海滨湿地景观过程模型与模拟研究——以江苏盐城保护区核心区为例

景观过程模型已经成为深化景观变化研究的重要手段并日益受到重视。海滨湿地是一种高度复杂和动态变化的景观系统，构建景观过程模型能够从机理上解释湿地景观时空动态演变过程。研究中，将盐城国家级自然保护区核心区划分为人工管理区和自然湿地区两种模式，根据2000、2006、2011 年3 个时相的景观资料以及野外生态监测数据，构建海滨湿地景观过程模型，结果如下：①利用空间代替时间并结合灰色关联分析，确定了海滨湿地景观演变的关键影响因素为土壤盐度和水分；②运用人工神经网络的方法，分析了土壤盐度和水分的空间分异，并通过图层叠加和分级统计，得出了土壤盐度和水分的阈值范围以及变化规律；③运用GIS-Matlab-CA技术构建了基于关键生态过程的海滨湿地景观过程模型，模型总体精度都在80%以上，Kappa 系数都在0.7 以上，一致性检验结果表明空间位置一致率决定了整个景观模拟的精度；④运用已构建好的模型对研究区2012-2020、2025 年的结果进行了模拟预测，预测结果表明，盐城海滨湿地芦苇沼泽和米草沼泽扩张、碱蓬沼泽面积减少的趋势将愈加明显。

Building Model Based on Processes and Simulating Landscape Change in the Coastal Wetlands—A Case Study of the Core Area in Yancheng Natural Reserve

Wetlands are one of the most important ecological systems on earth. Their unique ecosystem function has been sometimes called"the earth's filtration system". Coastal wetland is an important type of wetland, which is not only a rare ecosystem but is also often found in areas disturbed by human activities. Landscape changes have become the most significant feature of this fragile zone. Landscape process model has already become an important means for the further development on the research of landscape changes, and its importance is recognized increasingly. Costal wetland is kind of a highly complicit landscape system, with dynamic changes. The muddy coastal wetland in Yancheng is one of the most typically primitive coastal wetland for its significant strategic position in global biodiversity protection. Over the years, under the influence of the nature and human socio-economic development, the wetland landscape structure and function changed dramatically. In this case, constructing a landscape process model could be used to explain spatiotemporal dynamic change process of the wetland landscape on mechanism. The article divides the core area of Yancheng National Nature Reserve into two patterns: the human-managed one and the natural one. According to the landscape material and wild's ecological monitoring data of 2000, 2006 and 2011, the costal wetland landscape process has been constructed, and the outcome is as follows: 1) Substituting space for time, combined with grey relational analysis, it concludes that the key factor that influences costal wetland landscape change is the soil salinity and moisture. 2) Using artificial neural networks can achieve the space distribution of the soil salinity and moisture, meanwhile, through layer overlay and hierarchical statistical method, it figures out the scope of threshold value and change rules of the soil salinity and moisture. 3) Constructing costal wetland landscape process based on the key ecological process with the GIS-Matlab-CA technology and inspects its overall precision, Kappa coefficient and consistency, it turns out that the concordance rate of spatial location finally determines the accuracy of the whole landscape simulation. 4) Use the founded model to predict the outcome of 2012-2020 and 2025; it forecasts that the tendency of reed swamp, spartina marsh expansion and salsa swamp decrease in Yancheng costal wetland will become much more obvious.