A landscape pattern gradient analysis method coupling wavelet algorithm for balancing development and protection activities of the coastal zone: A case study

Abstract

Due to the economic importance and ecological vulnerability of coastal areas, the optimization of spatial patterns is important for the sustainable development of coastal zones. In this study, a new landscape pattern index gradient analysis method coupled with wavelet algorithm (GA-WA) was explored to balance the development and protection activities in the coastal zone. First, based on the classification results of remote sensing images, the gradient detection of the coastal zone was carried out using the cumulative moving window method, and the calculation of the landscape pattern index was carried out using Fragstats software to obtain the gradient change curve of the landscape pattern index. Secondly, the wavelet algorithm was applied to micro-analyze the above gradient change curves, and the results of multiscale analysis of the landscape pattern index in the study area were obtained. Finally, it is demonstrated how this method can help in practical decision making. It was found that:(1) The landscape pattern index gradient analysis method coupled with wavelet algorithms offers a new approach to balance the development and protection of coastal zones. This approach presents an innovative method for optimizing the ecological pattern of coastal zone landscapes from the perspective of macro–micro combination, which is not only applicable to coastal zones but also can be extended to other strip corridor landscapes.(2) The gradient analysis of the coastal zone landscape pattern index provides results at multiple spatial scales, which is superior to the analysis at a single spatial scale. The results of the gradient analysis method are finer and more stable, which solves the uncertainty problem of traditional landscape pattern index analysis. There is a significant difference between the landscape indices of the overall landscape and the gradient landscape, in which the AREA_MN index of the ecological land has the largest difference, reaching 110.18%.(3) Wavelet coefficient variance diagrams, real component contour diagrams, and modulus squared diagrams reveal the characteristics, cycles, and patterns of landscape dynamics changes, as well as the distribution at different gradient window scales. These tools help to identify the driving factors and development trends of landscape dynamics changes and solve problems of functional positioning in undeveloped areas, as well as the transformation of microscopic small parcels within developed areas. By analyzing regular cycles of landscape index changes, it is possible to determine the functional positioning of land and to position human intervention through peaks to guard against risky gradient windows.