Study On The Remote Sensing-Based Measurement Model Of Vegetation Resilience

Vegetation can maintain the stability of the entire ecosystem.Understanding the connotation of the vegetation resilience and quantitatively measuring the resilience can provide a theoretical basis for ecosystem management and ecological restoration.However,the existing measurement systems of the vegetation resilience lack unified monitoring indicators and calculation methods.Meanwhile,some measurement methods don't take the environmental variables into consideration and the results are subjective.Therefore,it is urgent to establish a vegetation resilience measurement system that can describe the dynamic changes of vegetation groups and the response characteristics of vegetation on disturbance.The multi-resolution,multi-temporal and multi-band of remote sensing images can meet the demands of land-surface monitoring.However,the number of observation platforms and data are increasing.Exploring multi-platform data fusion methods and establishing effective monitoring indicators are the keys to establish the vegetation resilience measurement system.Therefore,the general objective of this dissertation is to establish the monitoring indexes for vegetation ecosystem and remote sensing-based measurement model of vegetation resilience on the basis of the definition of vegetation resilience.In this dissertation,firstly,the vegetation disturbance factors and the influencing mechanism are summarized by the literature review.Then the mathematical modeling and statistical analysis methods are used to establish the remote sensing measurement model of vegetation resilience.The analysis method of resilience driving factors is also proposed.The drought monitoring index based on evapotranspiration has been proposed and been verified by using the climatical transect analysis method.For the noises and missing values in the time series data among model variables,the reconstruction methods are developed.Finally,taking the Northern Australia Tropical Transect,Shanxi Yicheng farmland ecosystem and Shaanxi Qinling forest ecosystem as the study areas,the model has applied to measure the vegetation resilience and explore the influencing factors of the vegetation resilience.The main conclusions are as follows:?1?The autoregressive model based on time series data can describe the dynamic changes and the memory effect of the vegetation ecosystem.According to the disturbance factors and the response of vegetation on water variation,this dissertation establishes the vegetation resilience measurement model based on ARx and lag correlation coefficient by taking vegetation index anomaly,temperature anomaly and drought monitoring index as model variables,and verifies the validity of the model through the variables correlation coefficient and stationarity test.Meanwhile,the Coupla function is used to establish the occurrence probability of extreme weather,and the driving factors of vegetation resilience changes are analyzed according to the tail correlation.?2?The drought index is an essential variable in the vegetation resilience model.Based on the land surface-atmospheric water balance process,a drought monitoring index CWDa based on the water deficit is proposed in this dissertation.The observations from the flux towers and the Australian Water Availability Project are used to calculate the CWDa.And the drought monitoring availability of the CWDa is tested in the Australian rainfall gradient zone.The results show that the CWDa can effectively detect the change of soil moisture and identify drought conditions in the humid and arid areas.The high sensitivity to deep soil moisture makes it beneficial to analyze the feedback of vegetation to the soil moisture changes in arid areas.The consistency of the CWDa established under different observation time ranges is higher,and the drought identification results are not affected by the number of observations.?3?For the missing values and noises in the time-series data used in the resilience measurement model,this dissertation proposes the reconstruction methods according to the attributes and the interannual difference of the indexes.The continuous meteorological observations and heat flux data can be obtained by the DINGO algorithm.For the vegetation index based on the remote sensing time series data,the Spline-Changing weight filtering method is proposed in order to judge the local minimum value.The method reduces the noises in the curve and also retains the reasonable fluctuation of the NDVI time series to describe the details of the dynamic changes of vegetation groups.Because the evapotranspiration data is greatly affected by meteorological conditions and annual differences,using the data around missing values to fill the gaps can obtain a higher estimation accuracy.?4?The results of driving factors analysis based on the remote sensing-based vegetation resilience measurement model show that the vegetation resilience in the Australian vegetation transect descents with the decrease of annual rainfall,and the vegetation community in the arid area is difficult to maintain steady-state under drought and extreme temperature.The vegetation resilience of the Yicheng's farmland ecosystem in Shanxi is affected by water deficit,while the Qinling's forest ecosystem in Shaanxi has high biomass and the vegetation community has not been disturbed obviously so that the change of the resilience is not significant.The case studies show that the vegetation resilience is affected by the structure of the vegetation community,the occurrence probability of extreme weather and the changes in rainfall.The vegetation groups with the higher tree-grass ratio and the C₄/C₃ ratio have stronger vegetation resilience.The vegetation groups in the arid and semi-arid regions are sensitive to climate change that it is necessary to reinforce the management of the ecosystems to reduce the risk of ecosystem steady-state transfer.This dissertation proposes a remote sensing measurement system for vegetation resilience based on environmental variables and vegetation response characteristics,which can use remote sensing data to efficiently and real-time monitor the changes of vegetation community resilience in different climatical areas and analyze the driving factors.It can help decision-makers and managers to choose and verify the ecological restoration strategies.This dissertation contains 39 figures,26 tables and 209 references.