[Description]

This dataset is curated for a manuscript published in Earth and Space Science by Hoyeon Shi and his colleagues in April 2024. 

Shi, H., Tonboe, R., Lee, M., Dybkjær, G., Sohn, J., Singha, S., & Baordo, F. (2024). A Simple and Robust CryoSat-2 Radar Freeboard Correction Method Dedicated to TFMRA50 for the Arctic Winter Snow Depth and Sea Ice Thickness Retrieval. Earth and Space Science, 11(10), e2024EA003715. https://doi.org/10.1029/2024EA003715

Here, version 2 is uploaded, corresponding to the revised manuscript during the revision. The main changes compared to version 1 are:
    1) Update of the CryoSat-2 radar freeboard dataset (from v2p4 to v2p6)
    2) Update of the coefficients for the radar freeboard correction equations
    3) Extension of the retrieval period for the CS2IS2 method (April is now included)
    4) Removal of OIB data points used for the regression from the validation datasets
    5) Inclusion of the Fram Strait mooring dataset in the validation dataset

It consists of three directories, each described below.

01_retrieval_results

This directory includes CryoSat-2-based monthly fields of Arctic sea ice thickness, snow depth, total freeboard, ice freeboards, ice draft, and bulk sea ice density for the winter months of the 2011-2022 period (January-March for alpha method and January-April for CS2IS2 method). Those variables are obtained using six combinations of two retrieval methods and three radar freeboard correction methods.

Retrieval methods

• alpha method: A simultaneous retrieval method based on Shi et al. (2020) and Shi et al. (2023), combining CryoSat-2, AVHRR, and AMSR data
• CS2IS2 method: A simultaneous retrieval method based on Kwok and Marcus (2018) and Kwok et al. (2020), combining CryoSat-2 and ICESat-2 data

Radar freeboard correction methods

• Wave speed correction method: Mallet et al. (2020)
• Empirical correction method: An empirical correction derived from the CS2_OIB_matchup data, using snow depth as a predictor
• Bias correction method: An empirical correction derived from the CS2_OIB_matchup data, doing bias correction

The datasets used for generating this dataset are as follows:

• CryoSat-2 
  - AWI CryoSat-2 sea ice thickness v2p6 (doi: 10.5281/zenodo.10044554)
• ICESat-2
  - NSIDC ATL20 dataset (doi: 10.5067/ATLAS/ATL20.004)
• AVHRR
  - Copernicus Marine Service's surface temperature datasets (doi: 10.48670/MOI-00130, doi: 10.48670/MOI-00123)
• AMSR
  - JAXA AMSR-E 6.9 GHz brightness temperature (doi: 10.57746/EO.01gs73ayng11rpwk7n54aynyj1)
  - JAXA AMSR2 6.9 GHz brightness temperature (doi: 10.57746/EO.01gs73b1nzeh3g66jr4p04mr0j)
• Auxiliary data
  - Sea ice concentration: OSI SAF (doi: 10.15770/EUM_SAF_OSI_0013, doi: 10.15770/EUM_SAF_OSI_0014)
  - Sea ice type: OSI SAF (doi: 10.15770/EUM_SAF_OSI_NRT_2006)

The naming convention is 'RetrievalMethod_CorrectionMethod_yyyymm.bin'. The 'RetrievalMethod' is either 'alpha' or 'CS2IS2', and the 'CorrectionMethod' is either 'WaveSpeed,' 'Empirical,' or 'BiasCorrection.' The data format is a 32-bit floating point array in the shape of 6 x 448 x 304 (25 km polar stereographic grid). The first dimension indicates the variables (in the order of snow depth (0), sea ice thickness (1), ice freeboard (2), total freeboard (3), sea ice draft (4), and bulk sea ice density (5)). For example, to read the sea ice thickness of January 2020 based on the alpha method with an empirical correction, you may write this Python command:

import numpy as np
data = np.fromfile('alpha_Empirical_202001.bin', dtype=np.float32).reshape(6,448,304)
hi = data[1,:,:]

The unit of thickness-related variable is cm, and the unit of density is kg/m3. The 25 km polar stereographic grid information is available on the NSIDC website (doi: 10.5067/N6INPBT8Y104).

02_valdiation data 

This directory includes reference data used for quality assessment of retrievals. There are three sub-directories:

'Mooring_draft_psn25_monthly' includes sea ice draft measurements from the moorings in the Beaufort Sea (https://www2.whoi.edu/site/beaufortgyre/data/mooring-data/), Fram Strait (doi: 10.21334/npolar.2022.b94cb848), and the Laptev Sea (doi: 10.1594/PANGAEA.912927, doi: 10.1594/PANGAEA.899275).

'OIB_SD_psn25_monthly' and 'OIB_TFB_psn25_monthly' include airborne snow depth and total freeboard measurements from NASA's Operation IceBridge campaign (doi: 10.5067/G519SHCKWQV6, doi: 10.5067/GRIXZ91DE0L9).

Original data were processed to become monthly gridded data to make a comparison with satellite retrievals. The OIB data points used for the regression were excluded when processing the monthly gridded data. The naming convention of each file is 'Var_yyyymm.bin,' where 'Var' is the variable name (SD: snow depth, TFB: total freeboard, Di: ice draft). For example, you can use the following code to read the OIB snow depth in March 2014.

import numpy as np
hs = np.fromfile('SD_201403.bin', dtype=np.float32).reshape(448,304)

03_CS2_OIB_matchup

This directory includes a match-up of AWI's CryoSat-2 L2P track data and OIB track data. The matching was done by resampling two high-resolution data on a coarser-resolution common grid (25 km polar stereographic grid) using a drop-in-a-bucket resampling method. The file format is CSV, and it is straightforward to understand when it is opened.

[Abbreviations]

AMSR: Advanced Microwave Scanning Radiometer
AVHRR: Advanced Very High Resolution Radiometer
AWI: Alfred Wegener Institute
CS2: CryoSat-2
JAXA: Japan Aerospace Exploration Agency
NASA: National Aeronautics and Space Administration
NSIDC: National Snow and Ice Data Center
OIB: Operation IceBridge
OSI SAF: Ocean and Sea Ice Satellite Application Facility