OLIVEIRA, J. J.; http://lattes.cnpq.br/7481208418394241; OLIVEIRA, João Jácome de.
Resumo:
The objective of this work, had to study the spatial variability of chemical
characteristic in the saline-sodic soil. The experiment was carried out in the Irrigated
Perimeter of Sao Goncalo, Sousa, Paraiba, in Experimental Station area of the Embrapa
Algodao. The experimental unit was of 2912m2, where were demarcated 182 observations
points, ordered in a square mesh of 4,0x4,0m. The soil samples were colletecd at the
profundity of 0-3 Ocm, in each experimental mesh point. The chemical determinations were
realized in the Salinity and Irrigation Laboratory, belonging Agricultural Engineering
Departament of Federal University of Paraiba, Campina Grande-PB. The data analysis
were realized through descriptive statistic and geoestatistic. The descriptive analysis were
realized with objective of to observe the data deportment usually of and the geoestatistics
analyses through of the semivariogram view of, to examine the spacial structure of
evaluated chemical variable. From of parameter estimate fitted to semivariogram of model
was, realized the krige, in the mesh of 2,0x2,0m, totalized 624 points, to construction of
isolines maps. By descriptive analysis showed that the data of pH presented variation fewer
(CV=11,06%), followed of CTC (CV=39,43%), PST (CV=43,06%), K (CV=48,24%), Na
(CV=65,93%), CE (CV=79,20%), Ca (CV=80,43%) and Mg (95,68%). The analysis of
frequency distribution, showed that the data of CTC, followed a normal distribution. The
data of CE, K and Mg followed a log-normal distribution. The another chemical variable
not fbllwed non type of distribuition evaluated. The geoestatistic analysis realized showed
that the models best of to semivariogram fited of data of chemical variable, PST, CE,
CTC, K, Ca and Na was the spherical. The data of pH and Mg not suggested structure of
spacial dependence, when analised, the assesment total of data, but, after the expuged of data considerate discrepancy, in case of Mg and data that stayed characterized a trend in
case of pH, was feasible to fit gaussiano and linear models, respectively, to pH and Mg.
The semivariogram analysis showed spatial structure with medium scope of 21,15,12, 12,
25, 27, 23 and 16m, respectively, to pH, CE, PST, K, Ca, Mg e Na. The chemical variable
PST, K and the Ca, suggested fort spatial dependence, while, the another suggested
moderate spatial dependence. The isolines maps, obtained the leave of estimated values by
krige of process, permited a visual of regions of variability minor and major for each
chemical variable analyzed, the what is important at the recommendation from mangeable
this regions, as also in the experimental delineatement of futuros works in the area.
