EarthClimatePhysProject_SeaLevelRise/create_sampler.py at main · adriansousapoza/EarthClimatePhysProject_SeaLevelRise · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Tue Oct 25 16:23:58 2022

@author: asp
"""

import os
import numpy as np
import array
from scipy.stats import norm, lognorm
import emcee
import pandas as pd

datafolder = os.path.abspath('/home/asp/Downloads/Earth and Climate Physics/Project/data')

def read_csv_file(file):
    data = pd.read_csv(f"{datafolder}/{file}").copy()

    ref = 0
    for i in range(len(data['time'])):
        data['time'][i] = data['time'][i]//100 + (data['time'][i] % 100)*1/12

    y = np.array(data['gm_tas'])[1:]
    x = np.array(data['time'])[1:]
    y_mean = np.zeros(len(y[6:-6]))
    x_mean = x[6:-6]
    for i in range(len(y_mean)):
        y_mean[i] = np.mean(y[i:i+12])
        if np.isclose(data['time'][i], 2000):
            ref = y_mean[i]

    y_mean = y_mean - ref
    y = y - ref

    return x,y,x_mean,y_mean

def read_dat_file(file, historical = True):
    run_to_plot = 1
    if historical:
        yr_strt = 1801
        yr_end = 2040
        yr_ref = 2000
    else:
        yr_strt = 2001
        yr_end = 2310
        yr_ref = 2020

    n_mnths = 12 * (yr_end - yr_strt + 1)
    n_ref = 12 * (yr_ref - yr_strt) + 1

    fid = open(f"{datafolder}/{file}", mode = 'rb')
    a = array.array("i")
    a.fromfile(fid, 1)
    b = a[0]
    n_vals = b // 4
    n_runs = n_vals // n_mnths

    tmp = array.array("f")
    tmp.fromfile(fid, b//4)
    tmp = np.asarray(tmp)
    strh_gm = np.reshape(tmp, (n_runs, n_mnths), 'F')
    strh_gm[strh_gm<-1.7e7] = "NaN"

    for i in range(n_runs):
        strh_gm[i,:] = strh_gm[i,:] - strh_gm[i,n_ref]

    x = np.zeros(n_mnths)
    for i in range(n_mnths):
        x[i] = (i + 0.5) / 12 + yr_strt

    y = strh_gm[run_to_plot-1,:][~np.isnan(strh_gm[run_to_plot-1,:])]
    x = x[~np.isnan(strh_gm[run_to_plot-1,:])]

    y_mean = np.zeros(len(y[6:-6]))
    x_mean = x[6:-6]
    for i in range(0,len(y_mean)):
        y_mean[i] = np.mean(y[i:i+12])

    return x,y,x_mean,y_mean

def steric_model(tas_timeseries, a, b, tau, S0):
    delta_t=1
    S = np.empty_like(tas_timeseries)
    S[0] = S0
    for i in range(1,len(tas_timeseries)):
        Seq = a*(tas_timeseries[i]) + b
        dSdt = (Seq-S[i-1])/tau
        S[i] = S[i-1] + dSdt*delta_t
    return S


def log_likelihood(theta):
    a, b, tau, S0, sigma_SL = theta
    model = steric_model(temperature, a,b,tau,S0)
    return np.sum(norm.logpdf(y_mean, loc=model, scale=sigma_SL))

def log_prior(theta):
    a, b, tau, S0, sigma = theta
    logp = lognorm.logpdf(sigma/0.03, s=0.5)
    logp = logp + lognorm.logpdf(tau/100, s=0.3)
    if np.isinf(logp) or np.isnan(logp):
        return -np.inf
    return logp

def log_post(theta):
    logp = log_prior(theta)
    if np.isinf(logp):
        return logp
    logp = logp + log_likelihood(theta)
    return logp


file = "raw_data/CMIP6 steric SSH/cmip6_CMIP_historical_strh_zostoga_gm.dat"
x,y,x_mean,y_mean = read_dat_file(file)

file = "raw_data/cmip6_tas_for_steric_analysis/historical/gm_tas_CMIP_historical_ACCESS-CM2_r1i1p1f1_gn.csv"
x2,y2,x_mean2,temperature = read_csv_file(file)

runs = 5000
pos = [2.2e-01, 1.3e-01, 1.2e+03, -6.4e-02, 3.3e-03] + 1e-4 * np.random.randn(32, 5)
pos = np.float64(pos)
nwalkers, ndim = pos.shape
sampler = emcee.EnsembleSampler(nwalkers, ndim, log_post)
sampler.run_mcmc(pos, runs, progress=True)
tau = sampler.get_autocorr_time()
print(tau)

samples = sampler.get_chain(flat = True)
flat_samples = sampler.get_chain(discard=1000, flat=True)

print(np.mean(flat_samples,axis=0))

np.savetxt("samples_dat_{}runs.csv".format(runs), samples, delimiter=",")
np.savetxt("samples_dat_disc_{}runs.csv".format(runs), flat_samples, delimiter=",")