Phantom Structure in Networked Aerospace Contact Systems: Typed Contacts for Certification Under Correlated Failures (deposit v0.2)

Certification of scheduled networked aerospace systems often reduces a contact plan to two reported quantities: structural reachability and routing efficiency, combined as a delivery ratio. This decomposition can misattribute structural loss to routing inefficiency when scheduled contacts share physical failure domains. The mechanism is phantom structure: temporal reachability that exists in the discrete-event contact oracle but is intermittently destroyed by correlated link failure. The empirical design separates three evidentiary tiers. A frozen Canberra Deep Space Communication Complex snapshot provides a reproducible worked example: 372 MRO/Odyssey contacts over 104.3 days, with 3.40–6.38 percent phantom structure in four selected windows at rho_RF = 0.40. A separate manifest-pinned rolling CDSCC archive provides the station-resolved pilot: 378 MRO/Odyssey contacts, 98 eligible daily window/source blocks, median phantom fraction 2.53 percent, and bootstrap mean 3.00 percent with 95 percent interval [2.61, 3.43] percent. The 2010–2018 FOIA schedule provides scale and cadence evidence only: 149,482 rows, including 24,228 MRO/Odyssey rows, but no DSS, antenna, complex, or equipment fields. Gateway and BPv7 results are reported by measurement definition rather than as context-free constants. The Gateway example is a shared-domain collapse-risk calculation. The BPv7 Certification/Diagnosis/Timing/Reproducibility Quad is reported by layer: 817 untagged payload bytes, 871 tagged C/D/T/R bytes, 889 protocol wrapper bytes, and 905–925 RFC9171 bytes depending on CRC policy.