Project Topics
Get Complete Project Material File(s) Now! »
A brief introduction to atmospheric aerosols
In general, an aerosol refers to a suspension of solid or liquid particles in air. In climate sciences, cloud particles are often differentiated from other particles, hence “aerosols” is often used to represent the suspensions of solid or liquid particles except the hydrometeors (Boucher, 2015). This definition is hereafter used in this thesis.
Atmospheric aerosols can be classified into primary aerosols and secondary aerosols according to whether the aerosols are emitted as particles directly (primary aerosols) or come from condensation of atmospheric gas-phase species during chemical processes (secondary aerosols). These gas-phase species are called aerosol precursors.
According to their source, atmospheric aerosols can be divided into natural aerosols and anthropogenic aerosols. Natural aerosols mainly include biogenic (organic) aerosols derived from volatile organic compounds emitted by terrestrial plants and marine algae, dust lifted by winds, sea spray aerosols, black carbon from wildfires. Apart from the above ones, volcanic eruptions can also emit a large quantity of sulphate aerosol precursors into the atmosphere including in the stratospheres into higher atmosphere. Anthropogenic aerosols are emitted from anthropogenic activities such as fossil fuel burning, mining, agriculture and land use change. These activities can result in emissions of precursors of sulphate, nitrate and organic aerosols, but also primary black carbon and organic aerosols. Fly ash and industrial dust are other types of anthropogenic aerosols. In natural conditions, atmospheric aerosols are often a mixture of different chemicals. These mixtures can be of different particles with pure chemical composition each (external mixture), or of particles composed by well mixed chemicals (internal mixture). The mixing can strongly affect the particle physics and lead to different aerosol properties (Boucher, 2015).
The various source and types, as well as complex mixture of aerosols result in a heterogeneous spatiotemporal distribution of aerosols (Figure 1.2). Industrialized regions such as East Asia, Europe and North America generally have higher aerosol concentration than other regions, as the intense human activities strongly increase the emission of anthropogenic aerosols and their precursors. Besides these regions, deserts such as Sahal also have high aerosol concentration. This is because that the non-vegetated ground in deserts can easily become a source of dust aerosols when wind is strong. Tropical forest regions may also have high aerosol concentrations because of the large production of organic aerosol precursors in these forests. In addition, wild fires can occasionally cause high concentration of black carbon aerosols near the burned regions.
Aerosol-radiation interactions and aerosol-cloud interactions
The most known and studied mechanisms how aerosols affect climate system are aerosol-radiation interactions and aerosol-cloud interactions (Heywood & Boucher 2000). Aerosol-radiation interactions happen through the scattering and absorption of (essentially solar) radiation by aerosols. Because part of the irradiance is absorbed or scattered back to outer space during this process, aerosol-radiation interaction often causes a cooler surface and lead to a cooler climate. An exception occurs if the aerosol is very absorbing (e.g. black carbon) and/or the land surface is covered by snow or ice, resulting in a high albedo. In such case, the absorption of this aerosol may warm the climate system although locally it may still cool the surface because levels of solar radiation at the surface decreases. Globally, aerosol-radiation interactions are suggested to cause a radiative forcing of –0.45 (–0.95 to +0.05) W m−2 (Boucher et al., 2013).
Aerosol-cloud interactions occur because the formation of clouds can be affected by aerosols (Figure 1.3). Aerosols as a suspension of particles often serve as cloud condensation nuclei (CCN), so the presence of anthropogenic aerosols result in an increase in cloud condensation nuclei concentration. This increase in CCN further changes the size distribution of cloud droplets, and in turn alters the distribution, reflectance and life time of clouds. For a given cloud water content, more and smaller droplets leads to higher cloud reflectance. Aerosol-cloud interactions are estimated to cause a radiatve forcing of –0.45 (–1.2 to 0.0) W m−2 globally (Boucher et al., 2013).
The aerosol radiative forcing is highly heterogeneous (Figure 1.4). Generally, the regions with strong negative aerosol radiative forcing (<−1 W m−2) are mainly distributed in East and South Asia, Europe, East US, where intense anthropogenic activities emitted large amount of sulphate and nitrate aerosols and aerosol precursors.
Impacts of aerosols on terrestrial C cycle
Aerosols can affect the terrestrial C cycle in multiple ways (Figure 1.5). The first way is through their contribution to changing climate. Generally, the negative radiative forcing caused by aerosols results in lower temperature and a weaker solar irradiance at land surface. The precipitation changes in response to aerosol emissions remain poorly understood. These changes in climate in turn alter the C fluxes. This impact is implicitly included in previous fully coupled simulations using earth system models, however, rarely quantified.
Jones et al. (2003) investigated the impacts of sulphate aerosols on the C cycle by comparing coupled simulations with and without sulphate aerosol forcing on HadCM3L. Their study considered the aerosol-radiation interaction and the aerosol impacts on cloud albedo, but omitted the impacts of aerosols on cloud lifetime. The results suggested that sulphate aerosols can increase global C sink through their cooling effect, which suppressed the increasing soil respiration under a warming climate. A similar study was performed by Mahowald et al. (2011) using the Community Climate System Model (CCSM3.1) model. However, their study showed a relatively small impact of aerosols on global C cycle and attributed it to the weaker climate-carbon feedback in their model. There is no consensus on the impacts of aerosol-induced climate change on the land C cycle.
Canopy light transmission and photosynthesis in the ORCHIDEE trunk
The ORCHIDEE_DF model is based on ORCHIDEE trunk revision 5453 (updated in September 2018). A general description of the physical processes related to energy and water balance, vegetation dynamics and biogeochemical processes in ORCHIDEE can be found in Krinner et al. (2005). The ORCHIDEE trunk version 5453 (hereafter referred to as trunk for simplicity) brings a number of improvements and photosynthesis parameters were recently re-calibrated against FLUXNET data (Baldocchi et al., 2001) and atmospheric CO2 observations for the IPSL Earth System Model (IPSL-CM6) and the CMIP6 simulations.
The leaf-scale photosynthesis calculation in the ORCHIDEE trunk version is based on the scheme of Yin and Struik (2009). This scheme is an adaptation of the biophysical model of Farquhar et al. (1980) with a specific parameterization of stomatal conductance. The Farquhar et al. model calculates assimilation (A) as the minimum of the Rubisco-limited rate of CO2 assimilation (Ac) and the electron transport-limited rate of CO2 assimilation (Aj): 𝐴=min{𝐴𝑐,𝐴𝑗} Eq. 3.1.
Here Ac is mainly affected by the maximum carboxylation capacity of Rubisco (Vcmax), which is temperature dependent (Yin and Struik., 2009), and the CO2 concentration at the carboxylation site (Cc): 𝐴𝑐= (𝐶𝑐−𝛤∗)𝑉𝑐𝑚𝑎𝑥𝐶𝑐+𝐾𝑚𝐶(1 + 𝑂/𝐾𝑚𝑂)−𝑅𝑑.
Light partitioning in ORCHIDEE_DF
The lack of light quality (diffuse light fraction) information in most forcing datasets is one of the main difficulties when simulating the diffuse light effect. Here we partition the half-hourly downward PAR, which can be derived from the shortwave radiation, into diffuse and direct components following the Weiss and Norman (1985) empirical equation. Compared with another empirical method (Spitters et al., 1986), we found that this method reproduces better the observed diffuse light fraction at the flux sites used in this study (see results and Figure S3.1). The diffuse PAR fraction (𝐹𝑑𝑓𝑃𝐴𝑅) above the canopy is estimated as: 𝐹𝑑𝑓𝑃𝐴𝑅=1−𝑃𝐴𝑅𝑝,𝑑𝑟𝑃𝐴𝑅𝑝(1−(𝑎−𝑅𝑏)23).
Canopy light transmission in ORCHIDEE_DF
In ORCHIDEE_DF, we use the same stratification of canopy as in the trunk version (Eq. 3.4). But for the light transmission, we use a two-stream radiative transfer model following Spitters (1986). For convenience, we use radiation and I in this section to refer to the PPFD derived from the light partitioning step.
An assumption of the model is that leaves are bi-Lambertian surfaces for radiation, i.e. the reflection and transmission are isotropic. This reflection and transmission are together referred to as leaf scattering. This assumption implies that once direct radiation encounters a leaf, it gets either absorbed or scattered as diffuse light. While for diffuse radiation, the scattered light remains diffuse. The scattering coefficient, σ, is assumed equal to 0.2 following Spitters (1986), meaning 20% of the light encountering a leaf is scattered (80% is absorbed).
Based on this assumption, the radiation penetrating the canopy can be divided into three components (Figure 3.2): the direct light which has not been intercepted by leaves (𝐼𝑑𝑟,𝑑𝑟), the diffuse light generated by leaf scattering of intercepted direct light (𝐼𝑑𝑟,𝑑𝑓), and the diffuse light in the canopy provided by the TOC diffuse radiation (𝐼𝑑𝑓). It should be noted that the diffuse light generated by multiple times of scattering of the direct light is grouped into 𝐼𝑑𝑟,𝑑𝑓, while those from the scattering of TOC diffuse radiation belong to 𝐼𝑑𝑓 (Figure 3.2). The sum of 𝐼𝑑𝑟,𝑑𝑟 and 𝐼𝑑𝑟,𝑑𝑓 hereafter noted as 𝐼𝑑𝑟 represents the total radiation in each canopy later derived from the TOC direct radiation, hereafter 𝐼𝑑𝑟,0.
Flux data and site level simulations
To evaluate ORCHIDEE_DF, we collected flux site measurements from the La Thuile dataset, which includes 965 site-year observations from in total 252 sites. Because our ORCHIDEE simulations assumes that the ecosystems are in equilibrium and do not experience disturbances (e.g. logging, fire), we selected flux sites without strong disturbances during the last 10 years. For sites that also provided growing season LAI information, we also removed forests site with LAI<2, which may be considered as sparse forests with understory vegetation. In the end, observations of 655 site-years from 159 sites were retained (Table S3.2). The annual climate of the sites spans from −9 to 27 oC in temperature, and from 67 mm yr−1 to over 3000 mm yr−1 in precipitation (Figure S3.2), which is representative to most of the climate conditions over the globe. The dataset provides in situ meteorology, net ecosystem exchange (NEE), gross primary productivity (GPP), and data quality information at 30-min time steps. The GPP provided by this dataset is partitioned from NEE and gap filled using the method of Reichstein et al. (2005). Specifically, 64 of the 159 sites provided measurements of both total and diffuse PPFD, which allows us to evaluate the light partitioning parametrization (Eqs 9-20). The gaps and missing variables in meteorology are filled using the approach from Vuichard and Papale (2015) to meet the model input requirements.
Table of contents :
Table of contents
CHAPTER 1. GENERAL INTRODUCTION
1.1 THE TERRESTRIAL CARBON CYCLE
1.1.1 Main processes in the terrestrial carbon cycle
1.1.2 Factors affecting the terrestrial C cycle
1.1.3 Climate-Carbon feedback
1.2 ATMOSPHERIC AEROSOLS
1.2.1 A brief introduction to atmospheric aerosols
1.2.2 Aerosol-radiation interactions and aerosol-cloud interactions
1.2.3 Impacts of aerosols on terrestrial C cycle
1.3 METHODS INVESTIGATING AEROSOL IMPACTS.
1.4 THE OBJECTIVES OF THE THESIS
CHAPTER 2. IMPACTS OF ANTHROPOGENIC AEROSOLS ON LAND C FLUXES THROUGH CHANGING CLIMATE
SUMMARY
CHAPTER 3. MODELLING THE IMPACTS OF DIFFUSE LIGHT FRACTION ON PHOTOSYNTHESIS
SUMMARY
3.1 INTRODUCTION
3.2 DATA AND METHOD
3.2.1 Model description
3.2.2 Flux data and site level simulations
3.2.3 Analyses
3.3 RESULTS
3.3.1 Diffuse light fraction
3.3.2 General model performance
3.3.3 Effects of diffuse light on GPP and LUE
3.3.4 Interactions between diffuse light and environmental factors
3.4 DISCUSSION
3.4.1 Improvement of ORCHIDEE_DF
3.4.2 Factors affecting diffuse light effect on GPP
3.4.3 Uncertainties and Limitations
3.5 CONCLUSION
SUPPORTING INFORMATION FOR CHAPTER 3
CHAPTER 4. AEROSOL IMPACTS ON THE LAND CARBON CYCLE THROUGH CHANGING DIFFUSE RADIATION
SUMMARY
4.1 INTRODUCTION
4.2 DATA AND METHODS
4.2.1 ORCHIDEE_DF model
4.2.2 Forcing data and Experimental design
4.3 RESULTS
4.3.1 CRUJRA simulations
4.3.2 IPSL simulations
4.4 DISCUSSION
4.4.1 Fdf response to anthropogenic aerosols
4.4.2 Methods to reconstruct NoAA Fdf
4.4.3 Global impacts of diffuse radiation on C fluxes
4.4.4 The impacts of Radiation quality vs Radiation quantity
4.4.5 The main factors causing C flux changes
4.4.6 The impacts of volcanoes
4.4.7 Uncertainties
4.5 CONCLUSION
CHAPTER 5. CONCLUSION AND PERSPECTIVES
REFERENCES
