizing the spatial and temporal distribution of atmospheric moisture, and its measurement plays a significant role in the analysis of the earth-atmosphere radiation balance, climate change studies, and satellite radiometric calibration and validation. Sun photometers have the advantages of measuring CWV over the entire atmospheric column, providing high temporal resolution, and enabling continuous daytime observations. However, the calibration coefficient of the 936 nm water-vapor absorption channel may drift with instrument aging and the changing of environmental conditions. To address the limitations of conventional calibration methods, which are time-consuming and difficult to coordinate with continuous observations, this study investigates the field calibration of the 936 nm water-vapor channel of a sun photometer to improve CWV retrieval accuracy and support related operational applications. Methods Based on the observation data from a CE318 sun photometer deployed at the Dunhuang radiometric calibration site, field calibration of the 936 nm water-vapor channel of the sun photometer was carried out using an improved Langley method. First, the principles of the single-channel (water-vapor channel) method and the dual-channel (water-vapor channel and non-water-vapor channel) method were introduced. Then, water-vapor transmittance parameters were calculated using LBLRTM v12.9 and the HITRAN 2016 spectroscopic database to obtain the corresponding model parameters. Subsequently, the observation data acquired under relatively clear-sky conditions on August 8 and 9, 2018, were selected for fitting to derive the calibration coefficient of the 936 nm channel, and the effects of atmospheric stability, water-vapor transmittance parameters, and estimation errors of aerosol optical depth at 936 nm on the calibration results were further analyzed. Finally, the field calibration results were applied to CWV retrieval and validated against radiosonde observations. Results and Discussion The results showed that under the selected clear-sky conditions, both the single-channel and dualchannel methods achieved good fitting performance, with correlation coefficients exceeding 0.99. The calibration coefficients derived from the two methods were in good agreement, with a relative difference of about 1%. Compared to the standard calibration coefficients, the deviations for the calibration coefficients obtained on August 8, 2018 were relatively small, whereas those on August 9 were noticeably larger, indicating that the accuracy of field calibration is sensitive to atmospheric stability during the observation period. Further analysis revealed that the main factors affecting the calibration results include atmospheric stability, the value of the water-vapor transmittance parameter b, and the accuracy of aerosol optical depth estimation at 936 nm. Among these factors, utilizing more realistic atmospheric profiles and more accurate aerosol parameters can help reduce systematic errors. When the calibration coefficients obtained on August 8, 2018 were applied to CWV retrieval, the retrieved results were highly consistent with radiosonde observations, indicating that the field calibration results are reliable under stable atmospheric conditions. Conclusions It is proved in this work that the field calibration method for the 936 nm water-vapor channel of the sun photometer based on the imporved Langley method is feasible. Both the single-channel and dual-channel methods can provide reliable calibration results and meet the basic accuracy requirements for CWV retrieval and related satellite applications. Howvever, the performance of field calibration is strongly influenced by atmospheric conditions, and clear and stable skies are essential for ensuring calibration accuracy. Future work should establish a more representative lookup table of water-vapor transmittance parameters based on seasonal radiosonde profiles over the Dunhuang region, and further optimize the calibration of non-water-vapor channels and aerosol interpolation methods to improve the stability of field calibration and the accuracy of CWV retrieval.

Key words: sun photometer, water-vapor channel calibration, improved Langley method, columnar water-vapor, aerosol optical thickness