Land Cover Dynamics And Its Effect On Landscape Pattern And Ecosystem Service Values In Pingtan Island,China

Land cover is a very important cross-cutting environmental variable,which is at the center of different environmental concerns such as hydrology,climate,soil,biodiversity,natural disaster,and biogeochemical cycles.The availability and sustainability of ecosystem services,which are the basis of human survival,are also growing concern.Accordingly,land cover change?LCC?Information has been given much emphasizes by several national and international programs.Therefore,either spectral or class-based,monitoring of land cover is essential for agencies responsible for environmental policy formulation and its enforcement,as well as to examine the impacts of existing policies.In Pingtan Island,the accelerated socio-economic development together with the limited spatial space of the island has posed LCC and coastal sea reclamation.The socio-economic development has been hastened,especially since the setup of Pingtan comprehensive experimental zone,which is a general economic development plan with the aim of making Pingtan an international tourism island and free trade zone.To our knowledge,however,few efforts have been made for detailed quantification of the changes in land cover,its patch arrangements and driving factors in Pingtan island.Therefore,this study focused on comprehensive quantification of LCC,the spatial composition and configuration of land cover patches,LCC drivers and the effects of LCC on the ecosystem services in four periods;1984-1996(1^st),1996-2007(2^nd),2007-2017(3^rd)and?1984-2017??the entire period?.With the help of remote sensing and geographic information system?GIS?technologies,multispectral Landsat imageries of 1984?TM?,1996?TM?,2007?TM?and 2017?OLI?were acquired to produce four land cover maps with eight classes and to drive other spatiotemporal characteristics of LCC.Also,the benefit transfer method was adopted for the estimation of ecosystem service values?ESV?.The effects of socioeconomic factors on the change of major land cover types were examined using partial least square regression model.The results indicated that,84.38%?27175ha?of the total area was covered by farmland,shrub land,grassland and forest land during the beginning of the study?1984?,of which farmland alone accounted for 53.2%.Since then,the continuous shrinkage of farmland and expansion of urban land were the most noticeable.The remaining land cover categories did not have unidirectional trends in their changes.Regardless of the period-to-period dynamism,an overall positive net change was found in the area of forest,shrub land and wetland,whereas a net decrease in that of water bodies,grassland and bare land?1984-2017?.By the end of the study period?2017?,the major classes were in the order of farmland?31.66%?,urban land?23.46?,forest land?18.62?and shrub land?12.95%?,which altogether covered 86.69%of the island.Similar to other parts of China,urban sprawl is the main triggering factors for the change of other land covers,especially for the significant shrinkage of farmland.In all intertemporal periods,the intensity of urban land change was within the category of rapid expansion?0.4<CII<0.7?.Even though there has been a tendency of urban land densification as reflected by the general trends of different density indices?vertical development,GDP density and residential area density?,urban expansion mainly in the form of edge and leapfrog was dominant during the study period.The direction and extent of classes-to-class transitions were clearly understood in land cover transition matrices.In the 33 years period considered,10965ha of the island was stable,of which60.85%,12.64%and 9.4%were farmland,forest land and urban land respectively.The rest part of the island has experienced transformation.The dominant land cover transition was farmland conversion to urban land followed by forest land.Swap change was the main form of spatial dynamics among classes that accounts for 52.57%of the total change.Furthermore,the land cover transition was influenced by the size of each class by which the wider the class,the greater contribution to total change?in both gain and loss dimensions?.For example,85%to total change was contributed by some dominant classes that are farmland?32.93%?,urban land(16.76%,shrub land?16.36%?and forest land?13.01%?.Whereas,bare land with the smallest size played insignificant roles in the land cover transitions.The rate of increase in urban land and forest land were far greater than their rate of decrease,and hence these classes were the most expanded types with the net gain of 5926ha and 3682ha respectively.By contrast,farmland,grassland and water bodies decreased a lot with a net change of-6936ha,-1743ha and-1088ha respectively that correspond to the loss of their initial extent in 1984 by 40.48%,47.81%and 65.98%.These changes were also supported by the corresponding extent/condition of transformation?ET&CT?and the hotspots of change.Landscape pattern is the other important aspect of and focuses in land change studies that provide further insights into the spatiotemporal arrangements and compositions of patches.Based on landscape level metrics,the study area has experienced great changes in its landscape.Wavy trends were observed in their intertemporal changes.Considering the entire period of the study?1984-20147?,the landscape has become spatially disintegrated as evidenced with an increase in the number of patches?NP?from 2257 in1984 to 32812 in 2017.Along with patches fragmentation and prevalence of smaller patches,total edge also increased from 107035 meters?m?to 134311m in the respective years.The edge density was 333.482 meters/hectare?m/ha?in 1984,which also increased towards 2017 at an annual growth rate of 2.575 m/ha.An upward trend in Shannon's Diversity Index also signifies the continued increase in landscape diversity or patch richness.Due to the increased level of landscape fragmentation,the mean patch size reduced from 1.422 ha?1984?to 0.978ha?2017?.The variability of size among patches and largest patch index decreased continuously.The shape of patches has also become more simple?regular?towards 2017 as inferred from the downward trends in the area-weighted mean patch fractal dimension?AWMPFD?and area weighted mean shape?AWMSI?.As for class level landscape metrics,big variations were found among land cover classes and from period-to-period within each class.However,general nature of the landscape in the last33years demonstrated that there was an increase in the fragmentation level of major land cover classes that include farmland,urban land,forest land,shrub land and grassland.Among these classes,the relative variability in patches size about the mean increased in urban land and shrub land,but it declined in most of the remaining classes.Moreover,the shape complexity of patches increased in urban land,while it reduced and become more regular in grassland,shrub land and farmland.Land cover change?LCC?has a direct effect on ecosystem service values?ESV?.The information regarding the trend of ESV in relation to LCC has been increasingly reported as valuable to reflect the human-nature interactions and for alternative environmental management.In this study,the total ESV increased and decreased alternatively from the first to the last period.Except for forest land,the negative trends in ESV of other categories were more frequent over different periods of the study.Regulating services and cultural services have the highest and the least contributions respectively for the total ESV in all years considered.On the overall?1984-2017?,there was a loss in value of food production services,waste treatment and hydrology regulation,which cumulatively was less than that of the gain from the rest of sub-ecosystem services.Consequently,the net change in total ESV monetarily was 0.11million USD,which was an increase by 0.39%of the total ESV in the initial year?1984?.This small net gain was almost entirely due to the expansion of forest land that has higher per unit value coefficients of ecosystem services.There are different drivers that contributed for LCC,and hence its effect on landscape pattern and ecosystem services.The effects of physical factors on LCC were less important.The largest part of Pingtan was plain so that the influence of topography was minimal as compared to anthropogenic factors.On the whole,the areas under vegetation cover such as forest,shrub and grassland increased with elevation and slope,which may be due to a limited human influence on hilly and mountainous areas.By contrast,man-modified landscapes such as farmland and urban land distributed primary on the lowland plain areas.The presence of strong wind with frequent typhoons was also the other physical factor that positively affected the preservation of forest and shrub land in that plantation of trees as a shelterbelt has been the common practices,especially around the coastal areas as a shelterbelt.The chief drivers of LCC in Pingtan were socioeconomic development and policy factors.Some of the socioeconomic factors include population growth,total GDP,GDP from primary,secondary,construction and tertiary industry,per capita GDP,fixed asset investment,real estate development and others.Based on PLSR models,the effects of these factors on LCC varies from some to significant levels.Particularly,urban and forest land were affected positively while both grassland and farmland were affected negatively.Based on different parameters of the model quality,PLSR modes generated in this study can be used for simulation of LCC with high precisions.Policies of ecological protection,economic development,farmland protection and land use were also important drivers of LCC.For instance,the grain for green project as a national strategy contributes to the protection of existing forest land and afforestation of ecologically fragile areas.Weak enforcement of farmland protection policy was one of the causes of continuous shrinkage of farmland.Since 1978,the market-oriented economic reform of China also accelerated socio-economic development,and urban sprawl mainly on farmland and grassland.The notable effect of this economic reform on LCC was in the coastal areas where Pingtan Island is the one.Pingtan comprehensive experimental zone is another economic development plan since 2009 with the aim of making Pingtan an international tourism island and free trade zone.This further attracts investors and accelerate real estate development,construction of port and industrial zones with far-reaching impacts on urban expansion at the expenses of natural and semi natural categories such as farmland,water bodies,grassland and bare land.Even though socioeconomic pressure is very high in coastal China,there are no specific coastal management policies and the stakeholders were not actively participated in the development and implementation of environmental policies.In general,this study comprehensively measured LCC,landscape Patten,the effects of LCC on ESV,and the drivers of LCC,which all are very helpful for planning land management rationally and as baseline database for future studies.