Comparative pore water-sediment studies are crucial to trace biogeochemical and early diagenetic interactions between mineral and fluid phases. They are commonly conducted in environments either stable throughout longer timespans or with high sediment production rates, which create a depositional record with long time diagenetic alterations between the sediment and contemporaneous pore water. In contrast, shallow evaporative ponds with varying hydro chemistry, create a record with shorter reaction pathways and, thus, shorter duration of diagenetic processes.This work presents a pore water-sediment study within the Cinq Cases pool system, a shallow and saline water body with occasional marine influx, located on the Aldabra atoll in the western Indian Ocean. Sediments span ca. 3800 years, with three environmental stages: (i) an initial palustrine environment (Unit III), (ii) a slow marine flooding, with cyanobacteria and sponge blooms (Unit II), and (iii) lagoon flooding, including oxic conditions within the sediment (Unit I). The pore water shows a salinity increase with depth, inconsistent with the palustrine facies and anoxic bacterial community of the deepest lithological Unit (III). Low magnesium calcite cements in the upper part of Unit III, indicate ancient meteoric diagenesis, while empty cell envelopes of coccoid cyanobacteria in Unit II point to an ancient bloom, not represented in the 16S rRNA analysis. Unit I shows a gradual change from a marine-lagoonal to hypersaline environment, indicated by a shift from a foraminifera to an ostracod dominated environment. Thus, carbonate deposits of small water bodies are subject to frequent changes in diagenetic environments, reflected by three different proxies as partially overlapping but different time scales: (i) sediments reflect ancient processes (ii) pore waters are influenced by recent processes, and (iii) bacterial communities reflect an overlay between ancient and recent processes. To precisely distinguish between these time scales, frequently repeated and spatially variable measurements are necessary.