N 2 O station timeseries 2006-2012

Task 15.2: Modeling of N2O
modeling
JRC-IES (P. Bergamaschi, E. Koffi, M. Alexe, G. Maenhout)
MPG (U. Karstens)
CEA (M. Saunois, I. Mappe-Fogaing, P. Bousquet)
MET (A. Manning)
ECN (A. Vermeulen)
measurements
InGOS: InGOS NA2
LSCE (RAMCES, aircraft): M. Schmidt, M. Ramonet
NOAA: E. Dlugokencky
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
1
N2O inversions
station list
a priori inventory
period
InGOS data
NOAA+LSCE
flask
S1-N2O
N2O_001
EDGARv4.2FT-InGOS
2006-2012
2014 release
x
S2-N2O
N2O_002
EDGARv4.2FT-InGOS
2010-2012
2014 release
x
S3-N2O
N2O_002
no a priori
2010-2012
2014 release
x
S4-N2O
N2O_004
EDGARv4.2FT-InGOS
2010-2012
2014 release
-
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
2
European N2O emissions 2006-2011 S1
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
3
European N2O emissions 2010-2011 S1
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
4
European N2O emissions 2010-2011 S2
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
5
European N2O emissions 2010-2011 S4
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
6
European N2O emissions 2010-2011 S3
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
7
Global N2O emissions 2010-2011 S2
TM5 a priori:
global total: 24.0 Tg N2O yr-1
TM5 S2-N2O:
global total: 25.7 Tg N2O yr-1
TRANSCOM [Thompson et al., 2014]
global total (2006-2009):
22.2-30.5 Tg N2O yr-1
LMDZ S2-N2O:
global total: 12.1 Tg N2O yr-1
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
8
European N2O emissions - country totals NWE
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
9
European N2O emissions - country totals NEE
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
10
N2O station timeseries 2006-2012 (1)
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
N2O station timeseries 2006-2012 (2)
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
12
N2O station timeseries 2006-2012 (3)
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
13
N2O station timeseries 2006-2012 (4)
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
14
N2O station timeseries 2006-2012 (5)
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
15
N2O station timeseries 2006-2012 (6)
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
16
N2O station timeseries 2006-2012 (7)
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
17
N2O station timeseries 2006-2012 (8)
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
18
N2O station timeseries 2006-2012 (9)
bias (OBS-TM5)
JFJ
2006
-0.7
2007
-0.4
2008
-0.1
2009
-0.1
2010
0.3
2011
0.4
2012
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
19
N2O station timeseries 2006-2012 (10)
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
20
N2O station timeseries 2006-2012 (11)
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
21
N2O station timeseries LSCE flask sites
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
22
N2O station timeseries LSCE flask sites
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
23
N2O station timeseries 2011 ZEP
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
24
N2O station timeseries 2011 PAL
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
25
N2O station timeseries 2011 MHD
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
26
N2O station timeseries 2011 BI5
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
27
N2O station timeseries 2011 CB4
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
28
N2O station timeseries 2011 OX3
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
29
N2O station timeseries 2011 HEI
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
30
N2O station timeseries 2011 GIF
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
31
N2O station timeseries 2011 TR4
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
32
N2O station timeseries 2011 SCH
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
33
N2O station timeseries 2011 JFJ
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
34
N2O station timeseries 2011 PUY
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
35
statistics S1-N2O 2006-2011
all data (24h)
assimilated stations
validation stations
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
36
statistics S1-N2O 2010-2011
all data (24h)
assimilated stations
validation stations
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
37
statistics S1-N2O 2010-2011 (SEL)
surface stations: [12:00-15:00]
mountain stations: [0:00-3:00]
assimilated stations
validation stations
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
38
statistics S2-N2O 2010-2011
all data (24h)
assimilated stations
validation stations
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
39
statistics S4-N2O 2010-2011
all data (24h)
assimilated stations
validation stations
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
40
conclusions
• first results from first N2O inversions series
• model issues:
- deficiencies to simulate diurnal cycle, especially at continental surface stations (with
low sampling heights above ground) (TM5,....): nocturnal enrichment, timing (decrease
during daytime)
-> vertical gradients (and their temporal evolution)
-> validation of vertical profiles essential (profiles at towers, aircraft)
- LMDZ very low global total / European N2O emissions
- treatment of errors
• measurment issues:
- several stations have suspicous deviations from MHD base line
- time dependent bias of JFJ
- bias of LSCE flask data vs. INGOS / NOAA
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
InGOS uncertainty estimates
InGOS uncertainties:
(1) repeatability
(2) reproducibility
(3) lab internal scale consistency
(4) flask comparison uncertainty
(5) scale transfer uncertainty
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
statistical uncertainties
Statistical uncertainties
The statistical uncertainty of the individual measurement is given by the "repeatability".
This should be directly used as observation error (diagonal element in observations error
co-variance matrix; no temporal correlation).
-> implemented / used in TM5-4DVAR, STILT, LMDZ(?)
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
systematic uncertainties
(1) use estimates of systematic errors for data selection
E.g. a threshold could be set, and measurements, for which the systematic errors exceed this threshold,
will be rejected (i.e. not used in the inversion).
(2) treatment as temporally correlated errors
in principal possible, e.g. in 4DVAR systems [Järvinen et al., 1999].
however, not yet implemented in any (?) of the InGOS inverse modelling systems;
major programming effort, probably beyond the InGOS resources.
(3) Error inflation.
A simplified method to treat temporally correlated errors is the use of error inflation (e.g. [Chevallier,
2007]). Errors are treated as the random errors (described by diagonal element in observations error
co-variance), but scaled ("inflated") by a factor taking into account the temporal correlation (~
𝑛 ; n:
number of observations within temporal correlation time scale) .
(4) Bias correction (N2O only)
Ideally, the inversions should be performed without bias correction. In case that bias corrections are
applied in the inversion, the "scale transfer uncertainty" should be used as estimate of maximum bias
correction in models.
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
systematic uncertainties
- Temporal correlation time scale (for option (2) / (3)): How to estimate realistic temporal correlation
time scales ?
- Combination of different systematic error components. How to combine the different systematic error
components ? Problem is also that not all error components are reported for all stations. (2) + (3) only
available for stations with 'target gas' measurements. (4) only available for stations with parallel flask
measurements.
- Lab-internal scale consistency: This is considered as indicator for internal consistency of long-term
record (potentially concentration-dependent non-correctable error in the data) – not quantitative
- flask comparison uncertainty: if small then indicator for good data, if large, then reason not
immediately clear (difference between InGOS and flask data could also be due to errors of flask
measurements)
- use of NOAA + LSCE flask samples: How to treat errors of these flask measurements ? NOAA CH4 flask
data include "analysis_uncertainty" (but NOAA N2O and LSCE CH4 / N2O data don't include this yet).
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014
JRA3 deliverables
Deliverable Title
Delivery
Date
D15.1
Improved bottom up inventories for European CH4 emissions
12
D15.2
Improved bottom up inventories for European N2O emissions
12
D15.3
222Rn emission inventory (parameterized by soil type, porosity, moisture and water table depth
18
D15.4
Comparison of 222Rn simulations based on new 222Rn inventory (D15.3) with observations
24
D15.5
Comparison of simulated and observed boundary layer height
24
D15.6
Provision of 3D CH4 fields from CH4 inversions for comparison with FTIR and satellite data
24
D15.7
Model assessment of the potential to use δ13CH4
24
D15.8
Analysis of sensitivity of the InGOS network to European emissions
36
D15.9
European CH4 inversions using improved CH4 measurements from INGOS WP 2 and 3 (NA2 and NA3)
42
D15.10
European N2O inversions using improved N2O measurements from INGOS WP 2 and 3 (NA2 and NA3)
42
D15.11
European halocarbon inversions using improved halocarbon measurements at InGOS stations
42
D15.12
Detailed model intercomparisons and analysis of European CH4 emissions
45
D15.13
Model intercomparisons and analysis of European N2O emissions based on results from D15.10
45
D15.14
Model intercomparison and analysis of European emissions of important halocarbons with large GWPs
45
InGOS 2nd Periodic Meeting
Firenze, 14-16 October 2014