A change in seasonal flooding cycles in the Amazon may negatively impact nesting success of the Giant South American Turtle (Podocnemis expansa). Our aim was to devise a technique that could be replicated in the entire Amazon basin, for monitoring alterations in fluvial cycles and their effects on turtle nest mortality. We mapped the spatial distribution and height of P. expansa nests and tested the effects of different inundation scenarios within the Trombetas River Biological Reserve, Para state, Brazil. We also used historical data on water level and hatchling production to test whether the sharp decline in the Trombetas River P. expansa population over the past thirty years was related to detected changes in the flood pulse. Our models indicate that an increase of 1.5m in the water level is sufficient to decrease the time of exposure to less than the minimum required for incubation and hatching (55 days above the water) in 50% of the nesting area. This model explains the low hatchling production in dry seasons when the total nesting site exposure was less than 200 days. Since 1971, there was an average decline of 15 days per decade in sandbank exposure during the nesting season (a total of 62 days from 1971 to 2015). However, the decrease in sandbank exposure was not significantly correlated with the sharp decline in hatchling production. Changes to the water cycle in combination with the main sources of decline (overharvest, construction of dams, and dredging of riverbeds) might have an accumulative effect on P. expansa populations.