| As of 2022,the green tide disaster in the Yellow Sea has been continuously occurring for 16 years.During the large-scale outbreak of the Yellow Sea green tide,the affected area can reach tens of thousands of square kilometers,damaging the marine ecological environment and affecting the survival and reproduction of other marine organisms,causing significant social and economic impacts,making it the most serious ecological disaster in the Yellow Sea.The occurrence of the green tide disaster is a product of human activities changing the natural environmental conditions,with various physical,chemical,and biological factors working together to promote the outbreak of the green tide.The impact of human activities in the nearshore waters,the continuous input of terrestrial nutrients,global warming and climate events,the unique characteristics of green tide algae,the interactions between organisms and microorganisms,as well as the dynamic changes and interactions of these factors,collectively determine the characteristics and scale of green tide outbreaks.At the same time,the long-term outbreak of the green tide is also affecting and changing the marine ecological environment.When green tides outbreak,they have the characteristics of wide distribution range,long duration,and macroscopic drift.Satellite remote sensing has become an important means of dynamic monitoring of green tides due to its advantages such as large observation time span,wide observation range,and easy access to data.With the progress of information science and technology and the accumulation of large amounts of data,traditional remote sensing data analysis methods represented by image screening,manual interpretation,and image processing are insufficient in real-time monitoring and processing of green tide disasters,as well as in analyzing the long-term trend and outbreak mechanism of green tide disasters through large-scale,long-term,and multi-source data interaction.It is urgent to explore new methods for ecological environment monitoring and management from the perspective of big data.This study will establish an automatic extraction model for green tide based on the Google Earth Engine(GEE),which is a platform for massive data storage and online data analysis,so that it is possible to quickly and accurately extract and analyze large-scale and long-term green tide information.Based on the long-term changes in various ecological environmental factors,analyze the impact of environmental factors on green tide outbreaks and the impact of green tide outbreaks on the environment.Main research contents:(1)Based on the GEE platform,an adaptive maximum between-class variance(Ostu)threshold method green tide extraction model is established.This model includes the differentiation of green tide algae,prolifera,and golden tide algae,Sargassum horneri,as well as an adaptive mixed pixel decomposition module,to achieve fast and accurate green tide information extraction.Compare and evaluate the model extraction results with the single threshold method extraction results,and further analyze and discuss the stability and reliability of the model extraction long-term results.(2)Based on the results of the automatic extraction model of green tides established in this study,the spatiotemporal distribution patterns of green tides in the Yellow Sea from 2007 to 2021 were studied.(3)Investigate the long-term trends of environmental data such as sea surface temperature,sea surface salinity,precipitation,sea surface wind,sea surface current,photosynthetically active radiation,and water quality,as well as the impact of environmental factors on the outbreak and dissipation of green tide.(4)Explore the relationship between long-term outbreaks of green tide and phytoplankton using Chlorophyll-a concentration as an indicator,and analyze the impact of green tide outbreaks on the environment.Main conclusions:(1)A green tide extraction model is established based on the GEE platform combined with adaptive Ostu threshold method,which integrates remote sensing image preprocessing,cloud screening and removal,index calculation,threshold selection,and mixed pixel decomposition.This model reduces interference from human factors,has higher accuracy than a single threshold method.The long-term extraction results have certain stability and reliability,achieving fast and accurate extraction of long-term green tide information.(2)There are significant interannual differences in the spatiotemporal distribution of the Yellow Sea green tide from 2007 to 2021: in terms of scale,it has fluctuated in a wave like pattern over the past 15 years,and the amplitude of this fluctuation has increased since 2015;In terms of temporal distribution,both the initial date and outbreak date show a trend of advancement,while the dissipation date does not show a significant trend of advancement;In terms of spatial distribution,there are significant interannual differences in the range of affected sea areas over the years.The larger the distribution area in a year,the more the distribution area extends southward and eastward.(3)The sea surface temperature and precipitation in the early stage are the main factors affecting the inter-annual differences in the outbreak scale of green tides,while the high sea surface salinity and abundant precipitation in the dissipation period are sufficient conditions for prolonging the process of green tide disappearance.In the long run,the sea surface temperature,sea surface salinity,and photosynthetically active radiation are all on the rise,and the long-term changes in various environmental factors still provide conditions for the large-scale outbreak of green tides.(4)Green tides have a significant inhibitory effect on the growth of phytoplankton during the outbreak.In the early discovered areas of green tides,there is a significant correlation between Chlorophyll-a concentration in January and the scale of green tide outbreaks in the current year,which can be used for early prediction of green tide outbreaks.Innovation points:(1)A green tide extraction model based on GEE was established,and a new mixed pixel decomposition method was proposed.(2)The long-term trends of ecological factors and their long-term relationship with green tides were studied from a long-term and large-scale perspective. |
