Water clarity changes in Lake Taihu over 36 years based on Landsat TM and OLI observations
Water clarity changes in Lake Taihu over 36 years based on Landsat TM and OLI observations
复制标题
基于Landsat TM和OLI观测的太湖水体透明度36年变化
DOI：
10.1016/j.jag.2021.102457
复制
发表时间：
2021
影响因子：
7.5
通讯作者：
Bing Zhang
中科院分区：
文献类型：
--
作者：
Ziyao Yin;Junsheng Li;Yao Liu;Ya Xie;Junsheng Li;Shenglei Wang;Xiao Sun;Bing Zhang
·1984年至2019年，利用Landsat对太湖的Secchi圆盘深度（Z sd）进行了跟踪。·建立了基于准解析算法的TM Z sd估计模型。太湖Z sd变化的转折点出现在1998年前后。·Z sd的增加主要是由于1984年至1998年风速的下降。·藻类水华的增加可能是1998年至2019年Z sd下降的原因。20世纪80年代改革开放以来，随着我国经济社会的快速发展，地处长江三角洲最大城市群中部的太湖水质发生了巨大变化。在这里，1984年和2013年发射的Landsat-5 TM和Landsat-8 OLI卫星用于跟踪Secchi盘深度（Z sd）的36年变化，这是一个关键的水质和湖泊健康指标。首先，利用太湖12个航次299个采样点的实测Z sd和TM/OLI波段模拟遥感反射率（Rrs），对基于准解析算法（QAA）的Z sd估算模型和常用经验模型进行了标定。然后，利用69对与TM影像同步的Zsd实测数据和40对与OLI影像同步的Zsd实测数据对Zsd估算模型进行了验证。对于基于QA的方法，这已成功地应用于OLI数据，我们开发了一种方法，用于建模的R RS在TM中缺失的443 nm波段使用虚拟波段估计算法，这使得应用程序的方法TM数据。通过与其他经校准的经验模型进行比较，基于QAA的模型对于OLI和TM的准确性被确定为高，平均无偏相对误差<24%。其次，利用QAA模型反演了1984 - 2019年太湖Z sd的时空变化特征，并分析了其驱动因素。结果表明，从1984年到1998年，这主要是由风速的减小引起的Z sd的增加。Z sd的变化在1998年左右出现转折点，从1999年到2019年呈下降趋势，这主要是由于与低平均风速和气温升高相关的蓝藻水华频率和面积增加。该研究首次分析了太湖近36年来Z sd的时空变化特征，其结果和驱动力分析可为当地水资源保护提供参考。
• Secchi disk depth ( Z sd ) in Lake Taihu was tracked with Landsat from 1984 to 2019. • The quasi-analytical algorithm-based Z sd estimation model for TM was developed. • The turning point for Z sd changes in Lake Taihu occurred around 1998. • The increased Z sd was mainly due to the decreased wind speed from 1984 to 1998. • The increased algal bloom may be responsible for the Z sd decline from 1998 to 2019. The water quality in Lake Taihu, which is located in the middle of the largest group of cities on the Yangtze River delta of China, has changed greatly with rapid economic and social development since the reform and opening up of China in the 1980s. Here, Landsat-5 TM and Landsat-8 OLI satellites launched in 1984 and 2013 were used to track 36-year changes in Secchi disk depth ( Z sd ), which is a key water quality and lake health indicator. First, the quasi-analytical algorithm (QAA)-based Z sd estimation model and commonly used empirical models were calibrated using in situ measured Z sd and TM/OLI band-simulated remote sensing reflectance ( R rs ) obtained from 299 sampling points in 12 field cruises in Lake Taihu. Afterward, another 69 pairs of in situ measured Z sd data synchronized with TM images and 40 pairs of in situ measured Z sd data synchronized with OLI images were used to validate the calibrated Z sd estimation models. For the QAA-based method, which has been successfully applied to OLI data, we developed an approach for modeling the R rs in the missing 443-nm band in TM using a virtual-band estimator algorithm; this enables the application of the method to TM data. By comparing with other calibrated empirical models, the accuracy of the QAA-based model was determined to be high for both OLI and TM, with an average unbiased relative error of < 24%. Second, we employed the QAA-based model to retrieve the spatial and temporal Z sd variation characteristics of Lake Taihu from 1984 to 2019, and analyzed the driving factors. The results suggest an increase in Z sd from 1984 to 1998, which was mainly caused by a decrease in wind speed. There was a turning point for Z sd changes around 1998, with a downtrend from 1999 to 2019, which was primarily attributed to increase in the frequency and area of cyanobacterial blooms associated with the low average wind speed and increasing air temperatures. This study provides the first analysis of the spatial and temporal characteristics of Z sd in Lake Taihu over the past 36 years, and the Z sd results and driving force analyses could facilitate local water resource conservation.