Analyzing Land-Use Changes in Tourism Development Area: A Case Study of Cultural World Heritage Sites in Java Island, Indonesia

    Development and tourism activities have the potential to damage tourism sites, especially World Heritage sites (WHS), and endanger ecosystem services in general. The increase in tourism requires more space for facilities and activities, with a subsequent loss of natural ecosystems and their services (Lorilla et al., 2018). For example, based on satellite imagery data, it was found that the urban and agricultural areas increased in size around the largest historical tourist sites in Jordan (Ababneh et al., 2019). Changes in land use that occur due to population growth, infrastructural expansions, environmental pollution, and responses to tourist expectations, such as by constructing accommodation facilities (hotels, motels) or new roads, could lead to irreversible damage and ecological equilibrium disruption (Ayhan et al., 2020). The construction of infrastructure and transportation to support facilities and accommodations for tourism development induce vegetation deterioration and bring ecological risks (Yang et al., 2018). Various facilities that can support tourism activities tend to cluster (Widaningrum et al., 2017;  Widaningrum et al., 2018), resulting in an area with high social and economic interaction becoming even denser. Another study in a tourism-education and business city also showed a dependence on transportation facilities to fulfill community needs, although the availability and quality of these facilities still need to be improved (Purba et al., 2017). Tourism destinations included in the WHS categories are important to study. Despite the aim of increasing tourism flow at WHS, the government must also ensure that heritage sites in their administrative areas are protected and preserved under the mission of UNESCO's World Heritage Center (2018). Tourism activities and infrastructure development are considered to negatively influence tourism sites, specifically natural and cultural WHS properties in South-East Asia (UNESCO, 2012). Therefore, it is crucial to have an infrastructure development plan to support sustainable tourism and to ensure ecosystems can continue to provide services. Cities that cope with the challenges of both city development and tourism development have a greater opportunity to attain sustainability (Berawi, 2018).

   Appropriate planning requires detailed and accurate information to support the decision-making process. Current technological developments support the fulfillment of relevant data availability, including for spatial data. Spatial data are essential to gain knowledge about the spatial characteristics of an entity, as well as its relationship with other entities. The proximity characteristic refers to the first law of geography from Tobler (1970). The availability of spatial data allows for analyzing and simulating changes in land use to estimate the condition of land cover at a specific time, even for future timepoints.

        This study aims to analyze changes in land use in the areas around tourism sites. The analysis was conducted to provide information for policymakers and investors so that they can make the right decisions in the infrastructure development plan. Notably for areas whose economic sector is being developed, but also in particular, have areas that must be preserved. Employing spatial data from several sources, this study involved managing and conducting the initial processing with the geographical information system approach, using ArcGIS® Desktop 10.0 for Windows®. A future land use simulation (FLUS) model was developed to simulate land use changes based on spatial-temporal data on areas around tourist destinations in general and especially around cultural WHS. Previous studies showed that infrastructure development (including transportation facilities) and topographical conditions significantly affected changes in land use in tourism areas (including WHS). This study examined land use changes with the proximity of built-up areas, transportation facilities (airport, main road, railway station), tourism sites, and topographical factors (elevation, slope, aspect) as the driving factors using the artificial neural network (ANN) approach. The spatial data allocation of prediction results was carried out using the cellular automata (CA) approach. The results of the spatial allocation were used as a basic map for the analysis of the proximity between specified land use and tourism locations, which were managed in PostgreSQL. Proximity analysis utilized lines of code/SQL queries to search for spatial elements that were close to a specific perimeter surrounding each tourism site.

1.1. Spatial-Temporal Analysis

   Current economic conditions that are increasingly service-oriented and continuously changing require a modern analytical approach to decision-making that considers risk, such as modeling and simulation, and that also includes spatial and temporal contexts in its analysis (Chan, 2011). Tobler’s first law of geography states that ‘‘everything is related to everything else, but near things are more related than distant things” (Tobler, 1970). According to this law, the context of land use implies that the surroundings of a location are related to the land use at that location, but close surroundings have a stronger influence than more remote surroundings. The neighboring influence on land uses built into CA-based land use models based on the neighborhood effects (Vliet et al., 2009).

        Mas et al. (2014) previously compared CLUE-S, Dinamica EGO, and IDRISI, and one aspect emphasized was the process of determining land demand. Dinamica and IDRISI (CA_Markov and LCM) use the Markov matrix to calculate the quantity of land use change, while the CLUE-S model estimates the land use demand outside the model. The recently developed FLUS model has been widely used for land use change analysis since 2017. The FLUS model also uses both a CA approach and the ANN approach to identify the relationships between driving factors and land use patterns (Liu et al., 2017b). The FLUS and CLUE-S models can both be integrated with dynamic system simulation results, namely the land demand quantity, but the FLUS model also has tools to predict the amount of land demand using the Markov chain approach (Liang et al., 2018). Additionally, in the FLUS model, constraint growth or restricted areas can be integrated into the model. Based on these strengths, this study used the FLUS model as a basis for analyzing land changes in the tourism area.