Effect of snow on soil erosion - snowpack dynamics

Introduction

The overall goal of the project is to extend the Revised Universal Soil Loss Equation RUSLE (Renard, 1997) by including the effects of snowfall and snowmelt on soil erosion. The first step is to distinguish between precipitation in the form of rain (erosive) or snow (non-erosive). Additionally, snow-covered areas are detected at the end of each month using freely available satellite data (e.g. Sentinel-2). These two steps precede the presented code and have already been automated.

The aim of the presented code is to calculate monthly snowpack dynamics and to identify monthly snowmelt to later incorporate its erosive forces. The results are written as georeferenced raster files (.tif format) and can be accessed by all common GIS programs. Optionally, the user can plot various statistical values (e.g. snow cover) over time. If only a subarea is of interest, this can be defined by a shapefile. The previously calculated snowfall (mm) and the identified snow cover serve as input data. The following figure provides a highly simplified overview. Details can be found in the code diagram and the code description.

Requirements

Libraries

Python libraries: geo_utils, gdal, matplotlib.pyplot, numpy, pandas, rasterstats, scipy

The geo_utils package can be downloaded here and/or the path where the package is saved can be defined in the config.py file.

Standard libraries: glob, logging, os, sys, time

Input Data

The following rasters have to be provided to run the code:

1. Raster of measured snow depth
2. Raster of satellite data of the snow cover with the values 0 and 1, describing if there is snow (1) or if there is not (0)
3. Shapefile defining the zone for the statistical calculations

Both rasters listed in 1. and 2. have to be provided as .tif rasters and have to include the month and the year in the following way:

*yy_mm.tif

Code diagram

Code description

config.py

File where all required libraries and global variables are loaded. The following input arguments can be defined by the user:

Input argument	Type	Description
snow_path_mm	STRING	Path of measured snow depth rasters
snow_cover_path	STRING	Path of rasters of satellite data describing the snow cover
path_results	STRING	Path of the results folders
path_geo_utils	STRING	Path where geo_utils package is stored
shape_zone	STRING	Path of shapefile used for zonal statistics
statistical_param	STRING	Statistical parameter which is/are plotted
plot_statistics	BOOL	Dis- or enable plotting
plot_result	STRING	Path of folder for the plotting result

log.py

File where the logger and its streamhandler are set up and the basic logging configurations can be changed. The following three functions are defined to record when an arbitrary function is entered and exited.

wrapper()

Wrapper function to define which functions are called before and after another function.

Input argument	Type	Description
pre	STRING	Function which is called before the wrapped function
post	STRING	Function which is callled after the wrapped function

entering()

Logs at what time a function is entered.

exiting()

Logs at what time a function is exited.

data_management.py

File where DataManagement class is stored. This class contains useful and robust methods for managing output folders, extracting information from raster files and writing raster files using the previously extracted information. The class provides the following methods:

__init__()

Assigns values to class attributes when a new instance is initiated.

Input argument	Type	Description
path	STRING	Path needed to manage output folders
filename	STRING	Filename

folder_creation()

Method creates folder at instantiated path, if it does not already exists.

return: None

get_date()

Method gets the month and the year from the instantiated filenames.

return: INTEGERs sm_year,sm_month

create_date_string()

Method which returns a date string by calling the get_date method.

return: STRING datestring

get_proj_data()

Method which get the projection and geotransformation from a raster file (osgeo.gdal.Dataset)

return: TUPLE gt, STRING proj

save_raster()

Static Method which creates and saves raster-file (.tif) from an existing array using a defined projection and geotransformation data.

Input argument	Type	Description
res_path	STRING	Path and result filename
array	ARRAY	Values to rasterize
gt	TUPLE	gdal.DataSet.GetGeoTransform object
proj	STRING	gdal.DataSet.GetProjection object

return: None

array_calculations.py

File where ArrayCalculations class is stored. This class is for mathematical operations on arrays to calculate snow depths. It provides the following methods:

__init__()

Assigns values to class attributes when a new instance is initiated.

Input argument	Type	Description
snow_start_of_period	ARRAY	Snow depth at start of period
snow_cover	ARRAY	Satellite data of snow cover
snow_measured	ARRAY	Measured snow depth

snow_at_end()

Calculates snow depth at the end of the period by multiplying the snow depth at the start of the period with the satellite data of the snow cover.

return: ARRAY snow_end_of_period

snowmelt()

Calculates snow depth acting as snowmelt by substracting the actual snow depth at the end of the period from the calculated snowsum at the beginning of the period.

Input argument	Type	Description
snow_end_of_period	ARRAY	Actual snow depth at end of a period

return: ARRAY snowmelt

snow_at_start()

Calculates snow depth at the beginning of a period by summing up the actual snow depth transferred from the previous month and the measured snow depth.

Input argument	Type	Description
snow_end_of_period	ARRAY	Actual snow depth at end of a period

return: ARRAY snow_start_of_period

compare_data.py

File where CompareData class is stored. This class is for comparing different properties of arrays, objects and rasters. It provides the following methods:

__init__()

Assigns values to class attributes when new instance is initiated.

Input argument	Type	Description
array_one	ARRAY	Arbitrary array to be compared
array_two	ARRAY	Abitrary array to be compared
raster_one_path	STRING	Path of arbitrary raster to be compared
raster_two_path	STRING	Path of arbitrary raster to be compared

compare_shape()

Compares the number of columns and the number of rows of two arrays.

number_of_items()

Static method which compared the number of items of two objects.

Input argument	Type	Description
object_one	LIST, STRING, etc.	Object on which the len() function can be applied
object_two	LIST, STRING, etc.	Object on which the len() function can be applied

compare_geotransform()

Rounds geotransformation values of two rasters off to four decimal places and compares them.

compare_projection()

Compares the projection of two rasters.

zon_statistics.py

File where ZonStatistics class is stored. This Class contains useful methods to calculate and plot zonal statistics, including a custom statistic for calculating the (snow) cover of a user-defined zone. It provides following methods:

__init__()

Assigns values to class attributes when new instance is initiated.

Input argument	Type	Description
path_raster	LIST	Paths to raster files
shape	STRING	Path to shapefile defining the zone for calculating the statistics
datelist	LIST	Datestrings, format yy_mm
parameter	STRING	Definition of statistical parameter to be plotted

get_zon_statistics()

Method which creates a dataframe containing necessary information to plot the results. User defines the statistical value which should be plotted in config.py (here: Coverage).

return: DATAFRAME df_statistics

calc_zon_statistics()

Method which calculates several statistical values.

Input argument	Type	Description
list_entry	INTEGER	Identification of the raster to calculate the statistical values

return: LIST stats

plot_zon_statistics()

Method which plots the desired statistical value over time and writes a .png-image. Method can be disabled in config.py (plot_statistic=False).

coverage()

Custom statistic to calculate the percentage of values above zero (e.g. snow coverage, or areas of snow melt).

Input argument	Type	Description
raster_array	MASKED ARRAY	Contains raster array

return: INTEGER calc_coverage

fun.py

The following basic functions and functions wrapping up different class methods are stored in this file:

snowdepth()

Calculates the snow depth acting as snowmelt, the snow depth at the end of a period and the snow depth at the beginning of the next period by instanciating an object of the RasterCalculations class.

Input argument	Type	Description
snow_at_start	ARRAY	Snow depth at beginning of the period
measured_snow_next_period	ARRAY	Measured snow depth at the beginning of the next period
satellite_data	ARRAY	Satellite data of the snow cover

return: ARRAYs snow_at_end_array, snowmelt_array, snow_start_array

check_data()

Compares the size of two arrays, the geotransformation and projections of two rasters and the number of items in two objects by instanciating an object of the CompareData class.

Input argument	Type	Description
array_one	ARRAY	Arbitrary array to be compared
array_two	ARRAY	Arbitrary array to be compared
raster_one_path	STRING	Path of arbitrary raster to be compared
raster_two_path	STRING	Path of arbitrary raster to be compared
object_one	LIST, STRING, etc.	Object on which the len() function can be applied
object_two	LIST, STRING, etc.	Object on which the len() function can be applied

compare_date()

Extracts the dates from the filenames of two rasters and compares them by instantiating an object of the DataManagement class.

Input argument	Type	Description
filename_one	STR	Filename of raster one
filename_two	STR	Filename of raster two

create_lists()

Function which creates three empty lists.

return: LISTs list1, list2, list3

append2list()

Appends objects to lists.

Input argument	Type	Description
list1	LIST	Arbitrary list
list2	LIST	Arbitrary list
list3	LIST	Arbitrary list
object1	ARRAY, INT, etc.	Object which is appendable to a list
object2	ARRAY, INT, etc.	Object which is appendable to a list
object3	ARRAY, INT, etc.	Object which is appendable to a list

return: LISTs list1, list2, list3

get_path_from_list()

Function receives 2 list objects and returns two file paths.

Input argument	Type	Description
list1_object	STRING	Indexed string object from list
list2_object	STRING	Indexed string object from list

return: STRINGs path1, path2

main.py

File where the most relevant functions are stored, including the main() function.

raster2list()

Reads an arbitrary number of raster files from 2 different folders, extracts the date and the corresponding raster arrays and saves them into lists for further calculations.

Input argument	Type	Description
list_rasterpaths1	LIST	Paths to raster files
list_rasterpaths2	LIST	Paths to raster files

return: LISTs date_list, array_list1, array_list2

snowcalc_over_list()

Calculates snow depths for different time periods by iterating over arrays stored in lists.

Input argument	Type	Description
initial_snow	ARRAY	Snow depth at start of first time period
satellite_data	LIST	Arrays of snow cover data
measured_snow_next_period	LIST	Arrays of measured snow depths

return: LISTs snow_end_month, snow_melt