Análise exploratória da correlação entre espectrorradiometro de campo e variaveis de solo e planta em vinhedo

Abstract

The objective of this work was to evaluate the replacement of conventional procedures of physical-chemical laboratory analysis of agronomic parameters in viticulture by remote sensing techniques through their correlation with biological variables and agronomic attributes of Vitis vinifera L. A sample grid with 74 points on 3 ha of vines was established. For reflectance readings, the FieldSpec Hand Held 2 portable spectrometer was used and 17 vegetation indices (VI) (SR, NDVI, mNDVI, SGR, RGR, NPCI, SRPI, NPQI, SIPI, PI1, PI2, PI3, EVI, PRI, PSRI, PWI e ARI) were calculated, in two field reading methods, by leaf clip (CP) and by proximal sensing (SP). The VI were related by Pearson's correlation coefficient, at 5% level, with the variables of apparent electrical conductivity (ECa), soil chemistry and physics, chemical analysis of plant tissue, oenological and phytotechnical parameters. Based on the obtained correlations it was possible to affirm that the use of IV instead of the physical / chemical analysis of agronomic parameters is viable, provided that they reach at least one correlation coefficient with moderate magnitude, where we found 36 moderate coefficients. The VI and the analyzed soil variables that presented the highest Pearson correlation coefficient and at least moderate magnitudes for the 0- 20cm depth were mNDVI-SP with Effective CTC (0.508), SR-SP with CTC pH 7,0 (0.534), RGR-SP with K (-0.447), NPQI-SP with Cu (-0.472), P1-SP with K (0.546), Pl2-SP with CTC pH 7.0 (-0.545), and PRI-CP which correlated with both CTC pH 7.0 (0.572) as well as K (-0.565). Already for the depth of 20-40 cm we obtained only the mNDVI-CP with the K (-0.534). For the relationship between IV and plant tissue variables the best results were ARI-CP with B (-0.402), Pl3-SP with K (-0.506) and ARI-SP with K (0.566).