Introduction

Greenland surface runoff is repackaged from:

Paper: Mankoff (2020) http://doi.org/10.5194/essd-12-2811-2020.
Code (original): https://github.com/GEUS-Glaciology-and-Climate/freshwater
Data: http://doi.org/10.22008/promice/freshwater

Data example

Printout

<xarray.Dataset> Size: 203kB
Dimensions:                         (region: 7, time: 876)
Coordinates:
  * region                          (region) int32 28B 1 2 3 4 5 6 7
  * time                            (time) datetime64[ns] 7kB 1950-01-01 ... ...
Data variables:
    subglacial_discharge_from_land  (region, time) float64 49kB ...
    surface_runoff_from_land        (region, time) float64 49kB ...
    subglacial_discharge_from_ice   (region, time) float64 49kB ...
    surface_runoff_from_ice         (region, time) float64 49kB ...
    min_discharge_depth             (region) float32 28B ...
    region_name                     (region) <U2 56B ...
Attributes:
    title:        Ice sheet runoff by Mougniot region
    history:      TBD
    Conventions:  CF-1.8
    DOI:          https://doi.org/10.5281/zenodo.14020895

Minimum runoff depth per ROI

import xarray as xr
ds = xr.open_dataset('dat/GL_runoff.nc')
ds['min_discharge_depth'].to_dataframe()

region	min_discharge_depth
1	-661
2	-824
3	-1341
4	-1094
5	-1158
6	-643
7	-243

Plots

Annual runoff by type for all of Greenland

Example showing annual data for

Land runoff that enters the ocean subglacially (at depth)
Land runoff that enters the ocean at the surface
Ice runoff that enters the ocean subglacially
Ice runoff that enters the ocean at the surface

import xarray as xr
ds = xr.open_dataset('dat/GL_runoff.nc').load()
df = ds.sum(dim='region')\
       .drop_vars(['min_discharge_depth'])\
       .sel({'time':slice('1950-01-01','2021-01-01')})\
       .resample({'time':'YS'})\
       .sum()\
       .to_dataframe()

df = df * 86400 / 1E9 # convert from m^3/s -> ~ Gt/year
ax = df.plot(drawstyle='steps-post')
_ = ax.set_ylabel('Runoff [Gt$^{3}$ year^{-1}$]')

Region runoff for some years

import xarray as xr
ds = xr.open_dataset('dat/GL_runoff.nc')
df = ds.drop_vars(['min_discharge_depth'])\
       .sel({'time':slice('2017-01-01','2021-01-01')})\
       ['subglacial_discharge_from_ice']\
       .to_dataframe()['subglacial_discharge_from_ice']

df = df.unstack().T
df.columns = ds['region_name'].values
ax = df.plot(drawstyle='steps-post')
_ = ax.set_ylabel('Runoff [m$^{3}$ s$^{-1}$]')

Table: Annual ice-only runoff per region for Greenland

import xarray as xr
ds = xr.open_dataset('dat/GL_runoff.nc').load()
df = ds\
    .drop_vars(['min_discharge_depth'])\
    .sel({'time':slice('2000-01-01','2099-12-31')})\
    .resample({'time':'YS'})\
    .sum()\
    .to_dataframe()

df = df[['subglacial_discharge_from_ice','surface_runoff_from_ice']].sum(axis='columns')
df = df * 86400 / 1E9 # convert from m^3/s -> ~ Gt/year
df = df.unstack()
df.columns = [str(_[0]) + ' ['+_[1]+']' for _ in zip(ds['region'].values, ds['region_name'].values)]
df['GL'] = df.sum(axis='columns')
df.loc['Mean'] = df.mean(axis='rows')
df.index = [str(_)[0:4] for _ in df.index]
df.round()

	1 [CE]	2 [CW]	3 [NE]	4 [NO]	5 [NW]	6 [SE]	7 [SW]	GL
2000	48	37	36	33	39	75	128	395
2001	37	36	41	40	49	72	103	377
2002	71	40	68	46	42	74	105	446
2003	58	54	75	60	66	83	168	565
2004	70	49	48	31	44	81	127	449
2005	61	51	65	57	59	96	123	511
2006	53	43	36	30	54	77	118	412
2007	65	58	49	34	58	85	162	511
2008	68	38	80	67	66	85	112	516
2009	54	38	56	49	55	63	84	398
2010	85	72	58	46	74	128	204	668
2011	51	56	58	65	76	70	139	514
2012	74	86	88	82	100	111	254	794
2013	54	38	52	28	39	58	94	364
2014	64	51	53	47	63	89	135	502
2015	55	42	68	72	68	61	95	460
2016	71	56	81	63	64	94	159	588
2017	58	36	68	41	45	80	117	446
2018	50	40	34	33	37	66	103	363
2019	84	80	104	90	90	89	170	707
2020	59	41	77	75	54	83	138	526
2021	82	54	72	52	66	106	158	590
2022	53	45	45	49	58	92	145	488
Mean	62	50	61	52	59	83	137	504

df.describe().round()

	1 [CE]	2 [CW]	3 [NE]	4 [NO]	5 [NW]	6 [SE]	7 [SW]	GL
count	23	23	23	23	23	23	23	23
mean	62	50	61	52	59	83	137	504
std	12	14	18	18	16	17	39	111
min	37	36	34	28	37	58	84	363
25%	53	39	49	37	47	73	109	429
50%	59	45	58	49	58	83	128	502
75%	70	55	74	64	66	91	158	546
max	85	86	104	90	100	128	254	794

Fetch data

Warning: ~20 GB

mkdir -p tmp/greenland_runoff
pushd tmp/greenland_runoff
wget -r -e robots=off -nH --cut-dirs=3 --content-disposition "https://dataverse.geus.dk/api/datasets/:persistentId/dirindex?persistentId=doi:10.22008/FK2/XKQVL7"
popd

Combine land and ice runoff and bin by ROI

Create one dataset that includes

Land runoff (melted snow, and rain on land)
Ice runoff (melted snow, melted ice, and rain on the ice sheet that does not refreeze)

Note that land runoff can enter fjords via streams at their surface, or route subglacially and then mix with subglacial discharge at depth in fjords at the grounding line of marine-terminating glaciers. Ice runoff can also enter fjords at depth at the grounding line of marine-terminating glaciers, or at the surface from sub-aerial streams from land terminating glaciers. Metadata will be used to track each of these sources and destinations.

Runoff will be binned by Mouginot regions in Greenland. See Mouginot (2019) https://doi.org/10.7280/d1wt11. This metadata is already include in the product downloaded above.

import xarray as xr
import numpy as np

ds_land = xr.open_mfdataset('tmp/greenland_runoff/freshwater/land/MAR.nc', chunks='auto')
ds_ice = xr.open_mfdataset('tmp/greenland_runoff/freshwater/ice/MAR.nc', chunks='auto')

ds_land = ds_land.reset_coords('alt')
ds_ice = ds_ice.reset_coords('alt')

# Land runoff by destination (use alt metadata = BedMachine depth at outlet location)
land_sub = ds_land.where(ds_land['alt'] < 0).groupby('M2019_region').sum()['discharge'].resample({'time':'MS'}).sum()
land_surf = ds_land.where(ds_land['alt'] >= 0).groupby('M2019_region').sum()['discharge'].resample({'time':'MS'}).sum()

# Max depth of discharge within each region    
sub_depth = ds_land.where(ds_land['alt'] < 0).groupby('M2019_region').min()['alt']

# Ice runoff by destination
ice_sub = ds_ice.where(ds_ice['coast_alt'] < 0).groupby('M2019_region').sum()['discharge'].resample({'time':'MS'}).sum()
ice_surf = ds_ice.where(ds_ice['coast_alt'] >= 0).groupby('M2019_region').sum()['discharge'].resample({'time':'MS'}).sum()

ds = xr.Dataset({'subglacial_discharge_from_land':land_sub,
                 'surface_runoff_from_land':land_surf,
                 'subglacial_discharge_from_ice':ice_sub,
                 'surface_runoff_from_ice':ice_surf,
                 'min_discharge_depth':sub_depth})

# # scale from m^3/s -> km^3/sample_period
# scale = 86400 / 1E9
# # /1E9:  m^3 -> km^3
# # 86400: nuber of seconds in a day. Because sum() used above, works for month or year or any resample period
# scale_items = ['subglacial_discharge_from_land','surface_runoff_from_land',
#                'subglacial_discharge_from_ice','surface_runoff_from_ice']
# ds[scale_items] = ds[scale_items] * scale                 

print(ds)

<xarray.Dataset> Size: 203kB
Dimensions:                         (M2019_region: 7, time: 876)
Coordinates:
  * M2019_region                    (M2019_region) object 56B 'CE' 'CW' ... 'SW'
  * time                            (time) datetime64[ns] 7kB 1950-01-01 ... ...
Data variables:
    subglacial_discharge_from_land  (M2019_region, time) float64 49kB dask.array<chunksize=(7, 19), meta=np.ndarray>
    surface_runoff_from_land        (M2019_region, time) float64 49kB dask.array<chunksize=(7, 19), meta=np.ndarray>
    subglacial_discharge_from_ice   (M2019_region, time) float64 49kB dask.array<chunksize=(7, 46), meta=np.ndarray>
    surface_runoff_from_ice         (M2019_region, time) float64 49kB dask.array<chunksize=(7, 46), meta=np.ndarray>
    min_discharge_depth             (M2019_region) float32 28B dask.array<chunksize=(7,), meta=np.ndarray>

Write it out using Dask so it works on lightweight machines. This takes a few hours.

delayed_obj = ds.to_netcdf('tmp/GL_runoff.nc', compute=False)
from dask.diagnostics import ProgressBar
with ProgressBar():
    results = delayed_obj.compute()

[########################################] | 100% Completed | 110.26 s

Add metadata, extend back to 1840, etc.

ds = xr.open_dataset('tmp/GL_runoff.nc')

ds = ds.rename({'M2019_region':'region'})
ds['region_name'] = ds['region']
ds['region'] = np.arange(7).astype(np.int32) + 1

items = ['subglacial_discharge_from_land','surface_runoff_from_land',
         'subglacial_discharge_from_ice','surface_runoff_from_ice']
for i in items:
    ds[i].attrs['standard_name'] = 'water_volume_transport_into_sea_water_from_rivers'
    ds[i].attrs['units'] = 'm3 s-1'

ds['time'].attrs['long_name'] = 'time'
ds['region'].attrs['long_name'] = 'Mougniot (2019) region'

ds.attrs['title'] = 'Ice sheet runoff by Mougniot region'
ds.attrs['history'] = 'TBD'
ds.attrs['Conventions'] = 'CF-1.8'
ds.attrs['DOI'] = 'https://doi.org/10.5281/zenodo.14020895'
    
comp = dict(zlib=True, complevel=5)
encoding = {var: comp for var in items}
encoding['time'] = {'dtype': 'i4'}

!rm dat/GL_runoff.nc    
ds.to_netcdf('dat/GL_runoff.nc', encoding=encoding)
!ncdump -h dat/GL_runoff.nc

netcdf GL_runoff {
dimensions:
	region = 7 ;
	time = 876 ;
variables:
	int region(region) ;
		region:long_name = "Mougniot (2019) region" ;
	int time(time) ;
		time:long_name = "time" ;
		time:units = "days since 1950-01-01 00:00:00" ;
		time:calendar = "proleptic_gregorian" ;
	double subglacial_discharge_from_land(region, time) ;
		subglacial_discharge_from_land:_FillValue = NaN ;
		subglacial_discharge_from_land:long_name = "MAR discharge" ;
		subglacial_discharge_from_land:standard_name = "water_volume_transport_into_sea_water_from_rivers" ;
		subglacial_discharge_from_land:units = "m3 s-1" ;
	double surface_runoff_from_land(region, time) ;
		surface_runoff_from_land:_FillValue = NaN ;
		surface_runoff_from_land:long_name = "MAR discharge" ;
		surface_runoff_from_land:standard_name = "water_volume_transport_into_sea_water_from_rivers" ;
		surface_runoff_from_land:units = "m3 s-1" ;
	double subglacial_discharge_from_ice(region, time) ;
		subglacial_discharge_from_ice:_FillValue = NaN ;
		subglacial_discharge_from_ice:long_name = "MAR discharge" ;
		subglacial_discharge_from_ice:standard_name = "water_volume_transport_into_sea_water_from_rivers" ;
		subglacial_discharge_from_ice:units = "m3 s-1" ;
	double surface_runoff_from_ice(region, time) ;
		surface_runoff_from_ice:_FillValue = NaN ;
		surface_runoff_from_ice:long_name = "MAR discharge" ;
		surface_runoff_from_ice:standard_name = "water_volume_transport_into_sea_water_from_rivers" ;
		surface_runoff_from_ice:units = "m3 s-1" ;
	float min_discharge_depth(region) ;
		min_discharge_depth:_FillValue = NaNf ;
		min_discharge_depth:long_name = "height_above_mean_sea_level" ;
		min_discharge_depth:standard_name = "altitude" ;
		min_discharge_depth:units = "m" ;
		min_discharge_depth:positive = "up" ;
		min_discharge_depth:axis = "Z" ;
	string region_name(region) ;

// global attributes:
		:title = "Ice sheet runoff by Mougniot region" ;
		:history = "TBD" ;
		:Conventions = "CF-1.8" ;
		:DOI = "https://doi.org/10.5281/zenodo.14020895" ;
}

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

GL_runoff.org

GL_runoff.org

Table of contents

Introduction

Data example

Printout

Minimum runoff depth per ROI

Plots

Annual runoff by type for all of Greenland

Region runoff for some years

Table: Annual ice-only runoff per region for Greenland

Fetch data

Combine land and ice runoff and bin by ROI

Files

GL_runoff.org

Latest commit

History

GL_runoff.org

File metadata and controls

Table of contents

Introduction

Data example

Printout

Minimum runoff depth per ROI

Plots

Annual runoff by type for all of Greenland

Region runoff for some years

Table: Annual ice-only runoff per region for Greenland

Fetch data

Combine land and ice runoff and bin by ROI