Integrated GIS–Socioeconomic Framework for Accelerating Salt Self-Sufficiency in Indonesia

2.1 Framework integration architecture

This study establishes a novel hybrid methodological framework that dynamically integrates geospatial suitability, discrete-choice behavioral modeling, and macro-structural simulations. Addressing a major gap identified in the literature regarding the absence of dynamic feedback loops between static spatial constraints and behavioral adoption parameters over time [15, 16], the proposed workflow formalizes a bidirectional input-output data bridge among three analytical components. First, the GIS-MCDA output classified as Highly Suitable (S1) is operationalized as the absolute capacity constraint parameter ($Max$_$Land$$_{S 1}$) for the potential land stock variable within the Vensim SDs model. Second, the Binary Logistic Regression results transform empirical Odds Ratios (OR) for cooperative membership (OR = 4.482) and extension access (OR = 2.586) into normalized weighting coefficients ($w_i$) to calibrate the dynamic Technology Adoption Rate ($T A R_t$) equation, ensuring that institutional support evolves as an endogenous driver within the simulation framework.

2.2 Study sites and spatial data boundaries

Empirical field data and spatial observations were gathered across the 2024/2025 production cycles, bounding the research scope strictly to ten primary salt-producing provinces in Indonesia: Aceh, DI Yogyakarta, West Java, Central Java, East Java, Bali, West Nusa Tenggara, East Nusa Tenggara, South Sulawesi, and Gorontalo. Spatial data boundaries utilize secondary geospatial layers acquired at a 30-meter pixel resolution, projected using the World Geodetic System 1984 (WGS 84) coordinate reference system. The following are the Biophysical Suitability Criteria and AHP Weights, presented in Table 1.

Table 1. Biophysical suitability criteria and Analytical Hierarchy Process (AHP) weights

2.3 Geographic Information System-based Multi-Criteria Decision Analysis

To accurately identify potential zones for economic cluster expansion, we executed a GIS-MCDA utilizing an Analytical Hierarchy Process (AHP) matrix to calculate parameter weights [12, 13]. Biophysical datasets were filtered through explicit threshold classification rules to divide land suitability into four standardized tiers: Highly Suitable ($S 1$), Moderately Suitable ($S 2$), Marginally Suitable ($S 3$), and Unsuitable ($N$) [17, 18].

The suitability model incorporates five biophysical criteria: salinity, annual rainfall, land slope, distance to the coastline, and average air temperature. We evaluated the pairwise comparison matrices against expert validation protocols, strictly maintaining an AHP Consistency Ratio ($CR$) below 0.10 to eliminate systematic weighting bias [19]. Final map boundaries and regional acreage totals were calculated using weighted linear combination overlays within QGIS software.

Pairwise comparison matrices were verified against an expert-judgment consensus model, strictly maintaining a $CR$ below 0.10 to eliminate systematic bias. The final suitability map layer was generated via a Weighted Linear Combination (WLC) aggregation overlay, with total spatial geometry area calculations executed within QGIS software.

2.4 Socio-demographic sampling and statistical modeling

Primary socioeconomic data were compiled through a purposive sampling strategy targeting smallholder solar salt farmers across the ten target provinces. To ensure statistical power and modeling precision, we applied a strict data-screening protocol to the field survey pool. Out of 300 original farmer observations, 25 profiles were permanently excluded due to incomplete answers, internal tracking contradictions, or extreme multivariate outliers. This screening process yielded a finalized dataset of 275 valid observations for inferential evaluation.

We constructed a binary logistic regression model to evaluate individual technology adoption decisions, setting geomembrane adoption as the binary dependent variable ($Geomembrane$ $=1$, $Traditional$ $Clay$ $Bed$ $=0$). Predictor variables include farmer age, education level, landholding size, cooperative participation, and access to extension counseling. Model stability was validated using Variance Inflation Factors (VIF), which were restricted below 2.5 to rule out multicollinearity, while overall predictive fit was verified via the Hosmer-Lemeshow test.

2.5 Climate-controlled productivity evaluation

Direct comparative $t$-tests comparing traditional methods and geomembrane systems face severe confounding risks from multi-provincial climate variations, seasonal lengths, and differing baseline pond layouts. To isolate true technological performance gains from environmental noise, this study replaces simple unstratified testing with a climate-controlled analysis.

We implemented multiple linear regression modeling alongside stratified $t$-tests to evaluate seasonal productivity (expressed in tons/ha/season). By embedding provincial climate parameters-specifically seasonal rainfall totals and average regional temperatures-alongside operational farm characteristics (such as cultivated land area size) as continuous control variables, the model isolates the net technological impact of geomembranes. Performance gains are reported with 95% confidence intervals (CI) and practical effect sizes (Cohen's $d$) to ensure mathematical rigor.

2.6 System dynamics simulation setup

To project the long-term macroeconomic impacts of the proposed integrated cluster interventions, we developed a mathematical Stock and Flow simulation model using Vensim PLE software. To resolve the timeline contradictions identified in the original draft text, both the model architecture and all reported simulation outcomes are structurally locked to a uniform, long-term planning horizon spanning from 2025 to 2035. The system boundary coordinates causal loops and dynamic interactions between salt production capacity, technology adoption velocity, cooperative participation rates, market prices, and net smallholder household incomes. The simulation was parameterized using the spatial constraints derived from the GIS-MCDA step and calibrated utilizing the socio-institutional adoption weights generated by the logistic regression model.

4.1 The SEGAR conceptual model strategy

The empirical findings from this study validate the design of the Integrated and Sustainable Community Salt Business Area (SEGAR Cluster Model). Identifying 15,500 hectares of highly suitable ($S 1$) land provides a concrete geographical foundation that prevents the fragmented, uncoordinated land-use practices common in previous development schemes. By targeting volume-based clusters in areas like East Java and East Nusa Tenggara, and value-based clusters in land-constrained areas like Aceh and Yogyakarta, the framework aligns physical land capability with regional market dynamics. The SDs simulation demonstrates that this structured clustering approach yields a consistent 60% to 75% increase in net farmer income by the 2035 horizon, proving that linking spatial data directly to socio-institutional scaling mechanisms outperforms isolated technological or financial interventions.

4.2 Technical and environmental mechanisms of smart salt farming

The significant productivity leap from 65.8 to 105.3 tons/ha/season demonstrated in our results is fundamentally driven by specific biophysical and thermodynamic changes engineered by geomembrane linings. Traditional clay beds suffer from high thermal dissipation into the underlying soil substrate, slowing down evaporation rates. In contrast, geomembrane linings maximize solar radiation absorption and trap heat within the brine layer, maintaining significantly higher water temperatures and accelerating salt crystallization rates [3, 20, 21].

From an environmental standpoint, the geomembrane serves as an impermeable barrier that blocks direct contact between the processing seawater and the ground soil. This configuration isolates the crystallization process from soil-borne heavy metals and pollutants, yielding high-purity salt that meets the standard of $\geq$95% NaCl required for food and industrial substitution. Concurrently, this barrier function prevents localized soil salinization and mitigates subsurface leakage risks in zones characterized by high groundwater vulnerability [20, 22, 23].

Crucially, modernizing salt fields with geomembranes shifts production towards a sustainable, zero-waste operation. Traditional farms often discard bittern as a waste stream, causing localized marine toxicity. The smooth, non-porous surface of geomembrane systems simplifies the collection, concentration, and systematic reuse of bittern within a closed loop, aligning directly with zero-liquid-discharge (ZLD) principles [24, 25]. This method allows for the recovery of valuable chemical byproducts, such as magnesium and potassium salts, minimizing the industrial footprint of smallholder operations while diversifying secondary revenue streams [25, 26].

4.3 Cross-national cluster evaluation

To contextualize Indonesia's salt sector transformation, Table 5 compares the governance and scale efficiencies of the proposed SEGAR model against dominant international salt industry architectures. Linkages within smallholder-driven cooperative networks, such as those found in Gujarat, India, highlight the power of decentralized clustering to foster collective efficiency and flexible market specialization among small enterprises [27, 28]. However, as observed in both Indian and Indonesian traditional sectors, smallholder-only clusters frequently face severe scale-efficiency ceilings due to restricted technical capital, fragmented landholdings, and weak financial backing [29, 30].

Conversely, large-scale corporate clusters in Shandong, China, and Western Australia optimize scale efficiency and global cost-competitiveness through heavy research and development investments, advanced automation, and deep vertical integration [31]. Yet, these highly centralized corporate models carry high social costs, frequently creating monopolistic landscapes that displace traditional artisanal producers and reduce local community inclusivity [32]. The following is the Cross-National Cluster Evaluation, presented in Table 5.

Table 5. Cross-national cluster evaluation

The proposed framework resolves this trade-off. By integrating high-Technology Readiness Level (TRL) geomembrane modernizations into an inclusive cooperative governance framework, the SEGAR model achieves the technological precision and quality standards of corporate systems without sacrificing smallholder land ownership or coastal social equity.

4.4 Policy roadmap and national flagship research agenda

Our binary logistic regression model identified cooperative membership as the strongest predictor of technology uptake ($Odds Ratio$ $=4.482$), proving that technology transfer cannot rely on linear, individualistic dissemination models. Implementing the multi-year Nusantara Sea Salt Blueprint (NSSB) requires moving away from conventional top-down policies that exclude community perspectives, which historically cause marginalization, resistance, and conflicts within marine spatial planning frameworks [35-37]. Sectoral policy fragmentation between fisheries management, industrial trade regulations, and agrarian land reforms routinely undermines long-term developmental blueprints [38, 39].

To counter these structural bottlenecks, the proposed fishbone research roadmap operates as a polycentric, non-linear innovation system. By embedding academic research centers, the National Research Agency (BRIN), and smallholder cooperatives into an interconnected multi-stakeholder network, the framework bridges central governance institutions and local coastal actors [40, 41]. Cooperative networks serve as institutional hubs that handle data verification, pool capital to mitigate economic stressors, and deliver continuous technical training, resolving the education and awareness deficits that typically halt technology transfer in developing maritime economies [42-44].

The following is the Proposing Nusantara Sea Salt Blueprint (NSSB) Research Flagship Fishbone for Indonesian Salt Independence, presented in Figure 4.

4.jpg

Figure 4. Proposing Nusantara Sea Salt Blueprint (NSSB) Research Flagship Fishbone for Indonesian Salt Independence

4.5 Framework limitations and future paths

Despite its robust integration, several boundaries limit the immediate generalizability of this framework. First, while geomembrane technology significantly elevates salt yields and purities, long-term material exposure to solar UV radiation and hypersaline friction introduces risks of polymer degradation and subsequent microplastic contamination in the harvested salt [22, 45]. Future flagship research must prioritize the material science engineering of bio-based or highly durable solar evaporators to eliminate microplastic shedding risks [26].

Second, the high initial capital investment required for geomembrane installation and cluster infrastructure remains a steep entry barrier for independent, small-scale operators living under acute socioeconomic stressors [21, 46, 47]. Consequently, the success of the SDs projection remains highly dependent on sustained government financial subsidies and targeted credit incentives through institutional banking channels.

Finally, our cross-sectional research design captures socio-demographic and environmental conditions within a single time bracket. Future research directions should deploy longitudinal tracking and agent-based modeling to better map shifting socio-cultural traditions, community trust dynamics, and the evolving impacts of macro-stressors like sea-level rise and saltwater intrusion on smallholder behavioral adaptation.