滩涂湿地入侵种互花米草植被覆盖度的高空间分辨率遥感估算

互花米草是沿海滩涂生态系统的重要入侵物种,其分布状况和覆盖度是湿地生态研究的重要参数和基础。以宁德三沙湾(三都澳)滩涂湿地为研究区,以SPOT6 6 m空间分辨率卫星影像为数据源,对互花米草分布和植被覆盖度进行研究,并与同期10 cm空间分辨率无人机影像进行比较验证。结果表明,影像覆盖区域内互花米草总面积为20.19 km~2,其中蕉城区互花米草分布较广,面积为9.63 km~2,占研究区互花米草总面积的47.70%。互花米草植被覆盖度整体上以40%-60%和60%-80%的中、高覆盖度分布为主,其分布面积分别为5.44 km~2和4.95 km~2,占互花米草总分布面积的26.92%和24.52%,而40%以下的低覆盖度和80%以上较高覆盖度分布相对较少。SPOT6遥感影像估算得到的互花米草植被覆盖度具有较好的精度,与无人机影像值之间的均方根误差RMSE为0.117,线性回归决定系数R~2为0.918,可用于滩涂湿地植被覆盖度分析。

Estimating the Spartina alterniflora fractional vegetation cover using high spatial resolution remote sensing in a coastal wetland

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.