Grand Challenges in Central Europe: The Relationship of Food Security, Climate Change, and Energy Use

: Pursuing various sustainable development goals is posing new challenges for societies, policymakers, and researchers alike. This study implements an exploratory approach to address the complexity of food security and nuance its relationship with other grand challenges, such as energy use and climate change, in Central European countries. A multiple factor analysis (MFA) suggests that the three pillars of food security relate di ﬀ erently to climate change: food a ﬀ ordability and food accessibility positively correlate with climate change, while food quality has a negative association with temperature rise. However, if countries switched to renewable energy resources, all three pillars of food security could be achieved simultaneously. The study also underlines regional inequalities regarding grand challenges and emphasizes the need for innovative local solutions, i.e., advances in agriculture systems, educational programs, and the development of environmental technologies that consider social and economic issues.


Introduction
The National Academies of Science, Engineering and Medicine (NASEM) have raised pressing global and grand challenges that environmental scientists and engineers have uniquely put together for advanced support [1]. The report highlights issues regarding the sustainability of food, fossil energy, and renewable (solar, water, biomass) resources; climate change control; pollution and waste; efficient, healthy and resilient cities; as well as informed decisions and policies. However, future goals of sustainable development face several hindrances due to the interaction of complex socio-economic issues.
Several studies attempt to explore the interactions and find the optimal balance among crucial social, economic, and environmental areas influencing the future of humanity [2][3][4]. Academics have long been analyzing new scenarios to explore how the world may change in the rest of the 21st century. In the 2000s, a promising effort was made to predict the trajectories of population change, to more nuanced findings on how key SDGs could be solved at both a regional and global level. We rely on data collected in Central European (CE) countries between 2012 and 2018. A multiple factor analysis (MFA) is applied to compute the relationship among blocks of variables. The advantage of MFA is calculating correlations between the indices in each pillar while also taking regional disparities within CE into account.

Data and Variables
Several systematically collected groups of indicators attempt to measure the effects of human activity on the state of the planet. In this paper, groups of variables related to climate change, energy use, and food security, including food affordability (FAF), food accessibility (FAC), and food quality (FQ), were carefully selected. Table 1 presents the variables and their descriptions for CE countries. CE covers territories based on collective historical, social, and cultural identity [25]. However, it is often divided into West-Central Europe and East-Central Europe [26]. The latter contains the Visegrád Group (V4), countries that are strategic alliance partners and strongly integrated economic counterparts. The following countries were analyzed: Austria, Belgium, Germany, the Netherlands, Switzerland, and the V4 ones (the Czech Republic, Hungary, Poland, and Slovakia). The climate change pillar contains various World Development Indicators (WDI) from the World Bank Dataset [27] and the Harmonized World Soil Database (HWSD) [28], collected by the Food and Agricultural Organization (FAO). The pillar includes measures of air pollution (PM2.5 mean annual exposure), greenhouse gas emissions (CO 2 grass-and cropland), environment-related loss of forest area and soil erosion. Temperature rise, which is also essential to measure a country's vulnerability to climate change, was added here from the Global Food Security Index (GFSI) of the Economist Intelligence Units (EIU) Database [29] and standardized to the extent possible to facilitate cross-country comparisons.
The energy usage pillar involves the energy intensity level of primary (fossil) energy, which is the ratio between energy supply and Gross Domestic Product (GDP) measured at purchasing power parity (PPP) [27]. The energy output and consumption before transformation to similar end-user fuels (renewable electricity and refined petroleum products) and other energy indices were included in this pillar from the Eurostat Energy Database [30]-namely, final energy from combustible renewables and waste, such as solid biomass and animal products, gas and liquid from biomass, and municipal waste.
Data for food security indicators were drawn mainly from the European Intelligence Unit [29] and the Food and Agricultural Organization (FAO) [31]. Food security is defined here as the state in which people at all times have physical, social, and economic access to sufficient and nutritious food that meets their dietary needs for a healthy and active life [32]. This paper uses GFSI and FAO indicators, which are the most widely used measurements of food security at the national level. To improve the transparency and validity of the pillars, insecurity indicators are added from the World Bank Global Consumption Dataset [33] and the Tariff Online Facility of World Trade Organization (WTO) [34].
FAF measures the capacity and costs of a country's people to pay for food under normal circumstances and at times of food-related shocks. For instance, GDP per capita and food consumption expenditures of consumers to purchase food. Thus, FAF contains the agricultural import tariffs and food dependence control vulnerability to external price shocks. FAC influences the supply and the ease of access to food. FAC denotes the sufficiency of the national food supply, the risk of supply disruption, the agricultural infrastructure to expand agricultural output, the local and innovation capacity to reduce food loss as well as political stability. Finally, FQ contains the variety, nutritional quality, and availability of average diets. This category is sometimes referred to as 'utilization' because it explores the energy and nutrient intake by individuals and the diversity of the diet [31].
The selected pillars also reveal the progress of the environmental targets set by the 2030 Agenda for Sustainable Development [35]. SDGs require efforts to promote renewed policies and approaches regarding the great challenges. For example, SDG 2 relates to 'End hunger, achieve food security and improved nutrition and promote sustainable agriculture' [36]. The zero-hunger challenge supports the prevention of children's stunted growth (under the age of two), full access to adequate food all year round, sustainable food production systems, increase in smallholder productivity and income, and calling for zero loss or a waste of foods. SDG 7.1 promotes ensuring access to affordable, reliable, and renewable energy resources and services, SDG 11.6 contributes to reducing the adverse environmental effect of urban development, SDG 14.5 supports to conserving coastal and maritime areas, and SDG 15.5 aims to protect and prevent the extinction of threatened species.

Statistical Analysis
MFA was first introduced by Thurstone [37] and later described as well by Escoffier and Pagès [38]. The method is useful for analyzing a group of inter-correlating variables divided into blocks. In the first stage, a traditional principal component analysis (PCA) was performed on each variable block, and in the second stage, the blocks were made comparable through normalization by using the square root of the first eigenvalue obtained from the separate PCAs [39]. In the final stage of the analysis, a global PCA was performed on the normalized blocks of variables. Observations are represented in a lower (usually two) dimensional map; the coordinates are called factor scores. MFA provides a unique Energies 2020, 13, 5422 5 of 16 concept of partial factor scores which make it possible to position each observation by taking different groups of variables into account.
Besides balancing variable groups, MFA provides results specific to the group structure of the set of variables. For instance, the detailed groups of variables can give synthetic images and factors from separate analyses. MFA is also suitable for graphically displaying observations and their relations and hence building diverse clusters [40]. MFA is helpful to analyze different types of observations described by various groups of variables, and the method is even more valuable when the data set is large and complex [41]. It derives an integrated representation of the remarks and the relationships among groups of examined variables. The analysis was performed by FactoMiner, an R software package [42] for multivariate analysis [43].

Results
An MFA was conducted using the variables in Table 1. MFA first computes a series of PCAs. The relationship of the components with the global analysis was explored by computing loadings (correlations) between the components of each pillar (climate change, energy resources, and food security) and the global analysis. In this study, the analysis consists of five (T) datasets called blocks. Each block is a (I × J[t]) rectangular data matrix denoted by Y[t], where I is the number of observations and J[t] is the number of variables of the t-th block. Each data matrix is pre-processed (centered and normalized) and denoted by X[t]. Each observation is assigned a 'mass' which reflects its importance and is stored in an I × I diagonal matrix (M). The normalized blocks are concatenated into an I × T matrix called the global data matrix (Z) [38].
A standard PCA was used to estimate the singular value decomposition of the global data Z matrix: where U and V are the left and right singular vectors of Z, respectively, and ∆ is the diagonal matrix of the singular values. The global (F) factor scores are obtained as: where each row represents an observation, and each column is a component (dimension). The eigenvalue of the first dimension (DIM1) corresponds to 37.8% of the inertia ( Figure 1a). DIM1 is associated with air pollution, dietary diversity, port infrastructure, and GDP per capita (productivity). The second dimension (DIM2) corresponds to 13.9% of the inertia. DIM2 relatively strongly correlates with renewable energy resources, temperature rise, dietary diversity, and expenditures on agricultural R&D ( Figure 1b).

Validation of Results
The validation of MFA components was carried out using several methods ( Table 2): random sampling with replacement (bootstrapping), a permutation test within each block to preserve exchangeability, and exhaustive (leave-one-out) [44] and non-exhaustive (split-half) estimation [45]. These cross-validation (rotation estimation) techniques are crucial for assessing the accuracy of any predictive model in practice [46].
where each row represents an observation, and each column is a component (dimension). The eigenvalue of the first dimension (DIM1) corresponds to 37.8% of the inertia (Figure 1a). DIM1 is associated with air pollution, dietary diversity, port infrastructure, and GDP per capita (productivity). The second dimension (DIM2) corresponds to 13.9% of the inertia. DIM2 relatively strongly correlates with renewable energy resources, temperature rise, dietary diversity, and expenditures on agricultural R&D (Figure 1b).

Validation of Results
The validation of MFA components was carried out using several methods ( Table 2): random sampling with replacement (bootstrapping), a permutation test within each block to preserve  First, a bootstrap simulation was performed (N = 1000) with repetition for all indicators within each iteration. In the case of permutation tests, the permutation of the objects was applied within each block separately, resulting in a random dataset. The critical values were evaluated by the distribution of the explained variances of the MFA components. The two-sided p-values were calculated from the distributions. The null hypothesis of the test concerns to the initially explained variances are not different from the simulated ones. Both results show that the first and second components (DIM1 and DIM2) explain the largest proportion of the variance and are stable.
Regarding the split-half test, the dataset was halved into equal parts and a separate MFA was performed on each part. This process was repeated N times; the explained variance of MFA components was recorded and compared with one another by a Wilcoxon signed-rank sample test [47]. The lack of significance indicates no statistical difference between the two splits, and the second component is the most stable. In the leave-one-out cross-validation test, the MFA was executed nine times (there are nine countries in the dataset, and one country was left out each time from the analysis) to estimate the variation coefficients, i.e., the standard deviation divided by the mean. The value of the coefficient did not exceed the critical 20% in any case. Based on the results, the selected first two components were proved to be stable during validation tests and also explain a sufficient amount of variance.

Global Space and Partial Analysis
The global analysis reveals the common structure of the examined space regarding the pillars (groups of variables) in the model. It shows the relationship between the variables projecting the data set for global analysis. The projection matrix (Figure 2) contains the contributions of global factor scores (percentages). The global space map can be divided into four quadrants based on the two dimensions specified by the MFA. Each quadrant contains countries with different food security, climate, and energy profiles. The first quadrant (i.e., Austria, Germany, and Switzerland) can be characterized by a high level of renewable energy output and consumption, GDP per capita, food supply, and diet diversity. Germany, loading closer to the centroid, is associated with a high percentage of forest area and final energy consumption. The V4-country group forms a cluster on the other pole of the first dimension, i.e., in the second and third quadrants. The second quadrant contains Hungary, which is more strongly associated with soil erosion and temperature rise than the other CE countries. Hungary is dominated by agricultural landscapes in the plains and intensive agricultural, especially maize production, which can be damaged by high temperatures and low rainfall. Slovakia is on the border of the second and third quadrants, described by relatively high urban absorption compared to Poland and the Czech Republic due to its topography. In Slovakia, as mountain forests cover most of the The global space map can be divided into four quadrants based on the two dimensions specified by the MFA. Each quadrant contains countries with different food security, climate, and energy profiles. The first quadrant (i.e., Austria, Germany, and Switzerland) can be characterized by a high level of renewable energy output and consumption, GDP per capita, food supply, and diet diversity. Germany, loading closer to the centroid, is associated with a high percentage of forest area and final energy consumption. The V4-country group forms a cluster on the other pole of the first dimension, i.e., in the second and third quadrants. The second quadrant contains Hungary, which is more strongly Energies 2020, 13, 5422 8 of 16 associated with soil erosion and temperature rise than the other CE countries. Hungary is dominated by agricultural landscapes in the plains and intensive agricultural, especially maize production, which can be damaged by high temperatures and low rainfall. Slovakia is on the border of the second and third quadrants, described by relatively high urban absorption compared to Poland and the Czech Republic due to its topography. In Slovakia, as mountain forests cover most of the country, the agricultural production is limited by the availability of arable land. The third quadrant (Poland and the Czech Republic) is characterized by relatively high food consumption, improved energy intensity level, and political stability. Finally, the fourth quadrant (the Netherlands and Belgium) can be described by more intense public expenditure in agricultural R&D, and better food (protein) quality. Figure 2 suggests a link between the consumption of energy resources, climate change, and food security issues. For instance, the energy intensity level of primary resources (e.g., energy supply per GDP) is associated with the climate change pillar (high air pollution, soil erosion, and temperature rise), high FAF (i.e., food consumption and import dependency), and high FAC (i.e., population growth, political stability, and urban absorption). At the other pole of DIM1, GDP per capita, as a FAF indicator, is also related to better food accessibility (i.e., higher expenditure in agriculture R&D, food supply, and port infrastructure) and improved food quality (FQ) (i.e., protein quality and dietary diversity). Regarding the second dimension (DIM2), temperature rise seems to correlate with renewable energy consumption and electricity output negatively. Thus, the final energy consumption from biomass and renewables is negatively associated with GHG emissions and forest land changes. Figure 3 shows each CE country as a single point based on the first two principal components of the global analysis. The partial points position each country from the perspective of the five different pillars. Hence, the individual and regional differences, as well as the overall status of CE countries, can be evaluated.  The V4 group seems to be more strongly affected by energy usage and climate change than by food security issues. Lower levels of renewable energy consumption and electricity generation can also be observed in all V4 countries. According to the partial findings, Hungary is the most exposed to the effects of climate change (especially rising temperatures). In Slovakia and Poland, on the other hand, food accessibility (urban absorption) is more distinctive. In terms of FAF, Hungary is in the worst situation, while the Czech Republic is facing major FQ issues (dietary diversity and protein quality).
In the second dimension, Austria, Germany, and Switzerland are described by high renewable energy consumption and electricity output. The Netherlands, one of the drivers of environmental protection and sustainable food policy in the EU [48], stands out for its lower risk of climate change The V4 group seems to be more strongly affected by energy usage and climate change than by food security issues. Lower levels of renewable energy consumption and electricity generation can also be observed in all V4 countries. According to the partial findings, Hungary is the most exposed to the effects of climate change (especially rising temperatures). In Slovakia and Poland, on the other hand, Energies 2020, 13, 5422 9 of 16 food accessibility (urban absorption) is more distinctive. In terms of FAF, Hungary is in the worst situation, while the Czech Republic is facing major FQ issues (dietary diversity and protein quality).
In the second dimension, Austria, Germany, and Switzerland are described by high renewable energy consumption and electricity output. The Netherlands, one of the drivers of environmental protection and sustainable food policy in the EU [48], stands out for its lower risk of climate change and higher food protein quality and dietary diversity.

Discussion
A key finding of the MFA is that food security can be sustained if policy measures support a stable macro-environment [49], including improved food supply, better infrastructure, and less import dependency as a precursor to higher productivity. The sustainable food economy can contribute to economic growth by increasing investments (physical and human capital) and reducing macro (price and political) instability [50]. On the other hand, stability supports poverty alleviation by reducing vulnerability to sudden shocks of food prices or food availability [51]. A crucial outcome of a rural-focused regional growth policy is to achieve food security when economic growth has raised the poor above the direct poverty line and to stabilize food prices and stocks and prevent exogenous shocks [52].
Besides, food intake plays a crucial role in productivity growth, supported by better protein quality, dietary diversity, and health in the human body [53]. The findings of the current study are consistent with those of Deolalikar [54], who found that adequate childhood nutrition can also improve educational attainment and economic growth per capita. Population growth as a socio-economic indicator is positively associated with food consumption, which supports the idea that an increased and healthier food consumption reduces mortality rates [55] and may upsurge life expectancy [56]. However, in developed countries, increased food consumption is a major cause of obesity and can reduce the expected lifetime [57].
The global MFA revealed that climate change is negatively related to dietary availability and protein quality. Interestingly, the results suggest that global warming may adversely affect water availability, crop yields, and reduce food and environmental quality in the future [58]. However, better FAC is driven by population growth, urbanization, and industrialization, and increased food consumption is a challenge to sustain the required level of food quality. The findings of the present study are consistent with Khan et al. [59], who reviewed water management and crop production in China and pointed to the need to integrate climate, energy, food, environment, and population considerations. Hepperly [60] argued that increased agricultural production leads to higher energy consumption, especially fossil fuels, and thus to higher levels of greenhouse gas emissions, pollution, deforestation, and deteriorating water and food quality.
The extent of deliberate hydrological changes in Central Europe has been enormous, from small streams and ditches to the largest rivers. However, the current droughts are not only due to climate change. The long-term frequency of extreme droughts has been stable and severe in the last two or three centuries, which were also typical in the coastal part of Poland and Silesia [61]. Another important aspect is the change in plant water conditions due to the partial closure of the stomata. The impact of transient water scarcity and shortage with increasing CO 2 concentration leads to improved water-use efficiency (WUE) and less water consumption of plants due to limited transpiration [62]. In addition, the complex effect of climate change (not just rainfall) significantly limits the growth of timber (oak and pine tree stands) in Central Europe [63]. Other trees are sensitive to extreme weather events and environmental conditions, such as heatwaves [64] and late spring frost [65]. The effect of long-term climate change, extreme frosts, and sub-optimal temperatures on the earlier occurrence of flushing and flowering fruit (wine) stages [66] and on the phenology of vegetables, for instance, potato production [67], is transforming the distribution of food supply and waste in Central Europe undesirably. The Global Target for Sustainable Energy 2030 (Development Goal 7) is to increase the share of renewable energy sources in global energy production [68] and double the global rate of energy efficiency improvement [69]. Eco-efficiency leads to sustainability, which is closely linked to the separation of energy and material intensity, the prevention and management of food waste and air pollution, recycling, the widespread use of renewable energy, as well as the enhancement of the product life cycle and consolidation [70]. The findings of our study confirm the inverse relationship between (a) renewable energy consumption and electricity generation and (b) vulnerability to temperature rise and CO 2 emissions. Mouratiadou [71] also examined the effects of irrigation of bioenergy crops based on shared socioeconomic pathways (SSPs) and found it to be the most critical factor leading to significantly higher water demand due to climate change mitigation.
Climate change and local air pollution seem to be crucial to address regional energy policy issues in the case of CE countries. Our findings confirm that energy transformation could significantly reduce greenhouse gas emissions while ensuring sufficient energy is available through greater energy efficiency and a gradually increasing share of renewable energy sources [72]. Improving decarbonization in energy-intensive industries introduces efficiency (electrification) measures and savings from renewable energy technologies [73]. Agroforestry systems provide carbon sequestration and essential wood products against the greenhouse effect [74]. Furthermore, modern biomass heating applications and liquid biofuels are expected to double from the current energy supply level by 2050 [75]. Liu et al. [76] advocate the positive effects of biofuels on final total greenhouse gas emissions and show their advantages over fossil fuels. The production of biofuel-cellulosic biomass can also reduce crude oil consumption and pollution from fossil fuels [62].
However, air pollution is not limited to energy production and is closely linked to the food system. Along the food supply chain, agricultural production, processing, and distribution generate significant pollutants [77]. The excessive use of chemical fertilizers and animal husbandry are the primary sources of agricultural emissions [78]. In addition, the exhaust fumes from industrial waste and vehicles (aircraft, trucks) related to the food system contribute expressively to air pollution [79]. Air pollution not only reduces the supply of raw ingredients but can affect consumer demand and choice [80], and food supply will eventually change food prices [81].

Conclusions
The main objective of this study was to examine the interrelations of climate change, energy use, and food security, i.e., food affordability, accessibility, and quality, to shed light on novel research perspectives on grand challenges and Sustainable Development Goals (SDGs). A multiple factor analysis (MFA) was used to calculate correlations between the aforementioned pillars while also taking regional disparities within CE countries into account. The advantage of MFA is that it analyzes different types of observations described by various groups of variables, and the method is even more valuable when the data set is large and complex.
Contrary to previous approaches, we consider the complexity of food security, which is necessary for exploring the subtle interconnections of socio-economic and environmental issues. We found that the different pillars of food security are likely to have different impact on other sustainable development goals, namely: (a) higher FAF and FAC are associated with climate change and higher energy intensity level of primary resources; (b) better FQ couples with lower temperature rise and higher output per capita; (c) temperature rise is adversely related to renewable energy consumption and electricity output. At the regional level, we also found that (d) the V4 group seems to be relatively more affected by energy usage and climate change than by food security issues; (e) Austria, Germany, and Switzerland make better use of renewable energy; (f) the Netherlands stands out for its lower risk of climate change and higher food protein quality and dietary diversity.
The methodological implication is that researchers need to consider the various dimensions of complex sustainable development phenomena when they endeavor to measure and examine the impacts of grand challenges. The findings are also important for policymakers in Central Europe and globally, since the demand for safe and clean food and sustainable energy grows with a rising population worldwide.
Creative solutions are needed to maximize energy, food, and water supplies while reducing adverse climate impacts. For example, a unique concept of solar spectrum unbundling in food, energy, and water systems (SUFEWS) proposes to maximize crop production while simultaneously producing energy and managing waste supply by separating the solar spectrum on a plot of land. Gençer et al. [82] suggested that reflective parabolic troughs can be located above the field to accumulate solar energy from near-infrared and far-infrared light waves, while the desired solar spectrum for food production is passed to plants on the ground. Near-infrared light can be used to generate energy, and near-far and far-infrared power supply distillation or reverse osmosis is supported by hydropower treatment processes. Besides, electricity generated by solar panels can contribute to sustainable agricultural production or be exported to nearby residents.
Progress in agriculture systems is desirable, including crop and alternative food production methods, reduction of food waste, and changes in dietary diversity [83]. Optimizing productivity by managing irrigation is essential in, for example, paddy areas and can considerably reduce global methane and nitrous oxide emissions [84]. In addition, the usage of efficient agrochemicals, improved pest and disease forecasting, the adoption of modern wastewater treatment, the optimization of animal feed, and the improvement of the livestock environment should be assessed in order to improve the impact of agricultural and energy policies on food security and climate change mitigation [85][86][87].
Changing the diet can also be an effective tool to reduce air pollution due to the shift from animal husbandry to crop production, as raising ruminants has a greater impact on air pollution [88]. Meanwhile, meat-free protein products are diversifying, including innovative plant-based products that can be grown from animal and plant tissue cells in culture. Such products significantly increase the affordability of food and, if accepted by consumers, can reduce the constant demand for animals [89], thereby reducing the soil, energy, and water needs of animal proteins [90] and the associated adverse effects of environmental and climate change, while increasing the availability of nutrients [91].
This study calls for the improvement of educational outcomes related to sustainable food, climate change, and energy challenges in practice. Environmental education programs should rely on the integration of strict energy-saving system maintenance, big data science, and decision analysis, and redesign them into sustainable engineering projects. Therefore, the cooperation of scholars and social science experts is essential to understand the social, cultural, economic, regional, and political contexts of environmental challenges [1]. The curricula of engineering programs could be enriched with topics related to grand challenges and their interrelations, in addition to traditional focus areas, such as climate, energy, and air pollution.
There are some limitations to this study related to the method and variables chosen. The most important is omitted-variable bias, as the studied pillar variables reflect the authors' partly subjective choice. In general, sampling error may have a stronger bias on PCA results than on the correlations among observations [92]. Another potential problem is that the scope of the study appears to be broad; however, the inclusion of several indices is required to map the subtle interactions of seemingly unrelated development goals.
To conclude, exploring the effects and linkages of grand challenges should not rely on simplistic research questions [93]. Goals related to sustainable development are difficult to delineate and implement simultaneously due to their complexity and interdependence. Hence, future research has to be carried out across different disciplines and theories (e.g., economic growth, circular economy, sustainability, resource management, and climate theories) to retrace and develop indicators that reflect the status quo of our planet and potential trajectories. For example, researchers may consider the impacts of government incentives, community development, and environmental activities. The integration of sustainable economic growth, population growth, crop production, climate change, energy use, and water supply analyses is essential for a more comprehensive and reliable assessment of the food security pillars.