Chlorophyll And Primary Productivity Estimation Based On Remote Sensing In Yellow Sea

Color remote sensing data has a large horizontal scale and near-synchronizationinstantaneous characteristics, it can monitor the regional distribution and dynamic changes of theocean color elements effetively. Inversed chlorophyll concentration using satellite data is aneffective way to assess the pollution of the marine environment, especially to predict and detectthe red tide, and so on. Using ocean remote sensing data to estimate the ocean carbon storage ishelpful to understand the marine biogeochemical carbon cycle and then the global carbon cycleand climate change, which is significant to study marine ecology and global carbon cycle.In this study, the mathematical methods including statistical analysis and neural network areused to establish the chlorophyll concentration inversion model for the Yellow sea case II watersusing MODIS images and measured spectral data. The results show that the nonlinear model isslightly better than the linear model; neural network model is superior to the statistical model;the best method for chlorophyll concentration estimation is the Erf-BP neural network model thatis an improved algorithm of traditional back propagation neural network, this work provides atheoretical basis for the chlorophyll concentration estimation and data analysis in-depth.Remote sensing model for chlorophyll concentration estiamtion is applied to themulti-temporal remote sensing data to gets the different temporal chlorophyll concentrationdistribution in order to analyse the spatial and temporal variation of the Yellow sea chlorophyllconcentration. The statistical results derived from the each month chlorophyll concentrationmaps of2010year show that the minimum, maximum and mean chlorophyll concentration ofeach month is0.320mg/m~3,23.696mg/m~3and3.371mg/m~3respectively. The chlorophyllconcentration is lower in the center areas than that in the coastal areas, and is lower in thenorthern yellow sea than that in the southern areas. This result is consistent with the previousstudies. The spatial and temporal variations of the chlorophyll in different periods were discussedby visual interpretation. It is found that the chlorophyll concentration level of the whole studyarea is higher in spring and early summer due the nutrient control and has a slightly reduce insummer and an increasing trend in autumn; in winter, because of the effect of the temperature,the chlorophyll concentration is lower than other seasons.The chlorophyll concentration isdivided into nine levels according to the statistical characteristics of the chlorophyllconcentration maps, and its temporal and spatial variation is analysed using the gravity centermethod. The gravity center positions of the multi-temporal chlorophyll concentration on eachlevel were calculated and ploted to study the chlorophyll temporal and spatial variationcharacteristics in study area.Combined with the features of the study area and characteristics of the data, the VGPM(Vertically Generalized Production Model, VGPM) model is selected to estimate the ocean primary productivity, MODIS temperature data, MODIS Kd490(diffuse attenuation coefficientsof down-welling irradiance at490nm) data, SeaWIFS PAR (Photosynthetically Active Radiation,PAR) data and inversed chlorophyll concentration data were used as input parameters for theVGPM model to calculate the primary productivity. And then its seasonal variationcharacteristics were analysed.The results show that the seasonal variation of primary productivity in the yellow sea isobvious with a bimodal characteristics, which present a trend of spring> autumn> summer>winter. Contrasting the primary productivity in February2010, April, June, August, October andDecember, the maximum value is in June with the value of1342.802mgC/m~2d, followed byApril with the value of1157.993mgC/m~2d, October with the value of1092.379mgC/m~2d, theminimum primary productivity presents in winter with the value of873.609mgC/m~2d and786.622mgC/m~2d in December and February, respectively.In addition, the spatial distribution characteristics of the water-leaving radiance and thetransmission coefficient of the upward radiance just below the surface, as well as effects of roughsurface, incident light and observation geometry on these distribution characteristics werediscussed. The results shows that downward and upward irradiances just below the surfacedecreased by the increases of solar zenith angle, with a linear relations to the cosine value ofsolar zenith angle. Spatial distribution characteristic of the water-leaving radiance is obvious andhas nothing to do with wind-driven rough surface. A simple cubic equation was developed hereto fit the changes of water-leaving radiance with viewing zenith angle. On the flat surface, thetransmission coefficient of the upward radiance just below the surface does not change with theviewing azimuth angle, however, reduces by the viewing zenith angle increasing and is linearto its square tangent value. On the rough surface, the transmission coefficient of the upwardradiance just below the surface changes little with the viewing azimuth angle but obviously withthe viewing zenith angle. The relationship also exists between transmission coefficient and thesquare tangent value of viewing zenith angle.Choosing suitable mathematical methods, it is better to retrieve chlorophyll concentrationand other ocean color information of Case II waters based on remote sensing skills. Combinedwith the characteristics of large spatial scales and long temporal scales of remote sensing data, itcan achieve the temporal and spatial variation of ocean color elements in study area. The VGPMmodel can be used to estimate the primary productivity, using the specific chlorophyllconcentration retrieval algorithm based on a specific study area to reduce the estimation error ofthe chlorophyll concentration. Light attenuation coefficient was used to calculate the euphoticdepth, which can some extent reduce the estimation error of the primary productivity and avoidthe limitation of applicability of VGPM model in Case II waters. These studies provide a reference for carrying out the marine optical survey and marine biogeochemical investigationand study by remote sensing tools.