We have predicted the flow permeability and its spatial distribution for the Longmaxi shale gas reservoir using microseismicity induced during hydraulic-fracturing stimulation. In the time-of-occurrence versus distance-from-injector plot, we find that microseismic points exhibit a parabolic envelope, which we interpret as a triggering front. This reveals that fluid pressure diffusion is at least one of the underlying mechanisms of microseismicity generation. We derive the large-scale equivalent diffusivity from the triggering front plot and thereafter obtain a 3D diffusivity map of the heterogeneous reservoir by solving an eikonal-like equation suggested previously. During this process, we apply kriging interpolation to increase the density of sparsely distributed microseismic points. The resulting diffusivity ranges between 1.0 and [Formula: see text] with the peak probability attained at [Formula: see text], which is consistent with the estimate we obtain from the triggering front analysis. We transform the diffusivity map into a permeability map using three different theories of fluid pressure diffusion in porous media: the seismicity-based reservoir characterization method (SBRC) based on Biot’s theory of poroelasticity, the quasirigid medium approximation (QRMA), and the deformable medium approximation (DMA) based on the de la Cruz-Spanos theory. The permeability according to QRMA is slightly higher than that from SBRC, yet we observe no significant difference. However, these estimates are by one order of magnitude higher compared with the permeability estimate from DMA. Furthermore, the permeability from all three theories is much higher than that from previously reported core sample measurements. We interpret this as the difference between large-scale equivalent and matrix permeability and therefore lend weight to the hypothesis that there exist highly conducting fluid pathways, such as natural fractures.

Evaporation of water from porous media is essential for a large variety of applications, whereas the opacity of porous matrix imposes considerable challenges in unveiling complicated phase‐change phenomena. Air invasion was previously reported as the major desaturation mechanism, while cavitation in porous media is not well studied. Herein we characterize the transient distribution and evaporation of water in homogeneous tight porous media with nondestructive nuclear magnetic resonance (NMR) and magnetic resonance imaging (MRI). By monitoring the amplitude change of NMR transverse relaxation time T2, we investigate the dynamical pore filling status and water content during evaporation. Interestingly, we find that the T2 spectrum shifts immediately after the evaporation starts, indicating the emptying of big pores from the entire medium. Disconnected void clusters at different depths in the porous medium are also observed from MRI scanning and optical images. These observations indicate the emergence of cavitation across the entire porous media along with the evaporation from open surface. Cavitation occurs when the water is stretched to metastable state by large capillary pressure from the evaporating meniscus. By studying the evaporation from hydrophilic membrane‐separated porous media, we further demonstrate the existence of cavitation‐associated evaporation. The preferential water vaporization from the bottom part can still be found from T2 spectrum analysis and optical imaging even when the water‐permeable membrane cuts off possible air invasion. Our findings confirm cavitation‐associated evaporation is one of the primary mechanisms for tight porous media, which provides valuable guidance for evaporation and moisture control.