Acurácia de modelos digitais de elevação e processos hidrossedimentológicos em microbacia hidrográfica no semiárido

Abstract

The use of Remote Sensing (SR) helps in interpreting landscape features, especially at the watershed scale. Remotely Piloted Aircraft (RPA) is in evidence for the generation of Digital Elevation Models (DEM) with low execution cost and higher data accuracy. DEM’s represent the three-dimensional variability of terrain providing essential information for hydrology, erosion and watershed characterization studies. Water erosion is intensified by the interaction of edaphoclimatic factors and accelerated by human interference. Studying erosion on an erosion parcel scale helps to understand hydrosedimentological processes on larger scales. In an attempt to mitigate the effects of erosion, conservationist practices are presented as alternatives to conventional management in semi-arid agriculture. Thus, the objective was to evaluate the quality of digital elevation models for the morphometric characterization of the watershed and to identify the influence of soil management on runoff generation, soil losses and nutrients. The research was conducted in the Curu Valley Experimental Microbasin (MEVC), located in the Curu Valley Experimental Basin (BEVC), whose exutory coordinates are latitude 3 ° 48 '16.15' 'and longitude 39 ° 20' 47,28”, belonging to an experimental farm of the Federal University of Ceará, in the city of Pentecoste, Ceará. DEM’s obtained by SRTM and ALOS PALSAR data, and MDEs obtained by RPA with flight height of 100, 150 and 200 m were used in 2017 and 2019 expeditions. The aerofogrammetric data obtained by Argsoft PhotoScan software and the MDEs framing were evaluated. As for the Cartographic Accuracy Standard (PAC), the influence of the DEMs on the morphometric characterization of the SEMC was analyzed. To determine the erosion in plot scale, four different soil coverings were used: CN12- DTF fallow 12 years ago, CN43- DTS fallow 43 years ago, CPL-cowpea cultivation with leiras and CPq - cultivation of cowpea towards slope after burning. The period studied ranged from February to July 2019. Surface runoff volume was quantified at each erosive event, and samples were collected to determine sediment and nutrient loss. Thirteen erosive events were monitored, with runoff of at least one plot. The alignment of the images by RPA was influenced by the Dry Tropical Forest (FTSS) that make up mostly the cover of the watershed. SRTM, ALOS and RPA elevation data at 200 m presented average errors of -3.72; -3.90 and 3.99 in relation to the validation points, respectively. The best framing for PEC was for the 100 m RDE MDE in the 2017 expedition, which was rated class A on the 1: 2000 scale. The other models to be evaluated as class A would have to be in scales above 1: 10,000. High spatial resolution DEM’s more accurately generated the watershed drainage lines. The watershed was characterized as elongated shape and with low tendency to flood, in all automatic delimitations. Regarding the influence of management on runoff generation, it was observed that the most erosive events were concentrated in April and July. The CN43 plot presented only one runoff during the whole study period and the soil loss in the conservation practice plot was lower than in the DTS fallow plot 12 years ago. The largest soil and nutrient losses were registered in the conventionally managed plot. The use of conservation practices in the semiarid region has been shown to have great potential for reducing soil and nutrient loss, however, had no effect on reducing the total volume of runoff.