Urban river and lake systems show important ecological function, and eutrophication frequently occurs and results from human activities due to the limited self-regulating ability. Since nitrate (NO3-) is one of the key factors causing water eutrophication, its rapid qualification plays critical role in the eutrophication control and management. In this study, water samples were collected from typical water bodies from Nanjing in different seasons, and Fourier transform infrared attenuated total reflectance spectroscopy (FTIR-ATR) was employed for the quantitative determination of NO3- coupled with algorithms of deconvolution and partial least squares regression (PLSR). Results indicated that the typical absorption band of NO3- at 1500-1200 cm-1 was observed and the intensity of the band around 1360 cm-1 was positively correlated with the concentration of NO3- through spectra deconvolution. PLSR models were established based on the deconvolution spectra, which were excellent with the correlation coefficients (R2) of more than 0.8886 and the ratio of prediction to deviation (RPD) of more than 2.76; it was found that the carbonate in water might impact the prediction due to its absorption around 1450 cm-1, but the prediction model performed well in condition that the carbonate content in a low level with less than 10 mg L-1. Significant temporal and spatial variations of NO3- were observed in the typical water bodies, and the Qinhuai River having the highest NO3- content, which mainly was influenced by human activities, and the impact of water pH and temperature were not significantly observed. Therefore, FTIR-ATR combined with deconvolution and PLSR, allowed a rapid determination of NO3- in urban water bodies, providing an alternative option for the monitoring of nitrate in natural water body, which will benefit the prevention and control of eutrophication.
Keywords: Deconvolution; Eutrophication; FTIR-ATR; Natural water bodies; Nitrate; PLSR model.
Copyright © 2022 Elsevier Ltd. All rights reserved.