Biogeochemistry of the Continental Carbon Cycle

Workshop hosted by Project B2
Cooperative Research Centre for Greenhouse Accounting

Queensland Department of Natural Resources and Mines
Brisbane Laboratories
Indooroopilly, Queensland.

September 12 – 13, 2002

Dr Damian Barrett, Project Leader B2

Uncertainty and variability in the national carbon cycle

Large uncertainties exist in our understanding of the carbon cycle at regional scales. These uncertainties, which arise from the lack of understanding of fundamental processes, the existence of few data and observational difficulties, lead to an inability to accurately predict land-air fluxes of C.

The aim of research conducted in Project B2 is to identify and reduce uncertainties in the carbon cycle at the continental scale. The focus of this research is to generate a comprehensive understanding of the carbon cycle and the processes which link climate and the terrestrial biosphere on scales from seasons to centuries and catchments to continents. In this way, we may better anticipate changes in the continental carbon budget as it responds to long term variation in climate, disturbance and land use change.

This work compliments the National Carbon Accounting System of the Australian Greenhouse Office by achieving a better understanding of the effectiveness of C-sinks for Greenhouse Gas mitigation through a complete (“wall-to-wall”) analysis of pools and fluxes. Specific outputs of this research include:

1. examination of longevity of C sinks (the turnover time of C through the terrestrial biosphere),
2. identification of potential ‘leakage’ of carbon from pools not accounted for in national greenhouse gas inventories,
3. the role of climate variability in determining the dynamics of C sinks, and,
4. examination of the effects of disturbance, land management and land use change on C-emissions.

The outcomes of this research are an improved ability to identify potential threats to C-sinks arising from

• the sensitivity of soil organic matter decomposition to decadal to century variation in temperature and the implications of this for large scale carbon budgets and atmospheric CO2 concentration,
• the effect of afforestation on soil organic matter content at large scales,
• the quantification of emissions due to land use change in Australia and globally, and,
• the impact of climate variation on the management of C-sources and sinks.

The ‘Biogeochemistry of the Continental Carbon Cycle’ workshop held at the Queensland Department of Natural Resources and Mines in September 2002, aimed to better understand the current status of carbon cycle science, to identify where gaps remain and where new efforts are required and to determine the significance of these gaps for the C budget of the continent.

The specific aims of the workshop were to

• Facilitate developing an improved theoretical understanding of the interaction between carbon, climate, nutrients, disturbance and land cover change
• improve parameterization methods of carbon cycle models which incorporate multiple and disparate data sets of observations
• assist the development of new methods for analyzing the effects of multi-year variability in climate, natural disturbance and land use change on the net exchange of C between the atmosphere and the land surface
• continue development of a continental scale stratified sampling method for generating soil C and C-isotope data to constrain continental C-cycle models, and,
• improve the link between new remote sensing technology and large scale carbon cycle models using MODIS and Hyperspectral data

This workshop had its genesis at the 2002 CRC GA Annual Science Meeting, where participants identified the need for a workshop to focus on the science of the carbon cycle at the large scale. It was identified that there were a number of processes currently inadequately represented in terrestrial carbon cycle models which may have important impact on the terrestrial carbon budget. Furthermore, the role of climate variability and its interaction with the carbon cycle in determining uncertainty in the projections of C-sinks for Australia was highlighted. Finally, the need for a synthesis of the science of the carbon cycle was recognized with particular emphasis on uncertainty and variability at the large scale. The main uncertainty issues were

1. uncertainty in the theoretical understanding of processes governing the carbon cycle
2. development of new methods to improve parameter estimation and error propagation in models
3. sampling issues, and,
4. the role of remote sensing in providing time and space intensive observational constraints on models of the terrestrial C cycle
   

Reducing uncertainty in the carbon cycle: Process representation in models

During the workshop it was recognized that a number of important recent advances have occurred in our understanding of the processes governing the net exchange of carbon between the land surface and the atmosphere. These included

• an improved mathematical framework for modeling forest growth which incorporates self-thinning and disturbance, scaling the C-dynamics of forests in time and space between plot and region and an ability to link observed data with regional forest dynamics at any scale,
• establishing the functional relationship between irradiance, humidity and gross primary production of vegetation and its role in constraining NPP of large regions (eg. tropical savannas),
• the mutual constraints on carbon, water and nutrient fluxes imposed by energy balance and gas exchange processes and the stoichiometry of chemical transformations in terrestrial pools,
• identifying and understanding the processes controlling adsorption of C in soils and their importance in determining C-turnover,
• the dynamics of land degradation events and the interaction between climate, social and economic forces on the carbon budget of arid and semi-arid regions,
• improved methods of model parameterization using multiple sources of data (model-data fusion),
• relationships between future pasture production in tropical savannas and predicted climate change and CO2 fertilization, and,
• the improved availability of new remote sensing data products as constraints on large scale models.

Workshop outcomes

The workshop presentations available on the CRC-GA website (www.greenhouse.crc.org.au) expand on the details of the above work. Some important advances presented at the workshop include:

1. The similarity in magnitude of the estimated C-sink due to woody vegetation growth in central Queensland (53 g_C m^-2 y^-1; Burrows et al 2002) and the estimated net flux of C into the biosphere measured at Virginia Park by eddy correlation (82 g_C m^-2 y^-1; Leuning et al 2002).
2. The ability of AussieGrass (a spatial pasture production model) to distinguish variation in soil moisture and available nitrogen effects on plant production from BRDF, solar zenith angle and view angle effects in the NOAA-NASA AVHRR NDVI data (Carter),
3. The improved understanding of the interaction between land degradation drivers (climate, social and economic drivers), the resulting observed episodic nature of degradation events, and their effects on C-stocks of rangelands (McKeon),
4. A first attempt to examine the importance of erosion on the C-budget of arid and semi-arid regions of Australia (McTainsh). This analysis concluded that wind erosion is likely a significant term in the C-budget of surface soils from these regions. For example, a large dust storm event transporting 3 Mt of soil in the atmosphere, having an enrichment ratio (C content of dust to C content of soil) of ~5, and originating from an area of land 60Mha in size would remove 1g_C m^-2 from that area per event. In addition, more frequent smaller events having higher enrichment ratios (~10) may remove on average 33t(Dust) km-2 y^-1 leading to an annual loss of 3.3 g_C m^-2. An annual loss by soil erosion of 4 g_C m^-2 from a surface C-density of 25g_C m^-2 represents an annual loss of around 15% (assuming a bulk density of 800 kg m^-3 and a [C] of 1% in the top 0.5 cm). Alternatively, this loss rate would account for around 10% of annual litter production (assuming a litterfall rate of 40 g_C m^-2 y^-1 for arid regions).
   

Remaining uncertainties in the carbon cycle

While the workshop identified significant recent advances in our understanding of the C cycle, there are still areas of major uncertainty. These include:

• The scarcity of continental scale data sets collated using appropriate stratified sampling methodologies for use in calibration of large scale models
• The physical relationship between soil mineralogy and soil N and C dynamics
• The inability of large scale models to capture fast response events including rainfall-runoff events, N mineralization events and organic matter decomposition events,
• The inability of models to capture variation in plant tissue C:N ratios between validation sites in savannas,
• Elucidating the VPD – diffuse irradiance components of gross primary production and its impact on plant production at large scales
• The processes governing lateral transport of soil, carbon and nutrients (both wind and water) and their implications for C-budgets in semi-arid regions,
• The scaling of vegetation properties from fine to coarse scales taking into account the systematic errors in remote sensing data; particularly relating point data to time and spatial averages over heterogeneous vegetation and soil,
• The poor ability of pasture production models (Century, GRASP, MUSE and SAVANNA) to discriminate growth between poor and good rainfall years.
  

Where to next…?

The immediate next step identified during the workshop is to develop a manuscript for publication which details the analysis of the effect of dust transport on the C budget of Australian arid and semi-arid regions. It was proposed that this work be led by McTainsh, Raupach and McKeon and the ensuing paper will be titled “Assessment of the importance of dust transport to the continental C budget of Australia”. This preliminary analysis will be the precursor to further research on incorporating dust transport into the AussieGrass model through the QDNRM (McKeon). The outcome of this work will improve our understanding of the role aeolian transport in determining the C budget of Australia. Further research is required to incorporate saltation and dust uplift in an atmospheric transport model to examine the redistribution of C and sediment in the landscape. This is necessary to quantify the net flow of C (and nutrients) offshore from the continent. Further advances may be possible by developing partnerships with the Savanna CRC or Remote Sensing CRC with possible extension to C transport in rivers.

In late 2002 - early 2003, the visit to Canberra of Richard Betts (The Hadley Center for Climate Prediction and Research), and Steven Running (The University of Montana) presents an important opportunity to advance capabilities in large scale C-cycle modeling. It is proposed that a second Project B2 workshop be developed for early 2003 in Canberra which will focus on further improvement in process representation of large scale biogeochemical models by:

• development of disturbance routines particularly to cover emissions from fire, and,
• introduction of improved plant production routines in models using MODIS data products.

It was also proposed to link this work more closely with the Global Carbon Project of the International Geosphere Biosphere Program (IGBP), Foci 2 (Processes, Controls and Interactions) to contribute to international efforts in reducing uncertainties in the biogeochemistry of the C cycle at large scales. Finally, further development of large scale models in the CRC will contribute to the RangeASSESS spatial framework for analysis of carbon sequestration in rangelands in order to capitalize both on the large amount of information stored in these datasets from the first RangeASSESS workshop (held in Canberra, September 2000) and to extend and develop the capabilities of RangeASSESS as a large scale decision support system for C-sinks in rangelands.