Harvard Forest Research

Quantifying advective and nonstationary effects on eddy fluxes in the Ameriflux network

Principal Investigator: David Fitzjarrald
Atmosphere Scientific Research Ctr: Jan 01 2008 - Aug 15 2010:

Abstract:
What causes underestimation of nocturnal CO2 respiration under calm conditions at most Ameriflux towers? We investigate two aspects of this issue: a) Horizontal subcanopy advection of CO2. In earlier work (Staebler and Fitzjarrald, 2004, 2005), we showed that horizontal mean transport, not previously measured in detail to that time, accounts for some, but not all of the CO2 that goes unrecorded. Presence of relatively dense subcanopy air (“drainage forcing”) identified nights with missing flux problems (“deficit nights”) significantly better than did the commonly used u* criterion. With several surface observations near the forest floor, we found that three weeks of calm nights were needed to reduce errors sufficiently to make accurate estimates of the advective CO2 losses. To compensate for our limited number of sensors, we advocated a similarity profile approach to integrating concentrations in the vertical. Now we ask support to return to this problem with a more refined observational plan, one that has the potential to be made operational. Our aim is to make an experimental design that may be exported easily to other Ameriflux sites. b) Consequences of sporadic intermittent mixing that is poorly captured by conventional eddy covariance flux calculations. This follows from our previous studies of fluxes resulting from resonant wave motions in the canopy (Fitzjarrald and Moore, 1990; Cava et al., 2003); and

Methods. Our proposed topics fit into the concept of a ‘flux supersite’. In part a) we propose an improved subcanopy wind and scalar measurement network, one that is designed for long-term operation. The proposed deployment will feature vertical profiles at multiple subcanopy locations linked to a common IRGA. ‘Satellite sites’ distant from the tower will feature separate CO2 sensors. In part b) we propose to continue ‘data mining’, work currently being performed as part of NE-NIGEC, to review raw (e.g., 10Hz) eddy sensor data and develop a turbulent climatology of intermittent mixing events. The objective is to determine how often these events occur and whether or not their presence can make any significant difference in long-term NEE estimates. The project will be done at the Harvard Forest Ameriflux site, where we have a long data archive and previous experience.

: