Reconstructing missing block-level morphological evidence to support Urban renewal-unit delineation

Built-environment assessment for urban renewal often relies on zone-based categories, yet categories such as local climate zones are too coarse to read renewal-unit boundaries at the street-block scale. When block-level morphology records are incomplete, planners lose evidence on continuity, transition, and local distinctiveness. Using a controlled benchmark of 4,364 urban blocks in Nanjing, this study tests whether missing block-level feature values can be reconstructed into proxy evidence, defined here as reconstructed morphology that remains usable as supporting evidence rather than direct delineation output. Structural continuity refers to the retention of adjacency-based pattern and transition structure, whereas local morphological character refers to the preservation of block-scale combinations of density, openness, greenness, and form. A morphology-aware reconstruction benchmark compares alternative approaches under multiple masking scenarios to evaluate both local error recovery and planning-relevant structural preservation. Within this single-city controlled setting, Morpho-GNN provides the strongest overall proxy-evidence profile among the evaluated methods by preserving block pattern and neighbourhood coherence while keeping transition-sensitive structure interpretable. The findings suggest that morphology-aware reconstruction can provide planning-relevant supporting evidence for renewal-unit delineation under incomplete records, while remaining distinct from a validated operational delineation workflow.