2503

9 7 834
Anais XII Simpósio Brasileiro de Sensoriamento Remoto, Goiânia, Brasil, 16-21 abril 2005, INPE, p. 2503-2509.

Estimativa de cota do Rio Alto Taquari utilizando NDVI e precipitação
Rafael Galvan Barbosa Ferraz
William Tse Horng Liu
Edson Luis Santiami
Moacir Ademilson Stumpf
Laboratório De Geoprocessamento
Universidade Católica Dom Bosco,
Av. Tamandaré 6000, Campo Grande 79117-900-Ms e-mail: will@ucdb
Abstract
This study is part of the research work dealing with the Upper Paraguay River Basin hydrological analysis and Pantanal floods and drought prediction. In this study, a river water level prediction model was constructed for the head water region of the upper Taquari river sub basin. The model was based on the river water level (RWL) as a function of rainfall and NDVI (Normalized Difference Vegetation
Index) obtained from linear regression approach. Monthly rainfall and RWL data of 1981 to 2000 provided by the ANA and monthly composite NDVI data of the same period provided by the
NASA/GSFC were used. Data of 1981 to 1989 were used for model construction and 1995 to 2000 for model validation. The results show that the absolute mean error of 4,62% and 4,84% were obtained for simulated and validated data set respectively. It is concluded that the simple statistical model predicted quite well the RWL using rainfall and NDVI data.
Key Words: River water level, NDVI, statistical model, Taquari river, prediction, cota do rio, modelo estatistico, previsão, cheias.

1. Introdução
O Pantanal tem a maior área úmida continental do mundo que coleta e drena a água da Bacia do Alto Paraguai (BAP). O Pantanal possui o mais diversificado patrimônio da humanidade, abriga inúmeros tipos de flora e fauna. O ciclo anual distinto de inundação fornece recursos naturais ricos para peixes e animais selvagens e este ciclo é economicamente importante para a região e mais recentemente tem-se desenvolvido o turismo ecológico. Os rendimentos dos peixes podem variar o seu ciclo de