Determination of nitrate in seawater based on deep ultraviolet spectroscopy and optimized algorithm

In order to achieve high-precision, long-term and in-situ monitoring of nitrate concentration in seawater, a new method based on deep ultraviolet spectroscopy is presented to measure the absorbance of multi-component seawater samples at 200nm-400nm, and the measurement models are established using least square algorithm and kernel partial least square algorithm respectively. It is proved that kernel partial least square algorithm is better and the predicted concentration of nitrate is more accurately by comparing the two models. The research results show that the kernel partial least square algorithm can better extract the nonlinear relationship between the absorbance of different wavelengths and the nitrate concentration hidden in the spectrum of multi-component seawater, with better goodness of fit as well as prediction accuracy.