Neural Network predictions and ERA5 reference of integrated water vapour, and temperature and specific humidity profiles based on simulated microwave radiometer observations

This data set contains predictions of the Neural Network retrievals described in [1], where simulated microwave radiometer observations (brightness temperatures, TBs) from the evaluation data subset of [2] (years 2001, 2006, 2011, 2015) were used as input to the Neural Network. As described in Section 3.2 of [1], we trained an ensemble of 20 Neural Networks for each retrieved atmospheric quantity and applied them to the ERA5 evaluation data set to estimate the robustness of the retrievals with respect to random perturbations. The following atmospheric quantities were retrieved: 

• temperature profile (variable name 'temp_p', filename suffix 'temp_test_417'),
• boundary layer temperature profile (variable name 'temp_p', filename suffix 'temp_test_424'),
• specific humidity profile (variable name 'q_p', filename suffix 'q_test_472'),
• integrated water vapour (variable name 'iwv_p', filename suffix 'iwv_test_126')

The cryptic 3-digit filename suffixes represent different settings of the Neural Network retrieval. More information can be found in [3]. Variables that do not have the "_p" suffix are ERA5 data and used as reference to estimate errors of the retrievals by comparing them with the predictions. The dimension 'n_s' represents the ERA5 data sample number while the dimension 'n_rand' designates the ensemble of Neural Networks.

These files can be created when running run_NN_retrieval (contained in NN_retrieval.py, see [3]) with exec_type='20_runs' and eval_mode=True and test_id either "126", "417", "424" or "472". However, as this might take some hours, we provide them here.

[1]: Walbröl, A., Griesche, H. J., Mech, M., Crewell, S., and Ebell, K.: Combining low- and high-frequency microwave radiometer measurements from the MOSAiC expedition for enhanced water vapour products, Atmospheric Measurement Techniques, 17, 6223-6245, https://doi.org/10.5194/amt-17-6223-2024, 2024.

[2]: Walbröl, A., and Mech, M.: ERA5 based training, validation and evaluation data for retrievals combining 22-58 GHz with 175-340 GHz microwave radiometer measurements during MOSAiC (1.0.0). Zenodo. https://doi.org/10.5281/zenodo.10997365, 2024.

[3]: Walbröl, A.: Codes for: Combining low and high frequency microwave radiometer measurements from the MOSAiC expedition for enhanced water vapour products (1.0.1). Zenodo. https://doi.org/10.5281/zenodo.11123136, 2024.