def ITE(Dir_Basin, Name_NC_ET, Name_NC_LAI, Name_NC_P, Name_NC_RD, Name_NC_NDM,
Name_NC_LU, Startdate, Enddate, Simulation):
"""
This functions split the evapotranspiration into interception, transpiration, and evaporation.
Parameters
----------
Dir_Basin : str
Path to all the output data of the Basin
Name_NC_ET : str
Path to the .nc file containing ET data
Name_NC_LAI : str
Path to the .nc file containing LAI data
Name_NC_P : str
Path to the .nc file containing P data
Name_NC_RD : str
Path to the .nc file containing Rainy Days data
Name_NC_NDM : str
Path to the .nc file containing Normalized Dry Matter data
Name_NC_LU : str
Path to the .nc file containing Landuse data
Startdate : str
Contains the start date of the model 'yyyy-mm-dd'
Enddate : str
Contains the end date of the model 'yyyy-mm-dd'
Simulation : int
Defines the simulation
Returns
-------
I : array
Array[time, lat, lon] contains the interception data
T : array
Array[time, lat, lon] contains the transpiration data
E : array
Array[time, lat, lon] contains the evaporation data
"""
# import WA modules
import wa.General.raster_conversions as RC
import wa.Functions.Start.Get_Dictionaries as GD
# Define monthly dates
Dates = pd.date_range(Startdate, Enddate, freq="MS")
# Extract LU data from NetCDF file
LU = RC.Open_nc_array(Name_NC_LU, Var='LU')
# Create a mask to ignore non relevant pixels.
lulc_dict = GD.get_lulcs().keys()
mask = np.logical_or.reduce([LU == value for value in lulc_dict[:-1]])
mask3d = mask * np.ones(len(Dates))[:, None, None]
mask3d_neg = (mask3d - 1) * 9999
# Extract Evapotranspiration data from NetCDF file
ET = RC.Open_nc_array(Name_NC_ET, Var='ET')[-len(Dates):, :, :]
# Extract Leaf Area Index data from NetCDF file
LAI = RC.Open_nc_array(Name_NC_LAI, Var='LAI')[-len(Dates):, :, :]
# Extract Precipitation data from NetCDF file
P = RC.Open_nc_array(Name_NC_P, Var='Prec')[-len(Dates):, :, :]
# Extract Rainy Days data from NetCDF file
RD = RC.Open_nc_array(Name_NC_RD, Var='RD')[-len(Dates):, :, :]
# Extract Normalized Dry Matter data and time from NetCDF file
NDM = RC.Open_nc_array(Name_NC_NDM, Var='NDM')[-len(Dates):, :, :]
timeNDM = RC.Open_nc_array(Name_NC_NDM, Var='time')
# Create dictory to get every month and year for each timestep
datesNDMmonth = dict()
datesNDMyear = dict()
# Loop over all timestep
for i in range(0, len(timeNDM)):
# change toordinal to month and year
datesNDMmonth[i] = datetime.date.fromordinal(timeNDM[i]).month
datesNDMyear[i] = datetime.date.fromordinal(timeNDM[i]).year
# Calculate the max monthly NDM over the whole period
NDMmax = dict()
# loop over the months
for month in range(1, 13):
dimensions = []
# Define which dimension must be opened to get the same month
for dimension, monthdict in datesNDMmonth.items():
if monthdict == month:
dimensions = np.append(dimensions, dimension)
# Open those time dimension
NDMmonth = np.zeros([
np.size(dimensions),
int(np.shape(NDM)[1]),
int(np.shape(NDM)[2])
])
dimensions = np.int_(dimensions)
NDMmonth[:, :, :] = NDM[dimensions, :, :]
# Calculate the maximum over the month
NDMmax[month] = np.nanmax(NDMmonth, 0)
NDMmax_months = np.zeros(
[len(timeNDM),
int(np.shape(NDM)[1]),
int(np.shape(NDM)[2])])
# Create 3D array with NDMmax
for i in range(0, len(timeNDM)):
NDMmax_months[i, :, :] = np.nanmax(NDMmax[datesNDMmonth[i]])
# Create some variables needed to plot graphs.
et = np.array([])
i = np.array([])
t = np.array([])
# Change zero values in RD so we do not get errors
RD[RD == 0] = 0.001
LAI[LAI == 0] = 0.001
LAI[np.isnan(LAI)] = 0.1
# Calculate I
I = LAI * (1 -
np.power(1 + (P / RD) * (1 - np.exp(-0.5 * LAI)) *
(1 / LAI), -1)) * RD
# Set boundary
I[np.isnan(LAI)] = np.nan
# Calculate T
T = np.minimum(
(NDM / NDMmax_months), np.ones(np.shape(NDM))) * 0.95 * (ET - I)
# Mask Data
ET = ET * mask3d
T = T * mask3d
I = I * mask3d
ET[mask3d_neg < -1] = np.nan
T[mask3d_neg < -1] = np.nan
I[mask3d_neg < -1] = np.nan
# Calculate E
E = ET - T - I
# Calculate monthly averages
et = np.nanmean(ET.reshape(ET.shape[0], -1), 1)
i = np.nanmean(I.reshape(I.shape[0], -1), 1)
t = np.nanmean(T.reshape(T.shape[0], -1), 1)
# Plot graph of ET and E, T and I fractions.
fig = plt.figure(figsize=(10, 10))
plt.grid(b=True, which='Major', color='0.65', linestyle='--', zorder=0)
ax = fig.add_subplot(111)
ax.plot(Dates, et, color='k')
ax.patch.set_visible(False)
ax.set_title('Average ET and E, T and I fractions')
ax.set_ylabel('ET [mm/month]')
ax.patch.set_visible(True)
ax.fill_between(Dates, et, color='#a3db76', label='Evapotranspiration')
ax.fill_between(Dates, i + t, color='#6bb8cc', label='Transpiration')
ax.fill_between(Dates, i, color='#497e7c', label='Interception')
ax.scatter(Dates, et, color='k')
ax.legend(loc='upper left', fancybox=True, shadow=True)
fig.autofmt_xdate()
ax.set_xlim([Dates[0], Dates[-1]])
ax.set_ylim([0, max(et) * 1.2])
ax.set_xlabel('Time')
[r.set_zorder(10) for r in ax.spines.itervalues()]
# Define output folder and name for image
NamePic = "Sim%s_Mean_ET_E_T_I.jpg" % Simulation
Dir_Basin_Image = os.path.join(Dir_Basin, "Simulations", "Images")
if not os.path.exists(Dir_Basin_Image):
os.mkdir(Dir_Basin_Image)
# Save Images
plt.savefig(os.path.join(Dir_Basin_Image, NamePic))
return (I, T, E)
def ITE(Dir_Basin, nc_outname, Startdate, Enddate, Simulation):
"""
This functions split the evapotranspiration into interception, transpiration, and evaporation.
Parameters
----------
Dir_Basin : str
Path to all the output data of the Basin
nc_outname : str
Path to the .nc file containing all data
Startdate : str
Contains the start date of the model 'yyyy-mm-dd'
Enddate : str
Contains the end date of the model 'yyyy-mm-dd'
Simulation : int
Defines the simulation
Returns
-------
I : array
Array[time, lat, lon] contains the interception data
T : array
Array[time, lat, lon] contains the transpiration data
E : array
Array[time, lat, lon] contains the evaporation data
"""
# import WA modules
import wa.General.raster_conversions as RC
import wa.Functions.Start.Get_Dictionaries as GD
# Define monthly dates
Dates = pd.date_range(Startdate, Enddate, freq = "MS")
# Extract LU data from NetCDF file
LU = RC.Open_nc_array(nc_outname, Var = 'Landuse')
# Create a mask to ignore non relevant pixels.
lulc_dict = GD.get_lulcs().keys()
mask=np.logical_or.reduce([LU == value for value in lulc_dict[:-1]])
mask3d = mask * np.ones(len(Dates))[:,None,None]
mask3d_neg = (mask3d-1) * 9999
# Extract Evapotranspiration data from NetCDF file
ET = RC.Open_nc_array(nc_outname, 'Actual_Evapotranspiration', Startdate, Enddate)
# Extract Leaf Area Index data from NetCDF file
LAI = RC.Open_nc_array(nc_outname, 'LAI', Startdate, Enddate)
# Extract Precipitation data from NetCDF file
P = RC.Open_nc_array(nc_outname, 'Precipitation', Startdate, Enddate)
# Extract Rainy Days data from NetCDF file
RD = RC.Open_nc_array(nc_outname, 'Rainy_Days', Startdate, Enddate)
# Extract Normalized Dry Matter data and time from NetCDF file
NDM = RC.Open_nc_array(nc_outname, 'Normalized_Dry_Matter', Startdate, Enddate)
timeNDM = RC.Open_nc_array(nc_outname, 'time')
# Create dictory to get every month and year for each timestep
datesNDMmonth = dict()
datesNDMyear = dict()
# Loop over all timestep
for i in range(0,len(timeNDM)):
# change toordinal to month and year
datesNDMmonth[i] = datetime.date.fromordinal(timeNDM[i]).month
datesNDMyear[i] = datetime.date.fromordinal(timeNDM[i]).year
# Calculate the max monthly NDM over the whole period
NDMmax = dict()
# loop over the months
for month in range(1,13):
dimensions = []
# Define which dimension must be opened to get the same month
for dimension, monthdict in datesNDMmonth.items():
if monthdict == month:
dimensions = np.append(dimensions,dimension)
# Open those time dimension
NDMmonth = np.zeros([np.size(dimensions), int(np.shape(NDM)[1]), int(np.shape(NDM)[2])])
dimensions = np.int_(dimensions)
NDMmonth[:,:,:] = NDM[dimensions, :,:]
# Calculate the maximum over the month
NDMmax[month] = np.nanmax(NDMmonth,0)
NDMmax_months = np.zeros([len(timeNDM), int(np.shape(NDM)[1]), int(np.shape(NDM)[2])])
# Create 3D array with NDMmax
for i in range(0,len(timeNDM)):
NDMmax_months[i,:,:] = np.nanmax(NDMmax[datesNDMmonth[i]])
# Create some variables needed to plot graphs.
et = np.array([])
i = np.array([])
t = np.array([])
# Change zero values in RD so we do not get errors
RD[RD==0] = 0.001
LAI[LAI==0] = 0.001
LAI[np.isnan(LAI)] = 0.1
# Calculate I
I = LAI * (1 - np.power(1 + (P/RD) * (1 - np.exp(-0.5 * LAI)) * (1/LAI),-1)) * RD
# Set boundary
I[np.isnan(LAI)] = np.nan
# Calculate T
T = np.minimum((NDM/NDMmax_months),np.ones(np.shape(NDM))) * 0.95 * (ET - I)
# Mask Data
ET = ET * mask3d
T = T * mask3d
I = I * mask3d
ET[mask3d_neg<-1] = np.nan
T[mask3d_neg<-1] = np.nan
I[mask3d_neg<-1] = np.nan
# Calculate E
E = ET - T - I
# Calculate monthly averages
et = np.nanmean(ET.reshape(ET.shape[0], -1),1)
i = np.nanmean(I.reshape(I.shape[0], -1),1)
t = np.nanmean(T.reshape(T.shape[0], -1),1)
# Plot graph of ET and E, T and I fractions.
fig = plt.figure(figsize = (10,10))
plt.grid(b=True, which='Major', color='0.65',linestyle='--', zorder = 0)
ax = fig.add_subplot(111)
ax.plot(Dates, et, color = 'k')
ax.patch.set_visible(False)
ax.set_title('Average ET and E, T and I fractions')
ax.set_ylabel('ET [mm/month]')
ax.patch.set_visible(True)
ax.fill_between(Dates, et, color = '#a3db76', label = 'Evapotranspiration')
ax.fill_between(Dates, i + t , color = '#6bb8cc', label = 'Transpiration')
ax.fill_between(Dates, i , color = '#497e7c', label = 'Interception')
ax.scatter(Dates, et, color = 'k')
ax.legend(loc = 'upper left',fancybox=True, shadow=True)
fig.autofmt_xdate()
ax.set_xlim([Dates[0], Dates[-1]])
ax.set_ylim([0, max(et) * 1.2])
ax.set_xlabel('Time')
[r.set_zorder(10) for r in ax.spines.itervalues()]
# Define output folder and name for image
NamePic = "Sim%s_Mean_ET_E_T_I.jpg" %Simulation
Dir_Basin_Image = os.path.join(Dir_Basin, "Simulations", "Simulation_%d" % Simulation, "Images")
if not os.path.exists(Dir_Basin_Image):
os.mkdir(Dir_Basin_Image)
# Save Images
plt.savefig(os.path.join(Dir_Basin_Image,NamePic))
return(I, T, E)
def Create(Dir_Basin, Simulation, Basin, Startdate, Enddate, Name_NC_LU,
DataCube_I, DataCube_T, DataCube_E, Example_dataset):
"""
This functions create the CSV files for the sheets
Parameters
----------
Dir_Basin : str
Path to all the output data of the Basin
Simulation : int
Defines the simulation
Basin : str
Name of the basin
Startdate : str
Contains the start date of the model 'yyyy-mm-dd'
Enddate : str
Contains the end date of the model 'yyyy-mm-dd'
Name_NC_LU : str
Path to the .nc file containing the LU data
DataCube_I : array
3D array [time, lat, lon] containing the interception data
DataCube_T : array
3D array [time, lat, lon] containing the transpiration data
DataCube_E : array
3D array [time, lat, lon] containing the evaporation data
Example_dataset : str
Data path to the example tiff file containing the right amount of pixels and projection
Returns
-------
Data_Path_CSV : str
Data path pointing to the CSV output files
"""
# import WA modules
import wa.Functions.Start.Get_Dictionaries as GD
import wa.General.raster_conversions as RC
from wa.Functions import Start
# Create output folder for CSV files
Data_Path_CSV = os.path.join(Dir_Basin, "Simulations",
"Simulation_%d" % Simulation, "Sheet_2",
"CSV")
if not os.path.exists(Data_Path_CSV):
os.mkdir(Data_Path_CSV)
# Open LULC map
LULC = RC.Open_nc_array(Name_NC_LU, 'LU')
# Set the months
Dates = pd.date_range(Startdate, Enddate, freq="MS")
# Define whole years
YearsStart = pd.date_range(Startdate, Enddate, freq="AS")
YearsEnd = pd.date_range(Startdate, Enddate, freq="A")
if len(YearsStart) > 0 and len(YearsEnd) > 0:
Years = range(int(YearsStart[0].year), int(YearsEnd[-1].year + 1))
Start_Year = np.argwhere(str(YearsStart[0])[0:10] == Dates)[0][0]
else:
Years = []
# Calculate the area for each pixel in square meters
area_in_m2 = Start.Area_converter.Degrees_to_m2(Example_dataset)
# Create Beneficial Maps
lulc_dict = GD.get_lulcs()
# Get all the LULC values
Values_LULC = np.unique(LULC)
# Create new Benefial arrays
T_ben_array = np.zeros(np.shape(LULC))
E_ben_array = np.zeros(np.shape(LULC))
I_ben_array = np.zeros(np.shape(LULC))
agriculture_array = np.zeros(np.shape(LULC))
environment_array = np.zeros(np.shape(LULC))
economic_array = np.zeros(np.shape(LULC))
energy_array = np.zeros(np.shape(LULC))
leisure_array = np.zeros(np.shape(LULC))
# Loop over LULC values and set benefial fractions
for Value_LULC in Values_LULC:
if Value_LULC in lulc_dict.keys():
T_ben = lulc_dict[Value_LULC][3]
E_ben = lulc_dict[Value_LULC][4]
I_ben = lulc_dict[Value_LULC][5]
agriculture = lulc_dict[Value_LULC][6]
environment = lulc_dict[Value_LULC][7]
economic = lulc_dict[Value_LULC][8]
energy = lulc_dict[Value_LULC][9]
leisure = lulc_dict[Value_LULC][10]
T_ben_array[LULC == Value_LULC] = T_ben / 100.
E_ben_array[LULC == Value_LULC] = E_ben / 100.
I_ben_array[LULC == Value_LULC] = I_ben / 100.
agriculture_array[LULC == Value_LULC] = agriculture / 100.
environment_array[LULC == Value_LULC] = environment / 100.
economic_array[LULC == Value_LULC] = economic / 100.
energy_array[LULC == Value_LULC] = energy / 100.
leisure_array[LULC == Value_LULC] = leisure / 100.
# Open sheet 2 dict
sheet2_classes_dict = GD.get_sheet2_classes()
# Convert data from mm/month to km3/month
I_km3 = np.einsum('ij,kij->kij', area_in_m2, DataCube_I) / 1e12
E_km3 = np.einsum('ij,kij->kij', area_in_m2, DataCube_E) / 1e12
T_km3 = np.einsum('ij,kij->kij', area_in_m2, DataCube_T) / 1e12
# Calculate beneficial I, E, and T
Iben_km3 = np.einsum('ij,kij->kij', I_ben_array, I_km3)
Eben_km3 = np.einsum('ij,kij->kij', E_ben_array, E_km3)
Tben_km3 = np.einsum('ij,kij->kij', T_ben_array, T_km3)
ETben_tot_km3 = Iben_km3 + Eben_km3 + Tben_km3
# Determine service contribution
agriculture_km3 = np.einsum('ij,kij->kij', agriculture_array,
ETben_tot_km3)
environment_km3 = np.einsum('ij,kij->kij', environment_array,
ETben_tot_km3)
economic_km3 = np.einsum('ij,kij->kij', economic_array, ETben_tot_km3)
energy_km3 = np.einsum('ij,kij->kij', energy_array, ETben_tot_km3)
leisure_km3 = np.einsum('ij,kij->kij', leisure_array, ETben_tot_km3)
# Create empty arrays
DataT = np.zeros([29, len(Dates)])
DataI = np.zeros([29, len(Dates)])
DataE = np.zeros([29, len(Dates)])
DataBT = np.zeros([29, len(Dates)])
DataBI = np.zeros([29, len(Dates)])
DataBE = np.zeros([29, len(Dates)])
DataAgriculture = np.zeros([29, len(Dates)])
DataEnvironment = np.zeros([29, len(Dates)])
DataEconomic = np.zeros([29, len(Dates)])
DataEnergy = np.zeros([29, len(Dates)])
DataLeisure = np.zeros([29, len(Dates)])
i = 0
# Loop over the LULC by using the Sheet 2 dictionary
for LAND_USE in sheet2_classes_dict.keys():
for CLASS in sheet2_classes_dict[LAND_USE].keys():
lulcs = sheet2_classes_dict[LAND_USE][CLASS]
# Create a mask to ignore non relevant pixels.
mask = np.logical_or.reduce([LULC == value for value in lulcs])
mask3d = mask * np.ones(len(Dates))[:, None, None]
# Calculate the spatial sum of the different parameters.
T_LU_tot = np.nansum(np.nansum((T_km3 * mask3d), 1), 1)
I_LU_tot = np.nansum(np.nansum((I_km3 * mask3d), 1), 1)
E_LU_tot = np.nansum(np.nansum((E_km3 * mask3d), 1), 1)
BT_LU_tot = np.nansum(np.nansum((Tben_km3 * mask3d), 1), 1)
BI_LU_tot = np.nansum(np.nansum((Iben_km3 * mask3d), 1), 1)
BE_LU_tot = np.nansum(np.nansum((Eben_km3 * mask3d), 1), 1)
Agriculture_LU_tot = np.nansum(
np.nansum((agriculture_km3 * mask3d), 1), 1)
Environment_LU_tot = np.nansum(
np.nansum((environment_km3 * mask3d), 1), 1)
Economic_LU_tot = np.nansum(np.nansum((economic_km3 * mask3d), 1),
1)
Energy_LU_tot = np.nansum(np.nansum((energy_km3 * mask3d), 1), 1)
Leisure_LU_tot = np.nansum(np.nansum((leisure_km3 * mask3d), 1), 1)
DataT[i, :] = T_LU_tot
DataBT[i, :] = BT_LU_tot
DataI[i, :] = I_LU_tot
DataBI[i, :] = BI_LU_tot
DataE[i, :] = E_LU_tot
DataBE[i, :] = BE_LU_tot
DataAgriculture[i, :] = Agriculture_LU_tot
DataEnvironment[i, :] = Environment_LU_tot
DataEconomic[i, :] = Economic_LU_tot
DataEnergy[i, :] = Energy_LU_tot
DataLeisure[i, :] = Leisure_LU_tot
i += 1
# Calculate non benefial components
DataNBT = DataT - DataBT
DataNBI = DataI - DataBI
DataNBE = DataE - DataBE
DataNB_tot = DataNBT + DataNBI + DataNBE
# Create CSV
first_row = [
'LAND_USE', 'CLASS', 'TRANSPIRATION', 'WATER', 'SOIL', 'INTERCEPTION',
'AGRICULTURE', 'ENVIRONMENT', 'ECONOMY', 'ENERGY', 'LEISURE',
'NON_BENEFICIAL'
]
i = 0
# Create monthly CSV
for Date in Dates:
# Create csv-file.
csv_filename = os.path.join(
Data_Path_CSV, 'Sheet2_Sim%d_%s_%d_%02d.csv' %
(Simulation, Basin, Date.year, Date.month))
csv_file = open(csv_filename, 'wb')
writer = csv.writer(csv_file, delimiter=';')
writer.writerow(first_row)
j = 0
# Loop over landuse and class
for LAND_USE in sheet2_classes_dict.keys():
for CLASS in sheet2_classes_dict[LAND_USE].keys():
# Get the value of the current class and landuse
Transpiration = DataT[j, i]
Evaporation = DataE[j, i]
Interception = DataI[j, i]
Agriculture = DataAgriculture[j, i]
Environment = DataEnvironment[j, i]
Economic = DataEconomic[j, i]
Energy = DataEnergy[j, i]
Leisure = DataLeisure[j, i]
Non_beneficial = DataNB_tot[j, i]
# Set special cases.
if np.any([
CLASS == 'Natural water bodies',
CLASS == 'Managed water bodies'
]):
Soil_evaporation = 0
Water_evaporation = Evaporation
else:
Soil_evaporation = Evaporation
Water_evaporation = 0
# Create the row to be written
row = [
LAND_USE, CLASS,
"{0:.2f}".format(np.nansum([0, Transpiration])),
"{0:.2f}".format(np.nansum([0, Water_evaporation])),
"{0:.2f}".format(np.nansum([0, Soil_evaporation])),
"{0:.2f}".format(np.nansum([0, Interception])),
"{0:.2f}".format(np.nansum([0, Agriculture])),
"{0:.2f}".format(np.nansum([0, Environment])),
"{0:.2f}".format(np.nansum([0, Economic])),
"{0:.2f}".format(np.nansum([0, Energy])), "{0:.2f}".format(
np.nansum([0, Leisure])),
"{0:.2f}".format(np.nansum([0, Non_beneficial]))
]
# Write the row.
writer.writerow(row)
j += 1
# Close the csv-file.
csv_file.close()
i += 1
# Create yearly CSV
i = 0
for Year in Years:
# Create csv-file.
csv_filename = os.path.join(
Data_Path_CSV,
'Sheet2_Sim%d_%s_%d.csv' % (Simulation, Basin, Year))
csv_file = open(csv_filename, 'wb')
writer = csv.writer(csv_file, delimiter=';')
writer.writerow(first_row)
j = 0
# Loop over landuse and class
for LAND_USE in sheet2_classes_dict.keys():
for CLASS in sheet2_classes_dict[LAND_USE].keys():
# Get the yearly value of the current class and landuse
Transpiration = np.sum(DataT[j, Start_Year:Start_Year + 12])
Evaporation = np.sum(DataE[j, Start_Year:Start_Year + 12])
Interception = np.sum(DataI[j, Start_Year:Start_Year + 12])
Agriculture = np.sum(DataAgriculture[j, Start_Year:Start_Year +
12])
Environment = np.sum(DataEnvironment[j, Start_Year:Start_Year +
12])
Economic = np.sum(DataEconomic[j, Start_Year:Start_Year + 12])
Energy = np.sum(DataEnergy[j, Start_Year:Start_Year + 12])
Leisure = np.sum(DataLeisure[j, Start_Year:Start_Year + 12])
Non_beneficial = np.sum(DataNB_tot[j,
Start_Year:Start_Year + 12])
# Set special cases.
if np.any([
CLASS == 'Natural water bodies',
CLASS == 'Managed water bodies'
]):
Soil_evaporation = 0
Water_evaporation = Evaporation
else:
Soil_evaporation = Evaporation
Water_evaporation = 0
# Create the row to be written
row = [
LAND_USE, CLASS,
"{0:.2f}".format(np.nansum([0, Transpiration])),
"{0:.2f}".format(np.nansum([0, Water_evaporation])),
"{0:.2f}".format(np.nansum([0, Soil_evaporation])),
"{0:.2f}".format(np.nansum([0, Interception])),
"{0:.2f}".format(np.nansum([0, Agriculture])),
"{0:.2f}".format(np.nansum([0, Environment])),
"{0:.2f}".format(np.nansum([0, Economic])),
"{0:.2f}".format(np.nansum([0, Energy])), "{0:.2f}".format(
np.nansum([0, Leisure])),
"{0:.2f}".format(np.nansum([0, Non_beneficial]))
]
# Write the row.
writer.writerow(row)
j += 1
# Close the csv-file.
csv_file.close()
i += 1
Start_Year += 12
return (Data_Path_CSV)
def Create(Dir_Basin, Simulation, Basin, Startdate, Enddate, nc_outname, Example_dataset):
"""
This functions create the CSV files for the sheets
Parameters
----------
Dir_Basin : str
Path to all the output data of the Basin
Simulation : int
Defines the simulation
Basin : str
Name of the basin
Startdate : str
Contains the start date of the model 'yyyy-mm-dd'
Enddate : str
Contains the end date of the model 'yyyy-mm-dd'
nc_outname : str
Path to the .nc file containing the data
Example_dataset : str
Data path to the example tiff file containing the right amount of pixels and projection
Returns
-------
Data_Path_CSV : str
Data path pointing to the CSV output files
"""
# import WA modules
import wa.Functions.Start.Get_Dictionaries as GD
import wa.General.raster_conversions as RC
from wa.Functions import Start
# Create output folder for CSV files
Data_Path_CSV = os.path.join(Dir_Basin, "Simulations", "Simulation_%d" %Simulation, "CSV")
if not os.path.exists(Data_Path_CSV):
os.mkdir(Data_Path_CSV)
# Open LULC map
LULC = RC.Open_nc_array(nc_outname, 'Landuse')
# Open I, T, E
DataCube_I = RC.Open_nc_array(nc_outname, 'Interception', Startdate, Enddate)
DataCube_T = RC.Open_nc_array(nc_outname, 'Transpiration', Startdate, Enddate)
DataCube_E = RC.Open_nc_array(nc_outname, 'Evaporation', Startdate, Enddate)
# Set the months
Dates = pd.date_range(Startdate, Enddate, freq = "MS")
# Define whole years
YearsStart = pd.date_range(Startdate, Enddate, freq = "AS")
YearsEnd = pd.date_range(Startdate, Enddate, freq = "A")
if len(YearsStart) > 0 and len(YearsEnd) > 0:
Years = range(int(YearsStart[0].year), int(YearsEnd[-1].year + 1))
Start_Year = np.argwhere(str(YearsStart[0])[0:10]==Dates)[0][0]
else:
Years = []
# Calculate the area for each pixel in square meters
area_in_m2 = Start.Area_converter.Degrees_to_m2(Example_dataset)
# Create Beneficial Maps
lulc_dict = GD.get_lulcs()
# Get all the LULC values
Values_LULC = np.unique(LULC)
# Create new Benefial arrays
T_ben_array = np.zeros(np.shape(LULC))
E_ben_array = np.zeros(np.shape(LULC))
I_ben_array = np.zeros(np.shape(LULC))
agriculture_array = np.zeros(np.shape(LULC))
environment_array= np.zeros(np.shape(LULC))
economic_array = np.zeros(np.shape(LULC))
energy_array = np.zeros(np.shape(LULC))
leisure_array = np.zeros(np.shape(LULC))
# Loop over LULC values and set benefial fractions
for Value_LULC in Values_LULC:
if Value_LULC in lulc_dict.keys():
T_ben = lulc_dict[Value_LULC][3]
E_ben = lulc_dict[Value_LULC][4]
I_ben = lulc_dict[Value_LULC][5]
agriculture = lulc_dict[Value_LULC][6]
environment = lulc_dict[Value_LULC][7]
economic = lulc_dict[Value_LULC][8]
energy = lulc_dict[Value_LULC][9]
leisure = lulc_dict[Value_LULC][10]
T_ben_array[LULC == Value_LULC] = T_ben/100.
E_ben_array[LULC == Value_LULC] = E_ben/100.
I_ben_array[LULC == Value_LULC] = I_ben/100.
agriculture_array[LULC == Value_LULC] = agriculture/100.
environment_array[LULC == Value_LULC] = environment/100.
economic_array[LULC == Value_LULC] = economic /100.
energy_array[LULC == Value_LULC] = energy/100.
leisure_array[LULC == Value_LULC] = leisure /100.
# Open sheet 2 dict
sheet2_classes_dict = GD.get_sheet2_classes()
# Convert data from mm/month to km3/month
I_km3 = np.einsum('ij,kij->kij', area_in_m2, DataCube_I)/ 1e12
E_km3 = np.einsum('ij,kij->kij', area_in_m2, DataCube_E)/ 1e12
T_km3 = np.einsum('ij,kij->kij', area_in_m2, DataCube_T)/ 1e12
# Calculate beneficial I, E, and T
Iben_km3 = np.einsum('ij,kij->kij', I_ben_array, I_km3)
Eben_km3 = np.einsum('ij,kij->kij', E_ben_array, E_km3)
Tben_km3 = np.einsum('ij,kij->kij', T_ben_array, T_km3)
ETben_tot_km3 = Iben_km3 + Eben_km3 + Tben_km3
# Determine service contribution
agriculture_km3 = np.einsum('ij,kij->kij', agriculture_array, ETben_tot_km3)
environment_km3 = np.einsum('ij,kij->kij', environment_array, ETben_tot_km3)
economic_km3 = np.einsum('ij,kij->kij', economic_array, ETben_tot_km3)
energy_km3 = np.einsum('ij,kij->kij', energy_array, ETben_tot_km3)
leisure_km3 = np.einsum('ij,kij->kij', leisure_array, ETben_tot_km3)
# Create empty arrays
DataT = np.zeros([29,len(Dates)])
DataI = np.zeros([29,len(Dates)])
DataE = np.zeros([29,len(Dates)])
DataBT = np.zeros([29,len(Dates)])
DataBI = np.zeros([29,len(Dates)])
DataBE = np.zeros([29,len(Dates)])
DataAgriculture = np.zeros([29,len(Dates)])
DataEnvironment = np.zeros([29,len(Dates)])
DataEconomic = np.zeros([29,len(Dates)])
DataEnergy = np.zeros([29,len(Dates)])
DataLeisure = np.zeros([29,len(Dates)])
i = 0
# Loop over the LULC by using the Sheet 2 dictionary
for LAND_USE in sheet2_classes_dict.keys():
for CLASS in sheet2_classes_dict[LAND_USE].keys():
lulcs = sheet2_classes_dict[LAND_USE][CLASS]
# Create a mask to ignore non relevant pixels.
mask=np.logical_or.reduce([LULC == value for value in lulcs])
mask3d = mask * np.ones(len(Dates))[:,None,None]
# Calculate the spatial sum of the different parameters.
T_LU_tot = np.nansum(np.nansum((T_km3 * mask3d),1),1)
I_LU_tot = np.nansum(np.nansum((I_km3 * mask3d),1),1)
E_LU_tot = np.nansum(np.nansum((E_km3 * mask3d),1),1)
BT_LU_tot = np.nansum(np.nansum((Tben_km3 * mask3d),1),1)
BI_LU_tot = np.nansum(np.nansum((Iben_km3 * mask3d),1),1)
BE_LU_tot = np.nansum(np.nansum((Eben_km3 * mask3d),1),1)
Agriculture_LU_tot = np.nansum(np.nansum((agriculture_km3 * mask3d),1),1)
Environment_LU_tot = np.nansum(np.nansum((environment_km3 * mask3d),1),1)
Economic_LU_tot = np.nansum(np.nansum((economic_km3 * mask3d),1),1)
Energy_LU_tot = np.nansum(np.nansum((energy_km3 * mask3d),1),1)
Leisure_LU_tot = np.nansum(np.nansum((leisure_km3 * mask3d),1),1)
DataT[i,:] = T_LU_tot
DataBT[i,:] = BT_LU_tot
DataI[i,:] = I_LU_tot
DataBI[i,:] = BI_LU_tot
DataE[i,:] = E_LU_tot
DataBE[i,:] = BE_LU_tot
DataAgriculture[i,:] = Agriculture_LU_tot
DataEnvironment[i,:] = Environment_LU_tot
DataEconomic[i,:] = Economic_LU_tot
DataEnergy[i,:] = Energy_LU_tot
DataLeisure[i,:] = Leisure_LU_tot
i += 1
# Calculate non benefial components
DataNBT = DataT - DataBT
DataNBI = DataI - DataBI
DataNBE = DataE - DataBE
DataNB_tot = DataNBT + DataNBI + DataNBE
# Create CSV
first_row = ['LAND_USE', 'CLASS', 'TRANSPIRATION', 'WATER', 'SOIL', 'INTERCEPTION', 'AGRICULTURE', 'ENVIRONMENT', 'ECONOMY', 'ENERGY', 'LEISURE', 'NON_BENEFICIAL']
i = 0
# Create monthly CSV
for Date in Dates:
# Create csv-file.
csv_filename = os.path.join(Data_Path_CSV, 'Sheet2_Sim%d_%s_%d_%02d.csv' %(Simulation, Basin, Date.year, Date.month))
csv_file = open(csv_filename, 'wb')
writer = csv.writer(csv_file, delimiter=';')
writer.writerow(first_row)
j = 0
# Loop over landuse and class
for LAND_USE in sheet2_classes_dict.keys():
for CLASS in sheet2_classes_dict[LAND_USE].keys():
# Get the value of the current class and landuse
Transpiration = DataT[j,i]
Evaporation = DataE[j,i]
Interception = DataI[j,i]
Agriculture = DataAgriculture[j,i]
Environment = DataEnvironment[j,i]
Economic = DataEconomic[j,i]
Energy = DataEnergy[j,i]
Leisure = DataLeisure[j,i]
Non_beneficial = DataNB_tot[j,i]
# Set special cases.
if np.any([CLASS == 'Natural water bodies', CLASS == 'Managed water bodies']):
Soil_evaporation = 0
Water_evaporation = Evaporation
else:
Soil_evaporation = Evaporation
Water_evaporation = 0
# Create the row to be written
row = [LAND_USE, CLASS, "{0:.2f}".format(np.nansum([0, Transpiration])), "{0:.2f}".format(np.nansum([0, Water_evaporation])), "{0:.2f}".format(np.nansum([0, Soil_evaporation])), "{0:.2f}".format(np.nansum([0, Interception])), "{0:.2f}".format(np.nansum([0, Agriculture])), "{0:.2f}".format(np.nansum([0, Environment])), "{0:.2f}".format(np.nansum([0, Economic])), "{0:.2f}".format(np.nansum([0, Energy])), "{0:.2f}".format(np.nansum([0, Leisure])), "{0:.2f}".format(np.nansum([0, Non_beneficial]))]
# Write the row.
writer.writerow(row)
j += 1
# Close the csv-file.
csv_file.close()
i += 1
# Create yearly CSV
i = 0
for Year in Years:
# Create csv-file.
csv_filename = os.path.join(Data_Path_CSV, 'Sheet2_Sim%d_%s_%d.csv' %(Simulation, Basin, Year))
csv_file = open(csv_filename, 'wb')
writer = csv.writer(csv_file, delimiter=';')
writer.writerow(first_row)
j = 0
# Loop over landuse and class
for LAND_USE in sheet2_classes_dict.keys():
for CLASS in sheet2_classes_dict[LAND_USE].keys():
# Get the yearly value of the current class and landuse
Transpiration = np.sum(DataT[j,Start_Year:Start_Year+12])
Evaporation = np.sum(DataE[j,Start_Year:Start_Year+12])
Interception = np.sum(DataI[j,Start_Year:Start_Year+12])
Agriculture = np.sum(DataAgriculture[j,Start_Year:Start_Year+12])
Environment = np.sum(DataEnvironment[j,Start_Year:Start_Year+12])
Economic = np.sum(DataEconomic[j,Start_Year:Start_Year+12])
Energy = np.sum(DataEnergy[j,Start_Year:Start_Year+12])
Leisure = np.sum(DataLeisure[j,Start_Year:Start_Year+12])
Non_beneficial = np.sum(DataNB_tot[j,Start_Year:Start_Year+12])
# Set special cases.
if np.any([CLASS == 'Natural water bodies', CLASS == 'Managed water bodies']):
Soil_evaporation = 0
Water_evaporation = Evaporation
else:
Soil_evaporation = Evaporation
Water_evaporation = 0
# Create the row to be written
row = [LAND_USE, CLASS, "{0:.2f}".format(np.nansum([0, Transpiration])), "{0:.2f}".format(np.nansum([0, Water_evaporation])), "{0:.2f}".format(np.nansum([0, Soil_evaporation])), "{0:.2f}".format(np.nansum([0, Interception])), "{0:.2f}".format(np.nansum([0, Agriculture])), "{0:.2f}".format(np.nansum([0, Environment])), "{0:.2f}".format(np.nansum([0, Economic])), "{0:.2f}".format(np.nansum([0, Energy])), "{0:.2f}".format(np.nansum([0, Leisure])), "{0:.2f}".format(np.nansum([0, Non_beneficial]))]
# Write the row.
writer.writerow(row)
j += 1
# Close the csv-file.
csv_file.close()
i += 1
Start_Year += 12
return(Data_Path_CSV)