3.1. Study Area
Liuqiu Islande, which is located at 22°33′86″ N and 120°36′98″ E, is in the southwest of the port of Pingtung County and is the only coral island in Taiwan. The Liuqiu Island has an area of 6801 square kilometers, and has a population of around 13,000 residents. Reefs are rare among the tourist islands of Asia. Liuqiu Island has a rich terrain, marine ecosystems, and historical monuments. In addition, it is Taiwan's only winter tourist island and has various attractions including natural resources, cultural heritage, traditions, and festivals. A variety of activities including snorkeling and water sports are also available. Scholars have assessed the impact of recreation on the biodiversity [
8] and recreation experience [
9] of Liuqiu Island. Negative impacts generated by numerous visitors and recreational activities may affect the biodiversity and reduce the biological populations in the intertidal zones. According to statistics from the Tourism Bureau (2016) [
41], the annual number of visitors to Liuqiu Island increased by approximately 54% from 260 thousand tourists in 2010 to 400 thousand in 2015. Thus, tourism has become a crucial industry on Liuqiu Island. As such, local residents have increasingly been providing lodging and recreational facilities in order to meet the growing demand from both local and international visitors. However, the construction of such facilities can cause substantial negative impacts to the environment, including to water and land resources, because the terrain and geology of this ecologically sensitive island are relatively fragile. With this in mind, assessments of the sustainability of tourist and recreational facilities that can guide the relevant decision-makers with respect to efficient resource usage are urgently required to ensure that the environment and its ecosystems are protected even as operational efficiency is increased and associated costs are reduced.
3.3. Ecosystem Service Value Model
Costanza et al. (1997) [
42] defined ecosystem services as “the benefits human populations derive, directly or indirectly, from ecosystem functions.” According to the Millennium Ecosystem Assessment (MEA) published by the United Nations in 2005, ecosystem services comprise supply services, regulating services, cultural services, and support services [
43]. In this study, the ecosystem service value was assessed using the money value assessment method. Using the models proposed by Costanza et al. (1997) [
42], Mamattursun et al. (2010) [
44], and Sawut et al. (2013) [
45], we calculated the ecosystem service value in the research area.
We estimated the equivalence factor of the ecosystem service values for the various types of land use and land cover on Liuqiu Island according to the following assessment model [
39], as presented in Formula (1):
where the total ecosystem service value is denoted by ESV; the distribution area of the
ith type of land use and cover (gha) is denoted by
; the equivalence factor for the
jth item of ecosystem goods and services provided by the
ith ecosystem is denoted by
; the production per unit area or the ecosystem service value coefficient is denoted by
; a K coefficient of regional differences is denoted by
; a K regional service support coefficient is denoted by
; the type of land use and cover in a particular ecosystem is denoted by
i; and the ecosystem service category is denoted by
j.
3.4. Evaluation Items of the EF Model
Due to the ongoing and progressively increasing interest in sustainable development, interested researchers and institutions around the world have gradually developed a range of tools and indicators that can be utilized in order to evaluate sustainable development efforts. Ideally, such tools and indicators should provide accurate and reasonable reflections of the actual environmental features which they refer to while also providing effective analyses of resource consumption as well as explorations of the relationships among distinct types of environmental impacts. In general terms, the indicators and measurement models of sustainable development currently in use, whether developed domestically or internationally, all have their own particular features. The majority of them succeed in terms of giving due consideration to how various aspects of societies, economies, ecologies, and environments affect sustainable development. However, the existing indicators and measurement models nonetheless exhibit several key shortcomings. First, some of the indicators and measurement models appear to be too complex to provide accurate reflections of the specific connotations of sustainable development, even as the dynamic indicators of sustainable development suggested in previous research appear to be insufficient. Second, a number of the existing indicators and their related measurement models were based on comprehensive systems, making the quantification of the indicators difficult and even impossible, which in turn makes them relatively inoperable with respect to practical purposes. Third, data accessibility problems affect several more of the indicators and measurement models, making them difficult to apply in various contexts.
Rees first proposed the EF model in 1992 [
46]. The key aspect of the model consists of its capacity to measure and compare human environmental demands in relation to the biosphere’s capacity to replenish its resources and yield the services demanded. In a subsequent study, Wackernagel and Rees (2000) [
47] suggested that the magnitude of EF is inversely proportional to the per capita usable area of biologically productive land while also being directly proportional to the environmental impact. Due to the fact that it is relatively easy to both understand and calculate, EF has become a quantitative indicator that is now widely utilized in the field of ecological economics. A number of studies conducted in Taiwan have investigated the EF model in terms of its basic concepts, theoretical hypotheses, assessment methods, deficiencies, and empirical applications [
48,
49,
50], in addition to developing a number of EF-related theories and estimation methods. Relatedly, Lee and Peng conducted a study in 2014 [
51] in which they expanded upon earlier research in order to analyze Taiwan’s EF from 2008 to 2011. In a study the following year, Lee [
52] expanded upon that preceding study still further by conducting a time-series analysis in order to evaluate the land footprint, carbon footprint, and water footprint in Taiwan from 2000 to 2011. Among that study’s specific findings was the finding indicate that Taiwan’s land footprint declined from 5.39 gha in 2000 to 3.63 gha in 2011.
In the present study, we sought to evaluate the EF of Liuqiu Island from 2010 and 2015, adopting the EF concept previously suggested by Gössling et al. (2002) [
53] and subsequently utilized by Martin-Cejas and Sanchez (2010) [
54] as the theoretical framework through which to do so. Using this approach, the evaluation items were categorized according to three types, namely, transportation EF, activity EF, and food and fiber consumption EF items, in order to assess the impacts of EF on the island’s environment. The transportation EF category was further divided in terms of two key features: (a) the built-up area of transportation facilities utilized by travelers (i.e., road areas and parking lot areas) and (b) the energy consumed for the purpose of transportation during travel activities. The computation of the activity EF category was also broken down in terms of two key aspects: (a) the built-up land areas within various types of scenic areas, including tourist trails, highways, and scenic view spaces and (b) the area transformed through fossil-based energy consumption, such as the areas in touring scenic sites transformed by vehicle usage. The food and fiber consumption EF category, meanwhile, consisted of three key features: (a) the land area occupied by food and beverage service facilities (for example, restaurants and beverage sellers), (b) the total area of biologically productive land transformed as a result of by tourists through the consumption of foods, and (c) the total area of biologically productive land transformed by tourists through the consumption of fiber. The sources of data and main evaluation items for each of these categories are listed in
Table 1.
The general formulas for calculating the EF and ecological carrying capacity (EC) are presented in (2) and (3) as
where EF denotes the total EF (gha);
N denotes the total population;
ef denotes the per capita EF (gha);
aai denotes the per capita biologically productive area (gha) converted to the
ith traded commodity type;
ci denotes the per capita consumption (kg) of the
ith commodity type;
pi denotes the average global productive capacity [kg/(t/gha)] of the
ith consumer good type;
rj and
yj denote the equivalence factor (EQF) and yield factor (YF) of the
jth land type, respectively,
j denotes the corresponding type of land use or cover; EC denotes the total ecological capacity,
ecj denotes the ecological carrying capacity per capita; and
Aj denotes the per capita area of the
jth land type in the region [
39].
3.5. Ecological Footprint Model
A national footprint model aimed at classifying biologically productive land into six types, namely, grazing, crop, fishing, forest, carbon uptake, and built-up land, has been developed by the Global Footprint Network. Because these six types of land have differing levels of biological productivity, their areas are weighted such that they can be represented using a unit of measurement known as the global hectare, or “gha”, with 1 gha of a given land type representing an area equivalent in terms of biological productivity to 1 gha of any other type. In effect, the global hectare can be used to quantify the biocapacity of the earth in a given year, with 1 gha representing the average productivity of the various types of biologically productive areas. The calculation utilized to convert raw land area values into gha values mainly relies on the terms EQF and YF.
EQF is used to evaluate the differences among the aforementioned six types of productive land; specifically, it represents the ratio of the average potential biological productivity of all global lands to the potential biological productivity of a specific land type. That said, because the lands in different regions and countries have different levels of available resources, the biological productivity of even the same type of land may vary across different regions, to say nothing of the variations across the different types of land themselves. Therefore, in order to make accurate comparisons among different regions, the area of each type of land in the region or regions under consideration must be converted into an equivalent area corresponding to the global average for biological productivity, and the conversion factor used for such conversion calculations is known as the YF.
The traditional EF model classifies the EF and ecological carrying capacity of land ecosystems into biological resource consumption (e.g., agricultural land, forest land, grassland, and fisheries) and energy consumption (carbon footprint) along with six other ecological system units [
39]. However, the EF model combines the ecological function of systems and services to calculate the ecological carrying capacity in consideration of the ecosystem services of various land ecosystem supply units. Therefore, this study determined the ecological carrying capacity by reclassifying the ecosystem service value of land ecosystems into the following categories: (1) agricultural ecosystems; (2) forest ecosystems; (3) grassland ecosystems; (4) settlement ecosystems; (5) fishery ecosystems (including lakes, rivers, and wetlands); and (6) unutilized land and the six other ecosystem units. By contrast, traditional EF estimates consider only food and raw material production to provide two ecological functions [
39].
In contrast, in the integrated EF model, an appropriate EQF and YF are employed in order to ascertain the ecological functions for the biological land productivity and the raw materials for food production rather than the ecological functions and the ecosystem service value. The YF was thus primarily used in this study for the purpose of representing the differences in ecosystem service values per unit area across various regions of Liuqiu Island, and was calculated as follows (4):
where YF
j represents the YF of the
jth ecosystem unit type, with
j = 1, 2, …, 6 indicating the six ecosystem unit types;
vj represents the function of the ecosystem service value per unit area of the
jth ecosystem type of a region; and
represents the function of the mean ecosystem service value per unit area of the
jth ecosystem type in Taiwan.
According to Formula (4), this study calculated the YF values of Liuqiu Island from 2010 to 2015 (
Table 2).
3.6. Ecological Analysis and Safety Evaluation
This study used the EF method to evaluate the use of ecological resources on Liuqiu Island to determine whether the current tourism levels exceed the land assimilative capacity of the area. The ecological remainder status is defined as a biocapacity value greater than zero (on the environmental resource supply side) assuming that the EF (on the environmental resource demand side) is neglected. A value of zero or lower indicates ecological deficit. Furthermore, to make EF capacity per unit area more accurately reflect environmental pressure, the ecological footprint index (EFI) was adopted to evaluate regional ecological security. EFI is also known as the EF pressure index and refers to the EF based on the ecological carrying capacity per unit area of a certain region. EFI levels are listed in
Table 3 and were calculated as follows (5):
where EC represents biocapacity. When 0 < EFI < l, the supply of ecological resource supply exceeds the corresponding demand, and the region is ecologically secure. When EFI = 1, the supply and demand of ecological resources are balanced, and the region is ecologically unstable. When EFI > 1, the pressure per unit biocapacity area exceeds the supporting capacity. Thus, the supply and demand are unbalanced, and ecological security is under threat; the greater EFI deviates from 1, the greater the degree of ecological insecurity. As shown in
Table 3, EFI < 0.5 is good to maintain a sustainable ecosystem. EFI = 0.5–0.8 and EFI = 0.8–1 indicate fair and poor situations, respectively. If EFI > 1.0, a long-term sustainable ecosystem is unlikely to be maintained.