Contact: Alek Petty / [email protected] / www.alekpetty.com

Code repository for producing sea ice thickness estimates from ICESat-2 freeboard data. The model code is written in the open source programming language Python (Python Software Foundation, https://www.python.org/). The main inputs needed for the thickness calculation are ATL10 sea ice freeboards from ICESat-2 and snow depth/density data (e.g. from the NESOSIM model). We also need to make a few other assumptions (e.g. ice density) which are discussed in the code.

v1.0: This initial repository was used to processes ICESat-2 (and also ICESat) thickness estimates for the first winter season of data collection (October 14th 2018 to April 30th 2019).

This code was originally developed on a linux server and has not been tested on other operating systems (I'm hoping to get round to that soon). If you are running this on linux and you are familiar with using conda to manage your Python environment you should be able to just import the conda environment included in this repository as:

conda env create -n py36envX -f environment.yml

then simply activate this environment to work in (to avoid using you own Python environment which wil likely have dependency issues)

source activate py36envX

In an effort to provide an environment file that is less build dependent and might work on any os, I also include a more basic environment file (just specifies the main libraries needed and not their versions etc). This can be used as:

conda env create -n py36envB -f environment_basic.yml

then simply activate this environment to work in (to avoid using you own Python environment which wil likely have dependency issues)

source activate py36envB

Alternatively, a requirements.txt file is also included for installing with pip, although this has not been tested extensively yet.

The /Code/ folder contains the primary scripts to process ATL10 freeboards (and also original ICESat freeboards) into sea ice thickness.

The batch_process_icesat2.py and batch_process_icesat.py scripts link to modules contained in common_functions.py that read in freeboard data, assign this to a Pandas dataframe, read in other relevant datasets and finally estimate sea ice thickness and its uncertainties. THe data are saved as netcdf files using xarray to enable the use of Dask/xarray to read in and analyze these large data files in parralel.

Also included are some plotting scripts: /Code/Plotting/ which can be used to generate most of the figures included in Petty et al., (2020).

The basic flow chart of this sea ice processing approach is shown in the flowchart below and is explained in more detail in Petty et al., (2020).

We use the ICESat-2 ATL10 sea ice freeboard product (designated Release 002) which is disseminated through the NSIDC (https://nsidc.org/data/atl10). The six beams are comprised of three beam pairs, with each beam pair containing a strong and weak beam which are separated by 90 m across-track and 2.5 km along-track, with each beam pair then separated by around 3.3 km across-track. Individual segment heights are produced from each beam using 150-photon aggregates, in an effort to produce heights with a precision of 2 cm or less over flat surfaces, as described in the ATL07 sea ice/sea surface height product description and ATBD . This results in segment lengths of around 10 m to 200 m for the strong beam (mean of around 15 m), and around 40 m to 800 m (mean of around 60 m) for the weak beam. We primarily utilize strong beam 1 for our processing but have carried out beam consistency checks to show this has only a small impact on our basin-scale distributions. In future releases we expect to provide thickness estimates using all of the beams.

We primarily make use of snow depth and density data from the NASA Eulerian Snow On Sea Ice Model (NESOSIM) v1.0; a new open-source snow budget model that is currently configured to simulate snow on sea ice across the Arctic Ocean through the accumulation season (Petty et al., 2018, https://github.com/akpetty/NESOSIM). For this time period we run NESOSIM using ERA-Interim (ERA-I) snowfall and winds (http://apps.ecmwf.int/datasets/data/interim_full_daily/), NASA Climate Data Record (CDR) sea ice concentrations (https://nsidc.org/data/g02202) and the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT) Ocean and Sea Ice Satellite Application Facility (OSI SAF, www.osi-saf.org) ice drifts (http://osisaf.met.no/p/ice/). ERA-I near-surface air temperatures are also used to set the initial (August) snow conditions.

We use the EUMETSAT OSI SAF sea ice type product to delineate our results between FYI and MYI. This is used mainly to derive modified Warren climatology snow depth information, but is also stored in the datafiles to enable post-processing. The data can be accessed at (http://www.osi-saf.org/?q=content/global-sea-ice-type-c)

We also utilize the NSIDC regional mask of the Arctic Ocean and its peripheral seas to delineate the results by Arctic region (region_n.msk in /AncData/ of this repo).

Data are in the process of being provided on the NSIDC. For now just the gridded data are available on the ICESat-2 project page: https://icesat-2.gsfc.nasa.gov/sea-ice-thickness-data. The plan is to provide this routinely for all future winter seasons (just for the months of October through to April when we have higher confidence of the snow loading and reduced likelihood of surface melt complexity).

A single NetCDF file is provided for each ATL10 granule covering the period October 14th 2018 to April 30th 2019. These datafiles include sea ice thickness and several other variables of expected interest to the user (snow depth/density, uncertainity etc).

As the data is provided in NetCDF format it can be easily read in with xarray or other python/netcdf libraries.

Gridded (25 km, NSIDC polar stereographic projection) monthly data are also provided. A single NetCDF file is provided for each month from November 2018 to April 2019.

An example monthly gridded sea ice thickness dataset is shown below.

Petty, A. A., N. T. Kurtz, R. Kwok, T. Markus, T. A. Neumann (2020), Winter Arctic sea ice thickness from ICESat‐2 freeboards, Journal of Geophysical Research: Oceans, 125, e2019JC015764. doi:10.1029/2019JC015764