INTERNATIONAL JOURNAL OF CLIMATOLOGY
Int. J. Climatol. 22: 1915–1933 (2002)
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/joc.856
VALIDATION AND RESULTS OF A SCALE MODEL OF DEW DEPOSITION
IN URBAN ENVIRONMENTS
K. RICHARDSa,* and T. R. OKEb
a Geography Department, University of Otago, Dunedin, New Zealand
b Geography Department, University of British Columbia, Vancouver, BC, Canada
Received 2 July 2001
Revised 6 August 2002
Accepted 6 August 2002
ABSTRACT
There is growing interest in urban dew and its significance in questions of urban climate and air pollution deposition,
but little research has been undertaken to study it. In this study, a generic, urban residential neighbourhood is modelled
out-of-doors at a scale of 0.125, using three wooden houses (1.08 m tall), a concrete pavement (1.0 m in width), a
grassed park (7.5 m in half-width) and several small trees (up to 1.5 m tall). The thermal inertia of each house is inflated,
according to the internal thermal mass (ITM) approach, so that nocturnal surface temperatures are conserved. First-order
validation was achieved through comparison with data collected at nearby full-scale sites in Vancouver, BC, Canada.
Moisture accumulation (measured by blotting on grass and by lysimetry) is found to be primarily controlled by nocturnal
weather conditions and the intrinsic nature of each substrate, e.g. dewfall is abundant on nights with few clouds and light
winds, and on surfaces such as grass and asphalt–shingle roofs, which cool rapidly after sunset. However, these responses
are modified by location effects related to the net radiation balance of the surface, which itself is strongly linked to site
geometry as expressed by sky view factor and whether surfaces are isolated from heat sources. The dominant mechanism
is argued to be the systematic increase in longwave radiation loss that is associated with increased sky view. Results
agree with those observed at the full scale and suggest that maps of sky view factor, and knowledge of dew at an open
site, can potentially be used to create maps of dew distribution in urban and other complex environments. Copyright 
2002 Royal Meteorological Society.
KEY WORDS: Canada; dew; model; scaling; thermal scaling; urban climate; urban dew
1. INTRODUCTION
The presence of dew is linked to many topics in the natural and applied sciences. Moisture from dew has a
role as a water source for plants and animals (Sharma, 1976), in the growth of fungal disease on crops (Huber
and Gillespie, 1992), in harvesting practices (Newton and Riley, 1964), and there is potential for dew to be
collected for human use, where other sources of fresh water are limited (Nikolayev et al., 1996).
In the case of cities, there is growing scientific interest in dew in relation to questions of urban humidity,
evaporation rates, near-surface temperature, air pollution deposition and the subsequent chemical damage
to plants and materials (Wisniewski, 1982; Okochi et al., 1998; Grimmond and Oke, 2000; Masson, 2000).
Despite this, little research has been undertaken to describe and explain meso- and micro-scale patterns
of urban dew, except work by Myers (1974), Mattsson (1979) and Richards (1999; Richards and Oke,
2000). Studies of dew tend to focus on rural areas (e.g. Newton and Riley, 1964; Sharma, 1976; Sudmeyer
et al., 1994); urban climate studies seldom address moisture fluxes beyond evaporation, and urban dew is
rarely measured.
* Correspondence to: K. Richards, Geography Department, University of Otago, P.O. Box 56, Dunedin, New Zealand;
e-mail: kr@geography.otago.ac.nz
Copyright  2002 Royal Meteorological Society

1916 K. RICHARDS AND T. R. OKE
Droplets of water that appear overnight on grass and other surfaces tend to be called ‘dew’, but it is
more rigorous to call this surface moisture because the source of the water may vary. Dew is strictly
surface water derived solely from condensation. It is commonly divided into dewfall, which is deposited
directly from vapour in the lower atmosphere, and distillation, the recent origin of which is moist soil or
adjacent wet leaves. Dewfall and distillation are similar in appearance, but they can be separated using
lysimetry (Slatyer and McIlroy, 1961) because the former is a net addition of water mass whereas the
latter is a redistribution between the soil and plant leaves in the lysimeter monolith. The amount and
spatial distribution of dew is primarily governed by surface temperature because this reflects the surface
energy balance that governs deposition rates. Dew is most common on nights with clear skies and light
winds, and when the lower atmosphere is moist (Garratt and Segal, 1988). Under such conditions, dew
deposition is enhanced because near-surface stability suppresses turbulent mixing and surface radiative
cooling dominates the surface energy balance. Dew is a small mass flux: accumulation on leaves seldom
exceeds 0.3–0.5 mm per night. However, this is not trivial in terms of the energy required to dry the
surface: at 10 °C, 1.24 MJ m−2 latent heat is required to evaporate 0.5 mm depth of water. On grass and
certain other plants, guttation may also be present, i.e. fluid exuded through leaf pores under internal
pressure. This occurs when leaves are fully turgid, surface air is close to saturation, and the rate of
water supply from the roots exceeds losses due to transpiration. The general consensus is that its rate
is a function of soil moisture and temperature, but the linking mechanisms are uncertain. It typically
forms large droplets on the tips of grass blades and its presence tends to coincide with dew (Hughes and
Brimblecombe, 1994).
Processes of dew formation are the same in urban and rural areas, but in cities the surface on which
dew forms is often more complex: a three-dimensional mosaic of many, often dissimilar, materials. In
contrast, rural dew is typically measured on a relatively extensive, homogeneous surface, e.g. a pasture
or crop canopy (Sharma, 1976; Sudmeyer et al., 1994). Scale modelling is consequently attractive in an
urban study because it permits the urban landscape to be reduced in complexity and size, and allows
structural elements (e.g. buildings and trees) to be more easily manipulated. Dew is difficult to create
artificially. Few have attempted to do so and methods are crude, e.g. outdoor refrigeration, droplets in a
fog chamber, or water sprayed onto leaves (Mulawa et al., 1986; Arends and Eenkhoorn, 1989; Fuentes and
Gillespie, 1992). Though adequate for investigating the consequences of dew, such as spore germination
or droplet chemistry, these approaches provide little information about its formation. Hardware models to
study dew deposition are rare and must operate out-of-doors where dew forms naturally (e.g. Mattsson,
1979). Out-of-doors models to study physical processes in built systems are more common (e.g. Davis
and Pearson, 1970; McPherson, 1980; McPherson et al., 1989; Wegner et al., 1988). None addresses dew
formation, but some provide insight into the surface temperature fields that are closely linked to processes of
dew deposition.
Cities seem to offer few sites favourable for dewfall compared with the vegetation in the countryside.
Buildings and pavements retain heat well into the night, and the urban core (‘downtown’) landscape
is associated with the urban heat island phenomenon. This is attributed in part to the large sensible
heat storage capacity of urban building materials (e.g. brick and steel), but also to anthropogenic heat
sources, increased shortwave radiation gain, and reduced longwave radiation loss in cities compared with
their rural surroundings (Oke, 1987). However, field evidence from Vancouver indicates that grass and
other vegetation (Spronken-Smith, 1994; Richards, 1999; Richards and Oke, 2000), and roofs (Richards,
1999; Richards and Oke, 2000) in urban residential (‘suburban’) areas can accumulate significant amounts
of dew when weather and site conditions are favourable. Further, ‘suburban’ landscapes in Vancouver
and many other North American cities possess a degree of regularity or repetitiveness (Schmid, 1988;
Schmid et al., 1991), so they can be modelled to a first order using relatively few generic landscape
elements (e.g. houses, streets, parks) arranged in representative configurations. The focus of this study
was consequently urban residential, rather than ‘downtown’ landscapes. In theory, if hardware modelling
can assess dew patterns within a scaled and simplified landscape, then the results can potentially be
extrapolated to larger areas of similar geometry. This paper presents the design, validation and results of
such a model.
Copyright  2002 Royal Meteorological Society Int. J. Climatol. 22: 1915–1933 (2002)