LCLU location-allocation with spatial contiguity and compactness

Indonesia, the most densely archipelagic state, has a population of 282 million, or 3.47% of the world's population. In 2022, Indonesia, in general, and Java, in particular, already experienced a biocapacity deficit based on a study by the Global Footprint Network. This study aims to optimize land cover/land use (LCLU) location selection. This ecological footprint (EF) is utilized as an LCLU allocation assessment based on primary demands. The EF analysis shows that paddy rice croplands and estate crop plantations are deficient. This study also conducted an LCLU suitability model using a support vector machine (SVM) via an error-correcting output codes (ECOC) classifier for multiclass learning. The ECOC obtains posterior probability values that indicate the suitability level. The overall accuracy was 82.81%, with a kappa coefficient value 0.772. The next step is LCLU location-allocation based on spatial contiguity and compactness (2C) simultaneously with the Pareto optimization approach. The verification results show that the algorithms can facilitate each 2C definition. The results show that estate crop plantations still exhibit deficits, but deflation exists. The deficit-cutting for paddy rice croplands reached 59,185.92 ha, and estate crop plantations reached 676,962.86 ha. The LCLU location candidates have limited choices because Java is already densely populated. There are limited choices of LCLU locations because Java is already densely populated. Despite the results, geometrically, the results of location selection are more compact, and there is a reduction in land fragmentation.