def plot_trend_tcl():
res = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/trends.csv', index_col=0)
ds_vod = io('SMOS_IC')
ds_tcl = io('TCL')
ds_agb = io('AGB')
lut = ds_vod.lut.reindex(res.index)
agb = ds_agb.read_img('AGB')
agb_err = ds_agb.read_img('AGB_err')
agb[agb_err > 160] = np.nan
date_from = '2010-01-01'
date_to = '2010-12-31'
vod = ds_vod.read_img('VOD', date_from=date_from, date_to=date_to)
invalid = (ds_vod.read_img('Flags', date_from=date_from, date_to=date_to) > 0) | \
(ds_vod.read_img('RMSE', date_from=date_from, date_to=date_to) > 8) | \
(ds_vod.read_img('VOD_StdErr', date_from=date_from, date_to=date_to) > 1.2)
vod[invalid] = np.nan
vod = np.nanmean(vod, axis=0)
invalid = np.where(np.isnan(agb) | np.isnan(vod))
agb[invalid] = np.nan
vod[invalid] = np.nan
valid = np.where(~np.isnan(agb))
p = np.poly1d(np.polyfit(vod[valid], agb[valid], 1))
fontsize = 16
f = plt.figure(figsize=(18, 10))
tmp = res[f'slope_vod_orig']
tmp[res[f'p_vod_orig'] > 0.05] = np.nan
tmp[tmp > 0] = np.nan
arr = np.full(ds_vod.lat.shape, np.nan)
arr[lut.row_ease, lut.col_ease] = tmp
plt.subplot(1, 3, 1)
im = plot_img(ds_vod.lon, ds_vod.lat, arr*12, cbrange=[-0.01,0], cmap=truncate_colormap(cc.cm.bjy, 0, 0.5), title='VOD trend [yr$^{-1}$]', fontsize=fontsize,
plot_cmap=True)
plt.subplot(1, 3, 2)
im = plot_img(ds_vod.lon, ds_vod.lat, p[0]*arr*(-12), cbrange=[-0.3,0], cmap=truncate_colormap(cc.cm.bjy, 0, 0.5), title='AGB trend [Mg/ha/yr]', fontsize=fontsize,
plot_cmap=True)
plt.subplot(1, 3, 3)
tcl = ds_tcl.read_img('TCL')
tcl[tcl < 0.07] = np.nan
im = plot_img(ds_tcl.lon, ds_tcl.lat, tcl, cbrange=[0.1,0.4], cmap=truncate_colormap(cc.cm.bjy_r, 0.5, 1), title='Tree cover loss [-]', fontsize=fontsize,
plot_cmap=True)
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/vod_trend_tcl.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
def plot_mask():
res = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/lagged_corr.csv', index_col=0)
ds = io('LAI')
lut = ds.lut.reindex(res.index)
fontsize = 16
t = np.full(ds.lat.shape, np.nan)
p = np.full(ds.lat.shape, np.nan)
r = np.full(ds.lat.shape, np.nan)
t[lut.row_ease, lut.col_ease] = res[f'R_VOD_T_-3']
p[lut.row_ease, lut.col_ease] = res[f'R_VOD_P_-3'] # possibly -2
r[lut.row_ease, lut.col_ease] = res[f'R_VOD_R_-2']
mask = np.full(ds.lat.shape, np.nan)
# mask[lut.row_ease, lut.col_ease] = 0
mask[(t < 0.5) & (p < 0.5) & (r < 0.5)] = 1
f = plt.figure(figsize=(6,10))
plot_img(ds.lon, ds.lat, mask, cbrange=None, cmap='viridis', title='Correlation mask', fontsize=fontsize,
plot_cmap=False)
plt.tight_layout()
plt.show()
def plot_bp_rss_stats():
res = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/breakpoints.csv', index_col=0)
ds = io('LAI')
lut = ds.lut.reindex(res.index)
fontsize = 15
variables = ['LAI', 'VOD']
f = plt.figure(figsize=(12, 10))
for n, variable in enumerate(variables):
cols = [col for col in res.columns.values if variable in col]
arr = np.full(ds.lat.shape, np.nan)
arr[lut.row_ease, lut.col_ease] = res[cols].max(axis=1) - res[cols].min(axis=1)
plt.subplot(1, 2, n+1)
im = plot_img(ds.lon, ds.lat, arr, cbrange=(0, 25), cmap='viridis_r', title=variable, fontsize=fontsize,
plot_cmap=True)
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/rss_range.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
def plot_trends():
res = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/trends.csv', index_col=0)
ds = io('LAI')
lut = ds.lut.reindex(res.index)
fontsize = 12
variables = ['lai', 'vod', 'temp', 'prec', 'rad']
titles = ['LAI [yr$^{-1}$]', 'VOD [yr$^{-1}$]', 'T [$^\circ$C / yr]', 'P [kg / m2 / mo / yr]', 'Rad. [W / m2 / yr]']
cmap = cc.cm.bjy
cbr = [-0.015, 0.015]
f = plt.figure(figsize=(18, 6))
for n, (variable, tit) in enumerate(zip(variables, titles)):
tmp = res[f'slope_{variable}']
tmp[res[f'p_{variable}'] > 0.05] = np.nan
arr = np.full(ds.lat.shape, np.nan)
arr[lut.row_ease, lut.col_ease] = tmp
plt.subplot(1, 5, n+1)
im = plot_img(ds.lon, ds.lat, arr, cbrange=cbr, cmap=cmap, title=tit, fontsize=fontsize,
plot_cmap=True)
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/trends_sig.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
def plot_best_corr_spatial_vod():
res = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/lagged_corr.csv', index_col=0)
ds = io('LAI')
lut = ds.lut.reindex(res.index)
fontsize = 12
tags = ['R_VOD_T', 'R_VOD_P', 'R_VOD_R']
tags_anom = ['R_anom_VOD_T', 'R_anom_VOD_P', 'R_anom_VOD_R']
lags = [-3, -3, -2]
lags_anom = [0, 0, -1]
titles = ['Temperature', 'Precipitation', 'Radiation']
cbr = (-0.9,0.9)
cbr_anom = (-0.4,0.4)
cmap = cc.cm.bjy
f = plt.figure(figsize=(10, 10))
for i, (tag, tag_anom, lag, lag_anom, tit) in enumerate(zip(tags, tags_anom, lags,lags_anom, titles)):
var = f'{tag}_{lag}'
var_anom = f'{tag_anom}_{lag_anom}'
arr = np.full(ds.lat.shape, np.nan)
arr[lut.row_ease, lut.col_ease] = res[var]
arr_anom = np.full(ds.lat.shape, np.nan)
arr_anom[lut.row_ease, lut.col_ease] = res[var_anom]
plt.subplot(2, 3, i+1)
im = plot_img(ds.lon, ds.lat, arr, cbrange=cbr, cmap=cmap, title=tit, fontsize=fontsize,
plot_cmap=True)
if i == 0:
plt.ylabel('Total signal', fontsize=fontsize)
plt.subplot(2, 3, i+4)
if i == 1:
cm = cc.cm.bjy_r
else:
cm = cmap
im = plot_img(ds.lon, ds.lat, arr_anom, cbrange=cbr_anom, cmap=cm, title='', fontsize=fontsize,
plot_cmap=True)
if i == 0:
plt.ylabel('Anomalies', fontsize=fontsize)
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/best_correlations.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
def plot_breakpoints():
res = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/breakpoints.csv', index_col=0)
ds = io('LAI')
lut = ds.lut.reindex(res.index)
fontsize = 15
variables = ['LAI', 'VOD']
cmap_d = colors.ListedColormap(sns.color_palette("viridis", n_colors=10))
cmap_d2 = colors.ListedColormap(sns.color_palette('viridis', n_colors=11))
# cmap_d = colors.ListedColormap(cm.get_cmap('jet')(np.linspace(0, 1, 10)))
# jet_12_colors = jet(np.linspace(0, 1, 10))
cbr = [2010, 2020]
f = plt.figure(figsize=(24, 10))
for n, variable in enumerate(variables):
cols = [col for col in res.columns.values if f'RSS_{variable}' in col]
year = [int(col.split('_')[-1]) for col in cols]
res[f'bp_{variable}'] = -1
min_rss = res[cols].min(axis=1)
for col, yr in zip(cols, year):
res[f'bp_{variable}'][res[col] == min_rss] = yr+1
arr = np.full(ds.lat.shape, np.nan)
arr[lut.row_ease, lut.col_ease] = res[f'bp_{variable}']
plt.subplot(1, 4, 2*n+1)
im = plot_img(ds.lon, ds.lat, arr, cbrange=cbr, cmap=cmap_d, title=f'{variable} (RSS)', fontsize=fontsize,
plot_cmap=True)
arr = np.full(ds.lat.shape, np.nan)
arr[lut.row_ease, lut.col_ease] = res[f'bfast_{variable.lower()}_b1']
plt.subplot(1, 4, 2*n+2)
im = plot_img(ds.lon, ds.lat, arr, cbrange=cbr, cmap=cmap_d, title=f'{variable} (BFAST)', fontsize=fontsize,
plot_cmap=True)
# plt.subplot(1, 3, 3)
# arr = np.full(ds.lat.shape, np.nan)
# arr[lut.row_ease, lut.col_ease] = res[f'bp_VOD'] - res[f'bp_LAI']
# im = plot_img(ds.lon, ds.lat, arr, cbrange=[-5, 5], cmap=cmap_d2, title='VOD - LAI', fontsize=fontsize,
# plot_cmap=True)
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/breakpoints_bfast.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
def plot_lagged_corr_spatial():
res = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/lagged_corr.csv', index_col=0)
ds = io('LAI')
lut = ds.lut.reindex(res.index)
fontsize = 12
variables = ['T', 'P', 'R']
titles = ['temperature', 'precipitation', 'radiation']
lags = [-5, -4, -3, -2, -1, 0]
cbr = (-0.3,0.4)
cmap = cc.cm.bjy
for met, tit in zip(variables, titles):
f = plt.figure(figsize=(22, 11))
plt.suptitle(f'Lagged correlation ({tit}) [-]', fontsize=fontsize+2, y=0.93)
for i, veg in enumerate(['LAI', 'VOD']):
for j, lag in enumerate(lags):
var = f'R_anom_{veg}_{met}_{lag}'
arr = np.full(ds.lat.shape, np.nan)
arr[lut.row_ease, lut.col_ease] = res[var]
plt.subplot(2, len(lags), i*len(lags) + j + 1)
title = f'lag: {lag} months' if i == 0 else ''
ylabel = veg if j == 0 else ''
im, m = plot_img(ds.lon, ds.lat, arr, cbrange=cbr, cmap=cmap, title=title, fontsize=fontsize,
plot_cmap=False, return_map=True)
plt.ylabel(ylabel, fontsize=fontsize)
x, y = m(-59, -48)
plt.text(x, y, f'm(+) = {np.nanmean(arr[arr>0]):.2f}', fontsize=fontsize)
x, y = m(-59, -51)
plt.text(x, y, f'm(a) = {np.nanmean(arr):.2f}', fontsize=fontsize)
x, y = m(-59, -54)
plt.text(x, y, f'm(-) = {np.nanmean(arr[arr<0]):.2f}', fontsize=fontsize)
plot_centered_cbar(f, im, len(lags), fontsize=fontsize, pad=0.02, wdth=0.02, hspace=0.01)
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/lagged_corr_anom_{met}.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
def plot_lagged_corr_spatial_LAI_VOD_SIF():
res = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/lagged_corr_w_sif.csv', index_col=0)
ds = io('LAI')
lut = ds.lut.reindex(res.index)
fontsize = 12
lags = [-2, -1, 0, 1, 2]
cbr = (-1,1)
cmap = cc.cm.bjy
f = plt.figure(figsize=(20, 12))
plt.suptitle(f'Lagged correlation (SIF) [-]', fontsize=fontsize+2, y=0.93)
for i, var in enumerate(['LAI_SIF', 'VOD_SIF']):
for j, lag in enumerate(lags):
tag = f'R_{var}_{lag}'
arr = np.full(ds.lat.shape, np.nan)
arr[lut.row_ease, lut.col_ease] = res[tag]
plt.subplot(2, len(lags), i*len(lags) + j + 1)
title = f'lag: {lag} months' if i == 0 else ''
ylabel = var.split('_')[0] if j == 0 else ''
im, m = plot_img(ds.lon, ds.lat, arr, cbrange=cbr, cmap=cmap, title=title, fontsize=fontsize,
plot_cmap=False, return_map=True)
plt.ylabel(ylabel, fontsize=fontsize)
x, y = m(-59, -48)
plt.text(x, y, f'm(+) = {np.nanmean(arr[arr>0]):.2f}', fontsize=fontsize)
x, y = m(-59, -51)
plt.text(x, y, f'm(a) = {np.nanmean(arr):.2f}', fontsize=fontsize)
x, y = m(-59, -54)
plt.text(x, y, f'm(-) = {np.nanmean(arr[arr<0]):.2f}', fontsize=fontsize)
plot_centered_cbar(f, im, len(lags), fontsize=fontsize, pad=0.02, wdth=0.02, hspace=0.01)
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/lagged_corr_SIF.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
def calc_trend_breakpoints():
fout = Path('/Users/u0116961/Documents/work/deforestation_paper/breakpoints.csv')
ds_lai = io('LAI')
ds_vod = io('SMOS_IC')
date_from = '2010-01-01'
date_to = '2019-12-31'
x0s = np.arange(2010, 2020)
rss_cols = [f'RSS_LAI_{x0}' for x0 in x0s] + [f'RSS_VOD_{x0}' for x0 in x0s] + \
['bfast_lai_b1', 'bfast_lai_b2', 'bfast_lai_m1', 'bfast_lai_m2' ] + \
['bfast_vod_b1', 'bfast_vod_b2', 'bfast_vod_m1', 'bfast_vod_m2' ]
start_date = datetime(2012, 1, 1)
bfast = BFASTMonitor(
start_date,
freq=365,
k=3,
hfrac=0.25,
trend=False,
level=0.05,
verbose=0,
backend='python',
device_id=0,
)
for i, _ in ds_lai.lut.iterrows():
print(f'{i} / {len(ds_lai.lut)}')
lai = ds_lai.read('LAI', i, date_from=date_from, date_to=date_to).dropna()
vod = ds_vod.read('VOD', i, date_from=date_from, date_to=date_to)
if len(vod) > 0:
invalid = (ds_vod.read('Flags', i, date_from=date_from, date_to=date_to) > 0) | \
(ds_vod.read('RMSE', i, date_from=date_from, date_to=date_to) > 8) | \
(ds_vod.read('VOD_StdErr', i, date_from=date_from, date_to=date_to) > 1.2)
vod[invalid] = np.nan
vod = vod.dropna()
if (len(lai) < 1) | (len(vod) < 1):
continue
mlai = pd.DataFrame(normalize(lai.resample('M').mean()), columns=['LAI'])
mvod = pd.DataFrame(normalize(vod.resample('M').mean()), columns=['VOD'])
rss = pd.DataFrame(columns=rss_cols, index=(i,))
data = np.reshape(mvod.values, (len(mvod), 1, 1))
dates = mvod.index.to_pydatetime()
bfast.fit(data, dates, n_chunks=5, nan_value=-32768)
rss.loc[i,'bfast_vod_m1'] = bfast.magnitudes[0][0]
if bfast.breaks[0][0] >= 0:
dt = dates[dates >= datetime(2011, 1, 1)][bfast.breaks[0][0]]
rss.loc[i,'bfast_vod_b1'] = dt.year + dt.month/12
if len(bfast.breaks) > 1:
dt = dates[dates > datetime(2011, 1, 1)][bfast.breaks.flatten()[1]]
rss.loc[i,'bfast_vod_b2'] = dt.year + dt.month/12
rss.loc[i,'bfast_vod_m2'] = bfast.magnitudes.flatten()[1]
data = np.reshape(mlai.values, (len(mlai), 1, 1))
dates = mlai.index.to_pydatetime()
bfast.fit(data, dates, n_chunks=None, nan_value=-32768)
rss.loc[i,'bfast_lai_m1'] = bfast.magnitudes[0][0]
if bfast.breaks[0][0] >= 0:
dt = dates[dates >= start_date][bfast.breaks[0][0]]
rss.loc[i,'bfast_lai_b1'] = dt.year + dt.month/12
# if len(bfast.breaks) > 1:
# dt = dates[dates > datetime(2011, 1, 1)][bfast.breaks.flatten()[1]]
# rss.loc[i,'bfast_lai_b2'] = dt.year + dt.month/12
# rss.loc[i,'bfast_lai_m2'] = bfast.magnitudes.flatten()[1]
for x0 in x0s:
mlai[f'LAI_trend_{x0}'] = np.nan
mvod[f'VOD_trend_{x0}'] = np.nan
lai_l = mlai[date_from:f'{x0}']['LAI']
lai_u = mlai[f'{x0 + 1}':date_to]['LAI']
vod_l = mvod[date_from:f'{x0}']['VOD']
vod_u = mvod[f'{x0 + 1}':date_to]['VOD']
for lai, vod in zip([lai_l, lai_u], [vod_l, vod_u]):
try:
mk_lai = mk.original_test(lai)
mk_vod = mk.original_test(vod)
x = (lai.index.year - lai.index.year.min()) * 12 + lai.index.month
mlai.loc[lai.index, f'LAI_trend_{x0}'] = mk_lai.intercept + mk_lai.slope * x.values
x = (vod.index.year - vod.index.year.min()) * 12 + vod.index.month
mvod.loc[vod.index, f'VOD_trend_{x0}'] = mk_vod.intercept + mk_vod.slope * x.values
except:
continue
rss.loc[i, f'RSS_LAI_{x0}'] = ((mlai[f'LAI'] - mlai[f'LAI_trend_{x0}']) ** 2).sum()
rss.loc[i, f'RSS_VOD_{x0}'] = ((mvod[f'VOD'] - mvod[f'VOD_trend_{x0}']) ** 2).sum()
if fout.exists():
rss.to_csv(fout, float_format='%0.4f', mode='a', header=False)
else:
rss.to_csv(fout, float_format='%0.4f')
def calc_trends():
fout = Path('/Users/u0116961/Documents/work/deforestation_paper/trends.csv')
ds_lai = io('LAI')
ds_vod = io('SMOS_IC')
ds_met = io('MERRA2')
date_from = '2010-01-01'
date_to = '2019-12-31'
for i, val in ds_lai.lut.iterrows():
print(f'{i} / {len(ds_lai.lut)}')
lai = ds_lai.read('LAI', i, date_from=date_from, date_to=date_to).dropna()
if len(lai) == 0:
continue
vod = ds_vod.read('VOD', i, date_from=date_from, date_to=date_to)
if len(vod) > 0:
invalid = (ds_vod.read('Flags', i, date_from=date_from, date_to=date_to) > 0) | \
(ds_vod.read('RMSE', i, date_from=date_from, date_to=date_to) > 8) | \
(ds_vod.read('VOD_StdErr', i, date_from=date_from, date_to=date_to) > 1.2)
vod[invalid] = np.nan
vod = vod.dropna()
if len(vod) == 0:
continue
temp = ds_met.read('T2M', i, date_from=date_from, date_to=date_to)
prec = ds_met.read('PRECTOTLAND', i, date_from=date_from, date_to=date_to)
rad = ds_met.read('LWLAND', i, date_from=date_from, date_to=date_to) + \
ds_met.read('SWLAND', i, date_from=date_from, date_to=date_to)
if len(temp)+len(prec)+len(rad) == 0:
continue
res = pd.DataFrame(index=(i,))
mk_lai = mk.original_test(normalize(lai.resample('M').mean()))
mk_vod = mk.original_test(normalize(vod.resample('M').mean()))
mk_rad = mk.original_test(normalize(rad.resample('M').mean()))
mk_temp = mk.original_test(normalize(temp.resample('M').mean()))
mk_prec = mk.original_test(normalize(prec.resample('M').sum()))
mk_vod_orig = mk.original_test(vod.resample('M').mean())
res[f'slope_lai'] = mk_lai.slope
res[f'slope_vod'] = mk_vod.slope
res[f'slope_rad'] = mk_rad.slope
res[f'slope_temp'] = mk_temp.slope
res[f'slope_prec'] = mk_prec.slope
res[f'slope_vod_orig'] = mk_vod_orig.slope
res[f'p_lai'] = mk_lai.p
res[f'p_vod'] = mk_vod.p
res[f'p_rad'] = mk_rad.p
res[f'p_temp'] = mk_temp.p
res[f'p_prec'] = mk_prec.p
res[f'p_vod_orig'] = mk_vod_orig.p
if fout.exists():
res.to_csv(fout, float_format='%0.8f', mode='a', header=False)
else:
res.to_csv(fout, float_format='%0.8f')
def plot_lagged_corr_boxplot():
res = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/lagged_corr_w_sif.csv', index_col=0)
ds = io('LAI')
lut = ds.lut.reindex(res.index)
variables = ['T', 'P', 'R']
titles = ['Temperature', 'Precipitation', 'Radiation']
datasets = ['LAI', 'SIF', 'VOD']
lags = np.arange(-6, 1, 1)
pos = [i + j for i in np.arange(1, len(variables) + 1) for j in np.linspace(-0.4, 0.4, len(lags))]
colors = [f'{s}' for n in np.arange(len(variables)) for s in np.linspace(0.2, 0.98, len(lags))]
# colors = [s for n in np.arange(len(variables)) for s in ['lightblue', 'lightgreen', 'coral']]
f = plt.figure(figsize=(13, 14))
fontsize = 14
for n, d in enumerate(datasets):
ax = plt.subplot(3, 1, n+1)
if n == 0:
axpos = ax.get_position()
data = list()
for v, t in zip(variables, titles):
for lag in lags:
var = f'R_anom_{d}_{v}_{lag}'
tmp = res[var].values
# tmp = tmp[tmp>0]
data.append(tmp[~np.isnan(tmp)])
box = ax.boxplot(data, whis=[10, 90], showfliers=False, positions=pos, widths=0.07, patch_artist=True)
for patch, color in zip(box['boxes'], colors):
patch.set(color='black', linewidth=2)
patch.set_facecolor(color)
for patch in box['medians']:
patch.set(color='black', linewidth=2)
for patch in box['whiskers']:
patch.set(color='black', linewidth=1)
plt.xlim(0.5, len(variables)+0.5)
if n < 2:
plt.xticks([], [], fontsize=fontsize)
else:
plt.xticks(np.arange(1,len(variables)+1), titles , fontsize=fontsize)
plt.ylim(-0.5,0.5)
plt.yticks(np.arange(-0.5,0.5,0.2))
# plt.ylim(0,0.9)
# plt.yticks(np.arange(0,1,0.2))
plt.axhline(color='k', linestyle='--', linewidth=1)
for i in np.arange(1,len(variables)):
plt.axvline(i + 0.5, linewidth=1.5, color='k')
plt.ylabel(d, fontsize=fontsize)
if n==0:
plt.title('Lagged correlation [-] (anomalies)',fontsize=fontsize)
f.subplots_adjust(hspace=0.1)
plt.figlegend((box['boxes'][0:len(lags)]), lags, 'upper right', title='Lag [m]',
bbox_to_anchor=(axpos.x1+0.09,axpos.y1+0.013), fontsize=fontsize-2)
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/lagged_corr_anom_boxplot_w_sif.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
def calc_lagged_corr():
fout = Path('/Users/u0116961/Documents/work/deforestation_paper/lagged_corr_w_sif.csv')
ds_lai = io('LAI')
ds_vod = io('SMOS_IC')
ds_met = io('MERRA2')
ds_sif = io('SIF')
date_from = '2010-01-01'
date_to = '2019-12-31'
for i, val in ds_lai.lut.iterrows():
print(f'{i} / {len(ds_lai.lut)}')
lai = ds_lai.read('LAI', i, date_from=date_from, date_to=date_to).dropna()
if len(lai) == 0:
continue
vod = ds_vod.read('VOD', i, date_from=date_from, date_to=date_to)
if (len(lai)>0) & (len(vod)>0):
invalid = (ds_vod.read('Flags', i, date_from=date_from, date_to=date_to) > 0) | \
(ds_vod.read('RMSE', i, date_from=date_from, date_to=date_to) > 8) | \
(ds_vod.read('VOD_StdErr', i, date_from=date_from, date_to=date_to) > 1.2)
vod[invalid] = np.nan
vod = vod.dropna()
if len(vod) == 0:
continue
sif = ds_sif.read('sif_dc', i, date_from=date_from, date_to=date_to)
invalid = (ds_sif.read('n', i, date_from=date_from, date_to=date_to) <= 1) | \
(ds_sif.read('cloud_fraction', i, date_from=date_from, date_to=date_to) > 0.7)
sif[invalid] = np.nan
sif = sif.dropna()
if len(sif) == 0:
continue
df_veg = pd.concat((lai, vod, sif), axis=1, keys=['LAI', 'VOD', 'SIF']).resample('M').mean().dropna()
for col in df_veg:
df_veg[f'{col}_anom'] = calc_anomaly(df_veg[col], method='harmonic', longterm=True, n=3)
temp = ds_met.read('T2M', i, date_from=date_from, date_to=date_to)
prec = ds_met.read('PRECTOTLAND', i, date_from=date_from, date_to=date_to)
rad = ds_met.read('LWLAND', i, date_from=date_from, date_to=date_to) + \
ds_met.read('SWLAND', i, date_from=date_from, date_to=date_to)
df_met = pd.concat((temp,prec,rad), axis=1, keys=['T','P','R']).resample('M').mean().dropna()
if len(df_met) == 0:
continue
df_met['T_anom'] = calc_anomaly(df_met['T'], method='harmonic', longterm=True, n=3)
df_met['P_anom'] = calc_anomaly(df_met['P'], method='harmonic', longterm=True, n=3)
df_met['R_anom'] = calc_anomaly(df_met['R'], method='harmonic', longterm=True, n=3)
tmp_df_met = df_met.copy()
tmp_df_veg = df_veg.reindex(df_met.index).copy()
tmp_df_veg.columns = tmp_df_veg.columns + '_nolag'
tmp_df_met = pd.concat((tmp_df_met, tmp_df_veg), axis=1)
tmp_df_met.index = np.arange(len(df_met))
res = pd.DataFrame(index=(i,))
for lag in np.arange(-6,7):
tmp_df_veg = df_veg.reindex(df_met.index)
tmp_df_veg.index = np.arange(len(tmp_df_veg)) + lag
corr = pd.concat((tmp_df_met, tmp_df_veg), axis=1).corr()
res[f'R_LAI_T_{lag}'] = corr['R']['LAI']
res[f'R_LAI_P_{lag}'] = corr['T']['LAI']
res[f'R_LAI_R_{lag}'] = corr['P']['LAI']
res[f'R_VOD_T_{lag}'] = corr['R']['VOD']
res[f'R_VOD_P_{lag}'] = corr['T']['VOD']
res[f'R_VOD_R_{lag}'] = corr['P']['VOD']
res[f'R_SIF_T_{lag}'] = corr['R']['SIF']
res[f'R_SIF_P_{lag}'] = corr['T']['SIF']
res[f'R_SIF_R_{lag}'] = corr['P']['SIF']
res[f'R_LAI_VOD_{lag}'] = corr['LAI_nolag']['VOD']
res[f'R_LAI_SIF_{lag}'] = corr['LAI_nolag']['SIF']
res[f'R_VOD_SIF_{lag}'] = corr['VOD_nolag']['SIF']
res[f'R_anom_LAI_T_{lag}'] = corr['R_anom']['LAI_anom']
res[f'R_anom_LAI_P_{lag}'] = corr['T_anom']['LAI_anom']
res[f'R_anom_LAI_R_{lag}'] = corr['P_anom']['LAI_anom']
res[f'R_anom_VOD_T_{lag}'] = corr['R_anom']['VOD_anom']
res[f'R_anom_VOD_P_{lag}'] = corr['T_anom']['VOD_anom']
res[f'R_anom_VOD_R_{lag}'] = corr['P_anom']['VOD_anom']
res[f'R_anom_SIF_T_{lag}'] = corr['R_anom']['SIF_anom']
res[f'R_anom_SIF_P_{lag}'] = corr['T_anom']['SIF_anom']
res[f'R_anom_SIF_R_{lag}'] = corr['P_anom']['SIF_anom']
res[f'R_anom_LAI_VOD_{lag}'] = corr['LAI_anom_nolag']['VOD_anom']
res[f'R_anom_LAI_SIF_{lag}'] = corr['LAI_anom_nolag']['SIF_anom']
res[f'R_anom_VOD_SIF_{lag}'] = corr['VOD_anom_nolag']['SIF_anom']
if fout.exists():
res.to_csv(fout, float_format='%0.4f', mode='a', header=False)
else:
res.to_csv(fout, float_format='%0.4f')
def mean_veg_met():
ds_lai = io('LAI')
ds_vod = io('SMOS_IC')
ds_met = io('MERRA2')
date_from = '2010-01-01'
date_to = '2019-12-31'
lai = np.nanmean(ds_lai.read_img('LAI', date_from=date_from, date_to=date_to), axis=0)
vod = ds_vod.read_img('VOD', date_from=date_from, date_to=date_to)
invalid = (ds_vod.read_img('Flags', date_from=date_from, date_to=date_to) > 0) | \
(ds_vod.read_img('RMSE', date_from=date_from, date_to=date_to) > 8) | \
(ds_vod.read_img('VOD_StdErr', date_from=date_from, date_to=date_to) > 1.2)
vod[invalid] = np.nan
vod = np.nanmean(vod, axis=0)
temp = np.nanmean(ds_met.read_img('T2M', date_from=date_from, date_to=date_to), axis=0)
prec = np.nanmean(ds_met.read_img('PRECTOTLAND', date_from=date_from, date_to=date_to), axis=0)
rad = np.nanmean(ds_met.read_img('LWLAND', date_from=date_from, date_to=date_to)+
ds_met.read_img('SWLAND', date_from=date_from, date_to=date_to), axis=0)
temp_ease = np.full(lai.shape, np.nan)
prec_ease = np.full(lai.shape, np.nan)
rad_ease = np.full(lai.shape, np.nan)
temp_ease[ds_lai.lut.row_ease, ds_lai.lut.col_ease] = temp[ds_lai.lut.row_merra, ds_lai.lut.col_merra]
prec_ease[ds_lai.lut.row_ease, ds_lai.lut.col_ease] = prec[ds_lai.lut.row_merra, ds_lai.lut.col_merra]
rad_ease[ds_lai.lut.row_ease, ds_lai.lut.col_ease] = rad[ds_lai.lut.row_merra, ds_lai.lut.col_merra]
# Collocation
invalid = np.where(np.isnan(vod) | np.isnan(lai) | np.isnan(temp_ease) | np.isnan(prec_ease) | np.isnan(rad_ease) )
lai[invalid] = np.nan
vod[invalid] = np.nan
temp_ease[invalid] = np.nan
prec_ease[invalid] = np.nan
rad_ease[invalid] = np.nan
fontsize = 12
f = plt.figure(figsize=(18,6))
# gs = gridspec.GridSpec(3, 3, height_ratios=[2,1,0.025], wspace=0.35, hspace=0.20)
plt.subplot(1, 5, 1)
plot_img(ds_lai.lon, ds_lai.lat, lai, cbrange=(0,6), cmap='viridis_r', title='LAI [-]', fontsize=fontsize)
plt.subplot(1, 5, 2)
plot_img(ds_lai.lon, ds_lai.lat, vod, cbrange=(0,0.8), cmap='viridis_r', title='VOD [-]', fontsize=fontsize)
plt.subplot(1, 5, 3)
plot_img(ds_lai.lon, ds_lai.lat, temp_ease, cbrange=(5, 30), cmap='viridis_r', title='T [$^\circ$C]', fontsize=fontsize)
plt.subplot(1, 5, 4)
plot_img(ds_lai.lon, ds_lai.lat, prec_ease * 1e3, cbrange=(0,2), cmap='viridis_r', title='P [kg / m2 / d] * 1e-2', fontsize=fontsize)
plt.subplot(1, 5, 5)
plot_img(ds_lai.lon, ds_lai.lat, rad_ease, cbrange=(60,160), cmap='viridis_r', title='Rad. [W / m2]', fontsize=fontsize)
# plt.tight_layout()
# plt.show()
fname = '/Users/u0116961/Documents/work/deforestation_paper/plots/mean_veg_met.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
def calc_pca_classes():
ds_met = io('MERRA2')
date_from = '2010-01-01'
date_to = '2019-12-31'
temp = np.nanpercentile(ds_met.read_img('T2M', date_from=date_from, date_to=date_to), [2, 25, 50, 75, 98], axis=0).reshape((5,-1))
prec = np.nanpercentile(ds_met.read_img('PRECTOTLAND', date_from=date_from, date_to=date_to), [2, 25, 50, 75, 98], axis=0).reshape((5,-1))
rad = np.nanpercentile(ds_met.read_img('LWLAND', date_from=date_from, date_to=date_to) +
ds_met.read_img('SWLAND', date_from=date_from, date_to=date_to), [2, 25, 50, 75, 98], axis=0).reshape((5,-1))
df = pd.DataFrame({'t_med': temp[2,:],
'p_med': prec[2,:],
'r_med': rad[2,:],
't_iqr': temp[3, :] - temp[1, :],
'p_iqr': prec[3, :] - prec[1, :],
'r_iqr': rad[3, :] - rad[1, :],
't_rng': temp[4, :] - temp[0, :],
'p_rng': prec[4, :] - prec[0, :],
'r_rng': rad[4, :] - rad[0, :]})
df = df.iloc[:,0:6]
df_nonan = df.dropna()
pca = PCA(n_components=len(df_nonan.columns))
df_nonan.loc[:,:] = pca.fit_transform(StandardScaler().fit_transform(df_nonan)).copy()
df_nonan.columns = np.arange(len(df_nonan.columns))+1
n_cl = 3
tmp = pd.concat((df_nonan.iloc[:,0:n_cl], df_nonan.iloc[:,0:n_cl]*-1), axis=1)
tmp.columns = np.arange(len(tmp.columns))+1
cl = tmp.idxmax(axis=1).reindex(df.index)
cl.name = 'PCA_class'
cl.to_csv(f'/Users/u0116961/Documents/work/deforestation_paper/PCA_classes.csv')
# ----- PLOT CLASSIFICATION -----
f = plt.figure(figsize=(6,8))
cm = colors.ListedColormap(sns.color_palette('muted',n_colors=n_cl*2))
# plt.subplot(2, 6, 7+n_cl+1)
arr = cl.values.reshape(ds_met.lat.shape)
im, cb = plot_img(ds_met.lon, ds_met.lat, arr, cbrange=[1, n_cl*2+1], cmap=cm, title='PCA classes', fontsize=16,
plot_cmap=True, return_cbar=True)
cb.set_ticks(np.arange(6) + 1.5)
cb.set_ticklabels(['1(+)', '2(+)', '3(+)', '1(-)', '2(-)', '3(-)'])
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/PCA_classes.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
df_pca = df_nonan.reindex(df.index)
df_pca.columns = [f'PCA-{i}' for i in np.arange(len(df_nonan.columns))+1]
f = plt.figure(figsize=(8,6))
exp_var = [pca.explained_variance_ratio_[0:i+1].sum() for i in range(len(pca.explained_variance_ratio_))]
print(pd.DataFrame(pca.components_,index=df_pca.columns, columns=df.columns.values).T)
plt.plot(np.arange(len(df_nonan.columns))+1, exp_var)
plt.plot([0,3],[exp_var[2], exp_var[2] ], color='black', linestyle='--', linewidth='1')
plt.xticks(np.arange(len(df_nonan.columns))+1, df_pca.columns.values)
plt.ylim(0.55,1.02)
plt.xlim(0.8,6.2)
plt.title('Cumulative fraction of variance explained')
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/PCA_var_explained.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
# ----- PLOT VARIABLE CONTRIBUTIONS -----
f, axes = plt.subplots(1, len(df_pca.columns), sharex=True, sharey=True, figsize=(12, 5))
variables = ['T (median)', 'P (median)', 'R (median)', 'T (IQR)', 'P (IQR)', 'R (IQR)']
n_vars = len(variables)
for i, ax in enumerate(axes):
ax.barh(np.arange(n_vars), (pca.components_[i, :][::-1]), orientation="horizontal", color="teal")
ax.set_yticks(np.arange(n_vars))
ax.set_title(f"PCA-{i + 1}", fontsize=12)
ax.set_xlim(-0.8,0.8)
ax.axvline(color='black', linestyle='--', linewidth=1)
ax.grid(c="lightgrey", linewidth=0.5, which="major")
ax.grid(c="lightgrey", linewidth=0.5, which="minor")
axes[0].set_yticklabels(variables[::-1])
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/PCA_contribution.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
# ----- PLOT VARS -----
f = plt.figure(figsize=(25,12))
fontsize = 18
for i, (col, tit) in enumerate(zip(df,variables)):
if i > 5:
continue
plt.subplot(2, 6, i+1)
arr = df[col].values.reshape(ds_met.lat.shape)
im = plot_img(ds_met.lon, ds_met.lat, arr, cbrange=None, cmap='viridis', title=tit, fontsize=fontsize,
plot_cmap=True)
# ----- PLOT FIRST FEW PCAS -----
for i, col in enumerate(df_pca):
# if i > n_cl:
# continue
plt.subplot(2, 6, i+7)
arr = df_pca[col].values.reshape(ds_met.lat.shape)
im = plot_img(ds_met.lon, ds_met.lat, arr, cbrange=[-2,2], cmap=cc.cm.bjy, title=col, fontsize=fontsize,
plot_cmap=True)
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/PCA_maps.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
def AGB_VOD_LAI():
print_agb_vod_lai = True
print_agb_change = False
print_agb_tcl = False
ds_agb = io('AGB')
ds_vod = io('SMOS_IC')
ds_met = io('MERRA2')
ds_lai = io('LAI')
ds_tcl = io('TCL')
cl = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/PCA_classes.csv', index_col=0)
cl_arr = np.full(ds_agb.lat.shape, np.nan)
cl_arr[ds_agb.lut.row_ease,ds_agb.lut.col_ease] = cl.values.reshape(ds_met.lat.shape)[ds_agb.lut.row_merra,ds_agb.lut.col_merra]
# cl = pd.Series(cl.values.reshape(ds_met.lat.shape)[ds_agb.lut.row_merra,ds_agb.lut.col_merra],
# index=ds_agb.lut.index)
trends = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/trends.csv', index_col=0)
lut = ds_agb.lut.reindex(trends.index)
sig = np.full(ds_agb.lat.shape, 1.)
sig[lut[trends[f'p_vod'] > 0.05].row_ease, lut[trends[f'p_vod'] > 0.05].col_ease] = np.nan
corr = pd.read_csv('/Users/u0116961/Documents/work/deforestation_paper/lagged_corr_w_sif.csv', index_col=0)
lut2 = ds_agb.lut.reindex(corr.index)
t = np.full(ds_agb.lat.shape, np.nan)
p = np.full(ds_agb.lat.shape, np.nan)
r = np.full(ds_agb.lat.shape, np.nan)
t[lut.row_ease, lut.col_ease] = corr[f'R_VOD_T_-3']
p[lut.row_ease, lut.col_ease] = corr[f'R_VOD_P_-3']
r[lut.row_ease, lut.col_ease] = corr[f'R_VOD_R_-2']
mask = np.full(ds_agb.lat.shape, np.nan)
mask[(t < 0.5) & (p < 0.5) & (r < 0.5)] = 1
tcl = ds_tcl.read_img('TCL')
agb = ds_agb.read_img('AGB')
agb_err = ds_agb.read_img('AGB_err')
agb[agb_err > 160] = np.nan
date_from = '2010-01-01'
date_to = '2010-12-31'
lai = np.nanmean(ds_lai.read_img('LAI', date_from=date_from, date_to=date_to), axis=0)
vod = ds_vod.read_img('VOD', date_from=date_from, date_to=date_to)
invalid = (ds_vod.read_img('Flags', date_from=date_from, date_to=date_to) > 0) | \
(ds_vod.read_img('RMSE', date_from=date_from, date_to=date_to) > 8) | \
(ds_vod.read_img('VOD_StdErr', date_from=date_from, date_to=date_to) > 1.2)
vod[invalid] = np.nan
vod2010 = np.nanmean(vod, axis=0)
date_from = '2019-01-01'
date_to = '2019-12-31'
vod = ds_vod.read_img('VOD', date_from=date_from, date_to=date_to)
invalid = (ds_vod.read_img('Flags', date_from=date_from, date_to=date_to) > 0) | \
(ds_vod.read_img('RMSE', date_from=date_from, date_to=date_to) > 8) | \
(ds_vod.read_img('VOD_StdErr', date_from=date_from, date_to=date_to) > 1.2)
vod[invalid] = np.nan
vod2019 = np.nanmean(vod, axis=0)
# sif = ds_sif.read_img('sif_dc', date_from=date_from, date_to=date_to)
# invalid = (ds_sif.read_img('n', date_from=date_from, date_to=date_to) <= 1) | \
# (ds_sif.read_img('cloud_fraction', date_from=date_from, date_to=date_to) > 0.7)
# sif[invalid] = np.nan
# sif = np.nanmean(sif, axis=0)
# sif_ease = np.full(vod2019.shape, np.nan)
# sif_ease[ds_sif.lut.row_ease, ds_sif.lut.col_ease] = sif[ds_sif.lut.row_sif, ds_sif.lut.col_sif]
# Collocation
invalid = np.where(np.isnan(agb) | np.isnan(vod2010) | np.isnan(vod2019) | np.isnan(lai))
agb[invalid] = np.nan
lai[invalid] = np.nan
vod2010[invalid] = np.nan
vod2019[invalid] = np.nan
tcl[invalid] = np.nan
for cls in np.unique(cl_arr[~np.isnan(cl_arr)]):
tmp_agb = agb.copy()
tmp_lai = lai.copy()
tmp_vod2010 = vod2010.copy()
tmp_agb[cl_arr != cls] = np.nan
tmp_lai[cl_arr != cls] = np.nan
tmp_vod2010[cl_arr != cls] = np.nan
valid = np.where(~np.isnan(tmp_agb))
fontsize = 10
if print_agb_vod_lai:
f = plt.figure(figsize=(16,11))
gs = gridspec.GridSpec(3, 4, height_ratios=[2,1,0.025], wspace=0.2, hspace=0.20)
plt.subplot(gs[0,0])
plot_img(ds_agb.lon, ds_agb.lat, tmp_agb, cbrange=(0,100), cmap='viridis_r', title='AGB [Mg/ha]', fontsize=fontsize)
plt.subplot(gs[0,1])
plot_img(ds_agb.lon, ds_agb.lat, tmp_lai, cbrange=(0,5), cmap='viridis_r', title='LAI [-]', fontsize=fontsize)
plt.subplot(gs[0,2])
plot_img(ds_agb.lon, ds_agb.lat, tmp_vod2010, cbrange=(0,0.8), cmap='viridis_r', title='VOD [-]', fontsize=fontsize)
p_order = 3
ax = plt.subplot(gs[1,1])
x, y = tmp_lai[valid], tmp_agb[valid]
ax.hexbin(x, y,
gridsize=35, bins='log',
cmap='viridis', mincnt=1)
xs = np.linspace(x.min(),x.max(), 100)
p = np.poly1d(np.polyfit(x, y, p_order))
# xlim = ax.get_xlim()
# ylim = ax.get_ylim()
xlim = [-0.2,6.2]
ylim = [-20,330]
plt.plot(xs, p(xs), linestyle='-.', linewidth=1, color='k')
plt.xlim(xlim)
plt.ylim(ylim)
corr = np.corrcoef(x, y)[0,1]
plt.title(f'R = {corr:.3f}', fontsize=fontsize)
plt.xlabel('LAI [-]', fontsize=fontsize)
plt.ylabel('AGB [Mg/ha]', fontsize=fontsize)
plt.xticks(fontsize=fontsize)
plt.yticks(fontsize=fontsize)
ax = plt.subplot(gs[1,2])
x, y = tmp_vod2010[valid], tmp_agb[valid]
ax.hexbin(x, y,
gridsize=35, bins='log',
cmap='viridis', mincnt=1)
xs = np.linspace(x.min(),x.max(), 100)
p = np.poly1d(np.polyfit(x, y, p_order))
# xlim = ax.get_xlim()
# ylim = ax.get_ylim()
xlim = [-0.2,1.4]
ylim = [-20,330]
plt.plot(xs, p(xs), linestyle='-.', linewidth=1, color='k')
plt.xlim(xlim)
plt.ylim(ylim)
corr = np.corrcoef(x, y)[0,1]
plt.title(f'R = {corr:.3f}', fontsize=fontsize)
plt.xlabel('VOD [-]', fontsize=fontsize)
# plt.ylabel('AGB [Mg/ha]', fontsize=fontsize)
plt.xticks(fontsize=fontsize)
plt.yticks([])
fname = f'/Users/u0116961/Documents/work/deforestation_paper/plots/AGB_LAI_VOD_cl{int(cls)}.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
p = np.poly1d(np.polyfit(vod2010[valid], agb[valid], 3))
diff = p(vod2019) - p(vod2010)
# Various masking options
# diff[np.isnan(sig)] = np.nan
# tcl[np.isnan(sig)] = np.nan
# diff[np.isnan(mask)] = np.nan
# tcl[np.isnan(mask)] = np.nan
# tcl[diff > 0] = np.nan
# diff[diff > 0] = np.nan
diff[tcl < 0.1] = np.nan
tcl[tcl < 0.1] = np.nan
valid = np.where(~np.isnan(diff))
print(np.corrcoef(diff[valid], tcl[valid]))
if print_agb_change:
f = plt.figure(figsize=(12,10))
plt.subplot(1, 2, 1)
plot_img(ds_agb.lon, ds_agb.lat, diff / 9, cbrange=(-4,4), cmap=cc.cm.bjy,
title='$\Delta$ AGB [Mg/ha/yr] (2019 - 2010)', fontsize=fontsize + 4)
plt.subplot(1, 2, 2)
plot_img(ds_agb.lon, ds_agb.lat, tcl, cbrange=(0.1 , 0.6), cmap=truncate_colormap(cc.cm.bjy_r,0.5,1), title='Tree cover loss [-] (2019 - 2010)', fontsize=fontsize+4)
fname = '/Users/u0116961/Documents/work/deforestation_paper/plots/AGB_change_2019_2010_diff_lt_0_p_lt_0.05.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
if print_agb_tcl:
f = plt.figure(figsize=(12,10))
fontsize += 6
x, y = diff[valid], tcl[valid]
plt.hexbin(x, y,
gridsize=100, bins='log',
cmap='viridis', mincnt=1, )
plt.axvline(color='k', linestyle='--', linewidth=1.5)
corr = np.corrcoef(x, y)[0,1]
plt.title(f'R = {corr:.3f}', fontsize=fontsize)
plt.xlabel('$\Delta$ AGB', fontsize=fontsize)
plt.ylabel('Tree cover loss', fontsize=fontsize)
plt.xticks(fontsize=fontsize)
plt.yticks(fontsize=fontsize)
plt.xlim([-50,50])
plt.ylim(0,0.8)
fname = '/Users/u0116961/Documents/work/deforestation_paper/plots/AGB_diff_vs_TCL_tcl_gt_0.1.png'
f.savefig(fname, dpi=300, bbox_inches='tight')
plt.close()
def plot_trend_ts():
ds_lai = io('LAI')
ds_vod = io('SMOS_IC')
date_from = '2010-01-01'
date_to = '2019-12-31'
lat, lon = -22.87578546915353, -62.311473240766794
i = np.argmin((ds_lai.lut.lat - lat) ** 2 + (ds_lai.lut.lon - lon) ** 2)
print(i)
# i = 3000
lai = ds_lai.read('LAI', i, date_from=date_from, date_to=date_to).dropna()
vod = ds_vod.read('VOD', i, date_from=date_from, date_to=date_to)
if len(vod) > 0:
invalid = (ds_vod.read('Flags', i, date_from=date_from, date_to=date_to) > 0) | \
(ds_vod.read('RMSE', i, date_from=date_from, date_to=date_to) > 8) | \
(ds_vod.read('VOD_StdErr', i, date_from=date_from, date_to=date_to) > 1.2)
vod[invalid] = np.nan
vod = vod.dropna()
mlai = pd.DataFrame(normalize(lai.resample('M').mean()), columns=['LAI'])
mvod = pd.DataFrame(normalize(vod.resample('M').mean()), columns=['VOD'])
x0s = np.arange(2010, 2020)
rss_lai = pd.DataFrame(columns=['RSS_LAI'], index=x0s)
rss_vod = pd.DataFrame(columns=['RSS_VOD'], index=x0s)
for x0 in x0s:
mlai[f'LAI_trend_{x0}'] = np.nan
mvod[f'VOD_trend_{x0}'] = np.nan
lai_l = mlai[date_from:f'{x0}']['LAI']
lai_u = mlai[f'{x0 + 1}':date_to]['LAI']
vod_l = mvod[date_from:f'{x0}']['VOD']
vod_u = mvod[f'{x0 + 1}':date_to]['VOD']
for lai, vod in zip([lai_l, lai_u], [vod_l, vod_u]):
try:
mk_lai = mk.original_test(lai)
mk_vod = mk.original_test(vod)
x = (lai.index.year - lai.index.year.min()) * 12 + lai.index.month
mlai.loc[lai.index, f'LAI_trend_{x0}'] = mk_lai.intercept + mk_lai.slope * x.values
x = (vod.index.year - vod.index.year.min()) * 12 + vod.index.month
mvod.loc[vod.index, f'VOD_trend_{x0}'] = mk_vod.intercept + mk_vod.slope * x.values
except:
continue
rss_lai.loc[x0, 'RSS_LAI'] = ((mlai[f'LAI'] - mlai[f'LAI_trend_{x0}']) ** 2).sum()
rss_vod.loc[x0, 'RSS_VOD'] = ((mvod[f'VOD'] - mvod[f'VOD_trend_{x0}']) ** 2).sum()
f = plt.figure(figsize=(16, 10))
ax1 = plt.subplot(2, 2, 1)
mlai[['LAI', f'LAI_trend_201{np.argmin(rss_lai)}', 'LAI_trend_2019']].plot(ax=ax1, legend=False)
# plt.ylim(mlai['LAI'].min(), mlai['LAI'].max())
ax2 = plt.subplot(2, 2, 2)
mvod[['VOD', f'VOD_trend_201{np.argmin(rss_vod)}', 'VOD_trend_2019']].plot(ax=ax2, legend=False)
# plt.ylim(mvod['VOD'].min(), mvod['VOD'].max())
ax3 = plt.subplot(2, 2, 3)
rss_lai.plot(ax=ax3)
plt.xlim(2010, 2020)
ax4 = plt.subplot(2, 2, 4)
rss_vod.plot(ax=ax4)
# plt.xlim(mvod.index.min(), mvod.index.max())
plt.xlim(2010, 2020)
# plt.ylim(0,100)
plt.tight_layout()
plt.show()