 |
| Titel |
A grid-based distributed flood forecasting model for use with weather radar data: Part 1. Formulation |
| VerfasserIn |
V. A. Bell, R. J. Moore |
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| ISSN |
1027-5606
|
| Digitales Dokument |
URL |
| Erschienen |
In: Hydrology and Earth System Sciences ; 2, no. 2/3 ; Nr. 2, no. 2/3, S.265-281 |
| Datensatznummer |
250000354
|
| Publikation (Nr.) |
 copernicus.org/hess-2-265-1998.pdf |
|
|
|
|
|
| Zusammenfassung |
| A practical methodology for distributed
rainfall-runoff modelling using grid square weather radar data is developed for use in
real-time flood forecasting. The model, called the Grid Model, is configured so as to
share the same grid as used by the weather radar, thereby exploiting the distributed
rainfall estimates to the full. Each grid square in the catchment is conceptualised as a
storage which receives water as precipitation and generates water by overflow and
drainage. This water is routed across the catchment using isochrone pathways. These are
derived from a digital terrain model assuming two fixed velocities of travel for land and
river pathways which are regarded as model parameters to be optimised. Translation of
water between isochrones is achieved using a discrete kinematic routing procedure,
parameterised through a single dimensionless wave speed parameter, which advects the water
and incorporates diffusion effects through the discrete space-time formulation. The basic
model routes overflow and drainage separately through a parallel system of kinematic
routing reaches, characterised by different wave speeds but using the same isochrone-based
space discretisation; these represent fast and slow pathways to the basin outlet,
respectively. A variant allows the slow pathway to have separate isochrones calculated
using Darcy velocities controlled by the hydraulic gradient as estimated by the local
gradient of the terrain. Runoff production within a grid square is controlled by its
absorption capacity which is parameterised through a simple linkage function to the mean
gradient in the square, as calculated from digital terrain data. This allows absorption
capacity to be specified differently for every grid square in the catchment through the
use of only two regional parameters and a DTM measurement of mean gradient for each
square. An extension of this basic idea to consider the distribution of gradient within
the square leads analytically to a Pareto distribution of absorption capacity, given a
power distribution of gradient within the square. The probability-distributed model theory
(Moore, 1985) can then be used directly to obtain the integrated runoff production for the
square for routing to the catchment outlet. justification for the simple linkage function
is in part sought through consideration of variants on the basic model where (i) runoff
production is based on a topographic index control on saturation and (ii) absorption
capacity is related to the Integrated Air Capacity of the soil, as obtained from soil
survey. An impervious area fraction is also introduced based on the use of Landsat
classified urban areas. The Grid Model and its variants are assessed in Part 2 (Bell and
Moore, 1998), first as simulation models and then as forecasting models, following the
development of updating procedures to accommodate recent observations of flow so as to
improve forecast performance in a real-time context. |
| |
|
| Teil von |
|
|
|
|
|
|
|
|