1. Daten Laden

2. Einladen, Säubern und Darstellen

In [1]:

# Module einladen
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from scipy.optimize import curve_fit

In [2]:

# Einladen und Darstellen der Daten in Numpy
heading = np.loadtxt('monthly_in_situ_co2_mlo.csv', 
                     skiprows = 54, 
                     max_rows = 3, 
                     dtype = str, 
                     delimiter = ',' )
data = np.loadtxt('monthly_in_situ_co2_mlo.csv', 
                  skiprows = 57, 
                  delimiter = ',')

In [3]:

# Säubern der Daten
data_clean = data[np.all(data != -99.99, axis=1)]

In [4]:

# Darstellen in Pandas 
df = pd.DataFrame(data_clean)
df

Out [4]:

	0	1	2	3	4	5	6	7	8	9
0	1958.0	3.0	21259.0	1958.2027	315.70	314.43	316.19	314.90	315.70	314.43
1	1958.0	4.0	21290.0	1958.2877	317.45	315.16	317.30	314.98	317.45	315.16
2	1958.0	5.0	21320.0	1958.3699	317.51	314.71	317.86	315.06	317.51	314.71
3	1958.0	7.0	21381.0	1958.5370	315.86	315.19	315.86	315.22	315.86	315.19
4	1958.0	8.0	21412.0	1958.6219	314.93	316.19	313.99	315.29	314.93	316.19
...	...	...	...	...	...	...	...	...	...	...
749	2021.0	1.0	44211.0	2021.0411	415.17	415.11	415.24	415.17	415.17	415.11
750	2021.0	2.0	44242.0	2021.1260	416.47	415.70	416.13	415.35	416.47	415.70
751	2021.0	3.0	44270.0	2021.2027	417.14	415.59	417.06	415.50	417.14	415.59
752	2021.0	4.0	44301.0	2021.2877	418.24	415.44	418.48	415.66	418.24	415.44
753	2021.0	5.0	44331.0	2021.3699	418.92	415.50	419.23	415.81	418.92	415.50

754 rows × 10 columns

In [5]:

# Achsen definieren
y_all = data_clean[:,4] # Spalte 4 (CO2 Gehalt)
x_all = data_clean[:,3] # Spalte 3 (Datum)

In [6]:

# Plotten
fix, ax = plt.subplots()
ax.plot(x_all,y_all, label = 'CO2 ppm per month')
ax.set_xlabel('Year')
ax.set_ylabel('CO2 Concentration (ppm)' % data[-1,1]);
ax.set_title('Mauna Loa Observatory, Hawaii - Montlhy Average Carbon Dioxide Concentration')
ax.legend()

Out [6]:

<matplotlib.legend.Legend at 0x7fc652cb48d0>

3. Data Analysieren

3.1 Datenverlauf in einem kürzeren Zeitfenster

In [7]:

# Zeitraum der Daten zuschneiden
after_1970 = data_clean[data_clean[:, 0] >= 1970] # nach 1970
data_short = after_1970[after_1970[:, 0] <= 1980] # vor 1980

In [8]:

# Darstellen in Pandas 
df_short = pd.DataFrame(data_short)
df_short

Out [8]:

	0	1	2	3	4	5	6	7	8	9
0	1970.0	1.0	25583.0	1970.0411	325.06	325.02	325.30	325.24	325.06	325.02
1	1970.0	2.0	25614.0	1970.1260	325.98	325.33	325.99	325.32	325.98	325.33
2	1970.0	3.0	25642.0	1970.2027	326.93	325.60	326.74	325.40	326.93	325.60
3	1970.0	4.0	25673.0	1970.2877	328.13	325.74	327.89	325.48	328.13	325.74
4	1970.0	5.0	25703.0	1970.3699	328.08	325.16	328.47	325.55	328.08	325.16
...	...	...	...	...	...	...	...	...	...	...
127	1980.0	8.0	29448.0	1980.6230	337.61	339.00	337.49	338.91	337.61	339.00
128	1980.0	9.0	29479.0	1980.7077	335.90	339.01	335.90	339.04	335.90	339.01
129	1980.0	10.0	29509.0	1980.7896	336.03	339.24	335.95	339.15	336.03	339.24
130	1980.0	11.0	29540.0	1980.8743	337.12	339.13	337.27	339.26	337.12	339.13
131	1980.0	12.0	29570.0	1980.9563	338.23	339.06	338.55	339.36	338.23	339.06

132 rows × 10 columns

In [9]:

# Achsen definieren
y_short = data_short[:,4]
x_short = data_short[:,3]

In [10]:

# Plotten des Zeitraums
fix, ax = plt.subplots()
ax.plot(x_short,y_short, label = 'CO2 ppm per month')
ax.set_xlabel('Date')
ax.set_ylabel('CO2 Concentration (ppm)' % data[-1,1]);
ax.set_title('Mauna Loa Observatory, Hawaii - Montlhy Average Carbon Dioxide Concentration')
ax.legend()

Out [10]:

<matplotlib.legend.Legend at 0x7fc652bc1b10>

3.2 Datenanpassung

In [11]:

# Funktion Definieren
def func_linear (x, a, f, m, c):
    return (a * np.sin(x * f)) + (m * x + c)

In [12]:

# Data fit
popt_short, pcov_short  = curve_fit(func_linear, x_short, y_short, p0=[10, 3.7, 5, 5])

In [13]:

# Plotten der Daten & Fit des Zeitraums
fig, ax = plt.subplots()
ax.plot(x_short, y_short, label = 'CO2 ppm per month') 
ax.plot(x_short, func_linear(x_short, popt_short[0], popt_short[1], popt_short[2], popt_short[3]), '.-r', label = 'data_fit')
ax.set_xlabel('Time')
ax.set_ylabel('CO2 Concentration (ppm)' % data[-1,1])
ax.set_title('Mauna Loa Observatory, Hawaii - Montlhy Average Carbon Dioxide Concentration')
plt.legend()

Out [13]:

<matplotlib.legend.Legend at 0x7fc652b47790>

3.3 Fitbereich erweitern

In [14]:

# Plotten
fig, ax = plt.subplots()
ax.plot(x_all, y_all, label = 'CO2 ppm per month') 
ax.plot(x_all, func_linear(x_all, popt_short[0], popt_short[1], popt_short[2], popt_short[3]), '.-r', label = 'data_fit') # x_all (Ganzer Zeitraum)
ax.set_xlabel('Time')
ax.set_ylabel('CO2 Concentration (ppm)' % data[-1,1])
ax.set_title('Mauna Loa Observatory, Hawaii - Montlhy Average Carbon Dioxide Concentration')
plt.legend()

Out [14]:

<matplotlib.legend.Legend at 0x7fc652ac6b50>

3.4 Zweite Datenanpassung

In [15]:

# Funktion definieren
def func_exp(x, b, xo, p, c):
    return b * ((x - xo)**p) + c

In [16]:

# Curve Fit
popt_all, pcov_all = curve_fit(func_exp, x_all, y_all, p0 = [5,0.1,1.5,5])

/opt/tljh/user/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: invalid value encountered in power
  This is separate from the ipykernel package so we can avoid doing imports until

In [17]:

# Plot
fig, ax = plt.subplots()
ax.plot(x_all, y_all, label = 'CO2 ppm per month') 
ax.plot(x_all, func_exp(x_all, popt_all[0], popt_all[1], popt_all[2], popt_all[3]), '.-r', label = 'data_fit')
ax.set_xlabel('Time')
ax.set_ylabel('CO2 Concentration (ppm)' % data[-1,1])
ax.set_title('Mauna Loa Observatory, Hawaii - Montlhy Average Carbon Dioxide Concentration')
plt.legend()

Out [17]:

<matplotlib.legend.Legend at 0x7fc652a57610>

3.5 Anpassung Kombinieren

In [18]:

# Funktion definieren
def func_exp_tri(x, b, xo, p, c, a, f):
    return (b * ((x - xo)**p) + c) + (a * np.sin(x * f))

In [19]:

# Curve Fit
popt_all, pcov_all = curve_fit(func_exp_tri, x_all, y_all, p0 = [5,0.1,1.5,5,10,3.7])

/opt/tljh/user/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: invalid value encountered in power
  This is separate from the ipykernel package so we can avoid doing imports until

In [20]:

# Plot
fig, ax = plt.subplots()
ax.plot(x_all, y_all, label = 'CO2 ppm per month') 
ax.plot(x_all, func_exp_tri(x_all, popt_all[0], popt_all[1], popt_all[2], popt_all[3], popt_all[4], popt_all[5]), '.-r', label = 'data_fit')
ax.set_xlabel('Time')
ax.set_ylabel('CO2 Concentration (ppm)' % data[-1,1])
ax.set_title('Mauna Loa Observatory, Hawaii - Montlhy Average Carbon Dioxide Concentration')
plt.legend()

Out [20]:

<matplotlib.legend.Legend at 0x7fc6529d8e50>

3.6 Ausgabe

In [21]:

# Ausgabe der Funktion
popt_all_rounded = np.round(popt_all, decimals = 2)
print('f(x) = ('+str(popt_all_rounded[0])+' * ((x - '+str(popt_all_rounded[1])+')**'+str(popt_all_rounded[2])+') + '+str(popt_all_rounded[3])+') + ('+str(popt_all_rounded[4])+' * np.sin(x * '+str(popt_all_rounded[5])+'))')

f(x) = (0.24 * ((x - 1958.2)**1.44) + 316.15) + (-0.06 * np.sin(x * 3.68))

4. Daten Extrapolieren

In [22]:

# Neuer Zeitraum
x_future_int = np.arange(2000,2101)

In [23]:

# Plot
fig, ax = plt.subplots()
ax.plot(x_future_int, func_exp_tri(x_future_int, popt_all[0], popt_all[1], popt_all[2], popt_all[3], popt_all[4], popt_all[5]), '.-r', label = 'data_fit')
ax.set_xlabel('Date')
ax.set_ylabel('CO2 Concentration (ppm)' % data[-1,1])
ax.set_title('Mauna Loa Observatory, Hawaii - Extrapolation - Montlhy Average Carbon Dioxide Concentration')
plt.legend()

Out [23]:

<matplotlib.legend.Legend at 0x7fc65295d1d0>

5. Klimaszenarien

In [24]:

# Datenstruktur
ipcc_year  = [2000,2005,2010,2020,2030,2040,2050,2060,2070,2080,2090,2100]
ipcc_rcp26 = [368.9,378.8,389.3,412.1,430.8,440.2,442.7,441.7,437.5,431.6,426.0,420.9]
ipcc_rcp45 = [368.9,378.8,389.1,411.1,435.0,460.8,486.5,508.9,524.3,531.1,533.7,538.4]
ipcc_rcp60 = [368.9,378.8,389.1,409.4,428.9,450.7,477.7,510.6,549.8,594.3,635.6,669.7]
ipcc_rcp85 = [368.9,378.8,389.3,415.8,448.8,489.4,540.5,603.5,677.1,758.2,844.8,935.9]

In [25]:

# Plot
fig, ax = plt.subplots()
ax.scatter(ipcc_year, ipcc_rcp26, c = 'b', label = 'RCP26')
ax.scatter(ipcc_year, ipcc_rcp45, c = 'g', label = 'RCP45')
ax.scatter(ipcc_year, ipcc_rcp60, c = 'r', label = 'RCP60')
ax.scatter(ipcc_year, ipcc_rcp85, c = 'c', label = 'RCP85')
ax.set_xlabel('Time')
ax.set_ylabel('CO2 Concentration (ppm)')
ax.set_title('5. IPCC Bericht - Abundances of the Well-Mixed Greenhouse Gases')
plt.legend()

Out [25]:

<matplotlib.legend.Legend at 0x7fc6528cf690>

6. Business as usual vs. Klimamodelle

In [26]:

# Jahre und Monate für Rechnung in  Dezimalzahlen umwandeln
year_month_dec = []
sum = 2000
for i in range (1,1201):
    sum = sum + (1/12)
    year_month_dec.append(np.round(sum, decimals = 2))

In [27]:

# Interpolation
rcp26_interp = np.interp(year_month_dec,ipcc_year,ipcc_rcp26,)
rcp45_interp = np.interp(year_month_dec,ipcc_year,ipcc_rcp45,)
rcp60_interp = np.interp(year_month_dec,ipcc_year,ipcc_rcp60,)
rcp85_interp = np.interp(year_month_dec,ipcc_year,ipcc_rcp85,)

In [28]:

# Neue X-Achse mit Korrekter Beschriftung
year_month = []
for i in range(2000,2100):
    for j in range(1,13):
        date = str(i) + '_' + str(j)
        year_month.append(date)

In [29]:

# Plot
fig, ax = plt.subplots()
ax.plot(year_month, rcp26_interp, c = 'b', label = 'RCP26')
ax.plot(year_month, rcp45_interp, c = 'g', label = 'RCP45')
ax.plot(year_month, rcp60_interp, c = 'r', label = 'RCP60')
ax.plot(year_month, rcp85_interp, c = 'c', label = 'RCP85')
ax.plot(year_month, func_exp_tri(np.round(year_month_dec, decimals = 0), popt_all[0], popt_all[1], popt_all[2], popt_all[3], popt_all[4], popt_all[5]), 'k', label = 'data_fit')
plt.xticks(np.arange(0, 1200, step=120))
plt.xticks(rotation = 'vertical')
ax.set_xlabel('Time')
ax.set_ylabel('CO2 Concentration (ppm)')
ax.set_title('5. IPCC Bericht - Abundances of the Well-Mixed Greenhouse Gases')
plt.legend()

# Kann etwas länger laden

Out [29]:

<matplotlib.legend.Legend at 0x7fc652820050>

In [ ]:

# RCP26 und RCP85 zeigen starke Unterschiede in den kommenden Jahren. Die selbst erstellte Projektion entspricht am ehesten RCP45. Hoffentlich wird es eher RCP26 :I

310 KiB Raw Permalink Blame History