Verification of parameterizations for clear sky downwelling longwave irradiance in the Arctic

• Main Authors: Pace, Giandomenico, Sarra, Alcide, Calì Quaglia, Filippo, Ciardini, Virginia, Iorio, Tatiana, Iaccarino, Antonio, Meloni, Daniela, Muscari, Giovanni, Scarchilli, Claudio
• Format: Text
• Language: English
• Published: 2023
• Subjects: Arctic; Greenland; Ice Sheet; Thule
• Online Access: https://doi.org/10.5194/amt-2023-181; https://amt.copernicus.org/preprints/amt-2023-181/

Ground-based high resolution observations of downward longwave irradiance (DLI), surface air temperature, water vapour surface partial pressure and column amount, zenith sky infrared (IR) radiance in the atmospheric window, and all-sky camera images are regularly obtained at the Thule High Arctic Atmospheric Observatory (THAAO, 76.5° N, 68.8° W), North-West Greenland. The datasets for the years 2017 and 2018 have been used to assess the performance of different empirical formulas to infer clear sky DLI. An algorithm to identify clear sky observations has been developed, based on value, variability, and persistence of zenith sky IR radiance. Seventeen different formulas to estimate DLI have been tested against the THAAO dataset, using the originally determined coefficients. The formulas which combine information on total column water vapour and surface air temperature appear to perform better than others, with a mean bias with respect to the measured DLI smaller than 1 W/m 2 and a root mean squared error (RMSE) around 6 W/m 2 . Some formulas, specifically developed for the Arctic, are found to produce poor statistical results; this is attributed partly to limitations in the originally used dataset, which does not cover a whole year, or is relative to very specific conditions (i.e., the ice sheet). The bias displays a significant improvement when the coefficients of the different formulas are calculated using the THAAO dataset. The presence of two full years of data allows the investigation of the inter-annual variability, and the use of different years for the determination of the coefficients and the evaluation of results. The smallest values of the bias and RMSE reach 0.1 W/m 2 and 5 W/m 2 , respectively. Overall, best results are found for formulas which use both surface parameters and total water vapour column, and have been developed from global datasets. Conversely, formulas which express the atmospheric emissivity as a linear function of the logarithm of the column integrated water vapour appear to poorly ...