Cheas 2006 Meeting

1
Cheas 2006 Meeting Marek Uliasz Estimation of
regional fluxes of CO2
2

• Problems of regional scale CO2 flux estimations
  by inversions
• limited domain
• domain coverage by tower data

3
ASSUMPTION SiB-RAMS is capable to
realistically reproduce diurnal cycle and
spatial distribution of CO2 (assimilation and
respiration) fluxes. Therefore, observation data
are used to correct those fluxes for errors in
atmospheric transport.
4
time independent corrections to be estimated from
concentration data for each inversion cycle
CO2 flux
respiration assimilation fluxes simulated by
SiB-RAMS
5
time independent corrections to be estimated from
concentration data for each inversion cycle
CO2 flux
respiration assimilation fluxes simulated by
SiB-RAMS
6
MODELING FRAMEWORK
SiB-RAMS
typically run with several nested grids covering
a continental scale
meteo fields
CO2 fields and fluxes
LPDM
run on any subdomain extracted from SiB-RAMS
influence functions
CO2 observations
inversion techniques
Bayesian
MLEF
corrected CO2 fluxes
corrected within each inversion cycle
7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
sunrise
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
influence functions for 396m WLEF tower
integrated over unit flux for 7x10 day inversion
cycles
22
Implementation for a given inversion cycle
C observed concentration k index over
observations (sampling times and towers) i
index over source grid cell (both respiration
assimilation fluxes) CR.A influence function
integrated with respiration assimilation
fluxes CIN background concentration combining
effect of the flow across lateral
boundaries and initial concentration at the cycle
start betas corrections to be estimated
23
INVERSION EXPERIMENTS
SiB-RAMS simulation 75 days starting on April
25th, 2004 on two nested grids (10 km grid
spacing on the finer grid)
24
INVERSION EXPERIMENTS
SiB-RAMS simulation 75 days starting on April
25th, 2004 on two nested grids (10 km grid
spacing on the finer grid)
LPDM and influence function domain 600x600km
centered at WLEF tower
25
INVERSION EXPERIMENTS
SiB-RAMS simulation 75 days starting on April
25th, 2004 on two nested grids (10 km grid
spacing on the finer grid)
LPDM and influence function domain 600x600km
centered at WLEF tower
Concentration pseudo-data were generated for
WLEF and the ring of towers from SiB-RAMS
assimilation and respiration fluxes using
correction values of 1
26
INVERSION EXPERIMENTS
SiB-RAMS simulation 75 days starting on April
25th, 2004 on two nested grids (10 km grid
spacing on the finer grid)
LPDM and influence function domain 600x600km
centered at WLEF tower
Concentration pseudo-data were generated for
WLEF and the ring of towers from SiB-RAMS
assimilation and respiration fluxes using
correction values of 1
Model-data mismatch error was assumed to be
higher for lower towers 1 ppm for towersgt100m,
1.5 ppm for towers gt 50m, and 3 ppm for towers lt
50m and very high values for short towers during
nighttime
27
INVERSION EXPERIMENTS
SiB-RAMS simulation 75 days starting on April
25th, 2004 on two nested grids (10 km grid
spacing on the finer grid)
LPDM and influence function domain 600x600km
centered at WLEF tower
Concentration pseudo-data were generated for
WLEF and the ring of towers from SiB-RAMS
assimilation and respiration fluxes using
correction values of 1
Model-data mismatch error was assumed to be
higher for lower towers 1 ppm for towersgt100m,
1.5 ppm for towers gt 50m, and 3 ppm for towers lt
50m and very high values for short towers during
nighttime
7 x 10 day inversion cycles were performed using
Bayesian inversion technique with concentration
pseudo data (initial corrections 0.75 and their
standard deviations 0.1)
28
source area 20x20 km NW of WLEF
10 day (cycle) average
29
source area 20x20 km NW of WLEF
24 hour average
30
source area 20x20 km NW of WLEF
hourly average
31
NEE uncertainty reduction umol/m2/s
cycle 1
32
NEE uncertainty reduction umol/m2/s
cycle 2
33
NEE uncertainty reduction umol/m2/s
cycle 3
34
NEE uncertainty reduction umol/m2/s
cycle 4
35
NEE uncertainty reduction umol/m2/s
cycle 5
36
NEE uncertainty reduction umol/m2/s
cycle 6
37
NEE uncertainty reduction umol/m2/s
cycle 7
38
NEE UNCERTAINTY INITIAL, WLEF, RING
aggregation of source areas
39
NEE UNCERTAINTY INITIAL, WLEF, RING
aggregation of source areas
40
NEE UNCERTAINTY INITIAL, WLEF, RING
aggregation of source areas
41
Implementation for a given inversion cycle
C observed concentration k index over
observations (sampling times and towers) i
index over source grid cell (both respiration
assimilation fluxes) CR.A influence function
integrated with respiration assimilation
fluxes CIN background concentration combining
effect of the flow across lateral
boundaries and initial concentration at the cycle
start betas corrections to be estimated
42
Implementation for a given inversion cycle
C observed concentration k index over
observations (sampling times and towers) i
index over source grid cell (both respiration
assimilation fluxes) l - index over time
intervals CR.A influence function integrated
with respiration assimilation fluxes CIN
background concentration combining effect of the
flow across lateral boundaries and
initial concentration at the cycle start betas
corrections to be estimated
43
Inversion experiments

• Pseudo-data experiments
• The ring of towers (Bayesian, MLEF)
• US continental scale (MLEF)
• Real data experiments
• The ring of towers (Bayesian)

new SiB-RAMS simulations
44
RUC-LPDM

• Influence functions to be integrated with user
  provided CO2 fluxes