【作者】 孙永光；

【导师】 李秀珍；

【作者基本信息】 华东师范大学 ， 生态学， 2011， 博士

【摘要】 围垦区作为复杂的自然系统和人为系统作用的综合区域,其景观结构和功能呈现特殊的复杂性,本研究以长江口海岸带不同年限围垦区作为研究对象,利用数量生态学方法,构建能够体现围垦区建设先进性和可操作性要求的围垦区建设综合指标体系(景观、土壤、植被)和将景观结构和功能进行耦合研究的方法,利用CLUE-S模型对围垦区景观动态未来发展趋势进行情景模拟,为河口地区滩涂湿地围垦建设提供科学决策依据。研究内容和获得结论如下：基于1987、1995、2000和2006年长江口TM、ETM遥感影像,利用RS、GIS与数学统计方法,对长江口不同围垦区(3个试验区)土地利用动态进行了分析。结果表明：3个试验区间土地利用多样性指数的变化无显著性差异；各试验区土地利用年变化速率在空间上不存在显著差异,在时间上具有显著差异的土地类型为大棚用地和建筑用地；3个试验区的土地利用程度具有一致性,三区土地利用转移方向的差异也不显著,转移强度相似。长江口不同围垦区的土地利用动态在空间上的不显著差异主要受围垦工程和围垦时间的影响。采用“空间代时间”和主成分分析(PCA)综合评价模型的方法,对长江口奉贤不同围垦年限土地利用动态特征进行了综合评价分析。利用南汇围垦区土地利用动态变化对综合评价的结果进行验证。结果表明：(1)PCA土地利用综合评价指数(F)与农耕用地正相关(水田>大棚用地>旱地>林地>园地),与反映滨海特征的土地利用类型(养殖塘>草地>开放水域>光滩)呈负相关；(2)土地利用PCA综合评价指数(F)随围垦年限增加呈对数增长(R2=0.5119)并在围垦后40年左右达到稳定态。回归分析表明F值能很好反映土地利用程度指数(L)和多样性指数(GM)的变化趋势。借助多元方差分析(MANOVA).地统计分析(GA)、主成分分析(PCA)和典范对应分析(CCA)方法,研究了不同围垦年限土壤理化性质、土壤综合功能的分异及控制因子。结论：1)土壤水分、盐分、电导率和土壤粒度随着围垦年限的增加表现为下降趋势,有机质为上升趋势；速效磷在围垦初期上升,到围垦100年左右又表现为下降趋势；硝氮、氨氮和pH值则在不同围垦年限间变化不大；土壤理化性质在围垦后30年左右达到稳定态；2)在PCA分析中,反映土壤养分因子(硝氮、有机质、速效P、氨氮、-pH和-土壤粒度)的权重值(总计,0.97)要高于土壤限制因子(盐分、电导率和水分)的权重值(总计,0.48),Fs值越高土壤质量越好；3)不同土地利用类型在土壤功能熟化过程中起到不同的作用；4)CCA分析表明,围垦区土壤理化性质主要受到围垦时间的影响,其次受到土地利用方式的影响；土壤功能综合指数Fs与围垦时间的相关性(0.1905)大于土地利用方式(-0.1161)的影响。以2009年7月植被67个样方调查数据、2009年4月调查216样点土壤属性数据和2006年TM(30m)数据作为基础,应用双向指示种分析法(TWINSPAN)、多元方差分析(MANOVA)、除趋势典范对应分析(DCCA)和典范对应分析(CCA)方法,研究了杭州湾不同围垦年限垦区自然植被群落组成及特征分异。结果表明：1)围垦区植被群落以中生植物群落和沼生—中生过渡群落为主,随着围垦年限的变化呈梯度性分布,发现物种丰富度、物种多样性和地上生物量随着围垦年限的增加而增大,但在围垦中后期(>30年)会受到土地利用方式影响而出现反弹现象(下降)；2)物种丰富度指数和物种多样性指数的变化规律与围垦年限的变异趋势相一致,随着土地利用程度的升高其出现先上升后下降的趋势。而物种生态优势度指数则出现下降趋势,地上生物量则略有增加；3)土地利用方式(r=0.55,p<0.05)、土壤水分(|r|=0.53,p<0.05)、土壤盐分(|r|=0.43,p<0.05)、围垦时间(r=0.40,p<0.05)是影响围垦后不同植被群落分布的主要因子。海岸围垦区景观动态变化是人-地关系复杂作用的结果,以1990年、2000年、2009年TM影像、社会统计数据、野外调查数据作为数据源。借助RDA和PCA分析了海岸带景观动态变化驱动力因子贡献度,并利用kappa指数对CLUE-S模型所选驱动力模拟效力进行验证。结果表明：海岸带围垦区景观动态变化人为驱动因子贡献度(57.10%)>自然驱动因子贡献度(42.90%)。CLUE-S模拟正确率可达82%,主导景观类型耕地、未利用地和养殖塘kappa指数均大于0.75,CLUE-S模拟效果较为理想,所选驱动力很好地模拟了规则景观突变的空间分布特征,进而对围垦区2010-2020年未来发展进行了情景模拟,模拟结果表明CLUE-S模型能够较为有效地模拟围垦区景观格局的规则性、集中性和呈带性分布,同时为围垦区管理决策提供了参考依据。更多还原

【Abstract】 The reclaimed zone or polder land has complicated landscape structure and functions due to the mixed influence from natural and anthropological factors.Based on the reclamation zones of different years at the Changjiang River (Yangze River) Estuary, a synthesized indicator system,including landscape, soil and vegetation factors, was established to combine landscape structure and functions together. CLUE-S model was employed to simulate the future scenarios in the reclamation zones. Following are the main results:1) Based on the TM and ETM image of the Changjiang River Estuary obtained from 1987,1995,2000 and 2006, landuse dynamics of 3 typical pilot study areas were analyzed with GIS and statistical methods. The results showed that the 3 study areas were not significantly different in terms of land use diversity indicated by Gibbs-Mirtin diversity index. The land use change rate had no spatial difference among the 3 study areas. But the change rate of Greenhouse land and built-up land were found to be significantly different among time scales, land use levels of the 3 study areas were similar with each other, with similar land transformation trends and intensities.The insignificant difference among the 3 study areas might be caused by reclamation projects and time of reclamation.2) Fengxian area was select for the study or the land use dynamics along temporal series, with combined methods of "spatial sere substituting for time sere" and Principle Component Analysis, while the 2 study areas from Nanhui were used for verification.The results indicated:(1)The comprehensive evaluation index based on PCA(F)was positively related to farmlands(Paddy field>Greenhouse land>dry land>forest land>orchards),and negatively related to coastal land cover types(breeding ponds>grass land>open water>bare flat).(2)The F-value increased with time logarithmically (R2:0.5119), and reached stable around 40 years after reclamation.Regression analysis showed that F-value can reflect the trends for both land use levels(L) and diversity index (GM).3)With multivariate analysis of variance(MANOVA),Geo-statistical analysis(GA), PCA and Canonical Correspondence analysis, soil properties and their possible control factors were analized among zones with different reclamation years.(1)With the increase of reclamation time. Soil moisture, salinity, conductivity and partical size decreased,While soil organic matter increased;Soil available phosphorus increased first, and decreased after 100 years of reclamation.No change trend was found for NO3-N, NH4+-N and pH along time series. Soil properties reached stable around 30 years after reclamation.(2) According to the PCA results, the weight values for soil nutrient factors(such as NO3--N, SOM,SAP,NH4+-N,-pH and-soil partical size) were higher than those for soil limitation factors(such as soil salinity, conductivity and soil moisture). Higher Fs values indicated better soil quality. (3) Different land use types contributed differently during the soil maturation process; (4) CCA results indicated that soil properties were mainly affected by reclamation time, and the influence from land use was secondary. The comprehensive soil function index Fs is more related to reclamation time (r=0.1905) than with land use (r=-0.1161).4) Based on the 67 quadrats data obtained in July 2009,216 soil sampling points data obtained in April 2009, as well as the TM data of 2006, the natural vegetation zones were analized with TWINSPAN, MANOVA, DCCA and CCA methods. The results indicate that:(1) the vegetation community in the reclamation zones was mainly composed of mesophytes and helophytes-mesophytes. With the increase of reclamation time, the communities showed a transitional gradient trend; Species richness (MA), species diversity (H) and above ground biomass also increased. Nevertheless, they appeared to decline slightly in the middle and late reclamation period (>30 years), being affected by land use types, (ii) The change in species richness(MA) and species diversity (H) with reclamation time also tended to increase at first and then decreased with the rise in land use degree. Species dominance (D), however, tended to decline, and above-ground biomass increased slightly, (iii) The distribution of the vegetation community was mainly influenced by the following factors:land use patterns (r=0.55, p<0.05), soil moisture (r=0.53, p<0.05), soil salinity (r=0.43, p<0.05) and reclamation time (r=0.40, p<0.05).5) Landscape dynamics of the reclamation zone is resulted from the interaction between human and natural factors. Based on the TM image data of 1990,2000 and 2009, the driving forces for landscape dynamics in the reclamation zone was analized with RDA and PCA methods. Kappa index and CLUE-S model were also employed to verify the effectiveness of the driving factors selected.Results indicated that:The contribution from human induced factors (57.10%) is higher than that from natural factors (42.9%) to the landscape dynamics in the reclamation area.The prediction accuracy from CLUE-S model reached 82%, While the kappa index for major landscape types such as farmland,unused land and breeding ponds were all higher than 0.75.Therefore the CLUE-S model could be used for future scenarios simulation, and the driving factors selected for this study were also effective in depicting the spatial distribution of landscape transformation.Future development scenarios of the study area were simulated for the next 10 years(2010-2020), which can provide scientific support for decision making in the newly reclaimed area.更多还原