Abstract:
Objectives Lakes on the Tibetan Plateau are sensitive indicators of global climate change. Monitoring the rapid extent changes of those lakes, driven by factors like precipitation and glacier melt, requires satellite imagery with high temporal resolution. The moderate resolution imager (MERSI-II) aboard FengYun-3D (FY-3D) satellite provides extensive coverage every day, but its capability for monitoring plateau lakes at its native 250 m resolution has not been systematically evaluated. This paper aims to assess the performance of MERSI-II imagery for dynamic lake monitoring on the Tibetan Plateau and explore its potential for enhancement through super-resolution techniques.
Methods Lake area and water boundaries extracted from MERSI-II images are compared against reference data derived from contemporaneous third-party datasets and Landsat-8 operational land imager imagery. Furthermore, a spatial super-resolution technique is applied to enhance the MERSI-II imagery from 250 m to 150 m, and its monitoring performance is re-evaluated to quantify the improvement.
Results At the native spatial resolution of 250 m, MERSI-II imagery achieves an average accuracy of 95.12% in lake area inversion and an average boundary fitness of 91.21% in water extraction. Owing to its high temporal resolution, MERSI-II imagery proves highly valuable for long-term and dynamic water monitoring applications. After super-resolving the imagery from 250 m to 150 m, the performances are further improved. The average accuracy for lake area inversion is increased by 2.62% to 97.74%, and the average boundary fitness is increased by 4.44% to 95.65%.
Conclusions The findings demonstrate that FY-3D MERSI-II imagery, with its high temporal resolution and potential for spatial enhancement, holds great value for long-term, highly dynamic lake monitoring applications.
下载:
