中文摘要：

利用太湖水体藻蓝素的实测数据,基于蓝藻的光谱特征分析,选择MODIS250m分辨率的卫星遥感影像,建立了藻蓝素估测模型.研究表明,该模型可以较为准确地识别新生蓝藻水华,辅助提取新生蓝藻水华的覆盖区.在新生蓝藻水华的覆盖区内,藻蓝素的定测估算已经失去实际意义,没有必要讨论估测的精度高低.在新生蓝藻水华的覆盖区外,藻蓝素的遥感估测精度取决于藻蓝素浓度的高低以及藻蓝素与叶绿素的定量关系,即当藻蓝素浓度〈35μg/L时,模型的平均相对估测误差约为134%;藻蓝素浓度〉35μg/L时,平均相对估测误差降至31%;但对于那些藻蓝素的浓度〉35μg/L,且藻蓝素浓度与叶绿素a浓度的比值〈8的湖区而言,藻蓝素浓度模型的相对估测误差约为29%.

英文摘要：

On the basis of the spectral characteristics of cynobacteria, a robust semi-empirical model was developed to estimate phycocyanin concentration （Cpc） using MODIS imagery with a spatial resolution of 250m. The floating neonatal cyanobacteria could be identified well and truly by the estimation model of Cpc. So it was easy to retrieve the coverage of floating neonatal cyanobacteria from MODIS imagery by the model, where there was not any actual meaning to quantitatively estimate Cpc and it was certainly not necessary to pay attention to its estimation precision. In the outside of the newly cyanobacterial blooms-covered area, the Cpc estimation precision was highly determined by phycocyanin pigment content and the ratio of Cpc to chlorophyll-a concentration （Chl-a）. When Cpc was less than 35μg/L, the relative estimation error of the model was about 134% on average; and when it was more than 35μg/L, the relative error was reduced to 31%. However, when Cpc was more than 35μg/L and the ratio of Cpc to Chl-a was less than 8, the relative error was only up to 29% on average.