Fuzzy Hydrologic Model in Tropical Watershed - Advances in Data, Methods, Models and Their Applications in Geoscience - page 285

14
Fuzzy Hydrologic Model in
Tropical Watershed
Aurélio Azevedo Barreto-Neto
Federal Institute of Espírito Santo,
Brazil
1. Introduction
Natural elements such as soil, geology and vegetation are usually represented as map
classes, whose boundaries are sharply defined. However not all entries in geo-objects
datasets are sharp and this is true both for their attribute values as well as for their spatial
distribution (Burrough, 1996).
This traditional representation between geo-objects is considered as an oversimplification of
a more complex pattern. In some conditions, these boundaries are recognized more easily
because are associated to significant and abrupt land changes, such as situations in which
the boundaries are located in river banks, in geologic phenomenon (intrusions, flaws,
fractures) or associated with sudden relief variations (Burrough, 1986). Apart from these
situations, the boundaries are associated to uncertainties caused by limited observations
(Hadzilacos, 1996). In all these cases, fuzzy methods are more suitable than boolean logic.
Zadeh (1965) developed fuzzy set theory allowing the mathematical modeling in zones of
imprecisions and uncertainties. Fuzzy set theory is a generalization of the boolean logic to
situations where data are modelled by entities whose attributes have zones of gradual
transition, rather than sharp boundaries. Studies of natural phenomena and natural objects
demonstrated that the use of boolean logic is an inadequate method and brings much
inferior results (Burrough, 1986).
The objective of this study was to develop a fuzzy rule-based modelling to predict runoff
in a watershed using the Soil Conservation Service Curve Number (SCSCN) model (SCS,
1972).
Although the SCSCN model was developed primarily based on small watersheds, it can be
applied in medium and large watersheds, with a diversified variety of soils and vegetation,
if integrated to a geographical information system (GIS) (Johnson & Miller, 1997; Thompson,
1999).
2. Study area
The study area is the Quilombo River watershed, located in Ribeira Valley, South of the
State of São Paulo, Brazil (Fig. 1). The land-cover of the area is composed of Atlantic forest
(dominant) and pasture. The choice to study this watershed was driven by the availability of
soil map, rain record gage, and stream discharge record gage.