珠江三角洲东部常平地区景观异质性研究

A STUDY OF LANDSCAPE HETEROGENEITY FOR THE CHANGPING AREA IN THE EASTERN PART OF ZHUJIANG DELTA

本文以１９８８年～１９９６年６个时段景观遥感类型图为主要信息源，利用信息熵法和空隙度指数法对珠江三角洲东部常平地区的景观异质性进行了研究。重点探讨了景观组分的空间分布差异和异质性动态变化情况。结果显示，开发区主要呈宏观异质性分布，果园、农田和林地呈微观异质性分布，城镇和水体的异质性分布特征均出现显著变化。此外，研究结果还表明在同一地区同时使用上述两种主要方法进行研究。

Heterogeneity is the most essential landscape characteristic and one of the most important objects of landscape research. In this paper, we use six types of landscape remote sensing maps along with a new procedure that combines average information levels and a lacunarity index, to the analysis of the landscape heterogeneity of the eastern part of the Zhujiang delta. Heterogeneity types and spatial distributions of landscape elements in working areas are very significant. Arable land and orchards are microheterogeneous elements with a homogeneous distribution and a low heterogeneity level. Developing areas are macroheterogeneous elements with luster distributions and high heterogeneity levels. A woodland looks like a microheterogeneous element, but its distribution is more uneven. Natural environmental changes and human disturbances are the main causes of the significant differences in heterogeneity and spatial distribution among the landscape elements. Homogeneous management of agricultural production converts arable land and orchards into nonheterogeneous elements, but as areas decrease there is a gradual increase in their heterogeneity. Vigorous human disturbances in a local area, like non-agricultural land exploitation, causes the difference in the spatial distribution of towns and other developing areas. Since geomorphology is a medi-scale impact factor, the woodland has more of a cluster distribution pattern because of its effect. Water's high heterogeneity level is produced by a river's linear characteristics after most of its surface area is changed by construction. Its results can be used for classifying the microheterogeneity of landscape elements. The lacunarity index method, although losing some detailed structural information, provides results that are not impacted by any local special element distribution. We also gain important structural information such as area change and autocorrelation. Using the two methods simultaneously in heterogeneity studies compensates for the weaknesses of each other and improves the results.