A Case Study in the Mentougou District in Beijing

sustainability

Article Changes of Ecosystem Services and Landscape Patterns in Mountainous Areas: A Case Study in the Mentougou District in Beijing

Yang Yi 1,2,3 , Mingchang Shi 3,*, Chunjiang Liu 1,2,*, Bin Wang 3 , Hongzhang Kang 1,2 and Xinli Hu 1,2 1 School of Agriculture and Biology, Shanghai Jiao Tong University, Shanghai 200240, China; [email protected] (Y.Y.); [email protected] (H.K.); [email protected] (X.H.) 2 Shanghai Urban Forest Ecosystem Research Station, State Forestry Administration, Shanghai 200240, China 3 Beijing Engineering Research Center of Soil and Water Conservation, Beijing Forestry University, Beijing 100083, China; [email protected] * Correspondence: [email protected] (M.S.); [email protected] (C.L.); Tel.: +86-10-62336614 (M.S.)

Received: 16 August 2018; Accepted: 27 September 2018; Published: 15 October 2018

Abstract: Land use types have been strongly modiﬁed across mountainous areas. This has substantially altered the patterns and processes of ecosystems and the components of ecosystem services (ESs), and could in turn impact the sustainable development. In the mountainous Mentougou district of Beijing, we explored the changes in land use type (cropland, orchard, forested land, scrubland, grassland, bare land, water bodies, wasteland and built-up land), landscape patterns and ESs as well as their interactions during the past 30 years (1985–2014). The ESs included water yield (WY), carbon stocks (CS) and soil retention rate (SR). The results showed that 23.65% of the land use changed and the wasteland decreased by 80.87%. As for ESs, WY decreased by 47.32% since the year 2000, probably due to the increases in temperature and evapotranspiration. Although the decrease of forested land led to the decrease of CS, the increase of vegetation coverage improved SR. CS decreased by 0.99%from 1990 to 2014, and SR increased by 1.38% from 1985 to 2014. Landscape patterns became fragmented and dispersed, and MPS and CS, SHDI and SR were signiﬁcantly negatively correlated. IJI and CS was positively correlated. This indicated that landscape patterns were highly correlated with ESs. In order to maintain the sustainable development of ESs, we should not only plan land use types, but also consider the rationality of landscape patterns.

Keywords: ecosystem services; landscape patterns; land use types; mountainous areas; Mentougou

1. Introduction Ecosystem services (ESs) are the beneﬁts or goods that human beings receive, directly or indirectly, from ecosystems [1]. These transactions, emerging from natural environments, generate the needed conditions for human survival [2,3]. ESs are an important bridge between natural ecosystems and human well-being. Global population growth, with the associated socioeconomic development, is a major driver of the land use change (e.g., through agricultural activities, built-up areas and mining, etc.) and ecosystems (e.g., species diversity, climate change, water quality, etc.). This simultaneously changes the energy ﬂow and material cycle of the ecosystems, thereby affecting ESs and human well-being [4,5]. The Millennium Ecosystem Assessment estimates that more than half of the ESs are fading, and this trend will not slow down [3]. Landscape patterns are an indicator of the ecological process [6]. The interactions between landscape patterns and ecological processes drive the overall dynamics of the landscape and present certain landscape functional characteristics [7,8]. This function is related to human needs and

Sustainability 2018, 10, 3689; doi:10.3390/su10103689 www.mdpi.com/journal/sustainability Sustainability 2018, 10, 3689 2 of 17 constitutes the core of human life support system. Changes to landscape patterns are mainly reflected in the change of the type, shape, size, quantity and space combination of the landscape patch, which affects the ecological processes [9,10]. In the last 20 years, the scientific problems related to ESs had attracted the attention of many scholars in the field of geography, ecology and economics [11]. At present, the research is mainly focused on the impact of land use types and climate change on ESs [12]. In fact, landscape patterns will profoundly affect ESs. The changes in landscape composition directly effect of ESs [13]. Landscape pattern changes indirectly affect ESs by changing the ecological processes [14]. At the same time, the transformation of ESs is reflected not only in the changes of landscape space, but also in the spatial and temporal changes of the material circulation and energy flow in the landscape [5,13,15]. In mountainous areas, there are abundant natural resources such as forests, minerals and water, which provide the basic material living conditions for 10% of the people in the world [16,17]. In China, the mountainous areas account for 69% of the land area where 56% of the population live [18,19]. Mountainous areas provide a variety of ESs, such as carbon sequestration, water conservation, and plant resources, etc. [20,21]. With the rapid development of the global economy, environmental problems are becoming more and more serious in mountainous regions [15]. This not only leads to ecological problems, but also hinders economic development in mountainous areas [22,23]. Mentougou district is located in the northwestern part of Beijing municipality, and of sixteen districts, it is the only one for which the whole territory is mountainous. This district is considered as an important ecological barrier for Beijing, playing an essential role in carbon sequestration, biodiversity protection, water conservation, and recreation [24]. During the past 30 years, Mentougou district has been under the pressure of the rapid urbanization and climate change, leading to ecological and environmental problems, such as air pollution, the expansion of built-up areas, and an excessive number of tourists [25]. In order to protect the environment in mountainous areas in Beijing, which is the capital, a series of ecological projects (e.g., Grain for Green, Controlling Area of Wind and Sand Sources, and Ecological Restoration of Rivers, etc.) were adopted by the government [26,27]. With these projects, the functioning of ecosystems was greatly improved in Mentougou district. This paper evaluated the temporal and spatial changes of ESs (water yield, carbon stocks and soil retention rate) in relation to the changes in landscape patterns. The objectives of this study were, (1) to quantify the changes of land use types, weather and socio-economy in the study area, (2) to examine the spatial distribution and quantitative changes of ESs and landscape patterns, and (3) to explore the relationship between ESs and landscape patterns.

2. Study Area and Methods

2.1. Study Area Mentougou district is located in the north-western part (115◦250 E–116◦100 E, 39◦480 N–40◦100 N) of the Beijing in North China and covers a total area of 1455 km2 (Figure1). This area has a mid-latitude continental monsoon climate, and the climate in the eastern plains and western mountainous areas is very different [28]. The average temperature in the eastern plain is 11.7 ◦C, and the average temperature in the west is 10.2 ◦C The average annual precipitation is 520.13 mm (1985–2014 a). The annual average sunshine time is about 2470 h. The precipitation increases gradually from the west to the east. Due to the inﬂuence of the mid latitude atmospheric circulation and the monsoon, precipitation changes greatly [29]. Sustainability 2018, 10, x 3689 FOR PEER REVIEW 33 of of 17 17

FigureFigure 1. 1. LocationLocation of of Mentougou Mentougou district district and and topography. topography. In Mentougou, the main native tree species are Quercus variabilis, Pinus tabuliformis Carr., In Mentougou, the main native tree species are Quercus variabilis, Pinus tabuliformis Carr., Platycladus orientalis (L.) Franco, Larix gmelinii Rupr, etc. [28]. The forests are generally distributed in Platycladus orientalis (L.) Franco, Larix gmelinii Rupr, etc. [28]. The forests are generally distributed in the mountainous areas above the altitude of 1000 m. In areas lower than 1000 m, plant communities the mountainous areas above the altitude of 1000 m. In areas lower than 1000 m, plant communities are dominated by scrubs (e.g., Oxytropis aciphylla Ledeb., Lespedeza bicolor Turcz. and Hibiscus are dominated by scrubs (e.g., Oxytropis aciphylla Ledeb., Lespedeza bicolor Turcz. and Hibiscus syriacus L., etc.) [30]. The soils are divided into three main soil types (mountain meadow, mountain syriacus L., etc.) [30]. The soils are divided into three main soil types (mountain meadow, mountain brown soil and cinnamon soil) (Figure2). brown soil and cinnamon soil) (Figure 2).

Figure 2. SoilSoil distribution distribution of Mentougou Mentougou district.

Sustainability 2018, 10, 3689 4 of 17 Sustainability 2018, 10, x FOR PEER REVIEW 4 of 17

2.2. Land Use Classification Classification and Sub-WatershedSub‐Watershed Division In this study, wewe usedused datadata fromfrom thethe LandsatLandsat Operational Land Imager (OLI) (bands 1, 2, 4, 5, and 7) remote sensing images for nearly 30 years (1985, 1990, 1995, 2000, 2005, 2005, 2010 2010 and and 2014) 2014) [31,32]. [31,32]. We referredreferred toto thethe landland useuse inin Beijing,Beijing, andand thethe ChineseChinese classificationclassification system, and classifiedclassified the land use, into nine types via supervised classification,classification, using raster maps generated at a 30 m spatial resolution [[27,33].27,33]. The The land land use included cropland, orchard, forested land, scrubland, grassland, bare land, water bodies, wasteland (including abandoned abandoned lands, lands, bare bare soil and bare vegetation and sparse vegetation.)vegetation.) and and built built-up‐up land. land. We We chose chose 450 450 point point (each (each land land use use type type had had 50 points) representative points distributed throughout the study area using stratifiedstratified random sampling to examine thethe accuracy accuracy of theof interpretationthe interpretation using data using from data Google from Earth Google (Google Earth Earth high-resolution(Google Earth images)high‐resolution (10 × 10 images) m) (before (10 2005)× 10 m) and (before field investigations 2005) and field (after investigations 2005), all above (after 85%. 2005), We all also above used 85%. the ArcSWATWe also used software the toArcSWAT divide the software Mentougou to divide into 39 the sub-watersheds, Mentougou whichinto 39 provided sub‐watersheds, the basis forwhich the calculationprovided the of basis the ESs for and the thecalculation landscape of the patterns ESs and (Figure the landscape3). patterns (Figure 3).

FigureFigure 3. SubSub-watersheds‐watersheds of of Mentougou Mentougou district (made by ArcSWAT software).software). 2.3. Quantiﬁcation of Landscape Patterns 2.3. Quantification of Landscape Patterns To assess changes in the structural characteristics of landscape patterns at the sub-watershed To assess changes in the structural characteristics of landscape patterns at the sub‐watershed level, we selected the mean patch size (MPS), largest patch index (LPI), interspersion and juxtaposition level, we selected the mean patch size (MPS), largest patch index (LPI), interspersion and index (IJI), Shannon’s diversity Index (SHDI) and aggregation index (AI), to characterize landscape juxtaposition index (IJI), Shannonʹs diversity Index (SHDI) and aggregation index (AI), to patterns [27,34]. All calculations were extracted from the FRAGSTATS 4.2 software manual. characterize landscape patterns [27,34]. All calculations were extracted from the FRAGSTATS 4.2 2.4.software Methods manual. to Quantify the Ecosystem Services

2.4. MethodsConsidering to Quantify the local the conditions,Ecosystem Services the feasibility of the assessment method, and the availability of data, we chose the subset of the InVEST model and focused on reporting the services for water Considering the local conditions, the feasibility of the assessment method, and the availability yield and carbon stocks. We used RUSLE (universal soil loss equation) model to simulate the soil of data, we chose the subset of the InVEST model and focused on reporting the services for water retention rate and employed the InVEST model to estimate the ESs according to the land use types and yield and carbon stocks. We used RUSLE (universal soil loss equation) model to simulate the soil biophysical data (digital elevation model, climate regulation, potential evapotranspiration, soil types retention rate and employed the InVEST model to estimate the ESs according to the land use types and soil roughness). The simulation of ESs were based on sub-watersheds. The InVEST model clearly and biophysical data (digital elevation model, climate regulation, potential evapotranspiration, soil indicated that water yield must be calculated in sub-watersheds, which was convincing and accurate, types and soil roughness). The simulation of ESs were based on sub‐watersheds. The InVEST model compared with calculations using raster units [35]. Therefore, in order to analyze the relationship clearly indicated that water yield must be calculated in sub‐watersheds, which was convincing and between landscape patterns, the value of the carbon storage and soil retention rate was ﬁnally selected accurate, compared with calculations using raster units [35]. Therefore, in order to analyze the to be sub-watersheds. relationship between landscape patterns, the value of the carbon storage and soil retention rate was finally selected to be sub‐watersheds.

Sustainability 2018, 10, 3689 5 of 17

2.4.1. Water Yield The water yield model in the InVEST model is a simpliﬁed hydrologic model without considering the surface, underground and basal ﬂow. Based on the theoretical basis of water balance, the Budyko curve and annual precipitation were used for calculation [36]. The water yield of each space unit was equal to the difference between the annual precipitation and actual evapotranspiration. Therefore, the water yield in this model included not only the surface runoff, but also soil water content, canopy interception and litter water holding capacity. Yxj = 1 − AETxj/Px × Px (1) AETxj/Px = 1 + ωx + Rxj / 1 + ωx + 1/Rxj (2)

ωx = Z × (AWXx/Px) (3) Rxj = kxj × ETox /Px (4) where AETxj is the actual annual evapotranspiration and Px is the average annual precipitation. Where Yjx is the annual water yield; AETxj/Px approximates the Budyko curve developed by Zhang et al. [37]. Rxj is the Budyko aridity index; Kxj is the coefﬁcient of vegetation evapotranspiration; ET0x is the potential evapotranspiration; and ωx is a dimensionless ratio characterizing the natural climate and soil properties.

2.4.2. Soil Retention Rate The equation (RUSLE) can predict the actual annual amount of soil erosion [38]. The amount of soil loss will be blocked by vegetation, the reduction of soil loss caused by vegetation is deﬁned as soil retention (SEDR). Soil retention was expressed by the subtraction between the potential soil loss under no vegetation coverage (SEDRET) and the actual soil loss (RUSLE). Soil retention is inﬂuenced by rainfall intensity, slope and slope length, thus soil retention cannot effectively explain the contribution of the ecosystem services to soil conservation [39]. In order to analyze real soil conservation function of the ecosystem, we used the rate of soil retention and soil loss under no vegetation coverage (soil retention rate) as a measure of the ecosystem contribution to soil conservation.

RUSLE = R × K × LS × C × P (5)

SEDRET = R × K × LS (6)

SEDR = SEDRET − USLE (7) SEDR F = × 100% (8) SEDRET where RUSLE is the amount of soil erosion, SEDRET is the potential soil loss under no vegetation coverage, SEDT is the soil loss under vegetation coverage; R is the rainfall erosivity, K is the soil erodibility factor, LS is the slope length-gradient factor, C is the crop management factor and P is the support practice factor.

2.4.3. Carbon Stocks The InVEST model quantified the amount of carbon stored and sequestered based on four carbon pools: below ground biomass, above ground biomass, soil organic matter and dead organic matter [40,41]. The carbon stocks in vegetation were estimated by multiplying vegetation carbon density by the area of each land use type which were derived from the results of local studies. The vegetation carbon density was cropland (30.96 tC·ha−1), orchard (30.96 tC·ha−1), forested land (150.40 tC·ha−1), scrubland (118.42 tC·ha−1), grassland (96.68 tC·ha−1), bare land (10.00 tC·ha−1), Sustainability 2018, 10, 3689 6 of 17 waterSustainability bodies 2018 (0, 10 tC, x FOR·ha− PEER1), wasteland REVIEW (96.68 tC·ha−1) and built-up land (0 tC·ha−1), respectively,6 of 17 according to Huang [30]. 2.5. Statistical Analysis 2.5. StatisticalWe used Analysis data from seven periods and 39 sub‐watersheds to analyze the relationship of landscapeWe used patterns data from and sevenESs (WY, periods CS, andSR), 39 a total sub-watersheds of 273 data. to Multivariate analyze the relationshipanalysis was of performed landscape patternsto explore and the ESs influence (WY, CS, of SR), the a landscape total of 273 pattern data. Multivariate factors (MPS, analysis LPI, IJI, was SHDI performed and AI) to on explore the ESs. the influenceAfter standardization, of the landscape a pattern redundancy factors (MPS, analysis LPI, IJI,(RDA) SHDI andwas AI) applied on the ESs.(the After largest standardization, detrended acorrespondence redundancy analysis analysis (RDA) gradient was appliedlength values (the largest were detrended all shorter correspondence than 3.0). After analysis 499 randomly gradient lengtharranged values Monte were Carlo all shorter displacement than 3.0). tests, After the 499 significance randomly arranged of all normalized Monte Carlo axis displacement eigenvalues tests,was theobtained. significance We performed of all normalized RDA and axis acquired eigenvalues four was discriminant obtained. functions We performed for ESs RDA and and landscape, acquired fourrespectively. discriminant The functionsfirst two forfunctions ESs and were landscape, extracted respectively. and used The in bi first‐plots. two The functions correlation were extracted analysis and usedRDA in were bi-plots. conducted The correlation using IBM analysis SPSS and (version RDA were20, Chicago, conducted IL, using USA), IBM and SPSS the (version RDA was 20, Chicago,performed IL, using USA), CANOCO and the RDA (version was 4.5, performed Ithaca, NY, using USA). CANOCO (version 4.5, Ithaca, NY, USA).

3. Results

3.1. Land Use Types, Weather andand SocioeconomicSocioeconomic ChangesChanges fromfrom 19851985 toto 20142014 inin MentougouMentougou

The study area experienced 344.10 km2 of land change, and 23.65% of the land changed from 1985 to 2014 (Figure4 4).). ForestedForested landland andand scrublandscrubland accountedaccounted forfor moremore thanthan 75%75% ofof thethe studystudy area. The dynamicdynamic change change in in land land use use in thein lastthe 30last years, 30 years, from largestfrom largest to smallest, to smallest, it was wasteland it was wasteland (80.87%), cropland(80.87%), cropland (66.97%), (66.97%), bare land bare (41.63%), land (41.63%), built-up built land‐up (32.40%), land (32.40%), scrubland scrubland (27.78%), (27.78%), water bodies water (11.80%),bodies (11.80%), forested forested land (7.18%), land (7.18%), orchard orchard (6.93%) (6.93%) and and grassland grassland (5.33%). (5.33%). The The trend trend of of land land use changes was was the the continuous continuous expansion expansion of of built built-up‐up land land and and shrunken shrunken of cropland. of cropland. At the At same the sametime, time,the wasteland the wasteland decreased decreased significantly. signiﬁcantly. Orchards Orchards continued continued to todecrease decrease before before 2000, 2000, and and increased increased after 2000. Scrubland continued to increase. Bare Bare land, land, water water bodies bodies and and forested land decreased, and grassland increased slightly.

Figure 4. Changes in land use types from 1985 to 2014. The data in the color histogram represent the area of land use types in the study period (1985, 1990, 1995, 2000, 2005, 2010 and 2014) from left to right. Figure 4. Changes in land use types from 1985 to 2014. The data in the color histogram represent the BUL refers to built-up land, OC to orchard, CL to cropland, WL to wasteland, WB to water bodies, area of land use types in the study period (1985, 1990, 1995, 2000, 2005, 2010 and 2014) from left to GL to grassland, BL to bare land, FL to forested land, and SL to scrubland. right. BUL refers to built‐up land, OC to orchard, CL to cropland, WL to wasteland, WB to water Duringbodies, GL the to study grassland, period, BL to the bare wasteland land, FL decreasedto forested land, by 80.87% and SL (101.66 to scrubland. km2), of which 66.83% was converted to scrubland and 8.84% was converted to orchards (Table1). Built-up land has increased During the study period, the wasteland decreased by 80.87% (101.66 km2), of which 66.83% was by 32.40% (21.11 km2). This increased built-up land are mainly expanded by occupying cropland converted to scrubland and 8.84% was converted to orchards (Table 1). Built‐up land has increased (15.49 km2), orchard (2.58 km2) and wasteland (2.42 km2). During the study period, the scrubland by 32.40% (21.11 km2). This increased built‐up land are mainly expanded by occupying cropland tended to increase (224.94 km2), mainly through the conversion of other land use types. The amount (15.49 km2), orchard (2.58 km2) and wasteland (2.42 km2). During the study period, the scrubland tended to increase (224.94 km2), mainly through the conversion of other land use types. The amount of forested land showed a decreasing trend (120.46 km2), mainly due to its conversion to scrubland. The water area ﬂuctuated little. Sustainability 2018, 10, 3689 7 of 17 of forested land showed a decreasing trend (120.46 km2), mainly due to its conversion to scrubland. The water area ﬂuctuated little.

Table 1. Transition matrix of changes of land use type between 1985 (columns) and 2014 (rows) (km2).

Sustainability 2018, 10, x FOR PEER REVIEW 7 of 17 BL OC WB CL GL WL BL FL SL Total BL 65.16Table 1. Transition matrix of changes of land use type between 1985 (columns) and 2014 (rows) 65.16 OC(km 2.582). 39.59 1.77 13.23 57.17 WB 0.62 16.42 0.05 0.94 0.62 18.65 BL OC WB CL GL WL BL FL SL Total CL 15.49 2.51 7.98 0.12 3.96 30.06 BL 65.16 65.16 GL 13.79 3.28 17.07 OC 2.58 39.59 1.77 13.23 57.17 WL 2.42WB 11.11 0.62 16.42 0.13 0.05 0.94 1.73 0.62 23.3 0.37 2.6418.65 84.01 125.71 BLCL 15.49 2.51 7.98 1.240.12 0.853.96 30.06 2.09 FL GL 13.79 3.28 479.1217.07 120.46 599.58 SLWL 2.42 11.11 0.03 0.13 1.73 23.3 0.37 2.64 84.01 74.79125.71 464.57 539.39 Total 86.27BL 53.21 16.45 9.931.24 17.70 24.040.85 1.22 556.552.09 689.51 1454.88 FL 479.12 120.46 599.58 SL 0.03 74.79 464.57 539.39 The weatherTotal and socio-economic86.27 53.21 16.45 trends 9.93 in17.70 the Mentougou24.04 1.22 556.55 were analyzed689.51 1454.88 with time series data between 1985 and 2014 obtained from the statistical yearbook throughout the study region (Figure5). The regionalThe weather weather conditionsand socio‐economic of the trends Mentougou in the Mentougou showed were a warming analyzed trend. with time Temperatures series data rose between 1985 and 2014 obtained from the statistical yearbook throughout the study region (Figure steadily over the past 30 years, with an average increase rate of 0.0923 ◦C/year (p < 0.01). During the 5). The regional weather conditions of the Mentougou showed a warming trend. Temperatures rose −1 same period,steadily the over estimated the past 30 precipitation years, with an average also increased increase rate 2.0069 of 0.0923 mm ·°C/yeara . Overall, (p < 0.01). the During weather the in the study areasame showed period, athe warming estimated and precipitation wetting trend.also increased The GDP 2.0069 showed mm∙a−1 a. Overall, signiﬁcant the weather exponential in the growth, with anstudy average area increase showed a of warming 1.1658 timesand wetting each year trend. (p The< 0.01). GDP Theshowed registered a significant population exponential decreased from 247,200growth, in with 1985 an to average 233,900 increase in 2000, of than1.1658 increased times each to year 249,098 (p < 0.01). in 2014. The Theregistered increase population in population indicateddecreased the increase from 247,200 in demand in 1985 for to 233,900 built-up in 2000, land than and increased cropland, to which249,098 hadin 2014. a direct The increase impact in on land population indicated the increase in demand for built‐up land and cropland, which had a direct use types,impact and on indirectly land use types, affected and ESs.indirectly affected ESs.

Figure 5.FigureTemporal 5. Temporal variation variation in the in the weather weather (mean (mean annual precipitation, precipitation, mean mean annual annual temperature), temperature), and socio-economicand socio‐economic indicators indicators (gross (gross domestic domestic production,production, registered registered population) population) in Mentougou in Mentougou district from 1985 to 2014. district from 1985 to 2014. Sustainability 2018, 10, x FOR PEER REVIEW 8 of 17 Sustainability 2018, 10, 3689 8 of 17 3.2. Changes in Landscape Patterns

3.2. ChangesLPI generally in Landscape showed Patterns a decreasing trend before 2010, decreasing by 3.82%, and a slight increased after 2010. The MPS was 21.12 ha in 1985, decreasing to 19.45 ha in 2005, and increasing to LPI generally showed a decreasing trend before 2010, decreasing by 3.82%, and a slight increased 20.03 ha in 2014. This showed that the landscape patterns gradually fragmented. IJI and SHDI after 2010. The MPS was 21.12 ha in 1985, decreasing to 19.45 ha in 2005, and increasing to 20.03 ha continued to decrease by 16.07% and 10.71% respectively, from 1990 to 2014. It showed that the in 2014. This showed that the landscape patterns gradually fragmented. IJI and SHDI continued to landscape patterns showed increasingly typical zonal distribution characteristics, but the diversity decrease by 16.07% and 10.71% respectively, from 1990 to 2014. It showed that the landscape patterns of landscape patterns was decreasing gradually. The AI generally showed a decreased trend, showed increasingly typical zonal distribution characteristics, but the diversity of landscape patterns indicating that the landscape patterns in the study area were gradually dispersed (Table 2). was decreasing gradually. The AI generally showed a decreased trend, indicating that the landscape patterns in the studyTable area 2. Landscape were gradually indices dispersed of Mentougou (Table2 ).district from 1985 to 2014.

Table 2. LandscapeLPI (%) indices MPS of Mentougou(ha) IJI (%) district SHDI from 1985AI (%) to 2014. 1985 14.10 21.12 58.75 1.39 97.07 LPI (%) MPS (ha) IJI (%) SHDI AI (%) 1990 14.24 19.71 63.54 1.40 97.08 19851995 14.10 14.14 21.1219.48 63.04 58.75 1.36 1.3997.07 97.07 1990 14.24 19.71 63.54 1.40 97.08 2000 11.30 19.55 59.09 1.34 97.03 1995 14.14 19.48 63.04 1.36 97.07 20002005 11.30 11.83 19.5519.45 58.53 59.09 1.31 1.3497.08 97.03 20052010 11.83 10.82 19.4519.94 53.57 58.53 1.27 1.3196.86 97.08 20102014 10.82 11.93 19.9420.03 53.33 53.57 1.25 1.2796.96 96.86 2014 11.93 20.03 53.33 1.25 96.96 The high value of MPS was usually distributed in the central region, usually more than 20 ha. The Thehigh high value value of LPI of was MPS located was usually in the distributedsoutheastern in part the centralof Mentougou region, (Figure usually 6). more After than 1995, 20 ha.the TheLPI high in the value western of LPI region was located was increased. in the southeastern Most of the part regions of Mentougou had a high (Figure value6). area After of 1995, the IJI, the while LPI inthe the low western value region areas waswere increased. mainly distributed Most of the regionsin some had parts a high of the value central area of and the northern IJI, while regions, the low valueusually areas below were 40%. mainly The distributeddistribution in of some the IJI parts in most of the regions central showed and northern a trend regions, with first usually an increase below 40%.from The 1985 distribution to 2005 and of thethen IJI a in decrease most regions from showed 2005 to a trend2014. withThe ﬁrsthigh an value increase of SHDI from 1985was tomainly 2005 anddistributed then a decrease in the eastern from 2005 region, to 2014. usually The highmore value than of1.0. SHDI The was high mainly value distributedof AI was inin thethe eastern central region,region. usually more than 1.0. The high value of AI was in the central region.

Figure 6. Cont. Sustainability 2018,, 10,, 3689x FOR PEER REVIEW 9 of 17 Sustainability 2018, 10, x FOR PEER REVIEW 9 of 17

Figure 6. The landscape patterns in sub‐watersheds from 1985 to 2014. Each row represents the Figure 6. The landscape patterns in sub-watersheds from 1985 to 2014. Each row represents the change changeFigure of6. differentThe landscape landscape patterns indices in insub sub‐watersheds‐watersheds from in the1985 same to 2014.year. EachEach rowcolumn represents represents the of different landscape indices in sub-watersheds in the same year. Each column represents the change thechange change of differentof the landscape landscape indices indices in sub in ‐subwatersheds‐watersheds in different in the same years. year. Each column represents of the landscape indices in sub-watersheds in different years. the change of the landscape indices in sub‐watersheds in different years. 3.3. Changes in Multiple Ecosystem Services 3.3. Changes in Multiple Ecosystem Services 3.3.1. Changes in Water Yield 3.3.1. Changes in Water Yield The water yield (WY) decreased by 44.20% from 1985 to 2014 (Figure7 7).). TheThe WYWY reachedreached thethe highestThe value water in yield 2010. (WY) This This decreasedis is probably probably by due due44.20% to the from weather 1985 to in 2014 this (Figure period 7).(high The rainfall WY reached and low the temperaturehighest value inducedinduced in 2010. evapotranspiration). evapotranspiration). This is probably Indue In addition, addition, to the theweather the WY WY in inin 2000 2000this was periodwas the the second(high second rainfall highest highest and value value low of theoftemperature the study study period period induced (5.33 (5.33 evapotranspiration).× ×10 108 m8 m3).3). The The amount amount In addition, ofof WYWY was wasthe generallyWYgenerally in 2000 increasedincreased was the beforesecondbefore 2000, highest and value the WYof the was study decreased period after (5.33 2000. × 10 8 In m 2014,3). The the amount WY reached of WY the was lowest generally value increased of the whole before research 2000, and period. the ItWY decreaseddecreased was decreased 3.313.31 ×× 10after88 mm 2000.33 (62.10%)(62.10%) In 2014, compared compared the WY reachedwith with 2000. 2000. the The Thelowest significant signiﬁcant value of decrease decreasethe whole in in research WY might period. be relatedIt decreased to the 3.31 changes × 10 in8 m land3 (62.10%) use type compared and weather. with From 2000. The1985 significant to 2000, the decrease average in precipitation WY might inbe Mentougourelated to the ranged changes from in land480.0 use mm type to 594.8and weather. mm, then From signiﬁcantlysignificantly 1985 to 2000, decreased the average to 264.2 precipitation mm in 2014 in (FigureMentougou7 7).). Meanwhile,Meanwhile, ranged from thethe 480.0 annualannual mm average averageto 594.8 potential potentialmm, then evapotranspiration evapotranspirationsignificantly decreased inin MentougouMentougou to 264.2 mm droppeddropped in 2014 from(Figure 776.30 7). Meanwhile, mm to 721.88 the mm annual from 1985average to 2000, potential then increasedevapotranspiration to 763.24 mm in Mentougouin 2014. The increasedropped offrom potential 776.30 evapotranspirationmm to 721.88 mm fromwas related 1985 to to 2000, the thenincrease increased in temperature to 763.24 mmand inthe 2014. changes The increase in land useof potential type. evapotranspiration was related to the increase in temperature and the changes in land use type.

Figure 7. Cont. Sustainability 2018, 10, 3689 10 of 17 Sustainability 2018, 10, x FOR PEER REVIEW 10 of 17 Sustainability 2018, 10, x FOR PEER REVIEW 10 of 17

Figure 7. The changes of WY (water yield) on different time scales. ( (aa––gg)) were were the the water yield of Figure 7. The changes of WY (water yield) on different time scales. (a–g) were the water yield of subsub-watersheds‐watersheds from 1985 to 2014, and (h) the amount of water yield in Mentougou. sub‐watersheds from 1985 to 2014, and (h) the amount of water yield in Mentougou. 3.3.2. Changes Changes in Carbon Stocks 3.3.2. Changes in Carbon Stocks The changes ofof CSCS in in the the Mentougou Mentougou are are illustrated illustrated in in Figure Figure8. During 8. During the studythe study period, period, CS ﬁrst CS The changes of CS in the Mentougou are illustrated in Figure 8. During the study period, CS firstincreased increased and thenand decreased,then decreased, ﬁnally finally increasing increasing slightly slightly after 2010. after From 2010. 1985 From to 1995, 1985 CSto increased1995, CS first increased4 and then decreased, finally increasing slightly after 2010. From 1985 to 1995, CS increasedby 24.61 × by10 24.61t (1.44%). × 104 This t (1.44%). showed This that showed the ecological that the service ecological function service of CS infunction Mentougou of CS was in increased by 24.61 × 104 t (1.44%). This showed that the ecological service function of CS in Mentougouincreased, which was increased, was mainly which related was to mainly the decrease related ofto croplandthe decrease and of wasteland cropland and wasteland the increase and of Mentougou was increased, which was mainly related to the decrease of4 cropland and wasteland4 and thescrubland. increase In of 1995–2000, scrubland. 2000–2005 In 1995–2000, and 2005–2010, 2000–2005 CS and decreased 2005–2010, by 13.85 CS ×decreased10 t (0.80%), by 13.85 5.09 × 10104 tt the increase of 4scrubland. In 1995–2000, 2000–2005 and 2005–2010, CS decreased by 13.85  104 t (0.80%),(0.30%) and5.09 2.43 104× t 10(0.30%)t (0.14%), and 2.43 respectively.  104 t (0.14%), This isrespectively. mainly because This is the mainly area of because forested the land area has of (0.80%), 5.09  104 t (0.30%) and 2.43  104 t (0.14%), respectively. This is mainly because the area of forestedbeen decreased. land has Although been decreased. the area Although of scrubland the area and orchardof scrubland had increased, and orchard but had the increased, forested land but hadthe forested land has been decreased. Although the area of scrubland and orchard had increased, but the forestedthe greatest land CS had compared the greatest with CS other compared land use with types, other and land its use decrease types, had and led its todecrease the decline had led in CS. to forested land had the greatest CS compared with other4 land use types, and its decrease had led to theFrom decline 2010 toin 2014, CS. From the CS 2010 increased to 2014, slightly the CS by increased 4.53 × 10 slightlyt (0.27%). by This4.53  was 104 closely t (0.27%). related This to was the the decline in CS. From 2010 to 2014, the CS increased slightly by 4.53  104 t (0.27%). This was closelyincrease related in forested to the land increase and thein forested decrease land of wasteland and the decrease in this period. of wasteland in this period. closely related to the increase in forested land and the decrease of wasteland in this period.

Figure 8. The changes of SC (soil stocks) on different time scales. (a–g) were the soil stocks of Figure 8. The changes of SC (soil stocks) on different time scales. (a–g) were the soil stocks of sub-watershedsFigure 8. from The 1985 changes to 2014, of andSC (soil (h) the stocks) amount on ofdifferent soil stocks time in Mentougou.scales. (a–g) were the soil stocks of sub‐watersheds from 1985 to 2014, and (h) the amount of soil stocks in Mentougou. sub‐watersheds from 1985 to 2014, and (h) the amount of soil stocks in Mentougou.

Sustainability 2018, 10, 3689 11 of 17

Sustainability 2018, 10, x FOR PEER REVIEW 11 of 17 3.3.3. Changes in Soil Retention Rate 3.3.3. Changes in Soil Retention Rate The spatial distribution of SR was relatively stable, with a higher value usually in the north and The spatial distribution of SR was relatively stable, with a higher value usually in the north and south of the mountainous areas (Figure9). The SR in the seven periods was 97.35%, 97.83%, 96.00%, south of the mountainous areas (Figure 9). The SR in the seven periods was 97.35%, 97.83%, 96.00%, 98.40%, 97.93%, 98.72% and 98.70%, respectively. The lowest value for the whole study period was 98.40%, 97.93%, 98.72% and 98.70%, respectively. The lowest value for the whole study period was in in 1995. The SR was relatively stable after 2000. The highest value was in 2010. Compared with 1995. The SR was relatively stable after 2000. The highest value was in 2010. Compared with 1995, 1995, the SR increased by 2.93% in 2000 and 2.72% in 2014. Generally speaking, the soil conservation the SR increased by 2.93% in 2000 and 2.72% in 2014. Generally speaking, the soil conservation function has improved in recent years. Soil conservation increased by 1.35% from 1985 to 2014. function has improved in recent years. Soil conservation increased by 1.35% from 1985 to 2014.

Figure 9. The changes in SR (soil retention rate) on different time scales. (a–g) were the soil retention Figure 9. The changes in SR (soil retention rate) on different time scales. (a–g) were the soil retention rate of the sub-watersheds from 1985 to 2014, and (h) the amount of soil retention rate in Mentougou. rate of the sub‐watersheds from 1985 to 2014, and (h) the amount of soil retention rate in Mentougou. 3.4. The Relationship of Ecosystem Services and Landscape Patterns 3.4. The Relationship of Ecosystem Services and Landscape Patterns This paper drafted a three-dimensional coordinate system, in which the XYZ axis respectively representedThis paper the drafted changes a ofthree carbon‐dimensional stocks, soil coordinate retention system, rate and in which water the yield XYZ from axis 1985 respectively to 2014, asrepresented shown in Figurethe changes 10. The of coordinate carbon stocks, points soil of theretention levels inrate ecosystem and water services yield (ESs)from were 1985 distributedto 2014, as onshown the diagonal in Figure body 10. The diagonal coordinate of the three-dimensionalpoints of the levels coordinate in ecosystem system. services In the (ESs) past were 30 years, distributed the ESs hadon the changed diagonal considerably. body diagonal In general, of the three the‐ SRdimensional after 2000 coordinate was higher system. than that In the before past 2000, 30 years, but thethe amountESs had ofchanged WY and considerably. CS showed theIn general, opposite the result. SR after 2000 was higher than that before 2000, but the amountWe usedof WY the and landscape CS showed indices the opposite and ESs result. of 39 sub-watersheds in seven periods (1985–2014) to carryWe out used the RDA the landscape analysis. Theindices overall and test ESs of of the 39 canonicalsub‐watersheds relationship in seven was periods significant (1985–2014) (p < 0.01). to Thecarry cumulative out the RDA percentage analysis. varianceThe overall of test ESs of explained the canonical on the relationship first axis was of the significant RDA was (p 10.20%< 0.01). (eigenvalues:The cumulative 0.102), percentage the second variance axis of theESs RDAexplained was 1.30%on the (eigenvalues: first axis of the 0.013). RDA As was shown 10.20% in Figure(eigenvalues: 11, the landscape0.102), the pattern second factors axis of related the RDA to the was height 1.30% of (eigenvalues: the first ordination 0.013). axis As shown were IJI in and Figure LPI, while11, the MPS, landscape SHDI and pattern AI were factors negatively related correlated. to the height In the of correlationthe first ordination analysis, thereaxis were was aIJI significant and LPI, negativewhile MPS, correlation SHDI betweenand AI MPSwere andnegatively CS (F = −correlated.0.311, p < 0.001),In the and correlation a significant analysis, positive there correlation was a betweensignificant IJI negative and CS correlation (F = 0.280, betweenp < 0.001), MPS SHDI and CS and (F SR = −0.311, were significantlyp < 0.001), and negatively a significant correlated positive (Fcorrelation = −0.124, betweenp < 0.05). IJI In theand ESs CS of (F the = 0.280, sub-watersheds, p < 0.001), CS SHDI was and a relatively SR were obvious significantly indicator. negatively MPS, IJI andcorrelated SHDI were(F = − the0.124, landscape p < 0.05). pattern In the factors ESs that of the had sub great‐watersheds, impact on ESs.CS was a relatively obvious indicator. MPS, IJI and SHDI were the landscape pattern factors that had great impact on ESs. Sustainability 2018, 10, 3689 12 of 17 Sustainability 2018, 10, x FOR PEER REVIEW 12 of 17 Sustainability 2018, 10, x FOR PEER REVIEW 12 of 17

FigureFigure 10. 10. TheThe ecosystem servicesservices (ESs)(ESs) in in Mentougou Mentougou from from 1985 1985 to to 2014. 2014. Each Each point point represents represents a time. a Figure 10. The ecosystem services (ESs) in Mentougou from 1985 to 2014. Each point represents a time.In the In coordinate the coordinate system, system, the X the-axis X‐ isaxis the is amount the amount of carbon of carbon stocks, stocks, the Y-axis the Y is‐axis the is amount the amount of water of time. In the coordinate system, the X‐axis is the amount of carbon stocks, the Y‐axis is the amount of wateryield, yield, and the andZ -axisthe Z is‐axis the is amount the amount of soil of retention soil retention rate. rate. water yield, and the Z‐axis is the amount of soil retention rate.

Figure 11. Biplot diagram of the RDA between the ESs (WY, CS, SR) and landscape patterns (LPI, MPS, Figure 11. Biplot diagram of the RDA between the ESs (WY, CS, SR) and landscape patterns (LPI, FigureIJI, SHDI, 11. AI).Biplot diagram of the RDA between the ESs (WY, CS, SR) and landscape patterns (LPI, MPS, IJI, SHDI, AI). 4. DiscussionMPS, IJI, SHDI, AI). 4. Discussion 4.4.1. Discussion Impact of Land Use Type and Weather Changes on Ecosystem Services 4.1. Impact of Land Use Type and Weather Changes on Ecosystem Services 4.1. ImpactOver theof Land past Use 30 Type years, and 23.65% Weather of theChanges land on use Ecosystem types had Services changed (Table1). Of these, the area of scrubland,Over the past wasteland 30 years, and 23.65% forested of the land land were use the types main had land changed use types (Table that 1). affectedOf these, ESs. the area A large of Over the past 30 years, 23.65% of the land use types had changed (Table 1). Of these, the area of scrubland,amount of wasteland wasteland convertedand forested into land scrubland, were the contributing main land to use CS. types At the that same affected time, a ESs. large A amount large scrubland, wasteland and forested land were the main land use types that affected ESs. A large amountof forested of wasteland land changed. converted Forested into land,scrubland, as the contributing most effective to landCS. At use the type same in termstime, a of large carbon amount stock amount of wasteland converted into scrubland, contributing to CS. At the same time, a large amount ofcapacity, forested was land one changed. of the reasons Forested for land, the decrease as the most of CS effective in the studyland use area type after in 1995. terms of carbon stock of forested land changed. Forested land, as the most effective land use type in terms of carbon stock capacity,In general, was one the of landthe reasons use types for werethe decrease covered of from CS lowin the vegetation study area to after high 1995. vegetation, resulting in capacity, was one of the reasons for the decrease of CS in the study area after 1995. increasedIn general, evapotranspiration the land use types and were reduced covered WY from [42,43 low]. This vegetation result was to high also vegetation, veriﬁed in ourresulting research. in In general, the land use types were covered from low vegetation to high vegetation, resulting in increased evapotranspiration and reduced WY [42,43]. This result was also verified in our research. increased evapotranspiration and reduced WY [42,43]. This result was also verified in our research. Sustainability 2018, 10, 3689 13 of 17

Except for the special weather in Beijing in 2010, there was an abnormally high value of WY. Taking 2000 as a turning point, it presented a trend of ﬁrst increase and then decrease. This was probably due to the Grain for Green program in the study area in 2000. The increase of scrubland, forested land and orchard was probably due to the increases in evapotranspiration, resulting in low WY. The SR in the study area showed a gradual increasing trend. This was probably due to the ecological construction and vegetation coverage increaseing in the study area. The increased grassland, scrubland and forested land could increase SR. However, at the same time, this would also increase the evapotranspiration. Moreover, although the scrubland had increased, the forested land was generally decreased. Forested land as the largest carbon reserve led to the reduction of CS. The trade-off of ESs is a question worth pondering. In addition, climate change has a great impact on ESs [36,44]. Abnormal temperature has signiﬁcant effects on ESs. In the study of the effects of rainfall and temperature on crops, forests and biological activities, researchers have found that warm weather would reduces grain yield, and increases the frequency of forest pests, and the early pollination of bees [45–47].

4.2. Mechanisms Underlying the Impact of Landscape Pattern Changes on Ecosystem Services The evolution of landscape patterns is closely related to the change of ESs [14]. However, it is often neglected in many studies of land use types and weather change [36,43,48]. In the study, three indices of landscape indices (MPS, IJI and SHDI) were significantly correlated with CS and SR (Figure 11). This indicated that the landscape patterns were closely related to ESs. The same conclusion was also confirmed for farming and fertilization, which change the physical and chemical properties of soil, affecting the storage of organic matter and increasing the occurrence of non-point source pollution. Ultimately, this affected the value of the ESs [47,49]. Landscape patterns changes could also be related to the alteration of hydrological cycles, microclimates in the watershed, and soil nutrients [50,51]. Landscape patterns and ESs were significantly correlated [34,52,53]. The positive and negative relationships between these landscape indices and different types of ESs were different in different places. For example, the relationship between landscape patterns and ESs in the South Bay area of Hangzhou Bay indicated that increasing SHDI was conducive to improving the overall ESs [54]. There was a negative correlation between SHDI and ESs in the typical Karst area of Northwest Guangxi. In this study, SHDI was negatively correlated with the soil conservation function. The main reason was that the decrease of SHDI was mainly due to the reduction of wasteland in the study period [55]. The wasteland was mainly transformed into scrubland and forested land with stronger soil conservation capacity, thus SR was improved. In this study, MPS was negatively correlated with CS, which was different from most studies. The main reason for this was a large amount of broken small block scrubland around forested areas. This may result in smaller MPS in scrubland and forested land. It could be seen that the landscape patterns were closely related to the change of ESs and the positive or negative correlation was made according to local conditions.

4.3. Strategies for the Sustainable Use of Ecosystem Services Rational land use planning should take account of the value and sustainable development of multiple ESs and human well-being. This is still a huge challenge and could not really be put into action in many places. In the past 30 years, the built-up land in the study area had increased by 32.40%. We found that 73.38% of these increased built-up lands came from cropland. Moreover, these increased built-up lands were mainly concentrated in the few plain areas in the study area. Generally speaking, people usually choose fertile land for farming. Therefore, the productivity of land had been largely draining away [26,56]. Many ecological control projects have been implemented in the study area, but the overall tendency for forested land was still in decline [57]. Matthew et al. emphasized the impact of fragmentation in landscape patterns on ESs, suggesting that the fragmentation of landscape patterns would fragment biological pathways [52]. For example, the gene exchange of the population is limited by the population of the habitat patches [58,59]. Sustainability 2018, 10, 3689 14 of 17

The connectivity and diversity of landscape would also have a significant impact on ESs [34,53]. Our conclusion also proved that the landscape pattern was the main factor affecting ESs. When people consume one or several ESs, they will intentionally or unintentionally influence the provision of other ESs [60]. This changes the trade-off phenomenon of ESs were changed [61]. It is very important to balance the relationship of this tradeoff. In this study, the increase of scrubland and forested land would have a positive impact on CS. However, it also increased evapotranspiration, resulting in reduced WY. The same trade-off is also reflected in human activities, such as agricultural reclamation and deforestation. These activities, on the one hand, increase services such as food supply and timber supply. On the other hand, it also led to a decline in carbon fixation, soil and water conservation [12]. This contradictory relationship makes the trade-off of ESs increasingly prominent.

5. Conclusions Our results showed that 23.65% of the land use types in Mentougou district, a mountainous area in Beijing, changed from 1985 to 2014, with a largest fraction of change being for forested land and scrubland. Overall, the water yield decreased roughly 47.32% after year 2000. Carbon storage generally decreased. The soil retention rate increased by 1.38%. This was probably due to the increase in temperature and evapotranspiration, wasteland reduction and the increase of vegetation coverage. The landscape patterns were dispersed and fragmented. MPS and CS, IJI and CS were significantly correlated (p < 0.001), and SHDI and SR were significantly correlated (p < 0.05). Two important measures should receive attention to maintain the sustainable development of ESs in mountainous areas, (1) rational management of land use types. The loss of basic cropland and the destruction of forested land should be avoided. The influence of evapotranspiration on WY should also be considered during afforestation, and (2) The rationality of landscape patterns should be considered. Landscape patterns affect ESs, but it is often neglected in current models and research. Some landscape pattern indices were significant correlated with ESs.

Author Contributions: Y.Y. conceived and designed the study with the support of C.L. and M.S. All co-authors drafted and revised the article together. All authors read and approved the final manuscript. Funding: This research was funded by [name of funder] grant number [the National Key R&D Program of China] And the APC was funded by [2017YFC0505501]. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Costanza, R.; D’Arge, R.; de Groot, R.; Farber, S.; Grasso, M.; Hannon, B.; Limburg, K.; Naeem, S.; O’Neill, R.V.; Paruelo, J.; et al. The value of the world’s ecosystem services and natural capital. Nature 1997, 387, 253–260. [CrossRef] 2. Millennium Ecosystem Assessment (MA). Ecosystems and Human Well-Being: A Framework for Assessment; Island Press: Washington, DC, USA, 2003. 3. Millennium Ecosystem Assessment (MA). Ecosystems and Human Well-Being: Synthesis; Island Press: Washington, DC, USA, 2005. 4. Daily, G.C. Nature’s Services; Island Press: Washington, DC, USA, 1997. 5. Antle, J.M.; Stoorvogel, J.J. Predicting the Supply of Ecosystem Services from Agriculture. Am. J. Agric. Econ. 2006, 88, 1174–1180. [CrossRef] 6. Syphard, A.D.; Franklin, J. Spatial aggregation effects on the simulation of landscape pattern and ecological processes in southern California plant communities. Ecol. Model. 2004, 180, 21–40. [CrossRef] 7. Zhang, H.; Yang, G.; Huang, Q.; Li, G.; Chen, B.; Xin, X. Analysis on Dynamic Characteristics of Landscape Patterns in Hailer and around Areas BT—Computer and Computing Technologies in Agriculture IV; Li, D., Liu, Y., Chen, Y., Eds.; Springer: Berlin/Heidelberg, Germany, 2011; pp. 250–260. 8. Gu, D.; Zhang, Y.; Fu, J.; Zhang, X. The landscape pattern characteristics of coastal wetlands in Jiaozhou Bay Under the Impact of Human Activities. Environ. Monit. Assess. 2007, 124, 361–370. [CrossRef][PubMed] Sustainability 2018, 10, 3689 15 of 17

9. Peng, J.; Xie, P.; Liu, Y.; Ma, J. Urban thermal environment dynamics and associated landscape pattern factors: A case study in the Beijing metropolitan region. Remote Sens. Environ. 2016, 173, 145–155. [CrossRef] 10. Gage, E.A.; Cooper, D.J. Relationships between landscape pattern metrics, vertical structure and surface urban Heat Island formation in a Colorado suburb. Urban Ecosyst. 2017, 20, 1229–1238. [CrossRef] 11. Bateman, I.J.; Harwood, A.R.; Mace, G.M.; Watson, R.T.; Abson, D.J.; Andrews, B.; Binner, A.; Crowe, A.; Day, B.H.; Dugdale, S.; et al. Bringing Ecosystem Services into Economic Decision-Making: Land Use in the United Kingdom. Science 2013, 341, 45–50. [CrossRef][PubMed] 12. Pickard, B.R.; Van Berkel, D.; Petrasova, A.; Meentemeyer, R.K. Forecasts of urbanization scenarios reveal trade-offs between landscape change and ecosystem services. Landsc. Ecol. 2017, 32, 617–634. [CrossRef] 13. Wu, J.; Zhao, Y.; Yu, C.; Luo, L.; Pan, Y. Land management influences trade-offs and the total supply of ecosystem services in alpine grassland in Tibet, China. J. Environ. Manag. 2017, 193, 70–78. [CrossRef] [PubMed] 14. Wu, J. Landscape sustainability science: Ecosystem services and human well-being in changing landscapes. Landsc. Ecol. 2013, 28, 999–1023. [CrossRef] 15. Castro, A.J.; Martín-López, B.; López, E.; Plieninger, T.; Alcaraz-Segura, D.; Vaughn, C.C.; Cabello, J. Do protected areas networks ensure the supply of ecosystem services? Spatial patterns of two nature reserve systems in semi-arid Spain. Appl. Geogr. 2015, 60, 1–9. [CrossRef] 16. Deng, W.; Fang, Y.P.; Tang, W. The strategic influence and development orientation of mountain urbanization in China. Bull. Chin. Acad. Sci. 2013, 28, 66–73. (In Chinese) 17. Sewall, J.O.; Riihimaki, C.A.; Kadegis, J. Orbital control, climate seasonality, and landscape evolution in the Quaternary Rocky Mountains. Geomorphology 2015, 250, 89–94. [CrossRef] 18. United Nations. Report of the World Summit on Sustainable Development. Environ. Politics 2002, 45, 82–221. 19. Zhou, S.X. Speeding up the construction of mountain areas is an inevitable choice to achieve sustainable development. Speech at the international commemorative meeting of mountain year. For. China 2003, 1, 10–11. (In Chinese) 20. Briner, S.; Elkin, C.; Huber, R. Evaluating the relative impact of climate and economic changes on forest and agricultural ecosystem services in mountain regions. J. Environ. Manag. 2013, 129, 414–422. [CrossRef] [PubMed] 21. Fürst, C.; Frank, S.; Witt, A.; Koschke, L.; Makeschin, F. Assessment of the effects of forest land use strategies on the provision of ecosystem services at regional scale. J. Environ. Manag. 2013, 127, S96–S116. [CrossRef] [PubMed] 22. Zeleke, G.; Hurni, H. Implications of Land Use and Land Cover Dynamics for Mountain Resource Degradation in the Northwestern Ethiopian Highlands. Mt. Res. Dev. 2001, 21, 184–191. [CrossRef] 23. Wu, J.; Goldberg, S.D.; Mortimer, P.E.; Xu, J. Soil respiration under three different land use types in a tropical mountain region of China. J. Mt. Sci. 2016, 13, 416–423. [CrossRef] 24. Bai, Y.; Wang, R.; Jin, J. Water eco-service assessment and compensation in a coal mining region—A case study in the Mentougou District in Beijing. Ecol. Complex. 2011, 8, 144–152. [CrossRef] 25. Du, J.; Thill, J.C.; Peiser, R.B.; Feng, C. Urban land market and land-use changes in post-reform China: A case study of Beijing. Landsc. Urban Plan. 2014, 124, 118–128. [CrossRef] 26. Du, J.; Thill, J.C.; Peiser, R.B. Land pricing and its impact on land use efficiency in post-land-reform China: A case study of Beijing. Cities 2016, 50, 68–74. [CrossRef] 27. Yi, Y.; Zhao, Y.; Ding, G.; Gao, G.; Shi, M.; Cao, Y. Effects of Urbanization on Landscape Patterns in a Mountainous Area: A Case Study in the Mentougou District, Beijing, China. Sustainability 2016, 8, 1190. [CrossRef] 28. Zhu, T.F. Land Use/Cover Change and Its Water Resources Effect in North China: A Case Study of Mentougou District, Beijing City; China Agricultural University: Beijing, China, 2014. (In Chinese) 29. Beijing Municipal Bureau of Statistics. Beijing Statistical Yearbook; China Statistic Press: Beijing, China, 1985, 1990, 1995, 2000, 2005, 2010, 2014. (In Chinese) 30. Huang, C.H. Research on Ecosystem Service Function Based on InVEST Model; Beijing Forestry University: Beijing, China, 2014. (In Chinese) 31. United States Geological Survey. Remote Sensing Images. Available online: https://www.usgs.gov/ (accessed on 17 April 2018). Sustainability 2018, 10, 3689 16 of 17

32. China’s National Standard (GB/T 21010-2017). Current Land Use Condition Classiﬁcation; China Zhijiang Publishing House: Beijing, China, 2017. 33. Chinese Academy of Sciences. Geo Spatial Data Cloud. Available online: http://www.gscloud.cn/ (accessed on 15 August 2018). 34. Xiao, R.; Wang, G.; Zhang, Q.; Zhang, Z. Multi-scale analysis of relationship between landscape pattern and urban river water quality in different seasons. Sci. Rep. 2016, 6, 25250. [CrossRef][PubMed] 35. InVEST 2.2.4 User’s Guide. Available online: http://data.naturalcapitalproject.org/nightly-build/invest- users-guide/html/ (accessed on 15 August 2018). 36. Fu, Q.; Li, B.; Hou, Y.; Bi, X.; Zhang, X. Effects of land use and climate change on ecosystem services in Central Asia’s arid regions: A case study in Altay Prefecture, China. Sci. Total Environ. 2017, 607–608, 633–646. [CrossRef][PubMed] 37. Zhang, L.; Dawes, W.R.; Walker, G.R. Response of mean annual evapotranspiration to vegetation changes at catchment scale. Water Resour. Res. 2001, 37, 701–708. [CrossRef] 38. Wischmeier, W.H.; Smith, D.D. Predicting Rainfall Erosion Losses—A Guide to Conservation Planning; Department of Agriculture; Island Press: Washington, DC, USA, 1978. 39. Graﬁus, D.R.; Corstanje, R.; Warren, P.H.; Evans, K.L.; Hancock, S.; Harris, J.A. The impact of land use/land cover scale on modelling urban ecosystem services. Landsc. Ecol. 2016, 31, 1509–1522. [CrossRef] 40. Zhang, C.; Ju, W.; Chen, J.; Zan, M.; Li, D.; Zhou, Y.; Wang, X. China’s forest biomass carbon sink based on seven inventories from 1973 to 2008. Clim. Chang. 2013, 118, 933–948. [CrossRef] 41. Edmondson, J.L.; Davies, Z.G.; McCormack, S.A.; Gaston, K.J.; Leake, J.R. Land-cover effects on soil organic carbon stocks in a European city. Sci. Total Environ. 2014, 472, 444–453. [CrossRef][PubMed] 42. Wang, Z.; Mao, D.; Li, L.; Jia, M.; Dong, Z.; Miao, Z.; Ren, C.; Song, C. Quantifying changes in multiple ecosystem services during 1992–2012 in the Sanjiang Plain of China. Sci. Total Environ. 2015, 514, 119–130. [CrossRef][PubMed] 43. Gao, J.; Li, F.; Gao, H.; Zhou, C.; Zhang, X. The impact of land-use changes on water-related ecosystem services: A study of the Guishui River Basin, Beijing, China. J. Clean. Prod. 2017, 163, S148–S155. [CrossRef] 44. Grimm, N.B.; Faeth, S.H.; Golubiewski, N.E.; Redman, C.L.; Wu, J.; Bai, X.; Briggs, J.M. Global Change and the Ecology of Cities. Science 2008, 319, 756–760. [CrossRef][PubMed] 45. Lobell, D.B.; Schlenker, W.; Costa-Roberts, J. Climate Trends and Global Crop Production Since 1980. Science 2011, 333, 616–620. [CrossRef][PubMed] 46. Trumbore, S.; Brando, P.; Hartmann, H. Forest health and global change. Science 2015, 349, 814–818. [CrossRef][PubMed] 47. Bárcena, J.F.; Gómez, A.G.; García, A.; Álvarez, C.; Juanes, J.A. Quantifying and mapping the vulnerability of estuaries to point-source pollution using a multi-metric assessment: The Estuarine Vulnerability Index (EVI). Ecol. Indic. 2017, 76, 159–169. [CrossRef] 48. Sun, X.; Li, F. Spatiotemporal assessment and trade-offs of multiple ecosystem services based on land use changes in Zengcheng, China. Sci. Total Environ. 2017, 609, 1569–1581. [CrossRef][PubMed] 49. Díaz, F.J.; O0Geen, A.T.; Dahlgren, R.A. Agricultural pollutant removal by constructed wetlands: Implications for water management and design. Agric. Water Manag. 2012, 104, 171–183. [CrossRef] 50. Tripathi, K.; Singh, B. Species diversity and vegetation structure across various strata in natural and plantation forests in Katerniaghat Wildlife Sanctuary, North India. Trop. Ecol. 2009, 50, 191–200. 51. Meng, L.Z.; Yang, X.D.; Martin, K.; Gan, J.M.; Liu, Y.H.; Gong, W.C. Movement patterns of selected insect groups between natural forest, open land and rubber plantation in a tropical landscape (southern Yunnan, SW China). J. Insect Conserv. 2016, 20, 363–371. [CrossRef] 52. Mônica, B.S.; Tamminga, K.R.; Tangari, V.R. A Method for Gauging Landscape Change as a Prelude to Urban Watershed Regeneration: The Case of the Carioca River, Rio de Janeiro. Sustainability 2012, 4, 2054–2098. 53. Hao, R.; Yu, D.; Liu, Y.; Liu, Y.; Qiao, J.; Wang, X.; Du, J. Impacts of changes in climate and landscape pattern on ecosystem services. Sci. Total Environ. 2017, 579, 718–728. [CrossRef][PubMed] 54. Xia, D. The Change and Driving Force of Wetland Ecosystem Service Value on the South Bank of Hangzhou Bay; Zhejiang University: Hangzhou, China, 2012. (In Chinese) 55. Fan, F.D.; Luo, J.; Wang, K.L.; Chen, H.S.; Zhang, W. The ecosystem service function of the Karst region of Northwest Guangxi and the spatial analysis. Chin. J. Ecol. 2011, 30, 804–809. (In Chinese) Sustainability 2018, 10, 3689 17 of 17

56. Zhou, D.; Zhao, S.; Zhu, C. The Grain for Green Project induced land cover change in the Loess Plateau: A case study with Ansai County, Shanxi Province, China. Ecol. Indic. 2012, 23, 88–94. [CrossRef] 57. Dai, E.; Wang, X.; Zhu, J.; Xi, W. Quantifying ecosystem service trade-offs for plantation forest management to beneﬁt provisioning and regulating services. Ecol. Evol. 2017, 7, 7807–7821. [CrossRef][PubMed] 58. Ceballos, G.; Ehrlich, P.R.; Barnosky, A.D.; Garcia, A.; Pringle, R.M.; Palmer, T.M. Accelerated modern human-induced species losses: Entering the sixth mass extinction. Sci. Adv. 2015, 1, e1400253. [CrossRef] [PubMed] 59. Qiu, L.; Tao, T.T.; Han, S.R.; Yang, W.Y.; Luan, X.L.; Qiu, Y.N.; Liu, M.S.; Xu, C. Effects of local landscape fragmentation on species richness at a macroecological scale. Acta Ecol. Sin. 2017, 37, 7595–7603. 60. Jayachandran, S.; de Laat, J.; Lambin, E.F.; Stanton, C.Y.; Audy, R.; Thomas, N.E. Cash for carbon: A randomized trial of payments for ecosystem services to reduce deforestation. Science 2017, 357, 267–273. [CrossRef][PubMed] 61. Niu, X.; Wang, B.; Liu, S.; Liu, C.; Wei, W.; Kauppi, P.E. Economical assessment of forest ecosystem services in China: Characteristics and implications. Ecol. Complex. 2012, 11, 1–11. [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/).