Currently, the U.S. and China are the two largest national economic entities in the world. However, it is noticeable that the two countries have considerably different strategies for economic development, environmental protection and land supply in coastal zones. In order to understand the coastline dynamics, land use land cover (LULC) changes and land management policies in the U.S. and China, a case study of the Tampa Bay (TB) watershed, Florida, U.S., and Xiangshan Harbor (XH), Zhejiang Province, China was conducted. The two areas possess similar humid subtropical climate and dense population, but experienced different anthropogenic impacts. TB sat at a developed stage with sound environmental laws, regulations and projects to preserve natural landscapes. XH was at a developing stage and focused more on an economic development in the last 30 years. Comparing the LULC change patterns and the major driving forces for the changes between the two study areas, governments and public could know what factors cause the land use conversion and how to preserve the natural landscapes. A new water index called the weighted normalized water index (WNDWI) was proposed to extract coastlines in TB and XH since current water indices could not classify turbid water bodies and shadow areas well. Two threshold methods (i.e., Otsu threshold method and multiple thresholds method) were implemented to find an optimal threshold to segment the water from the land. The experiments demonstrate that the WNDWI algorithm can achieve high accuracies to classify water from land with an optimal threshold in the two study sites. Coastlines in 1985, 1995, 2005 and 2015 in TB and XH were extracted and the changes were detected and highlighted. The results indicate that coastlines in TB were mostly stable, while those in XH had been undergoing intensive human interferences, indicating that XH was at a developing stage. Major anthropogenic impacts on XH coastlines are land reclamation and aquaculture, resulting in an impacted area of approximately 20.3 km2. The land cover maps of TB and XH in 1985 (1986), 1995, 2005 and 2015 were produced by classifying Landsat images using the random forest algorithm. The reflectance distributions of the land cover types indicate that it is difficult to classify agricultural land, rangeland, upland forest and wetland if using the optical bands only from a single Landsat image. Multi-seasonal image composites and the land surface temperature (LST) band were involved in image classification to achieve higher accuracies. The overall accuracies (OAs) of the land cover map of TB in 2015 and that of XH in 2005 were increased by 5.14% and 4.33% after adding the LST band. The OAs of the four years’ land cover maps of TB range from 81.14% to 83.43%, whereas those of XH vary from 84.67% to 87.67%. According to the experimental results, the total urban area increased by 11.8% in TB, while that in XH increased 138.9% during the last 30 years. Wetland in TB reduced by 8.3% while that in XH reduced 49.0%. The results of logistic regression analysis indicated that the density of wetland is a major driver for urban growth in TB with a strong negative impact while the relationship is opposite in XH. It is worth noting that XH has been undergoing a rapid urbanization and industrialization process with a vast amount of natural landscapes converted to urban areas, whereas TB has already passed the developing stage and issued environmental laws and programs to preserve natural landscapes from human exploitation. In terms of preserving natural landscapes and protecting the vulnerable coastal environment for our next generation, the coastal planning decision makers in XH should not only consider economic values and short-term benefits but also integrate values of ecological, social, and cultural and long-term benefits when making coastal management decisions.
Identifer | oai:union.ndltd.org:USF/oai:scholarcommons.usf.edu:etd-8489 |
Date | 28 June 2018 |
Creators | Guo, Qiandong |
Publisher | Scholar Commons |
Source Sets | University of South Flordia |
Detected Language | English |
Type | text |
Format | application/pdf |
Source | Graduate Theses and Dissertations |
Page generated in 0.0015 seconds