Introduction

This report is a tale of two meetings about the terrestrial part of the global carbon cycle. My attendance at the first (Ottowa, February 2000) was funded by the EOC, while the IGBP funded attendance at the second (Lisbon, May 2000). The report fulfils a funding requirement for the EOC but I hope it will be useful not only for the EOC but also for the BWG and for Divisional colleagues.

Both meetings were the result of greatly increased international interest in the global carbon cycle, and in particular its terrestrial components. The two meetings were tightly linked in both content and organisation, the Lisbon meeting being in essence a direct successor to the Ottowa meeting. It has been very helpful to delay this report until after the Lisbon meeting (just completed) as a much better perspective on the international effort is now available.

Ottowa Meeting (February 2000)

Purpose
The meeting was called by GTOS (the Global Terrestrial Observation System) to develop a global plan for terrestrial carbon observation and long-term monitoring. The chair and main organiser was Josef Cihlar (Canadian National Center for Remote Sensing).

Attendance
This was an invitation-only workshop of about 30 people, including earth-observation and carbon-cycle scientists and a good representation by agency personnel. The attendance was heavily skewed towards North America: there were no European scientific representatives because of an impending EU proposal deadline, and only three Asian or Australasian representatives (two from Japan, plus myself). The resulting imbalance was an impediment, but the intention was to redress this at the Lisbon meeting reported below.

Presentations
The meeting was not oriented around formal presentations but several speakers were designated to introduce aspects of the subject and outline the programs of individual regions. Here is an incomplete list:

• Scott Denning's talk on atmospheric inverse approaches highlighted gaps in the current CO₂ (and other gas) concentration observing network (not enough continental and tropical observations) and possible solutions (tall towers and "virtual tall towers"). It was characteristically well structured and well received.
• Several speakers addressed policy issues. In particular Rene Gommes (FAO) emphasised that a terrestrial carbon observation is relevant to several international agendas in addition to greenhouse gases, the UNFCCC and the Kyoto Protocol: he outlined the Convention to Combat Desertification (CCD) and the Convention on Biodiversity (CBD).
• The US, European, Canadian, Japanese and Australian carbon cycle programs were outlined. I spoke about the Australian (mainly Biosphere Working Group) program in this context, generally receiving good feedback. The Australian "multiple-constraint" approach (envisaged by the CSIRO Biosphere Working Group) was seen as a pathfinder.

Plenary Discussions and Small Groups:
The meeting decided (partly at my urging) to amend the program and spend some time in plenary identifying the reasons why the world needs a Terrestrial Carbon Observing System (TCOS) and who the clients for such a system would be. The organisers had assumed before the meeting that this was obvious, but it turned out not to be so! The resulting "compelling reasons" included (1) scientific understanding of the C cycle; (2) support for the UNFCCC and its potential successors; (3) monitoring for surprise behaviour in the earth system; (4) inventory compliance (though a comprehensive TCOS would be only marginally relevant to the Kyoto Protocol because of the highly selective carbon accounting mandated by the KP); and (5) providing information to constrain and give context to land use and land management policy at regional levels.

The next plenary was about the synthesis framework (I chaired it). Francis Bretherton was vocal in urging attention to a "synthesis hierarchy" involving systematic gathering, processing, analysis, archiving, production of high-level information, and delivery to clients; the analogy with the weather network was examined. This was an excellent session for keeping dreams and reality in some connection.

Three breakout groups then considered the details of the observations, respectively the atmospheric observations, surface flux and stock observations, and remote sensing. The result was a series of lists of variables, requirements for precision and potential gaps, reflected in detail in the meeting report. I am not certain with hindsight that these lists represented much more than the sum of the personal wishes and biases of the participants, and I suspect that a different group would come up with a rather different list.

Outcomes
Most attendees judged the meeting to be successful. In my view, the main progress was the three-part recognition (widely endorsed) that:

• Carbon cycle observation requires multiple measurement approaches, including (1) atmospheric concentrations, (2) eddy fluxes, (3) in-situ vegetation stocks and growth rates, and (4) a variety of remotely sensed observations by a variety of together with measurements.
• These measurements need to be combined within a carefully designed integrative or synthesis framework, rather than merely being used as a series of different views or measurements of the same thing (the net land-air exchange of carbon, being the net contribution of the terrestrial carbon cycle to the atmosphere).
• A TCOS ultimately requires an operational rather than a research base, geared to gathering, processing, analysing and archiving data, producing high-level information, and delivering this to clients.

It is not a coincidence that the first two concepts are at the heart of the Science Plan of the CSIRO Biosphere Working Group.

The output of the Ottowa meeting was a comprehensive (over 100 pages) draft report on observation requirements for a "Terrestrial Carbon Observing System" (TCOS). This report is ultimately destined for the Steering Committees of GTOS and then the IGOS-P (Integrated Global Observing System Partnership). The report has subsequently been criticised as too long(its precis runs to thirty pages). The executive summary (two pages) is reproduced in an Appendix to this Report.

Lisbon Meeting (May 2000)

Purpose
This meeting was called by IGBP to formulate a global science research plan for the (terrestrial) carbon cycle. Capable leadership was provided by Berrian Moore (Chair of the IGBP Scientific Committee) and Will Steffen (Executive Director, IGBP). The meeting had a broader focus than the Ottowa meeting, being concerned with the development of a science plan to study the terrestrial carbon cycle in the context of the entire carbon cycle (embracing atmospheric, oceanic and terrestrial processes, and contemporary and historic time frames). The emphasis was on research requiring an internationally coordinated research effort, over and above questions that can be tackled successfully by individuals or small teams.

Attendance
Like the Ottowa meeting, attendance was by invitation. About 60 scientists and a smattering of agency personnel attended, with a good balance between North America and Europe, and some representation outside these blocs (Roger Francey, Ian Enting, Peter Rayner and myself from Australia, three scientists from Japan and one from China).

Presentations
Like the Ottowa meeting the emphasis was not on set-piece presentations but there were a number of introductory talks. Among these (again not a full list) were:

• A talk by Colin Prentice on carbon cycle dynamics, starting from the global C budgets in the IPCC Third Assessment. All atmospheric budget methods (CO₂ / O₂, CO₂ / ¹³CO₂, and others) now agree on a global balance of 6.3±0.6 PgC emissions, -2.3±0.8 PgC ocean uptake, an annual mean of 3.3±0.2 PgC accumulation in the atmosphere and a net terrestrial biospheric sink of -0.7±1.0 PgC, made up of +1.6±0.8 PgC emissions from land use change and a terrestrial uptake (sequestration sink) of -2.3±1.3 PgC (1989-98 IPCC figures). However the interannual variability is large and is mainly in the terrestrial biospheric term, causing a nearly equal and opposite change in the atmospheric accumulation term (to which the other three terms must sum). The terrestrial sink is now agreed to be in the northern temperate latitudes. However there is intense terrestrial biosphere activity in the tropics, with a large likely sink due to C sequestration being counterbalanced by a nearly equal source due to biomass burning and land clearing.
• Later in the meeting, Roger Francey presented brand new isotopic evidence that the interannual variability in the terrestrial sink (and hence the rise in atmospheric CO₂) is 60% caused by variability in biomass burning, which is highly climate-dependent. This was perhaps the show-stopper result of the meeting.
• An excellent talk on the role of institutions in global change was given by Virginia Walsh, Program officer of the Institutional Dimensions Project in IHDP (International Human Dimensions Program on Global Environmental Change; this is a sister program to IGBP and WCRP, these three being the global environmental change programs of ICSU, the International Council of Scientific Unions; but that's enough of the acronym maze).
• Francis Bretherton emphasised again (as at Ottowa) the importance of distinguishing between campaign-style research and long-term operational monitoring, and the special needs of the latter which are usually not appreciated by the research community.
• Chris Field gave the clearest talk of the meeting on "Integrating Observations, Models and Process Studies". He began by identifying the fundamental objectives of global C cycle science, coming up with a list comparable with (but not identical to) the equivalent list from the Ottowa meeting (see above). He then looked at possible causes of the current C sequestration in the terrestrial biosphere, identifying C fertilisation, N deposition, climate change, forest regrowth post-disturbance, fire suppression, land use change to agriculture, improved agriculture, decreased deforestation, sediment burial in rivers. Attributing these processes in magnitude, space and time is a large part of the problem. He called for a strategy based on "multiple levels of consistency and constraint" (echoes of the BWG plan again!). He saw the main tasks of international efforts as being to promote (1) activities needing international coordination, and (2) activities where progress anywhere depends on progress everywhere (such as atmospheric concentration monitoring). He identified five modest goals for the international effort in terrestrial carbon which would collectively make a big difference: rationalise flux networks, harmonise forest inventories, expedite atmospheric concentration measurements, complete land use histories, and develop tools for allowing interaction between different approaches. As I said to him afterwards, it was a talk I wish I'd given.
• Bob Scholes gave a delightful talk on soil C, in the manner of "all you need to know in 10 minutes". He concluded that the cost of a rational soil survey rises logarithmically with the area surveyed, because of a combination of sampling theory and travel distance constraints. Hence it is not much more expensive to do a continent than a catchment, or the world than a continent. By contrast, most soil surveys are small-region efforts which are hard to patch together because of methodological incompatibilities (Elisabeth Bui and the ASRIS group know this very well!).
• At the end of the meeting there was a key talk by Minister Ferriara, Portuguese Minister for Agriculture and the chair of the EU expert group developing a position on terrestrial sinks under the LUCF components of the Kyoto Protocol, for COP 6 in November. She gave a very well-informed and lucid account of the likely EU position, which in brief is that there are major problems in definition and measurement of LUCF sinks, including the definition of "direct human induced"; that measurement burdens on poorer countries are likely to be severe; and therefore that a very limited definition of sinks will be favoured by the EU. They will not support the inclusion of sinks in the CDM (Clean Development Mechanism). This position is in sharp contrast with that of the Australian Government and the AGO; see the AGO Issues Paper "Terrestrial Sinks and the Kyoto Protocol" and the BWG response to it.

Small Groups
The small groups were organised around four themes: (1) spatial variation in the terrestrial C cycle, (2) temporal variation in the terrestrial C cycle, (3) human perturbations especially land use change, and (4) dynamics of the C cycle into the future. I chaired the first group. Each group spent about a day and a half producing quite a detailed report on the main research issues in its area and on the needs for international coordination to answer these questions (bearing Chris Field's criteria in mind). The reports contained a lot of overlap, especially (and not surprisingly) between Groups 1 and 2. First impressions were that the written reports were of excellent quality given the process that por5oduced them, and that they will form a solid foundation for a coordinated international effort.

Outcomes
The main output is a relatively brief (< 50 page) report, to be synthesised from the small group reports by Kathy Hibbard who will work in Canberra at the GCTE Office to do this during July. The report is ultimately destined for ICSU (International Council of Scientific Unions), after it is merged with the outcomes of a similar meeting on the ocean carbon cycle.

Highlights as I see them at this time (immediately after the meeting) were:

• Overall (though informal) endorsement of the major outcomes of the Ottowa meeting as described above;
• Emphasis on the need for maintenance and improvement of existing observational systems, especially of atmospheric concentrations;
• New results on interannual variability, its terrestrial biospheric attribution and the role of biomass burning;
• An emerging tension, both political and scientific, between terrestrial sinks and reduced emissions as strategies for lowering the rate of increase of atmospheric CO₂. The argument that "sinks buy time" is under challenge, especially from the EU community who see sinks as a displacement activity. Our main task here is to improve the science, including measurement technologies, realistic and verifiable sequestration technologies, and forecasts of genuine sink potential and (importantly) longevity. Will the carbon stay put without intensive subsequent management?
• The meeting was conscious of the need to consider the terrestrial biosphere, the oceans and the atmosphere together (there is only one carbon cycle). Though there were few oceanographers present, the next meeting in this series (New Hampshire, October) will bring the terrestrial and ocean communities together.

Executive Summary

In response to an increasing interest in terrestrial carbon and following a proposal led by the Global Terrestrial Observing System (GTOS), the Integrated Global Observing Strategy Partnership (IGOS-P) approved terrestrial carbon cycle as its second major theme in November, 1999. This report presents results of a follow-up Terrestrial Carbon Observation (TCO) Synthesis Workshop, organised by GTOS in collaboration with IGBP and other IGOS-P members for 8-11 February, 2000 in Ottawa, Canada. The workshop was designed to summarise existing information and observation requirements regarding terrestrial carbon, conduct initial evaluation of existing data or observations in relation to the requirements, identify major gaps or deficiencies, and propose solutions.

Existing stated requirements for terrestrial carbon information were reviewed in several areas, including international conventions, scientific understanding of the global carbon cycle and assessments of its evolution (current and into the future), and land management. Based on these, the needed observations were analysed with a view to satisfy a ‘dual constraint’ methodology for estimating terrestrial carbon fluxes, based respectively on a) local ecosystem models scaled up with satellite data, and b) atmospheric model inversions using concentration measurements of atmospheric CO₂ and other tracer gases. Existing observations, gaps, and needed improvements were also discussed.

To meet TCO needs, the concept an observing system was considered. Such a system will contribute to the integrated understanding and human management of the global carbon cycle- through systematic, long-term monitoring of the terrestrial exchanges of greenhouse gases, especially CO₂, and the associated changes in carbon stocks. The goal is to obtain estimates through the use of models that synthesise information from several types of measurements: atmospheric CO₂ and other gases, surface fluxes, ecological, and remote sensing. These estimates will be provided with known and decreasing uncertainty, by systematic cross-checking of independent approaches and by designed expansions of current measurement networks. The information products will be of value not only at the global and regional levels, but also for land management and assessment in support of sustainable development at the national level. Ultimately, an integrated global observing strategy should provide near-real-time diagnosis of carbon sources and sinks at high resolution in both space and time that simultaneously satisfies all the data constraints (in situ, remotely sensed, and atmospheric) at multiple scales. Such a system will be more than a set of observations; rather, it will constitute a carbon cycle data assimilation system analogous to the observing systems currently used for temperature, precipitation etc. in operational weather prediction.

Based on the presentations and discussions, the following conclusions were reached:

1. Information on the global distribution of terrestrial carbon sinks and sources is essential for policy and scientific purposes in four areas: reporting for multilateral environmental agreements; understanding of the carbon cycle; assessment of global change trends and impacts; and the management of ecosystem resources at local to regional levels.
2. A dual observation and modeling approach, based respectively on the inversion of atmospheric observations and on the use of satellite data and ecosystem models, is capable of achieving accurate information on the distribution of carbon sources and sinks at all scales from landscape to global.
3. Many components needed for terrestrial carbon observations are well understood. Some are in place, others need to be augmented, and all need to be placed in a consistent, functioning framework.
4. To be effective, such a framework must incorporate both international co-ordination and national implementation as essential components.

The following recommendations are made to IGOS-P by workshop participants:

1. Seek endorsement for the TCO system concept.
2. If adopted, modify the proposed evolution strategy as appropriate and take steps to its implementation. These should include an integrated approach to data distribution, quality control, and archiving; arrangements for the generation of core products; and clarifications regarding the responsibilities of agencies in the planning, development, and performance assessment of these activities.
3. Ensure continuation of existing satellite observations important to TCO into the foreseeable future. Accelerate the development and deployment of new satellite observation technology, including lidars for vegetation biomass, canopy structure, and atmospheric CO2 concentration.
4. Expand the system of flux networks and ensure adequate geographic coverage, continuity of observations, and co-ordination.
5. Improve the access and use of existing (non-flux) sites and national data sets for TCO purposes.
6. Review and further refine the strategy for the development of the dual constraint concept, and ensure active participation of the hydrological community in this process.
7. Give high funding priority to research and development of instruments, observation methods, and models related to carbon cycle observations.
8. In the evolution of terrestrial carbon observations, maintain close linkages with the ocean carbon cycle observation community.
9. Issues relating to scaling, gridded data sets, emissions, and others identified at this workshop should be examined by a broader scientific community in order to understand the implications for global terrestrial carbon observations.