基于模式识别的景观格局分析与尺度转换研究框架

A framework on landscape pattern analysis and scale change by using pattern recognition approach

景观格局分析和尺度转换是生态学和地理学研究的核心问题之一。尽管多年来投入了大量的精力,然而由于尺度效应的复杂性和尺度转换过程的不确定性,仍然没有找到合适的方法。目前关于尺度转换的研究存在两个误区,其一是过分重视空间尺度的转换,忽略了过程尺度的转换;其二是过分强调了对不同尺度间数量关系的外推与转换,忽略了不同尺度间生态规律的外推与转换。在系统分析了景观格局与尺度转换研究现状、存在问题和难点的基础上,提出了借用模式识别的原理和方法,针对特定的生态过程,开展景观格局分析与尺度转换的思路。认为尺度转换的关键在于通过识别不同尺度上影响生态过程的主导因子,找到各个尺度上“格局(环境因子空间组合)-过程”、尺度间“格局-格局”之间的对应关系,通过建立“环境-格局-过程”模式识别数据库,就能够建立不同尺度之间基于模式识别的尺度转换方法。

Landscape ecology mainly deal with the realtionship between landscape pattern and ecological processes. However, because of the difficulty on ecological process quantification from various scales, few achievements were made. To clearlify the relationship between landscape pattern and ecological processes at different scales, a key bottleneck to be resolved was the methodology on how to extend ecological rule from small scales to large scales. Due to the complexity of scale effects and the uncertainty of scale change, few scientific methods on scale change were established despite years of substantial research efforts. Landscape pattern analysis and scale change are one of the core issues in ecology and geography. However, due to the complexity of scale impact and the uncertainty of scale change, an appropriate method is yet to be found even though much research effort has been given toward this issue. There are existing two misunderstandings in scale change. The first one tends to overly emphasize the effect of spatial scale change while ignore the effect of process scale change. The second one is to pay too much attention on the extrapolation of quantitative relationships among different scales, but lack of understanding about the ecological impact of different scales. This research points out a new idea, based on the systematic analyses on the previous studies on landscape pattern and scale change, that is to conduct landscape pattern analyses and scale change for a particular ecological process using pattern recognition approach. We believe that the key issue in scale change is to establish relationships between the pattern and process for each scale and the relationships between the patterns of different scales through identifying dominant factors of ecological processes at each scale. A new scaling mechanism can be established after the database of environment-patternprocess pattern recognition was built.