# README - Code and Data Repository
# Tedstone and Machguth, Increasing surface runoff from Greenland's firn areas
# Nature Climate Change, 2022

In a hurry?
The most useful datasets for potential re-use are as follows:

* Shapefiles of annual and reference period visible runoff limits, +/- 1 MAD. See runoff_limit_polys/. 
* Coordinates of the annual maximum runoff limit every 1 km for the whole study area: rlim_annual_maxm/xytpd*.csv. One file for each of median, -1MAD, +1MAD, maximum likely runoff limit.

For panels of the main Figures which are underlain by tabular data, please see their associated 'source data' files available through Nature Climate Change's online version of the manuscript.


Full description of repository:

> bg_gis_data: contains various 'background' GIS-type data which supports analysis in the manuscript.
    - 100k_boxes: shapefiles containing (1) the limits of each 100 km zone, (2) for each zone the 100 member slices.
    - gris_only_outline: shapefile of the Greenland Ice Sheet-only outline.
    - impossible_rlim_zones: manually delimited 'impossible' detection zones, plus these zones conformed to the ice sheet margin using gris_only_outline.
    - greenland_region_bbox: basic shapefile of the region surrounding Greenland. Only used for clipping NaturalEarth outlines.
    - manual_region_divides: shapefile with division of ice sheet into the regions used in the manuscript.

> code:
    - river_detection: contains all scripts used for image processing and for setting up a PostgreSQL/PostGIS database containing all the runoff limits in all_filtered_csv. This database is used for most of the analysis.
    - analysis_plotting: contains all scripts for data reduction, analysis and plotting.
    - Both folders contain .yml files describing the conda environment in which they were run.

> gee_manual_rlims: 
    - shapefiles of manual runoff limit delineations for the two reference periods in each of the NE and NW areas (see also manuscript Figure 1).

> all_filtered_csv: all the valid runoff limit retrievals in CSV files, organised by quadrant of the ice sheet and then into jobs with <5% cloud cover and 5-50% cloud cover. 

> rlim_annual_maxm:
    - 100km_box_medians.csv: annual time series of visible runoff limit elevation in each 100 km zone/box.
    - xytpd*.csv: annual picked coordinates of visible runoff limit in each 1 km slice (later used to generate polygons).

> runoff_limit_polys:
    - Annual polygons for -1MAD, median, +1MAD provided as shapefiles
    - Polygons of the two reference periods (1985-1992, 2013-2020)
    - RACMO and MAR rasterized masks of annual runoff areas and reference periods.

> seasonality: 
    - region_*_binned_freqs.csv: Counts of data in each 20 m elevation band, corresponding to the data underlying the right hand panels of Extended Data Fig. 9.
    - region_polynomial_d2_coefs.csv: Coefficients estimated for the 2-degree polynomial of each region.
    - annual_emax_region_poly2.csv: Estimated maximum visible runoff limit elevation by slice, annually.

> excess_melt:
    - netcdf files of excess melt generated from RACMO and MAR outputs.

> pixel_counts:
    - master_*_count_matchgimp_1km.tif: counts every 1 km of observation availiability, annually.

> job_ctrl: contains the control/configuration text files used to run the main `river_detection` workflow and to merge each of the four quadrants into master maps. Explanations of the parameters used in these files can be found in `river_detection/README.md`.