After withdrawing its settlements and ground forces in 2005, Israel exercised con-
trol over Gaza by means of aerial occupation. Despite the magnitude and exceptional
duration of Gaza’s sustained aerial bombardment, some 20 years and counting, there
is limited understanding of the spatiality of violence across episodes, or its links to the
current episode. The puzzle in this respect, concerns why and in what ways violence was
discriminate with respect to targets in select episodes but not in others? Our analysis
combines geo-coded data on aerial attacks from all major military operations in Gaza,
including 2008-2009, 2014, and 2023-2025 with data on casualties and socio-economic
indicators. We find that aerial violence intensified in scale, duration, lethality, and de-
struction, while simultaneously becoming more indiscriminate over time. Across episodes,
we identify two main targeting biases. In earlier episodes, a socio-economic targeting bias
with selective casualties clustering in locales characterized by higher wealth and educa-
tion. In later episodes, we identify the targeting of civilian infrastructure across Gaza,
particularly with respect to health facilities. Additionally, we find that bombardment
varies as a function of topography and urban morphology, with sites located at higher
altitudes or in more densely built-up areas being targeted more frequently. We argue
that the inability to inflict serious damage to the capacity of armed groups in Gaza
resulted in the abandonment of selectivity, in violence that was less contained and far
from discriminate, with the logic of aerial occupation—control, containment, periodic
killing——effectively transformed into full-out ethnic cleansing.

Read the full article:

Currently Under Review.