Spatiotemporal Fusion Of LAI Data And Its Impact On Noah-MP Soil Temperature And Humidity Simulations

The Leaf Area Index(LAI)is one of the important vegetation indicators to explore global climate change,as well as one of the important parameters to characterize the dynamic change of vegetation canopy and land surface model.However,due to the influence of multiple factors such as atmospheric environment and single data source,problems such as low quality of LAI data,discontinuity of spatiotemporal data,and lack of data appear,which affect the application of LAI in various models.In order to improve the data quality and spatial resolution of LAI,this study proposes a method of LAI reconstruction based on spatio-temporal fusion.Firstly,based on land cover type and quality control file,MODIS LAI and MuSyQLAI are replace with low quality by pixel quality analysis method.Then,based on the STNLFFM model(A spatio-temporal integrated temperature fusion model),S-G filtering,and comprehensive filtering of annual series outlier detection,spatio-temporal fusion and time series reconstruction of LAI data are performed.We apply the fusion LAI data to the Noah-MP model under the static vegetation,and design comparative experiment to explore the effect of LAI changes on simulation of soil temperature and moisture.The main conclusions of the study are as follows:(1)The results show that the fusion LAI have high consistency and similarity with high-resolution LAI.The correlation coefficients are both higher than 0.7,and the root mean square error is lower than 1.2,the average structural similarity is 0.81.The evaluation by observation data shows that the correlation R is 0.75 and the root mean square error is 0.92.Cross-validation analysis from time-space continuity and time-space consistency,the fusion LAI has a significant increase in the proportion of spatial continuity and inversion,and the loss frequency of temporal continuity has decreased.The fusion LAI has reasonable fluctuation amplitude and frequency,have consistent changes with other LAI products in seasons and different land cover types.The spatio-temporal resolution is improved to 5d,500m.Meantime,the smoothness of time series curve is better,with less noise,and the overall quality of LAI data is better.(2)The temporal and spatial variation characteristics of LAI generated by integration show that the overall LAI timing sequence curve from 2010 to 2015 was relatively smooth,with prominent features of peak value and valley value,without obvious noise and inflection point.LAI values in the growing season of vegetation were higher from April to October,while those in the non-growing season gradually decreased,and the annual variation was significantly different.The inter-annual variation of LAI values of different land cover types showed an upward trend.The spatial variation trend of LAI gradually decreases from southeast to northwest,and the order of seasonal variation is:Summer>autumn>spring>winter.The largest LAI value in growing season is mainly distributed in the eastern coastal areas,and the LAI value in non growing season declines obviously.The LAI value in most regions of China is in the range of medium variation and stability,and its variation trend is mainly upward.(3)To analyze the influence of fusion LAI on surface soil temperature and humidity,multiple sets of comparative experiments(model original,original MOIDS LAI,reconstructed MODIS LAI,Mu Sy Q LAI,fusion LAI)were designed to simulate the soil temperature and humidity in China in 2014.The simulation results of soil temperature show that the spatial distribution of soil temperature in China is roughly in three steps,decreasing from southeast to northwest.The soil temperature of the fusion LAI is 0.09?higher than the original model,and the root mean square error is reduced.The lower the latitude is,the better the effect.On the time scale,the fusion LAI simulation is better than the original simulation of the model in the growing season.The root mean square error is 0.083?lower than the original model,and the average absolute deviation is reduced by 0.5?.But in the non-growing season,there is little difference between the model and the original simulation results.The soil moisture simulation results show that the fusion LAI simulation results improves the original underestimation of the model,and the spatial distribution of soil moisture gradually increases from northwest to southeast.Compared with the original results of the model,most of the fusion LAI results are in the state of increased moisture.The moisture decreases in the northwest,southwest,northeast,and eastern coastal areas,and the proportion of sites with a root mean square error of less than0.1 m~3/m~3 increases,while the proportion of sites with root mean square error more than 0.12m~3/m~3 decreases.The simulation effect of soil moisture has been improved.In time scale,the soil moisture simulated by LAI is closer to the measured data,and its root mean square error,deviation,and correlation coefficient are better than other data simulation results.