Boundary-Layer Meteorology, 93 (2): 287-312, November 1999
Scale modelling is used to
determine the relative contribution of heat transfer processes to the nocturnal
cooling of urban parks and the characteristic temporal and spatial variation of
surface temperature. Validation is achieved using a hardware model-to-numerical
model-to-field observation chain of comparisons. For the calm
case, modelling shows that urban-park differences of sky view factor (ψs) and
thermal admittance (μ) are the relevant
properties governing the park cool island (PCI) effect. Reduction in sky
view factor by buildings and trees decreases the drain of longwave radiation
from the surface to the sky. Thus park areas near the perimeter where there may
be a line of buildings or trees, or even sites within a park containing tree
clumps or individual trees, generally cool less than open areas. The edge
effect applies within distances of about 2.2 to 3.5 times the height of the
border obstruction, i.e., to have any part of the park cooling at the maximum
rate a square park must be at least twice these dimensions in width. Although
the central areas of parks larger than this will experience greater cooling
they will accumulate a larger volume of cold air that may make it possible for
them to initiate a thermal circulation and extend the influence of the park
into the surrounding city. Given real world values of ψs and μ
it seems likely that radiation and conduction play almost equal roles in
nocturnal PCI development. Evaporation is not a significant cooling mechanism
in the nocturnal calm case but by day it is probably critical in establishing a
PCI by sunset. It is likely that conditions that favour
PCI by day (tree shade, soil wetness) retard PCI growth at night. The present
work, which only deals with PCI growth, cannot predict which type of park will
be coolest at night. Complete specification of nocturnal PCI magnitude requires
knowledge of the PCI at sunset, and this depends on daytime energetics.