Presented at the Global Ocean Salinity and the Water Cycle WorkshopIn recent decades, noticeable decadal salinity trends have emerged at the ocean surface, reflecting the changes in the ocean water cycle, strength of the ocean circulation, and changes in the ocean transports of salt. A number of potential causes for the observed changes have been suggested, including those related to anthropogenic aerosol forcing as well as natural climate fluctuations. Here we explore how the recent decadal salinity trends fit into the long-term picture, focusing on those ocean regions where the decadal changes oppose the long-term trends. For such regions, we estimate the likelihood of âunusualâ decadal salinity changes in the context of the historical record, contemporary estimates, and future projections. The analysis is based on a large suite of salinity estimates, including data from the modern satellite salinity observations missions, multi-decadal historical in situ observations, ocean model/data synthesis, and centennial realizations from global climate models. We test a hypothesis that under the influence of internal variability the likelihood of strong salinification (or freshening, depending on a region) is plausible in the presence of an opposing secular trend due to anthropogenic forcing. Computed probabilities that are close to zero indicate that the internal variability is negligible compared to the background trend. Alternatively, if the internal variability is large and the background trend is negligible, the chance for a decadal salinity trend of an opposite sign is random, with probability around 0.5. For example, we find that the chance of occurrence of the strong decadal salinification recently observed in the subpolar North Atlantic, despite an increase in freshwater supply and long-term freshening, is about 0.1 â a plausible, although rare, event. Finally, contrasting changes in salinity with those in freshwater fluxes, our findings suggest that the time of emergence for anthropogenic trends in the ocean water cycle might be substantially earlier than that for surface salinity. In particular, for the ocean water cycle twenty years might be sufficient for the anthropogenic signals to rise above the ânoiseâ level of natural variability.