Temporal and Spatial Heterogeneities of Surface Fluxes in an Urban Environment

In this study, we examine the regional climatology of the water cycle in urban environments through analyses of the surface energy balance. Analyses center on long-term observations (2007 - 2011) and land-surface model simulations of latent and sensible heat flux for the Princeton University campus. Model analyses are based on the Noah Land Surface Model (LSM), which is widely used for problems involving coupled land-atmospheric interactions. The research site is characterized by a mixture of grassland, trees, and urban surfaces. Partitioning of net radiation between latent and sensible heat flux plays an important role in determining the regional rainfall climatology. This research is motivated by the following question: How does temporal variability of soil moisture and vegetation state affect the partitioning of net radiation between latent and sensible heat flux in a heterogeneous urban environment? We use turbulent-flux measurements at a 5-minute time scale from an eddy covariance station, as well as measurements of the surface radiation balance (upwelling and downwelling longwave and shortwave), meteorological variables, CO2 concentration, soil moisture, and precipitation. Additionally, a water vapor-CO2 (q-c) flux similarity method developed by Scanlon and Kustas (2010) was implemented to partition the measured latent heat flux at the eddy covariance station into bare soil evaporation and transpiration components; results from this partitioning are used to examine the Noah land-surface model formulation. We show that: 1.) soil moisture plays an important role in the surface energy balance, even for this heterogeneous urban environment, 2.) the Noah LSM does a poor job of capturing soil moisture variability, due largely to inadequate representation of vertical structure of urban soils, 3.) seasonal variation of vegetation state is important for the surface energy balance, and 4.) the bare-soil evaporation simulated by the Noah LSM exhibits a positive correlation with the values from the flux partitioning performed using q-c similarity theory.