Study the changes in wetland landscape patterns using remote sensing images – A case study of Jiulongjiangkou, Fujian, China

• Published in: IOP conference series. Earth and environmental science 2023-04, Vol.1171 (1), p.12049
• Main Authors: Liu, Zhong, Zhu, Lirong, Chen, Zhongsheng, Huang, Xiuping, Lin, Chih-Wei
• Format: Article
• Language: English
• Subjects: Deep learning; Image acquisition; Image resolution; Landsat; Landsat satellites; Pattern analysis; Remote sensing; Satellite imagery; Spatial discrimination learning; Spatial resolution; Storm runoff; Urban development; Urban sprawl; Wetlands; Wind speed
• ISSN: 1755-1307; 1755-1315
• DOI: 10.1088/1755-1315/1171/1/012049

Remote sensing images play an important role in landscape pattern analysis. Medium/low-resolution remote sensing images are widely used because they are freely available. However, sometimes the results are not precise enough due to the data quality. Therefore, this study utilized the 30m medium-resolution remote sensing images acquired by the Landsat-8 satellite, which is associated with super-resolution technology of deep learning to upsample the spatial resolution from 30 * 30 m 2 to about 7.5 * 7.5 m 2 and perform artificial interpretation on the upsampling images. Several indexes of the landscape pattern were used to analyze the landscape composition and landscape pattern evolution trend of the Jiulongjiangkou wetland in Fujian Province from 2013 to 2021. The results showed that: (1) The total wetland area decreased from 101.26 km 2 to 92.65 km 2 . (2) The silt beach has the most dramatic change; its area is decreased by 6.53 km 2 . (3) The speed of urban expansion is obvious. The industrial and mining lands increased from 9.15 km 2 to 22.15 km 2 , and the traffic land increased from 1.9 km 2 to 3.46 km 2 . The study found that wind speed, temperature, radiation index, storm runoff, and human activities are the five main driving factors affecting wetland change.