slr/distributions.py

@@ -0,0 +1,130 @@
+"""Calculate probability distributions for IPCC sea level rise forecasts.
+
+This will calculate the values required to generate triangular distributions,
+i.e. 'min', 'mode', and 'max' in the `numpy.random.triang()` function.
+
+Reads:
+    'IPCC AR6.xlsx'
+
+Writes:
+    'triang-values.csv'
+
+D. Howe
+d.howe@wrl.unsw.edu.au
+2022-05-05
+"""
+import os
+import re
+import numpy as np
+import pandas as pd
+from scipy import stats, optimize
+import matplotlib.pyplot as plt
+
+PLOT = False
+
+
+def norm_cdf(x, loc, scale):
+    """Calculate cumulative density function, using normal distribution."""
+    return stats.norm(loc=loc, scale=scale).cdf(x)
+
+
+def triang_cdf(x, loc, scale, c):
+    """Calculate cumulative density function, using triangular distribution."""
+    return stats.triang(loc=loc, scale=scale, c=c).cdf(x)
+
+
+# Read data
+df = pd.read_excel('IPCC AR6.xlsx', index_col=[0, 1, 2, 3, 4])
+df = df.sort_index()
+dff = df.loc[838, 'total', 'medium', 'ssp585'].T
+dff.index.name = 'year'
+percentiles = dff.columns.values / 100
+
+for i, row in dff.iterrows():
+    values = row.values
+
+    # Fit normal distribution
+    loc, scale = optimize.curve_fit(norm_cdf, values, percentiles)[0]
+    p_norm = {'loc': loc, 'scale': scale}
+
+    # Fit triangular distribution
+    loc, scale, c = optimize.curve_fit(triang_cdf,
+                                       values,
+                                       percentiles,
+                                       p0=[values[0] - 0.1, 0.5, 0.5])[0]
+    p_triang = {'loc': loc, 'scale': scale, 'c': c}
+
+    # Get triangular distribution parameters
+    left = p_triang['loc']
+    centre = p_triang['loc'] + p_triang['scale'] * p_triang['c']
+    right = p_triang['loc'] + p_triang['scale']
+
+    dff.loc[i, 'min'] = left
+    dff.loc[i, 'mode'] = centre
+    dff.loc[i, 'max'] = right
+
+    if PLOT:
+        fig, ax = plt.subplots(1, 2, figsize=(10, 3))
+
+        x_min = stats.triang.ppf(0.01, **p_triang) - 0.2
+        x_max = stats.triang.ppf(0.99, **p_triang) + 0.2
+        x = np.linspace(x_min, x_max, num=1000)
+
+        ax[0].plot(x, 100 * stats.norm.cdf(x, **p_norm))
+        ax[0].plot(x, 100 * stats.triang.cdf(x, **p_triang))
+        ax[0].plot(values, 100 * percentiles, '.', c='#444444')
+
+        ax[1].plot(x, stats.norm.pdf(x, **p_norm), label='Normal')
+        ax[1].plot(x, stats.triang.pdf(x, **p_triang), label='Triangular')
+        ax[1].plot([], [], '.', c='#444444', label='IPCC data')
+        ax[1].legend()
+
+        ax[1].axvline(x=left, c='C3')
+        ax[1].axvline(x=centre, c='C3')
+        ax[1].axvline(x=right, c='C3')
+
+        ax[0].set_ylabel('Percentile', labelpad=10)
+        ax[0].set_title('Cumulative distribution')
+        ax[1].set_title('Probability density')
+
+        ax[0].annotate(i, (-0.3, 1),
+                       xycoords='axes fraction',
+                       clip_on=False,
+                       size=14)
+        for a in ax:
+
+            a.set_xlabel('SLR (m)', labelpad=10)
+            a.spines['top'].set_visible(False)
+            a.spines['right'].set_visible(False)
+
+        plt.show()
+
+# Make SLR relative to 2020 level (at the 50th percentile)
+dff -= dff.loc[2020, 'mode']
+
+# Save distribution parameters
+dff[['min', 'mode', 'max']].to_csv('triang-values.csv', float_format='%0.3f')
+
+if PLOT:
+    # Plot all triangular distributions
+    fig, ax = plt.subplots(1, 1, figsize=(8, 4))
+
+    cmap = plt.cm.get_cmap('RdBu_r', len(dff))
+    c = list(cmap(range(cmap.N)))
+
+    j = -1
+    for i, row in dff.iterrows():
+        j += 1
+        ax.plot(row[['min', 'mode', 'max']], [0, 1, 0], c=c[j])
+        if j % 2 == 0:
+            ax.annotate(f' {i}', (row['mode'], 1),
+                        ha='center',
+                        va='bottom',
+                        rotation=90)
+
+    ax.set_xlabel('SLR (m)', labelpad=10)
+    ax.set_ylabel('Probability density (-)', labelpad=10)
+
+    ax.spines['top'].set_visible(False)
+    ax.spines['right'].set_visible(False)
+    plt.show()

slr/generate_slr_timeseries.py

@@ -1,106 +0,0 @@
-import os
-import re
-import numpy as np
-import pandas as pd
-from scipy import stats
-import matplotlib.pyplot as plt
-
-PLOT = False
-START_YEAR = 2020
-END_YEAR = 2100
-
-n_runs = 100000
-
-df = pd.read_csv('cauchy-values.csv', index_col=0)
-years = np.arange(START_YEAR, END_YEAR + 1)
-
-# Squeeze distribution to zero in 2020
-df.loc[2020, 'scale'] = 0.0001
-df.loc[2020, 'min'] = df.loc[2020, 'loc'] - 0.0001
-df.loc[2020, 'max'] = df.loc[2020, 'loc'] + 0.0001
-
-# Interpolate intermediate values
-df = df.reindex(years).interpolate(method='cubic')
-
-# Prepare array for SLR values
-slr = np.zeros([len(years), n_runs], dtype=float)
-
-for i, (year, row) in enumerate(df.iterrows()):
-    # Get probability distribution
-    dist = stats.cauchy(loc=row['loc'], scale=row['scale'])
-
-    # Generate random samples
-    for factor in range(2, 10):
-        s_raw = dist.rvs(n_runs * factor)
-
-        # Take first samples within valid range
-        s = s_raw[(s_raw > row['min']) & (s_raw < row['max'])]
-
-        if len(s) > n_runs:
-            break  # Success
-        else:
-            continue  # We need more samples, so try larger factor
-
-    # Add the requried number of samples
-    slr[i] = s[:n_runs]
-
-# Sort each row to make SLR trajectories smooth
-slr = np.sort(slr, axis=1)
-
-# Randomise run order (column-wise)
-slr = np.random.permutation(slr.T).T
-
-# Set first year to zero
-slr[0, :] = df.loc[2020, 'loc']
-
-# Plot first few trajectories
-if PLOT:
-    fig, ax = plt.subplots(1,
-                           3,
-                           figsize=(12, 5),
-                           sharey=True,
-                           gridspec_kw={
-                               'wspace': 0.05,
-                               'width_ratios': [3, 1, 1]
-                           })
-
-    ax[0].plot(years, slr[:, :100], c='#444444', lw=0.2)
-    ax[1].hist(slr[-1, :],
-               bins=100,
-               fc='#cccccc',
-               ec='#aaaaaa',
-               orientation='horizontal')
-
-    i = len(years) - 1
-    dff = df.T.loc['5':'95', years[i]]
-    ax[2].hist(
-        slr[i, :],
-        bins=100,
-        fc='#cccccc',
-        ec='#aaaaaa',
-        orientation='horizontal',
-        cumulative=True,
-    )
-    ax[2].plot(dff.index.astype(int) / 100 * n_runs,
-               dff.values,
-               'o',
-               c='C3',
-               label='IPCC AR6 data')
-
-    ax[0].set_xlim(right=years[i])
-
-    ax[0].set_title(f'SLR trajectories\n(first 100 out of {n_runs:,} runs)')
-    ax[1].set_title(f'Probability\ndistribution\nin year {years[i]}')
-    ax[2].set_title(f'Cumulative\ndistribution\nin year {years[i]}')
-    ax[0].set_ylabel('SLR (m)', labelpad=10)
-
-    ax[2].legend()
-
-    ax[0].spines['top'].set_visible(False)
-    ax[0].spines['right'].set_visible(False)
-
-    for a in ax[1:]:
-        a.spines['top'].set_visible(False)
-        a.spines['right'].set_visible(False)
-        a.spines['bottom'].set_visible(False)
-        a.xaxis.set_visible(False)

slr/get_slr_distributions.py

@@ -1,161 +0,0 @@
-"""Calculate probability distributions for IPCC sea level rise forecasts.
-
-This will calculate the values required to Cauchy Distributions for
-different SLR projections.
-
-Reads:
-    'IPCC AR6.xlsx'
-
-Writes:
-    'cauchy-values.csv'
-
-D. Howe
-d.howe@wrl.unsw.edu.au
-2022-05-12
-"""
-import os
-import re
-import numpy as np
-import pandas as pd
-from scipy import stats, optimize
-import matplotlib.pyplot as plt
-
-PLOT = True
-
-
-def cauchy_cdf(x, loc, scale):
-    """Calculate cumulative density function, using Cauchy distribution."""
-    return stats.cauchy(loc=loc, scale=scale).cdf(x)
-
-
-# Read data
-df = pd.read_excel('IPCC AR6.xlsx', index_col=[0, 1, 2, 3, 4])
-df = df.sort_index()
-
-# Use all 'medium' confidence scenarios for intermediate quantiles
-scenarios = ['ssp119', 'ssp126', 'ssp245', 'ssp370', 'ssp585']
-dff = df.loc[838, 'total', 'medium', scenarios].groupby('quantile').mean()
-
-# Use ssp119/ssp585 for 5th and 95th quantiles
-dff.loc[5] = df.loc[838, 'total', 'medium', 'ssp119', 5]
-dff.loc[95] = df.loc[838, 'total', 'medium', 'ssp585', 95]
-dff = dff.T
-
-dff.index.name = 'year'
-percentiles = dff.columns.values / 100
-
-for i, row in dff.iterrows():
-    values = row.values
-
-    x_min = row[5] + (row[5] - row[50])
-    x_max = row[95] + (row[95] - row[50])
-    x = np.linspace(x_min, x_max, num=1000)
-
-    # Fit distribution
-    loc, scale = optimize.curve_fit(cauchy_cdf, values, percentiles)[0]
-    p = {'loc': loc, 'scale': scale}
-
-    dff.loc[i, 'loc'] = loc
-    dff.loc[i, 'scale'] = scale
-    dff.loc[i, 'min'] = x_min
-    dff.loc[i, 'max'] = x_max
-
-    if not PLOT:
-        continue
-
-    fig, ax = plt.subplots(1, 2, figsize=(10, 3))
-
-    ax[0].plot(x, 100 * stats.cauchy.cdf(x, **p))
-    ax[0].plot(values, 100 * percentiles, '.', c='#444444')
-
-    ax[1].plot(x, stats.cauchy.pdf(x, **p), label='Cauchy')
-    ax[1].plot([], [], '.', c='#444444', label='IPCC data')
-    ax[1].legend()
-
-    ax[0].set_ylabel('Percentile', labelpad=10)
-    ax[0].set_title('Cumulative distribution')
-    ax[1].set_title('Probability density')
-
-    ax[0].annotate(i, (-0.3, 1),
-                   xycoords='axes fraction',
-                   clip_on=False,
-                   size=14)
-    for a in ax:
-
-        a.set_xlabel('SLR (m)', labelpad=10)
-        a.spines['top'].set_visible(False)
-        a.spines['right'].set_visible(False)
-
-    plt.show()
-
-# Save distribution parameters
-dff.to_csv('cauchy-values.csv', float_format='%g')
-
-if PLOT:
-    # Plot all distributions
-    fig, ax = plt.subplots(1, 1, figsize=(8, 4))
-
-    cmap = plt.cm.get_cmap('RdBu_r', len(dff))
-    c = list(cmap(range(cmap.N)))
-
-    j = -1
-    for i, row in dff.iterrows():
-        j += 1
-
-        x = np.linspace(row['min'], row['max'], num=1000)
-
-        y = stats.cauchy(loc=row['loc'], scale=row['scale']).pdf(x)
-        ax.plot(x, y * row['scale'], c=c[j])
-        if j % 2 == 0:
-            ax.annotate(f' {i}', (x[y.argmax()], 0.32),
-                        ha='center',
-                        va='bottom',
-                        clip_on=False,
-                        rotation=90)
-
-    ax.set_ylim(bottom=0, top=0.35)
-    ax.set_yticks([])
-    ax.set_xlabel('SLR (m)', labelpad=10)
-    ax.set_ylabel('Normalised probability density (-)', labelpad=10)
-
-    ax.spines['top'].set_visible(False)
-    ax.spines['right'].set_visible(False)
-    plt.show()
-
-# Show histgrams with values clipped to our specific range
-if PLOT:
-    for i, row in dff[::4].iterrows():
-        fig, ax = plt.subplots(1, 1, figsize=(10, 2))
-
-        # Generate random samples in Cauchy distribution
-        r = stats.cauchy(loc=row['loc'], scale=row['scale']).rvs(1000000)
-
-        # Clip to our range
-        rc = np.clip(r, a_min=row['min'], a_max=row['max'])
-        f = (r != rc).sum() / len(r)  # Fraction outside range
-
-        ax.hist(rc, 100)
-
-        ym = ax.get_ylim()[1]  # Maximum y value
-
-        ax.axvline(x=row['min'], c='C3')
-        ax.axvline(x=row[5], c='C3')
-        ax.axvline(x=row[50], c='C3')
-        ax.axvline(x=row[95], c='C3')
-        ax.axvline(x=row['max'], c='C3')
-
-        ax.annotate(' P_min', (row['min'], ym))
-        ax.annotate(' P5', (row[5], ym))
-        ax.annotate(' P50', (row[50], ym))
-        ax.annotate(' P95', (row[95], ym))
-        ax.annotate(' P_max', (row['max'], ym))
-
-        ax.annotate(f' Samples clipped = {100 * f:0.1f}%', (x_max, ym / 2))
-
-        ax.set_title(row.name, x=0)
-        ax.set_yticks([])
-        ax.set_xlabel('SLR (m)', labelpad=10)
-        ax.spines['top'].set_visible(False)
-        ax.spines['right'].set_visible(False)
-
-        plt.show()