Obtenção de níveis e declividade da linha d'água a partir de altimetria por satélite no Rio São Francisco

Abstract

Given the challenges of producing conventional hydrological data, remote sensing data are a complementary and/or alternative source of information in hydrology. Satellite altimetry (SA) is one of the techniques that provides data with potential use in hydrological and hydraulic studies. Altimetric space missions are still focused on obtaining ocean levels, ice caps and sea ice, however numerous studies have demonstrated over almost three decades the viability of its use in continental waters, especially in large rivers. The consolidation of altimetric data in medium/small watercourses remains a challenge, as working with narrow watercourses (width less than 1 km) implies a series of limitations, due to the characteristics of altimeter operation and data processing. However, technological advances in altimeters and the good number of planned missions make the use of SA data even more promising. In this context, this research aims to evaluate the performance of multimission SA in a medium-sized tropical river (São Francisco River) and its potential use to obtain hydraulic variables. Five altimetric missions (Envisat, Saral, Jason, Sentinel-3 and Cryosat) and in situ data from twelve fluviometric stations (FS or EF in Portuguese) with altimetric leveling (WGS84) in a 1400 km stretch of the São Francisco River were used, which enabled a statistical analysis in absolute terms, rarely found in the literature. Furthermore, the in situ data were interpolated in order to produce data synchronized to the location and time of the satellite's passage over the river, called VSs (virtual stations or EV in Portuguese), which reduced the RMSE by 0.29 m. In all, 76 EVs were analyzed for the fixed orbit satellites and 850 points for the Cryosat satellite (derivative orbit). Based on the data from these EVs, the performance analysis for the different satellites was conducted and the characteristic bias of each of them was estimated, something unprecedented for a medium-sized tropical river. Other conclusions could be drawn: the environmental characteristics of the surroundings of the EVs determine the level of errors for different satellites causing great amplitude of the deviations; the Sentinel-3 satellites present superior performance to the others, which is attributed to the SAR (synthetic aperture radar) technology combined with the operation model that uses previous information on the target's altitude (“open loop tracking mode”) and, finally, there is a general tendency for satellites to overestimate water levels. The slope of the waterline is a very valuable variable in hydrological and hydraulic studies, but it is difficult to obtain with acceptable levels of precision. In the absence of available local information, data from the SRTM mission (Shuttle Radar Topography Mission) are typically used, but its spatial resolution is 90 m. In this context, the consisted data of leveled EFs and bias-free data from dozens of analyzed EVs were used to obtain longitudinal and slope profiles of the waterline, in different temporal scales, separately and together. The results achieved, especially by the model of EFs merged with EVs, were compatible with the values mentioned in the literature for the Middle RSF stretch, but with a better spatial resolution than SRTM.