Micro-meteorologists began studying local advection decades ago, and have usually focused on idealised cases, involving geometrical simplification. Such an idealised case is the airflow off an extensive region of bare, dry soil onto wet soil. This is a problem in only two spatial dimensions (height z, downwind distance x) provided the line of dry/wet discontinuity (lying at x=0) stretches far in the crosswind (y) direction, so that lateral (y) gradients are negligible. One speaks of an "internal boundary layer," a layer of modified air lying above the downwind (damp, in this case) surface, and growing slowly in its depth d(x) as distance x downwind from the discontinuity increases.
An early and very influential experiment on local advection was undertaken by Rider, Philip and Bradley (1963), who studied the flow from tarmac to grass at Canberra airport. The experiment was analysed in the context of J.R. Philip's analytical solution for local advection, which (not too unreasonably) assumed power law profiles for windspeed and eddy diffusivity (eg. mean windspeed U=U1 (z/z1) n where U1 is the windspeed at reference height z1), but adopted what is for most circumstances a too highly over-simplified lower boundary condition on temperature and humidity.
With the transition to numerical modelling of micrometeorological flow, came the possibility of more realistic treatment. Taylor (1970) modelled flow over surface changes, using a first order (K-theory) closure, in which the eddy diffusivity (K) and mean windspeed (U) were allowed to vary according to the local temperature stratification, which varied as the wind passed across the landscape. A still more general treatment was given by Rao, Wyngaard and Cote (1974a,b; hereafter RWC), who modelled local advection using second order closure, so that buoyancy effects arose more naturally through the governing equations for vertical velocity variance sw2 and other properties. Skipping forward twenty years (and making no reference to many intervening contributions to our understanding of local advection), Bink (1996) showed that the RWC model agrees rather well with the profiles of mean temperature and humidity observed in the cool, moist internal boundary layer observed on the Crau plain, in southern France, downstream of a dry-moist transition.
Wilson et al. (2001) computed mean potential temperature profiles for La Crau Run 42, from an implementation of the RWC equations solved using Patankar's SIMPLE method. Note the unstably-stratified upstream temperature profile (T decreasing with increasing height z), and the growth of a deepening cooler layer with increasing distance (given on the figure in metres) over the moist surface.
Keith McNaugton and Johannes Laubach studied the cool, moist internal boundary layer over irrigated rice, several hundred metres downwind of a hot, arid region in Australia. They showed that the "brief-averaged" (ie. circa 30 sec mean) evaporation rate varies in step with the "slow" changes in horizontal windspeed, changes originating from the largest boundary-layer eddies advecting from the convectively-unstable plain upstream. Theirs is one of only a very few studies to time-resolve the variations in the near-ground fluxes, and perhaps the first to explicitly focus on those changes, emphasizing the rather different boundary conditions that are imposed on different species. Their study suggests that the theoretical framework for describing these "slow" fluctuations in surface fluxes - which by the way result in non-equality of the eddy diffusivities for different species - must at a minimum resolve at least two spectral bands: the "slow, passive" cycles modulating the fluxes, and the faster "active" fluctuations ("passive" refers to the fact that the slow u fluctuations are not correlated with vertical speeed, so carry no shearing stress).
References:
Back to John D. Wilson home page.
Back to the Earth & Atmospheric Sciences home page.
