水体叶绿素含量的遥感定量模型

QUANTITATIVE ANALYSIS ON CHLOROPHYLL-A CONCENTRATION IN TAIHU LAKE USING THEMATIC MAPPER DATA

利用1992年7月25日陆地卫星TM数据与该日（准）同步湖面多点观测资料,对太湖北部水体叶绿素含量与不同波段遥感值的关系进行了定量分析,发现TM3与Chl-a含量有负幂次关系,TM4、TM7则分别与Chl-a含量有较好的正、负线性相关性;据此选择TM3、TM4和TM7等多个敏感波段值作为自变量,通过对它们的线性和非线性变换,进行叶绿素定量指标的组合和筛选试验;并依据太阳入射辐射的变化提出对遥感定量指标进行时相和纬度订正的通用方法;在此基础上,建立了叶绿素含量的遥感定量模型。利用这一模型计算得到的1992年10月29日太湖梅粱湾水域多点叶绿素含量很好地拟合了实测结果。统计分析和验证结果表明,本文选定的水体叶绿素含量遥感定量指标的相关性好、显著水平高,而且进行了时相和纬度订正,当天气条件相似时,该模型可应用于其它湖区和其他时相的遥感定量分析。

The LANDSAT Thematic Mappper (TM) data on July 25, 1992 were used together with concurrent measurement of Chlorophyll-a concentration in the northern part of Taihu Lake to relate the TM data quantitatively to Chl-a. Firstly, the statistical relationships be-tween TM data on each band at sample sites and measured Chl-a were analyzed. It is shown that the third band (TM3) data power-law decays with the increase of Chl-a, while the forth band (TM4) and the seventh band (TM7) separately show the positive and negative linear correlations with the Chl-a concentration. These sensitive bands (TM3, TM4 and TM7) were selected as independent variables, whose optimum combinations were examined through their linearly and non-linearly alterations in order to obtain a new quantitative index for Chl-a esti-mation. Meanwhile, a method of revising this index to remove the influences of ground irradi-ation changes is presented in this paper. It is found that there is a close linear interrelation be-tween the revised quantitative index and Chl-a concentration. Based on this interrelation, the final Chl-a estimation model was developed in this paper. The Chl-a values measured at several sample sites in Meiliang Bay on Oct. 29, 1992 were coincident with the calculated ones using this model. The statistical analysis and checking results show that this Chl-a estimation model can be applied to images of different dates and areas on similar weather condition because through irradiation amendment the factors of time and latitude have little interference with this model.