Resolve the North Carolina transport–power ordering under unattributed outages

Determine whether transport outages truly exceeded power outages in North Carolina during Hurricane Helene after resolving the unattributed residual, given that assigning more than 66 percent of the residual to power would reverse the observed ordering.

Background

The FCC Disaster Information Reporting System cause columns do not form a strict partition of reported cell-site outages. In North Carolina, 6.9 percent of site-days are unattributed overall, and the paper reports transport at 52.2 percent of attributed site-days versus power at 47.3 percent.

The authors show that the damage conclusion is robust under an adversarial allocation of the residual, but the transport–power ordering is not: assigning more than 66 percent of the residual to power would make power the larger cause. They report no evidence for such a skew, but state that it cannot be excluded and therefore characterize the North Carolina ordering as unresolved.

References

None of this excludes a power-skewed residual in North Carolina; it establishes that we see no sign of one.

— The Towers Were Standing: A Cause Decomposition of Cellular Outages During Hurricane Helene  (2609.10944 - Olukola et al., 10 Sep 2026) in Section 4.1, subsection “What the unattributed residual can and cannot do to this”; Section 7.1, “Limitations of the source”