PIK2 is a coastal mega-development in Indonesia, converting thousands of hectares of cropland and tree cover into built environment. Satellite data from 2017–2024 records the transformation. The question is not whether enclosure happens — it visibly does — but whether mathematical tools from statistical mechanics and information geometry can quantify its structure with precision that narrative description cannot.
The answer is yes, and the precision reveals structure. The transformation pulse — the rate of landscape change — peaked at 0.405 radians per year during 2019–2020, a measurement of phase-transition-like behavior in land use. Expected absorption times: 46 years for cropland, 38.1 years for tree cover to be converted to built area. The built environment self-retains at 96.4% — once converted, land stays converted. This is an absorbing state in the Markov chain sense.
Note (2026-11-13; read against the paper's full text): the essay presents the absorbing state as a finding that once converted, land stays converted. The paper says its absorbing-chain formulation imposes perfect irreversibility on the built area, which it calls an idealization, because the empirical leakage is 2 to 4 percent (which it reads as construction churn); the 96.4 percent is a pooled self-retention rate. The absorption times of 46.0 years for cropland and 38.1 years for tree cover are what the chain implies if the current transition dynamics persisted, from eight years of Sentinel-2 data, while the paper itself describes the enclosure as contested and non-stationary. It also lists that the chain treats pixels independently, that the land-cover product may misclassify, and that the analysis is confined to the PIK2 area. The essay's 'phase-transition-like' wording is not the paper's: it describes the 0.405 radians per year as a Fisher-Rao velocity pulse that coincides with major construction.
The percolation analysis is the sharpest finding. The giant connected component of built area persists at occupation probabilities far below the random percolation threshold. Random development at the same density would fragment into disconnected patches; planned development forms a connected whole. The connectivity is evidence of coordination — spatial growth that follows a plan, not stochastic accumulation.
The fractal dimension of urban boundaries increases over time, meaning the development's edge becomes more complex as it expands. The boundary is not a smooth frontier advancing uniformly but an irregular surface with increasing roughness — fingers of development reaching into unconverted land.
Information geometry measures the shape of change, not just the fact of it. The tools make visible what satellite images show but don't quantify: the rate, the irreversibility, the coordination, and the increasing complexity of the enclosure process.