Variabilidade espacial e representatividade amostral do carbono e parâmetros fisiológicos em sistema silvipastoril

Abstract

For a correct evaluation of the organic carbon dynamics in the soil-plant system is necessary to improve the collection methodologies due to their spatial variability. Thus, the objective was to propose a sampling methodology considering the spatial variability of soil organic carbon and physiological parameters in a silvopastoral system implanted in the northern region of Minas Gerais. The collections were performed in January 2018 in a silvopastoral system with double row arrangement (with E. urograndis planted in 3 x 2m spacing) and alleys (grown with B. brizanta cv. Marandu) at 14m spacing between rows. They were randomized four sampling transects within the system (with a usable area of 844m2), and they were taken samples at depths of 0-5, 5-10 and 10-20 cm, linearly, between the two rows, totaling 20 sampling points. Each transect was divided into side A (left row and left side of the alley) and side B (right row and right side of the alley), both with 10 meters long. Inside the alleys were carried out physiological evaluations on the marandu-grass plants using gas analyzer for infrared (IRGA) and determined stomatal conductance (Gs), transpiratory rate (E) and carbon dioxide assimilation rate (CO2) or photosynthetic rate (A). To identify the best sampling strategy for future experiments in silvopastoral system, it was used the modified maximum curvature method and the Hatheway method. According to the modified maximum curvature method for TOC analysis, sampling should be performed at seven points (four samples in the row and three in the alley). Already for Cmic of the soil in silvopastoral system, sampling should be performed at 11 sampling points (four random points in the alley and six points in the rows). The samples for physiological parameters of marandu-grass within the silvopastoral system should be composed by collection at 13 random points within the area (alley). Using the Hatheway method, due to the spatial variability of the silvopastoral system, it is not recommended to install experiments with high precision (minimum significant difference (DMS) close to 20%) since that there is a need to increase the number of samples collected in reason of the high estimates of coefficients of variation. Thus, considering an experiment in randomized block design (DBC) with 5 treatments and 4 replications, with 8 samples in the alley and 3 in the row it is possible to distinguish treatments with DMS below 23% for TOC and 71% for Cmic, while 23 evaluations within the plot make possible to distinguish of treatments with DMS below 80% for the estimated physiological parameters.