Research on Potential Evaluation and Sustainable Development of Rural Biomass Energy in Gansu Province of China

Home , Anning District, Baiyin, Hezuo, Jiayuguan City, Liangzhou District, Lintan County

sustainability

Article Research on Potential Evaluation and Sustainable Development of Rural Biomass Energy in Gansu Province of China

Sheng Zhong 1,*, Shuwen Niu 1,2 and Yipeng Wang 1

1 College of Earth and Environmental Sciences, Lanzhou University, 222 Tianshui S Rd, Chengguan Qu, Lanzhou 730000, China; [email protected] (S.N.); [email protected] (Y.W.) 2 Key Laboratory of Western China’s Environmental Systems (Ministry of Education), Lanzhou University, 222 Tianshui S Rd, Chengguan Qu, Lanzhou 730000, China * Correspondence: [email protected]; Tel.: +86-150-0250-8523

Received: 20 September 2018; Accepted: 17 October 2018; Published: 20 October 2018

Abstract: The development and utilization of renewable energy is an important way to solve the environmental dilemma. Biomass energy is a kind of renewable energy and one of the most widely distributed and easily accessible energy forms. It has currently become a main direction of renewable energy development. This paper took Gansu Province of China as the research object to calculate its theoretical reserves of biomass energy resources and then evaluate its potential of biomass energy development by using TOPSIS method under different agricultural development and geographical environmental conditions. Spatial autocorrelation analysis was also performed to reveal the spatial distribution and temporal evolution of the potential of biomass energy development in Gansu Province. The results show that: (1) The total reserves of biomass energy resources from agricultural wastes in Gansu Province reach 7.28 × 107 t/year, with equivalent biogas production of about 10 3 1.95 × 10 m /year. (2) In most counties of Gansu Province, the Ci value is smaller than 0.5000, indicating that the potential of biomass energy development is relatively low in Gansu Province. (3) The spatial agglomeration of biomass energy development potential occurs mainly in the Hexi area, the Gannan area and the Loess Plateau area of East Gansu Province. (4) There is an area with obvious high-low (H-L) agglomeration of biomass energy development potential to the north-west side of the Gannan area with low-low (L-L) agglomeration of biomass energy development potential. It is a key zone to help drive biomass energy development in the Gannan area. (5) The spatial range of positive correlation (high-high and low-low agglomeration) areas shrunk during the evaluation period.

Keywords: biomass energy; rural energy; potential evaluation; spatial-temporal analysis; growth pole; Gansu province; the northwest region of China

1. Introduction The development and utilization of renewable energy is an important way to solve the environmental dilemma [1]. It could help promote economic diversiﬁcation, raise productivity and enhance environmental quality and energy justice [2]. Biomass energy is a kind of renewable energy among many other renewable energy forms (wind, solar, hydraulic, geothermal, etc.) [3] and is also one of the most widely distributed and easily accessible energy forms. Biomass is almost the most important source of renewable energy in rural areas. Therefore, biomass energy has become a main direction of renewable energy development.

Sustainability 2018, 10, 3800; doi:10.3390/su10103800 www.mdpi.com/journal/sustainability Sustainability 2018, 10, 3800 2 of 20

Biomass energy is one of the oldest energy resources used by humankind [4]. As humans just learned how to use fire, biomass energy began to be used for cooking and lighting. After the rise of agriculture, agricultural residues played an important role in energy supply [5]. To date, in many remote areas, biomass energy still dominates the energy sector, particularly as the main source of energy for cooking and heating [6]. For example, in Sub-Saharan Africa, people relying on the traditional use of biomass to obtain energy account for 80% of total population [6,7]. Biomass energy mainly comes from surplus agricultural land, agricultural residues and wastes and forestry residues [8]. The main processes for obtaining energy from biomass include direct combustion, pyrolysis, gasification, hydro gasification, liquefaction, anaerobic digestion, alcoholic fermentation and trans-esterification [9]. Developing biomass energy is conducive to the sustainable development of human society as it can solve problems such as environmental degradation and resource depletion [10]. Europe is one of the regions where biomass energy was developed relatively earlier than in other areas. AEBIOM estimated a potential in the EU at about 78 billion Nm3 of biomethane, of which 58.9 billion Nm3 derived from agriculture (27.2 billion Nm3 from crops, 10 billion Nm3 from straw, 20.5 billion Nm3 from manure and 1.2 billion Nm3 from landscape management) and 19 billion Nm3 from waste (10.0 billion Nm3 from MSW, 3.0 billion Nm3 from industrial waste and 6.0 billion Nm3 from sewage sludge). From this potential, 46 billion Nm3 could be used until 2020 [11]. A spatial information system-based approach was used to evaluate biogas potential in Europe and spatial data on European-wide livestock and poultry were used for analysis. Results showed that the theoretical biogas potential of manure was estimated at 26 billion m3 biomethane in Europe (23 billion m3 biomethane in the EU) and the realistic biogas potential, counting on collectible manure, was assessed at 18 billion m3 biomethane in Europe (16 billion m3 biomethane in the EU) [12]. Fuel wood from forestry residues plays an important role in the development of biomass energy in Europe. If household fuelwood is included in energy wood, then wood biomass could satisfy 2–18% of world’s primary energy needs in 2050 [13]. In Turkey, fuel wood is considered the most noteworthy biomass energy because its share of the total energy production is as high as 14% in Turkey [14]. In Southeast Asia, especially Malaysia, oil palm waste is the most important source of biomass energy. It is estimated that Malaysia has the potential to generate around 15 billion m3 of biogas annually [15] and oil palm waste accounts for around 98.7% of total biomass energy potential in Sabah of Malaysia [16]. China is also a country with high biomass energy potential. In China, annual biogas potential from agricultural wastes is approximately (3350.58 ± 669.28) × 108 m3 (equal to 239.22 ± 47.78 million tons of standard coal) and such potential was underutilized in the past [17]. Among them, only livestock and poultry dung are expected to provide 110 billion m3 biogas equivalent by 2020 [18]. Besides, the total potential from crop residues (30%) and energy crops (70%) is equivalent to 1/4 of the total annual oil consumption of China [19]. Northwest China is an economically underdeveloped and ecologically fragile region, where the development of biomass energy is urgently needed to improve residents’ quality of life. At present, however, little attention has been paid to the development and utilization of biomass energy in this region. Some scholars have explored the extension and environmental impact of biomass energy development projects in this region [20]. However, there is almost no research on the assessment of biomass energy resource reserves and exploitation potential in this region, which is an obvious research gap that needs to be filled. Moreover, most scholars tend to calculate the specific value of theoretical biomass energy resource reserves in a region and then take this value as the actual potential of biomass energy development in the corresponding region. There are also some scholars who use spatial data to evaluate biomass energy development potential [12] but they only calculate the theoretical reserves of biomass energy resources in different areas and then explore the spatial distribution of biomass energy resources reserves. Current research rarely considers the influence of the agricultural development and geographical conditions on the actual potential of biomass energy development in a region. The calculated theoretical biomass resource reserves cannot reflect the actual potential of biomass energy development in a certain area, thus the actual biomass energy development potential should be evaluated. Therefore, this paper is devoted to filling above research gaps. Sustainability 2018, 10, 3800 3 of 20

Sustainability 2018, 10, x FOR PEER REVIEW 3 of 21 2. Overview of Study Region 2. Overview of Study Region Gansu Province is located at northwest China, near Shaanxi Province in the east, Qinghai Province and SichuanGansu Province isin located the south, at Xinjiangnorthwest Province China, near in the Shaanxi west and Province Ningxia in Province the east, and Qinghai Inner MongoliaProvince and Autonomous Sichuan Province region in in the the south, north Xinjiang (Figure 1Province). Gansu in the Province west and has Ningxia a total Province land area and of 4.56Inner× Mongolia105 km2, mostly Autonomous located region on the in second the north ladder (Figure of China’s 1). Gansu terrain, Province with has an average a total land elevation area of of 4.56 × 105 km2, mostly located on the second ladder of China’s terrain, with an average elevation of 1500–2000 m. Its length from east to west was more than 1600 km. Its shape is like a bone, on the 1500–2000 m. Its length from east to west was more than 1600 km. Its shape is like a bone, on the northwest border of the Qinghai-Xizang Plateau. Since ancient times, it has been a main route from the northwest border of the Qinghai-Xizang Plateau. Since ancient times, it has been a main route from Western region to the Central Plains. With obvious geographical advantages, it is now an important the Western region to the Central Plains. With obvious geographical advantages, it is now an route for the implementation of “Belt and Road” strategy. important route for the implementation of “Belt and Road” strategy.

FigureFigure 1.1. GeographicalGeographical location map o off Gansu Province. Province.

GansuGansu ProvinceProvince isis inin aa regionregion wherewhere threethree major plateaus (the Loess Loess Plateau, Plateau, the the Qinghai Qinghai-Tibet-Tibet PlateauPlateau and and the the Inner Inner Mongolia Mongolia Plateau) Plateau) meet, thus it it has complex complex and and diverse diverse terrain terrain and and geomorphology.geomorphology. FromFrom easteast toto west,west, it can be roughly d dividedivided into into four four distinct distinct geographical geographical units. units. (1)(1) The The Longdong Longdong Loess Loess PlateauPlateau area,area, locatedlocated inin the central and eastern eastern part part of of Gansu Gansu Province, Province, west west ofof the the Shaanxi-Gansu Shaanxi-Gansu borderborder andand easteast ofof Wushaoling,Wushaoling, is a typical loess hilly hilly area. area. (2) (2) The The Longnan Longnan MountainMountain area, area, located located in in the the south south of Gansu of Gansu Province Province and andsouth south of Weihe of Weihe River, River, is part is of part the western of the extensionwestern extension of the Qinling of the Mountains. Qinling Mountains. There are There dense are mountains dense mountains in this area, in this the area, mountains the mountains are high andare thehigh valleys and the are valleys deep. (3)are Thedeep. Gannan (3) The Plateau Gannan area, Plateau close area, to the close Qinghai-Tibet to the Qinghai Plateau-Tibet (called Plateau the roof(called of the the world),roof of the is a world), typical is plateau a typical area plateau with aarea high with topography a high topography and high altitude.and high (4)altitude. The Hexi (4) CorridorThe Hexi area, Corridor with area, Wushaoling with Wushaoling at its east at endits east and end Gansu-Xinjiang and Gansu-Xinjiang border border at its westat its west end, end, is an elongatedis an elongated area witharea with length length of 1000 of 1000 km. km. The The terrain terrain is is flat, flat, about about 1000–1500 1000–1500 mm above the the sea sea level. level. TheThe main main terrain terrain consists consists ofof floodplainfloodplain inin frontfront of mountains, floodplain floodplain in the the river river and and desert desert terrain. terrain. ItsIts southern southern part part is is the the QilianQilian Mountains,Mountains, mostmost ofof whichwhich are over 3500 m above the the sea sea level. level. GansuGansu Province’sProvince’s economic economic development development lags lags behind behind that that in many in many other other regions regions of China. of China. The GDP of Gansu Province was 679.03 billion CNY in 2015, ranking 27th out of 31 provinces, The GDP of Gansu Province was 679.03 billion CNY in 2015, ranking 27th out of 31 provinces, autonomous regions and municipalities in the country. Its per capita GDP was only 26,165.30 CNY in 2015 and was the lowest in China. The average urbanization rate in Gansu Province was 46.37% in Sustainability 2018, 10, 3800 4 of 20 autonomous regions and municipalities in the country. Its per capita GDP was only 26,165.30 CNY in 2015 and was the lowest in China. The average urbanization rate in Gansu Province was 46.37% in 2015, which was lower than the national average of 56.10%. The per capita disposable income of urban residents was 23,767.10 CNY in 2015, which was far below the national average of 31,195.00 CNY. The primary industry output value was 95.41 billion CNY in 2015, ranking 23rd in the country. The per capita disposable income of rural residents was 6936.20 CNY in 2015, which was also far below the national average of 11,421.70 CNY.

3. Materials and Methods

3.1. Data Sources The primary data of agricultural development, agricultural production and physical geography in each county and district of Gansu Province in 1997–2015 were obtained. The main data sources are as follows: (1) Statistical yearbooks of provinces and cities, such as “Gansu Development Yearbook” (1998–2016), “Lanzhou Statistical Yearbook” and so forth. (2) Statistical bulletins of cities, districts and counties, such as the Wuwei Municipal Statistical Bulletin on National Economic and Social Development, the Liangzhou District Statistical Bulletin and so forth. (3) Authoritative literatures. Some of the data used in this study were from authoritative literatures, such as the “Atlas of China” published by China Map Publishing House and so forth.

3.2. Index System Establishment The index system for evaluating biomass energy development potential was constructed (Table1). The index system consists of an evaluation content system (agricultural development foundation and biomass energy resource endowment) and an evaluation impact system (physical geographical elements).

Table 1. Index system.

Index System Items Indexes Units Primary industry output value CNY Agricultural Agricultural population People development Per capita disposable income in rural areas CNY foundation Agricultural acreage mu Evaluation content Total power of agricultural machinery kw system Grain total output t Biomass energy Cash crops total output t resource Pig total output Pigs endowment Sheep total output Sheep Cow total output Cows Unit altitude difference m Physical Evaluation impact Annual precipitation mm geographical system Annual mean temperature ◦C elements Annual sunshine duration h Unit altitude difference = Relief Amplitude/total land area, Relief Amplitude = local highest elevation-local lowest elevation.

In this index system, the evaluation of agricultural development foundation and biomass energy resource endowment enables to obtain the theoretical potential of biomass energy development in various regions of Gansu Province. Note, however, that this theoretical potential is limited by the physical geography of corresponding region. Therefore, to more accurately evaluate the actual potential of biomass energy development in a region, the inﬂuence of physical geographical elements should be considered. The physical geographical elements do not change signiﬁcantly in a short period; thus, they were assumed here to be constant during the evaluation period. Sustainability 2018, 10, 3800 5 of 20

3.3. Research Ideas and Methods

3.3.1.Sustainability Research 2018 Ideas, 10, x FOR PEER REVIEW 5 of 21

TheThe research research ideas ideas are are shown shown as as aa ﬂowflow chartchart (Figure(Figure 22))

FigureFigure 2.2. TheThe ﬂowflow chart of of research research ideas ideas..

3.3.2.3.3.2 Methodology. Methodology (1)(1 IEW) IEW & & TOPSIS TOPSIS DecisionsDecisions can can be be made made in in mathematical mathematical language to to obtain obtain the the best best solution solution to to problems. problems. However,However, in in actual actual analysis analysis and and decision-making decision-making process, conflicting conflicting situations situations often often need need to to be be considered.considered. For For example, example, in this in study, this study, rich reserves rich reserves of biomass of biomass energy resources energy resources and good and agricultural good developmentagricultural foundationdevelopment can foundation promote can the promote potential the of potential biomass of energy biomass development energy development but which but can bewhich restricted can be by restricted geographical by geographical environmental environmental elements elements at the same at the time. same This time. situation This situation is called is a multi-criteriacalled a multi decision-making-criteria decision- problem,making problem, which needs which to needs be solved to be solved by multi-criteria by multi-criteria decision-making decision- making method. Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), method. Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), VlseKriterijumska VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR), Analytic Hierarchy Process (AHP) Optimizacija I Kompromisno Resenje (VIKOR), Analytic Hierarchy Process (AHP) and so on are and so on are available methods. These methods have been widely used in project planning, content available methods. These methods have been widely used in project planning, content evaluation and evaluation and risk assessment in different fields [21–24]. In the field of energy research, these risk assessment in different fields [21–24]. In the field of energy research, these methods are also used methods are also used to evaluate energy development potential [25], as well as energy consumption toand evaluate performance energy development [26]. The feasibility potential of [ these25], as methods well as energy in this consumptionfield has been and widely performance confirmed. [26 ]. TheNotably, feasibility VIKOR of these is a multi methods-criteria in thisdecision field-making has been method widely with confirmed. compromise Notably, scheme, VIKOR which is is a multi-criteriainconsistent with decision-making the objective of method this paper. with Compared compromise with scheme,AHP, TOPSIS which method is inconsistent is more operable with the objectiveand it is of easier this paper.to understand Compared the quantization with AHP, TOPSISresults of method this approach. is more Therefore, operable TOPSIS and it ismethod easier to understandwas used thehere quantization to evaluate the results potential of this of approach. biomass energy Therefore, development TOPSIS inmethod rural areas was used of Gansu here to evaluateProvince the at potential county scale. of biomass The basic energy principle development of TOPSIS in method rural areas is to ofcarry Gansu out Provincesystematic at evaluation county scale. Theof basic different principle projects of TOPSIS based on method different is to attributes carry out by systematic constructing evaluation the original of different index projects matrix and based onmaking different use attributes of the multi by constructing-scheme decision the original-making index method matrix [27,2 and8]. The making aim isuse to offind the out multi-scheme the best decision-makingsolution and to solve method the system [27,28]. problem The aim optimally is to find [29 out]. The the specific best solution steps of the and TOPSIS to solve method the system are problemas follows. optimally [29]. The specific steps of the TOPSIS method are as follows. 1 ①Raw Raw data data standardization. standardization. The The rawraw datadata of biomass energy energy evaluation evaluation indexes indexes of ofthe the counties counties ofof Gansu Gansu Province Province in in each each year year werewere obtainedobtained and the evaluation evaluation matrix matrix of of n nrowsrows (samples) (samples) and and m m columns (indexes) was formed: X = (Xij)m  n . The raw data were standardized by the Spannweite Sustainability 2018, 10, 3800 6 of 20

columns (indexes) was formed: X = (Xij)m×n. The raw data were standardized by the Spannweite Standardization to eliminate the impact of dimensions and magnitudes of indexes on the ﬁnal results. 0 0 The obtained matrix was X = (X ij)m×n. 2 Weight determination. IEW (Information Entropy Weight) method was used to determine the weight of index [30]. The information entropy values of the standardized indexes were calculated with the following formula: n eij = −K∑ f ij ln f ij (1) i=1 where K (K = 1/ln n) is a constant, which is related to the number of evaluation objects n. fij is calculated as follows: 0 1+X ij fij = m (2) 0 ∑ (1 + X ij) i=1 Index information utility value G is the difference between 1 and the index information entropy value: Gij = 1 − eij (3) Then, the weight of the index can be calculated as follows:

Gij 1 − eij Wij = n = n (4) ∑ Gij n − ∑ eij j=1 j=1

This objective weighting process is purely based on unbiased data and is thus able to overcome the deﬁciencies of subjective weighting methods [31,32]. 3 Ideal reference solution and anti-ideal reference solution. The Ideal reference solution and anti-ideal reference solution are determined as follows [29]:  maxv ( for stimulant)  ij V+ = i (5) i minv ( for destimulant)  ij i  minv ( for stimulant)  ij V− = i (6) i maxv ( for destimulant)  ij i

4 Ci value calculation. The weighted Euclidean distance was used to measure the difference between the normalized value of each index and the worst (optimal) solution. v u n − u − 2 di = t∑ wj(xij − xj ) (7) j=1

v u n + u + 2 di = t∑ wj(xij − xj ) (8) j=1

+ The smaller the di value is, the closer the evaluation object is to the optimal solution and the better − the development condition is, whereas the smaller the di value is, the closer the evaluation object is to the worst solution and the worse the development condition is [33,34]. and can be integrated as Ci [35].

− di ci = − + (9) di + di Sustainability 2018, 10, 3800 7 of 20

The value of Ci ranges between 0 and 1. The closer the Ci value is to 1, the higher the evaluation level is; the closer the Ci value is to 0, the lower the evaluation level is. (2) Spatial autocorrelation Spatial autocorrelation analysis is a common spatial analysis method that has been widely used in the study of geography. It is used to explore the temporal and spatial variation of geographical elements and it can effectively solve problems that traditional spatial statistical methods cannot directly solve. In this study, the spatial autocorrelation method was used to analyze the potential of biomass energy development in Gansu province. The speciﬁc steps of this method are as follows: 1 Spatial weight matrix construction. Spatial weight matrix is the basis of the spatial autocorrelation method. In this study, a two-dimensional symmetric spatial weight matrix based on the domain rule was established as follows: ( 1 County I is adjacent to county j W = (10) ij 0 County I is not adjacent to county j

  w11 w12 ··· w1n  ···   w21 w22 w2n    (11)  ············  wn1 wn2 ··· wnn

2 Global autocorrelation Global autocorrelation analysis reﬂects the total spatial concentration and distribution of biomass energy development potential in Gansu Province, which is usually measured by global Moran’s I. The formula is as follows [36]:

n n ! n n n !2 2 I = n∑ ∑ wij(xi − x)(xi − x) ∑ ∑ wij ∑(xi − x) (12) i=1 j=1 i=1 j=1 i=1 where xi and xj are the Ci values of i and j samples in the study area, respectively. I value is between [−1, 1]. The closer the I value is to 1, the more positive the correlation of Ci value is in space. The closer the I value is to −1, the more negative the correlation of Ci value is in space. If I value is close to 0, this indicates that Ci value has no correlation in space [37]. Standardized Z value can be used to determine whether the index I has significant autocorrelation in space. I − E(I) Z(I) = p (13) VAR(I) where E(I) and VAR(I) are theoretical expectation and theoretical variance of I value, respectively. For a given level of significance, if Z is positive and significant, there is a positive spatial autocorrelation. If Z is negative and significant, there is a negative spatial autocorrelation. If Z value is close to 0, then the sample shows a random independent distribution. 3 Local autocorrelation Local autocorrelation analysis focuses on the correlation of research object in local space and clearly identifies the location where spatial agglomeration occurs and changes. Moran scatter plot and LISA cluster map can be used for local autocorrelation analysis. The Moran scatter plot is a two-dimensional presentation of data z and its spatial lag factor Wz, with (Wz, z) as the coordinate [38]. It distributes the data points of Ci into four different quadrants. The first quadrant to the fourth quadrant correspond to four different spatial patterns: high-high (HH) agglomeration, low-high (LH) agglomeration, low-low (LL) agglomeration and high-low (HL) agglomeration, respectively. H (L) indicates that the observed value is higher (lower) than their average value. Sustainability 2018, 10, 3800 8 of 20

The LISA cluster map generated by the local Moran’ I can reﬂect the spatial agglomeration characteristics more intuitively and measure the correlation degree between cities. The local Moran’ I is calculated as follows: n Ii = (xi − x)∑ Wij xj − x (14) j=1

4. Results

4.1. Biomass Energy Resources By collecting relevant data of agricultural production in Gansu Province, we calculated the total reserves of biomass energy resources from agricultural wastes by using the ratio of grain to straw, the coefficient of discharge of livestock and poultry effluents and so forth. The resource composition and regional distribution of biomass energy in Gansu Province were preliminarily analyzed. This provides a basis for the research of biomass energy development and utilization in Gansu province. RGS (the ratio of grain to straw) was derived from the value published in 2015 by the National Development and Reform Commission of China and the Ministry of Agriculture of China in “Notice on the Final Evaluation of Comprehensive Utilization Planning of Crop Straw” (2015, 3264). In this notice, according to the current situation and characteristics of agricultural development in different regions of China, different RGS were determined and used. Among them, the RGS for all kinds of crops in the northwest farming area of China were (kg/kg): corn 1.52, wheat 1.23, beans 1.07 and potato 1.22. DCLP (the discharge coefficient of livestock and poultry effluents) was obtained from the published “The Manual of the First National Survey of Pollution Sources of Livestock and Poultry Breeding.” The specific values were (kg/day): pigs 2.37, sheep 2.60, beef cattle 15.01 and cow 32.86. The DCLP of cow and beef cattle are different, thus the coefficient was determined to be 20.00 kg/day in this paper. The results show that the total reserves of biomass energy resources from agricultural wastes in Gansu Province reach 7.28 × 107 t/year, with equivalent biogas production of about 1.95 × 1010 m3/year (Table2). The amount of equivalent biogas production per capita in Gansu Province is 748.80 m3/year. The biogas consumption of a family of three is generally in the range of 200–250 m3/year. Clearly, the biomass energy resources in Gansu Province are relatively abundant. If they could be effectively exploited, they are theoretically sufficient to meet the energy demands in rural areas of Gansu Province.

Table 2. Total reserves, per capita reserves and geographical distribution of biomass energy resources in Gansu Province.

The Hexi The Longzhong The Longdong The Longnan The Gannan Total or Area *1 Area *2 Area *3 Area *4 Area *5 Average Total reserves 2.57 × 107 1.55 × 107 1.30 × 107 1.06 × 107 8.15 × 106 7.28 × 107 (t) Biogas equivalent 6.81 × 109 3.89 × 109 3.50 × 109 2.92 × 109 2.34 × 109 1.95 × 1010 (m3) Per capita reserves 1398.11 484.57 700.32 447.26 784.36 748.77 (m3) *1 Including Jiayuguan city, Jinchang city, Wuwei city, Zhangye city, Jiuquan city. *2 Including Lanzhou city, Baiyin city, Dingxi city. *3 Including Pingliang city, Qingyang city. *4 Including Tianshui city, Longnan city. *5 Including Linxia prefecture, Gannan prefecture.

The reserves of biomass energy resources in the Hexi area are 2.57 × 107 t/year, accounting for more than 35% of the total reserves of biomass energy resources in Gansu Province. The total population of this area is only about 4.8 million, thus the per capita biomass energy reserves can reach about 1400 m3/year. Clearly, the Hexi area is the region with the most abundant reserves of biomass energy resources in Gansu Province. The reserves of biomass energy resources in the Longzhong area are second only to those in the Hexi area and are 1.55 × 107 t/year, accounting for 20% of the total biomass energy resource reserves in Gansu Province. The total population in the Longzhong Sustainability 2018, 10, 3800 9 of 20 area is 8.04 million, far more than that in the Hexi area. Therefore, the per capita biomass energy reserves in the Longzhong area is only about 485 m3/year and far lower than that in the Hexi area. The total reserves of biomass energy resources in the Longdong area are slightly less than those in the Longzhong area, accounting for about 18% of the total reserves in Gansu Province. However, the total population in the Longdong region is 5 million, much less than that in the Longzhong area. Therefore, the per capita biomass energy reserves in the Longdong region reaches 700 m3/year and is more than that in the Longzhong area. The biomass energy resources reserves in the Longnan area and the Gannan area account for 15% and 12% of the total reserves in Gansu Province, respectively. The total population in Longnan area is second only to that in the Longzhong area and is about 6.5 million. Its per capita biomass energy resource reserves is the lowest in Gansu Province, only about 447 m3/year. In contrast, although the biomass energy resource reserves in the Gannan area are the least in Gansu Province, its total population is less than 3 million and thus its per capita biomass energy resource reserves is relatively high, about 750 m3/year, which is second only to that of the Hexi area.

4.2. Evaluation of Biomass Energy Development Potential

4.2.1. Weight Calculation Results The index weight was determined by IEW. Physical geography is only a factor affecting the potential of biomass energy but not a decisive factor. Therefore, during calculation, the weight of physical geographical elements should be calculated separately and then included into the calculation results of evaluation content system. Then, the ﬁnal results were obtained. In the evaluation content system (Figure3), the indexes that have the greatest impact on the potential of biomass energy development are x10, x9 and x3, with weights of 0.1063, 0.1031 and 0.1002, respectively. This is due to the following two aspects. First, livestock and poultry manure are the most available resources to develop biomass energy, thus the related indexes should have large weight values. Second, the development of biomass energy and its market-oriented supply depend on certain economic conditions. As the income of rural residents reaches a certain level, they will choose to purchase biogas or biomass molding fuel (BMF) with relatively high prices. In other words, we should promote the comprehensive development and marketization of biomass energy only when most of the residents have the ﬁnancial ability to purchase biomass energy. In the evaluation impact system, y2 and y1 have the greatest impact on the potential of biomass energy development, with weights of 0.2702 and 0.2460, respectively. These two indexes, one affecting the agricultural production and the other affecting the large-scale exploitation of biomass energy, are important factors that limit the developmentSustainability 2018 of, local10, x FOR biomass PEER REVIEW energy. 10 of 21

FigureFigure 3. 3.Weight WeightValues Values of of EachEach Index.Index.

In the evaluation content system (Figure 3), the indexes that have the greatest impact on the potential of biomass energy development are x10, x9 and x3, with weights of 0.1063, 0.1031 and 0.1002, respectively. This is due to the following two aspects. First, livestock and poultry manure are the most available resources to develop biomass energy, thus the related indexes should have large weight values. Second, the development of biomass energy and its market-oriented supply depend on certain economic conditions. As the income of rural residents reaches a certain level, they will choose to purchase biogas or biomass molding fuel (BMF) with relatively high prices. In other words, we should promote the comprehensive development and marketization of biomass energy only when most of the residents have the financial ability to purchase biomass energy. In the evaluation impact system, y2 and y1 have the greatest impact on the potential of biomass energy development, with weights of 0.2702 and 0.2460, respectively. These two indexes, one affecting the agricultural production and the other affecting the large-scale exploitation of biomass energy, are important factors that limit the development of local biomass energy.

4.2.2. Ci Values Calculation and Sorting Results

The Ci value of each county was determined by TOPSIS method. Taking the year of 2015 as an example, we calculated the Ci values and obtained their rankings (Table 3).

Table 3. Ci values of counties in Gansu province and their rankings at 2015.

4.2.2. Ci Values Calculation and Sorting Results

The Ci value of each county was determined by TOPSIS method. Taking the year of 2015 as an example, we calculated the Ci values and obtained their rankings (Table3).

Table 3. Ci values of counties in Gansu province and their rankings at 2015.

County Ci Rank County Ci Rank County Ci Rank Liang Z 0.7385 1 Jing N 0.4316 31 Xi H 0.3725 61 Gan Z 0.5702 2 Zhuang L 0.4296 32 Lin X (co-) 0.3724 62 Hui N 0.5557 3 Shan D 0.4292 33 Gao L 0.3721 63 Su Z 0.5207 4 Xi F 0.4266 34 Jin C 0.3720 64 Min Q 0.5152 5 Lin Z 0.4242 35 Xia H 0.3698 65 Zhen Y 0.5111 6 Yong D 0.4240 36 Cheng G 0.3632 66 Jing Y 0.4995 7 Zheng N 0.4234 37 Guang H 0.3595 67 Jin T 0.4956 8 Qing S 0.4218 38 Yong J 0.3570 68 Kong T 0.4951 9 Gua Z 0.4217 39 Die B 0.3546 69 Huan 0.4905 10 Gao T 0.4207 40 Su B 0.3541 70 Ning 0.4834 11 Li 0.4191 41 Zhuo N 0.3519 71 An D 0.4812 12 Ma Q 0.4160 42 Liang D 0.3505 72 Lin T 0.4796 13 He S 0.4144 43 Tian Z 0.3501 73 Jing T 0.4730 14 Hui 0.4142 44 Bai Y 0.3480 74 Ling T 0.4694 15 Min 0.4125 45 He Z 0.3479 75 Jing C 0.4625 16 Yu M 0.4096 46 Jia Y G 0.3463 76 Qin A 0.4607 17 Hua C 0.4015 47 Ji S S 0.3405 77 Gu L 0.4568 18 Wei Y 0.4009 48 Ping C 0.3378 78 Gan G 0.4551 19 Chong X 0.4006 49 Xi G 0.3361 79 Wu S 0.4509 20 Kang 0.3970 50 A K S 0.3346 80 Wu D 0.4472 21 Cheng 0.3929 51 Zhou Q 0.3343 81 Mai J 0.4439 22 Dong X Z 0.3902 52 Zhang 0.3342 82 Qing C 0.4426 23 Kang L 0.3859 53 He Z 0.3268 83 Yu Z 0.4410 24 Hua T 0.3857 54 Lin T 0.3232 84 Yong C 0.4409 25 Su N 0.3843 55 Qi L H 0.3168 85 Tong W 0.4404 26 Wen 0.3794 56 Lin X (ci-) 0.2990 86 Min L 0.4396 27 Lu Q 0.3784 57 An N 0.1934 87 Qin Z 0.4354 28 Zhang J C 0.3757 58 ------Long X 0.4342 29 Dang C 0.3753 59 ------Dun H 0.4316 30 Hong G 0.3735 60 ------

The Ci values of most counties were smaller than 0.5000 in 2015, indicating that the potential of rural biomass energy development was generally low in Gansu Province. This agrees with the fact that the ecological environment is poor in Gansu Province and agricultural development lags behind that in many other regions of China. Among all regions of Gansu Province, Liangzhou District had the highest Ci value, reaching 0.7385, far higher than those of other counties. In addition, most of the high-value counties/districts were concentrated in the Hexi Corridor area, such as Liangzhou District, Ganzhou District, Suzhou District, Minqin County and so forth. Among the low-value counties/districts, many such as Linxia city, Lintan county and the Hezuo city (except Anning District, Qilihe District and Xigu District, where the Ci value was too low due to the high level of urbanization) were concentrated in the Gannan area. Therefore, Gannan area can be preliminarily seen as a low potential concentration area.

4.2.3. Ci Value Interannual Variation

The average Ci value of Gansu Province at each year was calculated to observe the temporal variation of Ci values during the evaluation period (Figure4). Sustainability 2018, 10, x FOR PEER REVIEW 11 of 21

Qin A 0.4607 17 Hua C 0.4015 47 Ji S S 0.3405 77 Gu L 0.4568 18 Wei Y 0.4009 48 Ping C 0.3378 78 Gan G 0.4551 19 Chong X 0.4006 49 Xi G 0.3361 79 Wu S 0.4509 20 Kang 0.3970 50 A K S 0.3346 80 Wu D 0.4472 21 Cheng 0.3929 51 Zhou Q 0.3343 81 Mai J 0.4439 22 Dong X Z 0.3902 52 Zhang 0.3342 82 Qing C 0.4426 23 Kang L 0.3859 53 He Z 0.3268 83 Yu Z 0.4410 24 Hua T 0.3857 54 Lin T 0.3232 84 Yong C 0.4409 25 Su N 0.3843 55 Qi L H 0.3168 85 Tong W 0.4404 26 Wen 0.3794 56 Lin X (ci-) 0.2990 86 Min L 0.4396 27 Lu Q 0.3784 57 An N 0.1934 87 Qin Z 0.4354 28 Zhang J C 0.3757 58 ------Long X 0.4342 29 Dang C 0.3753 59 ------Dun H 0.4316 30 Hong G 0.3735 60 ------

4.2.3. Ci Value Interannual Variation

SustainabilityThe2018 average, 10, 3800 Ci value of Gansu Province at each year was calculated to observe the tempora11l of 20 variation of Ci values during the evaluation period (Figure 4).

FigureFigure 4. Interannual4. Interannual variation variation ofof average CCi valuei value (1997 (1997–2015).–2015).

The averageThe averageCi value Ci value of of Gansu Gansu Province Province fluctuatedfluctuated a a lot lot with with time time but but it was it was in a ingeneral a general trend trend of decreasingof decreasing first first and and then then rising. rising. Specifically, Specifically, the the entire entire evaluation evaluation period period can be candivided be dividedinto two into two phases:phases: (1) 1997–2006 is a period of Ci value fluctuating and generally declining. During this period, the (1) 1997–2006 is a period of Ci value fluctuating and generally declining. During this period, the Ci value fluctuated greatly and was in a downward trend. This is because during this period Gansu C value fluctuated greatly and was in a downward trend. This is because during this period Gansu i Province generally paid more attention to the development of the second and third industries and Province generally paid more attention to the development of the second and third industries and the agricultural development was relatively lagged behind, leading to a generally declining trend of agricultural production. (2) 2007–2015 is a period of Ci value steadily increasing. At around 2008, the Chinese government began to make great efforts to solve the “three rural problems” and invested many financial and material resources to solve the rural and agricultural development problems. The government of Gansu Province also gradually recognized the importance of agricultural development and began to pay more attention to agricultural production; thus, the agricultural production capacity of Gansu Province began to rise.

4.3. Spatial and Temporal Analysis

4.3.1. Global Autocorrelation Analysis (1) Global Moran’s I and signiﬁcance test

The global Moran’s I of Ci value from 1997 to 2015 was calculated with GeoDa 0.95i software and significance test was also carried out. The global Moran’s I passed the significance test under the level of 0.01. This shows that there is a significant positive spatial autocorrelation of Ci value in Gansu province at county scale and there is a strong spatial dependence between counties in terms of Ci value (Table4).

Table 4. Global Moran’s I and its signiﬁcance test results (1997–2015).

Year Global Moran’s I E[I] SD p-Value Z-Score 1997 0.2240 −0.0116 0.0446 0.003 5.3215 1998 0.2383 −0.0116 0.0464 0.003 5.3795 1999 0.2188 −0.0116 0.0436 0.002 5.3017 2000 0.2430 −0.0116 0.0455 0.002 5.5394 2001 0.2183 −0.0116 0.0455 0.002 5.1219 2002 0.1903 −0.0116 0.0437 0.003 4.6302 2003 0.1918 −0.0116 0.0454 0.002 4.4970 2004 0.1945 −0.0116 0.0439 0.001 4.6523 2005 0.1827 −0.0116 0.0445 0.002 4.3858 Sustainability 2018, 10, 3800 12 of 20

Table 4. Cont.

Year Global Moran’s I E[I] SD p-Value Z-Score 2006 0.1804 −0.0116 0.0446 0.001 4.3255 2007 0.1805 −0.0116 0.0427 0.001 4.4982 2008 0.1668 −0.0116 0.0431 0.002 4.1407 2009 0.1767 −0.0116 0.0438 0.001 4.2892 2010 0.1720 −0.0116 0.0454 0.002 4.0990 2011 0.1742 −0.0116 0.0460 0.003 4.0056 2012 0.1718 −0.0116 0.0444 0.002 4.1692 2013 0.1708 −0.0116 0.0450 0.003 3.9575 2014 0.1621 −0.0116 0.0447 0.002 3.9047 2015 0.1726 −0.0116 0.0454 0.003 3.9679

(2) Interannual variation of global Moran’s I Although the global Moran’s I ﬂuctuated slightly during the evaluation period, it showed a downward trend in general (Figure5). Especially after 2002, there was an obvious declining trend. This shows that the spatial correlation of biomass energy potential in Gansu Province decreased from 1997 to 2015 and the spatial dependence of various regions in terms of biomass energy development potentialSustainability also decreased. 2018, 10, x FOR PEER REVIEW 13 of 21

FigureFigure 5. Interannual5. Interannual variation variation ofof global Moran’s Moran’s I (1997I (1997–2015).–2015).

(3) Spatial(3) Spatial agglomeration agglomeration characteristics characteristics (Moran (Moran scatter scatter plot) The characteristicsThe characteristics of spatial of spatial agglomeration agglomeration and and the trend the trend of evolution of evolution can be can intuitively be intuitively reflected by Moranreflected scatter by Moran plot. Inscatter this plot. paper, In wethis chosepaper, sixwe timechose points six time (1997, points 2000, (1997, 2004, 2000, 2007, 2004, 20112007, and2011 2015) to observeand 2015) the spatial to observe agglomeration the spatial agglomeration characteristics characteristics and evolution and of evolution the potential of the of biomasspotential energy of biomass energy development in rural areas of Gansu Province (Figure 6). development in rural areas of Gansu Province (Figure6). First, there are more counties in the first and third quadrants in the Moran scatter plot than in First,the other there two are quadrants more counties at each time in the point first (Figure and third 6). In quadrants other words, in high the Moran-high (H scatter-H) agglomeration plot than in the otherand two low quadrants-low (L-L) at each agglomeration time point patterns (Figure are6). Invery other common. words, In high-high contrast, less (H-H) counties agglomeration are in the and low-lowseco (L-L)nd and agglomeration fourth quadrants patterns of the are Moran very scatter common. plot In (Figure contrast, 6). This less countiesmeans the are high in the-low second (H-L) and fourthagglomeration quadrants of and the low Moran-high scatter (L-H) plotagglomeration (Figure6). patterns This means are relatively the high-low less common (H-L) agglomeration in Gansu and low-highProvince. (L-H)In general, agglomeration there is a significant patterns positive are relatively spatial correlatio less commonn of biomass in Gansu energy Province. development In general, therepotential is a significant in Gansu positive Province. spatial Second, correlation the number of of biomass counties energy in the first development and third quadrants potential of in the Gansu Province.Moran Second, scatter plot the tend number to decrease of counties with time in the and first that and in the third third quadrants quadrant seems of the to Moran decrease scatter even plot more. In comparison, the number of counties in the second and fourth quadrants tends to increase tend to decrease with time and that in the third quadrant seems to decrease even more. In comparison, with time. This indicates the decrease of the global Moran’s I of Ci value and the weakening of the positive spatial correlation of biomass energy development potential in Gansu Province. Sustainability 2018, 10, 3800 13 of 20 the number of counties in the second and fourth quadrants tends to increase with time. This indicates the decrease of the global Moran’s I of Ci value and the weakening of the positive spatial correlation of biomassSustainability energy 2018 development, 10, x FOR PEER potential REVIEW in Gansu Province. 14 of 21

Figure 6. Moran scatter plot at each time point. Figure 6. Moran scatter plot at each time point. 4.3.2. Local Autocorrelation Analysis (1) Local4.3.2 Moran’s. Local AutocorrelationI (LISA cluster Analysis map) In(1) order Local toMoran’s further I (LISA highlight cluster the map) spatial heterogeneity of biomass energy development potential at county scaleIn order in Gansu to further Province, highlight LISA the indexspatialwas heterogeneity used to clearly of biomass identify energy the development location where potential spatial agglomerationat county scale of biomass in Gansu energy Province, development LISA index potential was used occurs to clearly and identify intuitively the location reﬂect thewhere evolution spatial of biomassagglomeration energy development of biomass energy potential development in different potential regions occurs of Gansu and Provinceintuitively (Figure reflect the7). evolution Theof biomass spatial energy agglomeration development of biomass potential energy in different development regions of potential Gansu Province occurred (Fi mainlygure 7). in the Hexi area, the GannanThe spatial area agglomeration and the Loess of biomass Plateau energy area of development East Gansu potential Province occurred (Figure 7mainly). The in Hexi the Hexi area is a traditionalarea, the agriculturalGannan area and area the in Loess Gansu Plateau Province. area of Its East agricultural Gansu Province development (Figure 7). leads The Hexi to richness area is of a traditional agricultural area in Gansu Province. Its agricultural development leads to richness of biomass energy resources reserves, thus it is one of the main areas where biomass energy development biomass energy resources reserves, thus it is one of the main areas where biomass energy potential shows a H-H agglomeration pattern. The Loess Plateau of East Gansu Province has a development potential shows a H-H agglomeration pattern. The Loess Plateau of East Gansu ﬂat terrain,Province abundant has a flat sunshine, terrain, abundant large area sunshine, of cultivated large area land of and cultivated rich rural land labor and rich resources, rural labor which provideresources, a good which foundation provide fora good agricultural foundation development. for agricultural Therefore, development. it isTherefore, also an areait is also showing an H-Harea agglomeration showing H-H of agglomeration biomass energy of biomass development energy development potential. In potential. contrast, In the contrast, Gannan the area Gannan shows area shows significant L-L agglomeration of biomass energy development potential and low Ci values Sustainability 2018, 10, 3800 14 of 20

significant L-L agglomeration of biomass energy development potential and low Ci values are not only concentrated but also spread to cover a wide spatial range. The reason is that although animal husbandry in this area is relatively developed, its agricultural development is poor, labor resources are scarce, the topography is uneven and the rural residents live in a scattered way. It is thus difficult to achieve the centralized utilization of biomass energy. It is worth noting that there is an area with obvious H-L agglomeration of biomass energy development potential to the north-west side of the Gannan area. It is a typical transition zone, including several counties with high Ci values, together with the counties with low Ci values in the Gannan area to form a H-L agglomeration pattern of biomass energy development potential. This transition zone should be identified as a key zone to help develop biomass energy in the Gannan area characterized by L-L agglomeration of biomass energy Sustainabilitydevelopment 2018 potential, 10, x FOR PEER in the REVIEW future. 16 of 21

FigureFigure 7. 7. LISALISA cluster cluster map map.. The spatial range of positive correlation (H-H and L-L agglomeration) regions shrunk during the (2) Local Geary’s coefficient evaluation period, while that of negative correlation (H-L and L-H agglomeration) regions fluctuated. TheCompared number of with H-H the agglomeration local Moran’s regions I, the local decreased Geary’s from coefficient 15 in 1997 can todetect 10 in spatial 2015 and agglomeration the number moreof L-L accurately agglomeration [32]. regions decreased even more, from 30 in 1997 to 23 in 2015. It shows that with agriculturalAt each of and the technological six time points, development no region had in a GansuZ value Province, significant the at difference0.01 and 0.0001 in biomass levels (Figure energy 8)development. The Z value potential of most regions between was rural significant regions wasat the gradually level of 0.001. decreasing. These regions were characterized by strongThere spatial was no agglomeration obvious spatial of migrationbiomass energy of agglomeration development pattern potential of biomass but the energyspatial developmentrange of the agglomerationpotential during pattern the evaluation shrunk during period. the Only evaluation the H-H period. agglomeration It shows pattern that with experienced the agricultural a small development in Gansu Province, the regional difference in the potential of biomass energy development is decreasing. In addition, some regions had a Z value significant at 0.05 level. The degree of agglomeration of these regions was not so strong and the spatial distribution of these regions was relatively scattered. Although there has been a decrease in number of regions with agglomeration of biomass energy development potential, the decrease is not significant. By comparing the local Moran’s I with the local Geary’s coefficient, we can see that the spatial agglomeration revealed by the latter is more accurate than that revealed by the former at different significance levels [32]. On the whole, however, the difference is insignificant. Sustainability 2018, 10, 3800 15 of 20 change. This change is mainly reflected in the migration of H-H agglomeration pattern from the eastern part to the central part of Gansu Province. The eastern part of Gansu Province is an important high potential area of biomass energy development. H-H agglomeration pattern is prevalent in this region but gradually weakens. In recent years, the H-H agglomeration pattern has gradually appeared in the central part of Gansu Province and this trend is continuing. This reflects a spatial migration of agglomeration pattern of biomass energy development potential. (2) Local Geary’s coefficient Compared with the local Moran’s I, the local Geary’s coefficient can detect spatial agglomeration more accurately [32]. At each of the six time points, no region had a Z value significant at 0.01 and 0.0001 levels (Figure8). The Z value of most regions was significant at the level of 0.001. These regions were characterized by strong spatial agglomeration of biomass energy development potential but the spatial range of the agglomeration pattern shrunk during the evaluation period. It shows that with the agricultural development in Gansu Province, the regional difference in the potential of biomass energy development is decreasing. In addition, some regions had a Z value significant at 0.05 level. The degree of agglomeration of these regions was not so strong and the spatial distribution of these regions was relatively scattered. Although there has been a decrease in number of regions with agglomeration of biomass energy development potential, the decrease is not significant. By comparing the local Moran’s I with the local Geary’s coefficient, we can see that the spatial agglomeration revealed by the latter is more accurate than that revealed by the former at different significance levels [32]. On the whole, however,Sustainability the 2018 difference, 10, x FOR is PEER insignificant. REVIEW 17 of 21

FigureFigure 8.8. Geary’sGeary’s coefficient coefﬁcient graph graph..

5. Conclusions and Recommendations

5.1. Conclusions On the basis of the calculation of biomass energy resources, this paper considers, for the first time, the influence of agricultural development and geographical environmental factors on biomass energy development and evaluates the actual potential of biomass energy development in Gansu Province. The results indicate that the biomass energy development potential is generally low in Gansu Province. But this paper find that the low development potential in the Gannan area is not due to the low reserves of biomass energy resources but due to the comprehensive influence of factors such as poor agricultural development, scarce labor resources, uneven topography and so forth. In addition, the existence of a transitional zone for biomass energy development is also found for the first time in Gansu Province. This zone is significant for developing the abundant biomass energy resources and improving the biomass energy development potential in the Gannan area. This is the main innovation and contribution of this paper. On this basis, the conclusions of this paper are as follows: (1) In most counties of Gansu Province, the Ci value is smaller than 0.5000, indicating that the biomass energy development potential is generally low. (2) The spatial agglomeration of biomass energy development potential occurs mainly in the Hexi area, the Gannan area and the Loess Plateau area of East Gansu Province. The central region of the Hexi area and the Loess Plateau area of East Gansu Province are the main areas with H-H agglomeration of biomass energy development potential. The Gannan area shows a very significant L-L agglomeration of biomass energy development potential. Sustainability 2018, 10, 3800 16 of 20

5. Conclusions and Recommendations

5.1. Conclusions On the basis of the calculation of biomass energy resources, this paper considers, for the first time, the influence of agricultural development and geographical environmental factors on biomass energy development and evaluates the actual potential of biomass energy development in Gansu Province. The results indicate that the biomass energy development potential is generally low in Gansu Province. But this paper find that the low development potential in the Gannan area is not due to the low reserves of biomass energy resources but due to the comprehensive influence of factors such as poor agricultural development, scarce labor resources, uneven topography and so forth. In addition, the existence of a transitional zone for biomass energy development is also found for the first time in Gansu Province. This zone is significant for developing the abundant biomass energy resources and improving the biomass energy development potential in the Gannan area. This is the main innovation and contribution of this paper. On this basis, the conclusions of this paper are as follows: (1) In most counties of Gansu Province, the Ci value is smaller than 0.5000, indicating that the biomass energy development potential is generally low. (2) The spatial agglomeration of biomass energy development potential occurs mainly in the Hexi area, the Gannan area and the Loess Plateau area of East Gansu Province. The central region of the Hexi area and the Loess Plateau area of East Gansu Province are the main areas with H-H agglomeration of biomass energy development potential. The Gannan area shows a very significant L-L agglomeration of biomass energy development potential. (3) There is an area with obvious H-L agglomeration of biomass energy development potential to the north-west side of the Gannan area. It is a typical transition zone, including several counties with high Ci values, together with counties with low Ci values in the Gannan area to form a H-L agglomeration pattern. This transition zone can play a key role in driving the development of biomass energy in the Gannan area. (4) The spatial range of positive correlation (H-H and L-L agglomeration) regions shrunk during the evaluation period. The number of H-H agglomeration regions decreased from 15 in 1997 to 10 in 2015 and the number of L-L agglomeration regions decreased even more, from 30 in 1997 to 23 in 2015. It shows that with the agricultural and technological development in Gansu Province, the difference in biomass energy development among rural areas is gradually decreasing.

5.2. Recommendations (1) AAA and SMC To develop biomass energy in different regions, we should follow two principles: Act According to One’s Adaptability (AAA) and Suit One’s Measures to Local Conditions (SMC). AAA refers to considering the local agricultural development foundation and the income level of rural residents and then making appropriate choice regarding development scale and development model. SMC refers to taking into account the influence of natural geographical conditions on the development of biomass energy, respecting nature and rationally making plans. We need to consider whether a region has enough natural resources and financial ability to develop biomass energy, which is the premise of the implementation of biomass energy development projects. First, if a region has abundant reserves of biomass energy resources and at the same time the income levels of rural residents enable them to purchase relatively expensive commodity energy, then biomass energy development projects can be carried out in this region. Second, if a region has abundant biomass energy resources but the income level of rural residents is low, commercial biomass energy development projects, such as agricultural recycling project, compressed biomass fuel production project and so forth, are suitable for implementation. These projects can not only help exploit the rich biomass energy resources in the region but also help Sustainability 2018, 10, 3800 17 of 20 increase the income of local farmers. When the income level of local farmers enables then to purchase commodity energy, medium and large-scale biogas projects (MLBPs) can then be carried out. Third, if a region has no abundant biomass energy resources but the income level of rural residents is high, then there is an urgent need for concentrated supply of energy. Based on the good local economic development, modern rural communities can be built to make residents live in a centralized way and then farms and plantations can be built near the communities. By doing these, the daily living needs of community residents can be satisfied. More importantly, organic wastes from farms and plantations can be used to produce biomass energy through MLBPs so that residents can be supplied with stable and clean energy. Fourth, if a region has no abundant biomass energy resources and the income level of rural residents is low, then biomass energy development projects should not be considered in such situation. Furthermore, we must take into account the influence of local natural geographical conditions on biomass energy development and then rationally make plans. If the local area is dominated by flat terrain and the rural residents live in a concentrated way (a large rural community), then a large-scale biomass energy development project might be selected and carried out. If the terrain of the local area is relatively uneven, the residents live in a scattered way and the land for construction is very limited, then it is almost impossible to construct a large-scale rural community and thus medium-sized biomass energy development projects should be selected and carried out. In addition to topography, temperature, precipitation and sunshine duration are also important physical geographical factors that should be considered. These factors will not only affect agricultural production and further affect reserves of biomass energy resources but also affect the development of biomass energy. For example, biogas production (anaerobic fermentation) requires appropriate temperature and adequate sunlight. (2) Growth pole drives biomass energy development in the Gannan area This study reveals that the Gannan area is the largest and most typical area with L-L agglomeration of biomass energy development potential in Gansu Province. However, this is not a result of a lack of reserves of biomass energy resources in this area but a result of its uneven terrain and sparsely distributed population. In other words, the theoretical potential of biomass energy development in the Gannan area is considerable but the actual potential is low due to physical geographical factors. Meanwhile, it is also found that there is an area with obvious H-L agglomeration of biomass energy development potential to the northwestern side of the Gannan area. This is an important and key zone (growth pole) to help transform the theoretical potential of biomass energy development in the Gannan area into actual potential. Therefore, it is necessary to give full play to the growth pole function of this key zone, specifically from the following two aspects: First, biomass energy development projects can be tentatively carried out in this zone. This zone is adjacent to the Gannan area and they have similar physical geographical features. Pilot-scale biomass energy development projects can be first laid out in this zone to find potential problems and accumulate relevant experience. On this basis, the projects can be gradually popularized to the Gannan area. Second, efforts should be made to improve the biomass energy development potential in the Gannan area. Counties with rich biomass energy resources, concentrated distribution of population and high-income level of rural residents should be identified. Then, these regions can be cultivated into areas with high biomass energy development potential. On this basis, the development of biomass energy in the whole Gannan area can be gradually promoted. In combination with the above analysis, it was found that Linxia County, Yongjing County and Zhouqu County all have the potential to become new growth poles of biomass energy development in the Gannan area. (3) Agricultural modernization The development and utilization of biomass energy and agricultural modernization are inseparable. With the progression of agricultural modernization in Gansu Province, agricultural productivity has been improved and many agricultural biomass resources need to be utilized. At the same Sustainability 2018, 10, 3800 18 of 20 time, by concentrated utilization of agricultural biomass resources, agricultural organic wastes can be reduced, which is conducive to the modernization of agriculture. In addition, biomass energy, represented by biogas and BMF, is clean, renewable, stable and affordable, thus its wide application can promote the transformation of rural energy structure in Gansu Province. Furthermore, the use of biomass energy can improve the living and production environment and the quality of life in rural areas. Therefore, the development and utilization of biomass energy must be closely related to the modernization of agriculture. Specifically, we can start from the following aspects. First, Gansu Province should actively introduce modern agricultural production technologies, promote intensive agricultural production and improve the level of agricultural mechanization. Modern agricultural production technologies can not only improve agricultural production capacity and increase the reserves of biomass energy resources but also facilitate the collection and centralized utilization of biomass energy resources so as to easily reduce agricultural organic wastes. Second, Gansu Province should develop recycling agriculture and promote the construction of ecological agriculture (eco-agriculture). Eco-agriculture is a modern and efficient agricultural development model that considers economic, environmental and social benefits. The recycling agriculture is one of the important contents of eco-agriculture. For example, the typical “pig-bog-fruit” project is a relatively common model of ecological recycling agriculture. This model can give full play to the advantages of biomass energy and achieve the goal of “low mining, high utilization, low emission and recycling.” It not only improves the efficiency of biomass energy utilization but also produces good economic, environmental and social benefits. Third, Gansu Province should advocate and popularize modern lifestyles. Modernization of rural lifestyle is also one of the important contents of agricultural modernization and mainly includes three aspects: urbanization of life, community-based living and environmental cleanliness. The realization of these three aspects must be based on a centralized supply of stable and clean commodity energy. Since natural gas can hardly be popularized in rural areas and electricity is expensive, biomass energy, represented by biogas, has become the best choice for centralized supply of commodity energy in rural areas. Therefore, advocating the modernization of lifestyles in rural areas of Gansu Province can not only improve the quality of life but also provides a suitable way to utilize biomass energy. It is beneficial to promote the market-oriented biomass energy development in Gansu Province in an all-round way. This section is not mandatory but can be added to the manuscript if the discussion is unusually long or complex.

Author Contributions: Conceptualization, S.Z. and S.N.; methodology, S.Z. and S.N.; software, S.Z.; validation, S.Z., S.N. and Y.W.; formal analysis, S.Z.; investigation, S.Z., S.N. and Y.W.; resources, S.N.; data curation, S.Z. and S.N.; writing—original draft preparation, S.Z.; writing—review and editing, S.Z. and S.N.; visualization, S.Z.; supervision, S.N.; project administration, S.Z. and S.N.; funding acquisition, S.Z. and S.N. Funding: This resertch was funded by The "Belt and Road" special project, Lanzhou university, China. The grant number is 236000/841040. Conﬂicts of Interest: No conﬂict of interest exits in the submission of this manuscript and manuscript is approved by all authors for publication. I would like to declare on behalf of my co-authors that the work described was original research that has not been published previously and not under consideration for publication elsewhere, in whole or in part. All the authors listed have approved the manuscript that is enclosed.

References

1. Zhao, X.; Luo, D. Driving force of rising renewable energy in China: Environment, regulation and employment. Renew. Sustain. Energy Rev. 2017, 68, 48–56. [CrossRef] 2. Samuel, A.; Edem, K.M.K.; Alfred, A. Renewable and non-renewable energy, regime type and economic growth. Renew. Energy 2018, 125, 755–767. 3. Guozhu, M.; Ning, H.; Lu, C.; Hongmei, W. Research on biomass energy and environment from the past to the future: A bibliometric analysis. Sci. Total Environ. 2018, 635, 1081–1090. Sustainability 2018, 10, 3800 19 of 20

4. Víctor, C.G.; Stefano, M.; Emanuela, P.; Marco, N.; François, M. Optimal use of biomass in large-scale energy systems: Insights for energy policy. Energy 2017, 137, 789–797. 5. Wang, X.J.; Lu, X.G.; Yang, G.H. Development process and probable future transformations of rural biogas in China. Renew. Sustain. Energy Rev. 2016, 55, 703–712. [CrossRef] 6. Franziska, S.; Siwa, M.; Manfred, Z. Policies for a Sustainable Biomass Energy Sector in Malawi: Enhancing Energy and Food Security Simultaneously. World Dev. 2018, 103, 14–26. 7. International Energy Agency (IEA). World Energy Outlook 2016; OECD/IEA: Paris, France, 2016. 8. IPCC. Intergovernmental Panel on Climate Change, Special Reports on Renewable Energy Sources. Summary for Policy Makers. In Proceedings of the 11th Session of Working Group III of the IPCC, Abu Dhabi, United Arab Emirates, 10–13 May 2011. 9. Nikola, B.; Neven, V.; Barbara, J.; Vanja, J.; Ana, M.; Mateja, G.; Tajana, K. Evaluation of Croatian agricultural solid biomass energy potential. Renew. Sustain. Energy Rev. 2018, 93, 225–230. 10. Xu, X.L.; Chen, H.H. Examining the efficiency of biomass energy: Evidence from the Chinese recycling industry. Energy Policy 2018, 119, 77–86. [CrossRef] 11. AEBIOM: Biogas Roadmap for Europe; The European Biomass Association: Brussels, Belgium, 2018. 12. Nicolae, S.; Fernando, F.; Jean-François, D.; Fabio, M.; Vicenzo, M. A spatial analysis of biogas potential from manure in Europe. Renew. Sustain. Energy Rev. 2018, 94, 915–930. 13. Lauri, P.; Havlík, P.; Kindermann, G.; Forsell, N.; Böttcher, H.; Obersteiner, M. Woody biomass energy potential in 2050. Energy Policy 2014, 66, 19–31. [CrossRef] 14. Toklu, E.; Kalogirou, S.A.; Christodoulides, P. Biomass energy potential and utilization in Turkey. Renew. Energy 2017, 107, 235–244. [CrossRef] 15. Ozturk, M.; Saba, N.; Hakeem, K.R.; Jawaid, M.; Ibrahim, F.H. Biomass and Bioenergy: An Overview of the development potential in Turkey and Malaysia. Renew. Sustain. Energy Rev. 2017, 79, 1285–1302. [CrossRef] 16. Suzuki, K.; Tsuji, N.; Shirai, Y.; Hassan, M.A.; Osaki, M. Evaluation of biomass energy potential towards achieving sustainability in biomass energy utilization in Sabah, Malaysia. Biomass Bioenergy 2017, 97, 149–154. [CrossRef] 17. Zhang, T.; Yang, Y.; Xie, D. Insights into the production potential and trends of China’s rural biogas. Int. J. Energy Res. 2015, 39, 227–237. [CrossRef] 18. Zhao, X.; Chang, H.; Peng, S. China’s biomass gas production capacity and carbon emission reduction potential. China Environ. Sci. 2018, 38, 3151–3159. 19. Qin, Z.C.; Zhuang, Q.L.; Cai, X.M.; He, Y.; Huang, Y.; Jiang, D.; Lin, E.; Liu, Y.; Tang, Y.; Wang, M.Q. Biomass and biofuels in China: Toward bioenergy resource potentials and their impacts on the environment. Renew. Sustain. Energy Rev. 2018, 82, 2387–2400. [CrossRef] 20. Wang, Y.; Wu, X.H.; Tong, X.G. Life cycle assessment of large-scale and household biogas plants in northwest China. J. Clean. Prod. 2018, 192, 221–235. [CrossRef] 21. Javad, S.; Fereydoon, S.; Ali, A.J.; Mohammad, A.G.; Farzin, S. Application of SAW, TOPSIS and fuzzy TOPSIS models in cultivation priority planning for maize, rapeseed and soybean crops. Geoderma 2018, 310, 178–190. 22. Mohammad, Y. Risk assessment based on novel intuitionistic fuzzy-hybrid-modified TOPSIS approach. Saf. Sci. 2018, 110, 438–448. 23. Han, H.; Silvana, T. A fuzzy TOPSIS method for performance evaluation of reverse logistics in social commerce platforms. Expert Syst. Appl. 2018, 103, 133–145. [CrossRef] 24. Mateusz, P.; Danuta, M.; Małgorzata, Ł.; Mariusz, B.; Kesra, N. TOPSIS and VIKOR methods in study of sustainable development in the EU countries. Procedia Comput. Sci. 2018, 126, 1683–1692. [CrossRef] 25. Dan, Y.; Lan, Z.; Caihong, Z. Evaluation of Regional Development potential of Forest Biomass Energy based on Entropy weight TOPSIS. J. Beijing For. Univ. (Soc. Sci.) 2016, 15, 50–55. 26. Bruno, M.S.; Leoni, P.G.; Lucila, M.S.C. Performance Evaluation of Green Suppliers using Entropy-TOPSIS-F. J. Clean. Prod. 2018.[CrossRef] 27. Zyoud, S.H.; Fuchs-Hanusch, D. A bibliometric-based survey on AHP and TOPSIS techniques. Expert Syst. Appl. 2017, 78, 158–181. [CrossRef] 28. Sun, G.; Guan, X.; Yi, X.; Zhou, Z. An innovative TOPSIS approach based on hesitant fuzzy correlation coefficient and its applications. Appl. Soft Comput. 2018, 68, 249–267. [CrossRef] Sustainability 2018, 10, 3800 20 of 20

29. Meng, D.; Shao, C.; Zhu, L. Ethylene Cracking Furnace TOPSIS Energy Efficiency Evaluation Method Based on Dynamic Energy Efficiency Baselines. Energy 2018, 156, 620–634. [CrossRef] 30. Wang, E.; Alp, N.; Shi, J.; Wang, C.; Zhang, X.; Chen, H. Multi-criteria building energy performance benchmarking through variable clustering based compromise TOPSIS with objective entropy weighting. Energy 2017, 125, 197–210. [CrossRef] 31. Hossny, A.; Valid, H. Dataset normalization for low carbon cities in a multi-criteria evaluation model. Data Brief 2018, 18, 1111–1116. 32. Jing, R.; Zhu, X.; Zhu, Z.; Wang, W.; Meng, C.; Shah, N.; Li, N.; Zhao, Y. A multi-objective optimization and multi-criteria evaluation integrated framework for distributed energy system optimal planning. Energy Convers. Manag. 2018, 166, 445–462. [CrossRef] 33. Wang, Z.; Liang, L.; Fang, C.; Zhuang, R. Spatial and temporal Evolution characteristics and influencing factors of Ecological Security pattern of Beijing-Tianjin-Hebei Urban agglomeration. Acta Ecol. Sin. 2018, 38, 4132–4144. 34. Yuan, J.; Zhang, J.; Wang, T.; Du, J. Comprehensive Evaluation and weight sensitivity Analysis of Energy Saving in Thermal Power Plant based on TOPSIS Grey correlation projection method. J. Chin. Soc. Power Eng. 2015, 35, 404–411. 35. Ümran, ¸S.;Eren, M.; Shiraz, S.E.; Gezder, V.; ¸Sengül,A.B. Fuzzy TOPSIS method for ranking renewable energy supply systems in Turkey. Renew. Energy 2015, 75, 617–625. 36. Balducci, F.; Ferrara, A. Using urban environmental policy data to understand the domains of smartness: An analysis of spatial autocorrelation for all the Italian chief towns. Ecol. Indic. 2018, 89, 386–396. [CrossRef] 37. Melecky, L. Spatial Autocorrelation Method for Local Analysis of The EU. Procedia Econ. Financ. 2015, 23, 1102–1109. [CrossRef] 38. Chen, J.; Xu, C.; Song, M.; Liu, X. Driving factors of China’s energy productivity and its spatial character: Evidence from 248 cities. Ecol. Indic. 2018, 90, 18–27. [CrossRef]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).