1 Introduction

While renewable resources in nature can be restored to a certain extent, humans are increasingly consuming natural resources. However, one of the most decisive points in resource consumption is how quickly these resources are used. With a concept that started to be developed in the 1990s, today nature’s self-renewal process is expressed in terms of ECF and available bio-capacity. To ascertain a community’s environmental impact, the ECF, a metric that delineates the requisite bio-productive acreage—whether terrestrial or aquatic—necessary for the regeneration of consumed renewables and the assimilation of generated waste under the purview of current technological practices. Conversely, bio-capacity is the gauge of biological materials that an area—be it tillable land, grazing grounds, forest expanses, or productive ocean tracts—can supply. When the ECF surpasses bio-capacity, it is possible to speak of an ecological deficit (Schaefer et al., 2006). This ecological deficit can arise for three reasons. Initially, a nation might rely on the importation of naturally renewable resources, tapping into supplies it consumes yet does not cultivate domestically. This can lead to a deficit due to that country’s dependence on external ecosystem services. Second, unsustainable agricultural practices, overgrazing, overfishing, or deforestation may be using bio-capacity to the extent that it cannot replenish itself. Third, the ECF may increase due to carbon footprinting if CO2 is released into the atmosphere quicker than the natural absorption pace (Niccolucci et al., 2011). Reducing the ECF hinges on a nuanced trifecta: the scaling down of population density within a specific locale, the curtailment of individual consumption, and the enhancement of resource efficiency (Schaefer et al., 2006). The cases where the bio-capacity is larger than the ECF alone does not indicate that the country has a sustainable system. Because this bio-capacity is essential not only for human activities but also for the survival of other living things (Galli 2014).

Planet Earth grapples with grave threats—chief among them the erosion of diverse ecosystems and the diminishment of biodiversity. These crises have surged to the forefront as unchecked industrial expansion, the easing of trade constraints, the relentless pursuit of economic advancement, and a burgeoning global populace exert their toll. Of late, these elements coalesce, casting a shadow on human well-being and the vitality of natural habitats, further exacerbating the decline in biodiversity; all this is compounded by the insidious rise in tropospheric ozone levels (Bhuiyan et al., 2018). Although biodiversity is essential for the continuity of human life, it is rapidly being destroyed by humans. Forests, grasslands, wetlands, and other important ecosystems have been destroyed as a result of intense human mobility that started with the Industrial Revolution. Between 1970 and 2016, the inhabitant mass of mammalians, birds, amphibian genus, reptilians, and fishes decreased by 68% on average, and more than 85% of wetlands were lost. The biggest causes of biodiversity loss are the increment in international trade, excessive use of goods, and populace growth over the last 50 years, as well as urbanization. By the time 1970, earth’s ECF was lower than the world’s speed of renewal, while at least 56% of the world’s bio-capacity was overused to sustain a twenty-first century lifestyle. At an unprecedented pace, the fabric of the natural world undergoes a profound metamorphosis, with the forces of climate change propelling this rapid transformation even further (WWF, 2020). The increase in ECF increases pressure on bio-capacity and biodiversity. When the ECF exceeds bio-capacity, it causes CO2 to accumulate in the atmosphere and oceans, which is one of the most critical risks to biodiversity (GFN, 2016).

By 2050, inhabitants are expected to be approximately 10–11 billion, which implies that global food demand will increase (García-Oliveira 2022). The rural areas are increasingly marred by the swelling numbers of human and animal inhabitants. This surge exerts pressure on the land, precipitating its degradation, accelerating soil erosion, and amplifying the diffuse influx of pollutants. Forests and shrubs are being destroyed and wetlands are being invaded to meet energy demands from wood fuel, human settlement, and construction. Deforestation due to population pressure and changing plant species is loading increasingly sediments into rivers and lakes in the basin (Olago & Odada, 2007). Scarcity of soil and water resources, the reduction in biodiversity, increasing negative impacts of climate change, habitat destruction, biodiversity loss, land degradation, and pollution threaten sustainable food production (García-Oliveira 2022).

Before agriculture, the hunter-gatherer lifestyle fed about 4 million people around the world. But, modern agriculture now feeds around 8 billion people. Cereal production has duplicated in the last 40 years because of increased yields from new crops with the use of more fertilizers and pesticides (Tilman et al., 2002). Nevertheless, with a growing population and low-income levels, one of the most formidable obstacles confronting humankind… is meeting basic food needs. While increasing agricultural production, it is fundamental to minimize the harmful impacts of agriculture to maintain ecological balance. The harmful environmental impacts of agriculture are mainly caused by the transition of natural habitats into agricultural areas (Mózner et al., 2012). Land can be used for agriculture as long as it is not desert, tundra, rocky, or arctic, yet half of the land available for agriculture in the world is already under cultivation (Tilman et al., 2001). Moreover, the loss of natural areas, the increase in agricultural land leads to harmful amounts of substances, such as nitrogen and phosphorus in the soil. For these reasons, agricultural production has a substantial effect on the ecosystem (Tilman et al., 2002). In particular, excessive water consumption caused by increasing agricultural production has increased the stress on water resources (Martindale 2014). On the other hand, climate change and increased rapid utilization affect water resources and cause a decrease in agricultural production (Hanjra & Qureshi, 2010).

Livestock occupies a major part of the food production system. Grazing expanses weave across a staggering 26% of planet’s ice-free land, a mosaic of pastures and ranges. Contrastingly, the cultivation of forage crops envelops a full third of the globe’s arable soils. Within the tapestry of agriculture, the pastures and fields dedicated to livestock embracing a vast 70% of all farmland. When one pauses to consider the entirety of Earth’s land, livestock that have claimed 30% (FAO, 2006). However, 35% of agricultural products produced globally are used as animal feed. Despite usage as food, the use of greatly valuable cropland to produce animal feed can result in a net decrease in the earth’s possible food supply (Foley et al., 2011). The livestock industry is ranking within the top echelons of contributors to severe environmental issues at every conceivable level. From local brook to vast global stage, the presence of livestock heralds a cascade of pollutants—from the scattering of nutrients to the seepage of pathogens and the drug residues that slip into waterways. Animals and their byproducts discharge gases, playing a substantive role in the tapestry of climate change, while simultaneously driving alterations in land use patterns dictated by the escalating demand for feedstocks and grazing terrain (FAO, 2006). Enteric fermentation of ruminants is also a major source of emissions from agricultural production (UNEP 2022). In addition, livestock consumes a lot of water. To yield just one kilogram of grain-fed beef demands a volume of water that more than what’s needed for a kilogram of plant crops (Chapagain & Hoekstra, 2003). Every year, 4387 cubic kilometers of water are channeled towards nurturing animal feed. This sum represents nearly half − 41% to be exact- of water’s role in the agriculture (Heinke et al., 2020).

It is not always easy to expand the production of food crops from the land due to declining yields and competition for limited land and water resources. However, the need for nutritious food is rising due to a rapidly growing population, and individuals are demanding high-protein meat products as incomes rise and preferences diversify. The harvest of wild fisheries and ocean-farmed offerings making up 17% of the world’s meat. Beyond mere protein, these aquatic morsels are brimming with readily absorbed micronutrients and vital fatty acids that not so easily gleaned from the produce of the land (Costello et al., 2020). Aquaculture in a brisk ascendancy, it now provides with around half of the seafood (Clark et al., 2018). In the year 2020, the waters of the world yielded a record-breaking 214 million tons of aquatic life and plants. From this, a haul of 178 million tons of finned and shelled creatures coupled with 36 million tons of algae emerged. The portion of human consumption -algae aside- has now swelled to 20.2 kg for each person on Earth. This is a leap more than twofold from the modest average of 9.9 kg per capita in the sixties. Factors fueling this change include burgeoning incomes, the magnetic pull of urban lifestyles, advancements in harvest and handle procedures, and a shift in the dietary preferences. However, it is estimated that the livelihoods of around 600 million people rely at least partly on fisheries and aquaculture (FAO, 2022). Aquaculture stands as a beacon of biological efficiency in the realm of protein production when cast against its terrestrial counterparts. This is largely rooted in the prolific nature of aquatic lifeforms and their ability to convert feed into flesh with a frugality that land-based livestock cannot match (MacLeod et al., 2020). Apart from agriculture, seafood farming is increasing rapidly in developing countries and is considered as an significant source of protein and income. However, the potentially harmful impacts of aquaculture on the marine environment are a serious concern (Zhao et al., 2013). While aquaculture’s efficiency in the choir of protein provision, it’s not without its problematic points. The effects of this burgeoning industry extend beyond greenhouse gas emissions, water clarity and echoing concerns for the marine biodiversity (MacLeod et al., 2020). The swift swell of aquaculture’s influence has revealed problems such as environmental pollution, eutrophication, pharmaceutical remnants leach into the ocean, biodiversity lost. Amidst this, food security has become a real problem (Meng & Feagin, 2019).

All this makes it essential to inspect the effects of food production on the ECF. With the present analysis, the impacts of food production on the ECF in Turkey have been examined. However, agriculture, livestock, and aquaculture production, which are the sub-components of food, have been examined separately. In the next section, the literature review will be discussed, followed by the methodology and results section, and finally, the discussion section will be given.

2 Literature review

The consequences of food production on the ECF have recently began to be examined in the literature. Some of these studies have calculated food footprints and examines the burden of food depletion on the ECF (Kissinger, 2013). In addition, the effects of household dietary patterns on the ECF have also been examined (Collins & Fairchild, 2007). The ECF production of various food types and yield levels has also been studied (Chen et al., 2010). In addition, the literature has also examined temporal trends in the ECF left by food purchases, procurement and processing within the Brazilian urban sprawl. (da Silva et al., 2021). The cumulative footprint of food production on a global scale has also been calculated (Halpern et al., 2022).

The detrimental effects of agricultural production on ecological balance have been frequently analyzed regionally and globally in the economics literature. While some of these studies have addressed the whole agriculture in general (Alvarado et al., 2021; Aziz et al., 2020; Usman & Makhdum, 2021), others have analyzed only certain crops (Cerutti et al., 2010). Furthermore, the effects of agricultural water use on the ECF have also been studied (Cheng et al., 2022). Besides, the consequences of organic agriculture and traditional agriculture on ECF have also been studied comparatively (Lustigová & Kuskova, 2006). Moreover, the effects of agriculture and macroeconomic variables such as FDI and GDP on the ECF of agriculture have also been calculated for various countries (Udemba, 2021a, 2022). The impact of variables such as dissecting how threads like globalization, the appetite for renewable energy, and the vast mechanism of agricultural production on the ECF has been examined too (Pata, 2021; Salari et al., 2021). Additionally, the consequences of agriculture, commerce, and R&D on the ECF has also been investigated (Alvarado et al., 2021). However, there are also articles investigating the impact of livestock on the ECF (Arrieta et al., 2022; Ornelas-Villarreal et al., 2022). There are also studies analyzing the impacts of only a certain type of livestock on the ECF on the axis of climate change (Onyeneke et al., 2022). In addition, the mediate and immediate utilization of soft water and soil for meat plus dairy manufacture has also been examined using the land footprint (Bosire et al., 2015). Apart from this, studies on fisheries are also available in the literature (de Leo et al., 2014; Kong et al., 2021; Swartz et al., 2010). Therewithal, aquaculture has also been examined within the scope of ECF (Zhao et al., 2013). In the realm of investigation, the lens has often focused on aquaculture’s relation with carbon dioxide emissions (MacLeod et al., 2020) and carbon footprint (Jiang et al., 2022). The consequences of socioeconomic factors on the aquaculture footprint have also been investigated (Clark et al., 2018).

Environmental problems have come to the fore frequently in the economic literature for Turkey recently. When we view the studies on ECF, it is seen that the issue is generally associated with different macroeconomic variables. Some of these articles examined the relation between foreign commerce and globalization and ECF (Acar & Aşıcı, 2017; Bilgili et al., 2020; Dumrul & Kılıçarslan, 2020; Öcal et al., 2020; Apaydin 2020; Kirikkaleli et al., 2021; Güzel & Oluç, 2022). Moreover, the consequences of renewable energy on ECF have been investigated as well (Altay Topcu, 2021; Ersungur et al., 2022; Sharif et al., 2020). Furthermore, its relationship with financial development, GDP, and growth has also been examined (Beşe & Friday, 2022; Gülmez et al., 2020; Karasoy, 2021; Köksal et al., 2020; Udemba, 2020). There are also articles investigating the relation amongst FDI and the ECF (Bulut, 2021). Furthermore, its relationship with industrialization has also been examined (Destek, 2021). There are also articles analyzing the relation amongst military expenditures with the ECF (Gokmenoglu et al. 2021). In addition, the relationship among the ECF with tourism has also been examined (Godil et al., 2020). There are also articles investigating the relationship between corruption and democracy and the ECF (Özsoy, 2021; Ursavaş, 2021). The relation among environmental taxes and technological patents on the environment and the ECF has also been analyzed (Telatar & Birinci, 2022; Yavuz, 2021). Additionally, the consequences of income inequality on the ECF was examined (Aktürk & Gültekin, 2023).

Although there has been a lot of analysis on the ECF in late years, there is no study that examines the impacts of food production on the ECF in Turkey. However, agricultural production areas, which are one of the largest components of food production, are the second most foot printing resource in Turkey. In Fig. 1, Turkey’s ECF is analyzed under six different land type categories: carbon, cropland, forests, grazing lands, fishing grounds, and built-up land. The agricultural land footprint is the expanse needed to be used for human sustenance which grains, fibers, feed for animals, oil-bearing crops, and rubber. The agricultural land footprint, represented in orange in the figure, accounts for the largest share after the carbon footprint and has increased significantly since 1960. Most of the agricultural land footprint is food-related. The rest is largely because of the tobacco production and government expenditures (such as crop fibers and cotton for paper and cloth production). The footprint of grasslands, another area directly related to food production, is calculated by calculating the surface area used by livestock for meat, milk, leather, and wool products. The grassland footprint, represented in yellow in the figure, experienced a relative decline as livestock production declined in the 2000s, but has increased again in recent years as livestock production has accelerated with the increase in food security threats. In addition, fishing grounds refer to the calculated freshwater and marine areas required to produce and provide the seafood consumed. Although the footprint of fishing grounds, shown in light blue in the figure, remains quite limited compared to other areas, it has increased since the 1980s (GFN, 2012).

Fig. 1
figure 1

Source Global Footprint Network

Distribution of ECF by Land Types in Turkey.

Full size image

Turkey has become a net exporter in the agriculture-food sector by increasing its agricultural production by thousands of tons in the last 25 years. In addition, Turkey, which has experienced significant sectoral growth in animal husbandry, ranks 1st in Europe and amid the top 10 countries in the world in terms of the number of cattle and sheep, poultry meat and egg production. In addition, being a country surrounded by different seas provides great opportunities in aquaculture production diversity (MoAF, 2019). However, this production also brings ecological problems. Global food system emissions in 2021 have increased by 14% dated from 2001, now representing 30% of entire anthropogenetic emissions. Food production from agriculture and livestock constitutes almost half of the entire agricultural food system emissions. In the year 2021, the twin impact of deforestation and the methane from the enteric fermentation stood as colossal emitters of CO2, this causes over 40% of the agri-food systems’ carbon footprint. The methane birthed from animal manure management, its interment in soil, the casual drop of dung in pastures, and the final act of waste disposal in the agri-food played yet another critical role in this situation. Emissions for each person from agri-food organizations in Brazil and Argentina in 2021 were almost three times the global average. This was followed by the United States, Russian Federation, Indonesia, Mexico, Turkey and China. The year 2021 bore witness to stark contrasts in emissions footprints within the top ten agricultural producers. Indonesia and Brazil make their names high on the list, their emissions intensities soaring beyond twice the global mean. Meanwhile, China and Turkey swayed among the ten most best performers, their emissions contribute to environmental prudence (FAO, 2021). Although Turkey seems to be in a better position compared to other largest agricultural producer countries, ensuring sustainable food production is still out of the question. Amid global averages, Turkey is contributing to its ECF by increasingly its carbon dioxide emissions per capita rise, fossil fuel energy consumption, and withdrawals of fresh water (MoAF, 2021).

While the effect of food production on the ECF is excessive, there is a gap in the literature on the topic. While the literature was inquired, it was sighted that there was none detailed study on the subject and there were a very limited number of studies on food production. The present paper focuses on filling this gap and investigates the effects of food, agriculture, livestock, and aquaculture production on the ECF for the duration 1961–2018 Turkey. The paper has a unique value as the subject has not been analyzed for Turkey and examines various sub-sectors of food.

3 Methodology

3.1 Data and variables

In this paper, the impacts of food, agriculture, livestock, and aquaculture production on the ECF in Turkey are analyzed with annual data for the period 1961–2018. The explanations for the variables used to detect these elements are shown in Table 1.

Table 1 Variables applied in the model
Full size table

The variables used in the analysis were selected considering previous studies and data availability. The ECF has recently been largely used in the literature (Güzel & Oluç, 2022; Kirikkaleli et al., 2021). The Global Footprint Network contains detailed information on both ECF and biocapacity. Therefore, it is frequently used in the literature to provide data (Galli et al., 2016). However, the fact that the data for all countries end in 2018 constitutes the biggest limitation of the database.

3.2 Modelling

While constructing the analysis, not only the consequence of food production on the ECF was examined, but also the sub-sectors of food were included in the analysis. Thus, four different models, namely food production, agricultural production, livestock production, and aquaculture production, were created, and their impact on ECF was analyzed separately. The models established in this direction are as follows:

Model 1

$${\text{logef}}_{{\text{t}}} = \upbeta_{0} + \upbeta_{1} {\text{logfood}}_{{\text{t}}} + \upbeta_{2} {\text{loggdp}}_{{\text{t}}} + \upbeta_{3} {\text{logrural}}_{{\text{t}}} + \upbeta_{4} {\text{logpop}}_{{\text{t}}} + {\text{u}}_{{\text{t}}}$$
(1)

Model 2

$${\text{logef}}_{{\text{t}}} = \, \upbeta_{0} + \, \upbeta_{1} {\text{logcrop}}_{{\text{t}}} + \, \upbeta_{2} {\text{logfert}}_{{\text{t}}} + \, \upbeta_{3} {\text{loggdp}}_{{\text{t}}} + \, \upbeta_{4} {\text{logland}}_{{\text{t}}} + \, \upbeta_{5} {\text{logrural}}_{{\text{t}}} + {\text{ u}}_{{\text{t}}}$$
(2)

Model 3

$${\text{logef}}_{{\text{t}}} = \, \upbeta_{0} + \, \upbeta_{1} {\text{loglive}}_{{\text{t}}} + \, \upbeta_{2} {\text{loggdp}}_{{\text{t}}} + \, \upbeta_{3} {\text{logrural}}_{{\text{t}}} + \, \upbeta_{4} {\text{logpop}}_{{\text{t}}} + {\text{ u}}_{{\text{t}}}$$
(3)

Model 4

$${\text{logef}}_{{\text{t}}} = \, \upbeta_{0} + \, \upbeta_{1} {\text{logaqua}}_{{\text{t}}} + \, \upbeta_{2} {\text{loggdp}}_{{\text{t}}} + \, \upbeta_{3} {\text{logrural}}_{{\text{t}}} + \, \upbeta_{4} {\text{logage}}_{{\text{t}}} + \, \upbeta_{5} {\text{loglife}}_{{\text{t}}} + {\text{ u}}_{{\text{t}}}$$
(4)

While constructing the models, it was tried to be analyzed with variables related to the areas where production is carried out. Accordingly, in the first model, the food production index was used to examine general food production. In the second model, crop production index, agricultural land and fertilizer consumption were used to examine the agriculture effect. In the third model, livestock production index is used to represent livestock production. In the last model, aquaculture production was used to examine aquaculture production. Other variables in the models are used as explanatory variables. However, the difficulty in finding relevant long-run data for Turkey prevents the analysis with a larger data set in the models.

The major inputs to agricultural production include fertilizer use, agricultural land, and labor. Although the use of fertilizers increases agricultural production, large fertilizer inputs in production cause serious pollution of surface and ground waters through leakage (Lu et al., 2016). Groundwater, a courier beneath the soil, ferries both life-sustaining nutrients and pollutants across vast expanses. It sustains bodies of water, fueling the flow of lakes and propelling rivers. Yet, it also bears the potential to amplify health hazards, casting a web of risk over an array of life, from wild fauna to livestock and the human communities that bridge the natural world. Fertilizers and pesticides alter terrestrial habitats and reduce biodiversity, which in turn affects ecosystems (Mózner et al., 2012). For this reason, their impact on the environment has been oftentimes studied in the literature (Haverkort et al., 2014). As can be seen in Fig. 1, agricultural land has the second largest effect on the ECF. This area causes a growth in the ECF during both the manufacture of agricultural output and the manufacture of feed used in livestock. Therefore, it is widely used in the literature (Muoneke et al., 2022; Zhen & Du, 2017). Although every country wants to increase its GDP, the trade, transportation, urbanization, and industrialization that come with this growth cause pollution of the air, water, soil, etc. In this context, the consequence of national income per capita on ECF has been frequently examined in the literature (Sabir and Gorus 2019; Muoneke et al., 2022). Moreover, population growth has recently been seen as one of the essential factors of environmental pollution (Rehman et al., 2022). Therefore, there are several searches in the literature that analyze the effects of both the total population and the rural population on the environment (Clark et al., 2018; Rehman et al., 2022). Furthermore, the structure of the population also affects carbon emissions. There are analyses in the literature that analyze the relation among aging with environmental degradation (Li et al. 2019; Dimnwobi et al., 2021). Life expectancy at birth is frequently used in the literature (Samreen & Majeed, 2022). Life expectancy is also used as a quality indicator, and societies with increased life expectancy are expected to increase investments in the environment. However, this increase in expectations may also bring economic growth and more environmental degradation (Boukhelkhal, 2022).

4 Findings and discussion

Table 2 represents the descriptive statistics of the variables used within this research.

Table 2 Descriptive statistics
Full size table

ARDL framework is applied as the analytical instrument to probe the effects of food production on Turkey’s ECF. Pioneered by Pesaran and Shin (1999) and further refined by Pesaran et al. (2001), the ARDL model stands as a robust methodological approach, allowing for the elucidation in short and long periods dynamics within the intersecting variables. The superiority of this approach is that it gives consistent estimation outcomes in the long period even when the series are not stationary of the same order (Chen et al., 2019). Therefore, the Phillips-Perron (P-P) and Augmented Dickey-Fuller (ADF) unit root test was implemented before the analysis. Table 3 shows the outcomes of the P-P and ADF unit root tests, presenting empirical evidence on the stationarity of the data series under examination in the paper.

Table 3 P-P and ADF Results
Full size table

The analytical outcomes derived from the P-P and ADF tests indicate that the data series exhibit integration of order zero or one. Crucially, series doesn’t present integration of order two, or higher. This absence of higher-order integration corroborates the suitability of employing ARDL estimators within the investigative framework. Prior to implementing the bounds testing procedure, model selection criteria are applied to ascertain optimal lag lengths. The chosen lag marked according to the smallest criterion value and devoid of serial autocorrelation. Table 4 presents the Schwarz criterion (SCH) values and Breusch-Godfrey autocorrelation (LM) test statistics for all models.

Table 4 Appropriate lag length
Full size table

Subsequent to the determination of the appropriate lag length, the bounds testing approach to cointegration, as articulated by Pesaran et al. (2001), is applied. The table of critical values, as established by Pesaran et al. (2001), provides a benchmark for comparison with the F-statistics computed from the model. Interpretation of these statistics follows a structured criterion: F-statistic falling beneath the lower bound (LV) implies the lack of a cointegration relation among the series. If the F-statistic resides amid the bounds, the inference remains indeterminate. Conversely, an F-statistic that exceed the upper critical bound (UV) signals the entity of a cointegration relationship. The calculated F-statistic and these critical values (CV) are delineated in Table 5.

Table 5 Bound test results
Full size table

The empirical analysis reveals that the calculated F-statistic values exceed the upper critical thresholds as stipulated for the 1%, 5%, and 10% degrees of significance according to Pesaran et al. (2001). This elevates the models to a status of significance across each of these critical benchmarks, robustly pointing the entity of a cointegrated relation amid the variables under scrutiny. Recognition of such cointegration provides a firm foundation upon which to construct ARDL models. The constructed ARDL models, which serve to illuminate the long run interrelations within the series, are articulated as follows:

Model 1

$${\text{logef}}_{t} = \upbeta_{0} + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{1{\text{i}}}} {\text{logfood}}_{{\text{t - i}}} } + \sum\limits_{{\text{i = 1}}}^{{\text{m}}} {\upbeta_{{{\text{2i}}}} {\text{loggdp}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{3i}}}} {\text{logrural}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{4{\text{i}}}} {\text{logpop}}_{{\text{t - i}}} } + {\text{u}}_{{\text{t}}}$$
(5)

Model 2

$${\text{logef}}_{{\text{t}}} = \upbeta_{0} + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{1i}}}} {\text{logcrop}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\beta_{{{\text{2i}}}} {\text{logfert}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{3i}}}} {\text{loggdp}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{4i}}}} {\text{logland}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{5i}}}} {\text{logrural}}_{{\text{t - i}}} } + {\text{u}}_{{\text{t}}}$$
(6)

Model 3

$${\text{logef}}_{{\text{t}}} = \, \upbeta_{0} + \sum\limits_{{\text{i = 1}}}^{{\text{m}}} {\upbeta_{{{\text{1i}}}} {\text{loglive}}_{{\text{t - i}}} } + \sum\limits_{{\text{i = 1}}}^{{\text{m}}} {\upbeta_{{{\text{2i}}}} {\text{loggdp}}_{{\text{t - i}}} } + \sum\limits_{{\text{i = 1}}}^{{\text{m}}} {\upbeta_{{{\text{3i}}}} {\text{logrural}}_{{\text{t - i}}} } + \sum\limits_{{\text{i = 1}}}^{{\text{m}}} {\upbeta_{{{\text{4i}}}} {\text{logpop}}_{{\text{t - i}}} + {\text{u}}_{{\text{t}}} }$$
(7)

Model 4

$${\text{logef}}_{{\text{t}}} = \, \upbeta_{0} + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{1i}}}} {\text{logaqua}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{2i}}}} {\text{loggdp}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{3i}}}} {\text{logrural}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{4i}}}} {\text{logage}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{5i}}}} {\text{loglife}}_{{\text{t - i}}} } + {\text{u}}_{{\text{t}}}$$
(8)

Table 6 shows the long period coefficients obtained by ARDL models:

Table 6 Long run estimation results
Full size table

While the long period outcomes are analyzed, it is seen that all food production processes except aquaculture production increase the ECF in the long period. A 1% raise in food, agriculture, and livestock production raises the ECF by 0.678%, 0.504%, and 0.237%, respectively, while a 1% raise in aquaculture production falls the ECF by 0.081%. However, aquaculture production is found significant at the 10% significance level. Although there are no studies using the ECF in the literature for Turkey, there are a couple of analysis in the literature stating that food production leads to environmental pollution (Dogan & Karpuzcu, 2019; Yurtkuran 2021). Apart from this, studies generally focus on the consequences of environmental pollution on food production (Chandio et al., 2021). Therefore, the results obtained from the current study are of great importance. The great importance of agricultural production on the effect of food production on the ECF is seen from the coefficients. While the consequences of the factors used during agricultural production are very important in the emergence of this result, deforestation policies for the creation of agricultural land also cause a decrease in biodiversity and increase the harmful effects of agriculture (WWF, 2020). The current analysis shows that fertilizers and agricultural land used in agriculture increase the ECF. This raises the effect of agricultural production on the ECF. Although the impact of livestock is less than that of agriculture, it has an important consequence on the ECF. The main sources of pollution in livestock, which is a constantly developing sector in Turkey, are fertilizer and excreta, which are abundant in essential nutrients such as nitrogen, phosphorus, and potassium, these elements contribute significantly to amplifying crop productivity and enhancing the overall fertility of the soil. This fertilizer is mixed with surface and groundwater and causes serious damage to the environment (Yayli and Kiliç 2021). Although aquaculture production is a relatively new sector for Turkey, it has experienced rapid growth since 2000, yet Turkey still produces considerably less than its competitors (Ertör & Ortega-Cerdà, 2019). This difference in the signs of the coefficients may be due to the fact that aquaculture production is relatively new and quite low compared to sectors such as agriculture and livestock. The recent emergence of aquaculture production causes it to be more regulated by law. The fact that agriculture and livestock are carried out freely in every region gives it a little more freedom. However, the fact that a major part of aquaculture production is carried out in certain regions and with permission may have prevented it from increasing its ECF. Nevertheless, with the increase in production in this sector, the coefficient is expected to turn positive over time if protective measures are not taken.

Although fertilizer use was found to be insignificant in the analysis, there are studies in the literature that found that fertilizer use increases water and carbon dioxide footprints (Haverkort et al., 2014). In addition, the agricultural area is determined as the variable that affects the ECF the most in Model 2 in terms of the coefficient value. A 1% increment in the agricultural area raises the ECF by 0.948%. While the literature is examined, there are studies that find the ECF of agricultural land as the variable that increases the most in food production (Zhen & Du, 2017). When examining Fig. 1, it is seen that agricultural lands are the second area that induces the most ECF, and this supports the results of the analysis.

Consistent with the outcome of the analysis, income per capita expands the ECF in all four models in the long period. Accordingly, a 1% increment in income per capita, in the long period, expands the ECF by 0.439%, 0.304%, 0.447%, and 0.904%, respectively. There are studies in the literature that obtain similar results (Sabir and Gorus 2019). Apart from this, there are also several researches showing the negative results of economic growth on the environment (Rehman et al., 2022). According to neoclassical theory, the income elasticity of environmental protection is high, which means the demand on environmental sustainability is probably to raise as societies become increasingly wealthy. With the rise in prosperity, there emerges a shift in the demand for ecologically intensive resources, a trajectory that is anticipated to stabilize and eventually recede to sustainable thresholds owing to the escalating consciousness surrounding ecological imperatives. Consequently, the trajectory of prosperous advancement and overall developmental strides is poised to herald a potential mitigation of environmental harm and foster a diminution in the ECF over time (Clark et al., 2018). The fact that such a serious ECF is experienced with the increase in production shows that such a concern does not arise with the increase in income for Turkey. Moreover, according to the treadmill of production theory, capitalist systems always encourage economic growth but do not even consider slowing down the treadmill in the name of environmental sustainability (Obach, 2007). The rapid increase in food production in every sector despite its environmental impacts indicates that environmental sustainability is not at the forefront for Turkey. In general, developed countries and more advantaged groups within the country live in better-quality environmental conditions and are less exposed to pollutants than others (OECD, 2018). The reality is that Turkey is still a developing country and continuously encouraging production in order to move to the next level causes environmental degradation to be ignored.

In the analysis conducted pursuant to the existing analysis, it was found that both the total population and the rural population increase the ECF in three models. However, only in Model 2, the rural population diminishes the ECF. Accordingly, an increment of 1% in the population raises the ECF by 0.606% and 1.322%, respectively. Moreover, an increment of 1% in the rural population raises the ECF by 0.556%, 0.777%, and 0.771%, respectively, while it decreases it by 0.527% in Model 2. When the literature is investigated, it has occurred that the relationship has not been clearly revealed. While there are papers that discovered a positive relation between ECF and population (Udemba, 2021b), there are papers likewise that find a contradictory relationship (Dogan et al., 2020). However, there are studies that state while the total population increases the ECF, the rural population decreases it (Rehman et al., 2022). The mounting populace exerts a proportional upsurge in the requisition and utilization of natural resources, encompassing vital commodities like energy and sustenance. By the tenets of the ECF paradigm, this upswing delineates an amplification in the burden borne by burgeoning populations, culminating in the erosion of environmental quality and the exacerbation of ecological strain (Udemba, 2021b).

Finally, an increase of 1% in the age dependency ratio used in model 4 decreases the ECF by 1.415%, while life expectancy at birth increases by 2.551%. There are papers in the literature indicating that the aging of the population increases environmental pollution (Marquart-Pyatt, 2015; Li et al. 2019). However, there are also articles in the literature stating that the working population causes more environmental pollution (Dimnwobi et al., 2021). This variable may produce different results in different income groups (Wang et al., 2022). The contradictory relation can be detailed accordingly the young population, especially those with higher income, consumes more, is more active in social life, and easily directs their consumption by following new trends/fashion concepts. In addition, increasing life expectancy leads to increasing depletion of the world’s natural resources, while simultaneously increasing the human populace and reducing biodiversity (Samreen & Majeed, 2022). In the literature, studies suggest that life expectancy at birth increases the ECF (Boukhelkhal, 2022).

Then, some diagnostic tests were used to demonstrate the suitability of the models. Table 7 presents the diagnostic test outcomes for the long-run analysis.

Table 7 Diagnostic test results
Full size table

The suite of diagnostic evaluations—comprising the Jarque–Bera (J-B) test for normality, the Breusch-Godfrey (B-G) Serial Correlation LM test, the Breusch-Pagan-Godfrey (B-P-G) test for heteroscedasticity, and the Ramsey RESET for model specification—collectively ascertains that the model stands free by issues of these problems. Figure 2 elucidates the findings from the CUSUM and CUSUMQ tests, as presented by Brown et al. (1975), to scrupulously assess the model’s stability. The fluctuations within the confines of established critical boundaries substantiate the stability of the model’s coefficients in the long period and affirm their reliability at a 5% level of significance.

Fig. 2
figure 2

CUSUM and CUSUMQ test results of the models

Full size image

Subsequently deciding the long period relation, the short period relationship is determined by the error terms and the difference values of the series. The short-run models are as follows:

$$\Delta {\text{logef}}_{{\text{t}}} = \upbeta_{0} + \upbeta_{1} {\text{Eq}}_{{\text{t - 1}}} + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{2i}}}} \Delta {\text{logfood}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{3i} \Delta {\text{loggdp}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{4i}}}} \Delta {\text{logrural}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{5i}}}} \Delta {\text{logpop}}_{{\text{t - i}}} } + {\text{u}}_{{\text{t}}}$$
(9)
$$\begin{aligned} \Delta {\text{logef}}_{t} = & \upbeta_{0} + \upbeta_{1} {\text{Eq}}_{{\text{t - 1}}} + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{2i}}}} {\text{logcrop}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{3i}}}} \Delta {\text{logfert}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{4i}}}} \Delta {\text{loggdp}}_{{\text{t - i}}} } \\ & + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{5i}}}} \Delta {\text{logland}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{6i}}}} \Delta {\text{logrural}}_{{\text{t - i}}} } + {\text{u}}_{{\text{t}}} \\ \end{aligned}$$
(10)
$$\Delta {\text{logef}}_{{\text{t}}} = \upbeta_{0} + \upbeta_{1} {\text{Eq}}_{{\text{t - 1}}} + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{2i} {\text{loglive}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{3i} {\text{loggdp}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{4i}}}} {\text{logrural}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{5i}}}} {\text{logpop}}_{{\text{t - i}}} } + {\text{u}}_{{\text{t}}}$$
(11)
$$\begin{aligned} \Delta {\text{logef}}_{{\text{t}}} = & \upbeta_{0} + \upbeta_{1} {\text{Eq}}_{{\text{t - 1}}} + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{2i} {\text{logaqua}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\beta_{3i} {\text{loggdp}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{4i} {\text{logrural}}_{{\text{t - i}}} } \\ & + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{{{\text{5i}}}} {\text{logage}}_{{\text{t - i}}} } + \sum\limits_{{\text{(i = 1)}}}^{{\text{m}}} {\upbeta_{6i} {\text{loglife}}_{{\text{t - i}}} } + {\text{u}}_{{\text{t}}} \\ \end{aligned}$$
(12)

The term Eq. 1 represents the lagged value, specifically a one-period retardation, of the error series derived consequent to establishing the long-term relations, as detailed in Table 6. This variable quantifies the extent to which deviations from equilibrium in the short period are expected to realign within the long period trajectory. The coefficient of Eq. 1 is theoretically posited to bear significance and exhibit a negative sign, signifying the corrective convergence mechanism at play. An explication of the short period coefficient determines, alongside the coefficients of the error correction model, is presented in Table 8.

Table 8 Error correction models results
Full size table

The error term coefficients are determined to be negative and significant, in accordance with expectations. This result reveals that the deviation from equilibrium that may occur in the period under consideration is fixed in the following period. The coefficients of the error correction terms are − 1.434, − 1.753, − 1.154, and − 1.097 in the four models, respectively. This denotes that the error correction trajectory does not align in a monotonic fashion directly towards the equilibrium state, but instead exhibits oscillatory behavior that asymptotically dampens around the long period equilibrium. Upon the culmination of this adjustment process, convergence towards the equilibrium path is characterized by an accelerated pace (Narayan & Smyth, 2006).

5 Conclusion and policy recommendation

Reducing the ECF and governing the ecosystem remain one of the most important challenges before us in the next century to progress to sustainable development. In addition, the loss of biodiversity threatens food security. Therefore, it is very important to transform food production (WWF, 2020). Turkey, which is not a country rich in water resources, should transition to environmentally compatible good agricultural practices for the sustainability of its agricultural policies. In addition, Turkey’s population is constantly increasing due to both the migration wave and the natural birth rate. The increase in consumption that will come with this population will increase the ECF. In order to prevent this situation, large masses of people should be educated quickly and effectively, penalties for environmental crimes should be increased, and inspections should be tightened. Although there are laws for the protection of the environment in Turkey, the implementation and supervision of these laws are not carried out regularly enough. The real problem stems from this, not from the absence of laws. Therefore, strict inspections must be carried out to ensure that all sectors working in food production work in accordance with existing laws.

In Turkey, it is crucial to increase agricultural production efficiency while simultaneously shifting to eco-friendly and low-carbon production systems. Encouraging the consume of clean renewable energies for instance wind, solar, and biofuels, modernizing farmers’ irrigation systems, increasing organic farming, encouraging the use of solar tube wells for tunnel farming, shifting from conventional tillage to no-till farming, and minimizing fertilizer use are important for ensuring environmental protection (Raihan & Tuspekova, 2022). The government should encourage more unpolluted and low-carbon technology sources in the context of agricultural production, the integration of renewable energy sources includes solar, geothermal, biomass, biogas, and tidal, alongside the adoption of photovoltaic systems and wind turbines, each contributing to a diversified and sustainable energy matrix within the sector (Muoneke et al., 2022).

Enhancements in manure management practices are imperative throughout the continuum of handling, storage, processing, and its eventual application to agricultural fields. To avert the infiltration of manure into groundwater and surface water bodies, insulation of storage areas and silo foundations must be meticulously reinforced. The strategic redirection of manure from animal husbandry towards biogas generation promises dual dividends: an augmentation in energy resources and a commendable mitigation of nitrogenous emissions through the transformation of manure’s methane content into a viable source of energy (Yayli and Kiliç 2021). While bio-pesticides can be applied for biodiversity conservation, it is necessary to promote agroecological principles such as appropriate crop rotation, diversification of agricultural practices, ensuring optimum stocking rate for production, and promoting agroecological principles that can be efficient (Banerjee et al., 2021).

The unsystematic overgrazing of grasslands used for livestock raising has a serious impact on biodiversity. Therefore, government actions such as grassland monitoring and assessment, grassland restoration and management, and grassland protection are crucial to protect these areas. In addition, forests need to be protected both for biodiversity conservation and to prevent the ECF from increasing with the use of these areas for activities such as agriculture and livestock.

The fact that aquaculture production reduces ECF is a very pleasing indicator for Turkey. This proves that food production can be carried out without harming the environment if clean production is practiced. Therefore, good examples in aquaculture production should be extended to other areas. However, aquaculture is easier to control as it is produced in a more limited way in certain regions of the country. On the other hand, agriculture and animal husbandry are spread throughout the country and constitute the livelihood of a large part of the rural population. Therefore, it is not as easy to regulate and control as aquaculture production. Newly developed tools such as the farmer registration system will provide convenience in this respect. However, it does not seem possible to abandon old methods such as wild irrigation at once. In this respect, it is important to increase the frequency and spread of training activities carried out with local development agencies. Teaching modern methods, especially in irrigation and waste disposal, and forcing them to be applied when resistance is encountered may be a solution method.

Importing food products can also be a way of diminishing the domestic ECF. However, it is important to ensure sustainability and not jeopardize food security. Imports without sustainable resources reduce the resources allocated to production investment within the country over time and reduce the possibility of substitution against sudden import shocks. This can lead to food supply shortages in emergencies, triggering food inflation and causing serious economic and political consequences for the country.

In all models, GDP is found to increase the ECF. This situation shows that Turkey does not take environmental degradation into account sufficiently during its growth process. Turkey, a developing country, currently ignores environmental degradation in order to achieve a high growth rate. However, the consequences of these policies will emerge more harshly in the future. Therefore, it is going to be essential to limit environmental degradation while ensuring rapid growth.

This study has several limitations. First of all, it is quite hard to detect data for Turkey starting from the 1960s. Therefore, the current analysis was conducted with a very limited data set. In future studies, shortening the data set and repeating it with new variables will reveal the effects of different factors. Additionally, it is quite difficult to find specific data on the subject. For this reason, explanatory variables used in similar articles could not be used. Moreover, the limited number of studies similar to the current analysis of Turkey also reduces the possibility of comparison. In future studies, investigating the issue with methods other than ARDL will provide the opportunity for comparison.