Quantitative Inversion And Evaluation Of Coastal Mariculture Quality Parameters In Guangdong Province

In the past 20 years,the total output of mariculture in China has ranked first in the world.However,in the process of the continuous development of fishery production,due to the unreasonable feeding mode in the process of aquaculture,the coastal waters are facing the problems of water quality and eutrophication.In the past,fixed point sampling was used to monitor seawater quality at a time.It is time-consuming,labor-intensive and expensive,and it is difficult to reflect the water quality of the whole bay.In recent years,remote sensing technology has the characteristics of strong timeliness,wide range of monitoring and high monitoring precision.Remote sensing technology is gradually applied in water quality monitoring,which provides technical support for the study of dynamic change of water quality.The coastline in Guangdong is very long,and its cultivation technology is mature,and its sea farming area has reached the top of the country.Hailing Island is located in the western Guangdong fishing port area,one of the three big fishing ports in Guangdong.The variety of marine aquaculture is many and typical.Therefore,this study takes the near shore waters near Hailing Island as the research object.In May 2017,August 2017 and September 2017,by monitoring the water quality parameters such as chlorophyll A?Chl-a?,total phosphorus?TP?,ammonia nitrogen?NH₃-N?and chemical oxygen demand?COD?in pond culture area and cage culture area,then based on the sensitive bands of GF-1 remote sensing data,combined with BP neural network and regression analysis,a satellite remote sensing inversion model for mariculture water quality parameters was established.The temporal and spatial distribution of water quality parameters was analyzed by inversion results.The comprehensive pollution index P and the comprehensive nutritional status index TSI were used to evaluate the eutrophication of water bodies in pond culture and cage culture,and the time and space distribution of each index were analyzed.The main results are as follows:?1?The water quality quantitative inversion model was established by BP neural network or regression analysis in combination with the measured data of water quality and remote sensing image data at the same time in the aquaculture area.This study selects different modeling methods based on the characteristics of chlorophyll a,total phosphorus,ammonia nitrogen and COD,so as to make the initial precision of the whole quantitative model meet the requirements of inversion.?2?Using the GF-2 high spatial resolution remote sensing image as data source,we extract the fish ponds in pond aquaculture area by using the object extraction method.The object extraction process is based on the combination of different objects'spectral,texture and spatial characteristics rules,and the classification is effective.A total of 9401 fish ponds were taken from the research area near the Hailing embankment in Hailing Island.The total area was5636.26 hm².?3?The geotiffread function of MATLAB is used to read the remote sensing image,and the remote sensing image is converted into a two-dimensional matrix,and the trained BP neural network inversion model is used to retrieve the water quality.Then,using the longitude and latitude of the measured points to extract the inversion parameters of the water quality parameters of the inversion image,the relative error of each water quality parameter is compared.The inversion model of several water quality parameters is of good precision,which can be used in the quantitative inversion of water quality parameters in the aquaculture area of the study area,and provide a feasible and effective technical means for the water quality monitoring of mariculture.?4?Combined with the image of water quality parameters,the changes of water quality parameters in pond culture water and cage culture water in the study area were analyzed spatio-temporal.From May to August,the average content of chlorophyll,total phosphorus,COD,and ammonia nitrogen has increased.From August to September,the average content of these water quality parameters per unit area decreased.The experimental results accord with the basic law of culture,but in pond culture,the average content per unit area of each parameter is more than the standard of surface water quality stipulated by the state.Therefore,it needs to be paid more attention.?5?ArcGIS software was used to evaluate the eutrophication of water body in mariculture area with Comprehensive pollution index?P?and theTurbine Supervisory Instrumentation?TSI?.The distribution map of comprehensive pollution and the distribution map of comprehensive nutritional status in May,August and September were obtained.The spatio-temporal analysis of the comprehensive pollution and comprehensive nutritional status of pond aquaculture water and cage aquaculture water in the study area was carried out,and the trend of the trend of early rise and subsequent decline,was also consistent with the basic law of culture.The nutrient status index and the comprehensive pollution index of most of the pond aquaculture are very high,so the water quality needs to be concerned and improved.