Calculating the amount of atmospheric carbon dioxide absorbed by the oceans Helen Kettle & Chris Merchant School of GeoSciences, University of Edinburgh, UK. CASIX: Centre for Air-Sea Interactions and fluXes Background Calculating the transfer of CO 2 between the air and sea Accuracy of the final value - our role in CASIX CO 2 transfer = solubility x gas transfer velocity x [ pCO 2 sea – pCO 2 air ] Dr. Helen Kettle Dr. Chris Merchant H Kettle & C Merchant Systematic errors in global air-sea CO 2 flux caused by temporal averaging of sea-level pressure. Atmospheric Chemistry and Physics (in press). This varies in time and space. In general the carbon flux is out of the sea around the equator (source regions – shown in red) and into the sea in the higher latitudes (sink regions – shown in blue). This is the ability of CO 2 to dissolve in sea water. This varies with sea surface temperature (SST). SST is a very influential variable because it affects solubility and ocean circulation (cold water sinks removing CO 2 from the surface waters). Luckily we can measure SSTs over the whole planet using satellites. This is the rate at which CO 2 moves between the atmosphere and ocean. It is strongly influenced by wind speed. There are several different ways of calculating the transfer velocity (see figure below). Luckily wind speed over the oceans can also be measured by satellite. pCO 2 sea is a measure of the amount of CO 2 in the ocean. It depends upon the plants and animals living in the ocean and the sea surface temperature and salinity. A large amount of pCO 2 sea data have been collected by ships but we can also use satellite data (ocean colour). To model changes in the amount of carbon in the ocean a computer model describing interactions between the plants and animals in the sea is needed. pCO 2 air is a measure of the amount of CO 2 in the air. It depends upon the air pressure and humidity at the sea surface. The relative sizes of pCO 2 air and pCO 2 sea control whether the carbon goes into or out of the ocean. The final value for the amount of CO 2 absorbed by the oceans is very sensitive to the kind of data used in the above equation. Data may be from different sources (satellite, in-situ or modelled) and at different time and space intervals. The final value depends on SST, salinity, wind speed, air pressure, biology and chemistry. These factors are themselves dependent upon each other. This means that if averaged data are used then the effects of these dependencies may be ignored. 1. Increasing amounts of CO 2 in the atmosphere from fossil fuel burning and changes in land-use have caused the mean global air temperature to rise (see figure). 2. Some of the excess CO 2 emitted to the atmosphere is used for plant growth through photosynthesis and some is absorbed at the ocean surface and then transferred to deeper water. These are ‘carbon sinks’. 3. The increase in the amount of CO 2 in the atmosphere is less than would be expected by simply balancing our best guesses of the values of carbon sources and sinks. This implies we need to improve our calculations! 4. CASIX (NERC project) aims to calculate accurately the size of the ocean sink using satellite data and computer models which simulate the transfer of carbon by water movement, chemical interactions and through plants and animals in the ocean. A CASIX member recently produced a new equation that includes an additional component based on wave- breaking. David Woolf, 2005, Tellus (in press). The left figure shows how the global air-sea CO 2 flux varies with time and according to the type of pCO 2 air data used (i.e. 6 hourly, monthly or climatological). The right figure shows how the errors caused by time-averaging vary with time of year. [Red: WM99, black:W92 (gas transfer equations)]. The left and centre maps show mean errors in flux (mol C /m 2 /yr) caused by monthly averaging of pCO 2 air data for 2 different gas transfer parameterisations. The right map shows the covariation in wind and pressure (mb m/s). There is an obvious relation between flux errors and the covariation of wind and pressure. The figures above show the relation between the flux error caused by monthly averaging of pCO 2 air data and the wind-pressure covariation averaged over , using 2 gas transfer parameterisations. The large amount of scatter is due to the wide range of pCO 2 sea and sea surface temperature in regions with the same wind-pressure covariance. Example: The effect of using monthly averaged atmospheric CO 2 data on our estimation of the size of the ocean carbon sink. We found the best estimate of the amount of carbon absorbed by the oceans is actually overestimated by about 12%. This is because most calculations are done using long-term averaged pCO 2 air data - either monthly averages or averaged over many years (climatological). This means that the daily fluctuations in atmospheric carbon dioxide due to changes in air pressure are ignored. When air pressure is low, wind speed is generally high (and vice versa) - if time averaged data are used this negative correlation is ignored leading to large errors. See figures below for details. From