Advanced features

Land cover evolution

The land cover types in hydrobricks are defined by the user (see the hydro units section). Each hydro unit is thus internally defined by a total area and fractional land covers. These land covers can have a dynamic evolution. This evolution can be externally driven or internally computed.

Hydrobricks offers the following options:

1. Evolution set through csv file

2. Evolution computed from shapefiles

3. Evolution computed from shapefiles and delta-h method

4. Evolution computed from ice thickness and delta-h method

The definition of a land cover evolution does not replace the original definition of the hydro units, which need to be also provided to the function. The areas provided in the definition of the hydro units are the starting point of the model, and these changes will be enforced in due time. However, if some changes are defined for dates prior to the start of the modelling period, these changes will also be applied.

The two last options are specific to glacier land covers. They do not handle debris covers on glaciers.

Evolution set through csv file

One can provide the model with a timeseries of dates and new land cover areas, such as:

changes = actions.ActionLandCoverChange()
changes.load_from_csv(
    '/path/to/surface_changes_glacier_debris.csv',
    hydro_units, area_unit='km2', match_with='elevation'
)
model.add_action(changes)

The function changes.load_from_csv() can be called multiple times for different files. The corresponding csv file must contain the name of the land cover to change on the first row (for example here glacier_debris), the dates of these changes on the second row, and then the change time series. These changes list all hydro units that need to change; those that do not need to change should not be listed in the file. There are two ways to identify the hydro units: by elevation (match_with='elevation') or by ID (match_with='id'). In the following example, these changes start with the unit elevation and contain the time series of the area (here in km2) for every date given above.

Example of a csv file containing a land cover evolution.

bands,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris,glacier_debris
,01/08/2020,01/08/2025,01/08/2030,01/08/2035,01/08/2040,01/08/2045,01/08/2050,01/08/2055,01/08/2060,01/08/2065,01/08/2070,01/08/2075,01/08/2080,01/08/2085,01/08/2090,01/08/2095,01/08/2100
4274,0.013,0.003,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
4310,0.019,0.009,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
4346,0.052,0.042,0.032,0.022,0.012,0.002,0,0,0,0,0,0,0,0,0,0,0
4382,0.072,0.062,0.052,0.042,0.032,0.022,0.012,0.002,0,0,0,0,0,0,0,0,0
4418,0.129,0.119,0.109,0.099,0.089,0.079,0.069,0.059,0.049,0.039,0.029,0.019,0.009,0,0,0,0
4454,0.252,0.242,0.232,0.222,0.212,0.202,0.192,0.182,0.172,0.162,0.152,0.142,0.132,0.122,0.112,0.102,0.092
4490,0.288,0.278,0.268,0.258,0.248,0.238,0.228,0.218,0.208,0.198,0.188,0.178,0.168,0.158,0.148,0.138,0.128
4526,0.341,0.331,0.321,0.311,0.301,0.291,0.281,0.271,0.261,0.251,0.241,0.231,0.221,0.211,0.201,0.191,0.181
4562,0.613,0.603,0.593,0.583,0.573,0.563,0.553,0.543,0.533,0.523,0.513,0.503,0.493,0.483,0.473,0.463,0.453
4598,0.648,0.638,0.628,0.618,0.608,0.598,0.588,0.578,0.568,0.558,0.548,0.538,0.528,0.518,0.508,0.498,0.488
4634,0.618,0.608,0.598,0.588,0.578,0.568,0.558,0.548,0.538,0.528,0.518,0.508,0.498,0.488,0.478,0.468,0.458
4670,0.478,0.468,0.458,0.448,0.438,0.428,0.418,0.408,0.398,0.388,0.378,0.368,0.358,0.348,0.338,0.328,0.318
4706,0.306,0.296,0.286,0.276,0.266,0.256,0.246,0.236,0.226,0.216,0.206,0.196,0.186,0.176,0.166,0.156,0.146
4742,0.338,0.328,0.318,0.308,0.298,0.288,0.278,0.268,0.258,0.248,0.238,0.228,0.218,0.208,0.198,0.188,0.178
4778,0.199,0.189,0.179,0.169,0.159,0.149,0.139,0.129,0.119,0.109,0.099,0.089,0.079,0.069,0.059,0.049,0.039
4814,0.105,0.095,0.085,0.075,0.065,0.055,0.045,0.035,0.025,0.015,0.005,0,0,0,0,0,0
4850,0.051,0.041,0.031,0.021,0.011,0.001,0,0,0,0,0,0,0,0,0,0,0
4886,0.019,0.009,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
4922,0.008,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
4958,0.003,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0

There is no need to specify the corresponding changes in the generic ground land cover as it will be automatically computed to preserve the total hydro unit area.

Evolution computed from shapefiles

One can provide the model with a timeseries of dates and shapefiles, such as:

times = ['2008-01-01', '2010-01-01', '2016-01-01']
ice_glaciers = ['/path/to/Glacier_ice_2008.shp',
                '/path/to/Glacier_ice_2010.shp',
                '/path/to/Glacier_ice_2016.shp']
debris_glaciers = ['/path/to/Glacier_debris_2008.shp',
                   '/path/to/Glacier_debris_2010.shp',
                   '/path/to/Glacier_debris_2016.shp']
changes, changes_df = actions.ActionLandCoverChange.create_action_for_glaciers(
    study_area, times, ice_glaciers, debris_glaciers,
    with_debris=True, method='raster', interpolate_yearly=True)
model.add_action(changes)

This method also creates a dataframe that can then be exported as csv files, and reloaded in if needed using the first option:

changes_df[0].to_csv('/path/to/surface_changes_glacier_ice.csv', index=False)
changes_df[1].to_csv('/path/to/surface_changes_glacier_debris.csv', index=False)
changes_df[2].to_csv('/path/to/surface_changes_ground.csv', index=False)

And the hydrological units can also separately be initialized using the following lines:

hyd_units.initialize_from_land_cover_change('glacier_ice', changes_df[0])
hyd_units.initialize_from_land_cover_change('glacier_debris', changes_df[1])

Tips and tricks

If information about land cover evolution is only available for a date after the beginning of the simulation period, it is possible to assume a constant land cover by duplicating the first data and assigning it the simulation begining date. This evolution, is of course, debatable…

For example:

times = ['2005-01-01', '2008-01-01', '2010-01-01', '2016-01-01']
ice_glaciers = ['/path/to/Glacier_ice_2008.shp',
                '/path/to/Glacier_ice_2008.shp',
                '/path/to/Glacier_ice_2010.shp',
                '/path/to/Glacier_ice_2016.shp']
debris_glaciers = ['/path/to/Glacier_debris_2008.shp',
                   '/path/to/Glacier_debris_2008.shp',
                   '/path/to/Glacier_debris_2010.shp',
                   '/path/to/Glacier_debris_2016.shp']

Evolution computed from shapefiles and delta-h method

The delta-h method from Huss et al. (2010), implemented by Seibert et al. (2018) is also available in Hydrobricks. A contrario to the two first methods, in the delta-h approach the glacial evolution is not forced from the outside but decided by the modeled melt of the glacier. Hydrobricks compute the amount the glacier melted in the year, and retrieves the corresponding glacier area from the lookup table. This makes this method and the following most appropriate for future discharge modeling or past discharge data when no glacier extent timeseries are available, whereas the two first methods are most appropriate when glacier timeseries of glacier extents are available.

We recommend 10 glacier elevation bands per HRU elevation band.

elev_distance = 40 # 40 m HRU elevation band
study_area = catchment.Catchment(outline='path/to/watershed/shapefile.shp')
glacier_evolution = preprocessing.GlacierEvolutionDeltaH()
glacier_df = glacier_evolution.compute_initial_ice_thickness(
     study_area, ice_thickness = glacier_thickness,
     elevation_bands_distance = elev_distance / 10)
glacier_evolution.compute_lookup_table(update_width=False)

The glacier lookup table glacier_evolution can then be linked to Hydrobricks. At the beginning of October, the hydrological model will sum up all the glacier mass loss that occurred during the hydrological year and will modify the land cover according to the areas stored in the glacier lookup table:

changes = actions.ActionGlacierEvolutionDeltaH()
changes.load_from(glacier_evolution, land_cover='glacier',
                  update_month='October')

This method also creates a dataframe that can then be exported as csv files, and reloaded in if needed using the first option:

glacier_df.to_csv('/path/to/surface_changes_glacier.csv', index=False)

The glacier lookup table can be saved as a csv file:

glacier_evolution.save_as_csv('/path/to/results/folder/')

Evolution computed from ice thickness and delta-h method

References

• Seibert, J., Vis, M., Kohn, I., Weiler, M., & Stahl, K. (2018). Technical note: Representing glacier geometry changes in a semi-distributed hydrological model. Hydrology and Earth System Sciences.

• Huss, M., Jouvet, G., Farinotti, D., & Bauder, A. (2010). Future high-mountain hydrology: A new parameterization of glacier retreat. Hydrology and Earth System Sciences.

Note: Options and compatibility with radiation/aspect discretization

Glacier thickness-related options

The glacier evolution methods require appropriate configuration of the following two options:

glacier_infinite_storage: Boolean flag indicating whether glaciers have unlimited thickness (i.e., no thinning due to melt).

snow_ice_transformation: Rate at which snow transforms into ice, expressed in mm/day. The default value is 0.002 mm/day.

These options must be set depending on the method used for glacier evolution:

For methods 1 and 2 (CSV input or shapefile input), the computation only takes into account the area, and not the ice thickness. As such:

glacier_infinite_storage = True
snow_ice_transformation = False

For methods 3 and 4 (shapefiles with delta-h or ice thickness with delta-h), the computation relies on the ice thickness to compute the area. As such:

glacier_infinite_storage = False
snow_ice_transformation = True

They are specified during model initialization:

socont = models.Socont(...,
                       glacier_infinite_storage = glacier_infinite_storage,
                       snow_ice_transformation = snow_ice_transformation)

Snow redistribution option

Hydrobricks supports a snow redistribution mechanism based on the SnowSlide algorithm (Bernhardt & Schulz, 2010). This option simulates gravitational snow transport and can help improve snow distribution modeling across elevation bands and avoid ‘snow towers’.

Default example:

socont = models.Socont(soil_storage_nb = 2,
                       snow_redistribution = 'transport:snow_slide')

In addition, you must provide a connectivity CSV file describing the lateral redistribution pathways between hydro units:

hydro_units.set_connectivity(/path/to/connectivity.csv)

Resources:

Working example implementation

Script to compute the connectivity CSV

References

• Bernhardt, M., & Schulz, K. (2010). SnowSlide: A simple routine for calculating gravitational snow transport. Geophysical Research Letters.