Abstract
Aligning regional innovation capability with territorial space use efficiency (TSUE) is important for regional sustainable development. Using data from 2010 to 2020, this study examines their spatiotemporal patterns and matching-type transitions in the Pearl River Delta urban agglomeration through entropy-weighted TOPSIS, a super-efficiency slacks-based measure (SBM) model incorporating undesirable outputs, and four-quadrant analysis. The results show that: (1) Both innovation capability and TSUE increased overall, but innovation disparities widened while efficiency disparities narrowed. (2) The composition of matching types shifted toward a predominance of high–high and innovation-lagging units. A core–periphery pattern was evident, with high–high units concentrated along the Pearl River Estuary and innovation-lagging units widespread in peripheral areas. (3) Matching types exhibited substantial persistence, alongside prominent transitions from low–low to innovation-lagging and from efficiency-lagging to high–high. High–high and innovation-lagging units showed particularly strong persistence, although some reversals also occurred. These findings support differentiated and locally adapted spatial governance: high–high areas should enhance spatial quality and innovation diffusion; innovation-lagging areas should strengthen innovation capability; efficiency-lagging areas should prioritize site-specific industrial-land regeneration and functional restructuring; and low–low areas should improve infrastructure and industrial conditions while gradually cultivating innovation capacity.