Division of Farmland Use in East China at County level Based on Landscape Structure and Output Function——a Case of Pingdu City,Shandong Province

Jinfeng WU, Xiaoyan JIANG, Xiuhong WANG, Yuesheng GUO, Min LI

1. Key Laboratory of Land Surface Pattern and Simulation, Institute of Geographic Sciences and Natural Resources Research,CAS, Beijing 100101, China;

2. University of Chinese Academy of Sciences, Beijing 100049, China;

3. Pingdu Soil and Fertilizer Station, Pingdu 266700, China;

4. Shandong Soil and Fertilizer Station, Jinan 250100, China;

5 Qingdao Soil and Fertilizer Station, Qingdao 266071, China

Currently, the area of farmland in China is decreasing because of non-agricultural construction and ecological restoration,while people’s demands on farmland and agricultural products are increasing in terms of quality and quantity,due to population growth and increase of domestic consumption[1-2]. There fore, it is a crucial phase for China to reinforce eco-environment protection as economy grows rapidly. Farmlands are fundamental resources for human beings’ existence and development,as well as the source of food production[3-4].Besides ensuring national food security, farmland protection is a basic strategy of guaranteeing ecological safety and sustainable development[5-6].At present,it is urgent to carry out farmland protection projects,especially for farmlands in the east of China.

Confronted by limited land resources, land use division, providing rational trends of land use,is generally formulated on basis of history, status characters, changes and evolvement of regional lands[7-9].The research proposed landscape structure and output function-based farmland use division method at county level,with consideration of large-scale farmland divisions in China, including cropping regionalization based on different natural conditions and cropping systems[10-11], andagricultural natural resources and agricultural divisions based on land use types and farmland use trends[12-13].Domestically, for either land use division or farmland use division,most researches available are conducted as per qualitative and semi-quantitative analyses. In contrast, less attention is paid to quantitative research,research on structure or function-oriented land division by establishment of quantitative quotes in particular. The research selected land use factors including farmland structure, output, geographical location,et al.,and then performed a cluster analysis for farmland use division. From the perspective of application, the research method provides scientific references for scientific protection and rational use of county-level farmlands.

Research Objects and Methods

Introduction of research area

Pingdu City is situated in the west of Jiaodong peninsula at 119°31′30″-120°19′13″E,36°28′15″-37°02′46″N.It is the largest county-level city in Shandong Province, with an area of 3.27×105hm2and a population of 1 381 000 involving farming population of 1 202 000[14]. Daze Mountain is located in the north of Pingdu City, so the terrain of northern part is high, where the land use type is dominated by forests and grassland;while the terrain of southern part is low, where the land use type is dominated by farmland[15].The area of farmland, which is 1.78 ×105hm2，is large in Pingdu City，where the major crops include wheat, corn, peanut and vegetable. However, farmland reserve proves insufficient.For instance,undeveloped lands was only 1.2×104hm2,including intertidal zone, which just takes up 3.58%of the city area.What’s worse,the quality of most land soil is low, which has a high slope, and land slope is higher, so that the lands are not suitable for being developed into farmlands[16]. Pingdu has a temperate semi-humid climate, where the annual mean temperature is 11.9 ℃，the extremely maximal temperature is 38.6℃, and the extremely mini temperature is-17.9 ℃.The city is divided into 5 streets and 12 towns, of which Fengtai has the minimum area, with construction land over 2/3(Fig.1).Recently, with economy development,eco-environment is deteriorating, and farmlands are increasingly occupied by construction lands.Therefore,in order to get the maximized economic benefits, the farmland should be scientifically protected and rationally used.

Data

The land use of Pingdu City(1 ∶100 000) in 2013 was obtained based on interpretation of Landsat ETM, from USGS (http://glovis.usgs.gov/), which meets the demand of low cloud coverage (＜10%). Specifically,the images were corrected by ENVI4.8 to construct image interpretation symbol based on field survey and GIS in order to guarantee interpretations between human and machines. Subsequently, according to landscape functions, lands can be further divided into farmland, forest land, grassland, water,construction land and unused land by ArcGIS9.3. The sown area and yield per unit area of crop,peanut,and vegetable were from Pingdu Statistical Yearbook in 2014.

Research Method

It’s necessary to make comprehensive assessment and sort order on all factors related to farmland,when divided the use of farmland. And the cluster analysis is an important method,which can reflect relationships among multiple samples for researching classification and division of geographical phenomenon in a quantitative way[17-20]. Therefore,the research selected indices related to farmlands and carried out division of farmlands at county level as per hierarchical clustering method from the perspectives of landscape structure, farmland output and location of evaluated units.

Selection of clustering indices

Functional index of landscape structure The spatial pattern of regional farmland use is under influence of natural, social and economic factors. Fragstats is an ecological analysis software containing many landscape indexes. These indexes reflect their structural composition and feature of spatial allocation, highly enriched information of landscape structure. Hence, it is an approporate method of categorizing landscape patterns[21-24].The research selected class area (CA), percentage of landscape(PLAND), number of patches (NP),and Shannon’s diversity index(SHDI), and construct the percentage of a target farmland to the entire evaluation unit (%) (PLANDi),farmland area-to-construction land area ratio (CBRCA), the number of farmland patch-to-the number of construction land patch (CBRNP),farmland area-tothe area of other agricultural land(CORCA), the number of farmland patch/the number of patch of other agricultural land (CORNP), and Shannon’s diversity index (SHDI). Explanation was shown below:

Assuming i was given 1, 2 and 3,representing farmlands, construction lands and lands for other purposes,and j, refering to the number of patch,given from 1 to n.

where PLANDiis the areal percentage of a patch type i to the entire landscape(%,0＜PLAND＜=100). The closer the value to zero, the fewer the patch type.If the value is equal to 100,it suggests that the region is just composed of one patch.

where i represents land type (similarly hereinafter); CAiis an area of all patches of the ithtype(m2),followed by unitconversion to hectare.

NPi=ni, where NPirepresents the total number of the ithpatch;NP shows the spatial pattern of the landscape,whose value is of positive correlation with degree of fragmentation. The higher the NP, the higher the fragmentation degree.

where SHDI equals minus the sum,across all patch types, of the proportional abundance of each path type multiplied by that proportion (SHDI＞=0), which reflects landscape heterogeneity.In a landscape system,for example, the richer land use, the higher degree of fragmentation, and the higher SHDI.

Output function index The outputfunction shows agricultural production capacity and structure of a region,which provides references for regional farmland division. The research provided per unit yield indices of grain,peanut and vegetable on basis of farming and output of Pingdu City, reflecting output property in unit test regions,as follows:

where i is the type of crop;Ciis per unit area yield of crop i; Yiis total yield of crop i;Aiis the sowing area of crop i.

Functional index of spatial location With ArcGIS9.3[25], the geometric centers of assessment units were represented by x and y,respectively.

Cluster analysis Because original data are diversified in terms of units,dimensions and importance, it is necessary to normalize data and grade importance of indices before cluster analysis undergoing. Specifically, data normalization can be done conveniently in SPSS, and index importance was set by five-mark scoring. According to contribution, PLANDiwas 5, Ciwas 4,spatial location functional index was 3,SHDI was 2 and rest indices were graded 1. Hence, cluster index can be computed as the methods above, followed by a cluster analysis.The proximities among five streets and 12 towns were measured as per Euclidean distance.

Results and Analysis

Results of different areas

Considering the specialty cluster analysis results of Fengtai Street, due to a small area,low proportion of farmland to assessed units, extensive construction lands,cluster analysis results are special,so that it should be singled separately. However, in order to simplify the results, it would be classified into the category with similar property.

As shown in Fig.2, when between-class analysis was 18 ＜d ＜20,Baishahe Street,Nancun Town, Cuijiaji Town, Liaolan Town and Tonghe Town belonged to the same group;Guyan Town, Renzhao Town, Liyuan Street, Xinhe Town, Mingcun and Tianzhuang Town were the same group; Dazeshan Town, Dongge Street,Dianzi Town,Jiudian Town and Yunshan Town were the same group.But the Guxian Town and Renzhao Town in the middle east are discontinuous with Liyuan Town, Xinhe Town,Mingcun Town and Tianzhuang Town in the west in space, so that the two were grouped separately. Therefore,when the between-class distance was 15,the research regions can be classified into group 1 including Baishahe,Nancun, Cuijia, Liaolan and Tonghe,group 2 including Liyuan,Xinhe,Mingcun and Tianzhuang,group 3 including Guyan and Renzhao, and group 4 including Dazeshan,Dongge,Dianzi,Jiudian and Yunshan. Additionally,Fengtai Street was classified into group 1 considering status quo and development(Fig.3).

Division of farmland use and landscape structure and function

In accordance with the percentages of farmlands to division areas,the proportion of farmland area in group 3 reached the highest at 81.02%,followed by groups 1 and 2 at 72.37%and 79.85%,so that top priority should be given to areas of the three groups in planting at large scales. As for group 4, farmland area kept lower at 54.87%,but the SHDI value was the highest, indicating that the land use types are diversified.

Furthermore, CBRCAvalues of groups 1 and 2 were lower,suggestingthat the construction lands occupy more than farmlands there, so farmland protection should be reinforced;the lower CBRNPof groups 1 and 3 incorporated that the construction lands scattered, and the concentration degree and use rate should be improved at planing to control its expansion by occupying farmlands. As for the groups with the highest CORCAand CORNP, forestry, animal husbandry and aquaculture should be well protected for advancement to underscore land use and output diveristies, and improve agricultural spatial optimization and eco-servicing. In areas of group 4, it’s important to control the indiscriminate extension of farmland,maintain land use diversity, promote characteristic agriculture development and prevent water and soil losses. In general, the proportion of farmland area is in consistent with that of construction lands, suggesting that construction lands and farmlands are evenly distributed in the City. On the other hand, patch number of farmlands coincides with that of agricultural land, but th agricultural lands with small areas should be extended to some extents to improve farmland environment.

Table 1 Landscape structure and function indices of farmlands in different categories

Division and output function of farmlands

It is understandable that yields of grains,peanuts and vegetables are diversified because of varied farmland area, location and terrains, with variance coefficients of 75.43%, 68.31%and 107.63%, showed moderate variability of grain and peanut yields and high variability of vegetable. Specifically, yields of crops of group 1 maintained the highest; the yields of grains and peanuts of group 3 were the least;vegetable yield of group 2 was the least.Comparatively speaking,per unit area yield tended to have small differences.For example,the differences of per unit area yield of grain and peanut were just 7.68% and 8.52%. Still, the differences of per unit area yield of vegetable kept larger at 30.08%, with low variability.It can be concluded that the regions of groups 1 and 3 were domianted by vegetable growing, and groups 1 and 2 dominated by grain.Whta’s more, vegetable farming requires lots of fertilizers,organic fertilizers and pesticides, and it is important to control agricultural non-point pollutions(Fig.4).

Conclusions and Discussions

The research established the percentage of a target farmland relative to the entire evaluation unit(%)(PLANDi),farmland area-to-construction land area ratio (CBRCA), the number of farmland patch-to-the number of construction land patch (CBRNP), farmland area-to-the area of other agricultural land (CORCA),the number of farmland patch/the number of patch of other agricultural land (CORNP), and Shannon’s diversity index (SHDI), per unit area yield (Ci)and the center of geometry.After data normalization, cluster index was computed by a cluster analysis.The proximities among units were measured and farmlands in Pingdu were finally classified into 4 groups.

In accordance with landscape structures, it is obvious that construction lands are extensive in groups 1 and 2, where it is necessary to strengthen farmland protections,especially for highly-quality farmlands. In groups 1 and 3, construction lands scatter here and there,so that it is crucial to improve concentration degrees and use rate, and control farmland pollution possibly caused by constructions. In group 4, however, priority should be given to maintaining diversity of land use patterns, laying foundation for land use in other groups. In terms of farmland outputs, groups 1 and 3 are dominated by vegetable growing and groups 1 and 2 by grain.Therefore, it becomes important to intensity controlling of agricultural nonpoint pollution in these areas. As for group 4, attention should be paid to featured agriculture to avoid aimless expansion of farmland.

In conclusion, during urbanization process in eastern coastal agricultural regions, it is of importance to prevent fertile farmlands being occupied by construction lands and to enhance use rate of construction lands. Furthermore, agricultural non-point pollution should be detected constantly in intensively used farmlands,especially in vegetable fields. Additionally, land use diversity should be well protected in case of farmland expansion to improve farmland environment.

[1]LIU Y (刘玉),HAO XY (郝星耀),et al.Evaluation and zoning of cultivated ladn intensive use in Henan Province at county level (河南省耕地集约利用时空分异及分区研究)[J]. Scientia Geographica Sinica (地理科学), 2014(10):1218-1225.

[2]REN P(任平),WU T(吴涛),et al.Method of quantitative compensation for cultivated land conversion based on spatial characteristics of cultivated land protection value (基于耕地保护价值空间特征的非农化区域补偿方法)[J]. Transactions of the Chinese Society of Agricultural Engineering (农业工程学报),2014(20):277-287.

[3]LIU Y (刘玉), LIU YS (刘彦随), et al.Calculation of the integrated productive capacity and subarea utilization of cultivated land in alluvial plaiin area of Hasihe River(海河冲积平原区耕地综合产能核算及其分区利用)[J]. Resource Sciences (资源科学), 2009 (04): 598-603.

[4]YANG L(杨立),WANG BQ(王博祺),et al. A quantitative srudy on the performance of cultivated land protection in China since the reform and opening up based on the angel of protection (改革开放以来我国耕地保护绩效定量研究——基于数量保护的视角)[J].Journal of Agricultural Mechanization Research(农机化研究),2015(03):1-6.

[5]LIU YS(刘彦随),QIAO LY(乔陆印).Innovating system and policy of arable land conservation under the new-type urbanization in China (中国新型城镇化背景下耕地保护制度与政策创新)[J].Economic Geography(经济地理),2014(04):1-6.

[6]YANG L(杨立),WANG BQ(王博祺),et al. A quantitative study on the performance of cultivated land protection in China since the reform and opening up based on angel of protection(改革开放以来我国耕地保护绩效定量研究——基于数量保护的视角)[J].Journal of Agricultural Mechanization Research (农机化研究),2015:002.

[7]YANG ZS (杨子生),HAO XZ (郝性中).An approach on the several problems of land utilization regionalization(土地利用区划几个问题的探讨)[J]. Journal of Yunnan University (Natural Sciences Edition) (云南大学学报(自然科学版)),1995(04):363-368.

[8]ZHANG JX (张洁瑕),CHEN YQ (陈佑启), et al. Study on land use regionalization based on land-use function:case study of Jilin Province (基于土地利用功能的土地利用分区研究——以吉林省为例)[J].Journal of China Agricultural University (中国农业大学学报),2008(03):29-35.

[9]ZHAO RQ (赵荣钦), HUANG XJ (黄贤金), et al. Application of clustering analysis to land use zoning of coastal region in Jiangsu province(聚类分析在江苏沿海地区土地利用分区中的应用)[J].Transactions of the Chinese Society of Agricultural Engineering (农业工程学报),2010(06):310-314.

[10]The 6thdrafting group, the 5thChina farming regionalization, Chinese Academy of Agricultural Sciences(中国农业科学院5 中国种植业区划6 编写组).China farming regionalization (中国种植业区划). Beijing: China Agriculture Press(北京:农业出版社),1984.

[11]Climate and Weather Institute,Chinese Academy of Meteorological Sciences(中央气象局气象科学研究院天气气候研究所), Agricultural Meterological Institute of Nanjing Meteorological College (南京气象学院农业气象研究室).Regionalization of agricultural climate resoruces and farming system in China (我国农业气候资源与种植制度区划). Beijing: China Agricultural Press(北京:农业出版社),1981.

[12]SUN Q (孙强), WANG L (王乐), et al.SOFM network based integrated regionalization of cropland conversion pressures in China (基 于SOFM 网 络的中国耕地压力综合分区)[J]. Acta Scientiarum Naturalium Universitatis Pekinensis (北京大学学报(自然科学版)),2008(04):625-631.

[13]National agricultrual regionalization committee (全国农业区划委员会).Agricultural natural resources and agricultural regionalization in China(中国农业自然资源和农业区划). Beijing: China Agricultural Press (北京:农业出版社),1991

[14]LI YJ (李玉军). Analysis on spatialtemporal characteristics of Pingdu City based on RS andGIS(基于RS 与GIS的平度市土地利用时空变化分析)[D].Shandong Univeristy ( 山东大学),2008.

[15]ZHANG Y(张勇),LI XS(李显嵩),et al.Cultivated Land Changes and Affected Factors in East Coast Regions with Rapid Development——A case Study of Pingdu(东部沿海快速发展区耕地变化及影响因素分析——以山东青岛平度市为例)[J]. Territory & Natural Resources Study(国土与自然资源研究),2012(03):24-26.

[16]LI JX (李金玺), JIANG D (姜栋). The characteristics and use trends of lands in Pingdu City(浅谈平度市土地利用的特点及利用方向)[J].Land&Resources of Shandong Province(山东国土资源),2009(7):11-13.

[17]HE P (何鹏), ZHANG HR (张会儒).Study on factor analysis and selection of common landscape metrics(常用景观指数的因子分析和筛选方法研究)[J].Forest Research (林业科学研究),2009(04):470-474.

[18]HE YR (何原荣),ZHOU QS (周青山).Extraction and analysis of regional landscape indexes based on SPOT-5 images and fragstats software (基于SPOT 影像与Fragstats 软件的区域景观指数提取与分析)[J]. Hydrographic Surveying and Charting (海洋测绘),2008(01):18-21.

[19]SAURA S, TORNE J. Conefor Sensinode 2.2: a software package for quantifying the importance of habitat patches for landscape connectivity[J].Environmental Modelling & Software,2009(1):135-139.

[20]SAURA S, PASCUAL-Hortal L. A new habitat availability index to integrate connectivity in landscape conservation planning: comparison with existing indices and application to a case study[J]. Landscape and Urban Planning,2007(2):91-103.

[21]GAN YP(甘永萍).Application of cluster analysis in land use regionalization with a case of Hechi City, Guangxi Province(聚类分析方法在土地利用区划中的应用──以广西河池市为例)[J].Journal of Guangxi Teachers Education University (Natural Science Edition) (广西师院学报 (自然科学版)),1997(02):9-13.

[22]ZHENG JR(郑建蕊),JIANG WG(蒋卫国), et al. Selection of wetland landscape indices and analysis of landscape pattern in Dongting lake area(洞庭湖区湿地景观指数选取与格局分析)[J]. Resources and Environment in the Yangtze Basin (长江流域资源与环境),2010(03):305-310.

[23]CHEFAOUI R. M. Landscape metrics as indicators of coastal morphology: A multi-scale approach[J].Ecological Indicators,2014:139-147.

[24]WANG X, ZHENG D, et al. Land use change and its driving forces on the Tibetan Plateau during 1990-2000[J].Catena,2008(1):56-66.

[25]WANG WY(王文宇),DU MY(杜明义).Guide to ArcGIS mapping and spatial analysis(ArcGIS 制图和空间分析基础实验教程)[M]. Beijing: Surveying and Mapping Press (北京: 测绘出版社),2011.