Stochastic Modeling of Anomalous Transport and Temporal Dependence in Trajectories
stochastic processes; random walk; Brownian motion; anomalous diffusion; continuous-time random walk; CTRW; alpha-stable distributions; hydrocarbon reservoirs; seismic anomalies; Dinoponera quadriceps; time series; autocorrelation
This dissertation investigates different classes of stochastic processes and their applications in two contexts: the simulation of anomalies in hydrocarbon reservoirs and the analysis of individual trajectories of Dinoponera quadriceps. First, normal diffusion is introduced as a null model for transport, starting from the discrete random walk, the Brownian limit, and Fick’s laws. Then, anomalous diffusion regimes are discussed, with emphasis on continuous-time random walks (CTRW), alpha-stable distributions, and fractional diffusion equations. This framework is applied to the construction of synthetic anomaly models in reservoirs, aiming to represent different transport regimes associated with heterogeneity, anisotropy, and trapping in porous media. In the second part, trajectories of D. quadriceps, a solitary foraging ant species, are studied. Geometric simulations indicate that sectorial exploration of space can increase the average search efficiency compared with isotropic walkers, within the range of resolutions analyzed. Finally, increment series extracted from the observed trajectories are analyzed using autocorrelation and partial autocorrelation functions, with confidence intervals estimated by bootstrap. The results suggest weak antipersistence at some lags, indicating short-range temporal dependence, although no unique ARMA model can be identified from the available data.