Granular and spatially differentiated density thresholds for No Net Land Take: A scenario-based cellular automata analysis in French territories

The "No Net Land Take" (NNLT) objective – halting land artificialization by 2050 – poses a fundamental planning challenge: how to accommodate housing demand without new land take. This paper addresses two issues. First: by demonstrating that achieving Net No Land Take (NNLT) requires moving beyond uniform average density targets toward spatially differentiated density thresholds adapted to local urban contexts. Second: by establishing that spatial computing methodologies, particularly high-resolution cellular automata modeling, provide an appropriate framework for defining and testing such granular density thresholds under NNLT constraints. To investigate these issues, we employ the LucSim high-resolution cellular automata model to simulate five density scenarios up to 2050 across two contrasting French territories: the demographically dynamic peri-metropolitan Grand Rovaltain (GR) and the ecologically sensitive, tourism-driven Aire Gapençaise (AG). Simulations are calibrated with official 2050 housing targets from their respective Territorial Coherence Schemes (SCoT) and translated into six settlement entity‑specific density thresholds to test whether granular spatialization matters. Methodologically, cellular automata prove appropriate for this task because they spatially allocate densification under policy constraints while preserving local morphological realism. Our results show that higher-density scenarios can substantially limit land artificialization and protect high-quality soils. However, the effectiveness of densification strategies strongly depends on regional socio-economic and biophysical contexts. This study provides a quantitative, scenario-based framework for policymakers, emphasizing that context-sensitive densification is essential for reconciling urban development with environmental sustainability.