Environmental Engineering Reference
In-Depth Information
ers,
including Lin et al.
(2005) and Mirza
shown, which is interesting in that it was not
represented in any of the national or local NWP
models and did not register in any raingauge.
Nevertheless the event produced significant sur-
face flooding in the town of Tokoroa.
The operational solution to these difficulties is
to develop a coupled system for QPF that is driven
from the large scale by a global NWP model and
from the small scale by remote-sensed data from
radar and satellites via a nowcasting system. Then,
of course, this system needs to be linked to an
appropriate hydrologicalmodel to convert theQPF
to a flood forecast.
et al. (2008).
The estimation of rainfall distributions by radar
has a long history, described in Atlas (1990). Its
effective use, particularly in urban situations, has
a shorter history but has become relatively well
established in many places (see, e.g., Chapter 7 in
this volume, and Austin and Austin, 1974). The
estimation of rainfall amounts from satellite
images has been attempted from the early days of
meteorological satellites and the early work is
reviewed in Barrett and Martin (1981). Satellite
rainfall data are clearly required because even for a
small catchment the air masses that subsequently
produce the rainfall are very distant from the
catchment hours or days earlier. Using visible and
infrared (IR) geostationary data for the initial sit-
uation is clearly more realistic than assuming no
initial rain at all (Lovejoy and Austin 1978). More
recently, radars and microwave radiometers car-
ried by satellites now offer better rainfall pattern
measurements over large portions of the globe.
Estimates of global cloud water content based
onmicrowave radiometers at a resolution of about
50 km are now available several times each day
from the National Oceanic and Atmospheric Ad-
ministration (NOAA). These estimates depend on
the different radiative properties of the ground,
water vapour in the atmosphere, and ice and water
clouds. By using measurements in many different
frequencies in the microwave part of the spec-
trum, it is in principle possible to separate the
different components and thus estimate the path-
integrated rain, water vapour, water and ice
clouds. There are, however, some difficulties re-
maining, described in Horvath and Gentemann
(2007) and elsewhere, and this approach works
much better over ocean than over land. Higher-
resolution images of clouds are also available from
the Moderate Resolution Imaging Spectroradi-
ometer (MODIS) and Multi-angle Imaging Spec-
troRadiometer (MISR) instruments, and some
agreement is shown between their high-resolution
results and the lower-resolution microwave re-
sults for large clouds. Difficulties remain in re-
gions of partial cloud cover (Horvath and
Davies 2007). This is clearly an important area of
Hydrometeorology
It is clear that water in the form of rain, ice and
clouds is a major factor in the energy balance of
the global atmosphere. This is because water va-
pour is by far the major greenhouse gas in the
Earth's atmosphere and also because of the
huge effects clouds have on the global radiation
budget. The recognition of the intimate relation-
ship between the global energy balance and
global water cycle culminated in the establish-
ment of the international Global Energy and
Water Cycle Experiment (GEWEX) programme,
which is dedicated to providing a better under-
standing of the processes underpinning this link-
age (Lawford et al. 2004). Although NWP is
modelling a complex and chaotic set of processes,
it is still in essence a classical initial value prob-
lem. In spite of this, NWP models are often ini-
tialized with no clouds or rain. The models 'spin
up' and subsequently generate their own clouds
and rainfall patterns. Given the intimate relation-
ship between the energy of the atmosphere at any
time and the amount of liquid water and vapour it
would seem highly desirable to include cloud and
rainfall distributions in the data assimilated into
NWP models. It is therefore to be expected that
assimilation of high-resolution cloud water con-
tent and rainfall data from satellites and rain
radars should result in improvements in the pre-
diction of rain and severe weather in particular. In
fact this has been demonstrated bymany research-
