The Aswan reservoir seismicity is accepted as an example of reservoir-induced seismicity with the M 5.4 event of 1981, which occurred 15 years after the reservoir impoundment started. During the 1982-2001 period, the Aswan seismicity separates into shallow and deep seismic zones, between 0 and 15 and 15 and 30 km, respectively. These two seismic zones behave differently over time, as indicated by the seismicity rate, depth migration, b-value, and spatial clustering. For the deep events, the rate decreases with time, the depth remains at a constant or increasing depth, and the b-value is lower than for the shallow events, which migrate toward the surface with a close to constant seismicity rate. Of all the previous parameters, only the seismicity rate correlates with seasonal variations of the lake level. This is positive evidence for the Aswan reservoir seismicity to be reservoir triggered in the 1982-2001 period. Clustering over time and space expresses that numerous aftershock sequences are activated, that is, a minimum of 35% of the seismicity, similarly to tectonic seismicity. The observed decrease of the b-value with depth is also a property suggested for natural events. Our results suggest that the Aswan seismicity emerges both from the water-level loading and the interplay between induced earthquakes themselves through numerous standard aftershock sequences. This latter process induces stochastic fluctuations in the seismicity patterns that inhibit a recovery of a simple seismic response to water-level changes. Even though aftershock sequences dominate the temporal fluctuations of seismicity, this does not negate the importance of influence of the reservoir in sustaining this unusual long-lasting seismic episode. During the 1982-2001 period, the Aswan seismicity appears as a power-law relaxation response to the 1981 M 5.4 shock, this response being decorated by long-term (larger than 100 days) water-level changes.