Abstract:Spartina alterniflora is an important invasive species in coastal beach ecosystems, and its distribution and cover are basic parameters that affect the ecology of these wetlands. In this study, S. alterniflora distribution and fractional vegetation cover were investigated, using high spatial resolution satellite remote sensing images from SPOT6 (6 m), and low altitude Unmanned Aerial Vehicle (UAV) images (10 cm) from Sansha Bay (San Du Ao). The latter is a typical coastal wetland area in Ningde, Fujian Province, China. For the extraction of SPOT6 remote sensing information, support vector machine and maximum likelihood classification methods were explored. In addition, a geographic information system (GIS) technique was used to reduce errors. As a result, an accurate S. alterniflora distribution map was obtained. Furthermore, the normalized difference vegetation index (NDVI) of image pixels was calculated for the S. alterniflora distribution area. Subsequently, with the NDVI of pure S. alterniflora pixels and pure beach soil pixels, the fractional vegetation cover could be calculated. S. alterniflora area was 20.19 km2 in the total study region. S. alterniflora area in the Jiaocheng district was 9.63 km2 and was distributed mainly in Zhangwan and Southern Jiaocheng, accounting for 47.70% of the total study area. S. alterniflora area in Fuan County was 5.20 km2, was distributed mainly in Yantian Bay and Magang Bay, and accounted for 25.76%. The fractional vegetation cover of S. alterniflora was mostly a medium (40%-60%) and a high degree (60%-80%) cover. In contrast, much less of the S. alterniflora cover displayed a low degree, (<40%) or much higher degree (>80%), of fractional cover. Statistical tests indicated that the fractional vegetation cover area of 40%-60% and 60%-80% represented 5.44 km2 and 4.95 km2, respectively, and accounted for 26.92% and 24.52% of the total area, respectively. Overall, the fractional vegetation cover area of a medium and high degree was 10.39 km2, and accounted for 51.44% of the whole S. alterniflora area. Based on these results, it is necessary to increase the monitoring and additional studies on S. alterniflora in Sansha Bay. In addition, relevant measures should be taken to reduce biological invasion damage, and to preserve the efficiency of this coastal wetland's ecological functioning. To evaluate the estimation accuracy of the S. alterniflora vegetation cover, 26 sample sites were selected randomly according to the overlapping SPOT 6 and UAV image region. An accuracy analysis indicated a root mean square error (RMSE) of 0.117, and a determination coefficient R2 of 0.918. Therefore, our S. alterniflora fractional vegetation cover results, as estimated by SPOT6 high spatial resolution remote sensing, had a satisfactory precision. These results could therefore be used as a reference for ecological coastal wetland research.