Duan, Hongtao; Tao, Min; Loiselle, Steven Arthur; et al.
The occurrence and related risks from cyanobacterial blooms
have increased world-wide over the past 40 years. Information on the abundance
and distribution of cyanobacteria is fundamental to support risk assessment and
management activities. In the present study, an approach based on Empirical
Orthogonal Function (EOF) analysis was used to estimate the concentrations of
chlorophyll a (Chla) and the cyanobacterial biomarker pigment phycocyanin (PC)
using data from the MODerate resolution Imaging Spectroradiometer (MODIS) in
Lake Chaohu (China's fifth largest freshwater lake). The approach was developed
and tested using fourteen years (2000-2014) of MODIS images, which showed
significant spatial and temporal variability of the PC:Chla ratio, an indicator
of cyanobacterial dominance. The results had unbiased RMS uncertainties of
<60% for Chla ranging between 10 and 300 mu g/L, and unbiased RMS
uncertainties of <65% for PC between 10 and 500 mu g/L. Further analysis
showed the importance of nutrient and climate conditions for this dominance. Low
TN:TP ratios (<29:1) and elevated temperatures were found to influence the
seasonal shift of phytoplankton community. The resultant MODIS Chla and PC
products were then used for cyanobacterial risk mapping with a decision tree
classification model. The resulting Water Quality Decision Matrix (WQDM) was
designed to assist authorities in the identification of possible intake areas,
as well as specific months when higher frequency monitoring and more intense
water treatment would be required if the location of the present intake area
remained the same. Remote sensing cyanobacterial risk mapping provides a new
tool for reservoir and lake management programs.
(来源:WATER RESEARCH, 2017, 122: 455-470)
|