Spatial Dynamics of Digital–Intelligent Integration and Carbon Productivity: Land-Use Linkages in the Middle Yellow River

Regional averages may improve even when individual cities remain within the same coordination state. Using panel data for 40 cities in the middle reaches of the Yellow River from 2011 to 2022, this study distinguishes continuous changes in the corrected coordination degree D from discrete state transitions and further examines spatial clustering, neighborhood-conditioned mobility, socioeconomic heterogeneity, and land-use landscape associations. Under the baseline specification, mean D increased from 0.3074 in 2011 to 0.3790 in 2022. Among 440 adjacent-year transitions, 85.23% remained in the same state and 99.77% occurred within the same or a neighboring state. State persistence, however, did not imply an absence of progress. Among same-state transitions, 62.4% still recorded an increase in D, and the next-year upgrading probability was markedly higher for observations located closer to the next state threshold, reaching 29.9% in the closest distance quartile but only about 1% in each of the two most distant quartiles. Spatial mapping and Getis–Ord Gi* identify localized clusters of high and low coordination values, while spatial Markov results provide no consistent evidence that transition patterns depend on neighborhood conditions across alternative spatial specifications. Geographically and temporally weighted regression (GTWR) reveals substantial spatial and temporal variation in socioeconomic associations, although residual spatial autocorrelation remains in 2013 and 2019. Landscape Shape Index and Patch Density vary much more across cities than within cities over time, and their fixed-effects associations with D are not reproduced under first differencing. Taken together, the results indicate that coordination in the study area is characterized more by gradual within-state adjustment than by frequent movement across state boundaries. They also show that spatial clustering, local socioeconomic associations, and landscape structure capture different dimensions of this process and should not be interpreted as direct evidence of spillover or causal mechanisms.