def test_Bockmon_Data_CBsys(self):
# Measured data from paper
batch_A = {"S": 33.190, "TA": 2215.08, "DIC": 2015.72, "pH": 7.8796}
batch_B = {"S": 33.186, "TA": 2216.26, "DIC": 2141.94, "pH": 7.5541}
pH = np.array([batch_A["pH"], batch_B["pH"]])
TA = np.array([batch_A["TA"], batch_B["TA"]])
DIC = np.array([batch_A["DIC"], batch_B["DIC"]])
S = np.array([batch_A["S"], batch_B["S"]])
BT = 433.0
# Csys calculations
# TA from pH and DIC
cTA = CBsys(pHtot=pH, DIC=DIC, BT=BT, S_in=S)
# Calculate % differences from measured
dTA = 100 * (TA - cTA.TA) / TA
self.assertLess(max(abs(dTA)), 0.2, msg="TA from DIC and pH")
# pH from TA and DIC
cpH = CBsys(DIC=DIC, TA=TA, BT=BT, S_in=S)
# Calculate % differences from measured
dpH = 100 * (pH - cpH.pHtot) / pH
self.assertLess(max(abs(dpH)), 0.2, msg="pH from TA and DIC")
# DIC from pH and TA
cDIC = CBsys(pHtot=pH, TA=TA, BT=BT, S_in=S)
# Calculate % differences from measured
dDIC = 100 * (DIC - cDIC.DIC) / DIC
self.assertLess(max(abs(dDIC)), 0.2, msg="DIC from TA and pH")
return
def test_Bockmon_Data_CBsys(self):
# Measured data from paper
batch_A = {'S': 33.190, 'TA': 2215.08, 'DIC': 2015.72, 'pH': 7.8796}
batch_B = {'S': 33.186, 'TA': 2216.26, 'DIC': 2141.94, 'pH': 7.5541}
pH = np.array([batch_A['pH'], batch_B['pH']])
TA = np.array([batch_A['TA'], batch_B['TA']])
DIC = np.array([batch_A['DIC'], batch_B['DIC']])
S = np.array([batch_A['S'], batch_B['S']])
BT = 433.
# Csys calculations
# TA from pH and DIC
cTA = CBsys(pHtot=pH, DIC=DIC, BT=BT, S=S)
# Calculate % differences from measured
dTA = (100 * (TA - cTA.TA) / TA)
self.assertLess(max(abs(dTA)), 0.2, msg='TA from DIC and pH')
# pH from TA and DIC
cpH = CBsys(DIC=DIC, TA=TA, BT=BT, S=S)
# Calculate % differences from measured
dpH = (100 * (pH - cpH.pHtot) / pH)
self.assertLess(max(abs(dpH)), 0.2, msg='pH from TA and DIC')
# DIC from pH and TA
cDIC = CBsys(pHtot=pH, TA=TA, BT=BT, S=S)
# Calculate % differences from measured
dDIC = (100 * (DIC - cDIC.DIC) / DIC)
self.assertLess(max(abs(dDIC)), 0.2, msg='DIC from TA and pH')
return
def test_Lueker_Data_CBsys(self):
"""
Need to incorporate nutrients!
"""
ld = pd.read_csv("cbsyst/test_data/Lueker2000/Lueker2000_Table3.csv",
comment="#")
# Calculate using cbsys
# TA from DIC and fCO2
cTA = CBsys(
DIC=ld.DIC.values,
fCO2=ld.fCO2.values,
T_in=ld.Temp.values,
S_in=ld.Sal.values,
)
dTA = ld.TA - cTA.TA
dTA_median = np.median(dTA)
dTA_pc95 = np.percentile(dTA, [2.5, 97.5])
self.assertLessEqual(abs(dTA_median), 2.5, msg="TA Offset <= 2.5")
self.assertTrue(all(abs(dTA_pc95 - dTA_median) <= 16),
msg="TA 95% Conf <= 16")
# fCO2 from TA and DIC
cfCO2 = CBsys(TA=ld.TA.values,
DIC=ld.DIC.values,
T_in=ld.Temp.values,
S_in=ld.Sal.values)
dfCO2 = ld.fCO2 - cfCO2.fCO2
dfCO2_median = np.median(dfCO2)
# dfCO2_pc95 = np.percentile(dfCO2, [2.5, 97.5])
dfCO2_percent_offset = 100 * dfCO2 / ld.fCO2
self.assertLessEqual(dfCO2_median, 2.5, msg="fCO2 Offset <= 2.5")
self.assertLessEqual(np.std(dfCO2_percent_offset),
3,
msg="fCO2 STD within 3%")
# print(dfCO2_pc95)
# self.assertTrue(all(abs(dfCO2_pc95) <= 70), msg='fCO2 95% Conc <= 70')
# DIC from TA and fCO2
cDIC = CBsys(
TA=ld.TA.values,
fCO2=ld.fCO2.values,
T_in=ld.Temp.values,
S_in=ld.Sal.values,
)
dDIC = ld.DIC - cDIC.DIC
dDIC_median = np.median(dDIC)
dDIC_pc95 = np.percentile(dDIC, [2.5, 97.5])
self.assertLessEqual(abs(dDIC_median), 2, msg="DIC Offset <= 2")
self.assertTrue(all(abs(dDIC_pc95) <= 15), msg="DIC 95% Conc <= 15")
return
def test_GLODAPv2_CBsys(self):
"""
Test Csys against GLODAP data (n = 83,030).
Check median offsets are within acceptable limits.
Check 95% confidence of residuals are within acceptable limits.
"""
if not os.path.exists(
"cbsyst/test_data/GLODAP_data/GLODAPv2_pH_DIC_ALK_subset.csv"):
get_GLODAP(path="cbsyst/test_data/GLODAP_data/", leave_zip=True)
# load GLODAP data
gd = pd.read_csv(
"cbsyst/test_data/GLODAP_data/GLODAPv2_pH_DIC_ALK_subset.csv")
gd.dropna(
subset=[
"phtsinsitutp",
"temperature",
"salinity",
"tco2",
"talk",
"pressure",
"phosphate",
"silicate",
],
inplace=True,
)
gd.pressure /= 10 # convert pressure to bar
# exclude weird cruise 270 data
gd = gd.loc[gd.cruise != 270]
# calculate pH from TA and DIC
cpH = CBsys(
TA=gd.talk,
DIC=gd.tco2,
T_in=gd.temperature,
S_in=gd.salinity,
P_in=gd.pressure,
TP=gd.phosphate,
TSi=gd.silicate,
BT=415.7,
)
pH_resid = gd.phtsinsitutp - cpH.pHtot
pH_median = np.median(pH_resid)
pH_pc95 = np.percentile(pH_resid, [2.5, 97.5])
self.assertLessEqual(abs(pH_median), 0.005, msg="pH Offset <= 0.01")
self.assertTrue(all(abs(pH_pc95) <= 0.05), msg="pH 95% Conf <= 0.05")
# calculate TA from pH and DIC
cTA = CBsys(
pHtot=gd.phtsinsitutp,
DIC=gd.tco2,
T_in=gd.temperature,
S_in=gd.salinity,
P_in=gd.pressure,
TP=gd.phosphate,
TSi=gd.silicate,
BT=415.7,
)
TA_resid = gd.talk - cTA.TA
TA_median = np.median(TA_resid)
TA_pc95 = np.percentile(TA_resid, [2.5, 97.5])
self.assertLessEqual(abs(TA_median), 0.5, msg="TA Offset <= 2.5")
self.assertTrue(all(abs(TA_pc95) < 13), msg="TA 95% Conf <= 15")
# calculate DIC from TA and pH
cDIC = CBsys(
pHtot=gd.phtsinsitutp,
TA=gd.talk,
T_in=gd.temperature,
S_in=gd.salinity,
P_in=gd.pressure,
TP=gd.phosphate,
TSi=gd.silicate,
BT=415.7,
)
DIC_resid = gd.tco2 - cDIC.DIC
DIC_median = np.median(DIC_resid)
DIC_pc95 = np.percentile(DIC_resid, [2.5, 97.5])
self.assertLessEqual(abs(DIC_median), 0.5, msg="DIC Offset <= 2")
self.assertTrue(all(abs(DIC_pc95) < 13), msg="DIC 95% Conf <= 15")
return
def test_GLODAPv2_CBsys(self):
"""
Test Csys against GLODAP data (n = 83,030).
Check median offsets are within acceptable limits.
Check 95% confidence of residuals are within acceptable limits.
"""
# load GLODAP data
gd = pd.read_csv(
'cbsyst/test_data/GLODAP_data/GLODAPv2_pH_DIC_ALK_subset.csv')
gd.dropna(subset=[
'phtsinsitutp', 'temperature', 'salinity', 'tco2', 'talk',
'pressure', 'phosphate', 'silicate'
],
inplace=True)
gd.pressure /= 10 # convert pressure to bar
# exclude weird cruise 270 data
gd = gd.loc[gd.cruise != 270]
# calculate pH from TA and DIC
cpH = CBsys(TA=gd.talk,
DIC=gd.tco2,
T=gd.temperature,
S=gd.salinity,
P=gd.pressure,
TP=gd.phosphate,
TSi=gd.silicate,
BT=415.7)
pH_resid = gd.phtsinsitutp - cpH.pHtot
pH_median = np.median(pH_resid)
pH_pc95 = np.percentile(pH_resid, [2.5, 97.5])
self.assertLessEqual(abs(pH_median), 0.005, msg='pH Offset <= 0.01')
self.assertTrue(all(abs(pH_pc95) <= 0.05), msg='pH 95% Conf <= 0.05')
# calculate TA from pH and DIC
cTA = CBsys(pHtot=gd.phtsinsitutp,
DIC=gd.tco2,
T=gd.temperature,
S=gd.salinity,
P=gd.pressure,
TP=gd.phosphate,
TSi=gd.silicate,
BT=415.7)
TA_resid = gd.talk - cTA.TA
TA_median = np.median(TA_resid)
TA_pc95 = np.percentile(TA_resid, [2.5, 97.5])
self.assertLessEqual(abs(TA_median), 0.5, msg='TA Offset <= 2.5')
self.assertTrue(all(abs(TA_pc95) < 13), msg='TA 95% Conf <= 15')
# calculate DIC from TA and pH
cDIC = CBsys(pHtot=gd.phtsinsitutp,
TA=gd.talk,
T=gd.temperature,
S=gd.salinity,
P=gd.pressure,
TP=gd.phosphate,
TSi=gd.silicate,
BT=415.7)
DIC_resid = gd.tco2 - cDIC.DIC
DIC_median = np.median(DIC_resid)
DIC_pc95 = np.percentile(DIC_resid, [2.5, 97.5])
self.assertLessEqual(abs(DIC_median), 0.5, msg='DIC Offset <= 2')
self.assertTrue(all(abs(DIC_pc95) < 13), msg='DIC 95% Conf <= 15')
return