Seasonal Covariance of Baroclinicity and Ecosystem Metabolism

GCM projections of climate change depend critically on our understanding of the natural carbon cycle, and the spatiotemporal distribution of long term carbon sources and sinks over land and ocean. Inverse modeling combines tracer transport models with atmospheric CO2 observations and a priori information about CO2 exchange at the surface to determine source/sink distributions. Despite some uncertainties, inversions remain a powerful tool for carbon research and policy support, and more surface and space-based observations will undoubtedly improve reliability. Large uncertainty, however, remains in CO2 transport algorithms; in particular, covariance between weather and the atmospheric carbon cycle.

It has long been recognized, for example, that seasonal and diurnal covariance between terrestrial ecosystem metabolism and fine-scale vertical transport in the atmosphere is a strong determinant of vertical structure in CO2 (“rectifier effect”). Numerical treatment of this subgrid-scale process is a leading source of uncertainty in CO2 inverse models. Transport by baroclinic waves in the mid-latitude storm track is another poorly-resolved process which controls the distribution of atmospheric trace gases on synoptic to seasonal time scales. Like the PBL-modulated CO2 rectifier, synoptic transport involves strong vertical motion and is correlated with ecosystem metabolism because large-scale baroclinicity and photosynthesis are both driven seasonally by variations in solar radiation.

The general circulation of the atmosphere is decomposed into a zonally symmetric circulation comprised of time mean and eddy components. Forward atmospheric transport simulations using PCTM, weather driven by Goddard EOS Data Assimilation System, and process-based surface fluxes of CO2 for 2005 show that transient and stationary eddies in Northern Hemisphere (NH) midlatitudes are responsible for strong CO2 transport out of midlatitudes where biological carbon exchange is large into polar regions where carbon exchange is weak. During Northern Hemisphere (NH) winter when baroclinic activity peaks, eddies transport nearly twice as much carbon poleward within midlatitude storm tracks on moist isentropes as on dry isentropes.

The idea that atmospheric mixing disperses signals away from fluxes regions, thus reducing mean meridional concentration tendencies, is not new. When computed on sigma coordinates, Fung et al (1983) determine that column integrated eddy transport in mid-latitudes is down-gradient year-round. We perform similar transport analysis on moist theta and dry theta, in addition to hybrid sigma-pressure (the native grid of PCTM), and compare to Fung et al (1983). We arrive at the same conclusion that net mixing typically diffuses gradients, but find the dominant mechanism is sensitive to season and vertical coordinate. Similar to Fung et al (1983), when averaged on hybrid sigma-pressure, eddy transport is down-gradient and diffusive, acting to disperse concentration gradients. When averaged on isentropic surfaces, however, eddy transport becomes counter-gradient during NH winter, opposite in sign to transport on σ and sigma-pressure, with mean meridional transport compensating through down-gradient transport. From an isentropic perspective, synoptic eddies therefore act to maintain gradients where baroclinic instability is strong. Eddy variations of mass flux are shown to be much larger on moist theta compared to dry theta, sigma, and hybrid sigma-pressure, especially during winter months in the NH, such that correlated variations with CO2 are also much stronger, a likely explanation for the increased strength of eddy transport on moist and dry theta.

It has been known that meridional transport strongly mediates the seasonal cycle of CO2 in middle latitudes and amplifies it in high latitudes, accounting for most of the observed seasonality at sites like Barrow, Alaska and Alert, Canada. Previous comparison of satellite vegetation indices with atmospheric CO2 tendencies imply that atmospheric mixing must influence signals at high latitudes where vegetation is nonexistent. The role of eddies at high latitudes, however, is more important than previously perceived. Although the relative roles of eddy vs mean transport has much seasonality, the pattern at high latitudes is for eddies to dominate CO2 tendencies.

While the seasonal covariance of subgrid-scale transport with surface CO2 fluxes is interesting from an earth science point of view, it is also relevant from a flux estimation perspective. Such covariance between these two highly uncertain processes inevitably leads to errors in flux estimates if not represented correctly in transport models. Synoptic transport of trace gases by frontal systems in middle latitudes occurs on scales that are typically unresolved in global models, and like the CO2 rectifier may be a source of inversion errors. Additional simulations were run using identical surface fluxes to calculate transport at three different resolutions for four different products. We found dramatic differences in seasonal meridional transport of CO2 by stationary and transient waves. Misrepresentation of these transport processes in global models are inevitably aliased into errors in estimated fluxes in CO2 inversions.