Marsh Vegetation-hydrological Change Monitoring Based On Time Series Remote Sensing And Its Spatiotemporal Coupling Relationship

The interaction between marsh vegetation and hydrology has become one of the hot topics in ecosystem research.This paper takes the Honghe National Nature Reserve as the research area,and uses the Landsat series and Sentinel-1 SAR time series images from 1985 to 2021 as the data source.Based on Google Earth Engine,a high-precision spatio-temporal monitoring model of marsh vegetation and hydrology remote sensing is constructed using Land Trendr and CCDC algorithms.The Random Forest algorithm was introduced to classify the monitoring results of CCDC time series changes,and to explore the temporal-spatial dynamic evolution of marsh vegetation and hydrology in the past 36 years and the differences in phenological characteristics under the natural succession of marsh vegetation.Integrate Otsu algorithm and SDWI index to extract wetland water time series information.The CCDC-PLSR interpretation model of phenological characteristics and hydrology-meteorological factors of different marsh vegetation was constructed by partial least squares regression(PLSR)method,and the temporal-spatial response mechanism of phenological characteristics of different marsh vegetation to hydrology-meteorological factors was revealed.At the same time,an innovative coupling coordination model was constructed to quantitatively evaluate the relationship between interannual loss and restoration of vegetation and hydrology.The coupling relationship between marsh vegetation and hydrology is analyzed by using transfer entropy,and the driving factors that cause long-term changes in marsh vegetation are determined.Finally,combined with CCDC and multivariate stepwise regression(MLSR)algorithm,a response model(CCDC-MLSR)of loss and restoration of marsh vegetation and hydrology-meteorological factors was constructed to explore the mechanism of hydrology-meteorological factors on the annual loss and restoration of marsh vegetation.The research conclusions are as follows:(1)Using the Land Trendr algorithm to monitor the dynamic changes of marsh vegetation and hydrology,the accuracy rates of vegetation loss and restoration are 0.83 and 0.85,respectively,which proves that this method can effectively monitor the temporal and spatial evolution of marsh vegetation and hydrology.From 1985 to 2019,the marsh vegetation in the study area generally showed a recovery trend,and the recovery duration was ≥20 years.The disturbed vegetation is fragmented and marginalized,and the disturbance duration is less than 20 years.(2)The trend lines of the mean values of the differences between the two change monitoring results based on the CCDC and Landtrendr algorithms were distributedwithin the 95% concordance limit.It shows that the use of CCDC algorithm to monitor the phenological dynamic evolution of marsh vegetation and the annual loss and recovery has high reliability.The shallow-water marsh vegetation and the deep-water marsh vegetation suffered high frequency and high-amplitude losses from April to October,resulting in a significantly larger area of loss during the year than the area of recovery.(3)The phenological bimodal trajectory reflects the two growth processes of different marsh vegetation throughout the year.And the amplitude values of deep-water marsh vegetation and shallow-water marsh vegetation are greater than 500,indicating that aquatic vegetation has undergone high-intensity dynamic growth and has high vegetation growth throughout the year.The change amplitudes of shrubs and trees were lower,indicating that the development process of terrestrial vegetation was slower.(4)From 1985 to 2019,both marsh vegetation and hydrology were in a state of imbalance,but the loss of vegetation and hydrological changes showed a high level of coupling.The CCDC-MLSR model and transfer entropy demonstrate that alternations of hydrology and climate affect the loss and restoration of vegetation in deep and shallow marsh.It is demonstrated that hydrology-meteorological factors are the driving factors leading to the loss and restoration of marsh vegetation in each growing period.