Research On Evapotranspiration Inversion And Influencing Factors In The Yellow River Delta Based On ESTARFM Algorithm And SEBS Mode

Affected by global warming and human activities,the wetland area of the Yellow River Delta（YRD）was shrinking and degraded,and ecological problems such as soil salinization and biodiversity degradation were prominent.Evapotranspiration（ET）is an important parameter of climate system,carbon feedback and hydrological cycle.Accurately quantifying the temporal and spatial dynamics of ET is of great significance for understanding climate change and wetland water resources management.At present,existing ET data cannot provide a high spatio-temporal resolution,which cannot meet the requirements of evapotranspiration research in the Yellow River Delta area with a high degree of surface fragmentation.Therefore,this study obtained ET datasets,based on Enhanced Spatiotemporal Adaptive Reflectance Fusion Model（ESTARFM）and Surface Energy Balance System model（SEBS）,and analyzed the spatiotemporal changes of ET and its influencing factors in the Yellow River Delta.Firstly,,this paper estimates the ET data of the Yellow River Delta with a spatial and temporal resolution of 8 days and a spatial resolution of 30 meters from 2000 to 2020 by using ESTARFM algorithm and SEBS model,based on MODIS and Landsat remote sensing data;The feasibility and accuracy of the combination of the two models were evaluated by comparing the fusion ET and Landsat ET in the prediction date and verified the reliability of the ET data by using the observation data of the flux station and the evaporative dish data of the meteorological station.Secondly,this study analyzed the spatiotemporal variation trend of ET and its influencing factors in the Yellow River Delta,based on the fused ET dataset;Finally,support vector machine was used to classify the land use in the Yellow River Delta,we analyzed the land use change from 2000 to 2020,and explored the response of ET to climate change and land use change in the Yellow River Delta in different periods with meteorological data and ET dataset.The main results are as follows:（1）The ESTARFM algorithm has good applicability in the Yellow River Delta region where the ground surface is broken,and the fusion accuracy can meet the needs of subsequent research.The higher the input image quality and the closer to the prediction date,the higher the fusion accuracy.The ESTARFM algorithm and SEBS model can accurately simulate the ET during the prediction period.The fusion ET has a high consistency with the actual Landsat ET on the prediction date（R≥0.86,RMSE≤0.37mm/day）.Compared to Scheme II（calculating ET before fusion）,Scheme I（calculating ET after fusion）has a higher accuracy.The accuracy of the ET dataset is relatively reliable.The root mean square error between the sensible heat flux of farmland pixels calculated by the model and the observed data of the Yucheng farmland ecosystem is only 57.8W/m².The correlation between the ET of water pixels and the evaporative dish data of Kenli and Hekou meteorological stations is significant（R>0.85）.（2）From 2000 to 2020,the annual average ET value in the Yellow Ri ver Delta region was 878.2mm,and increased at a rate of 3.5 mm·a^-1;The a nnual variation of ET shows a strong seasonality,with a spatial decrease fro m coastal to inland as a whole.From 2000 to 2020,the increasing degree of landscape fragmentation in the study area led to the increasingly complex sp atial pattern of ET,and the overall level of ET in the area improved.During the study period,the area where ET increased accounted for about 76%,wit h a significant increase of 23%,mainly distributed along the coast of Laizho u Bay and Bohai Bay.Temperature,precipitation,and saturated water vapor p ressure all present a significant linear positive correlation with ET,with the h ighest correlation between temperature and ET（R²=0.81）.Because warmer air can hold more water vapor before condensation,rising temperatures also incr ease the demand for water vapor and precipitation in the atmosphere,and ET will increase as the temperature increases.The trend of monthly precipitation is consistent with that of monthly ET,showing a strong seasonality.Howeve r,there is a certain lag in the impact of precipitation on ET on the daily sca le.Generally,ET will reach a small peak within 2-3 days after rainfall occur s.In space,ET is significantly negatively correlated with surface albedo and temperature vegetation drought index,and the spatial pattern of ET is closely related to land use types.The annual ET size of different land use types in the Yellow River Delta is:water body>mudflat>forest>natural vegetation>far mland>construction land.（3）The contribution of climate change to ET trends mainly comes from temperature and precipitation.The average contributions of temperature and precipitation to ET from 2000 to 2020 were 1.2 mm·a^-1and 0.6 mm·a^-1.In addition to climate change,the biggest contribution to the ET trend in the Yellow River Delta from 2000 to 2020 is the land use change from natural vegetation,mudflat to water body.Although ET increased only 7.8 mm·a^-1and 6.6 mm·a^-1from natural vegetation and mudflat to water body,due to the large area of transformation,its contribution to ET trend is far greater than that of other surface features.The trend of ET in the Yellow River Delta during the period 2000-2020 is relatively consistent with the trend of land use change’s contribution to ET.Land use change has a greater impact on ET in the Yellow River Delta than climate change,and land use change dominates the trend of ET change in the Yellow River Delta during the period 2000-2020.