EXECUTIVE SUMMARY

TERRESTRIAL CARBON OBSERVATION (TCO) IS A COMPONENT OF A GLOBAL CARBON OBSERVATION THEME BEING DEVELOPED BY THE INTEGRATED GLOBAL OBSERVING STRATEGY PARTNERS (IGOS-P). TCO is led by the Global Terrestrial Observing System (GTOS) and will provide systematic, spatially and temporally detailed information on the global distribution of terrestrial carbon sources and sinks.

In situ data are difficult to acquire and use at the global level compared to satellite measurements. GTOS, the International Geosphere-Biosphere Programme (IGBP) and the Food and Agriculture Organization of the United Nations (FAO) sponsored a workshop in June 2001 (Frascati, Rome) to bring together scientific experts to:

assess the status of in situ components;
identify gaps between requirements and existing data measurements and discuss approaches to fill these gaps;
examine the use of existing data for TCO objectives.

Presentations, debates, and breakout discussions were used to identify the important roles of in situ data in TCO (Chapter 3). These included:

for use in surface carbon models (development, validation, definition of model parameters, and as input data);
as inputs for preparing satellite-derived products;
for use in atmospheric inversion modelling;
for qualitative and quantitative assessment of TCO output products.

Participants reviewed the current networks and mechanisms for collecting point and gridded observations of atmospheric and ecosystem parameters, and the status of existing terrestrial data sets. National forest inventories were identified as prime sources of existing information on carbon distribution. Lists of existing data sets were compiled (Appendix 3).

The workshop reached the following conclusions:

1. TCO requires data on many surface and atmospheric variables, taken at different sampling frequencies and using different methodologies. Both point and gridded data products are needed. Historical as well as ongoing current and future observations are essential. There is a variety of relevant existing in situ data. The quality and comprehensiveness of such data will be dependent on the data type and source; the intended geographic and temporal coverage, and the original purpose for which it was acquired.
2. Atmospheric composition and surface flux network observations are appropriate for estimating carbon source/sink distributions. However, these networks provide incomplete coverage, in terms of geographic/biome representation and of the variables measured. Most networks receive irregular national funding which causes a lack of continuity in research programmes.
3. The tiers concept of the Global Hierarchical Observing System Strategy (GHOST, www.fao.org/gtos/doc/pub13.pdf) provides a suitable framework for further evolution of in situ networks.
4. Data from surface ecological networks needs to be assembled and ecological networks should be encouraged to participate in carbon studies.
5. Large-scale (supra-national) gridded data sets needed for models are available for various parameters. In many cases the limitations of these data sets can be improved. Horizontal carbon transfers by rivers and through trade, higher spatial and temporal resolution on emissions data, improved soil databases and better information on atmospheric chemical inputs could be rapidly improved through collaboration with data holders and the scientific community.
6. Existing national forest inventories are a unique data source. Although the data has limitations for TCO, as it was developed for other needs, they still have a potential which has not been fully exploited. The suitability of inventory data for national and global TCO applications can be achieved through improved data product preparation and inventory procedures. Further collaboration with national and international agencies is required which should be initiated through formal contacts at the national level.
7. In situ data is essential for the generation of higher-level products from satellite measurements. This is mainly through the validation process (i.e. the calibration of measured radiances). Close collaboration between in situ networks and satellite programmes is therefore required.
8. Over time a variety of data will be required by TCO. A combination of a central data management system (TCO DISS) and a distributed system is needed to facilitate effective participation by both the producers and the users of the data sets.

The main overarching recommendations which emerged from the workshop are:

1. TCO should establish national acceptance and support to develop the in situ component. Important contacts include ministries responsible for monitoring networks, natural resource inventories, and environmental research programmes. Initial contacts could be national representatives to the United Nations Framework Convention on Climate Change (UNFCCC).
2. TCO should work closely with interested international organizations, regional and global research programmes, and with related research or pilot projects to develop observing capabilities, improved data sets and ensuring that the observing components (in situ and satellite) and models using the data evolve in a mutually supportive way.
3. TCO should vigorously pursue consolidation and improvements of in situ networks by supporting the programmes and initiatives of these networks at national and international levels. The initial emphasis should be on surface fluxes and atmospheric composition measurements, as well as on closer collaboration with key ecological networks and data providers.
4. TCO should actively promote initiatives aimed at generating improved regional or global gridded data sets. Lateral carbon transfers by rivers and through trade; higher resolution emissions information in space and time, and improved soil databases have been identified as the initial priorities. To this end, TCO should work closely with the scientific community, as well as international and national agencies that have access to the data.
5. A central data management office for TCO should be established to provide a data and information system and service (DISS).

Numerous specific recommendations have been made throughout the report on point data (page 22 and 26); gridded data sets (page 34); data product assembly issues (page 39), and data and information systems and services (section 40).