Geospatial assessment of seaweed cultivation suitability using ocean color data and multi-criteria decision analysis in the coastal waters of Lombok Island, Indonesia

The rapid expansion of Kappaphycus alvarezii cultivation in Indonesia poses significant sustainability challenges due to ad hoc site selection, which has contributed to disease outbreaks and exceeded ecological carrying capacities. This study employs ocean color remote sensing and Multi-Criteria Decision Analysis (MCDA) to develop a knowledge-driven geospatial model for identifying suitable habitat zones on Lombok Island. By integrating the Analytical Hierarchy Process (AHP) and Weighted Linear Combination (WLC), a biophysical hierarchy is established that prioritizes physiological survival over theoretical yield maximization. To enhance model robustness in the absence of independent yield data, a Map Removal Sensitivity Analysis (MRSA) was implemented to quantify the relative influence and spatial sensitivity of key environmental parameters. The results indicate sea surface temperature (21.46%) and salinity (15.33%) as the most influential variables, supporting a “survival-first” paradigm for mitigating ice-ice disease risks. The geospatial analysis reveals a spatial pattern of high suitability consistent with an “environmentally balanced” zone, characterized by depths of approximately 2–30 m and current velocities of 0.2–0.4 m/s. Sensitivity results further show that while SST exerts a dominant but spatially uniform control, salinity and dissolved oxygen drive localized variability, identifying regional-scale sensitivity gradients whose ecological interpretation is bounded by the resolution of the underlying data. Quantitative spatial validation against 275 georeferenced farm GPS locations shows that 96.7% of farms are concentrated in the southern high-suitability zone, against an area-based null expectation of 26.0% for that cluster (chi-square = 716.7, df = 2, p < 0.001). A per-suitability-class overlay further confirms that 98.9% of farms fall within High-suitability pixels and none within Low-suitability pixels against an area-based null of 66.9% (chi-square p < 0.001); Cohen’s κ against a random background is modest (0.08), reflecting the abundance of high-suitability habitat relative to the small number of realized farms. These findings provide a practical and transferable framework for supporting Allocated Zones for Aquaculture (AZAs) and identifying growing-season and risk-informed aquaculture sites in data-limited coastal regions. The resulting model constitutes a growing-season suitability index explicitly framed for the East Monsoon cultivation window (June–August), and provides a spatially explicit, satellite-informed decision-support tool for K. alvarezii site selection in Lombok Island waters.