Relating temporal and spatial structure of the nocturnal surface layer to landscape heterogeneity

The temporal behavior of surface variables at night strongly depends on course of turbulent intensity. Early in the 20th century, Taylor and Richardson introduced what became known as the Richardson number (Ri) to describe how the state of "just-no-turbulence" can be maintained. Given relatively steady synoptic forcing, variation in mixing can be modulated by local differences in the elements of the surface energy balance. In landscapes heterogeneous both in surface type and in topography, some sites are more prone to mixing events than others. Our study aims to quantify the relationship between common surface descriptors and turbulent mixing in patchy forested terrain with modest topographic gradients, with the ultimate concern to improve parameterizations of subgrid exchanges.

We look at surface variables using a dense network of 26 surface stations in an area of 20 km by 20 km, in the region of Albany, NY. Nights were classified in two classes: When turbulence is active across the network ("connected nights"), there is diminished surface cooling and variables are spatially homogeneous. A simple parameterization is adequate. The opposite situation ("transition nights") occurs as turbulence decays to such a degree that surface disconnects from upper levels. In this case local conditions control the surface variables. Strong cooling, scalar jumps and spatial heterogeneity throughout the night characterize the transition.

Intermittence plays a role during both transition and connected nights, given that the Richardson number is close to critical. Sporadic leads to enhanced horizontal gradients. During transition nights, turbulent events may start and decay often at one or more stations, leading to isolated abrupt transitions in temperature and moisture. In the absence of intermittence, the consequent gradients would decrease in the late hours due to a radiative adjustment. There are also cases when turbulent activity keeps most of the network connected, except for a few intermittent places, "pockets" of no-turbulence embedded in a mixed environment, with peculiar temperature and scalar concentrations. We quantify the effects of intermittence, determining spatial and temporal scales of intermittent events and relating them to topography and land cover.