Integrated modelling tool of geospatial criteria for eco-industrial park site selection

Abstract

Single multi-criteria decision-making (SMCDM) approaches are limited by inconsistencies in the evaluation of criteria weights, making it unreliable for industrial park (IP) site selection. This led to wrong industrial site choice, difficulty to attract industry symbiosis clusters and often resulted in brownfield industrial parks (BFIP) with excessive greenhouse gas emissions. Many BFIPs are being phased out in favour of eco-industrial parks (EIP) with favourable locations for industrial clusters to synergise and manage materials efficiently. Industrial site selection heavily depends on criteria weighting and ranking. This study aimed to develop an integrated multi-criteria decision-making (IMCDM) tool and MCDM-GIS model that would enable researchers to consolidate the advantages and eliminate the weaknesses of SMCDM. To address the mentioned limitations, the analytic hierarchy process (AHP), analytic network process (ANP), and fuzzy-analytic hierarchy process (F-AHP) tools were constructed using the eigenvalue, limit supermatrix and triangular fuzzy numbers. The spatial criteria weights and ranking of water bodies, roads, residential areas, existing industries, land surface temperature and slope were evaluated. The SMCDM priority vectors were alternately integrated to produce the IMCDM methods which were also used in assessing the criteria weights. All criteria weights were subjected to sensitivity analyses and standard deviation. To test the weighting consistency of the SMCDM, IMCDM and the model efficiency, the spatial criteria data of 2009 and 2019 of Tanjung Langsat Industrial Area (TLIA) were collected using the geographic information system (GIS) and screened by the Boolean logic. The Landsat-7 enhanced thematic mapper and the kompsat-3 imager obtained the land use land cover data through PLANMalaysia. The GIS prepared the Euclidean distance and reclassified raster layers. The single and integrated weights percent were separately overlaid in the MCDM-GIS model with the 2009 criteria dataset. The SMCDM and IMCDM approach identified the water bodies as suitable brownfield eco-industrial park (BF-EIP) sites. This shows tool inconsistency using sparse criteria because industries cannot be built inside water bodies. Using the 2019 data, the AHP, ANP and F-AHP identified 5%, 2% and 3% as very-highly-suitable sites all in the northern part of the TLIA. The small spots were found away from the existing industries' location when superimposed with the criteria layers. The integrated hierarchy network-fuzzy analytic process and hierarchy network analytic process methods identified vast sites of different suitability but included 12% part of the water bodies as low-suitable, hence considered as inconsistent. The hierarchy fuzzy-analytic process (H-FAP) and network fuzzy hierarchy-analytic process (NFh-AP) measured large different suitable sites with explicit identification not including water bodies, hence consistent and reliable tools. When overlaid with the criteria layer, the very-highly-suitable site was identified in the centre of the TLIA, falling in place with the existing industries. The integrated H-FAP and NFh-AP algorithms become consistent and the best because of the interplay of hierarchical, geometric ratio and networking tools coming from different groupings of paired comparison and uncertainty, as well as their weights being close to the averages of the criteria set as evaluated by the standard deviation. The IMCDM tools are consistent only with concentrated criteria. However, the SMCDM tools are weak with both the sparse and concentrated criteria. This can lead to the wrong choice of an industrial site. Both SMCDM and IMCDM measured the economic, environmental, and social attributes as the most important in supporting the criteria to achieve the BF-EIP site selection. The MCDM-GIS model is efficient as the outputs of suitable EIP site layers under different criteria weights and distinguished spatial data. The H-FAP, NFh-AP have been proven to be the consistent criteria weight assessment algorithms and a flexible MCDM–GIS is hereby presented to support the government, EIP investors/developers, and researchers. This is to achieve an easy 4IR-driven modelling process to select brownfields for EIPs.