Linking the evolution of the surface area (as quantified, e.g., through its spatial roughness) of minerals to their dissolution rate is a key aspect of mineral reactivity. Unraveling the nature of their main features requires relying on approaches yielding a quantitative characterization of the temporal evolution of surface topography/roughness. Here, a mechanically polished {104} calcite surface was dissolved at room temperature and at close-to-equilibrium conditions (Ω = 0.6) with an alkaline solution (pH = 8) across a temporal window of 8 days. Surface topography images were acquired daily using vertical scanning interferometry, the ensuing topography data being then embedded within a statistical analysis framework aimed at describing comprehensively the surface roughness evolution. The strongest system variations were observed after 1 day: the probability density function of surface roughness was observed to transition from being approximately Gaussian to being left-skewed and leptokurtic, exhibiting a dramatic increase in the variance and a significant change in the semi-variogram structure. After a relaxation time of approximately 2 days, the reacting surface appeared to attain a steady-state configuration, being characterized by the values of the statistical moments characterizing surface roughness that become virtually independent of time. Attempting to unravel the underlying dissolution mechanism, an original numerical model able to reproduce satisfactorily the statistical behavior observed experimentally was developed and tested. Our results suggest that under the investigated conditions, dissolution may be characterized as a spatially correlated random process, with the areas most exposed to the flowing fluid being prone to preferential dissolution. The numerical model was also used to obtain insights into the influences of the initial surface roughness and of the fluid composition on the steady-state statistical characterization of the surface roughness. Our results suggest that the influence of the initial surface roughness is limited. The present study suggests that potential empirical relations linking the surface roughness of the reacted crystals to the saturation state at which they dissolved may be developed, which would allow to back-estimate the reacting conditions only based on topography data.