Research On Diffuse Attenuation Coefficient Estimation Of Taihu Lake Based On The GOCI Images

Diffuse attenuation coefficient (Kd) is an important parameter in water body optics, and also an important affecting factor of the water ecological system. Remote sensing estimation of water diffuse attenuation coefficient can reveal the change of the regional water light field, and it is of great significance for the water quality and water environment monitor. Based on the in-situ measured data obtained in Taihu Lake in different seasons, the apparent optical properties (AOPs) and the inherent optical properties (IOPs) were analyzed. Meanwhile, the absorption coefficient of different water components and back-scattering coefficient were considered as impact factors to figure out the contribution rate. Based on these researches, empirical model and semi-analysis model were constructed with GOCI data to estimate the diffuse attenuation coefficient at 490 nm. The optimal model was used to GOCI images to estimate the diffuse attenuation coefficient at 490 nm of Taihu Lake. Then, the spatial and temporal variation characteristics of diffuse attenuation coefficient were analyzed based on the inversion images. The research obtains the following conclusions mainly:(1) The impact factors of diffuse attenuation coefficient of Taihu LakeThe main impact factor of diffuse attenuation coefficient at 490 nm is the non-algal particles absorption coefficient, with the contribution rate much higher than other factors; The second impact factor is the pigment particles absorbtion coefficient. The contribution of back-scattering coefficient and CDOM absorption coefficient to diffuse attenuation coefficient are relatively small. While, the contribution of the back-scattering coefficient is generally higher than the contribution of CDOM absorption coefficient.(2) Estimation models for diffuse attenuation coefficient by GOCI dataGOCI band 7 can reflect the feature of the suspended matter well. Thus, it is a good way to improve the accuracy by introducing band 7 into the diffuse attenuation coefficient estimation model. The combination model with GOCI bands 4,5 and 7, namely the near infrared band, the red band and the green band, has the highest precision and fitting result with R2 of 0.737. The verification result shows that the average relative error and root mean square error are 28.6% and 1.495, respectively. This model has a low sensitivity to the change of the diffuse attenuation coefficient values and a high universality. The estimation accuracy of semi-analysis model relies heavily on the estimation precision of inherent optical properties. The semi-analysis model built with the measured inherent optical properties has a higher accuracy, while the model built with the estimated inherent optical properties by QAA and Doron method has a lower accuracy.The synchronous in-situ data were used to validate the diffuse attenuation coefficient, which was estimated by applying the optimal empirical model to GOCI images of Taihu Lake. The result shows that the model has a high estimation precision, and is more suitable for GOCI image data to estimate diffuse attenuation coefficient of Taihu Lake with good universality.(3) Spatial and temporal variation characteristics of the diffuse attenuation coefficient of Taihu LakeThe diffuse attenuation coefficient of Taihu Lake was estimated from 8:00 am to 3:00 pm using GOCI images on May 14,2013, August 5,2014 and October 24,2014. The research shows that the range of diffuse attenuation coefficient of Taihu Lake distributes between 0 to 15 m-1, and that can be divided into three levels generally, i.e., the range of the low value between 0～4 m-1, the medium value between 4～8 m-1, and the high value between 8～15 m-1. While, occasionally, some values are higher than 15 m-1.The diffuse attenuation coefficient of Taihu Lake distributes with notable spatial and temporal difference. At the view of spatial distribution, the diffuse attenuation coefficient values get higher from the east to the west of Taihu Lake, showing a tendency that the west values are higher than the east values. From the perspective of time, the diurnal variation of the diffuse attenuation coefficient of Taihu Lake in different day presents a similar regularity, i.e., the diffuse attenuation coefficient is relatively high in the morning and then shows a tendency of gradually reduce with the passage of time.