Arctic albedo MODIS trend analysis

This is a code for the following paper

E Plekhanova, JS Kim, J Oehri, A Erb, C Schaaf, G Schaepman-Strub Mid-summer snow-free albedo across the Arctic tundra was mostly stable or increased over the past two decades (2022) Environmental Research Letters https://doi.org/10.1088/1748-9326/aca5a1

In this code, I preprocess and analyze trends across 2000-2022 of MODIS albedo data (MCD43A3). Then I prepare and analyze trends across 2000-2014 of the climate data from CMIP6 LSMs

prepare_MODIS_albedo_and_LSM_python

First, I prepare MCD43A3 broadband albedo with Albedo_MCD43A3_downoad_mosaic_crop and Quality_MCD43A2_downoad_mosaic_crop scripts. The structure of the scripts is the following

1. download MODIS raw albedo HDf files (or QA/snow-flag for each band) from EarthData server via pyModis package for 15th of July each year. For this step, you would need to sign-up to EarthData and insert your log-in info in the first chunk of code.
2. mosaic the HDF files for each year to a single TIFF mosaic and reproject it to the CAVM coordinate system on the 500m resolution
3. crop the resultant mosaic with the CAVM shapefile (above the Arctic treeline)

I also prepare the land surface models (LSMs) NetCDF files: reproject them, convert them to TIFF and crop them via LSM_NetCDF_to_tiff_reproject_and_crop script. The original LSMs were downloaded for 15th of July for years 2000-2014 of "land-hist" experiment. They are openly available at: https://esgf-node.llnl.gov/projects/cmip6/

r_analysis

• Filtering

  Excluding snow based on MCD43A2 snow flag and MOD10A2 snow product. Then filtering low quality pixels based on MCD43A2 quality flags.

• Calculation of slopes

  We calculated the slope with Theil–Sen estimator (Sen, 1968) and determined significance with Kendall’s tau statistic (p-value < 0.05). Slope and p-values were calculated for MODIS 2000-2022, MODIS 2000-2014 and land surface models from CMIP6.

• Figures

  Constructing figures for main text and supplementary