A major challenge in reaching a better understanding of global change is the integration of global carbon observations at different scales, made in the atmosphere, ocean and terrestrial domains.  This is essential to optimize efforts supporting national, regional and international policy related to the global carbon cycle.  The partners of the Integrated Global Observing Strategy (IGOS-P) representing all players in carbon cycle research and monitoring recognised this and produced, with the help of an international panels of experts, published theme reports on the Carbon Cycle (IGCO) and on Atmospheric Chemistry (IGACO).  These themes contain recommendations on how to more effectively coordinate and fill gaps in global Earth observations. 

A time series transect has been established in subantarctic surface water off the south east coast of New Zealand.  The 60 km long transect extends from the coast (45-46.20^oS 170-43.20^oE) to a station at 45-50.00^oS 171-30.00^oE. and sea surface temperature, salinity and pCO₂ have been measured bi-monthly since 1998 . SST, pCO₂ and pH of the subantarctic surface water show seasonal cycles that can be fitted with simple harmonic curves.  Temperature has a mean value of 10.4^oC, with an amplitude of 2.1^oC, the maximum occurring in late summer.  pCO₂ has a mean value of 360 matm, an amplitude of 10 matm, the maximum occurring in early spring.  The phase of the pCO₂ and temperature curves are offset by 158 days, indicating that change in sea water temperature is not the major factor affecting pCO₂ in this area.  The relative effects of temperature, biological utilization and air-sea gas exchange on the seasonal change in pCO₂ are determined using a simple model.  The model results reproduce the timing of the observed pCO₂, however the amplitude of the changes is not well reproduced.

The oxygen isotopic composition of atmospheric CO₂ can help constrain local- to global-scale biophysical processes and partition measured net ecosystem CO₂ fluxes into gross fluxes. Although current models still lack key features controlling gross ecosystem CO¹⁸O fluxes, considerable improvements have been achieved in the last four years. In this study we examine the influence on atmospheric CO¹⁸O of 1) a delayed seasonal cycle in soil water isotopes (relative to rain water) and 2) a new one-way flux model of night-time leaf respiration [Cernusak et al., 2004]. The latter covaries with enhanced night-time stomatal conductance, for which evidence arose recently [e.g. Snyder et al., 2003].

For the purpose of exploiting upcoming measurements of atmospheric CO₂ vertical profiles by aircrafts and continuous CO₂ data recorded along tall towers as part of the North American Carbon Plan (NACP), a direct carbon budgeting approach is being developed.

Satellite measurements of total column CO₂ can be used in inverse models to help isolate sources and sinks; however, using satellite concentrations in inversions may introduce spatial, temporal, and clear-sky errors. Using a coupled ecosystem-atmosphere model, we found that using satellite measurements to represent temporal averages will introduce large errors into the inversion and that inverse models must sample the concentrations at the same time as they are measured.  Spatial and local clear-sky errors are much smaller than the instrumental errors, although they increase with domain heterogeneity. Inverse models can minimize sampling errors by using homogenous regions and sampling the CO₂ concentrations at the same time as the satellite.

In order to further our understanding of the biophysical and biogeochemical mechanisms that control the fate of fossil fuel carbon emissions, we are simulating an hourly global atmospheric carbon dioxide concentration field ([CO₂]) for the year 2000 with realistic diurnal, synoptic and seasonal variability, including quantified errors.  In addition, we are simulating carbonyl sulfide (COS) for a continental mixed temperate forest to test a hypothesis that errors in seasonal simulations of CO₂ result from incorrect specification of springtime onset of photosynthesis rather than incorrect timing of ecosystem respiration.

Biomass burning is an important source of atmospheric CO₂, aerosols and chemically important gases. It is as important to global chemistry as industrial activities in the developed world [Crutzen and Andreae, 1990]. Biomass burning is a key component of the global carbon budget, currently releasing 2.6 GtC from fires in the tropical and subtropical ecosystems (van der Werf et al. [2003], to be compared to the 5.6 GtC released from fossil fuels) to the atmosphere each year, most of it being emitted in the form of carbon dioxide, although there is important spread amongst various estimates. Biomass burning contributes up to 40% of gross atmospheric CO₂ (IPCC, 2001), 38% of tropospheric O₃, and 10 % of CH₄.

The synoptic scale atmosphere-biosphere interaction can cause anomalies of ~10 ppm with length scale of ~1000 km in the monthly averaged surface CO₂ concentration. These anomalies may contribute to the errors and uncertainties of CO₂ inversion estimates.