Measuring the impact of global tropospheric ozone, carbon dioxide and sulfur dioxide concentrations on biodiversity loss

Environmental Research 160 (2018) 398–411

Contents lists available at ScienceDirect

Environmental Research journal homepage: www.elsevier.com/locate/envres

Measuring the impact of global tropospheric ozone, carbon dioxide and sulfur dioxide concentrations on biodiversity loss

MARK

⁎

Miraj Ahmed Bhuiyana, Haroon Ur Rashid Khanb, Khalid Zamanc, , Sanil S. Hishand a

School of Management, Wuhan University of Technology, Wuhan, Hubei, China School of Finance, College of Business and Public Management, Kean University, NJ, USA, Wenzhou-Kean University Campus, Wenzhou, PR China. c Department of Economics, University of Wah, Quaid Avenue, Wah Cantt, Pakistan d Faculty of Management, Universiti Teknologi Malaysia, 81310 Skudai, Johor, Malaysia b

A R T I C L E I N F O

A B S T R A C T

Keywords: Air pollutants Biological diversity Ecological footprint Panel cointegrating regressions

The aim of this study is to examine the impact of air pollutants, including mono-nitrogen oxides (NOx), nitrous oxide (N2O), sulfur dioxide (SO2), carbon dioxide emissions (CO2), and greenhouse gas (GHG) emissions on ecological footprint, habitat area, food supply, and biodiversity in a panel of thirty-four developed and developing countries, over the period of 1995–2014. The results reveal that NOx and SO2 emissions both have a negative relationship with ecological footprints, while N2O emission and real GDP per capita have a direct relationship with ecological footprints. NOx has a positive relationship with forest area, per capita food supply and biological diversity while CO2 emission and GHG emission have a negative impact on food production. N2O has a positive impact on forest area and biodiversity, while SO2 emissions have a negative relationship with them. SO2 emission has a direct relationship with per capita food production, while GDP per capita significantly affected per capita food production and food supply variability across countries. The overall results reveal that SO2, CO2, and GHG emissions affected potential habitat area, while SO2 and GHG emissions affected the biodiversity index. Trade liberalization policies considerably affected the potential habitat area and biological diversity in a panel of countries.

1. Introduction The ecosystem and biological diversity loss are the two most crucial challenges that faced by the planet due to rapid industrialization, trade liberalization policies, economic gains, massive population growth, etc. These factors severely damaged the tropospheric ozone that latterly influenced human health, ecosystems and biodiversity loss. One of the biggest discussions is going on in “The International Cooperative Programme on Effects of Air Pollution on Natural Vegetation and Crops” that separating the mapping for South and North Europe (ICP, 2015). This study is limited to the ecosystems and biodiversity loss that merely influenced by the mono-nitrogen oxides (NOx), nitrous oxide (N2O), sulfur dioxide (SO2), carbon dioxide emissions (CO2), and greenhouse gas (GHG) emissions respectively. In addition, this study used two growth measures, i.e., real economic growth and trade openness that cumbersome the certain ecosystems, including ecological footprint, potential habitat area (proxy by the forest area), food production per capita variability, and food supply per capita variability. This study constructed the biodiversity index by taking the different ecosystems for measuring the single relative weighted principal ⁎

Corresponding author. E-mail address: [email protected] (K. Zaman).

http://dx.doi.org/10.1016/j.envres.2017.10.013 Received 18 July 2017; Received in revised form 7 October 2017; Accepted 8 October 2017 0013-9351/ © 2017 Elsevier Inc. All rights reserved.

component, that's we labeled the ‘biodiversity index’. In search of the literature, we received plenty of research work that is related to the scope of the study; however, this study is unique in a sense, as it is used diversified portfolio of ecological biodiversity, which is influenced by the numerous air pollutants and growth factors across the globe. Wilson (1989) addressed the ‘threats to the biodiversity’ and concluded that global climate change is the significant contributor to the biodiversity loss in Polar Regions. McNeely (1992) argued that air pollution damages the biodiversity and ecosystems that merely due to the rapid industrialization process, which proven the fact that it provides greater payoffs to the society and for the economic development; however, the cost in the form of ecosystems and biodiversity loss is hidden that should be measured by ‘Five–I’ approaches i.e., investigation, information, incentives, integration, and international support. This ‘Five–I’ approach would provide the ecological justice for balancing our natural flora. Myers (1993) suggested some precautionary measures, including biological, ecological, and economic measures that should be balanced to conserve biological diversity. Kappelle et al. (1999) investigated the possible effects of climatic changes on biodiversity loss and argued that this relationship is complex, as the

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

(2014) confirmed environmental Kuznets curve (EKC) in an aggregated panel of countries by using the consistent time period from 1980 to 2009. The results show that EKC is widely visible across countries except high income countries, which is further followed in the form of high pollution during the transformation process from industrial activities to services sector, while low and middle income countries shows less pollution during the sectoral transformation process. The sustainability agenda is all across the viable policy agenda for environmental reforms. Stevens et al. (2014) conducted the survey on the acid grasslands in the UK for assessing the possible impact of atmospheric nitrogen deposition on species richness and confirmed that atmospheric nitrogen deposition reduces the richness of the species of acid grasslands in the UK. Shahbaz et al. (2014a) confirmed the EKC hypothesis in the context of Tunisian economy by using the data set from 1971 to 2010 and confirmed the long-run association between energy demand, per capita income, air pollution, and trade openness. The results emphasized the need of cleaner production technologies to adopt renewable energy mix for lessening the carbon emissions in a country. Shahbaz et al. (2014b) further investigated the long-run relationships between energy demand, industrialization, and air pollution in the context of Bangladesh by using the countrywide data from 1975 to 2010 and found an inverted U-shaped relationships between carbon emissions and industrialization, which further confirmed the finance led trade and trade led industrial development in a country. The results conclude that trade induce energy pollutants widely indicate the existence of ‘pollution haven’ hypothesis, thus the country's required sustainable policy efforts to reduce CO2 emissions through cleaner technology production. Salomon et al. (2016) concluded that reactive nitrogen compounds are one of the biggest threats to the atmosphere, soil, and water resources that ultimately damage the biological diversity and human health, which merely due to the agricultural practices and fuel combustion processes in a country. Ahmed et al. (2016) investigated the relationship between CO2 emissions, per capita income, and biomass energy consumption in a panel of 24 selected European countries for the period of 1980–2010 and confirmed an inverted U-shaped Kuznets curve between CO2 emissions and economic growth, while technological innovation substantial decreases CO2 emissions, which corresponds to digitalize renewable energy consumption for green growth. Thus, biomass energy is vital for sustainable actions for European countries. Uddin et al. (2016) examined the causal relationships between energy demand, air pollution, per capita income, and trade openness in the context of Sri Lanka for the period of 1971–2006 and confirmed the growth led carbon emissions, and growth led energy consumption in a country. The study emphasized the need of low carbon policies to support country's economic growth. Nguyen et al. (2017) determined the impact of investment and trade openness on energy demand, economic growth, and CO2 emissions in the context of China and India and found that Chinese economy is diversified by the economic gains of trade and investment, which largely explained the energy-growth-pollution nexus, however, this result is absent in case of India. The study concludes that investment activities and trade openness both associated with high economic growth and energy demand on the cost of environmental deterioration, thus sustainable environmental policies should largely included in global policy agenda. Sohag et al. (2017) considered a panel of middle income countries to determined the impact of sectoral value added and energy demand on CO2 emissions and confirmed that sectoral value added in the form of industrialization substantially deteriorate the countries natural environment, which further followed by energy demand that escalates CO2 emissions, however, the results does not establish any significant association between population growth and CO2 emissions in a panel of countries. Thus, the policies to reduce industrial emissions required renewable energy sources to promote sustainable development across countries. Charfeddine (2017) analyzed the environmental sustainability agenda in the context of Qatar economy by using the annul time series data from 1970 to 2015, for this purpose,

variability of climatic factors associated with the other environmental stressors that already cumbersome the biodiversity, therefore, there is substantial need to build the climate change model that evaluates the possible impact on biodiversity loss across the countries. Noss (1999) presented the forestry management framework for evaluating the potential habitat area for the forest species and identified the different factors that impact on the biodiversity loss, including simplification of the forest structure, forest patch sizes, increasing seclusion of patches, disruption of natural fire, and increased road infrastructures. The study further stresses the management of the ecological indicators that would helpful to conserve the forestry management around the globe. Greenfacts (2010) reports identified the five main threats to the global biodiversity, including insidious alien species, greenhouse gas emissions, carbon dioxide emission, forest area, and overexploitation of natural resources. Moreover, there are the following indirect factors that affected the biodiversity, including socio-economic and politicaltechnological factors. This report recognizes the potential threats to the biodiversity and possible mitigation strategies to preserve the global environment. Dietz and Adger (2003) examined the relationship between per capita income and biodiversity loss in the environmental Kuznets curve (EKC) framework in the panel of countries and found that there is an indirect relationship between per capita income and biodiversity, however, the result does not confirm the EKC hypothesis in the relationship between the rates of loss of habitat & species and per capita GDP in the region. Asafu-Adjaye (2003) investigated the dynamic relationship between economic growth and biodiversity loss in the panel of low income countries and found that economic growth damages both the ecosystem and biological diversity in the region. Binder and Neumayer (2005) investigated the possible impact of environmental pressure groups, i.e., environmental NGO (ENGO) groups on different pollution level across the globe. The results confirmed that ENGO has a considerable impact on the different air pollutants, including SO2, smoke, and concentration of heavy particulates in a cross-country region. Mozumder et al. (2006) investigated the relationship between GDP per capita and biodiversity loss in cross-countries setting and used different indicators of biodiversity including species, genetic and ecosystem drivers. The results do not support the EKC relationship between the variables. Biggs et al. (2008) explored the possible impacts of population growth, climate changes and future land use practices on the biodiversity loss in the Southern Africa, and conclude that the mitigation strategies for climatic factors, land–use changes, and reduction in the population growth strategies should required for aligning the biodiversity conservation in the region. Xiankai et al. (2008) investigated the global impact of nitrogen deposition on forest biological diversity and found that nitrogen deposition changes species diversity, however, disproportionate nitrogen may reduce the richness, abundance and even loses special species, that should be taken care while devising the conservation framework for forest biodiversity. Stevens et al. (2010) argued that excessive nitrogen deposition reduces the richness of the plant species which measured by the Soil Ph, in acid grasslands across the Europe. Paul and Uddin (2011) determined the long-run and casual relationships between energy and output dynamics in case of Bangladesh and confirmed an inverse relationship between output residuals and energy residuals, between output cycles, and energy cycles, and between output growth and energy growth. The Granger causality estimates confirmed the unidirectional causality running from output residuals to energy residuals but not vice versa. The study argued that energy conservation policies should be more supportive for broad-based growth that would helpful to integrate an energy based model in a country. Maes et al. (2012) investigated the interrelationship between ecosystems, biodiversity and conservation of habitat area in European scale and found that there is a significant and positive relationship between biodiversity indicators and ecosystem service supply, while habitat area linked with the conservation of biodiversity that had a higher potential of ecosystem supply in the region. Al Mamun et al. 399

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

helpful to develop an interactive environmental model, while unlike previous studies, this study constructed the biodiversity index, which is the relative weighted index of the following environmental factors, including ecological footprint, forest area, per capita food production and per capita food supply. This index served as a ‘response’ variable that influenced by number of environmental variables and growth factors in a panel of selected countries. The real contribution of the study is to used five factors of biological diversity loss, among which, ecological footprint is widely used indicator in the environmental literature, while forest area is used here to trace out the impact of environmental factors on natural habitat areas of species. The per capita food production and supply is closely linked with the forest area, as higher degree of variability in the food production is associated with the richness of species, as it provides food herbs and natural medicated herbs for health living. Finally, the study constructed weighted biodiversity index, which is relatively more pronounced with other environmental factors that is less explored area in environmental literature. The above discussion confirmed the strong nexus between the different socio-economic, biological & environmental factors and biodiversity loss across the globe. This study examined the impact of tropospheric ozone, carbon dioxide and sulfur dioxide concentration on biodiversity loss, coupled with the growth factors that simultaneously affected the ecosystems and biodiversity in the panel of thirty-four countries, during the period of 1995–2014. The study has a following sub-objectives i.e.,

the study selected the key promising economic and environmental factors, including, energy demand, country's per capita income, trade, financial development, and CO2 emissions. The study employed Markov Switching model and confirmed the regime dependent results in order to verify the EKC hypothesis for CO2 emissions and ecological carbon footprint, and U-shaped relationship for CO2 emissions and ecological footprint. The results provoked that ecological footprint and ecological CO2 footprint both introduced as a new environmental proxies that need serious policies to reduce environmental degradation by sustainable instruments in a country. Charfeddine and Mrabet (2017) further used ecological footprint as an environmental factor for a panel of 15 selected MENA countries and examined three important determinants of environmental degradation, including political institutional index, life expectancy at birth, and fertility rate. The results confirmed the EKC hypothesis for ecological footprint in the sample of oil exporting countries and for aggregated data. The results further indicated the Ushaped relationship for ecological footprint in non-oil exporting countries. The results conclude that political and social factors substantially improve the air quality indicator across countries. 1.1. Motivation of the study The real motivation of the study is to analyze the binding agreement of United Nation's Convention of Biological Diversity (CBD) across the diversified panel of countries, which further assess the achievement of sustainable development agenda through mainstreaming biodiversity by 2030. The biodiversity loss mainly contains the quality and quantity of precious species that affects the soil contents and nutrient cycles. The biodiversity loss leads to decrease natural medicated herbs that is vital for human health, thus, the importance of biodiversity is vital for health related research, which is affected by economic and environmental factors in order to reap maximum gains through unsustainable economic transformation on the cost of biodiversity loss. The importance of biodiversity attracts the environmentalists to analyzed conservation policies to developed sustainable broad-based environmental agenda for global prosperity. This study is one of the initiatives to develop an integrated environmental model, which includes number of promising economic and environmental factors that are aligned with the United Nations’ CBD policy. These factors include different national and global pollutants (i.e., CO2 emissions, GHG emissions, N2O emissions, NOx emissions, and SO2 emissions) and economic factors (i.e., per capita GDP and trade openness) which largely affect the ecosystem and global biological diversity. This study used different factors for ecosystem and biodiversity ((i.e., per capita food production and food supply variability, natural habitat area (forest area), ecological footprint, and relative weighted biodiversity index)) to analyzed the humanization process that largely responsible for biodiversity loss. The study emphasized the need of sustainable instruments to conserve biological diversity, which is the important aspect of United Nation's CBD policy. This integrated model will serve this purpose and it will propose certain policy actions for ‘green’ development.

i) To investigate the dynamic relationship between ecological footprint, air pollutants and growth factors across the countries. ii) To what extent has a potential habitat area affected by the air pollutants and growth factors. iii) To examine the long-run relationship between per capita food production variability, air pollutants and growth factors in the panel of selected countries. iv) To observe the changes in per capita food supply by the air pollutants and growth factors, v) To examine the impact of air pollutants, including NOx, N2O, SO2, CO2, and GHG emissions and growth factors on biodiversity index in the region. vi) To investigate the inter-temporal relationships between air pollutants and ecological biodiversity for the next ten year period. These objectives would be achieved by the sophisticated panel heterogeneous techniques including the panel heterogeneous unit root test, panel heterogeneous cointegration test, and panel Fully Modified OLS & panel Dynamic OLS estimators respectively. The study divided in to the following sections, i.e., after introduction in Section 1, Section 2 presented the overview of environmental pollutants and global biodiversity loss. Section 3 shows material and methods. Results discussed in Section 4. Final section concludes the study. 2. Overview of environmental pollutants and global biodiversity loss

1.2. Contribution of the study The literature suggested the following key proxies of environmental degradation, including, CO2 emissions (Arrow et al., 1995; Moomaw and Unruh, 1997; Soytas et al., 2007; Jalil and Mahmud, 2009; Lau et al., 2014; Al-Mulali et al., 2015; Zaman and Moemen, 2017, etc.), SO2 emissions (Kaufmann et al., 1998; Wang et al., 2016, etc.), N2O emissions (Rasli et al., 2017; Zambrano-Monserrate and Fernandez, 2017, etc.), GHG emissions (Møller et al., 2004; Hristov et al., 2013; Riahi et al., 2017, etc.), Ecological footprint (Gössling et al., 2002; Jorgeson, 2003; Ozturk et al., 2016; Charfeddine, 2017; Charfeddine and Mrabet, 2017, etc.), etc. Alike previous studies, this study also used the number of environmental factors, including CO2 emissions, GHG emissions, N2O emissions, NOx emissions, and SO2 emissions that

Environmental pollution is the chemical process of certain biological and physical components of atmospheric system that changed the natural beautification of air quality in to hazardous contamination (Bright, 2012a, 2012b). Environmental pollution has different types and structured composition, which includes i) soil pollutants that largely affected through poor agricultural practices and industrial waste materials. The most common types include heavy metals, pesticides, chemicals, explosive gases, radioactive materials, and asbestos ii) air pollutants largely channelized through two main streams, i.e., industrial activities and transportation, as it is evident that transportation is considered one of the chief factor that responsible for global air 400

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

of climate change is desirable policy options to device environmental sustainable policy framework to reduce environmental threats episodes of global greenhouse gas emissions (Vidal, 2017). Biodiversity loss is another potential threats by climate change and air pollutants that largely affected the existence of animal species, plant species, microorganisms and their genes, and ecosystem, including water, terrestrial and marine, which is necessary for fundamental building blocks for healthy living (Wanjui, 2013). The literature suggests different biodiversity proxies, including i) potential habitat area that proxy by ‘forest area’ (Zaman, 2017; Malik et al., 2016; Ozturk, 2015, 2016, 2017; Danielsen et al., 2009), ii) Global Environmental Facility (GEF) biodiversity benefits index measured by index value of no potential of biodiversity to maximum potential of biodiversity that is proposed by World Bank (Zaman, 2016, 2017), iii) invasive mammalian predators (Doherty et al., 2016), iv) marine environment (Van Dover et al., 2017), etc. The potential habitat destruction is the main factor of biodiversity loss, which is primarily responsible by different factors, including, high mass deforestation, rapid population, global warming and pollution. The convention of biological diversity (CBD) for biological conservation signed by more than 190 countries in 1992 at the Earth Summit in Rio de Janeiro and emphasized the need of preservation of biodiversity for human survival (Bright, 2012a, 2012b). The approximate figures of different biodiversity species including around 9 million types of different plants, animals, protists and fungi species inhabit the Earth corresponding with around 7 billion people, which is endangered with human made activities either directly and/or indirectly that dismantling the Earth's ecosystems and damaged the biological traits with distressing rate (Cardinale et al., 2012). The CBD agenda fairly distributed its positive outcomes for the conservation of biological diversity by sustainable actions and equitable sharing of the benefits accrued by the use of genetic resources. Thus it encourages sustainable future of the planet by collaboration of developed and developing countries on unified policy vista.

pollution. The most common air pollutants are tropospheric ozone, particulate matter, CO2 emissions, SO2 emissions, CO emissions, N2O emissions, NOx emissions, and lead iii) water pollutants largely deposit due to industrial waste disposition in the water, septic systems, and illegal dumping of solid waste. The common types include mercury, phosphorus, and bacterial pollution, iv) noise pollutants disrupts the one's quality of life through different sources, including large sound generated by aircraft, traffic volume, building and infrastructure constructions, loud music etc (Lake, 2017). The GHG emissions contain different trace gases in the earth's atmosphere, which contains CO2 emissions, methane (CH4) emissions, nitrous oxide (N2O) emissions, and fluorinated gases. The gases that corner heat waves in the atmosphere are called greenhouse gases. The earth's atmosphere contains different trace gases, among which CO2 emissions is an important trace gas available on earth's surface. This gas played a major role to change earth's temperature through GHG effects; hence it's considering a major GHG (Petty, 2004). There are number of ways through which CO2 emissions enter in to the atmosphere, i.e., fossil fuel combustion, chemical reactions, solid waste, etc. CH4 emission is largely attributed to the transportation and production of oil, coal, and natural gas. It's further resulted with different agricultural and livestock practices, and municipal solid waste that caused by GHG emissions. N2O emissions resulted with industrial and agricultural activities, solid waste and combustion of fossil fuel. The fluorinated gas caused by different industrial activities and it's considered the most potent GHG emissions that referred as potent global warming gases (EPA, 2015). The toxicological effects of notorious pollutants severely caused by number of unhygienic man-made activities, including i) massive population growth that triggered due to improper healthcare infrastructure, ii) improper sewage disposition due to large drain in the water resource system, iii) industrial waste and pollution that emerged new toxics air pollutants, iv) radioactive waste due to contamination of different unsustainable process, etc., these activities largely affect the global health living (Wasi et al., 2013). The increase use of combustion of fossil fuel in industrial activities leads to increase number of toxic emergent pollutants that affect the air quality and quality of life of the peoples. The transportation industry outweighs the positive effects and considers one of the prime responsible of global air pollution, which leads serious health infections. The contamination of water and soil is largely attributed due to high used of pesticides in the agricultural activities, which made by chemical process that harmful for the surrounding environment. Trade activities include free flow of goods and services between intra-trade domestic activities and/or intra-trade foreign activities that damage the natural environment. Finally, housing constructions considered another potential determinant of environmental pollution that depletes forestry, wildlife activities and different plant species. Thus, these factors are considered the prime causes of environmental degradation across countries (Rinkesh, 2017). Climate change is one of the crucial challenge to the globalized world that need to device certain sustainable policies to mitigate climate change impact on different economic activities, including biodiversity loss, ecosystem disturbance, agricultural productivity loss, soil erosions, and many socio-economic and environmental factors that affect largely with the low adaptation of climate change policies in countries economic profile. It is evident that developed countries less prone to the climate change due to better ability to manage climatic affects as compared to the developing countries (Wijaya, 2014). The most important threat of climate change on food security to the developed countries, especially for staple foods/crops, which is steadily suffering from serious drought, heat waves, and other extreme weather events that costs to the agriculture value added across the globe (McDonnell, 2016). The human induced climate change largely hit the poor countries by experiencing gradual rises sea-level, high intensity cyclones, warmer days long and nights, heavy rains, and longer heatstroke episodes. Thus, climate change mitigated policies and adaptation

3. Material and methods The study measured the dynamic impact of tropospheric ozone, CO2 and SO2 concentrations on biodiversity loss in the panel of thirty-four countries namely Australia, Austria, Belarus, Belgium, Bulgaria, Croatia, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Japan, Latvia, Lithuania, Luxembourg, Netherlands, New Zealand, Norway, Poland, Portugal, Russian Federation, Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey, Ukraine, UK, and the USA. These countries selected to evaluate the environmental performance indicators as indicated by ‘United Nations Convention on biological diversity framework’. The study used different ecological biodiversity indicators i.e., ecological footprint in hectares per person, forest area in 1000 ha, per capita food production variability in 1$ per person (constant 2004–06), and per capita food supply variability in kcal/capita/day, which is affected by the tropospheric ozone layer (NOx), stratospheric ozone layer (N2O), CO2 emissions, GHG emissions, and SO2 concentration including Land Use, Land-Use Change and Forestry in Gigagrams CO2 Equivalent. CO2 emissions is one of the GHG emissions, and have a considerable share in GHG emissions while GHG is typically not air pollutants, however, we study CO2 emissions separately with the GHG emission is that it causes global warming mainly due to burning fossil fuels and forest depletion. This study used the forest area as a proxy for the potential habitat area of species that exposed by an overdose of carbon emissions. The study further used two growth factors which served as intervening variables in between the relationship between environmental indicators and biodiversity loss, i.e., Gross Domestic Product – purchasing power parity in million constant 2005 international US$, and trade openness as percent of GDP. Both the variables simultaneously affected the ecological biodiversity coupled with the environmental indicators in the panel of thirty-four selected countries, over the 401

Environmental Research 160 (2018) 398–411 UNEP (2015a, 2015b) UNEP (2015a, 2015b) GDP may affected either positively or negatively biological diversity TOP may affected either positively or negatively biological diversity

Factor

Eigenvalue

Difference

Proportion

1 2 3 4

1.386732 1.018320 0.962909 0.632039

0.368412 0.055411 0.330870 –

0.3467 0.2546 0.2407 0.1580

Cumulative value 1.386732 2.405052 3.367961 4

PC 3 0.096850 0.741227 0.663427 0.032668

PC 4 0.633924 0.126524 −0.269061 0.713959

PCFPROD

PCFSPLY

1 0.150975

1

Panel B: Eigenvectors (loadings) Variable PC 1 PC 2 EFP −0.610766 0.464462 FAREA −0.222920 −0.620387 PCFPROD 0.303936 0.628565 PCFSPLY 0.696344 −0.065574 Panel C: Ordinary correlations EFP FAREA EFP 1 FAREA 0.015200 1 PCFPROD −0.006064 −0.039059 PCFSPLY −0.331691 −0.093424

Cumulative proportion 0.3467 0.6013 0.8420 1

Note: EFP indicates ecological footprint, FAREA indicates forest area, PCFPROD indicates per capita food production variability, and PCFSPLY indicates per capita food supply variability. PC1, PC2, PC3, and PC4 shows principal component factor 1 to factor 4 respectively.

period of 1995–2014. Table 1 shows the list of variables and their expected signs. Table 1 shows the list of the studied variables, measurement, symbols and expected outcomes. The study hypothesize that tropospheric ozone layer (NOx), stratospheric ozone layer (N2O), SO2, CO2, and GHG emissions badly affected the biological diversity, while growth factors, including economic growth and trade openness may have either a positive and/or negative impact on biological diversity in the panel of selected countries. Table 1 further shows the biological diversity index that is a relative weighted index of the combination of ecological footprint, forest area, food production variability per capita and food supply variability per capita. The study employed Principal Component Analysis (PCA) technique to construct the biological diversity index and presented the constructed index in Table 2. Table 2, Panel A shows the eigenvalue matrix, where the first factor have an eigenvalue of 1.386 which explained the proportion of 34.67%, second factor contain an eigenvalue of 1.018 that explained the proportion of 25.46%, third factor have an eigenvalue less than the unity, i.e., 0.962 contained the proportion of 24.07%, while fourth factor has an eigenvalue of 0.632 and explained 15.80% of the total variations in the given constructs. Table 2, Panel B shows the eigenvectors loading from Principal Component factor 1 (PC1) to PC4. The eigenvectors loading indicate that, except PC3, remaining vector loadings contain the negative loading values; therefore, the study used PC3 eigenvectors loading to construct the single relative weighted component of biodiversity index. Finally, Panel C shows the ordinary correlations between the ecological footprint, forest area, food production variability per capita and food supply variability per capita in the panel of countries. The results show the weak correlation between EFP and FAREA (r = 0.015), between EFP and PCFPROD (r = −0.006), and moderate relationship between EFP and PCFSPLY (r = −0.331) respectively. The other correlations tend to show the weak positive and/or negative correlation between the variables. The tabulation of PCA matrix gives unique factor analysis of four vital components of biodiversity, including forest area, ecological footprint, food production variability and food supply variability that helpful to develop a weighted biodiversity index, which gives robust econometric series for individually created biodiversity index that include all attributes of given biodiversity factors in a single variable. The economic instinct for PCA matrix derives a new index that flair the available biodiversity index in a single component for robust results. The study used the following linear regression equations in order to

GDP TOP

Purchasing Power Parity in Million Constant 2005 International US$ Exports + Imports /GDP

2015b) 2015b) 2015b) 2015b) 2015b) (2015a, (2015a, (2015a, (2015a, (2015a, UNEP UNEP UNEP UNEP UNEP CO2, GHG, N2O, NOx, and SO2 affected biological diversity Land Use, Land-Use Change and Forestry (LULC) in Gigagrams CO2 equivalent

2015b) 2015b) 2015b) 2015b) 2015b) (2015a, (2015a, (2015a, (2015a, (2015a, UNEP UNEP UNEP UNEP UNEP Hectares per Person 1000 ha I$ per person constant 2004–06 Kcal/Capita/Day Relative Weighted index comprises EFP, FAREA, PCFPROD, and PCFSPLY

EFP FAREA PCFPROD PCFSPLY BDINDEX indicators CO2 GHG N2O NOx SO2

Table 2 PCA for biodiversity index.

Dependent variables Ecological footprint Forest Area Per Capita Food Production Variability Per Capita Food Supply Variability Biodiversity Index Independent variables – environmental Carbon dioxide Emissions Total Greenhouse Gas Emissions Nitrous Oxide Emissions Mono-Nitrogen Oxides Sulfur Dioxide Emissions – Intervening variables Gross Domestic Product Trade Openness

Variables

Table 1 List of studied variables.

Symbol

Measurement

Possible Outcomes

Data Source

M. Ahmed Bhuiyan et al.

402

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

we conclude that the given variable series are level stationary, however, if the variable become insignificant, we come to the conclusion that there is a unit root problem, and the variable series are differenced stationary. IPS panel unit root tests assumed that the unit root differs to all the cross sections with the different lag lengths for different cross sections. It is evaluated the alternative hypothesis, i.e., at least one of the cross section is stationary. In a similar manner, the null hypothesis of IPS test states that the given panel series has a unit root. IPS panel unit root tests based on the t-bar statistics and the group means LM-bar (Lagrange Multiplier test). The average LM-bar statistics and t-bar statistics based upon the Monte Carlo experiments which provide better finite sample properties as compared to the LLC panel unit root test. Secondly, the study analyzed the long-run cointegration relationship between the variables. There are various forms of the residual based panel cointegration including Pedroni's (1999, 2004) residual based cointegration, Kao (1999) residual cointegration test and conventional Johansen cointegration test underlying with the Fisher test (Maddala and Wu, 1999). This study employed Kao (1999) residual cointegration technique to analyze the long-run cointegration relationship between the variables. Kao test is based upon the EngelGranger residual cointegration, which specified by the cross-section specific intercepts and homogeneous coefficients on the first-stage regressors. The Kao residual cointegration evaluated the null hypothesis of no cointegration, when the residual series eit should be stationary. Finally, the study adopted Fully Modified OLS (FMOLS) and Dynamic OLS (DOLS) residual based panel estimators (Pedroni, 2000, 2001) that produce asymptotically unbiased coefficient estimates with the normally distributed coefficients. Both the tests used for analyzing the long-run cointegrating relationship between the variables. The Fully Modified OLS estimator is a non-parametric approach that corrected the problems of possible autocorrelation and endogeneity from the given model, while the Dynamic OLS estimator is a parametric approach included the lagged first differenced terms in diverse lag length information criteria. The Schwarz information criteria (SIC) are being used in this study for observing the lead and lag values of the given variables. The Dynamic OLS estimator included lead and lag dynamics for correcting the problem of serial correlation and endogeneity parametrically.

evaluate the dynamic linkages between environmental factors and biodiversity loss in the following five distinct models i.e., Model -1: Environmental Factors and Ecological Footprint

ln (EFP )i, t = β0 + β1 ln (NOx )i, t + β2 ln (N 2O)i, t + β3 ln (SO2)i, t + β4 ln (GHG )i, t + β5 ln (GDP )i, t + β6 ln (TOP )i, t + εi, t (1) Model -11: Environmental Factors and Forest Area

ln (FAREA)i, t = β0 + β1 ln (NOx )i, t + β2 ln (N 2O)i, t + β3 ln (SO2)i, t + β4 ln (CO2)i, t + β5 ln (GHG )i, t + β6 ln (GDP )i, t +β7 ln (TOP )i, t + εi, t (2) Model -111: Environmental Factors and Food Production Variability Per Capita

ln (PCFPROD )i, t = β0 + β1 ln (NOx )i, t + β2 ln (N 2O )i, t + β3 ln (SO2)i, t + β4 ln (CO2)i, t + β5 ln (GHG )i, t + β6 ln (GDP )i, t + β7 ln (TOP )i, t + εi, t (3) Model -1V: Environmental Factors and Food Supply Variability Per Capita

ln (PCFSPLY )i, t = β0 + β1 ln (NOx )i, t + β2 ln (N 2O )i, t + β3 ln (SO2)i, t + β4 ln (CO2)i, t + β5 ln (GHG )i, t +β6 ln (GDP )i, t + β7 ln (TOP )i, t + εi, t (4) Model -V: Environmental Factors and Biodiversity Index

ln (BDINDEX )i, t = β0 + β1 ln (NOx )i, t + β2 ln (N 2O)i, t + β3 ln (SO2)i, t + β4 ln (CO2)i, t + β5 ln (GHG )i, t +β6 ln (GDP )i, t + β7 ln (TOP )i, t + εi, t (5) where, EFP indicates ecological footprint, FAREA indicates forest area, PCFPROD indicates per capita food production variability, PCFSPLY indicates per capita food supply variability, BDINDEX indicates biodiversity index, NOx indicates mono-nitrogen oxides, N2O indicates nitrous oxide, SO2 indicates sulfur dioxide emission, CO2 indicates carbon dioxide emission, GHG indicates greenhouse gas emissions, GDP indicates gross domestic product, TOP indicates trade openness, ‘ln’ indicates natural logarithm, ‘t’ indicates time period of the study starting from 1995 to 2014, 'i' indicates number of cross-section identifiers i.e., 34 countries, and ε is the white noise error term. The coefficient of β1 shows the tropospheric ozone by NOx emissions, the coefficient of β2 shows the stratospheric ozone by N2O emission, the coefficient of β3 , β4 , and β5 shows the SO2 emission, CO2 emission, and GHG emissions, while the coefficient of β6 and β7 shows the growth factors including GDP and trade openness respectively. Fig. 1 shows the plots of the studied variables in natural logarithmic form. The study followed the sequential order of panel econometric techniques in order to measure the long-run relationship between environmental factors and biodiversity loss in the panel of thirty-four countries. Firstly, the study used panel unit root test rather than individual unit root because of the higher power of panel tests. The unit root test is one of the prerequisites for analyzing the stationary properties for the variables in the modern time series analysis. The previous literature suggested the five diversified panel unit root tests, including Levin et al. (2002), Breitung (2000), Im et al. (2003), Fisher-type tests using ADF and PP tests (Maddala and Wu, 1999) and Choi (2001), and Hadri (2000). This study employed Im et al. (2003) panel unit root test for analyzing the stationary series of the variable both at a given level and at first difference. If the variable become significant at given level,

4. Results and discussions The study sequentially estimated the following panel econometric tests for robust inferences in between the environmental factors and biodiversity loss in the panel of thirty-four countries, i.e., descriptive statistics & correlation matrix, panel unit root test, Panel cointegration test, and panel cointegrating regression estimators. 4.1. Descriptive Statistics & Correlation Matrix Table 3 shows the descriptive statistics and correlation matrix for environmental factors and biodiversity loss in the panel of thirty-four countries across the globe. The study constructed the biodiversity index by relative weights for the ecological footprint, forest area, per capita food production variability, and per capita food supply variability. The biodiversity index contains the minimum weighted average value of 17.493 and the maximum value of 149,965.6, having an average value of 8058.839, with the standard deviation of 26,782.84. The CO2 emission has a minimum value of −10,332.80 and the maximum value of 5,279,850, with an average value of 334,939.6 Gigagrams CO2 equivalent. The average value of ecological footprint (EFP) is about 5.276 ha per person, with the maximum value of 12.195 ha per person and standard deviation of 1.729 ha per person. EFP has a positively skewed distribution with considerable peak of the distribution. Forest area comprises an average value of 43,470.80 thousand hectares, with the maximum value of 809,269 and standard deviation is about 14,453.24 thousand hectares. GDP has a minimum value of 11,404.10 403

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

Fig. 1. Plot of the studied variables. Source: UNEP (2015a, 2015b).

Gigagrams CO2 equivalent respectively. Per capita food production has a maximum variability of 81.700 dollar per person, with an average per dollar variability of 18.952. The average variability of per capita food supply is about 42.566 Kcal/Capita/Day. Finally, the minimum value of trade is 16.750 as a percentage of GDP, and the maximum value of 352.900% of GDP, with an average value of 91.617% of GDP. Table 3, Panel B shows the correlation matrix and found that trade openness is the sole factor that damages the biodiversity index and forest area, while the rest of the environmental factors connected with

million US $ and the maximum value of 13,518,200 US $, have an average value of 89,1504.7 million US $. The standard deviation of GDP is about 207,777.2 million US $, with the positively skewed distribution and has a considerable peak of the distribution. The mean value of GHG emissions is about 427,444.7 with the standard deviation of 105,594.9 Gigagrams CO2 equivalents. The N2O emission has a minimum value of 459.660 and the maximum value of 472,354.0 with an average value of 30,777.91 Gigagrams CO2 equivalent. The mean value of NOx and SO2 emission is about 1174.018 and 771.995

Table 3 Descriptive statistics and correlation matrix. Panel-A Mean Maximum Minimum Std. Dev. Skewness Kurtosis

BDINDEX 8058.839 149,965.6 17.49348 26,782.84 4.553758 23.58415

CO2 334,939.6 5,279,850 −10,332.80 850,913.8 4.652982 25.14767

Panel–B: Correlation matrix Probability BDINDEX CO2 BDINDEX 1 * CO2 0.470 1 EFP 0.015 0.230* FAREA 0.989* 0.470* GDP 0.374* 0.976* GHG 0.527* 0.997* N2O 0.561* 0.968* NOx 0.544* 0.970* PCFPROD −0.038 −0.083 PCFSPLY −0.093 −0.100 SO2 0.305* 0.913* TOP −0.251* −0.364*

EFP 5.276780 12.19520 1.986050 1.729494 0.836914 3.581646

FAREA 43,470.80 809,269.0 86.75000 14,453.4 4.553849 23.58486

GDP 891,504.7 13,518,200 11,404.10 207,777.2 4.463868 24.08763

GHG 427,444.7 6,414,840 −6668.850 105,594.9 4.568040 24.40605

N2O 30,777.91 472,354.0 459.6600 73,266.80 4.562895 24.23962

NOx 1174.018 21,346.40 0.440000 3065.101 4.990045 29.01214

PCFPROD 18.95254 81.70000 1.200000 15.17245 1.773980 6.374970

PCFSPLY 42.56628 140 0.001000 26.40169 0.915444 3.443563

SO2 771.9953 17,186.40 0.160000 2203.588 5.748086 37.44238

TOP 91.61730 352.9000 16.75000 49.80747 1.932872 9.414309

EFP

FAREA

GDP

GHG

N2O

NOX

PCFPROD

PCFSPLY

SO2

TOP

1 0.015 0.230* 0.220* 0.207* 0.220* −0.006 −0.331* 0.183* 0.335*

1 0.374* 0.527* 0.561* 0.544* −0.039 −0.093 0.305* −0.251*

1 0.966* 0.928* 0.909* −0.119* −0.134* 0.847* −0.385*

1 0.980* 0.978* −0.076 −0.099 0.906* −0.367*

1 0.973* −0.052 −0.074 0.895* −0.345*

1 −0.021 −0.090 0.946* −0.358*

1 0.150* 0.017 −0.0001

1 −0.036 0.009

1 −0.309*

1

Note: EFP indicates ecological footprint, FAREA indicates forest area, PCFPROD indicates per capita food production variability, PCFSPLY indicates per capita food supply variability, BDINDEX indicates biodiversity index, NOx indicates mono-nitrogen oxides, N2O indicates nitrous oxide, SO2 indicates sulfur dioxide emission, CO2 indicates carbon dioxide emission, GHG indicates greenhouse gas emissions, GDP indicates gross domestic product, and TOP indicates trade openness. * Indicates significant at 99% confidence interval.

404

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

Table 4 Percentile score of air pollutants. Percentiles

10%

20%

30%

40%

50%

60%

70%

80%

90%

CO2 GHG N2O NOx SO2

9012.293 14,954.67 1743.591 49.304 18.426

15,123.4 28,008.26 3801.512 91.088 34.01

31,266.75 48190.13 5518.279 147.537 63.426

42,263.36 57,862.54 7100.66 179.66 82.314

56,413.2 73,950.3 8315.33 214.675 145.3

111,967.8 132,100.4 10,382.6 297.2 316.014

237,780.5 300,557.1 18,636.42 893.398 554.649

334,410.2 467,818 29,373.76 1371.512 964.726

881,849.9 1,026,368 62,066.51 2083.017 1366.683

influenced by numerous studied variables, therefore, the variable’ series exhibit the non-stationary at ‘constant –level’, and ‘constant + trend’, however, the biodiversity index stationary at their first difference. Table 5 further indicates the divergence in the following pollutant factors, including NOx, N2O, CO2 and GHG emissions respectively. These pollutant series are not stationary at level (including constant, and constant + trend), however, these variables stationary at their first difference (both at constant, and constant + trend). The SO2 emissions shows the stationary series at ‘constant–level’, however, including trends in the level data, the result is evaporated, and the variable series stationary at their first difference. Finally, both the growth factors, including GDP and trade openness followed by the random walk, therefore, after decomposing the variable series with the first difference, these variables become to stationary. The overall panel heterogeneous unit root test reveals that all the studied variables, including environmental factors, biodiversity indicators, and growth factors is having non-stationary series at level, however, these variables become stationary at their first difference. The study safely conclude that the studied variables holding the property of integration, i.e., order of integration is one - I(1) variables. The importance of heterogeneous unit root in the modern time series by using panel data indicates that these variables influenced by the numerous factors, therefore, the dispersion in the variables series should be examined by the panel heterogeneous residual cointegration techniques.

the potential increase of the biodiversity index in the region. The ecological footprint is suffered by rapid economic growth (r = 0.230), climate change (r = 0.220), N2O emission (r = 0.220), SO2 emission (r = 0.183), and trade openness (r = 0.335) respectively. The per capita food production variability and food supply variability both affected by rapid economic growth, i.e., r = −0.199, and r = −0.134 respectively. However, as far as correlation magnitude is concerned, per capita food supply variability is affected greater than food production variability in the region. Table 4 shows the percentile score of air pollutant for ready reference. These statistics would helpful to understand the trend of the variables, correlation, and percentile of air pollutant, while, for considerate the functional relationship between the studied variables, the estimates of FMOLS and DOLS provide a sound parametric results for policy implications. 4.2. Panel heterogeneous unit root test Table 5 shows the result of panel heterogeneous unit root by IPS test. The results show that the ecological footprint having the stationary series at first difference, as IPS W-statistics at constant, and with the constant + trend significant at the 1% level. The result implies that ecological footprint is affected by numerous studied factors; therefore, there is the divergence of the variable series that stationary at their first difference. Similarly, the series of forest area becomes first differenced stationary. The other two biological diversity factors including per capita food production and food supply, both stationary at ‘constant–level’, however, the results have been disappearing at ‘constant + trend’. Both the variable’ series stationary at their first differenced which indicate the food variability in the region. The study constructed the index for biodiversity by using the weights to assign for EFP, FAREA, PCFPROD, and PCFSPLY. The series of biodiversity index

4.3. Panel heterogeneous cointegration test Table 6 shows the results of panel heterogeneous cointegration test proposed by Kao (1999) residual cointegration test. There are number of other panel cointegration techniques estimated in the previous studies, including Pedroni's cointegration test and Westerlund cointegration test (see, Kahia et al., 2016, 2017; Charfeddine and Mrabet, 2017 etc), however, both the tests may not be estimated due to more than 6 regressors in each model. In this study, all five models have more than 6

Table 5 Panel unit root test. Variables

EFP FAREA PCFPROD PCFSPLY BDINDEX NOx N2O SO2 CO2 GHG GDP TOP

Level

First difference

Constant

Constant+Trend

Constant

Constant+Trend

0.149 0.702 −3.923* −2.668* 0.088 −0.023 −0.669 −1.504*** −1.057 −0.350 1.487 −0.663

1.415 2.745 −0.840 −0.710 3.455 1.723 2.781 2.031 2.324 2.218 2.638 −0.606

−11.216* −4.398* −9.989* −8.905 −6.203* −7.896* −8.102* −7.551* −7.047* −7.606* −6.920* −10.209*

−8.942* −2.476* −7.237* −6.159* −4.153* −6.487* −6.151* −5.393* −5.231* −5.174* −4.207* −6.845*

Table 6 Kao residual cointegration test.

Note: EFP indicates ecological footprint, FAREA indicates forest area, PCFPROD indicates per capita food production variability, PCFSPLY indicates per capita food supply variability, BDINDEX indicates biodiversity index, NOx indicates mono-nitrogen oxides, N2O indicates nitrous oxide, SO2 indicates sulfur dioxide emission, CO2 indicates carbon dioxide emission, GHG indicates greenhouse gas emissions, GDP indicates gross domestic product, and TOP indicates trade openness. * and *** indicates significant at 99% and 90% confidence interval respectively.

Models

ADF statistics

Residual variance

Model -1: EFP NOX N2O SO2 CO2 GHG GDP TOP Model -11: FAREA NOX N2O SO2 CO2 GHG GDP TOP Model -111: PCFPROD NOX N2O SO2 CO2 GHG GDP TOP Model -IV: PCFSPLY NOX N2O SO2 CO2 GHG GDP TOP Model -V: BDINDEX NOX N2O SO2 CO2 GHG GDP TOP

−3.362*

0.290

−9.127

1.06E+08

−5.900

56.995

−4.869*

240.840

−9.112*

3624592

*

*

Note: EFP indicates ecological footprint, FAREA indicates forest area, PCFPROD indicates per capita food production variability, PCFSPLY indicates per capita food supply variability, BDINDEX indicates biodiversity index, NOx indicates mono-nitrogen oxides, N2O indicates nitrous oxide, SO2 indicates sulfur dioxide emission, CO2 indicates carbon dioxide emission, GHG indicates greenhouse gas emissions, GDP indicates gross domestic product, and TOP indicates trade openness. * Indicates significant at 99% confidence interval.

405

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

by the ecological footprint, while tropospheric ozone emission reduces the ecological footprints from the region. Besides that, in DOLS estimators, there are few other factors that influenced the ecological footprint, including nitrous oxide emission and sulfur dioxide emission, i.e., there is a significant and positive relationship between N2O and ecological footprint, while SO2 emission exerts the negative relationship with the ecological footprints. The higher value of adjusted Rsquared (i.e., 0.971) in the DOLS estimators as compared to the FMOLS estimator (i.e., 0.841) indicates that the model is well -explained by the environment and growth factors that considerably affected the ecological footprint in the region. The value of the standard error of regression is far less in the DOLS estimators (i.e., 0.099) as compared to the FMOLS estimator (i.e., 0.129) respectively. The sum of squared residual is far less in the DOLS estimator (i.e., 4.121) as compared to the FMOLS estimator (i.e., 8.221). These statistics would helpful to correct the serial correlation and endogeneity from parametric and non -parametric estimators. The results of Model -11 indicate that the forest area is affected by the SO2 emission and trade openness, however, trade openness has the largest magnitude in terms of influencing the forest area as compared to the SO2 emission in the FMOLS estimator. The NOx and N2O emissions, both have a significant and positive relationship with the forest area, however, stratospheric ozone –N2O emission have a greater share in terms of influencing the forest area as compared to the NOx in the region. The results of DOLS confirmed the positivity of NOx and N2O with the forest area, while there is the negative relationship of SO2 and trade openness with the forest area respectively. In both of the cointegrating estimators, there is a more elastic relationship of N2O and trade openness with the forest area, while the remaining variables exert the less elastic relationship between them. Besides that, in DOLS estimator, greenhouse gas emissions deteriorate the forest area that ultimately affected the biological diversity in terms of loss of potential habitat area in the region. The higher value of adjusted R-squared, low value of the standard error of regression, and less value of squared residual in the DOLS estimators indicate the higher power of parameters’ expression as compared to the FMOLS estimators, however, FMOLS exerted the longrun variance as compared to the DOLS estimator. The result of Model -111 shows that per capita food production is affected by rapid economic growth and CO2 emission, as if there is a 1% increase in the economic growth and CO2 emission, per capita food production decreases by 0.494% and 0.421% respectively. The GHG emissions exert the largest share in terms of influencing the food

regressors at least in each equations, thus, this study may only used and estimate the Kao cointegration test for long-run cointegrated relationship between the variables. The study evaluated all five models separately with the Kao residual cointegration technique that find the cointegration relationship with the given models by using the specified set of variables. The results show that the value of ADF statistics is −3.362, p < 0.001, with the residual variance of 0.290, which confirmed the cointegration relationship between the variables in the given Model −1. The low residual variance indicates that all the variables given in the Model −1 strongly cointegrated with each other; therefore, the null hypothesis of ‘no cointegration’ is rejected with the alternative hypothesis. Similarly, the value of ADF statistics for Model -11 is around −9.127, p < 0.001, with the residual variance of 1.06E+08. The low residual variance of Model -11 indicates that the variables’ series have a long-run cointegration relationship between them. Model -111 and Model -IV both confirmed the cointegration relationship between the variables’ series, as the value of ADF statistics is significant at the 1% level, and the residual variance of both the Model -111, and Model -IV are comparatively greater than the previous Models -1, and II respectively. Finally, the ADF statistic for the Model -V shows the value of −9.112, p < 0.001, with the residual variance of 3,624,592.0. The large residual variance as compared to remaining four models indicates the variability in the given variables series, therefore, the policies regarding the preservation of the ecological biodiversity should be carefully devised while using the per capita variability in the food production, food supply, and biodiversity index in the region. After confirmation of the cointegration relationship in all the given five models, the study estimated the parameters by FMOLS and DOLS estimators for conclusive outcomes. 4.4. Panel FMOLS and DOLS estimators Table 7 shows the result of panel FMOLS and a panel DOLS estimators for five given models. The results show that there is a significant and positive relationship between ecological footprint and economic growth in the FMOLS estimators, which implies that the ecological footprint is affected by the rapid economic growth of the region. The result is inverted in the relationship between NOx emission and ecological footprint, as increased NOx (i.e., tropospheric ozone emission) reduces the ecological footprints from the region. These results further confirmed by the DOLS estimators where economic growth influenced Table 7 Panel FMOLS and Panel DOLS parameter estimates. Variables

Log(NOx) Log(N2O) Log(SO2) Log(CO2) Log(GHG) Log(GDP) Log(TOP) Statistical tests R-squared Adjusted R-squared S.E. of Regression Long-run variance Mean dependent var. S.D. of dependent var. Sum squared residual

Log(EFP)

Log(FAREA)

Log(PCFPROD)

Log(PCFSPLY)

Log(BDINDEX)

FMOLS

DOLS

FMOLS

DOLS

FMOLS

DOLS

FMOLS

DOLS

FMOLS

DOLS

−0.189* 0.088 −0.045 – −0.011 0.241* −0.062

−0.166* 0.101** −0.070* – −0.002 0.311* 0.090

0.236*** 1.422* −0.506* −0.230 0.044 −0.211 −1.385*

0.242** 1.543* −0.395* −0.040 −0.563** 0.123 −1.073*

−0.043 −0.005 0.267* −0.421* 0.692** −0.494* 0.455

−0.012 −0.161 0.293* −0.312** 0.648* −0.377* 0.536*

0.742* −0.095 0.013 0.223 −0.289 −0.710* −0.276

0.733* −0.196 −0.021 0.168 0.137 −1.132* −0.100

0.208 1.433* −0.509* −0.217 0.052 −0.218 −1.401*

0.210*** 1.559* −0.400* −0.027 −0.558** 0.122 −1.091*

0.841 0.812 0.129 0.031 0.667 0.412 8.221

0.971 0.957 0.099 0.023 1.713 0.388 4.121

0.932 0.926 0.510 0.554 8.560 1.879 129.058

0.980 0.973 0.302 0.141 8.557 1.868 39.478

0.661 0.628 0.489 0.428 2.636 0.802 118.525

0.865 0.817 0.346 0.162 2.642 0.809 51.523

0.618 0.580 0.499 0.468 3.530 0.770 123.010

0.873 0.822 0.325 0.141 3.530 0.771 43.905

0.931 0.924 0.510 0.554 6.887 1.862 129.041

0.980 0.973 0.303 0.141 6.884 1.851 39.496

Note: EFP indicates ecological footprint, FAREA indicates forest area, PCFPROD indicates per capita food production variability, PCFSPLY indicates per capita food supply variability, BDINDEX indicates biodiversity index, NOx indicates mono-nitrogen oxides, N2O indicates nitrous oxide, SO2 indicates sulfur dioxide emission, CO2 indicates carbon dioxide emission, GHG indicates greenhouse gas emissions, GDP indicates gross domestic product, and TOP indicates trade openness.*, **, and *** indicates significant at 99%, 95%, and 90% confidence interval. Log represents natural logarithm. For DOLS estimation, lag and lead variables selected by Schwarz information criteria.

406

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

multicollinearity exists between all five given models, as the VIF values far less than the threshold values between 0 and 10. The interesting values of VIF in BDINDEX shows the less variance inflation as compared to the other four given models that shows the significance of the developed index in an effective manner. The robustness check is carried on for the panel of 47 countries, including 28 countries for high income panel countries, 12 for middle income panel countries, and 7 for low income countries and results are presented in Table 9 for ready reference. The results show that NOx emissions substantially increases forest area and biodiversity index in high-income countries, while it decreases per capita food production and food supply in middle income countries. In the panel of selected low income countries, NOx emissions increases ecological footprints, per capita food production and food supply, and biodiversity index, which correspond that NOx emissions largely damaged the biodiversity index in low-income countries that required sustainable policy framework for mitigation the losses of biodiversity by reducing NOx emissions across countries. The N2O emissions decrease ecological footprints from a panel of middle income and low income countries, while it increases per capita food production in high income and low income countries. The potential habitat area largely influenced by N2O emissions, thus, the policy to reduce N2O emissions to prevent the species richness in the ecosystem is desirable for longterm sustained growth across the globe. The SO2 emissions largely affected ecological footprint in a panel of high and low income countries, while it further influenced biodiversity in high and middle income countries, thus it is imperative to formulate long-term sulfur free policies to prevent the natural environment across countries. The carbon emissions highly influenced biodiversity and ecosystem in a panel of high, middle and low income countries and further induced the per capita food production and food supply in middle and low income countries that threatened the food industry by carbon footprint during the production and supply in a country. The GHG emissions largely influenced the ecological footprint in a panel of high, middle and low income countries, whereas, it further influenced biodiversity index in high and low income countries, which indicate that climate change threatened to the biodiversity and ecological footprint in a countries. The per capita income increases ecological footprint, forest area, and per capita food production and supply in high income panel of countries, while it influenced biodiversity in middle and low income countries. The high economic growth is achieved on the cost of biodiversity loss and damaged the natural environment is evident in middle and low income countries, which required sustainable development in the countries profile. Finally, trade openness jeopardize the increase ecological footprint and biodiversity loss in a panel of high and low income countries, thus, it is desirable to adopt sustainable trade options and cleaner technological methods to conserve natural environment and biodiversity.

production as compared to the SO2 emission in the FMOLS estimator. These results confirmed by the DOLS estimator including one more factors, that is, trade openness, which significantly increases the per capita food production in the region. The goodness- of –fit of the model indicates that environmental and growth factors explained 81.7% variation in the per capita food production in the DOLS estimator (while, it is 62.8% in FMOLS). The standard error of regression is less than in the DOLS as compared to the FMOLS estimator. The sum of squared residual is greater than in the FMOLS as compared to the DOLS estimator. Finally, the long-run variance is about 42.8% in the FMOLS estimator as compared to the DOLS estimator in the given model. The results of Table 7 for Model –IV illustrated that per capita food supply is affected by the rapid economic growth in the FMOLS estimator, as the magnitude to influence the per capita food supply is around 0.710% points. However, there is a significant and positive relationship between NOx and per capita food supply in the region. The same results have been obtained through DOLS estimator with the largest magnitude value of trade openness, that is, more elastic in nature, however, in case of NOx, the magnitude value to influence the per capita food supply is less than in the DOLS as compared to the FMOLS estimator. The higher value of adjusted R-squared, lower value of the standard error of the regression and lower sum of squared residual indicates the robust statistical inferences in DOLS estimator as compared to the FMOLS estimator. Finally, in the Model –V, biodiversity index is affected by the SO2 emission and trade openness in the FMOLS estimator, however, trade openness has the largest share in terms of influencing the biodiversity index as compared to the SO2 emission in the region. As if there is a 1% increase in the trade openness and SO2 emissions, biodiversity index affected by −1.401%, and −0.509% respectively. On the other hand, there is a significant and positive connectivity between N2O and biodiversity index, which implies that stratospheric ozone emission does not damage the biodiversity index, however, policy should be device caution with the care of tropospheric ozone emission in the region. In the DOLS estimator, biodiversity index influenced by the SO2 emission, GHG emissions and trade openness, however, trade openness has the largest share of influencing the biodiversity index, followed by the GHG emission and SO2 emission respectively. The relationship between trade openness and biodiversity index is more elastic, as the coefficient value exceeds the value of unity, while the remaining two variables, i.e., GHG emission and SO2 emission shows the less elastic relationship with the biodiversity index in the region. The result does not show any biological diversity loss due to the NOx, and N2O respectively, as both of the variables significantly increases the biodiversity index in the panel of selected countries. The remaining statistical tests confirmed the goodness -of -fit of the model, lowering the standard error of the regression and lower the sum of squared residual in DOLS estimator as compared to the FMOLS estimator. Table 8 shows the Variance Inflation Factor (VIF) of the five given models that presented in the Table 7 and found that the VIF values fall is in the range between 0 and 2 in EFP model, between 1 and 3 in FAREA model, PCFPROD model, and PCFSPLY model, and between 0 and 1 in BDINDEX. The results confirmed that there is no visible

4.5. Estimates of variance decomposition analysis Table 10 shows the estimates of the variance decomposition analysis of air pollutants. The results reveal that CO2 emissions would be the maximum share in order to influence potential habitat area in a form of forest area, i.e., 0.877%, followed by per capita food supply, i.e., 0.737%, ecological footprint, i.e., 0.100%, per capita food production i.e., 0.019%, and biodiversity index i.e., 0.060% over a next 10 year time period. The NOx emissions exert a maximum share to influence potential habitat area, followed by per capita food supply, per capita food production, ecological footprint and biodiversity index for the periods of 2015–2024. The N2O emission will influence greatly to the forest area, i.e., 1.576%, followed by per capita food supply, i.e., 0.938%, per capita food production, i.e., 0.104%, ecological footprint, i.e., 0.059%, and biodiversity index i.e., 0.004% for the next 10 year period. The SO2 and GHG emissions both greatly affect forest area, followed by per capita food supply, per capita food production,

Table 8 Estimates of Variance Inflation factor (VIF). Variables

EFP Model

FAREA Model

PCFPROD Model

PCFSPLY Model

BDINDEX

Log(NOx) Log(N2O) Log(SO2) Log(CO2) Log(GHG) Log(GDP) Log(TOP)

0.819 1.232 1.779 – 1.554 0.991 1.656

1.112 1.509 2.221 2.178 2.878 1.667 1.992

1.442 1.132 2.245 2.911 1.110 2.01 2.220

1.612 1.809 1.632 2.556 2.219 1.651 2.479

0.661 0.198 0.198 1.000 0.989 0.671 0.670

407

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

Table 9 Robustness check for high-income, middle-income, and low-income panel countries. Variables

High-Income Countries EFP

FAREA

PCFPRO

Middle Income Countries PCFSPLY

BDINDEX

Log(NOx) Ins + Ins Ins + Ins – + Ins Ins Log(N2O) Log(SO2) + Ins – – + Log(CO2) N/U – Ins Ins + Log(GHG) + – Ins Ins + Log(GDP) + + + + Ins Log(TOP) + – + Ins + Statistical tests R-Squared 0.676 0.787 0.687 0.443 0.701 0.623 0.745 0.634 0.412 0.687 Adjusted Rsquared Countries Australia, Austria, Belgium, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Japan, Latvia, Lithuania, Luxembourg, Netherlands, New Zealand, Norway, Poland, Portugal, Slovakia, Slovenia, Spain, Sweden, Switzerland, UK, and the USA (n = 28)

Low Income Countries

EFP

FAREA

PCFPRO

PCFSPLY

BDINDEX

EFP

FAREA

PCFPRO

PCFSPLY

BDINDEX

Ins – Ins N/U + Ins +

Ins – Ins Ins Ins + Ins

– Ins + + – _ +

– + – – _ Ins +

Ins Ins + + Ins + _

+ – + N/U + Ins Ins

Ins – + – Ins – –

+ + Ins + + Ins –

+ – + + + – –

+ – – + + + Ins

0.512 0.489

0.401 0.387

0.882 0.854

0.891 0.863

0.699 0.678

0.599 0.578

0.612 0.598

0.772 0.756

0.978 0.956

0.945 0.932

Belarus, Bulgaria, Croatia, Russian Federation, Turkey, Ukraine, Brazil, China, Malaysia, Pakistan, Philippines, and South Africa (n = 12)

Nepal, Tanzania, Uganda, Burkina Faso, Chad, Liberia, and Zimbabwe (n = 7)

Note: ‘+’ shows positive coefficient value, ‘-‘ shows negative coefficient value, ‘N/U’ shows not used CO2 in EFP model, ‘Ins’ shows insignificant coefficient value. The ‘+’ and ‘-‘ coefficient values are significant at 5% confidence interval. The FMOLS estimation technique is used for panel estimation.

biodiversity across the countries.

ecological footprint, and biodiversity index. The results indicate that air pollutants will have a larger effect on the potential habitat area while it would be least influenced to the biodiversity index in a panel of countries. The overall results of the study conclude that we may not ignore the fact that NOx and N2O, which considered as the tropospheric ozone and stratospheric ozone emissions damaging the ecological biodiversity across the globe, as we find some traces in the form of damaging the ecological footprints in this region. The policies should be device in order to preserve the ecological biodiversity which are the core objective of the ‘United Nations Convention of Biological Diversity Framework’ across the globe.

• The series of environmental indicators and specific growth factors

•

5. Conclusions This study focused on the key environmental pollutants and specific growth factors, which have a considerable impact on the biological diversity loss in the panel of thirty-four countries, during the period of 1995–2014. The study used different indicators in order to trace out the biological diversity, including ecological footprint, forest area, per capita food production variability, and per capita food supply variability. In addition, the study constructed the single relative weighed index for biodiversity which is the combination of all four prescribed biological diversity components in it. The study used a diverse set of environmental factors, including the tropospheric and stratospheric ozone depleting substances, i.e., NOx and N2O, SO2, CO2 and GHG emissions respectively. The study further used two specific growth factors, including real GDP and trade openness for robust inferences. The results of the study is connected with the dynamic long-run panel heterogeneous cointegration testing procedures, including panel heterogeneous unit root test, panel heterogeneous cointegration test, and panel FMOLS & panel DOLS estimators respectively. The following policy implications have been drawn from this exercise i.e.,

•

•

•

• Panel unit root test conclude that biological diversity factors fol-

lowed by the random walk hypothesis, as ecological footprint, forest area, per capita food production and food supply variability exhibit the non-stationary series, which is stabilized at their first differences. The results suggest that these variables affected by the numerous environmental and growth factors that should be limited by the sustainable policy instruments to preserve the ecological

• 408

tend to fluctuate over the period of time, which become stable at their first differences. The results imply that across the countries, these variables diverge from their steady state condition due to the massive reforms been held for limiting the climatic factors, atmospheric pollutants, and stable growth patterns at country level. However, there is still required the strong policy framework to normalize the natural flora by the strategic management decision for preventing the ‘dirty game of industrial production’ in the region. The results of panel cointegration conclude that the series of given variables in the ecological footprint model, forest area model, per capita food production model, per capita food supply model, and biodiversity index model confirmed the cointegration relationship between the variables in all the five given models. The results imply that these variables exhibit the long-run relationship over the period of time; therefore, the policies for preventing any biodiversity loss should be the need of the current scenario. The results of panel cointegrating estimators reveal that the tropospheric ozone – NOx and SO2 emission both damages the ecological footprint, while stratospheric ozone- N2O and GDP directly associated with the ecological footprint in the panel of selected countries. The results emphasize the need to limit the NOx and SO2 emissions in order to prevent from the biodiversity loss across the countries. Forest area is badly affected by the SO2 emissions, greenhouse gas emissions and trade liberalization policies, while both the NOx and N2O emissions directly associated with the forest area in the region. The results conclude that forest area is one of the potential habitat areas for the different species; therefore, the policies should be formulated in order to protect the habitat area for balancing the natural bonanza across the globe. Food production variability per capita indirectly connected with the CO2 emissions, and economic growth, while it is directly associated with the SO2 emissions, climatic factors, and trade liberalization policies. The trade liberalization policies should be pursued in order to increase the food production, while there is required a balanced economic growth to sustain our food resources. Per capita food supply is influenced with the N2O emissions, GDP per capita, and trade liberalization policies, while it is directly linked with the NOx emission in the region. Trade associated growth

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

Table 10 Variance decomposition analysis of air pollutants. Variance Decomposition of Δln(CO2): Period S.E. 2015 0.481379 2016 0.482458 2017 0.485686 2018 0.485778 2019 0.485862 2020 0.485865 2021 0.485867 2022 0.485867 2023 0.485867 2024 0.485867

Δln(CO2) 100.0000 99.55358 98.27677 98.23968 98.20761 98.20654 98.20584 98.20582 98.20580 98.20580

Δln(BDINDEX) 0.000000 0.013146 0.055529 0.058716 0.060055 0.060141 0.060140 0.060140 0.060142 0.060142

Δln(EFP) 0.000000 0.048968 0.091241 0.098820 0.099557 0.100148 0.100176 0.100181 0.100182 0.100182

Δln(FAREA) 0.000000 0.078840 0.836538 0.846709 0.876374 0.876666 0.877079 0.877082 0.877085 0.877085

Δln(PCFPROD) 0.000000 0.008042 0.007938 0.019265 0.019270 0.019379 0.019383 0.019386 0.019386 0.019387

Δln(PCFSPLY) 0.000000 0.297422 0.731980 0.736815 0.737136 0.737129 0.737382 0.737390 0.737406 0.737407

Variance Decomposition of Δln(NOx): Period S.E. 2015 0.510444 2016 0.511447 2017 0.521131 2018 0.521240 2019 0.521535 2020 0.521539 2021 0.521545 2022 0.521545 2023 0.521545 2024 0.521545

Δln(NOx) 100.0000 99.60853 95.94457 95.90521 95.81623 95.81456 95.81274 95.81266 95.81261 95.81260

Δln(BDINDEX) 0.000000 0.020677 0.043702 0.043766 0.045367 0.045407 0.045566 0.045566 0.045570 0.045570

Δln(EFP) 0.000000 0.006159 0.035403 0.053384 0.058460 0.059655 0.059724 0.059724 0.059724 0.059725

Δln(FAREA) 0.000000 0.064672 3.169031 3.181004 3.258670 3.258662 3.258623 3.258629 3.258640 3.258640

Δln(PCFPROD) 0.000000 0.037972 0.091027 0.097292 0.100385 0.100641 0.100764 0.100820 0.100822 0.100823

Δln(PCFSPLY) 0.000000 0.261993 0.716265 0.719340 0.720885 0.721080 0.722579 0.722603 0.722638 0.722638

Variance Decomposition of Δln(N2O): Period S.E. 0.412949 2016 0.414054 2017 0.418438 2018 0.418514 2019 0.418623 2020 0.418625 2021 0.418627 2022 0.418627 2023 0.418627 2024 0.418627

Δln(N2O) 100.0000 99.46802 97.39761 97.36224 97.31851 97.31751 97.31664 97.31661 97.31659 97.31659

Δln(BDINDEX) 0.000000 0.001398 0.002388 0.004235 0.004351 0.004370 0.004392 0.004392 0.004395 0.004395

Δln(EFP) 0.000000 0.007632 0.045486 0.056089 0.058755 0.059572 0.059601 0.059604 0.059604 0.059604

Δln(FAREA) 0.000000 0.042409 1.528492 1.534631 1.575603 1.575773 1.576096 1.576097 1.576098 1.576098

Δln(PCFPROD) 0.000000 0.091443 0.091782 0.104064 0.104515 0.104514 0.104574 0.104589 0.104592 0.104592

Δln(PCFSPLY) 0.000000 0.389096 0.934245 0.938741 0.938265 0.938257 0.938700 0.938706 0.938724 0.938725

Variance Decomposition of Δln(SO2): Period S.E. 2015 0.667745 2016 0.670974 2017 0.678287 2018 0.678443 2019 0.678620 2020 0.678623 2021 0.678626 2022 0.678626 2023 0.678626 2024 0.678626

Δln(SO2) 100.0000 99.10982 96.99148 96.94857 96.91191 96.91120 96.91043 96.91039 96.91037 96.91037

Δln(BDINDEX) 0.000000 0.020970 0.044233 0.045369 0.045350 0.045367 0.045390 0.045391 0.045393 0.045393

Δln(EFP) 0.000000 0.016788 0.052744 0.061714 0.063071 0.063580 0.063593 0.063593 0.063593 0.063594

Δln(FAREA) 0.000000 0.036905 1.587760 1.597642 1.631837 1.631932 1.632098 1.632098 1.632099 1.632099

Δln(PCFPROD) 0.000000 0.353914 0.347816 λ0.367011 0.368616 0.368646 0.368769 0.368797 0.368799 0.368800

Δln(PCFSPLY) 0.000000 0.461598 0.975964 0.979694 0.979221 0.979278 0.979718 0.979729 0.979743 0.979743

Variance Decomposition of Δln(GHG): Period S.E. 2015 0.447649 2016 0.448991 2017 0.452236 2018 0.452346 2019 0.452433 2020 0.452436 2021 0.452437 2022 0.452437 2023 0.452437 2024 0.452437

Δln(GHG) 100.0000 99.45129 98.03796 97.99037 97.95673 97.95554 97.95477 97.95474 97.95472 97.95471

Δln(BDINDEX) 0.000000 0.010262 0.031815 0.034734 0.035180 0.035210 0.035212 0.035212 0.035214 0.035214

Δln(EFP) 0.000000 0.031458 0.075001 0.083386 0.084584 0.085276 0.085306 0.085310 0.085310 0.085310

Δln(FAREA) 0.000000 0.094765 1.054755 1.070356 1.102657 1.103043 1.103373 1.103374 1.103375 1.103375

Δln(PCFPROD) 0.000000 0.078362 0.077705 0.093119 0.093094 0.093147 0.093152 0.093159 0.093159 0.093160

Δln(PCFSPLY) 0.000000 0.333859 0.722765 0.728030 0.727750 0.727781 0.728189 0.728208 0.728226 0.728226

•

policies harm the biological diversity coupled with the stratospheric ozone emission, therefore, the policy makers should formulate longterm growth policies that mitigate the air pollutants and supply healthy food across the countries. Biodiversity index is negatively influenced by the climate change, trade liberalization policies, and carbon dioxide emissions, while it is directly linked with the NOx, and N2O emissions in the panel of selected countries. There is substantial need to mitigate the climate change and carbon dioxide emissions that merely inclined with the free trade in the region. The policies should be redesigned in order

•

409

to preserve our natural base by the tropospheric ozone, carbon dioxide emissions and sulfur dioxide emissions respectively. Finally, SO2 emissions, CO2 emissions, GHG emissions, and trade openness largely influenced biodiversity in high income panel of countries, whereas SO2 emissions and economic growth jeopardize the biodiversity agenda in a panel of middle income countries. In a panel of low income countries, biodiversity is influenced by high mass carbon emissions, GHG emissions, and economic growth. The policies to reduce air pollutants should be re-define in terms of conserve natural environment and biodiversity and made policies as

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

1995. Economic growth, carrying capacity, and the environment. Ecol. Econ. 15 (2), 91–95. Asafu-Adjaye, J., 2003. Biodiversity loss and economic growth: a cross-country analysis. Contemp. Econ. Policy 21, 173–185. http://dx.doi.org/10.1093/cep/byg003. Biggs, R., Simons, H., Bakkenes, M., Scholes, R.J., Eickhout, B., van Vuuren, D., Alkemade, R., 2008. Scenarios of biodiversity loss in southern Africa in the 21st century. Glob. Environ. Change 18 (2), 296–309. Binder, S., Neumayer, E., 2005. Environmental pressure group strength and air pollution: an empirical analysis. Ecol. Econ. 55 (4), 527–538. Breitung, J., 2000. The local power of some unit root tests for panel data. In: In: Baltagi, B. (Ed.), Advances in Econometrics, Vol. 15: Nonstationary Panels, Panel Cointegration, and Dynamic Panels 15. JAI Press, Amsterdam, pp. 161–178. Bright, I. 2012b. Biodiversity Loss And Why We Should Be Concerned About It. Online available at: 〈http://www.tropical-rainforest-animals.com/biodiversity-loss.html〉. (Accessed 29 September 2017). Bright, I., 2012a. Environmental Pollution, Its Sources and Effects. Online available at: 〈http://www.tropical-rainforest-animals.com/Environmental-Pollution.html〉. (Accessed 29 September 2017). Cardinale, B.J., Duffy, J.E., Gonzalez, A., Hooper, D.U., Perrings, C., Venail, P., Kinzig, A.P., 2012. Biodiversity loss and its impact on humanity. Nature 486 (7401), 59. Charfeddine, L., 2017. The impact of energy consumption and economic development on Ecological Footprint and CO2 emissions: evidence from a Markov Switching Equilibrium Correction Model. Energy Econ. 65, 355–374. Charfeddine, L., Mrabet, Z., 2017. The impact of economic development and social-political factors on ecological footprint: a panel data analysis for 15 MENA countries. Renew. Sustain. Energy Rev. 76, 138–154. Choi, I., 2001. Unit root tests for panel data. J. Int. Money Financ. 20, 249–272. Danielsen, F., Beukema, H., Burgess, N.D., Parish, F., Brühl, C.A., Donald, P.F., Fitzherbert, E.B., 2009. Biofuel plantations on forested lands: double jeopardy for biodiversity and climate. Conserv. Biol. 23 (2), 348–358. Dietz, S., Adger, W.N., 2003. Economic growth, biodiversity loss and conservation effort. J. Environ. Manag. 68 (1), 23–35. Doherty, T.S., Glen, A.S., Nimmo, D.G., Ritchie, E.G., Dickman, C.R., 2016. Invasive predators and global biodiversity loss. Proc. Natl. Acad. Sci. USA 113 (40), 11261–11265. EPA, 1991. Evaluating Exposures to Toxic Air Pollutants: A Citizen’s Guide. EPA 450/390-023, March 1991. Online available at: 〈https://www3.epa.gov/airtoxics/3_90_ 023.html〉. (Accessed 14 October 2016). EPA, 2015. Overview of Greenhouse Gases. United States Environmental Protection Agency, online available at: 〈https://www.epa.gov/ghgemissions/overviewgreenhouse-gases〉. (Accessed 30 September 2017). Gössling, S., Hansson, C.B., Hörstmeier, O., Saggel, S., 2002. Ecological footprint analysis as a tool to assess tourism sustainability. Ecol. Econ. 43 (2), 199–211. Greenfacts, 2010. Biodiversity a global outlook. Chapter 6: What are the Main Threats to Biodiversity? Online available at: 〈http://www.greenfacts.org/en/globalbiodiversity-outlook/l-3/6-threat-biodiversity.htm〉. (Accessed 14 October 2016). Hadri, K., 2000. Testing for stationarity in heterogeneous panel data. Econom. J. 3, 148–161. Hristov, A. N., Oh, J., Lee, C., Meinen, R., Montes, F., Ott, T., ... & Yang, W., 2013. Mitigation of Greenhouse Gas Emissions in Livestock Production: A Review of Technical Options for Non-CO2 Emissions. FAO Animal Production and Health Paper No, 177, pp. 1–206. ICP, 2015. Impacts of Ozone Pollution on Biodiversity. The International Cooperative Programme, Centre for Ecology and Hydrology, Gwynedd, UK. Im, K.S., Pesaran, M., H., Shin, Y., 2003. Testing for unit roots in heterogeneous panels. J. Econ. 115, 53–74. Jalil, A., Mahmud, S.F., 2009. Environment Kuznets curve for CO2 emissions: a cointegration analysis for China. Energy Policy 37 (12), 5167–5172. Jorgenson, A.K., 2003. Consumption and environmental degradation: a cross-national analysis of the ecological footprint. Soc. Probl. 50 (3), 374–394. Kahia, M., Aïssa, M.S.B., Charfeddine, L., 2016. Impact of renewable and non-renewable energy consumption on economic growth: new evidence from the MENA Net Oil Exporting Countries (NOECs). Energy 116, 102–115. Kahia, M., Aïssa, M.S.B., Lanouar, C., 2017. Renewable and non-renewable energy useeconomic growth nexus: the case of MENA Net Oil Importing Countries. Renew. Sustain. Energy Rev. 71, 127–140. Kao, C., 1999. Spurious regression and residual-based tests for cointegration in panel data. J. Econ. 90, 1–44. Kappelle, M., Van Vuuren, M.M., Baas, P., 1999. Effects of climate change on biodiversity: a review and identification of key research issues. Biodivers. Conserv. 8 (10), 1383–1397. Kaufmann, R.K., Davidsdottir, B., Garnham, S., Pauly, P., 1998. The determinants of atmospheric SO2 concentrations: reconsidering the environmental Kuznets curve. Ecol. Econ. 25 (2), 209–220. Lake, R., 2017. Types of Environmental Pollutants. Online available at: 〈http://www. livestrong.com/article/221368-types-of-environmental-pollutants/〉. (Accessed 29 September 2017). Lau, L.S., Choong, C.K., Eng, Y.K., 2014. Investigation of the environmental Kuznets curve for carbon emissions in Malaysia: do foreign direct investment and trade matter? Energy Policy 68, 490–497. Levin, A., Lin, C.F., Chu, C., 2002. Unit root tests in panel data: asymptotic and finitesample properties. J. Econ. 108, 1–24. Maddala, G.S., Wu, S., 1999. A comparative study of unit root tests with panel data and a new simple test. Oxf. Bull. Econ. Stat. 61, 631–652. Maes, J., Paracchini, M.L., Zulian, G., Dunbar, M.B., Alkemade, R., 2012. Synergies and trade-offs between ecosystem service supply, biodiversity, and habitat conservation

per agreed treaty of convention of biodiversity to reduce biodiversity losses across the globe. The human exposure to air pollutants have a serious health hazard, though ‘National Ambient Air Quality Standards (NAAQS)’ promoted in order to prevent direct exposure of air pollutants to human health exclusively at outdoor while it is evident that human health are badly affected by indoor where the concentration of air pollutants become more volatile. Sexton and Ryan (1988) described the assessment of human exposure to air pollutants and emphasized the need of quantitative risk assessment that aimed to promote global health. EPA (1991) report described the four step exposure assessment which includes i) to identify the air pollutants that likely to be present in the air, ii) assessment of the concentration of air pollutants in the air, iii) evaluated the geographic area where these air pollutants concentrations are largely present, and iv) estimated the number of peoples that expose to be with air pollutants to check the regulatory standards of health in this region. WHO (2003) report argued that there are several alternative measures in which we may reduce the human exposure to air pollutants such as, to adopt the air quality guidelines and standard practicing at outdoor for minimizing health risks, to educate the general public about health risks due to contact with air pollutants, and used market driven instruments like imposing taxes on industries or firms that pollute air and water environment etc. these information would helpful to reduce human exposure to air pollutants and minimizes the risks accordingly. The 21st session of the Conference of the Parties (COP21) was held in Paris-France in 2015, which aims to legally binding the countries for universal agreement on climate protection to reduce global warming below 2 °C (UNEP, 2015a, 2015b). This initiative supports the countries to reduce anticipated hazards, absorbs climatic shocks, and redefines the development strategies to mitigate and adapt the climate risks. The number of stakeholders should play their role in order to provide climate change solutions to transform their policies for long-term sustainable development, including, public-private partnership, United Nation agencies, academic and research institutes, and other stakeholders that directly and/or indirectly associated with the climatic externalities. The COP21 is one of the sustainable development agenda to reduce climatic vulnerabilities through ‘early warning systems’ to prepare the countries for high massive disasters. In addition, ‘insurance and social protection’ regulatory framework should be device for climate and sustainability risks. Finally, to build better physical and natural infrastructure, it will helpful to reduce the risks of hazardous episodes of climate for long-term development (UN, 2015). Carbon emission is one of the GHG emissions that greatly cause global warming; therefore, this study separately examined the impact of carbon emissions on ecological and biological diversity. The carbon overdose in an atmosphere is caused by mainly burning fossil fuels or forest depletion, which continuously accumulated due to industrialization revolution. The findings of the study confirm that potential habitat area and biological diversity exposed by different air pollutants, including NOx, N2O, SO2, CO2, and GHG emissions. These air pollutants amplified with the specific growth factors, i.e., GDP and trade that increase the concentration of the pollutants due to unsustainable environmental reforms in a region. The clean air is the only desirable policy option for sustainable long-term growth of the region. References Ahmed, A., Uddin, G.S., Sohag, K., 2016. Biomass energy, technological progress and the environmental Kuznets curve: evidence from selected European countries. Biomass Bioenergy 90, 202–208. Al Mamun, M., Sohag, K., Mia, M.A.H., Uddin, G.S., Ozturk, I., 2014. Regional differences in the dynamic linkage between CO2 emissions, sectoral output and economic growth. Renew. Sustain. Energy Rev. 38, 1–11. Al-Mulali, U., Weng-Wai, C., Sheau-Ting, L., Mohammed, A.H., 2015. Investigating the environmental Kuznets curve (EKC) hypothesis by utilizing the ecological footprint as an indicator of environmental degradation. Ecol. Indic. 48, 315–323. Arrow, K., Bolin, B., Costanza, R., Dasgupta, P., Folke, C., Holling, C.S., Pimentel, D.,

410

Environmental Research 160 (2018) 398–411

M. Ahmed Bhuiyan et al.

Sustain. Energy Rev. 34, 325–336. Shahbaz, M., Uddin, G.S., Rehman, I.U., Imran, K., 2014b. Industrialization, electricity consumption and CO2 emissions in Bangladesh. Renew. Sustain. Energy Rev. 31, 575–586. Sohag, K., Al Mamun, M., Uddin, G.S., Ahmed, A.M., 2017. Sectoral output, energy use, and CO2 emission in middle-income countries. Environ. Sci. Pollut. Res. 24 (10), 9754–9764. Soytas, U., Sari, R., Ewing, B.T., 2007. Energy consumption, income, and carbon emissions in the United States. Ecol. Econ. 62 (3), 482–489. Stevens, C.J., Duprè, C., Dorland, E., Gaudnik, C., Gowing, D.J., Bleeker, A., Dise, N.B., 2010. Nitrogen deposition threatens species richness of grasslands across Europe. Environ. Pollut. 158 (9), 2940–2945. Stevens, C.J., Duprè, C., Dorland, E., Gaudnik, C., Gowing, D.J., Bleeker, A., ... & Dise, N. B., 2014. Biodiversity of acid grasslands in the Atlantic regions of Europe: the impact of nitrogen deposition. In: Nitrogen Deposition, Critical Loads and Biodiversity. Springer Netherlands, pp. 243–250. Uddin, M.G.S., Bidisha, S.H., Ozturk, I., 2016. Carbon emissions, energy consumption, and economic growth relationship in Sri Lanka. Energy Sources Part B: Econ. Plan. Policy 11 (3), 282–287. UN, 2015. UN Secretary-General’s Initiative Aims to Strengthen Climate Resilience of the World’s Most Vulnerable Countries and People. United Nations Sustainable Development Goals, Online available at: 〈http://www.un.org/ sustainabledevelopment/blog/2015/11/un-secretary-generals-initiative-aims-tostrengthen-climate-resilience-of-the-worlds-most-vulnerable-countries-and-people/〉. (Accessed 30 September 2017). UNEP, 2015a. COP21: Climate Change. United Nations Environment Protection, Online available at: 〈http://www.unep.org/climatechange/cop21〉 (Accessed 30 September 2017). UNEP, 2015b. The UNEP Environmental Data Explorer, as compiled from United Nations Environment Programme. United Nations Environment Programme, Geneva. Online available at: 〈http://ede.grid.unep.ch〉. (Accessed 12 October 2015). Van Dover, C.L., Ardron, J.A., Escobar, E., Gianni, M., Gjerde, K.M., Jaeckel, A., Smith, C.R., 2017. Biodiversity loss from deep-sea mining. Nat. Geosci. 10 (7), 464–465. Vidal, J., 2017. Climate Change Will Hit Poor Countries Hardest, Study Shows. Guardian News, Online available at: 〈https://www.theguardian.com/global-development/ 2013/sep/27/climate-change-poor-countries-ipcc〉. (Accessed 29 September 2017). Wang, Y., Han, R., Kubota, J., 2016. Is there an environmental Kuznets curve for SO2 emissions? A semi-parametric panel data analysis for China. Renew. Sustain. Energy Rev. 54, 1182–1188. Wanjui, J., 2013. Biodiversity conservation needs and method to conserve the biological diversity. J. Biodivers. Endanger. Species 1 (113), 1–2. Wasi, S., Tabrez, S., Ahmad, M., 2013. Toxicological effects of major environmental pollutants: an overview. Environ. Monit. Assess. 185 (3), 2585–2593. WHO, 2003. Role of Human Exposure Assessment in Air Quality Management. Report on the Joint Workshop of World Health Organization, Joint Research Center European Concerted Action Urban Air, Indoor Environment and Human Exposure, Bonn, Germany. Online available at: 〈http://www.euro.who.int/__data/assets/pdf_file/ 0003/164514/E79501.pdf〉. (Accessed 14 October 2016). Wijaya, A.S., 2014. Climate change, global warming and global inequity in developed and developing countries (analytical perspective, issue, problem and solution). In: IOP Conference Series: Earth and Environmental Science. IOP Publishing, Vol. 19, No. 1, p. 012008. Wilson, E.O., 1989. Threats to biodiversity. Sci. Am. 261 (3), 108–116. Xiankai, L., Jiangming, M., Shaofeng, D., 2008. Effects of nitrogen deposition on forest biodiversity. Acta Ecol. Sin. 28 (11), 5532–5548. Zaman, K., 2016. Confronting the biofuel-carbon emissions-biodiversity trilemma in a panel of biofuel consuming countries. Malays. J. Sci. 35 (2), 113–130. Zaman, K., 2017. Biofuel consumption, biodiversity, and the environmental Kuznets curve: trivariate analysis in a panel of biofuel consuming countries. Environ. Sci. Pollut. Res. http://dx.doi.org/10.1007/s11356-017-0087-y. (Forthcoming). Zaman, K., Abd-el Moemen, M., 2017. Energy consumption, carbon dioxide emissions and economic development: evaluating alternative and plausible environmental hypothesis for sustainable growth. Renew. Sustain. Energy Rev. 74, 1119–1130. Zambrano-Monserrate, M.A., Fernandez, M.A., 2017. An environmental Kuznets curve for N2O emissions in Germany: an ARDL approach. Nat. Resour. Forum 41 (2), 119–127.

status in Europe. Biol. Conserv. 155, 1–12. Malik, M.A.S., Shah, S.A., Zaman, K., 2016. Tourism in Austria: biodiversity, environmental sustainability, and growth issues. Environ. Sci. Pollut. Res. 23 (23), 24178–24194. McDonnell, T., 2016. Climate Change Could Have a Bigger Impact on Developed Countries’ Food Systems. Online available at: 〈http://grist.org/article/climatechange-could-have-a-bigger-impact-on-developed-countries-food-systems/〉. (Accessed 29 September 2017). McNeely, J.A., 1992. The sinking ark: pollution and the worldwide loss of biodiversity. Biodivers. Conserv. 1 (1), 2–18. Møller, H.B., Sommer, S.G., Ahring, B.K., 2004. Biological degradation and greenhouse gas emissions during pre-storage of liquid animal manure. J. Environ. Qual. 33 (1), 27–36. Moomaw, W.R., Unruh, G.C., 1997. Are environmental Kuznets curves misleading us? The case of CO2 emissions. Environ. Dev. Econ. 2 (4), 451–463. Mozumder, P., Berrens, R.P., Bohara, A.K., 2006. Is there an environmental Kuznets curve for the risk of biodiversity loss? J. Dev. Areas 39 (2), 175–190. Myers, N., 1993. Biodiversity and the precautionary principle. Ambio 22, 74–79. Nguyen, D.K., Sévi, B., Sjö, B., Uddin, G.S., 2017. The role of trade openness and investment in examining the energy-growth-pollution nexus: empirical evidence for China and India. Appl. Econ. 49 (40), 4083–4098. Noss, R.F., 1999. Assessing and monitoring forest biodiversity: a suggested framework and indicators. For. Ecol. Manag. 115 (2), 135–146. Ozturk, I., 2015. Sustainability in the food-energy-water nexus: evidence from BRICS (Brazil, the Russian Federation, India, China, and South Africa) countries. Energy 93, 999–1010. Ozturk, I., 2016. Biofuel, sustainability, and forest indicators' nexus in the panel generalized method of moments estimation: evidence from 12 developed and developing countries. Biofuels Bioprod. Bioref. 10 (2), 150–163. Ozturk, I., 2017. The dynamic relationship between agricultural sustainability and foodenergy-water poverty in a panel of selected Sub-Saharan African Countries. Energy Policy 107, 289–299. Ozturk, I., Al-Mulali, U., Saboori, B., 2016. Investigating the environmental Kuznets curve hypothesis: the role of tourism and ecological footprint. Environ. Sci. Pollut. Res. 23 (2), 1916–1928. Paul, B.P., Uddin, G.S., 2011. Energy and output dynamics in Bangladesh. Energy Econ. 33 (3), 480–487. Pedroni, P., 1999. Critical values for cointegration tests in heterogeneous panels with multiple regressors. Oxf. Bull. Econ. Stat. 61, 653–670. Pedroni, P., 2000. Fully modified OLS for heterogeneous cointegrated panels. In: Baltagi, H.B. (Ed.), Nonstationary Panels, Panel Cointegration and Dynamic Panels, Vol. 15 of Advances in Econometrics, pp. 93–130. Pedroni, P., 2001. Purchasing power parity tests in cointegrated panels. Rev. Econ. Stat. 83 (4), 727–731. Pedroni, P., 2004. Panel cointegration; asymptotic and finite sample properties of pooled time series tests with an application to the PPP hypothesis. Econom. Theory 20, 597–625. Petty, G.W., 2004. A first course in atmospheric radiation. Sundog Publ. 229–251. Rasli, A.M., Qureshi, M.I., Isah-Chikaji, A., Zaman, K., Ahmad, M., 2017. New toxics, race to the bottom and revised environmental Kuznets curve: the case of local and global pollutants. Renew. Sustain. Energy Rev. http://dx.doi.org/10.1016/j.rser.2017.08. 092. (Forthcoming). Riahi, K., Van Vuuren, D.P., Kriegler, E., Edmonds, J., O’neill, B.C., Fujimori, S., Lutz, W., 2017. The shared socioeconomic pathways and their energy, land use, and greenhouse gas emissions implications: an overview. Glob. Environ. Change 42, 153–168. Rinkesh, 2017. Environmental Pollution. Online available at: 〈https://www.conserveenergy-future.com/causes-and-effects-of-environmental-pollution.php〉. (Accessed 29 Sepetmber 2017). Salomon, M., Schmid, E., Volkens, A., Hey, C., Holm-Müller, K., Foth, H., 2016. Towards an integrated nitrogen strategy for Germany. Environ. Sci. Policy 55, 158–166. Sexton, K., Ryan, P.B., 1988. Assessment of human exposure to air pollution: methods, measurements, and models. In: Watson, Ann Y., Bates, Richard R., Kennedy, Donald (Eds.), Air Pollution, the Automobile, and Public Health. National Academy Press, New York, pp. 207–238. Shahbaz, M., Khraief, N., Uddin, G.S., Ozturk, I., 2014a. Environmental Kuznets curve in an open economy: a bounds testing and causality analysis for Tunisia. Renew.

411