Spatial distribution and covariate associations of a reconstructed urban wetland-park inventory in Harbin

Incomplete wetland-park inventories can limit spatial assessments of urban blue–green infrastructure. We analyzed a reconstructed inventory of 14 wetland-park records in Harbin, China, represented by analytical points and reported areas. Two nested subsets comprising 8 and 13 records were compared, with the additional record in the full inventory used to assess the sensitivity of point-pattern summaries. Standard deviational ellipses characterized directional structure, nearest-neighbor analysis assessed departures from a specified complete-spatial-randomness reference, and GeoDetector examined spatial correspondence with river density, slope, road density, and population density. The primary subsets showed broadly east–west orientations and an 8.26-km difference between their mean centres. None of the nearest-neighbor tests indicated a statistically significant departure from the specified reference. All single-factor GeoDetector q values were below 0.004, indicating weak measured correspondence under the sparse binary response and selected spatial specifications, rather than an absence of underlying associations. The findings provide a descriptive account of the reconstructed records, not evidence of verified municipal park development, causal spatial drivers, or policy effects.