Project Topics
Get Complete Project Material File(s) Now! »
Role of the greenhouse gases in global warming
Since the industrial revolution, human activities have been injecting into the atmosphere important quantity of carbon dioxide (36183 MtCO₂ in 2016, according to Global Carbon Atlas, www.globalcarbonatlas.org/) and other greenhouse gases such as methane (CH4) and nitrous oxide (N2O). The CO2 emissions are mainly related to fossil fuel combustion for industrial, domestic and transport energy needs. CH4 is mostly linked to agricultural practices (e.g., rice growing and enteric fermentation), waste decomposition, as well as oil and gas production. Whereas, N2O is emitted mostly from agricultural activities, with the use of mineral and animal fertilizers. Other new substances such as the Chlorofluorocarbons (CFC), hydrochlorofluorocarbons (HCFC), whose origin is totally anthropogenic, are characterized by a greenhouse gas effect that may exceed thousands of times the one of CO2 (Flanner et al., 2018). All these gases alter the global energy balance leading to the additional energy trapping near the surface. Other atmospheric components, such as aerosols, have a negative radiative forcing, which may lead to the atmospheric cooling (Figure I.2)
Figure I.2: Radiative forcing of the main anthropogenic (greenhouse gases, aerosol, short-lived gas) and natural (solar radiation) factors impacting the climate in 2011 compared to 1750 (5th Assessment Report of the IPCC 2013)
The contribution of the main greenhouse gases to the radiative forcing is presented in figure I.2. Since the beginning of the pre-industrial era, the additional radiative forcing caused by the anthropogenic activities is estimated to 2.3 W.m-2 (between 1.1 to -3.0 W.m-2, Myhre et al., 2013). According to the 5th Assessment Report of the IPCC (IPCC 2013), the CO2 is responsible for the highest radiative forcing that exceeds 1.5 W.m-2 (Figure I.2). CO2 contributes by more than 50% to the additional radiative forcing produced by the greenhouse gases, whereas the CH4 contribution is about 20%. The impact of the other atmospheric components and short-lived particles is limited (e.g. ozone, carbon monoxide, and the volatile organic compound). Each greenhouse gas has a different impact on the global warming, which depends on its own atmospheric residence time and its radiative forcing. In order to compare the relative contributions of these gases, we use the Global Warming Potential index (GWP). This index has been developed by Houghton and Jenkins (1990) in order to quantify the greenhouse effect of each gas compared to CO2. Greenhouse gas emissions are often calculated based on the amount of CO2 that would be required to produce a similar warming effect over a given time period. This is calculated by multiplying the amount of the emitted gas by its corresponding GWP index. The CO2 represents the reference value with a GWP index equal 1. For CH4, Etminan et al., (2016) estimated a GWP index 32 times higher than CO2 over a 100 year time period. The GWP index of N2O is estimated to 260 for the same period (Etminan et al., 2016).
Carbon budget
Carbon dioxide cycle
CO2 is the subject of many exchanges between land, ocean, and atmosphere. Figure I.3 represents the main carbon fluxes between the different reservoirs constituting the global carbon cycle.
The atmospheric concentrations of CO2, which was 280 ppm (parts per million by volume) during the pre-industrial era, increased to 380 ppm in 2011 (IPCC, 2013), and exceeded an average of 410 ppm across the entire month for the first time at Mauna Loa in last April. The rise of atmospheric CO2 concentrations is mainly related to the emissions by fossil fuel combustion estimated to 9.4 ± 0.5 GtC/yr (Le Quéré et al., 2017). The second largest anthropogenic source results from the land use changes, in particular, the deforestation estimated to 1.3 ± 0.7 GtC/yr (Le Quéré et al., 2017).
As shown in figure I.3, the global carbon cycle links together the atmosphere, the oceans, the land, and the fossil fuel reservoirs. The carbon fluxes are distributed, between these reservoirs, in different proportions. For example, the carbon fluxes between the atmosphere and the land are exchanged in both directions. The carbon is absorbed by the biomass due to the photosynthesis processes (123 PgC/year), and returns back to the atmosphere by surface emissions, fires, as well as plant, animal and microorganism respirations.
Figure I.3: Representation of the carbon cycle. The annual fluxes are estimated in PgC/year and averaged over the period 2000-2009. The amount of CO2 stored in the three reservoirs is expressed in PgC. The figure is taken from the 5th assessment report of the intergovernmental panel for the Climate Change (IPCC 2013)
The fluxes represented by black arrows on the figure I.3 show the net fluxes estimated for the pre-industrial era. The red arrows represent the additional fluxes emitted on average in the decade 2000-2009 by the human activities, including fossil fuel burning, land use changes and cement production. The additional anthropogenic carbon fluxes are estimated to 9 PgC/year. Of these 9 PgC/year, about 5 are absorbed by land and oceans, and 4 remain in the atmosphere, leading to the increase of the CO2 atmospheric concentrations which is precisely monitored at background observatories (Prather et al., 2012). If the magnitude of the oceanic and terrestrial sources and sinks are relatively well known on a global scale, their contributions at the sub-continental scale remain largely uncertain.
Methane cycle
During the pre-industrial era, the methane atmospheric concentration was about 700 ppb (parts per billion by volume), with a total emission of 215 TgCH4/year (Lelieveld et al., 2002). Since 1750, CH4 atmospheric concentrations increased by 150% (from 700 ppb) to a global mean value of 1853±2 ppb in 2016 (WMO Greenhouse Gas Bulletin N.13). The methane emissions can be separated into three types: natural, pyrogenic, leakages:
• Natural emissions are the result of fermentation reactions and methanogenesis processes of some microbes, produced from organic matter under low oxygen conditions. This category includes emissions from wetlands (e.g., peatlands, swamps, rice fields), termites, animals, landfill sites, wastewater, ruminants.
Pyrogenic sources result from incomplete combustions, from biomass fires or fossil fuels such as domestic biofuels.
Leakage emissions are caused by fossil fuel extraction and use (e.g., coal, natural gas, and oil industry).
The relative proportions of the different CH4 sources were estimated by Dlugokencky et al (2011) as presented in figure I.4, and recently revised by Saunois et al., (2016). The methane natural emissions were estimated to 218±47 TgCH4/year, whereas the anthropogenic emissions were estimated to 335±68 TgCH4/year. The anthropogenic activities include emissions from agriculture, waste treatment, biomass fires, transportation and fossil fuels combustion. At the global scale, the highest contribution was determined for wetlands with a total ranging between 177 and 284 TgCH4/year. The contributions of the geological sources, termites, hydrates and freshwater emissions are estimated respectively to 54±21, 12±10, 5±3, and 40±23 TgCH4/year. For the anthropogenic emissions, the most important contributions come from fossil fuels, waste management, rice and farming estimated to 95±10, 78±12, 37±3, 90±4 TgCH4/year respectively. The comparison between these sources showed the significant impact at the global scale of the wetlands, followed by the anthropogenic emissions. However, over the regional domain studied in this thesis (metropolitan France) some emissions, like the wetlands, biomass burning, rice cultivation, are much less important and can be neglected (Champeaux et al., 2005).
Figure I.4: Representation of the methane cycle. The annual fluxes are estimated in TgCH 4/year and averaged over the period 2000-2009. The amount of CH4 stored in the three reservoirs is expressed in TgCH4 .
CO2 and CH4 atmospheric measurerments
Direct measurement of the greenhouse gases has begun in 1958 by the sampling of the atmospheric concentrations of CO2 at Mauna Loa (Hawai, USA) with the initiative of C. D. Keeling (Keeling, 1960). Measurement networks have been gradually expanded around the world and extended to other greenhouse gases, including the methane since 1978 (Blake et al., 1982, Dlugokencky et al., 1994), with the objective to follow the evolution of the atmospheric concentrations of the principal GHG, and to monitor their long-term trends. Nowadays, both CO2 and CH4 concentrations are regularly measured in the atmosphere by several networks distributed around the world (Figure I.5). Two types of approaches are commonly adopted for the atmospheric sampling of the greenhouse gas concentrations:
Continuous measurements by in-situ instruments: this approach is based on direct measurements using instruments at surface sites or on mobile platforms (aircraft, ships). The main advantage of the continuous measurements relays on the ability to investigate the atmospheric variations at short term scales. The development of laser-based optical measuring instruments has allowed a strong development of measurement sites over the last 10 years.
Flask sampling coupled with in-lab analysis: such monitoring programs are generally developed at a frequency of few samples per month, which allows the characterization of long term trends and seasonal variations at background sites. The main advantage of the flask sampling is the possibility to perform multi-species measurements with limited infrastructure in the field.
Figure I.5 distinguishes several measurement networks. Panels a) and b), retrieved from the Earth System Research Laboratory website (https://www.esrl.noaa.gov/gmd/ccgg/globalview/), show the collaborative monitoring network led by NOAA/ESRL (USA). AGAGE (Advanced Global Atmosphere Gases Experiment) is also one of the oldest networks, providing measurements of different greenhouse gases (but not CO 2) since the beginning of the 1980s (Prinn et al., 2000; Cunnold et al. 2002). More recently, regional networks have also been developed, such as the RAMCES network in France presented in panel c) (Yver et al., 2011). The European Research Infrastructure network ICOS (Integrated Carbon Observation System, https://www.icos-ri.eu/icos-national-networks) aims to develop a dense and standardized monitoring network in Europe. Several data retrieved from the RAMCES and ICOS networks are used in this thesis.
The increasing number of measurement sites has made possible the development of methods dedicated to the estimation of the CO2 and CH4 surface fluxes at increasingly finer spatial scales (Bergamaschi et al., 2018, Kountouris et al., 2018).
Flux estimation approaches :
Bottom-up approach:
The greenhouse gas emissions from anthropogenic sectors can be estimated at different administrative scales (city, region, country) for policy makers, or on regular gridded scales for scientists, by using geo-referenced fields of socio-economic data and source-specific emission factors. For example, national emission inventories are reported every year by the countries to the United Nations Framework Convention on Climate Change (UNFCCC) and form the official data for international climate policies. In France, this activity is compiled by the Centre Interprofessionnel Technique d’Etudes de la Pollution Atmosphérique (CITEPA). Emissions are classified into different sectors (agriculture, transport, energy industries, residential, manufacturing combustion, industrial processes, waste, etc…), and their estimates must follow the guidelines established by the Intergovernmental Panel on Climate Change (IPCC). Similar activities are performed more and more frequently at regional and city scales. The bottom-up approach has significant uncertainties due to the incomplete accounting of all emitting sectors, and by the large uncertainties in the emission factors and activity statistics for many source sectors.
The greenhouse gases inventories currently reported by UNFCCC do not provide a complete picture of the global emissions since not all countries report their emissions every year. For this reason, comprehensive and consistent inventories are developed in addition to the national inventories reported to UNFCCC. This is for example the case of the Emissions Database for Global Atmospheric Research (EDGAR) (Janssens-Maenhout et al., 2017), which estimate anthropogenic emissions for all world countries (e.g. EDGARv4.3.2 FT2012 inventory available at http://edgar.jrc.ec.europa.eu/), or the European inventory provided by Institute for Energy Economics and the Rational Use of Energy (IER), University Stuttgart. Such inventories also have the advantage of providing information on regular spatial grids, which can be used as an input to the atmospheric simulations.
The inventories, like EDGAR or IER, generally do not cover all natural emission processes, like for example the CO2 exchange with the terrestrial ecosystems due to the plant and soil respiration, or the carbon uptake due to the photosynthesis. For those sectors, we may use biogeochemical models which often used remote sensing observations of the state of the vegetation and the weather. Such models are themselves validated by using direct measurement of atmospheric fluxes, which are very local and representative of an area less than 1 km2 (Schmid et al., 1994). Due to the strong spatio-temporal heterogeneities of the fluxes, the extrapolation of such measurements using biogeochemical models still faces significant uncertainties.
Top-down approach
The top-down approach provides an estimation of the surface fluxes using measured atmospheric concentrations, atmospheric models, and prior information of the surface fluxes. In this thesis, the quantification of the CO2 and CH4 fluxes over France will be performed using the Bayesian top-down approach (Tarantola et al. 2005) called hereafter by the inverse modelling. The robustness of this approach depends on the quality of the transport model that mix and transport the surface fluxes to be comparable with the measured concentration. The atmospheric concentration of a given gas represents the amount of fluxes transported in the atmosphere through different processes (e.g. horizontal and vertical mixing). As shown by Peylin et al (2002), the estimation of the CO2 fluxes at the regional scale can be uncertain in case of important transport errors. Consequently, the first step before the development of any inverse modelling framework should be the evaluation of the quality of the transport model used to build our inverse system.
Table of contents :
Chapter I: Introduction
I.1Global radiative balance:
I.2Role of the greenhouse gases in global warming
I.3Carbon budget
I.3.1Carbon dioxide cycle
I.3.2Methane cycle
I.4CO2 and CH4 atmospheric measurerments
I.5Flux estimation approaches :
I.5.1Bottom-up approach:
I.5.2Top-down approach
I.6Estimation of the regional fluxes
I.6.1Some techniques for flux optimization
I.6.2Estimation of CO2 fluxes
I.6.1Estimation of CH4 fluxes
I.7Objective and structure of this thesis
Chapter II: Identification of spikes associated with local sources in continuous time series of atmospheric CO, CO2 and CH4
II.1Summary
II.1.1Context of the study
II.1.2Material and methods
II.1.3Selection and the optimization of the spike detection methods
II.1.4Principal results
II.1.5Conclusions and implications
II.2Introduction
II.3Methodology
II.3.1Measurement sites and methods
II.3.1.1Measurement sites
II.3.1.2Measurement methods
II.3.2Spike detection algorithms
II.3.2.1Coefficient of variation (COV) method
II.3.2.2Standard deviation of the background (SD)
II.3.2.3Robust extraction of baseline signal (REBS)
II.4Results
II.4.1Optimization of the SD and REBS methods
II.4.1.1Sensitivity to the parameters of the SD method
II.4.1.2Sensitivity to the parameters of the REBS method
II.4.2Statistics of the three spike detection methods
II.4.1Comparison of SD and REBS methods to detect CH4 spikes at the PDM clean-air mountain station
II.4.2Comparison between automatic and manual spike detection
II.4.3Influence of the spike detection on hourly averages:
II.5Conclusion
Chapter III: Evaluation of the sensitivity of the transport model CHIMERE using different meteorological fields and surface fluxes for simulating the CO2 and the CH4 concentrations
III.1Introduction
III.2Methods
III.2.1CHIMERE atmospheric transport model
III.2.2Meteorological fields
III.2.2.1AROME
III.2.2.2ECMWF
III.2.1CO2 and CH4 surface fluxes
III.2.1.1Anthropogenic emissions
III.2.1.2Vegetation – atmosphere CO2 fluxes
III.2.2Atmospheric concentration measurements
III.2.3Ecosystem measurements
III.3results
III.3.1Comparison of the national totals and temporal distribution of IER and EDGAR anthropogenic fluxes
III.3.2Spatial differences between IER and EDGAR totals
III.3.3Temporal differences between IER and EDGAR
III.3.4Comparison of the biogenic CO2 fluxes between CTESSEL and VPRM
III.3.4.1Spatial distribution of the modeled fluxes for January and July
III.3.4.2Comparison between the modeled and the simulated NEE
III.3.5Sensitivity of the concentrations to the meteorological forcing
III.3.6Spatial distribution of the AROME/ECMWF differences
III.3.7Sensitivity of the concentrations to the surface fluxes
III.3.8Spatial distribution of the surface flux differences
III.4Conclusions
Chapter IV: The potential of a European network for the optimization the CO2 and the CH4 surface fluxes in France
IV.1Introduction
IV.2Methods
IV.2.1Inverse problem formalism
IV.2.1.1Inversion formalism
IV.2.1.2Inverse problem constraints
IV.2.1.3Regularization of the inverse problem
IV.2.2The solution of the inverse problem:
IV.2.3The inversion setup:
IV.2.3.1Estimation of the observations and prior variance-covariance matrices
IV.2.4The definition of the inverse problem
IV.2.4.1Control vector
IV.2.4.2Observation vector
IV.2.4.3Surface fluxes
IV.2.4.4Observation operator
IV.3Results
IV.3.1Inversion of the CH4 fluxes
IV.3.1.1Weight of the CH4 atmospheric observations in the inversion
IV.3.1.2Comparison of observation and prior flux errors with independent empirical estimates
IV.3.1.3Fit of posterior concentrations to observations
IV.3.1.4Emission regions constrained by the inversion
IV.3.1.5Spatial correlation of the flux errors
IV.3.1.6The spatio-temporal scales resolved by the inversion
IV.3.1.7Optimized fluxes
IV.3.2Inversion of the CO2 fluxes
IV.3.2.1Weight of the CO2 atmospheric observations in the inversion
IV.3.2.2Investigation of the observation and the prior flux errors
IV.3.2.3 Fit of posterior concentrations to observations
IV.3.2.4Flux regions constrained by the inversion
IV.3.2.5Spatial correlation of the anthropogenic and biogenic flux errors
IV.3.2.6The spatio-temporal scales resolved by the inversion
IV.3.2.7Optimized fluxes
IV.4Conclusions
Chapter V: Conclusions and perspectives :
V.1Conclusion
V.1.1Spike detection algorithms
V.1.2Evaluation of the simulated CO2 and CH4 concentrations
V.1.3Estimation of the CO2 and CH4 fluxes in France
V.2Perspectives
V.2.1Identification of the local contamination sources
V.2.2Atmospheric modeling
V.2.3Inverse modeling
Chapter VI: References
