Estimation Model Of Denitrification In Shallow Eutrophic Water Based On Hyperspectral Remote Sensing Inversion

With the development of society and industrialization,excessive input of nitrogen fertilizer has caused a series of environmental problems,including eutrophication of water body,acidification of soil and air pollution.Denitrification is the final step of the activated nitrogen cycle,it converts fixed nitrogen and artificially activated nitrogen in the ecosystem into nitrogen and returns it to the atmospheric nitrogen pool.In wetland ecosystem,denitrification can remove nitrate nitrogen in water,regulate water quality and prevent water eutrophication.The main product of denitrification is N₂.However,the background value of N₂ in the atmosphere is very high,up to 79%.Therefore,the determination of denitrification in such a high background environment requires a very high precision of the determination method,which is difficult to be achieved by ordinary methods.Previous studies have shown that in shallow eutrophic water,there is a correlation between Chl.a and water nitrogen concentration,and there is a significant linear correlation between CDOM and dissolved organic carbon（DOC）,while water nitrogen concentration,DOC and T_w are the main influencing factors of denitrification rate.Therefore,the relationship between the remote sensing satellite band and denitrification is expected to be established through the three remote sensing factors of Chl.a,CDOM and T_w,so as to realize the direct inversion of water denitrification rate by hyperspectral remote sensing.Based on the above assumptions,we took Taihu Lake as an example,a typical shallow eutrophic water body,and measured denitrification rate（Den）,Chl.a,CDOM and T_w through regular monitoring experiments with 48 sampling sites and 144 samples,and simultaneously observed the physicochemical properties of the water body and sediment,the quantitative relationship between denitrification rate and Chl.a,CDOM and T_w was determined.A mathematical model of denitrification rate was built with Chl.a,CDOM and T_w remote sensing variables,and the model was verified.In addition,the remote sensing data of Chl.a,CDOM and T_w of Taihu Lake from 2011 to 2020 were obtained from the National Data Center for Earth System Science.Based on the denitrification rate model established by us,the spatiotemporal variability and denitrification and denitrification capacity of the entire Taihu Lake basin were estimated.In this study,a method of denitrification inversion for eutrophic water by satellite remote sensing is proposed for the first time.The main principle of this study is that the chlorophyll a（Chl.a）,colored dissolved organic matter（CDOM）and temperature（T_w）of lake water can be retrieved by hyperspectral remote sensing technology.Early research has shown that in the eutrophication of shallow water,chlorophyll a and water body has a good linear correlation between nitrogen concentration,colored dissolved organic matter and dissolved carbon（DOC）has a good linear relationship,and water soluble carbon nitrogen concentration,concentration,temperature is the main influence factors of denitrification,accordingly,can build a remote sensing satellite mathematical relationship of wave and denitrification.Based on this,Taihu Lake,a typical shallow eutrophic water body,was taken as an example to measure denitrification rates（Den）,Chl.a,CDOM and T_w through 8 times of sampling,48 sampling sites and 144 samples in regular monitoring experiments.The physical and chemical properties of the water and sediments were simultaneously observed to determine the quantitative relationship between denitrification rate and Chl.a,CDOM and T_w.Three remote sensing variables Chl.a,CDOM and T_w were used to build the mathematical model of denitrification rate,and the model was verified and the uncertainty was analyzed.In addition,remote sensing data of Chl.a,CDOM and T_w of Taihu Lake from 2011 to 2020 were obtained from National Earth System Science Data Center,and the denitrification rate model established by us was used to estimate the temporal and spatial variability and denitrification and denitrification capacity of eutrophic water bodies in the entire Taihu Lake Basin.The main findings of this study are as follows:（1）The denitrification rate of Taihu Lake presents strong spatiotemporal heterogeneity.The denitrification rate in summer（June,July and August）was higher than that in winter（November,December and January）,and the denitrification rate in northwes was higher than that in eastern.The denitrification rates of Taihu Lake ranged from 21.32±11.18 to130.72±10.55μmol N₂-N·m^-2·h^-1,and the average denitrification rate of Taihu Lake was71.06μmol N₂-N·m^-2·h^-1.The maximum denitrification rate appeared in Meiliang Lake area on July 25,2020,and the minimum denitrification rate appeared in the lake center area on November 16,2020.（2）Based on the existing research understanding,an estimation model of denitrification rate and remote sensing factor variables（Chl.a,CDOM and T_w）in Taihu Lake was constructed,and the measured data（36 groups）were used for parameterization.The denitrification rate and three remote sensing variables（Chl.a,CDOM and T_w）were parameterized as follows:Den=-0.001*Chl.a²+0.431*Chl.a+9.87*CDOM+0.926*T_w+5.244;Parameterized with two remote sensing variables（Chl.a and T_w）:Den=-0.002*（Chl.a）²+0.789*（Chl.a）+0.385*（T_w）+41.493,the model can explain 70%and 57%of the variation of denitrification rate,respectively.Further application of observation data（another 12 groups）to verify the model has a good prediction effect.（3）Based on the hyperspectral remote sensing data of Chl.a and T_w in Taihu Lake,as well as the multivariate nonlinear estimation model of denitrification rate and remote sensing variables established in this study,the monthly average denitrification rate in 2019 and the annual average denitrification rate from 2011 to 2020 were calculated.In terms of monthly time,the denitrification rate of Taihu Lake mainly concentrated in May to July.In terms of annual time,the denitrification rate of Taihu Lake fluctuated little in the range of 55～60μmol N₂-N·m^-2·h^-1,except that the denitrification rate decreased significantly in 2014 and 2015.In terms of spatial distribution,denitrification rates in the northwest and center of Taihu Lake are generally higher than those in other lakes.From 2011 to 2020,the annual nitrogen consumption of Taihu Lake is in the range of 2.7～3.4×10⁴ t,and the denitrification rate is in the range of 50%～63%,which is basically consistent with the estimated results of previous application of income and expenditure method.This study is the first time to estimate the denitrification rate of eutrophic water by satellite remote sensing method.It is of great significance to establish the rule of nitrogen reduction in eutrophic water body,to set scientific environmental capacity target,and to reduce the environmental load of nitrogen.