Validating a Terrain-Derived Geodiversity Index Against Inventoried Geosites: A Lacustrine Blind Spot at Lake Matano, Indonesia

Terrain-derived geodiversity indices are increasingly used to screen aspiring geoparks, yet rarely tested statistically against independently inventoried geosites, and, to our knowledge, never in a lake-centred landscape where mining concessions and geoheritage overlap. This study built and tested such an index for the four districts surrounding Lake Matano, Southeast Asia's deepest lake and core of the proposed Matano and Malili Lake System geopark in Indonesia, where eight nickel mining permits intersect the study area. A normalised absolute Topographic Position Index and a neighbourhood ruggedness measure, derived from the 30 m ALOS World 3D model, were combined equally into a composite index (range 0-0.814). Of 46 inventoried tourism objects, 15 with explicit geological or geomorphological value were retained as geosites and compared with 46 random points using 150 m buffer maxima and one-sided Mann-Whitney tests. Geosites scored higher on average (0.174 vs 0.145) but not significantly (p = 0.290). Three geosites and, confirmed by overlay with the lake polygons, three random points fell on open water, returning zero by construction; excluding these six, 12 terrestrial geosites still outscored 43 random points on land (0.217 vs 0.155), approaching but not reaching significance (p = 0.067, r = 0.20). Global and local Moran's I showed no robust spatial clustering among inventoried values. The index thus ranks terrestrial landforms weakly in the expected direction and is structurally blind to lacustrine geoheritage, the very feature on which a lake-centred geopark's candidacy rests; such indices need supplementing with bathymetry, shoreline geometry or a separate lacustrine rubric. Keywords: geodiversity index; geomorphometry; geosite validation; lacustrine geoheritage; aspiring geopark; nickel mining