Time Series Comparisons: Fluxes versus Forcing Functions¶
In [4]:
import datetime
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from scipy import stats
import seaborn as sns
%matplotlib inline
In [5]:
plt.rcParams['font.size'] = 16
In [6]:
sns.set_palette(sns.color_palette("colorblind"))
myp = sns.color_palette('colorblind')
sns.palplot(myp)
In [7]:
def scaled_correlation(a, b, scale):
sum_correlations = 0
count = 0
for i in range(len(a)//scale):
r, _ = stats.pearsonr(
a[i * scale: (i + 1) * scale],
b[i * scale: (i + 1) * scale]
)
sum_correlations += r
count += 1
return sum_correlations/count
In [8]:
# Fluxes
shallow_flux = pd.read_csv('shallowflux.csv', parse_dates=[0], index_col=0)
deep_flux = pd.read_csv('deepflux.csv', parse_dates=[0], index_col=0)
baroclinic = shallow_flux + deep_flux
barotropic = shallow_flux - deep_flux
In [9]:
# Rivers
r2015 = pd.read_csv('SoG_runoff_2015.csv', parse_dates=True, index_col=0)
r2016 = pd.read_csv('SoG_runoff_2016.csv', index_col=0, parse_dates=True)
r2017 = pd.read_csv('SoG_runoff_2017.csv', index_col=0, parse_dates=True)
r2018 = pd.read_csv('SoG_runoff_2018.csv', index_col=0, parse_dates=True)
SoGrunoff = pd.concat([r2015, r2016, r2017, r2018])
In [10]:
# Wind
wind = pd.read_csv('day_avg_wind.csv', index_col=0)
wind.index = pd.to_datetime(wind.index, format="%Y-%m-%d")
low_pass_wind = wind.rolling(window=4, center=True).mean()
In [11]:
# SSH
ssh = pd.read_csv('low_pass_ssh.csv')
base = datetime.datetime(2015, 1, 1, tzinfo=datetime.timezone.utc)
date_list = [base + datetime.timedelta(days=x) for x in range(1461)]
ssh['time'] = date_list
ssh.set_index('time', inplace=True)
In [12]:
# Tides
low_pass_tide = pd.read_csv('low_pass_tide.csv', index_col=0)
low_pass_tide.index = pd.to_datetime(low_pass_tide.index, format="%Y-%m-%d")
In [13]:
# Density Forcing
sigma = pd.read_csv('sigma_2015_201806.csv', index_col=0)
m2015 = sigma.south - sigma.north
m2015.index = pd.to_datetime(m2015.index, format="%Y-%m-%d")
sigma = pd.read_csv('sigma_2016_201806.csv', index_col=0)
m2016 = sigma.south - sigma.north
m2016.index = pd.to_datetime(m2016.index, format="%Y-%m-%d")
sigma = pd.read_csv('sigma_2017_201806.csv', index_col=0)
m2017 = sigma.south - sigma.north
m2017.index = pd.to_datetime(m2017.index, format="%Y-%m-%d")
sigma = pd.read_csv('sigma_2018_201806.csv', index_col=0)
m2018 = sigma.south - sigma.north
m2018.index = pd.to_datetime(m2018.index, format="%Y-%m-%d")
goverrho=9.81/1000.
depthwidth = 50*10e3
densitydiff = pd.concat([m2015, m2016, m2017, m2018])
densityforcing = np.sqrt(goverrho*densitydiff[:])
In [14]:
densitydiff
Out[14]:
2015-01-01 00:00:00+00:00 188.117726
2015-01-02 00:00:00+00:00 195.545107
2015-01-03 00:00:00+00:00 196.789936
2015-01-04 00:00:00+00:00 189.535515
2015-01-05 00:00:00+00:00 178.682278
...
2018-12-27 00:00:00+00:00 145.119867
2018-12-28 00:00:00+00:00 151.433523
2018-12-29 00:00:00+00:00 152.565994
2018-12-30 00:00:00+00:00 175.501618
2018-12-31 00:00:00+00:00 192.270882
Length: 1461, dtype: float64
In [15]:
x = densitydiff[:]
y = baroclinic.transport
nansInArray = (np.isnan(y) | np.isnan(x))
print (np.corrcoef(x[~nansInArray], y[~nansInArray]))
[[1. 0.81211531] [0.81211531 1. ]]
In [16]:
max = 0
for lag in range(-30, 30):
value = baroclinic.transport.corr(densitydiff.shift(lag))
if value > max:
max = value
mlag = lag
plt.plot(lag, value, 'x')
print (mlag, max)
3 0.8481520997403567
In [17]:
x = densityforcing[:]
y = baroclinic.transport
nansInArray = (np.isnan(y) | np.isnan(x))
print (np.corrcoef(x[~nansInArray], y[~nansInArray]))
[[1. 0.8073329] [0.8073329 1. ]]
In [18]:
x = low_pass_tide.vozocrtx.values
y = baroclinic.transport.values
nansInArray = (np.isnan(y) | np.isnan(x))
print (np.corrcoef(x[~nansInArray], y[~nansInArray]))
[[ 1. -0.29698463] [-0.29698463 1. ]]
In [26]:
xmin = 0
sum = 0
icount = 0
for scale in range(7, 30):
for shift in range(scale):
correl = scaled_correlation(x[~nansInArray][shift:], y[~nansInArray][shift:], scale)
if correl < xmin:
print(scale, shift, correl)
xmin = correl
if scale == 20:
# print (shift, correl)
sum = sum + correl
icount = icount + 1
print (sum/icount)
7 0 -0.5183314140976558 8 4 -0.5186335080227122 8 5 -0.5346159072380634 8 6 -0.5375378857498109 9 0 -0.5590004743163116 10 3 -0.5755258249468537 11 7 -0.5853702512774374 12 4 -0.5969262677860706 12 5 -0.6058928340650297 16 5 -0.6071867464900502 16 6 -0.6191511355224563 16 7 -0.6231722340980251 20 0 -0.6276783535421163 20 16 -0.632815023096598 20 17 -0.6384214658739567 -0.5995265343015981
In [37]:
min = 0
tide = low_pass_tide.tz_localize(None)
trans = baroclinic.tz_localize(None)
for lag in range(-30, 30):
value = trans.transport.corr(tide.vozocrtx.shift(lag))
if value < min:
min = value
mlag = lag
plt.plot(lag, value, 'x')
print (mlag, min)
-1 -0.30143435753216796
In [19]:
x = SoGrunoff.rorunoff
y = baroclinic.transport
nansInArray = (np.isnan(y) | np.isnan(x))
print (np.corrcoef(x[~nansInArray], y[~nansInArray]))
[[1. 0.35128684] [0.35128684 1. ]]
In [55]:
plt.plot(x - 0.5)
plt.plot(y/1e5)
Out[55]:
[<matplotlib.lines.Line2D at 0x7f1342775dc0>]
In [53]:
max = 0
for lag in range(-300, 0):
trans = baroclinic.tz_localize(None)
run = SoGrunoff.tz_localize(None)
value = trans.transport.corr(run.rorunoff.shift(lag))
if value > max:
max = value
mlag = lag
plt.plot(lag, value, 'x')
print (mlag, max)
-300 0.67016011416756
In [45]:
max = 0
for lag in range(0, 90):
density = densitydiff.tz_localize(None)
run = SoGrunoff.tz_localize(None)
value = density.corr(run.rorunoff.shift(lag))
if value > max:
max = value
mlag = lag
plt.plot(lag, value, 'x')
print (mlag, max)
46 0.6724532507316303
In [21]:
x = low_pass_wind.wind.values
y = baroclinic.transport.values
nansInArray = (np.isnan(y) | np.isnan(x))
print (np.corrcoef(x[~nansInArray], y[~nansInArray]))
[[ 1. -0.5397062] [-0.5397062 1. ]]
In [22]:
x = low_pass_wind.wind.values
y = densitydiff[:]
nansInArray = (np.isnan(y) | np.isnan(x))
print (np.corrcoef(x[~nansInArray], y[~nansInArray]))
[[ 1. -0.70492634] [-0.70492634 1. ]]
In [28]:
x = low_pass_wind.wind.values
y = barotropic.transport[:]
nansInArray = (np.isnan(y) | np.isnan(x))
print (np.corrcoef(x[~nansInArray], y[~nansInArray]))
[[ 1. -0.10037231] [-0.10037231 1. ]]
In [33]:
xmin = 0
for scale in range(2, 70):
sum = 0
icount = 0
for shift in range(scale):
correl = scaled_correlation(x[~nansInArray][shift:], y[~nansInArray][shift:], scale)
sum = sum + correl
icount = icount + 1
if sum/icount < xmin:
print(scale, sum/icount)
xmin = sum/icount
3 -0.011200018230132993 4 -0.01530209390573908 5 -0.026683606610939277 6 -0.04022024811381387 7 -0.0578778196174747 8 -0.07628291639293615 9 -0.09361381397391588 10 -0.1097476588077088 11 -0.12507890844301658 12 -0.14026034549988617 13 -0.15441651090948086 14 -0.1673362139658757 15 -0.17891324497694414 16 -0.18913053500163052 17 -0.19809858245912165 18 -0.20615872229384838 19 -0.2133109838807255 20 -0.21967100456118377 21 -0.22556717745619304 22 -0.2308115035273648 23 -0.2353183066515692 24 -0.23949863820281828 25 -0.24285832514685637 26 -0.2459802834333914 27 -0.2487052993428646 28 -0.25085063018680104 29 -0.2528195876276133 30 -0.2548367904152905 31 -0.2561883192169029 32 -0.257328821901221 33 -0.2587949694314816 34 -0.25960303471316415 35 -0.2603447725722375 36 -0.2609890329226619 37 -0.2612785709992666 38 -0.26159524882139884 39 -0.26206059965569434 40 -0.26224686478185033 41 -0.2623122942137049 42 -0.26254528705057795 43 -0.2625837371726567 44 -0.2626039389070515 45 -0.26272242165853765 47 -0.26272377619876663 48 -0.26281605912290035
In [24]:
x = low_pass_wind.wind.values
y = ssh.sossheig
nansInArray = (np.isnan(y) | np.isnan(x))
print (np.corrcoef(x[~nansInArray], y[~nansInArray]))
[[1. 0.7561315] [0.7561315 1. ]]
In [26]:
x = barotropic.transport
y = ssh.sossheig
nansInArray = (np.isnan(y) | np.isnan(x))
print (np.corrcoef(x[~nansInArray], y[~nansInArray]))
[[ 1. -0.05010086] [-0.05010086 1. ]]
In [67]:
fig, axs = plt.subplots(4, 2, figsize=(16, 12))
fig.subplots_adjust(wspace=0.25)
mylabels = [['a)', 'e)'], ['b)', 'f)'], ['c)', 'g)'], ['d)', 'h)']]
for iax in range(2):
(barotropic.transport/1000).plot(ax=axs[0, iax], color=myp[0], label="Barotropic")
(baroclinic.transport/1000).plot(ax=axs[0, iax], color=myp[1], label="Baroclinic")
np.sqrt(low_pass_tide.vozocrtx).plot(ax=axs[1, iax], color=myp[2], label="Low-Pass Tidal Speed")
densityforcing.plot(ax=axs[1, iax], color=myp[3], label="Density Forcing")
(SoGrunoff.rorunoff/1000).plot(ax=axs[2, iax], color=myp[4], label="River Runoff")
low_pass_wind.wind.plot(ax=axs[3, iax], color=myp[5], label="Low-Pass Wind Velocity")
axb = axs[3, iax].twinx()
axb.set_ylabel('Sea Surface Height (m)',color=myp[9])
ssh.sossheig.plot(ax=axb, color=myp[9], label="Sea Surface Height")
axb.legend(loc=[0.55, 0.85], fontsize=13)
axb.set_ylim(None, 0.72)
axb.set_yticks([-0.2, 0, 0.2, 0.4, 0.6])
for ax in axs[:, 0]:
ax.set_xlim(datetime.datetime(2015, 1, 1), datetime.datetime(2017, 1, 1))
for ax in axs[:, 1]:
ax.set_xlim(datetime.datetime(2015, 3, 1), datetime.datetime(2015, 9, 1))
for ax in axs[1, :]:
ax.set_ylim(None, 2.35)
for ax in axs[3, :]:
ax.set_ylim(None, 16)
#axs[1, 0].text(datetime.datetime(2014, 9, 1), -1.0, 'Tidal Speed Squared (m$^2$ s$^{-2}$)',
# rotation=90, color=myp[2])
#axs[1, 1].text(datetime.datetime(2015, 1, 29), -1.0, 'Tidal Speed Squared (m$^2$ s$^{-2}$)',
# rotation=90, color=myp[2])
axs[1, 0].text(datetime.datetime(2014, 8, 1), 0.2, 'Tidal Speed (m s$^{-1}$)', rotation=90, color=myp[2])
axs[1, 1].text(datetime.datetime(2015, 1, 22), 0.2, 'Tidal Speed (m s$^{-1}$)', rotation=90, color=myp[2])
for i in range(2):
axs[1, i].set_ylim(bottom=0.2)
axs[3, i].set_xlabel('Date')
axs[0, i].set_ylabel('Flux (mSv)')
axs[1, i].set_ylabel('Density Forcing (m s$^{-1}$)', color=myp[3])
axs[2, i].set_ylabel("River Runoff\n (mSv)")
axs[2, i].set_ylim(bottom=1)
axs[3, i].set_ylabel('Wind Velocity (m s$^{-1}$)', color=myp[5])
for j in range(3):
axs[j, i].set_xticklabels([])
axs[j, i].set_xlabel('')
axs[j, i].legend(loc='best', fontsize=14)
for j in range(4):
axs[j, i].grid(axis='x')
axs[j, i].text(0.01, 0.02, mylabels[j][i], transform=axs[j, i].transAxes,
fontsize=14, verticalalignment='bottom', fontfamily='serif',
bbox=dict(facecolor='0.7', edgecolor='none', pad=3.0))
axs[0, 0].legend(loc=[0.37, 0.72], fontsize=13)
axs[0, 1].legend(loc=[0.02, 0.72], fontsize=13)
axs[1, 0].legend(loc=[0.02, 0.85], fontsize=13, ncol=2)
axs[1, 1].legend(loc=[0.02, 0.85], fontsize=13, ncol=2)
axs[2, 0].legend(loc=[0.02, 0.82], fontsize=13)
axs[2, 1].legend(loc=[0.02, 0.82], fontsize=13)
axs[3, 0].legend(loc=[0.02, 0.85], fontsize=13)
axs[3, 1].legend(loc=[0.02, 0.85], fontsize=13)
fig.tight_layout()
plt.savefig('Ancillary_v2.pdf')
plt.savefig('Ancillary_v2.png')
In [18]:
fig, axs = plt.subplots(2, 2, figsize=(16, 6))
fig.subplots_adjust(wspace=0.25)
for iax in range(2):
(baroclinic.transport/1000).plot(ax=axs[0, iax], color=myp[1], label="Baroclinic Flux")
low_pass_tide.vozocrtx.plot(ax=axs[1, iax], color=myp[2], label="Low-Pass Tidal Speed")
densityforcing.plot(ax=axs[1, iax], color=myp[3], label="Density Forcing")
for ax in axs[:, 0]:
ax.set_xlim(datetime.datetime(2015, 1, 1), datetime.datetime(2017, 1, 1))
for ax in axs[:, 1]:
ax.set_xlim(datetime.datetime(2015, 3, 1), datetime.datetime(2015, 9, 1))
for ax in axs[1, :]:
ax.set_ylim(None, 2.35)
axs[1, 0].text(datetime.datetime(2014, 8, 1), 0.0, 'Tidal Speed (m s$^{-1}$)', rotation=90, color=myp[2])
#axs[1, 1].text(datetime.datetime(2015, 1, 22), 0.0, 'Tidal Speed (m s$^{-1}$)', rotation=90, color=myp[2])
for i in range(2):
axs[1, i].set_xlabel('Date')
axs[0, i].set_xticklabels([])
axs[0, i].set_xlabel('')
axs[1, i].legend(loc='best', fontsize=14)
for j in range(2):
axs[j, i].grid(axis='x')
axs[0, 0].set_ylabel('Flux (mSv)')
axs[1, 0].set_ylabel('Density Forcing (m s$^{-1}$)', color=myp[3])
axs[0, 0].legend(loc=[0.37, 0.72], fontsize=13)
axs[0, 1].legend(loc=[0.02, 0.72], fontsize=13)
axs[1, 0].legend(loc=[0.02, 0.85], fontsize=13, ncol=2)
axs[1, 1].legend(loc=[0.02, 0.85], fontsize=13, ncol=2)
fig.tight_layout()
plt.savefig('Ancillary_CMOS2024.pdf')
plt.savefig('Ancillary_CMOS2024.png')
In [14]:
fig, axs = plt.subplots(2, 1, figsize=(8, 6))
fig.subplots_adjust(wspace=0.25)
#(barotropic.transport/1000).plot(ax=axs[0], color=myp[0], label="Barotropic")
(baroclinic.transport/1000).plot(ax=axs[0], color=myp[1], label="Baroclinic Flux")
low_pass_tide.vozocrtx.plot(ax=axs[1], color=myp[2], label="Low-Pass Tidal Speed")
densityforcing.plot(ax=axs[1], color=myp[3], label="Density Forcing")
for ax in axs:
ax.set_xlim(datetime.datetime(2015, 1, 1), datetime.datetime(2017, 1, 1))
axs[1].set_ylim(None, 2.35)
axs[1].text(datetime.datetime(2014, 8, 1), 0.0, 'Tidal Speed (m s$^{-1}$)', rotation=90, color=myp[2])
axs[1].set_xlabel('Date')
axs[0].set_ylabel('Flux (mSv)')
axs[1].set_ylabel('Density Forcing (m s$^{-1}$)', color=myp[3])
axs[0].set_xticklabels([])
axs[0].set_xlabel('')
axs[0].legend(loc='best', fontsize=14)
for j in range(2):
axs[j].grid(axis='x')
axs[0].legend(loc=[0.37, 0.72], fontsize=13)
axs[1].legend(loc=[0.02, 0.85], fontsize=13, ncol=2)
fig.tight_layout()
plt.savefig('Ancillary_CMOS2024.pdf')
plt.savefig('Ancillary_CMOS2024.png')
In [69]:
fig, axs = plt.subplots(4, 1, figsize=(15, 12))
fig.subplots_adjust(wspace=0.25)
mylabels = ['a)', 'b)', 'c)', 'd)']
(barotropic.transport/1000).plot(ax=axs[0], color=myp[0], label="Barotropic")
(baroclinic.transport/1000).plot(ax=axs[0], color=myp[1], label="Baroclinic")
#a1l1, = axs[1].plot(low_pass_tide.index, low_pass_tide.vozocrtx, color=myp[2], label="Low-Pass Tidal Speed")
np.sqrt(low_pass_tide.vozocrtx).plot(ax=axs[1], color=myp[2], label="Low-Pass Tidal Speed")
#a1l2, = axs[1].plot(densityforcing.index, densityforcing, color=myp[3], label="Density Forcing")
densityforcing.plot(ax=axs[1], color=myp[3], label="Density Forcing")
(SoGrunoff.rorunoff/1000).plot(ax=axs[2], color=myp[4], label="River Runoff")
low_pass_wind.wind.plot(ax=axs[3], color=myp[5], label="Low-Pass Wind Velocity")
axb = axs[3].twinx()
axb.set_ylabel('Sea Surface Height (m)',color=myp[9])
ssh.sossheig.plot(ax=axb, color=myp[9], label="Sea Surface Height")
axb.legend(loc=[0.775, 0.82], fontsize=13)
axb.set_ylim(None, 0.72)
for ax in axs:
ax.set_xlim(datetime.datetime(2015, 1, 1), datetime.datetime(2019, 1, 1))
axs[1].set_ylim(None, 2.45)
axs[3].set_ylim(None, 16)
axs[1].text(datetime.datetime(2014, 9, 15), 0.2, 'Tidal Speed (m s$^{-1}$)', rotation=90, color=myp[2])
axs[3].set_xlabel('Date')
axs[0].set_ylabel('Flux (mSv)')
axs[1].set_ylabel('Density Forcing (m s$^{-1}$)', color=myp[3])
axs[2].set_ylabel("River Runoff\n (mSv)")
axs[2].set_ylim(bottom=1)
axs[3].set_ylabel('Wind Velocity (m s$^{-1}$', color=myp[5])
for j in range(3):
axs[j].set_xticklabels([])
axs[j].set_xlabel('')
axs[j].legend(loc='best')
axs[0].legend(loc=(0.46, 0.69), fontsize=13)
axs[1].legend(loc=(0.05, 0.82), fontsize=13, ncol=2)
axs[2].legend(loc='best', fontsize=13)
axs[3].legend(loc=[0.05, 0.82], fontsize=13)
for j in range(4):
axs[j].grid(axis='x')
axs[j].text(0.005, 0.02, mylabels[j], transform=axs[j].transAxes,
fontsize=14, verticalalignment='bottom', fontfamily='serif',
bbox=dict(facecolor='0.7', edgecolor='none', pad=3.0))
plt.savefig('Ancillary_sup_v2.pdf')
plt.savefig('Ancillary_sup_v2.png')
In [11]:
plt.rcParams['font.size'] = 14
In [12]:
fig, axs = plt.subplots(4, 1, figsize=(18, 12))
fig.subplots_adjust(wspace=0.25)
(shallow_flux.transport/1000).plot(ax=axs[0], color=myp[0], label="Shallow Flux")
low_pass_tide.vozocrtx.plot(ax=axs[1], color=myp[2], label="Low-Pass Tidal Speed")
(SoGrunoff.rorunoff/1000).plot(ax=axs[2], color=myp[4], label="River Runoff")
low_pass_wind.wind.plot(ax=axs[3], color=myp[5], label="Low-Pass Wind Velocity")
for ax in axs[:]:
ax.set_xlim(datetime.datetime(2015, 1, 1), datetime.datetime(2019, 1, 1))
axs[3].set_xlabel('Date')
axs[0].set_ylabel('Flux (mSv)')
axs[1].set_ylabel('Low-Pass Tidal Speed (m s$^{-1}$)', color=myp[3])
axs[2].set_ylabel("River Runoff (mSv)")
axs[3].set_ylabel('Low-Pass Wind Velocity (m s$^{-1}$', color=myp[5])
for j in range(3):
axs[j].set_xticklabels([])
axs[j].set_xlabel('')
for j in range(4):
axs[j].grid(axis='x')
axs[0].text(datetime.datetime(2016, 10, 15), 50, 'Shallow Flux to\n Victoria Sill', color='b', fontsize=18);
axs[1].text(datetime.datetime(2016, 9, 1), 1.07, 'Tidal Speed', color='b', fontsize=18);
axs[2].text(datetime.datetime(2016, 9, 15), 12, 'FreshWater', color='b', fontsize=18);
axs[3].text(datetime.datetime(2016, 5, 15), 7, 'Wind', color='b', fontsize=18);
p1 = datetime.datetime(2015, 6, 11)
for ax in axs:
ax.axvline(x=p1, color='r')
p2 = datetime.datetime(2016, 8, 13)
for ax in axs:
ax.axvline(x=p2, color='r')
p3 = datetime.datetime(2017, 8, 3)
for ax in axs:
ax.axvline(x=p3, color='r')
p4 = datetime.datetime(2018, 7, 4)
for ax in axs:
ax.axvline(x=p4, color='r')
#fig.savefig('/home/sallen/MEOPAR/estuarine_flux_paper/Leblond_forcing.png')
In [13]:
## Just Look at Barotropic Flux
In [14]:
barotropic[0:2]
Out[14]:
| transport | |
|---|---|
| 2015-01-01 00:00:00+00:00 | NaN |
| 2015-01-02 00:00:00+00:00 | NaN |
In [15]:
fig, ax = plt.subplots(1, 1, figsize=(15, 4))
barotropic.plot(ax=ax);
In [16]:
fig, ax = plt.subplots(1, 1, figsize=(15, 4))
ax.plot(barotropic.index, np.cumsum(barotropic.transport));
In [17]:
## So consistently out
In [18]:
fig, ax = plt.subplots(1, 1, figsize=(15, 4))
barotropic.plot(ax=ax);
shallow_flux.plot(ax=ax);
deep_flux.plot(ax=ax);
#ax.set_xlim(datetime.datetime(2017, 1, 1), datetime.datetime(2017, 12, 31))
ax.grid();
In [19]:
fig, ax = plt.subplots(1, 1, figsize=(15, 4))
barotropic.rolling(window=30).mean().plot(ax=ax);
shallow_flux.rolling(window=30).mean().plot(ax=ax);
deep_flux.rolling(window=30).mean().plot(ax=ax);
#ax.set_xlim(datetime.datetime(2017, 1, 1), datetime.datetime(2017, 12, 31))
ax.grid();
In [20]:
fig, ax = plt.subplots(1, 1, figsize=(15, 4))
barotropic.resample('M').mean().plot(ax=ax);
shallow_flux.resample('M').mean().plot(ax=ax);
deep_flux.resample('M').mean().plot(ax=ax);
ax.set_xlim(datetime.datetime(2017, 1, 1), datetime.datetime(2017, 12, 31))
ax.grid();
In [21]:
print(23.8+8)
shallow_flux.mean()
31.8
Out[21]:
transport 32148.180265 dtype: float64
In [22]:
print(23.7+2.2)
deep_flux.mean()
25.9
Out[22]:
transport 26750.82219 dtype: float64
Yearly Values
In [25]:
SoGrunoff.groupby(SoGrunoff.index.year).mean()
Out[25]:
| rorunoff | |
|---|---|
| datetime | |
| 2015 | 4918.779428 |
| 2016 | 4854.659295 |
| 2017 | 4719.688140 |
| 2018 | 4726.770248 |
In [27]:
densityforcing.groupby(densityforcing.index.year).mean()
Out[27]:
2015 1.538397 2016 1.526091 2017 1.559422 2018 1.535584 dtype: float64
In [31]:
plt.plot(SoGrunoff.groupby(SoGrunoff.index.year).mean()/4825, densityforcing.groupby(densityforcing.index.year).mean(), 'o')
Out[31]:
[<matplotlib.lines.Line2D at 0x7f61e512d130>]
In [52]:
plt.plot(-ssh.groupby(ssh.index.year).sossheig.mean().values*
(SoGrunoff.groupby(SoGrunoff.index.year).mean().values.transpose()[0]-4000),
densityforcing.groupby(densityforcing.index.year).mean().values)
Out[52]:
[<matplotlib.lines.Line2D at 0x7f61e58b18e0>]
In [55]:
plt.plot(low_pass_wind.wind.groupby(low_pass_wind.index.year).mean(),
densityforcing.groupby(densityforcing.index.year).mean(), 'o')
Out[55]:
[<matplotlib.lines.Line2D at 0x7f61dd35d250>]
In [81]:
plt.plot(-low_pass_wind.wind.groupby(low_pass_wind.index.year).mean(), 'o-')
plt.plot((SoGrunoff.groupby(SoGrunoff.index.year).mean()-4800)/100, '+-')
#plt.plot(20*(densityforcing.groupby(densityforcing.index.year).mean()-1.52), 's-')
#plt.plot(np.arange(2015, 2019), -(SoGrunoff.groupby(SoGrunoff.index.year).mean().values.transpose()[0]-4800)/100
# -low_pass_wind.wind.groupby(low_pass_wind.index.year).mean().values, '*--')
Out[81]:
[<matplotlib.lines.Line2D at 0x7f61dc63b8b0>]
In [ ]: