scholarly journals Technical note: A high-resolution inverse modelling technique for estimating surface CO<sub>2</sub> fluxes based on the NIES-TM–FLEXPART coupled transport model and its adjoint

2021 ◽  
Vol 21 (2) ◽  
pp. 1245-1266
Author(s):  
Shamil Maksyutov ◽  
Tomohiro Oda ◽  
Makoto Saito ◽  
Rajesh Janardanan ◽  
Dmitry Belikov ◽  
...  
Keyword(s):  
High Resolution ◽  
Transport Model ◽  
Dispersion Model ◽  
Surface Flux ◽  
Particle Dispersion ◽  
Anthropogenic Emissions ◽  
Coupled Transport ◽  
Tracer Transport ◽  
Flux Corrections

Abstract. We developed a high-resolution surface flux inversion system based on the global Eulerian–Lagrangian coupled tracer transport model composed of the National Institute for Environmental Studies (NIES) transport model (TM; collectively NIES-TM) and the FLEXible PARTicle dispersion model (FLEXPART). The inversion system is named NTFVAR (NIES-TM–FLEXPART-variational) as it applies a variational optimization to estimate surface fluxes. We tested the system by estimating optimized corrections to natural surface CO2 fluxes to achieve the best fit to atmospheric CO2 data collected by the global in situ network as a necessary step towards the capability of estimating anthropogenic CO2 emissions. We employed the Lagrangian particle dispersion model (LPDM) FLEXPART to calculate surface flux footprints of CO2 observations at a spatial resolution of 0.1∘×0.1∘. The LPDM is coupled with a global atmospheric tracer transport model (NIES-TM). Our inversion technique uses an adjoint of the coupled transport model in an iterative optimization procedure. The flux error covariance operator was implemented via implicit diffusion. Biweekly flux corrections to prior flux fields were estimated for the years 2010–2012 from in situ CO2 data included in the Observation Package (ObsPack) data set. High-resolution prior flux fields were prepared using the Open-Data Inventory for Anthropogenic Carbon dioxide (ODIAC) for fossil fuel combustion, the Global Fire Assimilation System (GFAS) for biomass burning, the Vegetation Integrative SImulator for Trace gases (VISIT) model for terrestrial biosphere exchange, and the Ocean Tracer Transport Model (OTTM) for oceanic exchange. The terrestrial biospheric flux field was constructed using a vegetation mosaic map and a separate simulation of CO2 fluxes at a daily time step by the VISIT model for each vegetation type. The prior flux uncertainty for the terrestrial biosphere was scaled proportionally to the monthly mean gross primary production (GPP) by the Moderate Resolution Imaging Spectroradiometer (MODIS) MOD17 product. The inverse system calculates flux corrections to the prior fluxes in the form of a relatively smooth field multiplied by high-resolution patterns of the prior flux uncertainties for land and ocean, following the coastlines and fine-scale vegetation productivity gradients. The resulting flux estimates improved the fit to the observations taken at continuous observation sites, reproducing both the seasonal and short-term concentration variabilities including high CO2 concentration events associated with anthropogenic emissions. The use of a high-resolution atmospheric transport in global CO2 flux inversions has the advantage of better resolving the transported mixed signals from the anthropogenic and biospheric sources in densely populated continental regions. Thus, it has the potential to achieve better separation between fluxes from terrestrial ecosystems and strong localized sources, such as anthropogenic emissions and forest fires. Further improvements in the modelling system are needed as our posterior fit was better than that of the National Oceanic and Atmospheric Administration (NOAA)'s CarbonTracker for only a fraction of the monitoring sites, i.e. mostly at coastal and island locations where background and local flux signals are mixed.

scholarly journals Technical note: A high-resolution inverse modelling technique for estimating surface CO<sub>2</sub> fluxes based on the NIES-TM – FLEXPART coupled transport model and its adjoint

2020 ◽  
Author(s):  
Shamil Maksyutov ◽  
Tomohiro Oda ◽  
Makoto Saito ◽  
Rajesh Janardanan ◽  
Dmitry Belikov ◽  
...  
Keyword(s):  
High Resolution ◽  
Transport Model ◽  
Dispersion Model ◽  
Surface Flux ◽  
Particle Dispersion ◽  
Anthropogenic Emissions ◽  
Coupled Transport ◽  
Tracer Transport ◽  
Flux Corrections

Abstract. We developed a high-resolution surface flux inversion system based on the global Lagrangian–Eulerian coupled tracer transport model composed of National Institute for Environmental Studies Transport Model (NIES-TM) and FLEXible PARTicle dispersion model (FLEXPART). The inversion system is named NTFVAR (NIES-TM-FLEXPART-variational) as it applies variational optimisation to estimate surface fluxes. We tested the system by estimating optimized corrections to natural surface CO2 fluxes to achieve best fit to atmospheric CO2 data collected by the global in-situ network, as a necessary step towards capability of estimating anthropogenic CO2 emissions. We employ the Lagrangian particle dispersion model (LPDM) FLEXPART to calculate the surface flux footprints of CO2 observations at a 0.1° × 0.1° spatial resolution. The LPDM is coupled to a global atmospheric tracer transport model (NIES-TM). Our inversion technique uses an adjoint of the coupled transport model in an iterative optimization procedure. The flux error covariance operator is being implemented via implicit diffusion. Biweekly flux corrections to prior flux fields were estimated for the years 2010–2012 from in-situ CO2 data included in the Observation Package (ObsPack) dataset. High-resolution prior flux fields were prepared using Open-Data Inventory for Anthropogenic Carbon dioxide (ODIAC) for fossil fuel combustion, Global Fire Assimilation System (GFAS) for biomass burning, the Vegetation Integrative SImulator for Trace gases (VISIT) model for terrestrial biosphere exchange and Ocean Tracer Transport Model (OTTM) for oceanic exchange. The terrestrial biospheric flux field was constructed using a vegetation mosaic map and separate simulation of CO2 fluxes at daily time step by the VISIT model for each vegetation type. The prior flux uncertainty for terrestrial biosphere was scaled proportionally to the monthly mean Gross Primary Production (GPP) by the Moderate Resolution Imaging Spectroradiometer (MODIS) MOD17 product. The inverse system calculates flux corrections to the prior fluxes in the form of a relatively smooth field multiplied by high-resolution patterns of the prior flux uncertainties for land and ocean, following the coastlines and vegetation productivity gradients. The resulting flux estimates improve fit to the observations at continuous observations sites, reproducing both the seasonal variation and short-term concentration variability, including high CO2 concentration events associated with anthropogenic emissions. The use of high-resolution atmospheric transport in global CO2 flux inversion has the advantage of better resolving the transport from the mix of the anthropogenic and biospheric sources in densely populated continental regions and shows potential for better separation between fluxes from terrestrial ecosystems and strong localised sources such as anthropogenic emissions and forest fires. Further improvements in the modelling system are needed as the posterior fit is better than that by the National Oceanic and Atmospheric Administration (NOAA) CarbonTracker only for a fraction of the monitoring sites, mostly at coastal and island locations experiencing mix of background and local flux signals.

scholarly journals Inversion of long-lived trace gas emissions using combined Eulerian and Lagrangian chemical transport models

2011 ◽  
Vol 11 (18) ◽  
pp. 9887-9898 ◽  
Author(s):  
M. Rigby ◽  
A. J. Manning ◽  
R. G. Prinn
Keyword(s):  
High Resolution ◽  
Large Scale ◽  
Transport Model ◽  
Initial Conditions ◽  
Dispersion Model ◽  
Chemical Transport ◽  
Particle Dispersion ◽  
Global Network ◽  
Lagrangian Model ◽  
Chemical Transport Model

Abstract. We present a method for estimating emissions of long-lived trace gases from a sparse global network of high-frequency observatories, using both a global Eulerian chemical transport model and Lagrangian particle dispersion model. Emissions are derived in a single step after determining sensitivities of the observations to initial conditions, the high-resolution emissions field close to observation points, and larger regions further from the measurements. This method has the several advantages over inversions using one type of model alone, in that: high-resolution simulations can be carried out in limited domains close to the measurement sites, with lower resolution being used further from them; the influence of errors due to aggregation of emissions close to the measurement sites can be minimized; assumptions about boundary conditions to the Lagrangian model do not need to be made, since the entire emissions field is estimated; any combination of appropriate models can be used, with no code modification. Because the sensitivity to the entire emissions field is derived, the estimation can be carried out using traditional statistical methods without the need for multiple steps in the inversion. We demonstrate the utility of this approach by determining global SF6 emissions using measurements from the Advanced Global Atmospheric Gases Experiment (AGAGE) between 2007 and 2009. The global total and large-scale patterns of the derived emissions agree well with previous studies, whilst allowing emissions to be determined at higher resolution than has previously been possible, and improving the agreement between the modeled and observed mole fractions at some sites.

scholarly journals Regional CO<sub>2</sub> inversions with LUMIA, the Lund University Modular Inversion Algorithm, v1.0

2021 ◽  
Vol 14 (6) ◽  
pp. 3383-3406
Author(s):  
Guillaume Monteil ◽  
Marko Scholze
Keyword(s):  
Inverse Modeling ◽  
Transport Model ◽  
Dispersion Model ◽  
Particle Dispersion ◽  
Model Errors ◽  
Inversion Algorithm ◽  
In Situ Data ◽  
Increasing Demand ◽  
Modular Inversion

Abstract. Atmospheric inversions are used to derive constraints on the net sources and sinks of CO2 and other stable atmospheric tracers from their observed concentrations. The resolution and accuracy that the fluxes can be estimated with depends, among other factors, on the quality and density of the observational coverage, on the precision and accuracy of the transport model used by the inversion to relate fluxes to observations, and on the adaptation of the statistical approach to the problem studied. In recent years, there has been an increasing demand from stakeholders for inversions at higher spatial resolution (country scale), in particular in the framework of the Paris agreement. This step up in resolution is in theory enabled by the growing availability of observations from surface in situ networks (such as ICOS in Europe) and from remote sensing products (OCO-2, GOSAT-2). The increase in the resolution of inversions is also a necessary step to provide efficient feedback to the bottom-up modeling community (vegetation models, fossil fuel emission inventories, etc.). However, it calls for new developments in the inverse models: diversification of the inversion approaches, shift from global to regional inversions, and improvement in the computational efficiency. In this context, we developed LUMIA, the Lund University Modular Inversion Algorithm. LUMIA is a Python library for inverse modeling built around the central idea of modularity: it aims to be a platform that enables users to construct and experiment with new inverse modeling setups while remaining easy to use and maintain. It is in particular designed to be transport-model-agnostic, which should facilitate isolating the transport model errors from those introduced by the inversion setup itself. We have constructed a first regional inversion setup using the LUMIA framework to conduct regional CO2 inversions in Europe using in situ data from surface and tall-tower observation sites. The inversions rely on a new offline coupling between the regional high-resolution FLEXPART Lagrangian particle dispersion model and the global coarse-resolution TM5 transport model. This test setup is intended both as a demonstration and as a reference for comparison with future LUMIA developments. The aims of this paper are to present the LUMIA framework (motivations for building it, development principles and future prospects) and to describe and test this first implementation of regional CO2 inversions in LUMIA.

scholarly journals A global coupled Eulerian-Lagrangian model and 1 × 1 km CO<sub>2</sub> surface flux dataset for high-resolution atmospheric CO<sub>2</sub> transport simulations

2011 ◽  
Vol 4 (3) ◽  
pp. 2047-2080 ◽  
Author(s):  
A. Ganshin ◽  
T. Oda ◽  
M. Saito ◽  
S. Maksyutov ◽  
V. Valsala ◽  
...  
Keyword(s):  
High Resolution ◽  
High Frequency ◽  
Transport Model ◽  
Dispersion Model ◽  
Particle Dispersion ◽  
Surface Fluxes ◽  
Atmospheric Co2 ◽  
Lagrangian Model ◽  
Co2 Concentrations ◽  
Atmospheric Co2 Concentrations

Abstract. We designed a method to simulate atmospheric CO2 concentrations at several continuous observation sites around the globe using surface fluxes at a very high spatial resolution. The simulations presented in this study were performed using a Lagrangian particle dispersion model coupled to a global atmospheric tracer transport model with prescribed global surface CO2 flux maps at a 1 × 1 km resolution. The surface fluxes used in the simulations were prepared by assembling the individual components of terrestrial, oceanic and fossil fuel CO2 fluxes. This experimental setup (i.e., a transport model running at a medium resolution, coupled to a high-resolution Lagrangian particle dispersion model together with global surface fluxes at a very high resolution), which was designed to represent high-frequency variations in atmospheric CO2 concentration, has not been reported at a global scale previously. Two sensitivity experiments were performed: (a) using the global transport model without coupling to the Lagrangian dispersion model, and (b) using the coupled model with a reduced resolution of surface fluxes, in order to evaluate the performance of Eulerian-Lagrangian coupling and the role of high-resolution fluxes in simulating high-frequency variations in atmospheric CO2 concentrations. A correlation analysis between observed and simulated atmospheric CO2 concentrations at selected locations revealed that the inclusion of both Eulerian-Lagrangian coupling and high-resolution fluxes improves the high-frequency simulations of the model. The results highlight the potential of a coupled Eulerian-Lagrangian model in simulating high-frequency atmospheric CO2 concentrations at many locations worldwide. The model performs well in representing observations of atmospheric CO2 concentrations at high spatial and temporal resolutions, especially for coastal sites and sites located close to sources of large anthropogenic emissions. While this study focused on simulations of CO2 concentrations, the model could be used for other atmospheric compounds with known estimated emissions.

scholarly journals A global coupled Eulerian-Lagrangian model and 1 × 1 km CO<sub>2</sub> surface flux dataset for high-resolution atmospheric CO<sub>2</sub> transport simulations

2012 ◽  
Vol 5 (1) ◽  
pp. 231-243 ◽  
Author(s):  
A. Ganshin ◽  
T. Oda ◽  
M. Saito ◽  
S. Maksyutov ◽  
V. Valsala ◽  
...  
Keyword(s):  
High Resolution ◽  
High Frequency ◽  
Transport Model ◽  
Dispersion Model ◽  
Particle Dispersion ◽  
Surface Fluxes ◽  
Atmospheric Co2 ◽  
Lagrangian Model ◽  
Co2 Concentrations ◽  
Atmospheric Co2 Concentrations

Abstract. We designed a method to simulate atmospheric CO2 concentrations at several continuous observation sites around the globe using surface fluxes at a very high spatial resolution. The simulations presented in this study were performed using the Global Eulerian-Lagrangian Coupled Atmospheric model (GELCA), comprising a Lagrangian particle dispersion model coupled to a global atmospheric tracer transport model with prescribed global surface CO2 flux maps at a 1 × 1 km resolution. The surface fluxes used in the simulations were prepared by assembling the individual components of terrestrial, oceanic and fossil fuel CO2 fluxes. This experimental setup (i.e. a transport model running at a medium resolution, coupled to a high-resolution Lagrangian particle dispersion model together with global surface fluxes at a very high resolution), which was designed to represent high-frequency variations in atmospheric CO2 concentration, has not been reported at a global scale previously. Two sensitivity experiments were performed: (a) using the global transport model without coupling to the Lagrangian dispersion model, and (b) using the coupled model with a reduced resolution of surface fluxes, in order to evaluate the performance of Eulerian-Lagrangian coupling and the role of high-resolution fluxes in simulating high-frequency variations in atmospheric CO2 concentrations. A correlation analysis between observed and simulated atmospheric CO2 concentrations at selected locations revealed that the inclusion of both Eulerian-Lagrangian coupling and high-resolution fluxes improves the high-frequency simulations of the model. The results highlight the potential of a coupled Eulerian-Lagrangian model in simulating high-frequency atmospheric CO2 concentrations at many locations worldwide. The model performs well in representing observations of atmospheric CO2 concentrations at high spatial and temporal resolutions, especially for coastal sites and sites located close to sources of large anthropogenic emissions. While this study focused on simulations of CO2 concentrations, the model could be used for other atmospheric compounds with known estimated emissions.

scholarly journals Simulation of variability in atmospheric carbon dioxide using a global coupled Eulerian – Lagrangian transport model

2011 ◽  
Vol 4 (2) ◽  
pp. 317-324 ◽  
Author(s):  
Y. Koyama ◽  
S. Maksyutov ◽  
H. Mukai ◽  
K. Thoning ◽  
P. Tans
Keyword(s):  
Atmospheric Chemistry ◽  
Transport Model ◽  
Dispersion Model ◽  
Regional Scale ◽  
Coupled Model ◽  
Particle Dispersion ◽  
Atmospheric Composition ◽  
Tracer Transport ◽  
Eulerian Model ◽  
Co2 Concentrations

Abstract. This study assesses the advantages of using a coupled atmospheric-tracer transport model, comprising a global Eulerian model and a global Lagrangian particle dispersion model, to improve the reproducibility of tracer-gas variations affected by the near-field surface emissions and transport around observation sites. The ability to resolve variability in atmospheric composition on an hourly time-scale and a spatial scale of several kilometers would be beneficial for analyzing data from continuous ground-based monitoring and from upcoming space-based observations. The coupled model yields an increase in the horizontal resolution of transport and fluxes, and has been tested in regional-scale studies of atmospheric chemistry. By applying the Lagrangian component to the global domain, we extend this approach to the global scale, thereby enabling computationally efficient global inverse modeling and data assimilation. To validate the coupled model, we compare model-simulated CO2 concentrations with continuous observations at three sites: two operated by the National Oceanic and Atmospheric Administration, USA, and one operated by the National Institute for Environmental Studies, Japan. As the goal of this study is limited to introducing the new modeling approach, we selected a transport simulation at these three sites to demonstrate how the model may perform at various geographical areas. The coupled model provides improved agreement between modeled and observed CO2 concentrations in comparison to the Eulerian model. In an area where variability in CO2 concentration is dominated by a fossil fuel signal, the correlation coefficient between modeled and observed concentrations increases by between 0.05 to 0.1 from the original values of 0.5–0.6 achieved with the Eulerian model.

Natural and anthropogenic methane emissions in West Siberia estimated using a wetland inventory, GOSAT and a regional tower network

2020 ◽  
Author(s):  
Shamil Maksyutov ◽  
Motoki Sasakawa ◽  
Rajesh Janardanan ◽  
Fenjuan Wang ◽  
Aki Tsuruta ◽  
...  
Keyword(s):  
High Resolution ◽  
West Siberia ◽  
Methane Emissions ◽  
Inverse Model ◽  
Anthropogenic Emissions ◽  
Taiga Zone ◽  
Tracer Transport ◽  
Bottom Up ◽  
Near Surface

&lt;p&gt;West Siberia contributes a large fraction of Russian methane emissions, with both natural emissions from peatlands and anthropogenic emissions by oil and gas industries. To quantify anthropogenic emissions with atmospheric observations and inventories, we must better understand the natural wetland emissions.&amp;#160; We combine high-resolution wetland mapping based on Landsat data for whole West Siberian lowland with a database of in situ flux measurements to derive bottom-up wetland emission estimates. We use a global high-resolution methane flux inversion based on a Lagrangian-Eulerian coupled tracer transport model to estimate methane emissions in West Siberia using atmospheric methane data collected at the Siberian GHG monitoring network JR-STATION, ZOTTO, data by the global in situ network and GOSAT satellite observations. High-resolution prior fluxes were prepared for anthropogenic emissions (EDGAR), biomass burning (GFAS), and wetlands (VISIT). A global high-resolution wetland emission dataset was constructed using 0.5-degree monthly emission data simulated by the VISIT model and wetland area fraction map by the Global Lake and Wetlands Database (GLWD). We estimate biweekly flux corrections to prior flux fields for 2010 to 2015. The inverse model optimizes corrections to two categories of fluxes: anthropogenic and natural (wetlands). Based on fitting the model simulations to the observations, the inverse model provides upward corrections to West Siberian anthropogenic emissions in winter and wetland emissions in summer. The use of high-resolution atmospheric transport in the flux inversion, when compared to low-resolution transport modeling, enables a better fit to observations in winter, when anthropogenic emissions dominate variability of the near-surface methane concentration. We estimate 15% higher anthropogenic emissions than EDGAR v.4.3.2 inventory for whole Russia, with most of the correction attributed to West Siberia and the European part of Russia. Comparison of the inversion estimates with the bottom-up wetland emission inventory for West Siberia suggests a need to adjust the wetland emissions to match observed north-south gradient of emissions with higher emissions in the southern taiga zone.&lt;/p&gt;

scholarly journals Inversion of long-lived trace gas emissions using combined Eulerian and Lagrangian chemical transport models

2011 ◽  
Vol 11 (5) ◽  
pp. 14689-14717 ◽  
Author(s):  
M. Rigby ◽  
A. J. Manning ◽  
R. G. Prinn
Keyword(s):  
High Resolution ◽  
Large Scale ◽  
Transport Model ◽  
Initial Conditions ◽  
Dispersion Model ◽  
Chemical Transport ◽  
Particle Dispersion ◽  
Global Network ◽  
Lagrangian Model ◽  
Chemical Transport Model

Abstract. We present a method for estimating emissions of long-lived trace gases from a sparse global network of high-frequency observatories, using both a global Eulerian chemical transport model and Lagrangian particle dispersion model. Emissions are derived in a single step after determining sensitivities of the observations to initial conditions, the high-resolution emissions field close to observation points, and larger regions further from the measurements. This method has the several advantages over inversions using one type of model alone, in that: high-resolution simulations can be carried out in limited domains close to the measurement sites, with lower resolution being used further from them; the influence of errors due to aggregation of emissions close to the measurement sites can be minimized; assumptions about boundary conditions to the Lagrangian model do not need to be made, since the entire emissions field is estimated; any combination of appropriate models can be used, with no code modification. Because the sensitivity to the entire emissions field is derived, the estimation can be carried out using traditional statistical methods without the need for multiple steps in the inversion. We demonstrate the utility of this approach by determining global SF6 emissions using measurements from the Advanced Global Atmospheric Gases Experiment (AGAGE) between 2007 and 2009. The global total and large-scale patterns of the derived emissions agree well with previous studies, whilst allowing emissions to be determined at higher resolution than has previously been possible, and improving the agreement between the modeled and observed mole fractions at some sites.

scholarly journals Greenhouse gas network design using backward Lagrangian particle dispersion modelling − Part 1: Methodology and Australian test case

2014 ◽  
Vol 14 (17) ◽  
pp. 9363-9378 ◽  
Author(s):  
T. Ziehn ◽  
A. Nickless ◽  
P. J. Rayner ◽  
R. M. Law ◽  
G. Roff ◽  
...  
Keyword(s):  
Network Design ◽  
Fossil Fuel ◽  
Dispersion Model ◽  
Surface Flux ◽  
Particle Dispersion ◽  
Surface Fluxes ◽  
Test Case ◽  
Lagrangian Particle ◽  
Gas Network ◽  
Australian Continent

Abstract. This paper describes the generation of optimal atmospheric measurement networks for determining carbon dioxide fluxes over Australia using inverse methods. A Lagrangian particle dispersion model is used in reverse mode together with a Bayesian inverse modelling framework to calculate the relationship between weekly surface fluxes, comprising contributions from the biosphere and fossil fuel combustion, and hourly concentration observations for the Australian continent. Meteorological driving fields are provided by the regional version of the Australian Community Climate and Earth System Simulator (ACCESS) at 12 km resolution at an hourly timescale. Prior uncertainties are derived on a weekly timescale for biosphere fluxes and fossil fuel emissions from high-resolution model runs using the Community Atmosphere Biosphere Land Exchange (CABLE) model and the Fossil Fuel Data Assimilation System (FFDAS) respectively. The influence from outside the modelled domain is investigated, but proves to be negligible for the network design. Existing ground-based measurement stations in Australia are assessed in terms of their ability to constrain local flux estimates from the land. We find that the six stations that are currently operational are already able to reduce the uncertainties on surface flux estimates by about 30%. A candidate list of 59 stations is generated based on logistic constraints and an incremental optimisation scheme is used to extend the network of existing stations. In order to achieve an uncertainty reduction of about 50%, we need to double the number of measurement stations in Australia. Assuming equal data uncertainties for all sites, new stations would be mainly located in the northern and eastern part of the continent.

Sign in / Sign up

Export Citation Format

Share Document