It is recognised that some proportion of South American large-scale topography has been generated by convection within the Earth's mantle. Yet, spatial and temporal patterns of dynamic topography remain poorly understood. Variation of present-day dynamic topography can be mapped in the oceans by calculating residual depths with respect to the well-known age-depth relationship. Along the margins abutting South America, anomalies with amplitudes of $\pm1$ km and wavelengths of $\sim10^{3}$ km are observed. Onshore, dynamic topography is investigated by analysing a range of disparate datasets. Positive long-wavelength free-air gravity anomalies and slow shear-wave mantle velocities correlate with high plateaux of the Borborema Province and the Central Andean Altiplano. Admittance analyses of these regions are used to gauge dynamic support. Admittance of $ > 20$ mGal km$^{-1}$ at wavelengths $ > $ 500 km suggests partial dynamic support. In this context, inverse modelling of longitudinal river profiles is applied to retrieve a continental-scale uplift history. Erosional parameters are calibrated against an independently derived uplift history of the Borborema Province that reveals uplift in the last 30 Ma. Results suggest that the bulk of South American regional topography grew during Cenozoic times. In the Central Andean Altiplano and Southern Patagonia, most uplift occurred in the last 20 Ma. In both areas, widespread Cenozoic magmatism suggests that youthful uplift might be related to asthenospheric upwelling. Uplift histories are used to predict sediment flux to the Amazon Fan, which reveals that onset of the delta is a direct consequence of intensified Andean uplift. Analysis of the Parna\'iba cratonic basin of northeast Brazil is carried out to evaluate long-term evolution of vertical motions and to understand the mechanisms driving this basin's subsidence. Joint interpretation of a deep seismic reflection profile that traverses the basin and receiver function analyses reveal a 3 km thick basin underlain by three crustal blocks. Moho depths of 38--43 km are observed beneath the Amazon craton west of the basin, whereas depths of 35 km are found underneath the Borborema Province to the east. The Moho is located at 38--42 km depth beneath the basin. Stratigraphic architecture from shallow seismic reflection profiles reveals undisturbed deposition between Palaeozoic and Mesozoic times. Rift-type structures are locally imaged. Subsidence analysis reveals thermally-driven subsidence with thermal time constants of $\sim$ 70--80 Ma. Assessment of crustal thickness variations indicates that minimal extension of up to 80 km, with small stretching factors (up to 1.15), is plausible beneath Parna\'iba. One- and two-dimensional strain rate histories suggest that pre-Silurian rifting followed by thermal subsidence is possible if a minimum of 1 km of syn-rift deposition occurred. Basin-wide erosional unconformities are observed throughout the sedimentary section and correlate with departures from long-term subsidence trends. These steps are interpreted as transient uplift events that led to development of ephemeral landscapes, suggesting that dynamic topography could have played a role in the evolution of this Phanerozoic basin.

[Truncated abstract] The southwest of Western Australia is an intraplate area classified as a stable continental region. It comprises predominantly Archaean and Proterozoic geology and has generally subdued topography. The region currently experiences significant seismicity in the Southwest Seismic Zone (SWSZ), which is one of the most seismically active areas in Australia and is thought to represent the highest seismic hazard of the region. In recent years, numerous scarps, potentially related to large palaeoearthquakes have been recognised not only within the SWSZ, but also in a broader region of the southwest of Australia. Palaeoseismological investigations of two of these scarps, the Dumbleyung and the Lort River scarps, confirm their association with surface-rupturing palaeoearthquakes and indicate events with likely maximum magnitudes of ~Mw 7.0 on faults of low to medium slip rates. Two trenches across the Dumbleyung Fault scarp revealed a thrust fault in alluvial sediments with two associated earthquakes in the last ca 24-60 ka. A possible Holocene age was recognised for the last recorded earthquake event exposed in these trenches. Two trenches across the Lort River Scarp show that this feature results from thrust faulting in the weathered gneissic country rock. These trenches exposed evidence of two events in the last ca 35 ka, with a likely late Pleistocene age for the last earthquake. On both sites, the earthquakes are interpreted as associated with the last phase of fault activity, which was likely been preceded by a long period of quiescence. Assessment of the earthquake hazard associated with large earthquakes at the Dumbleyung and Lort River Faults resulted in calculated peak ground accelerations of up to 2 g in the near-fault fields. Such earthquakes would significantly affect nearby towns such as Dumbleyung, Wagin, Katanning, and Esperance, but they are unlikely to cause any significant damage in Perth. The palaeoseismological investigations show that the earthquake activity in the southwest of Western Australia is not only confined to the SWSZ, as it has been considered in previous assessments of the seismic hazard, but that there is also potential for strong earthquakes across much of the region. The seismicity in the southwest of Western Australia appears to be transient and migratory. This is suggested by the lack of local relief associated with places of current seismicity and fault scarps, the widespread distribution of the fault scarps across the region, the increase in seismicity in the SWSZ following strong recent events, and the apparent long periods of earthquake recurrence at fault sites. Accordingly, the current seismicity in the SWSZ is inferred to be transient and probably associated with stress changes produced by the recent earthquakes. '...' This uplift could be associated with dynamic topography effects resulting from processes along the plate margins. The uplift is probably enhanced by a flexural response of the lithosphere to local differential loads and density contrast along the southern margin, a mechanism that may also help explain the occurrence of some earthquake activity. The results from this study, complemented by additional palaeoseismological studies must be included in future probabilistic assessments of the seismic hazard of the southwest of Western Australia.