Abstract
<title>Abstract</title> <p>Characterizing heterogeneity in carbonate reservoirs remains a persistent challenge because conventional statistical indicators often lack a direct geological interpretation. This study develops an integrated multiscale framework that links petrographic observations, electrofacies, four rock-typing schemes (Lucia, Lorenz, Winland, FZI) and a suite of statistical and geostatistical indicators (coefficient of variation (CV), Dykstra–Parsons coefficient (VDp), Standard Error (SE), Range of the variogram (RIV), semivariance) integrated with wireline log analysis (Velocity Deviation Log (VDL) and effective porosity), and with particular attention to anhydrite cementation as a key diagenetic control, to validate reservoir connectivity in a Persian Gulf carbonate reservoir. A fundamental decoupling between statistical and spatial heterogeneity is established: the most productive reservoir unit is spatially the most heterogeneous yet statistically intermediate, whereas a statistically highly heterogeneous unit is spatially homogeneous and non-reservoir. Log-derived PHIT–PHIE (Total Porosity-Effective porosity) separation and VDL acoustic patterns refine this assessment, distinguishing effective, connected porosity from isolated and non-reservoir features. The CV emerges as the most robust single indicator across four classification methods, correctly distinguishing outlier-driven from systematic heterogeneity; the Lorenz method is the sole exception, where SE is preferred. The RIV and semivariance add a critical spatial dimension, showing that low CV alone does not ensure spatial continuity. In contrast, high heterogeneity combined with high RIV denotes connected, flow-active heterogeneity, which is a distinctly positive attribute. Three fundamental patterns characterize the heterogeneity architecture: systematic heterogeneity (minimal PHIT-PHIE gap) defines the best reservoir targets; outlier-driven heterogeneity, primarily caused by anhydrite cementation and marked by a wide PHIT-PHIE gap, characterises non-reservoir units; and true homogeneity offers predictable flow behaviour. The workflow is transferable to multi-well studies, and the one-dimensional heterogeneity model provides a high-resolution vertical template for upscaling. These findings demonstrate that heterogeneity in carbonate reservoirs is not a monolithic property but a spectrum of geological realities, requiring a multi-indicator, petrophysically and petrographically-grounded assessment for reliable reservoir modelling.</p>