Detecting Spatiotemporal Features of Phosphorus Concentrations Using MODIS Images: A Case Study of Hongze Lake, China

Abstract

Spatiotemporal tendency of total phosphorus (TP) concentrations in Hongze Lake were estimated using multiple Moderate - Resolution Imaging Spectroradiometer/Aqua images (MODIS) from 2003 to 2017. Considering the non-significant spectral features of phosphorus, this study was designed to primarily link TP with sensitive water quality parameters. Suspended particulate matter (SPM) was identified because SPM contents are sensitive to TP in Hongze Lake and also have notable response wavelengths including red, blue, and green bands. Therefore, a TP estimation model was developed based on the band integration methods suitable for lake SPM. The results showed that: 1) the exponential model of 645 nm and 555 nm is the most suitable model for TP estimation in Hongze Lake, in which the R2 value for calculation and validation can both reach 0.7; 2) the best TP assessment accuracy always occurred when the SPM values were of a moderate level. The threshold values can be approximately designated as 40 ~ 70 mg/L; 3) the TP concentrations of the entire lake have apparently increased since 2012, and achieved a peak in 2014 because of substantial dredging activities; and 4) the center of the lake showed the highest TP concentrations, while the two bays in the northern and western lake showed the lowest TP concentra-tions. These spatial variations can be attributed to the integration of several factors including meteorological phenomena, land use change, and human disturbance such as dredging activities. This study further indicates that the Rayleigh-corrected reflectance of red band and green band derived from MODIS can be proposed as the suitable indicators for TP monitoring, especially in SPM-dominated lakes. In addition, more ideal modeling effect can be expected when lake SPM contents kept in a medium level (e.g. 40 ~ 70 mg/L within Hongze Lake).

Keywords: phosphorus, remote sensing, spatial and temporal tendency, lake suspended particulate matter, Hongze Lake basin