Geoscience Reference
In-Depth Information
orographic effects in objective analysis in the derivation of areal averages (Chua and Bras,
1982; Phillips
et al
., 1992). In a different approach, Daly
et al
. (1994) relied on digital
elevation data, including height and aspect, to derive local regression equations with which
to distribute point measurements of monthly and annual precipitation to regularly spaced
grid cells.
Some distribution functions
The areal distribution of precipitation over different time scales has been the subject
of intensive research and numerous relationships have been proposed in the literature
(see, for example, Court, 1961; Burns, 1964; Huff, 1966; Fogel and Duckstein, 1969)
to reduce the estimated point precipitation, when it is to represent a larger area. In some
of these, correlation studies were carried out relating the decay with increasing distance
from a gage at a central location (Huff and Shipp, 1969; Hutchinson, 1969). For design
purposes for use with point rainfall frequency data (and not with individual storms), in
the United States the reduction factors shown graphically in Figure 3.14, developed by
the US Weather Bureau (1957-1960; Miller, 1963; Myers and Zehr, 1980), have been
widely used. Various aspects of areal rainfall reduction procedures have subsequently
been studied by Rodriguez-Iturbe and Mejia (1974), Eagleson
et al
. (1987), Smith and
Karr (1990), Omolayo (1993), Sivapalan and Bloschl (1998), Asquith and Famiglietti
(2000), DeMichele
et al
. (2001), and Allen and DeGaetano (2005) among others. Several
of these investigations have shown that the reduction factor dependency on area is also a
function of the severity, i.e. the return period of the event. However, it was also generally
found that for several reasons the curves shown in Figure 3.14 are on the safe side, and
therefore are likely to result in a more conservative design. Thus actual precipitation
tends to decrease more rapidly with area covered, as the return period increases, than
indicated in the figure; also, convective type storms exhibit a more rapid decrease with
area than those of the stratiform type. Finally, although the reduction factors shown in the
figure appear to level off at around 1000 km
2
, it has been found that in fact the reduction
factor continues to decrease in an exponential manner as the area increases beyond that,
even up to 20 000 km
2
.
3.3.2
Temporal distribution
Precipitation is normally recorded on an hourly or daily basis, and in the data records
it can be reported over different averaging periods. The description of the evolution of
precipitation over time depends largely on the adopted temporal resolution. In applied
hydrology, a record of precipitation intensity with time for individual storm events is
commonly referred to as a
hyetograph
; a hyetograph is usually presented as a bar graph
with an hourly time step. The accumulated precipitation over time, is called a
mass curve
.
A
double mass curve
is a graph of seasonal or annual accumulated precipitation
at a given station plotted against the mean accumulated precipitation for a number of
neighboring or surrounding stations. Double mass analysis was introduced by Merriam
(1937) to check the consistency of the record at a station that has undergone changes
