基于斑块的红树林空间演变机理分析方法

现有红树林空间动态分析多是从整体的角度分析其面积的变化情况及其影响因素,着重于面积变化的定量分析和影响因素的定性分析,对其变化的发生途径缺乏深入分析,也不研究其斑块数量的变化动态。由于一定区域范围内的红树林由成百上千甚至更多空间上相互分离的斑块组成,各斑块边界和面积变化动态构成了区域红树林整体的空间分布变化动态,因此,只有深入摸清各个斑块的变化情况及其影响因素,才能对区域红树林整体变化情况作出全面、详细和准确的分析评估。提出了基于斑块的红树林空间演变机理分析方法,首先通过两期高空间分辨率遥感图像提取红树林空间分布信息,在GIS支持下采用叠置分析方法,根据前、后两期各个斑块的空间位置、形状和面积变化情况以及图像表征,逐一分析确定每个斑块变化的主要驱动因子(即变化原因,包括自然过程、围垦、养殖塘和盐田建设、工程建设和人工造林5种)和变化途径(即变化类型,包括稳定、扩张、萎缩、碎化、消失和新增6种),在此基础上构建斑块数量和面积变化的驱动因子-变化途径状态矩阵,通过总驱动量、总驱动率、净驱动量、净驱动率、趋势驱动率、总流量、总流率、净流量、净流率、趋势净流率和作用力等系列指标定量地评估红树林斑块数量和面积的变化动态。该方法不但能够定量地表达了各个驱动因子对红树林斑块数量和面积变化的影响程度,而且能够准确地阐明了红树林斑块数量和面积发生变化的途径,并且还能够准确地反映了每个驱动因子通过何种途径影响斑块数量和面积的变化,实现了红树林空间动态变化分析的定位化、定量化和精确化。

A patch-based method for mechanism analysis on spatial dynamics of mangrove distribution

The traditional analysis methods of mangrove spatial dynamic are generally top-down approaches that only summarize the overall dynamic trends and their underlining mechanism. Those methods can help reveal how much area of mangrove has been lost or gained and what possible drivers are but don't deal with the change procedures and the patch number dynamic. The fact is that a contiguous region of mangrove can be composed with hundreds or thousands of interconnected patches and its overall dynamics is a summation of changes in those individual patches. Therefore, it is important to study patch dynamics, including changes in patch boundary and area, in order to accurately understand and assess mangrove's overall dynamic processes and mechanisms at a landscape scale. In this paper we introduce a patch-based method for analyzing and monitoring temporal changes in spatial distribution of mangrove. We used two-time, high-resolution remote sensing imagery to quantify spatial distributions of mangrove and their changes over time. We used geographic information systems (GIS) to obtain patch-level spatial properties, including spatial position, shape, and area, at each time when the remote sensing data were acquired. We then compare patch-level mangrove between the two times to group mangrove patches into six categories: stable, expanded, shrank, fragmented, disappeared, and new patches. These six categories reflect the dynamic procedures for mangrove. Following similar remote sensing data interpretation, we identified five causes or drivers of mangrove dynamics, include nature process, inning, marineculture and saltern, construction, and plantation. We assigned each mangrove patch into one of the six dynamic procedures and one of the five drivers. By summing up mangrove patches of the same categories, we built a driver-procedure matrix for patch-numbrer and area, respectively. Thus we were able to calculate a series of indices, including the gross driving amount, gross driving rate, net driving amount, net driving rate, predicted driving rate, gross flowing amount, gross flowing rate, net flowing amount, net flowing rate, predicted flowing rate, and acting force, to assess mangrove patch dynamics in patch number and area. This analysis method is helpful to explicitly reveal changes in mangrove patches, in patch number or area, and what drivers trigger the changes and how the changes take place. The information obtained through such analysis provides important clues about the mechanisms involved in the changes in the mangrove landscape. This method means a site-specific, quantified, and precise analysis on mangrove's spatial distributions and has scientific values for the protection and restoration of mangrove.