Spatial-multivariate modelling of seagrass ecological quality index (SEQI) in Central to Eastern Indonesia

Indonesia is home to 660,156 ha of seagrass meadows, with 90% of these meadows located in the central to eastern zone, highlighting the importance of this area as a blue carbon reservoir in the country. Despite this extensive distribution, the ecological quality and spatial organization of seagrass ecosystems remain unevenly documented. We quantified the Seagrass Ecological Quality Index (SEQI) from five indicators: seagrass species richness (St), seagrass cover (Ct), water clarity (Wt), macroalgae cover (Mt), and epiphyte cover (Et), and combined spatial statistics with multivariate analyses to identify ecological gradients and local clusters. Spatial dependence was tested per region using Global Moran's I with adaptive neighborhood weights, followed by Local Moran's I (LISA) and Getis-Ord Gi* to map significant hotspots, coldspots, and outliers. Multivariate structure was evaluated with K-means clustering, PCA, and NMDS, and cluster robustness was assessed via silhouette width and permutation testing. Global Moran's I detected significant SEQI autocorrelation only in the central zone (I = 0.2632, p = 0.023), where LISA confirmed a clear local mosaic of High-High hotspots, Low-Low coldspots, and High-Low/Low-High outliers. K-means produced four clusters with weak global separation (silhouette = 0.158) but significant non-random structure (p = 0.0005), including one highly stable and distinctive cluster (silhouette = 0.773). PCA and NMDS indicated that SEQI variation is dominated by the seagrass species richness, seagrass cover, and water quality gradient. In contrast, the epiphyte cover represents an orthogonal stress axis, and the macroalga cover contributes minimally. Overall, the seagrass landscape in these zones is characterized by overlapping ecological gradients, with localized, statistically significant clustering restricted to the central zone. This highlights the hotspot areas for conservation, while the coldspot and outlier areas that degraded require targeted restoration to mitigate carbon-loss vulnerability.