The transition from flat to steep subduction in south Peru and its impact on seismicity and deformation

South Peru subduction is marked by a transition between flat slab, where the Nazca ridge enters into subduction, to dipping slab further South. Using a dense seismo geodetic network installed in the area together with Sentinel InSAR time series, we monitor in great details the seismic structure, the seismicity and the deformation in the area.The subduction of the Nazca ridge is associated with low interseismic locking as well as seismic swarms and repeaters on the interface likely indicative of the occurrence of shallow slow slip events. There, the overriding plate is characterized by a wide zone of deformation as shown by InSAR data and by crustal seismicity. The flat slab also exibits an intense intraslab seismicity that shows an intriguing correlation with the vertical surface deformation.Further south, the slab is dipping steeply and exhibits much less seismicity, maybe due to long lasting post-seismic relaxation following the 2001 Mw8.4 Arequipa earthquake and to high interseismic locking on the interface. Crustal seismicity is more localised: along the volcanic arc and associated tectonic structures, and along faults systems in the forearc.At the transition between flat to dipping slab, regular Mw~7+ earthquakes occur every ~5 years. The last one occurred in June 2024 and has been captured by our seismo-geodetic deployment. This Mw7.2 earthquake was preceded by a series of foreshocks, and followed by numerous aftershocks, both of which exhibit an intriguing extent down to 80km depth within the slab, likely guided by subducted oceanic structures along the edge of the Nazca ridge that mark the transition from flat to dipping slab. Our observations image the transition from flat to dipping slab in South Peru and its impact on the seismicity features, and on the upper plate deformation. We notably show that the subduction of the Nazca ridge cannot sustain the flat slab alone. Full waveform tomographic images instead show that the oceanic lithosphere is anomalously thin with asthenosphere upwelling, suggesting that it is thermally eroded by Easter hot spot, that contributes to the buoyancy of the flat segment. We also see that the flat slab has likely contributed to the delamination of the continental lithosphere that is almost absent, implying a significant viscous coupling between the slab and the overriding plate. The enhanced landward motion above the flat slab, seen by InSAR and GNSS, could be due to a viscous drag of the continental plate by the flat slab. Finally, the surface uplift imaged by InSAR can only partly be explained by a viscoelatic subduction model, including interseismic coupling on the interface and an elastic cold nose. Far inland, at about 250km from the trench, a secondary uplift zone correlates remarkably well with intraslab seismicity. We suggest that these intriguing features could be explained by the bending – unbending of the slab.