Temporal-spatial study on enclosure culture area in Yangcheng Lake with long-term landsat time series

Abstract

Accurate extraction of temporal-spatial information of the enclosure culture area is important to the protection and sustainable development of a lake. In this study, we presented a new method to extract enclosure culture area with long-term Landsat time series. First, we collected all the available Landsat 5 and 8 scenes over Yangcheng Lake from 1984 to 2017 (a total number of scenes of 396). Then, a new algorithm using spectral and texture information was proposed to extract the enclosure culture area. Furthermore, a temporal filter was designed to smooth the time-series results for eliminating the deviation caused by data inconsistency between years. Finally, the manually determined enclosure culture area based on the Google Earth high-resolution image was used as the ground truth reference. Results showed that the producer's accuracies of the final results were between 72.57% and 88.53%, whereas the user's accuracies were between 79.79% and 98.10% for three selected years (i.e., 2002, 2010, and 2015). The enclosure culture area in Yangcheng Lake had experienced five stages: no-enclosure culture (1984-1994), rapidly growing (1994-1998), peak (1999-2002), rapidly declining (2003-2006), and stability (2007-2017), with the largest area up to 100 km2 and the current area maintained to 30 km2. By studying the vegetation index (normalized difference vegetation index in this study) of the enclosure culture area, we found that the amount of float vegetation grown significantly after 2002. Comparison with the water quality data showed that the water quality of Yangcheng Lake had not improved significantly since it reached IV level in 2003, although the enclosure culture area had been decreasing since 2002. Therefore, in the development of lake culture, the government should insist on sustainable development and develop the lake economy while maintaining the lake water quality.