P1: A new real-time atmospheric CO2 forecast product
Anna Agusti-Panareda1, Sebastien Massart1, Gianpaolo Balsamo1, Anton Beljaars1, Souhail Boussetta1, Frederic Chevallier2, Richard J. Engelen1
1European
MACC-II has recently started producing forecasts of atmospheric CO2 abundance using the
infrastructure of the European Centre for Medium range Weather Forecasts (ECMWF)
Integrated Forecasting System (IFS). The CO2 forecasts are provided globally in near-realtime since October 2012 with a five-day lead time:
Centre for Medium range Weather Forecasts, Reading, UK, 2Laboratoire des Sciences du Climat et l'Environnement, Gif-sur-Yvette, France
Evaluation of the CO2 forecast
The phase and amplitude of the CO2 seasonal cycle vary with latitude. The model
is evaluated using the NOAA GLOBALVIEW-CO2 (2011) dataset which provides the
integrated effects of surface CO2 fluxes over large regions at different latitudinal
bands.
http://www.gmes-atmosphere.eu/d/services/gac/nrt/nrt_fields_co2
At first glance, the seasonal cycle phase and
amplitude and latitude dependence shown in Figure
3 appear to be reasonably well represented in the
forecast. However, there are clear discrepancies
between the forecast and GLOBALVIEW-CO2 product
in the northern hemisphere:
MACC-II
The pre-operational Monitoring of Atmospheric Composition and Climate - Interim
Implementation service (MACC-II, http://www.gmes-atmosphere.eu/) provides data records on
atmospheric composition for recent years, data for monitoring present conditions and forecasts
of the distribution of key constituents for a few days ahead.
♦ In the forecast, not enough CO2 is released before
and after the growing season (i.e. March to May and
October to December).
The forecast configuration
♦The onset of the CO2 sink associated with the
growing season starts too early in the forecast (e.g.
the sharp CO2 decrease in mid-latitudes depicted by
GLOBALVIEW-CO2 product in June starts in May in
the forecast). This also leads to a longer growing
season in the forecast.
• The CO2 forecast relies on
prescribed and modelled CO2
fluxes as well as the transport
from the IFS model.
Figure 1: Set up of real time CO2 forecast
• The IFS transport model is the
state-of-the-art Numerical
Weather Prediction (NWP)
model from ECMWF. It has a
semi-lagrangian advection
scheme, a mass flux
convection scheme and a
turbulence mixing scheme.
• Thanks to the CTESSEL Carbon module development in the IFS, as part of the Geoland-2 project
(http://www.gmes-geoland.info/), the net ecosystem exchange (NEE) is now available in nearreal-time (Boussetta et al., 2012). The biomass burning fluxes from the MACC-II Global Fire
Assimilation System (GFAS) are also used in near-real-time (Kaiser et al., 2012). The ocean fluxes
are from the Takahashi et al. (2009) climatology . The anthropogenic emissions are from the
EDGARv4.2 inventory (http://edgar.jrc.ec.europa.eu/) with an extrapolation of the last year
available (2008) based on a climatology of its global trend .
• The forecast is run at two resolutions: the resolution of ERA-Interim (T255,~77km, L60) and the
resolution of the operational NWP model (T1279,~16 km, L91).
The combination of these two factors is consistent
with the predominantly negative global annual bias
shown in Figure 2.
Figure 3: NOAA-GLOBALVIEW CO2 product for 2010 based on observations, (b) equivalent product based on the 24 hour
forecast of CO2 and (c) the difference between GLOBALVIEW and the forecast. The CO2 forecast has been sampled at the
same locations as the GLOBALVIEW observations and the same data processing described in Masarie and Tans (1995,
J.Geophys. Res., 100, 11593-11610) has been applied.
The cloudy warm conveyor belts in the mid-latitude low pressure systems are
associated with changes in temperature and solar radiation at the surface which
in turn increase the net ecosystem exchange. This increase can be linked with
the following:
• A decrease in the photosynthetic uptake following a decrease in radiation
(e.g. 3 and 7 September)
• An increase in ecosystem respiration following an increase in temperature
(e.g. 21 September)
• Both, a simultaneous decrease in the vegetation uptake and increase in
ecosystem respiration due to a concurrent decrease in radiation and increase
in temperature (e.g. 11 and 23 to 24 September)
Figure 5: (a) Daily mean biogenic fluxes: net
ecosystem exchange in cyan, photosynthetic
uptake in green and ecosystem respiration in red
[kg m-2 s-1]. (b) Daily mean 2-metre temperature
[K]. (c) Daily mean downward solar radiation at
the surface [J m-2] from the 24-hour forecast at
Park Falls in September 2010.
MACC-II collaborates with the Integrated Carbon Observation System (ICOS, https://icos-atc-demo.lsce.ipsl.fr/homepage) atmospheric network (Figure 6), which provides in-situ observations of CO2 in near-realtime. The observations help to validate the model and the model helps to interpret the observations.
The high and low resolution forecasts are monitored online in near-real-time with the
ICOS observations (Figure 7):
www.copernicus-atmosphere.eu/d/services/gac/verif/ghg/icos
With the high resolution, the CO2
forecast is closer to the observed CO2
particularly at:
Figure 6: Location of stations from the pre-operational
ICOS network providing observations of CO2 dry molar
fraction near the surface in near-real-time during 2012.
Applications for the CO2 forecast
• Coastal sites (MHD, IVI)
• Mountain sites (PUY)
• Sites close to anthropogenic
sources (CBW).
Figure 7: Time series of daily mean CO2 dry mole fraction [ppm] from the NRT CO2 forecast at low resolution (cyan) and high resolution (red) for 2012 at
different ICOS stations.
References
The NRT CO2 forecast can be used in several domains, for example:
• CO2 boundary conditions for regional modelling and flux inversions.
• Supporting the interpretation and the quality control of observations
•
•
•
•
The synoptic variability of CO2 associated with the passage of low pressure
systems is well captured by the forecast . Figures 3 and 4 illustrate this at the
NOAA/ESRL tall tower in Park Falls (Winsconsin, USA) in September 2010.
Monitoring of CO2 forecast in near-real time using the ICOS network: Sensitivity to model resolution
• As there are no CO2 observations used in the forecast, the global budget is not constrained. Thus,
there is a difference between the total flux from the model and the observed atmospheric
growth (ftp://ftp.cmdl.noaa.gov/ccg/co2/trends/co2_gr_gl.txt). This induces a global bias which
varies from year to year.
Figure 2: Time series of CO2 annual global budget
[ppm/year] of atmospheric growth estimated by
NOAA/ESRL (black) and total flux in the model (grey).
The different sources/sinks used in the model are
depicted by the coloured lines : anthropogenic fluxes
(purple), biomass burning (red), ocean sink (blue) and
net ecosystem exchange (green).
Figure 4: (a) CO2 dry molar fraction anomalies [ppm].
Areas above 392ppm at 10-m level (grey), and at 850
hPa level (cyan), and areas above 388ppm level at 500
hPa level and 300 hPa level. Black contours depict mean
seal level pressure and the location of Park Falls is
marked by red triangle. (b) ECMWF surface pressure
forecast [hPa]. (c) Daily mean CO2 dry molar fraction
[ppm] from the 24-hour forecast in cyan and
observations in black at the top level (396m) of the
NOAA/ESRL tall tower at Park Falls (Winsconsin, USA) in
September 2010 (Andrews et al., 2013, Atmos. Meas.
Tech. Discuss., 6, 1461-1553).
via monitoring activities.
Supporting the planning of field experiments.
Providing prior information for CO2 and CH4 satellite retrievals.
Assimilation of near real time CO2 observations (e.g. from the ICOS network). Figure 8: Atmospheric CO2 fields from the high resolution (left)
and low resolution (right) forecasts showing the landfall of
Improving the modelling of the radiative transfer and evapotranspiration in
Hurricane Sandy (2012). The shaded areas indicate dry molar
NWP analysis and forecast.
fractions larger than 392 ppm for different vertical levels (see
title).
•Boussetta, S., G. Balsamo, A. Beljaars, A. Agusti-Panareda, J.-C. Calvet, S. Lafont, M. Balzarolo, C. Jacobs, B. van den Hurk, P. Viterbo, L. Jarlan, G. van der Werf. Natural carbon dioxide exchanges in the ECMWF Integrated
Forecasting System: Implementation and Offline validation, J. Geophys. Res. (accepted).
•Kaiser, J.W., A. Heil, M. O. Andreae, A. Benedetti, N. Chubarova, L. Jones, J.-J. Morcrette, M. Razinger, M. G. Schultz, M. Suttie, and G. R. van der Werf. Biomass burning emissions estimated with a global fire assimilation
system based on observed fire radiative power, Biogeosciences, Vol. 9, Pages 527–554, 2012.
•Takahashi, T., S. C. Sutherland, R. Wanninkhof, C. Sweeney, et al. Climatological mean and decadal change in surface ocean pCO2, and net sea–air CO2 flux over the global oceans, Deep Sea Research Part II: Topical Studies in
Oceanography, Vol. 56, Issues 8–10, Pages 554–577, April 2009.
•GLOBALVIEW-CO2, 2011: Cooperative Atmospheric Data Integration Project - Carbon Dioxide. NOAA ESRL, Boulder, Colorado [Available at http://www.esrl.noaa.gov/gmd/ccgg/globalview/].
Acknowledgements MACC-II is funded by the European Commission under the Seventh Research Framework Programme. Thanks to Arlyn Andrews, Pieter Tans for providing the NOAA/ESRL data from tall tower at Parkfalls
(Winsconsin, USA). Thanks to Jérôme Tarniewicz (ICOS Atmospheric Thematic Center), Philippe Ciais and Michel Ramonet (Laboratoire des Sciences du Climat et l'Environnement) for providing the data for MACC-II from the
website at https://icos-atc-demo.lsce.ipsl.fr and Alex Vermeulen (Energy research Centre of the Netherlands), the Principal Investigator from Cabauw station. The authors acknowledge the European Commission for the support
of the preparatory phase of ICOS (2008–2013) and the Netherlands Ministry of IenM and ECN for the support of the observations at Cabauw.