Efficiency of Turbulent Transfer over an Urban Surface in Relation to Seasonal Changes in Source Area Patchiness

Much of our understanding of land-atmosphere interactions is rooted in direct flux measurements by the eddy covariance (EC) method. The EC method of analyzing in-situ data from flux towers has become the standard tool for monitoring turbulent heat exchange, trace gas emissions, and the water balance over land surfaces, including cities. The EC method has been developed for horizontally homogenous ecosystems like croplands and forests, where the source area has a uniform distribution of sources and sinks. This is not necessarily the case in urban ecosystems, where sources and sinks of momentum, sensible heat, water and carbon dioxide are dissimilar and can vary both horizontally, vertically, and temporally. The goal of this work is to establish a link between surface patchiness and the efficiency of the turbulent exchange of momentum and scalars. This will be useful in developing criteria for urban EC measurements to assess the quality and representativeness of measured fluxes. Most studies to date have explored the dissimilarity and efficiency of the turbulent exchange of momentum and scalars in cities as a function of atmospheric stability and local conditions - irrespective of potential seasonal changes.

This study uses a long-term (8 years) EC dataset from a flux tower located in a residential area in the City of Vancouver, BC, Canada (49.2ºN, Fluxnet ID "Ca-VSu"). The source area of the EC system is representative of a typical urban setting with detached single-family homes (average height: 5.3 m), and  tree coverage of 17.1 stems / ha (LCZ 6). Data was gathered from EC instrumentation at a height of 28.8 m using a sonic anemometer (CSAT-3, Campbell Scientific Inc.), and an open-path infrared-gas analyzer (Li-7500, Licor Inc.). Relationships between the urban surface patchiness and the turbulent exchange of momentum, sensible heat, water vapour, and CO₂ were established by calculating correlation coefficients between vertical wind and the turbulent entity of interest. By conditionally sampling correlation coefficients from particular wind directions, atmospheric stabilities, and times of year and day, seasonal changes are isolated. Attribution of fluxes to seasonally changing surface characteristics was done using source area models in combination with remotely sensed data (LIDAR, Aerial photos) and GIS data (traffic counts etc.).

Seasonal analyses showed a higher correlation coefficient of water vapour flux during spring, fall, and winter, relative to lower summertime values, when many pervious surfaces dried out, and only selected lawns were irrigated. The efficiency of water vapour fluxes was positively correlated to soil volumetric water content. Fluxes of CO₂ were found to exhibit higher correlation from wind directions where major traffic corridors exist, and CO₂ fluxes from residential sectors changed seasonally as home heating by natural gas (more uniformly distributed exhaust from chimneys) caused higher efficiencies during colder weather. Hence, the efficiency of CO₂ fluxes was positively correlated as a function of heating degree days in these sectors. Sensible heat was more efficiently exchanged in summer than winter, presumably due to increased shadowing in winter. The momentum transfer was slightly more efficient during leaves out compared to the leaves off season.