Journal of Harbin Institute of Technology(New Series), Vol. 16, No 3, 2009Landscape pattern and its restoration for coal mine subsidencearea in southern part of Tangshan in ChinaCHEN Ying, ZHANG Tie-min, BIAN Lu-feng, BO Li-jie, CUI Jian-yu, HU Lin陳穎張鐵民,邊錄豐,薄力杰,崔建宇,胡林(1. College of Resources and Environmental Sciences, China Agricultural University, Beijing 100193, China, hulin@2. Tangshan City Landscaping Administration Bureau, Tangshan 063000, ChilAbstract To investigate the ecosystem damage and environment pollution resulted from coal resources exploitation, based on remote sensing image, the landscape pattem was quantitatively analyzed with GIS and landscapepattern analysis software FRAGSTAtS by utilizing multiform landscape patten indexes, such as patch areapatch circumference, patch number, fractal dimension, fragmentation index, connectivity index and so on. Re-sults indicate that the restored landscape type occupies the control status and accounts for the majority of thewhole area, and needs to be further restored and adjusted because of its intemal body with smaller patcheslower fractal dimension and better connectivity. The landscape type without restoration has destroyed the ecologyand beauty of the whole landscape. In addition, some suggestive instructions for restoring various kinds of pat-ches are put forward according to the quantitative analysis.Key words: coal mine subsidence area; landscape pattern; ecological restoration; southem part of TangshanCLC number X1714Document code: AArticle ll:10059113(2009)034030346Coal is Chinas most main one-time energy, it playsWhen after analyzingan important role in the development of national econo- pattem, to set up a restoration model with garden land-e. To guarantee the normal development of the na- scape and leisure sports green space in the coal mine sub-nal economy by 2020, expected coal still accounts sience areas surrounding the city, is still on the way andfor about 70% of disposable energy. However, The lacks of related domestic research. Based on the Tangshanground subsidence caused by the exploitation will be- southem subsidence area as the research object, the com-of the ponent composition, quantitative analysis of landscalandscape ecosystem in subsidence land is imperative. cology theory frame, repair and not to have the land typeMany foreign scholars, such as John M, Santiago structure repair analysis, thus for the next step forwardSaura, Javier Manrtinez-Millan, have made some stud- constructive Suggestions on repair directionies respectively about how to restore and reconstruct theindustrial polluted areas to their steady ecosystem, how 1 Research Area Descriptionpeople's early activities have affected the landscape e-cosystem and what changes have taken place for theThe climate in Tangshan city is characterized bylandscape patterms under the influence of the society, warm temperate monsoon climate. It has cold and drywinter mostly influenced by north wind, hot and humidIn domestic, some scholars like Wang Yanylin, summer mostly influenced by southerly wind. The an-Zhang Shuli, Huang Yixiong, Bai Zhongke, have stud- nual average temperature is 11 C. The highest averageied in the area as well. They researched around the temperature in July is 25 C and the lowest in Januaryal reconstructionis-5.5℃. The anlandscape construction direction for social-economic- 623 mm. The freezing period begins in December andnatural complex ecosystem, landscape ecological con- ends in next March, with a frozen soil depth of 0.5struction of ormamental orchard in dump, the tec0.8 m. The frost-free period lasts more than 180 dgy and the model for the landscape ecological recon- The dominant wind is north wind, with a maximumstruction of waste lands, landscape heterogeneity of wind speed at 25 m/s. The average annual total sun-reclamation-area, system stability after ecological recla- shine2578 h tn 2891 h with a total sola中國煤化工Received 2007-12-03CNMHGSponsored by the National High Technology Research and Development Program of China( Grant No. 2005AA644030)and the National Key Technology RD Program in the llth Five Year Plan of China( Grant No. 2006 BAD16B09)303Journal of Harbin Institute of Technology( New Series), Vol. 16, No. 3, 2009The Tangshan southern subsidence is the part of 2.2 Determination of Landscape ClassificationKailuan coal mine which has a more than 100 years ofThe landscape in the study area was carefullymining history. It is located in downtown 670 m south classified into two grades according to the land typesand is the most influential coal mine subsidence in the and their restoration statuswhole city. The total study area is 28 km. The terrain1)Patch level was divided into restored componentin the plan scope is primarily smooth and the scope of unrestored component, populated country and buildingthe height changes is less than 1 m. Only western life site according to the present situation. Restored compo-rubbish surrounding terrain, the highest elevation 45 m nent included cultivated land, forest land, Nanhu Parkreached up, and surface subsidence area belongs to the and green space for sports. Unrestored component includ-lowest elevation of around 5 m. For more than ten ed refuse dump, fly ash, coal gangue dump, waste landyears, the local government has taken various measures and water area. Populated country and building site wereto transform the subsidence, such as govemment invest- studied as one landscape type.ment, raising funds from various social aspects, obliga2)Class level was based on patch level. It dividedtory and so on. Making full use of the existing condi- landscape types into nine components( Fig. 2): TItions, they have built the artificial plant communities cultivated land, T2-forest land, T3-green space foradapting to different section, different region and dif- sports, T4-Nanhu Park, T5-refused dump, T6-fly ashferent geological conditions. The plants ecological T7-waste land, T8-water area, T9-residential area andfunctions have thoroughly improved the ecological envi- building site. Among these components, T3 was reronment in the subsidence. Up to now, some land- stored into a golf course invested by enterprise. Byscape including cultivated land, forest land, Nanhu present, the site engineering of the first nine tunnelsPark, green space for sports, have been restored, But has been completed. It was the unique green space forome waste land, refuse dump, coal gangue dump and sports built on a coal mining wastelands in China. So itwater area are still not restoredbecame a new restoration form of wasted land transformation. T4, Nanhu Park has become a good ecological2 Methodslandscape after more than ten yearstreatment. Thisproject won the Chinese Human Settlement Environ2.1 Data Sourcement Example Awards in 2002 and Dubai the Best In-Based on 1: 10000 land use SPOT remote sensing ternational Example for Residential Environment Im-map( Fig. 1)from October to November in 2006 and provement Awards in 2004. It was annunciated to befield investigation data, the landscape indices were cal- the first batch of National Urban Wetland Park by theculated using remote sensing map vectorization soft- Ministry of Construction in 2005. Combined with localware-Arcgis and landscape pattem analysis software- governments planning strategy, such landscape modelFragstats. The landscape pattern and its planning di- dominated by garden landscape and green space forrection were studied according to such indices as patch sports would become the main landscapeoutenfractal dimension, ete, (18' patch area, patch numbessize, patch circumferencesubsidence areaCultivated landForest landGreen se for portsRefused dumpu Nanhu parkM Water area中國煤化工pemFig 1 Mapping from remote sensing image in the south of 2.3CNMHGMethodsape pattern indices of each patch type were304Journal of Harbin Institute of Technology( New Series), Vol. 16, No. 3, 200calculated].Theteristic of landscape pattern is circumference of the ith landscape type, A is total ar-was represented by some eigenvalues such as patch ea of the landscapes in study area. A, is area of the ithsize, circumference, size fraction, shape, distribu- landscape type, N is landscape fragmentation indexion, etc. The main characteristic values involved in Np is total patch number of landscape pattems, Nisthis study and their specific meanings were listed in ratio of landscape type area to its minimum patchTab. 1. Where n is total number of the patches in studarea, n, is patch number of the ith landscape type, PTab 1 Landscape pattern index[ 20)patch( Np)is the most basic spatial characteristic of landscape pattenPatch area(Ap)lects the difference of species, energy and nutrient inner land-Sun=nent. Patch size is the key to reflect landscape heterogeneityand will affect the species component and diversity in a landscapeFractional dimension describes the shape feature of a patch. It reflectsFractional dimension(Dr) Dr" InAthe irregularity of a geometric figure. When Fn is close to l the patch willbe more similar and become stronger against the disturbance.Fragmentation index(N, N..(NeFragmentation index means the degree of fragmentation in the landscapeResolution describes the discrete degree of element distribution in a land-Resolution( F)F=2(A/A)scape type. The higher the resolution is, the vaster the landscape disper-trol position in the whole landscape and formed a background of the study area( Tab. 2). Unrestored land3.1 Landscape Patterm Analysis of the Patch Level showed a higher fragmentation degree and the strongestThe total researched area is 28 km. Restored heterogeneityland includes cultivated land, forest land, green spaceFor the circumference of each restored landscapefor sports and Nanhu Park. Cultivated land was mainly type, there existed significant difference between rerestored on the basis of original subsidence area. For- stored land, unrestored land and residential area. Butest land was built through afforestation in a large scale the ratio of restored lands to the whole landscape andrganized by govemment or individual from 1986. To- there area were similarday 's Nanhu Park was built through enlarging the lakeThe fractional dimension values of three landscapelevel in the north, connecting small water level, impro- types were all lower than that of the whole landscapeving water quality, building landscape, and rationally which was 1. 313. It means that they had strong self-allocated plant community. In addition, a golf course similarity and great disturbance intensity. Among thesewas built around the westem water level invested by an three landscape types, residential areas had lower fracenterprise in 2004. Therefore a new landscape resto- tional dimension, which means that they had relativelyration model was basically formed for subsidence land structured shapes. Unrestored lands took the secondreformed to green space for sports. Based on area com- place and could be made into consideration as a wholeparison method, the restored land was 49. 06%, nearly in later restoration. The edge shapes of restored lands50% of the whole study area. It means that the new re- tended to complex for their different existing forms andstored landscape model has gradually occupied the con- distributionsTab 2 LaiPatch mumber Total area Area proportion Mean patch Total circumference Mean patch/blockcircumference/km dimensionMU中國煤化工8.05228.CNMHGWhole landscap28.4240.23I313.8422.552I.3I3Journal of Harbin Institute of Technology( New Series), Vol. 16, No.3, 20093. 2 Landscape Pattern Analysis of the Class Level Others were construction waste which destroyed theThe landscape was divided into nine landscape el- whole environmental landscape though having a littleement types with totally 123 patches based on the land- lower pollution degree. The landscape in the form ofscape types in the patch level. The area and circumfer- wasted land accounted an area of 2. 887 km withence distribution of each landscape patch were very im- 13 patches. Water area has an area of 1. 931 km withbalanced(Tab. 3). Residential area is biggest which 29 patches. The standard deviation of area and circum-implies the whole landscape is disturbed mostly by hu- ference between forest land and waste land was biggerman being, so its reformation should be done around than that of other types, showing a greater differenceland with an area of 7. 763 km, 27. 311% of total are- heterogeneity. The water area had more patches, imethuman beings'activities. Next came the cultivated the imbalance of patch size and the more significaa. Restored green space for sports, with an area of ler area and circumference. It scattered with small area0. 436 km, mainly included the already built golf and had the biggest Fragmentation index. Fragmentacourse, the water area and the grassland, which were tion index of residential area took the second place. Itdefined as three patches to study. Fly ash, an area of hadn't formed a regular settlement patten because of1. 308 kmaccounts for 4. 602%of the total area, with the scattered habitation condition. Refused dump hadone patch. Here is a huge source for air pollution. the biggest resolution index showing the highest diver-Nanhu Park was studied as a whole restored type with gence rate. The total resolution index showed that thethe area of 0. 899 km". Refused dump had 11 patches unrestored land distributed more disperse than the re-with the area of 0. 302 km. One patch was a stored landscape types. This indicated that the restora-0. 155 km of domestic waste heap which greatly affect- tion project had improved the vicious circle of theed the surrounding soil and water quality although a whole landscape greatlyepage control project had been done in advance.Tab 3 Landscape pattern characteristics of the class levelPatch number Area Standard Cireumfe. Standard Type area/ Fragmentation Fractional Resolution/blockdeviation/km? rence/km deviation/km total area/% index dimension index313.84220.6231.313TI cultivated land7.7630.47969.4543.36927.3110.6331.2313.8084.8460.55465.9665.77717.0490.7391.2626.101T3 Green space for sports0.4360.1581.5340.1061.1296.80811.4860.0437.2351.0620.3871.18997.894T Waste land0.58533.11610.1570.4571.21310.246.7941.24215.310T9 Residential area0.43979.4482.9410.9151.2453.672The patches'grades of nine landscape types were tion of patches at different level for each pattem typealculated according to the grain structure analysis. Tab. 4 indicated that all the nine types were mainlyThe grain structure of landscape patch refers to the composed by small patches, even in the restored larquantity combination conditions of patches with differ- scape types. The small patches came first, and thenent area and size level in the landscape patch 21.The came the large ones. And also they had similar distripatches in the study area was divided into 6 levels, i. bution( Fig. 1). Small patch with worse connected-e. min-patch( $0. 02 km), small patch(0. 02- ness might probably induce the decrease of species0. 2 km), middle patch(0. 2-0. 3 km), nakadai habitatpatch (0. 3-0. 4 km), large patch(0. 4-0.6 km)This中國煤化工pecies extinctionity which will af-and ultra-large patch(>0.6 km?'). Grain structure of fect thCNMHG and energy be-pattern types were described according to the distribu- tween patches to a great extent.Journal of Harbin Institute of Technology( New Series), VoL 16, No 3, 2009Tab 4 Grain structure of pattern types in the class levelLandscape patch typetalA≤0Qa0Q6TI cultivated LandT2 forest landT3 Green space for sportsT6 Fly ashWhole landscape315123The boundary effects of patches were revealed by logical benefit and the embody of their landscape val-heir circumference/area ratio. The larger the circum- ue, except forference in unit area was, the higher degree the patch justmenttype was cut along boundary, vice versa, the better thepatch connectivity was. Therefore the analysis of thedistribution and the judgment of the connectivity forrious landscape types will provide the scientific evi-dence for further landscape restoration( Fig. 3 ).Thewaste landboundary effect value of each patch showed in Fig. 3was made of integrated boundary effect value of patcheswith different size fraction, That water area has the lar-gest boundary effect meant that it had the highest cutaddition. it's a coal subsidence area. The imbalanceFig 3 The boundary effects of patchesnce ana scattered distribution of water area and didnt form a 4 Conclusion and Discussionwhole area. It would therefore hinder the execution ofcomprehensive treatment and the building of land1)As restored landscape types, Nanhu Park andcape Refused dump came to the second for it wasgreen space for sports had improved the landscape to amainly made of domestic waste and construction wastgreat degree indicated by some indices such as areawhich was related to the landscape pattem where resi- fractional dimension, patch scales, and so on. Theirto the thisas took up the majority. Wasted land came boundary range should be expanded to decrease theirdentiale all unrestored landscape boundary effect values and strengthen their connectivetypes. Fly ash and coal gangue dump had a patch ty. Such would let these two types possess a matrix atnumber of I so the shape relatively regular and it had tribute in the integrated landscape ecology and there-the lowest boundary effect value. Residential area and fore become a background of landscape patten, corbuilding site could be adjusted appropriately though trolling the mobility of ecology, material and energythere were some regions with too high boundary effect This would be extended as a new form of landscape ec-value. The same is to the forest land. The connectivity ological restoration for coal mining wastelands sur-was especially significant for not only was it used as aprotective corridor of farmland and a channel for meio-2)Though possessed a greater proportion in thefaunas, but also it was a key element in natural vegeta- whole study area, cultivated land was mainly composedtion landscape which could keep the connectivity of for-byh amall scaled field block, leadingest land and therefore assure the normal exertion of to a中國煤化工 ainst intensive utilNanhu Park, and green space for sports had better con- sheCNMHGanning of farmlandntegration and formMost region had good integrity which could a field-cubed, roadbe helpful to the full playing of their compositive eco-307Journal of Harbin Institute of Technology( New Series), Vol 16, No 3, 2009completed and land-leveled idyllic ecological pattern1996,10(2):62-68.( in Chinese)Such farmland makes the whole landscape with ecologi- [10 ]Fan Yinghong, Lu Zhaohua. Major ecological and ercal patch attributesmental problems and the ecological reconstruction te3)In further planning, one part of forest landgies of the coal mining areas in China. Acta Ecologica Sini-ca,2003(10):2144-2152.( in Chinese)could be used as a component of park landscape, com- [11]Li Xinju, Hu Zhenqi, Li Jing, et al. Research progress ofposing a background of the landscape. Another partreclaimed soil quality in mining subsidence area. Transac-distributed around farmland could be used as farmlandtions of the Chinese Society of Agricultural Engineeringshelterbelt network. Trees should be planted vertically2007,13(6):276-279.( in Chinese)to the main wind direction and matched species to the [12]Bi Rutian, Bai Zhongke, Li Hua, et al. Landscape changesites. Through enlarging and adjusting planting areaanalysis of reclamation land in opencast coal mine based orand range, forest land can serve as a corridor in the in-3S technology. Jourmal of China Coal Society, 2007(11):tegrated landscape ecology1157-1161.( in Chinese)4)With a series of technical measures, the[13]Yang Guojing, Xiao Duning. Forest landscape pattem andstored land, waste land, refused dump, coalfragmentation: A case study on Xishui Natural Reserve inQilian mountain. Chinese Journal of Ecology, 2003, 22dump and water area could be restored to park(5): 56-61.(in Chinese)space for sports, farmland, and forest land according to [14JJin Yingshan, Zhai Mingpu, Wang Chao, et al. Spatialtheir location and soil conditions. This will turn thedistribution pattems of scenic forest landscape in Xishanwaste into treasures by making rational, scientific andtain of Beijing. Joumal of Beijing Forestry Universityeffective use of land2007,29(3):74-80.( in Chinese)[15]Fu Bojie, Chen Liding. Landscape diversity types andReferencestheir ecological significance. Acta Geographical Sinica1996,51(5):454-461.( in Chinese)china's coal industry. Jourmal of Chinese Coal Economy [16]Liu Canran, Chen Lingzhi. Landscape scale fractal analy-sis of patch shape in the vegetation of the Beijing RegionCollege, 1996(1): 15-20.( in Chinese)Acta Phytoecologica Sinica, 2000, 24(2): 129-134. (in[2]Zhao Jingzhu. Dynamic indicator system for measurement ofpatial pattem in landscape ecology. Acta Ecologica Sini- [17]Ma Keming, Fu Bojie, Zhou Xiaofeng. Studies on speciesca,1990,10(2):182-186.( in Chinese)[3] Hanski 1. Metapopulation dynamics. Nature, 1998, 396and pattern diversities of the forest landscapes of dongling-han Mountain Region, Beijing, China. Acta Ecologica Sin-41-49a,1999,19(1):1-7.( in Chinese)[4] Li Habin, Wu Yegang. Quantitative methods in landscape [ 18]McGarigal K, Marks B J. FRAGSTATS: Spatial Pattemecology. Advances in Modern Ecology. Beijing: China Sci-ence and Technology Press, 1992.( in ChineseAnalysis Program for Quantifying Landscape Structure. Cor-allis: Oregon State University, 1993[5] Naveh Z, Lieberman A S. Landscape Ecology. Theory and [19]Shen Jin, Zhao Tiezheng. Brown land and green space netApplication. New York: Springer-Verlag, 1984[6] Wu Jianguo. Natural conservation and natural conservationwork: Special reutilization of the Nanhu coal mining settle.ment area in Tangshan. Architectural Journal, 2006(8)biology: Concepts and modelings. Advances in Modem E-28-30.( in Chinese)cology. Beijing: China Science and Technology Press, [20] Forman R TT, Godron M. Landscape Ecology. New1992.(in Chinese)York: John Wiley, 1986[7] Zhang Zhibin. The theoretical basis of biodiversity conser- [21]Guo Jinping, Xue Junjie, Li Zhiqiang. Studies on the dyvation. Principles and Methodologies in Biodiversity Studies. Beijing: China Science and Technology Press, 1994cs of patch size and grain structure of landscape elements in the forest restoration process. Acta Ecologica Sini( in Chinese)ca,2000,20(20):218-223.( in Chinese)[8] Bian Zhengfu, Zhang Yanping. Land use changes in [22 ]Fu Meichen, Chen Qiuji, Mi Jing, et al. Farm landscapeXuzhou coal mining area. Acta Geographica Si61(4):349-358.( in Chinese)design and planning and application of 3S. Transactions of[9] Li Qiang, Ci Longjun. Landscape heterogeneity and ecolog-Chinese Society of Agricultural Engineering, 2003, 19(1):196-199.( in Chinesical construction for the Shenmu-Fugu Dongsheng miningrea. Journal of Arid Land Resources and Environment中國煤化工CNMHG308