Stochastic simulation applied to temporal analysis of the potentiometric level in the surroundings of the Três Irmãos dam located in the urban area of Pereira Barreto (State of São Paulo, Brazil)

Abstract

The filling of the reservoir of Três Irmãos in Pereira Barreto (State of São Paulo, Brazil) caused the rise of the surrounding potentiometric level with geotechnical implications to the environment and to human occupation, imposing risks that must be mitigated. Data from control wells located along the reservoir banks were collected during monitoring campaigns before, during and after the reservoir filling, generating a set of potentiometric surfaces over time. This structure outlined a spatial-temporal analysis. The potentiometric surface was modeled by turning bands stochastic simulation in order to perform the analysis of variance using two simulations that represent opposite scenarios. The first scenario showed the lowest water level rise and the second showed the highest water level rise. This resulted not only in a three-dimensional model for water level rises but also allowed the interpretation of three periods of distinct and characteristic hydraulic behaviors: 1) an initial state of hydraulic equilibrium before the filling of the reservoir, 2) a momentary disequilibrium and heterogeneity during filling, and 3) the regaining of the balance when filling was completed. The algebraic analysis of the images was also performed by means of a filter defined to combine the two scenarios and enhance their contrasts. The application of this filter generated a map of variances that highlighted two regions: one with a wide variation located in the center-northeast of the area and another less susceptible to variations located southwest of the study area, respectively regions of highest and lowest potentiometric level variations.