Example 1: Basic usage

Fisrt of all, we need to install Docker/Docker Desktop.

Platform: arm64/amd64

[1]:

from pyclmuapp import usp_clmu
import matplotlib.pyplot as plt

Here we use the RUN_TYPE= "branch" to avoid the repeated spinup phases.

The case (usp_spinup) is run for 10 years to spinup the model.

How to get usp_spinup ?

usp_spinup = usp.run(
            case_name = "usp_spinup",
            SURF="surfdata.nc",
            FORCING="forcing.nc",
            RUN_STARTDATE = "2002-01-01",
            STOP_OPTION = "nyears",
            STOP_N = "10",
            RUN_TYPE= "coldstart",
        )
usp_spinup

[2]:

%%time
# initialize
usp = usp_clmu()
usp_london = usp.run(
            case_name = "example1",
            SURF="surfdata.nc",
            FORCING="forcing.nc",
            RUN_STARTDATE = "2012-01-01",
            STOP_OPTION = "nyears",
            STOP_N = "2",
            RUN_TYPE= "branch",
            RUN_REFCASE= "usp_spinup", # the case name of the spinup run
            RUN_REFDATE= "2012-01-01",
            hist_type = "'GRID','COLS'", # output type
        )
usp_london

Copying the file forcing.nc to the /Users/user/Documents/GitHub/pyclmuapp/docs/notebooks/usp/workdir/inputfolder/usp
CPU times: user 1.16 s, sys: 328 ms, total: 1.49 s
Wall time: 3min 36s

[2]:

['/Users/user/Documents/GitHub/pyclmuapp/docs/notebooks/usp/workdir/outputfolder/lnd/hist/example1_clm0_2024-11-23_14-53-19_clm.nc',
 '/Users/user/Documents/GitHub/pyclmuapp/docs/notebooks/usp/workdir/outputfolder/lnd/hist/example1_clm1_2024-11-23_14-53-19_clm.nc']

Plotting

Note: ploting with time will need the nc-time-axis package, which can be installed by pip install nc-time-axis or conda install nc-time-axis

[3]:

%%time
ds_usp=usp.nc_view(usp_london[1]).isel(time=slice(2, None))
#ds_usp['time'] = ds_usp['time'].dt.round('min')
ds_usp_hour = ds_usp['TSA'].isel(gridcell=0).groupby('time.hour')
mean_tsa = ds_usp_hour.mean('time')
var_tsa = ds_usp_hour.var('time')

ds_usp_hour = ds_usp['TBOT'].isel(gridcell=0).groupby('time.hour')
mean_tbot = ds_usp_hour.mean('time')
var_tbot = ds_usp_hour.var('time')

fig = plt.figure(figsize=(6, 5))

ax = fig.add_subplot(211)
ds_usp['TSA'].isel(gridcell=0).plot(ax=ax, label='TSA', color='#E02927')#["#E02927","#3964DF"]
ds_usp['TBOT'].isel(gridcell=0).plot(ax=ax, label='Tair', color='#3964DF')

ax.set_xlabel('Time', fontsize=14)
ax.set_ylabel('Temperature [K]', fontsize=14)
ax.set_title('a', loc='left', fontweight='bold')
ax.tick_params(axis='both', which='major', labelsize=12)
ax.legend(frameon=False, fontsize=12)

ax = fig.add_subplot(212)
mean_tsa.plot(ax=ax, label='TSA', color='#E02927')
mean_tbot.plot(ax=ax, label='Tair', color='#3964DF')
ax.set_xlabel('Local hour of day', fontsize=14)
ax.set_ylabel('Temperature [K]', fontsize=14)
ax.set_title('b', loc='left', fontweight='bold')
ax.tick_params(axis='both', which='major', labelsize=12)
ax.legend(frameon=False, fontsize=12)

plt.tight_layout()
plt.savefig('figs/exp1_usp.pdf', dpi=300)
plt.show()
../../_images/notebooks_usp_example1_usp_basic_5_0.png 
CPU times: user 1.92 s, sys: 512 ms, total: 2.44 s
Wall time: 3.32 s

plot other urban variables

radiation balance

\[Rnet(Q\text{*})=SWdown-SWup+LWdown-LWup\]
[4]:

fig = plt.figure(figsize=(12, 5))

axs = fig.subplots(2, 2)

ds_usp_h = ds_usp[['Rnet','SWdown','SWup','LWdown','LWup']].isel(gridcell=0).groupby('time.hour')
h = ds_usp_h.mean('time')
color_list = ["#AE02DE", "#65D8DF","#E02927","#E0C645","#3964DF"]

ds_usp_month = ds_usp[['Rnet','SWdown','SWup','LWdown','LWup']].isel(gridcell=0).groupby('time.month')
m = ds_usp_month.mean('time')

for i, var in enumerate(['Rnet','SWdown','SWup','LWdown','LWup']):
    if var == 'Rnet':
        label = 'Q*'
    else:
        label = var
    ds_usp[var].plot(ax=axs[0,0], label=label, color=color_list[i])
    m[var].plot(ax=axs[1,0], label=label, color=color_list[i], marker='o')
    h[var].plot(ax=axs[0,1], label=label, color=color_list[i], marker='o')

(ds_usp['Rnet'] - ds_usp['SWdown'] - ds_usp['LWdown']  + ds_usp['SWup'] + ds_usp['LWup'] ).plot(ax=axs[1,1], color='#3964DF')

xlabels = ['Time', 'Local hour of day', 'Month of year', 'Time']
ylabels = ['Flux [W/m$\mathrm{^{2}}$]', 'Flux [W/m$\mathrm{^{2}}$]', 'Flux [W/m$\mathrm{^{2}}$]', 'Residual [W/m$\mathrm{^{2}}$]']
titles = ['a', 'b', 'c', 'd']
for ax in axs.flat:
    ax.set_xlabel(xlabels.pop(0), fontsize=14)
    ax.set_ylabel(ylabels.pop(0), fontsize=14)
    ax.set_title(titles.pop(0), loc='left', fontweight='bold', fontsize=14)
    ax.tick_params(axis='both', which='major', labelsize=12)
axs[1,1].set_ylim(-5,5)
fig.legend(['Q*','SWdown','SWup','LWdown','LWup'],
           loc="upper center", frameon=False, ncol=5,
           bbox_to_anchor=(0.5, 1.05), fontsize=12)

plt.tight_layout()
plt.savefig('figs/exp1_usp_rad.pdf', dpi=300, bbox_inches='tight')
plt.show()
../../_images/notebooks_usp_example1_usp_basic_8_0.png

energy balance

\[Q\text{*} + Qanth=Qle+Qh+Qstor\]
\[Q\text{*}=Rnet\]
\[Qanth=WASTEHEAT+URBAN_{HEAT}\]
\[Qstor=FGR+URBAN_{HEAT}-URBAN_{AC}\]

  • Qanth: anthropogenic heat flux

  • Qle: latent heat flux

  • Qh: sensible heat flux

  • Qstor: storage heat flux

  • WASTEHEAT: wasteheat flux from heating and air conditioning flux

  • URBAN_HEAT: urban heating flux

  • URBAN_AC: heat flux from urban air conditioning

[5]:

fig = plt.figure(figsize=(12, 5))

axs = fig.subplots(2, 2)

#ds_usp['Qanth'] = ds_usp['URBAN_HEAT'] + ds_usp['WASTEHEAT'] - ds_usp['EFLXBUILD']
ds_usp['Qanth'] = ds_usp['WASTEHEAT'] + ds_usp['URBAN_HEAT']
ds_usp['Qstor'] = ds_usp['FGR'] + ds_usp['URBAN_HEAT'] - ds_usp['URBAN_AC']
ds_usp_h = ds_usp[['Rnet','Qle', 'Qh', 'Qstor','Qanth']].isel(gridcell=0).groupby('time.hour')
h = ds_usp_h.mean('time')
color_list = ["#AE02DE", "#65D8DF","#E02927","#E0C645","#3964DF"]

ds_usp_month = ds_usp[['Rnet','Qanth','Qle', 'Qh', 'Qstor']].isel(gridcell=0).groupby('time.month')
m = ds_usp_month.mean('time')

for i, var in enumerate(['Rnet','Qanth','Qle', 'Qh', 'Qstor']):
    if var == 'Rnet':
        label = 'Q*'
    else:
        label = var
    ds_usp[var].plot(ax=axs[0,0], label=label, color=color_list[i])
    m[var].plot(ax=axs[1,0], label=label, color=color_list[i], marker='o')
    h[var].plot(ax=axs[0,1], label=label, color=color_list[i], marker='o')

(ds_usp['Rnet']+ds_usp['Qanth'] - ds_usp['Qle'] - ds_usp['Qh'] - ds_usp['Qstor']).plot(ax=axs[1,1], color='#3964DF')

xlabels = ['Time', 'Local hour of day', 'Month of year', 'Time']
ylabels = ['Flux [W/m$\mathrm{^{2}}$]', 'Flux [W/m$\mathrm{^{2}}$]', 'Flux [W/m$\mathrm{^{2}}$]', 'Residual [W/m$\mathrm{^{2}}$]']
titles = ['a', 'b', 'c', 'd']
for ax in axs.flat:
    ax.set_xlabel(xlabels.pop(0), fontsize=14)
    ax.set_ylabel(ylabels.pop(0), fontsize=14)
    ax.set_title(titles.pop(0), loc='left', fontweight='bold', fontsize=14)
    ax.tick_params(axis='both', which='major', labelsize=12)
axs[1,1].set_ylim(-1,1)
fig.legend(['Q*','Qanth','Qle','Qh','Qstor'], loc='upper center', frameon=False, ncol=5, bbox_to_anchor=(0.5, 1.05), fontsize=12)

plt.tight_layout()
plt.savefig('figs/exp1_usp_energy.pdf', dpi=300, bbox_inches='tight')
plt.show()
../../_images/notebooks_usp_example1_usp_basic_10_0.png

buiding energy consumption

[6]:

fig = plt.figure(figsize=(6, 5))
ax = fig.add_subplot(211)
color_list = ["#AE02DE","#E02927","#3964DF"]
tot = ds_usp.sel(gridcell=0)
tot['TOTQANTH'] = tot['URBAN_HEAT'] + tot['WASTEHEAT']
tot = tot[['URBAN_HEAT','URBAN_AC','WASTEHEAT', 'TOTQANTH']].groupby('time.month').mean('time')
tot['URBAN_HEAT'].plot(ax=ax, label='Urban heating', marker='o', color=color_list.pop(0))
tot['URBAN_AC'].plot(ax=ax, label='Urban air conditioning', marker='o', color=color_list.pop(0))
tot['WASTEHEAT'].plot(ax=ax, label='Waste heat', marker='o', color=color_list.pop(0))
tot['TOTQANTH'].plot(ax=ax, label='Anthropogenic heat', marker='o', color='#E0C645')
ax.set_xlabel('Month of year', fontsize=14)
ax.set_ylabel('Heat flux [W/m$\mathrm{^{2}}$]', fontsize=14)
ax.set_title('a', loc='left', fontweight='bold', fontsize=14)
ax.tick_params(axis='both', which='major', labelsize=12)
ax.legend(frameon=False, fontsize=11, loc=(0.35,0.45))

ax = fig.add_subplot(212)
tot = ds_usp.sel(gridcell=0)['TBUILD'].groupby('time.month').mean('time').plot(ax=ax, marker='o')
ax.set_xlabel('Month of year', fontsize=14)
ax.set_ylabel('Internal building\nair temperature [K]', fontsize=14)
ax.tick_params(axis='both', which='major', labelsize=12)

plt.tight_layout()
plt.savefig('figs/exp1_usp_urban_bem.pdf', dpi=300)
plt.show()
../../_images/notebooks_usp_example1_usp_basic_12_0.png

column level temperature

[7]:

ds_usp_col=usp.nc_view(usp_london[0]).isel(time=slice(2, None))
ds_usp_col

[7]:

<xarray.Dataset> Size: 134MB
Dimensions:             (levgrnd: 25, levlak: 10, levdcmp: 1, time: 35086,
                         hist_interval: 2, lndgrid: 1, gridcell: 1,
                         landunit: 2, column: 6, pft: 6, levsoi: 20)
Coordinates:
  * levgrnd             (levgrnd) float32 100B 0.01 0.04 0.09 ... 28.87 42.0
  * levlak              (levlak) float32 40B 0.05 0.6 2.1 ... 25.6 34.33 44.78
  * levdcmp             (levdcmp) float32 4B 1.0
  * time                (time) datetime64[ns] 281kB 2012-01-01T01:30:00 ... 2...
Dimensions without coordinates: hist_interval, lndgrid, gridcell, landunit,
                                column, pft, levsoi
Data variables: (12/121)
    mcdate              (time) int32 140kB ...
    mcsec               (time) int32 140kB ...
    mdcur               (time) int32 140kB ...
    mscur               (time) int32 140kB ...
    nstep               (time) int32 140kB ...
    time_bounds         (time, hist_interval) datetime64[ns] 561kB ...
    ...                  ...
    URBAN_AC            (time, column) float32 842kB ...
    URBAN_HEAT          (time, column) float32 842kB ...
    WASTEHEAT           (time, column) float32 842kB ...
    WBT                 (time, column) float32 842kB ...
    Wind                (time, column) float32 842kB ...
    ZWT                 (time, column) float32 842kB ...
Attributes: (12/40)
    title:                                     CLM History file information
    comment:                                   NOTE: None of the variables ar...
    Conventions:                               CF-1.0
    history:                                   created on 11/23/24 14:50:56
    source:                                    Community Land Model CLM4.0
    hostname:                                  clmu-app
    ...                                        ...
    ctype_urban_pervious_road:                 75
    cft_c3_crop:                               1
    cft_c3_irrigated:                          2
    time_period_freq:                          minute_30
    Time_constant_3Dvars_filename:             ./example1.clm2.h0.2012-01-01-...
    Time_constant_3Dvars:                      ZSOI:DZSOI:WATSAT:SUCSAT:BSW:H...
[8]:

ds_usp_col=usp.nc_view(usp_london[0]).isel(time=slice(2, None))
ds_usp_col = ds_usp_col.assign_coords({'column':ds_usp_col['cols1d_itype_col']})
fig = plt.figure(figsize=(12, 5))
var_list = ['TSKIN','FGR','FSH','EFLX_LH_TOT']
title = ['a', 'b', 'c', 'd']
color_list = ["#AE02DE", "#65D8DF","#E02927","#E0C645","#3964DF"]
ylabel = {
    'TSKIN':'Skin temperature [K]',
    'FGR':'Ground heat flux [W/m$\mathrm{^{2}}$]',
    'FSH':'Sensible heat flux [W/m$\mathrm{^{2}}$]',
    'EFLX_LH_TOT':'Latent heat flux [W/m$\mathrm{^{2}}$]'
}
markers = ['o', 's', '^', 'v', 'd']
for i, var in enumerate(var_list):
    ax = fig.add_subplot(2, 2, i+1)

    for col in range(71,76):
        #ds_usp_col[var].sel(column=col).resample(time='M').mean()\
        ds_usp_col[var].sel(column=col).groupby('time.hour').mean('time')\
            .plot(ax=ax, color=color_list[col-71], marker=markers[col-71])

    ax.set_ylabel(ylabel[var], fontsize=12)
    ax.tick_params(axis='both', which='major', labelsize=12)
    if i in [2,3]:
        ax.set_xlabel('Local hour of day', fontsize=14)
    else:
        ax.set_xlabel('')
        ax.set_xticklabels([])
    #ax.legend(frameon=False)
    ax.set_title('')
    ax.set_title(title[i], loc='left', fontweight='bold', fontsize=14)

fig.legend(['Roof','Sunwall','Shadewall','Impervious road','Pervious road'],
           loc='upper center', frameon=False, ncol=5, bbox_to_anchor=(0.5, 1.05), fontsize=12)
fig.tight_layout()
fig.savefig('figs/exp1_usp_urban_col.pdf', dpi=300, bbox_inches='tight')
plt.show()
../../_images/notebooks_usp_example1_usp_basic_15_0.png

water balance

\[residual = RAIN + SNOW- EVPA - Qs - Qsb- DelSWE- DelSoilMoist - DelIntercept\]

  • RAIN: rain from atmosphere

  • SNOW: snow from atmosphere

  • EVPA: total evaporation. \(EVPA=QVEGT+QVEGE+QSOIL\)

  • QVEGT: canopy transpiration

  • QVEGE: canopy evaporation

  • QSOIL: Ground evaporation (soil/snow evaporation + soil/snow sublimation - dew)

  • Qs: total runoff. \(Qs=QOVER+QH2OSFC\)

  • Qsb: subsurface drainage. \(Qsb=QDRAI+QDRAI_PERCH\)

  • QDRAI: sub-surface drainage

  • QDRAI_PERCH: perched wt drainage

  • DelSWE: delta H2OSNO (snow depth (liquid water)) divided by time-step.

  • DelSoilMoist: delta SOILH2O (soil moisture) divided by time-step. SOILH2O is the sum of column level SOILLIQ and SOILICE weighed by pervious road (see following code).

  • DelIntercept: the delta intercept water of roof and pervious road divided by time-step

[9]:

ds_usp=usp.nc_view(usp_london[1])#.isel(time=slice(2, None))
ds_usp_col=usp.nc_view(usp_london[0])#.isel(time=slice(2, None))
ds_usp['EVPA'] = ds_usp['QVEGT'] + ds_usp['QVEGE'] + ds_usp['QSOIL']
ds_usp['Qs'] = ds_usp['QOVER'] + ds_usp['QH2OSFC']
ds_usp['Qsb'] = ds_usp['QDRAI'] + ds_usp['QDRAI_PERCH']
ds_usp['DelSWE'] =ds_usp['H2OSNO'].diff('time')/1800

ds_usp_col = ds_usp_col.assign_coords({'column':ds_usp_col['cols1d_itype_col']})
roof_fra = ds_usp_col['cols1d_wtgcell'].sel(column=71).values
road_imp_fra = ds_usp_col['cols1d_wtgcell'].sel(column=74).values
road_pre_fra = ds_usp_col['cols1d_wtgcell'].sel(column=75).values
ds_usp_col['SOILH2O'] = (1 - roof_fra) \
                        * road_pre_fra / (road_pre_fra + road_imp_fra) \
                        * (ds_usp_col['SOILLIQ'].sel(column=75) + ds_usp_col['SOILICE'].sel(column=75))
#ds_usp_col['SOILH2O'] = road_pre_fra * (ds_usp_col['SOILLIQ'].sel(column=75) + ds_usp_col['SOILICE'].sel(column=75))
ds_usp_col['Intercept'] = roof_fra *\
                            (ds_usp_col['SOILLIQ'].sel(column=71) + ds_usp_col['SOILICE'].sel(column=71))\
                        + (1-roof_fra)*road_imp_fra/(road_pre_fra+road_imp_fra)\
                        * (ds_usp_col['SOILLIQ'].sel(column=74) + ds_usp_col['SOILICE'].sel(column=74))

ds_usp['DelSoilMoist'] = ds_usp_col['SOILH2O'].sum('levsoi').diff('time')/1800
ds_usp['DelIntercept'] = ds_usp_col['Intercept'].sum('levsoi').diff('time')/1800

ds_usp = ds_usp.isel(time=slice(2, None))
water_list = ['RAIN', 'SNOW' ,
              'EVPA', 'Qs', 'Qsb', 'DelSoilMoist', 'DelIntercept', 'DelSWE']
#color_list = ["#E02927","#3964DF","#E0C645","#AE02DE","#65D8DF"]
ds_usp_month = ds_usp[water_list].isel(gridcell=0).groupby('time.month')
tot = ds_usp_month.mean('time')

fig = plt.figure(figsize=(6, 5))
ax = fig.add_subplot(211)
for i, var in enumerate(water_list):

    label = var
    tot[var].plot(ax=ax, marker='o', label=label,)
    ax.set_xlabel('Month', fontsize=14)
    ax.set_ylabel(var, fontsize=14)

ax.set_xlabel('')
ax.set_ylabel('Flux [mm/s]', fontsize=14)
ax.set_title('a', loc='left', fontweight='bold')
ax.legend(frameon=False, fontsize=10, loc='upper center', ncol=4, bbox_to_anchor=(0.5, 1.70))

ax = fig.add_subplot(212)

residual = ds_usp['RAIN'] + ds_usp['SNOW'] \
            - ds_usp['EVPA'] - ds_usp['Qs'] - ds_usp['Qsb'] \
            - ds_usp['DelSWE'] \
            - ds_usp['DelSoilMoist'] - ds_usp['DelIntercept']

residual.groupby('time.month').mean('time').plot(ax=ax, marker='o', label='Residual', color='#3964DF')
ax.set_xlabel('Month of year', fontsize=14)
ax.set_ylabel('Residual [mm/s]', fontsize=14)

ax.set_title('b', loc='left', fontweight='bold', fontsize=14)

ax.set_ylim(-1e12,1e12)

plt.tight_layout()
plt.savefig('figs/exp1_usp_water.pdf', dpi=300, bbox_inches='tight')
plt.show()
../../_images/notebooks_usp_example1_usp_basic_17_0.png

human heat stress index

[10]:

fig = plt.figure(figsize=(6, 5))
dis_list = ['HIA','SWBGT','WBT', 'HUMIDEX']

ax = fig.add_subplot(211)
mean = ds_usp['TSA'].sel(gridcell=0).groupby('time.month').mean('time')
std = ds_usp['TSA'].sel(gridcell=0).groupby('time.month').std('time')
mean.plot(ax=ax, marker='o',color='#E02927')
#ax.fill_between(mean['month'].values, mean - std, mean + std, alpha=0.2, color='#E02927')
ax.set_xlabel('')
ax.tick_params(axis='both', which='major', labelsize=12)
ax.set_ylabel('Air temperature [K]', fontsize=14)
ax.set_title('a', loc='left', fontweight='bold', fontsize=14)
ax.legend(frameon=False)
print('TSA', mean.values.round(2))

ax2 = ax.twinx()
mean = ds_usp['RH2M'].sel(gridcell=0).groupby('time.month').mean('time')
mean.plot(ax=ax2, marker='o', color='#3964DF')#color_list = ["#E02927","#3964DF"]
std = ds_usp['RH2M'].sel(gridcell=0).groupby('time.month').std('time')
#ax2.fill_between(mean['month'].values, mean - std, mean + std, alpha=0.2, color='#3964DF')
ax2.set_ylabel('Relative humidity [%]', color='#3964DF', fontsize=14)
ax2.tick_params(axis='y', labelcolor='#3964DF', labelsize=12)

ax = fig.add_subplot(212)
color_list = ["#AE02DE","#E02927","#3964DF","#E0C645"]
tot = ds_usp[dis_list].sel(gridcell=0)

for var in dis_list:
    mean = tot[var].groupby('time.month').mean('time')
    std = tot[var].groupby('time.month').std('time')
    color = color_list.pop(0)
    mean.plot(ax=ax, label=var, marker='o', color=color)
    print(var, mean.values.round(2))
    #ax.plot(mean, label=var, marker='o', color=color_list.pop(0))
    #ax.fill_between(mean['month'].values, mean - std, mean + std, alpha=0.2, color=color)

ax.set_ylim(0, 27)
ax.set_xlabel('Month of year', fontsize=14)
ax.set_ylabel('Heat stress index [°C]', fontsize=14)
ax.tick_params(axis='both', which='major', labelsize=12)
ax.set_title('b', loc='left', fontweight='bold', fontsize=14)
ax.legend(frameon=False, fontsize=10, ncol=4, loc='upper right')

plt.tight_layout()
plt.savefig('figs/exp1_usp_urban_heat_stress.pdf', dpi=300)
plt.show()

No artists with labels found to put in legend.  Note that artists whose label start with an underscore are ignored when legend() is called with no argument.

TSA [279.59 278.31 280.55 282.82 286.78 289.17 292.96 292.77 289.01 286.21
 281.78 280.45]
HIA [ 6.44  5.16  7.44  9.72 13.81 16.44 20.41 20.58 16.11 13.09  8.63  7.3 ]
SWBGT [10.56  9.43 10.98 12.51 15.65 17.64 20.53 20.55 17.61 15.81 12.2  11.27]
WBT [ 4.19  2.6   4.45  6.15  9.68 11.9  14.66 14.77 11.9  10.3   6.2   5.17]
HUMIDEX [ 5.08  3.23  5.87  8.48 13.7  16.99 21.82 21.82 16.91 13.82  7.83  6.26]
../../_images/notebooks_usp_example1_usp_basic_19_2.png

clean up the case files

[11]:

usp.case_clean()

# when you using ups.run(crun_type='usp-exec') --> see warmup section
# the case files in ./workdir/ouptutfolder/your_case, ./workdir/logfolder/your_case,
# and ./workdir/inputfoler/usp will be removed.

# when you using ups.run(crun_type='run') or ups.run(), the default crun_type is 'run'
# only ./workdir/inputfoler/usp will be removed.