Forest Height Estimation Method Using Space-borne X-band SAR Interferometry

As an important structural parameter,forest height is an indicator for estimating the potential of forest volume production,and is closely related to forest ground biomass(AGB)and carbon storage.Therefore,accurate and large-scale access to forest height information is of great significance for the fine management of forests,carbon cycle and scientific research on climate change.Interferometric synthetic aperture radar(In SAR)is sensitive to the vertical structure of vegetation and is one of the most potential remote sensing technologies to map forest height in large area.Terra SAR-X add-on for Digital Elevation Measurement(Tan DEM-X)is the first single-pass radar interferometer in space based on two radar satellites flying in close formation.The In SAR decorrelation of Tan DEM-X is mainly caused by volume scattering effect,so it is suitable for estimating forest parameters such as height and biomass.In this paper,Tan DEM-X single-polarizaiton interferometric data was used to study the SINC model(One kind of simplified Random Volume over Ground model)method(SINC method,in short)for forest height estimation,which is based on coherence coefficient.The forest in the north of Genhe city and Wangyedian forest farm in Chifeng city were used as experimental sites to analyze the performance of the SINC method for estimating forest height under different terrain conditions.The research content can be divided into the following three parts:(1)The performance for estimating forest height of the difference method(DIFF method,in short,taking the forest height as the difference between the DSM from Tan DEM-X and the DEM from Li DAR)and the SINC method was analyzed.The forest height estimation results were compared against stand average height and average dominant tree height computed from Li DAR derived canopy height model(CHM).It showes that the accuracy of the SINC method is better than the DIFF method.On the spatial scale of 30m×30m,the accuracy of the DIFF method for estimating stand average height and average dominant tree height was 77.52% and74.35%,respectively.The accuracy of the SINC method was 84.22% and 85.60%,respectively.On the spatial scale of 90m×90m,the estimation accuracy of the two methods was close.Compared with the DIFF method,the SINC method does not rely on high-precision DEM,and the estimation result is much more consistent with average dominant tree height.It does not need to be calibrated with the ground measured plot data,and has great value in the large-scale forest height mapping.(2)The effects of signal-to-noise ratio(SNR)decorrelation,spatial baseline,ground-to-volume scattering ratio and extinction coefficient on the estimation of the SINC method were analyzed,and the following conclusions were obtained: 1)SNR decorrelation is the main non-volumetric decorrelation contribution in Tan DEM-X data.Accurate calculation and removal of SNR decorrelation are essential for the SINC method,otherwise it will be taken as the volume decorrelation,so that the forest height is overestimated;2)When selecting the spatial baseline,baselines with height of ambiguity(Ho A)values in the order of 2?3 times of the local forest height level are preferable;3)For deciduous species,the extinction coefficient is lower in winter,which is more in line with the SINC method assumption,and the estimation results of forest height is more accurate;4)When the ground-to-volume scattering ratio is relatively large,using the SINC model to estimate forest height will cause an overestimation;and when the extinction coefficient is large,it will cause an underestimation.(3)For the estimation of forest height in complex terrain,the influence of terrain on the SINC method was analyzed,and a method based on the SINC model and Monte Carlo simulation to evaluate the reliability of forest height estimation was proposed.This method porvides a basis for more reasonable masking of areas affected by terrain and effective fusion of multi-track data estimation results.It has important reference value for large-area practical application of Tan DEM-X and similar In SAR data.In summary,this study established a technical process for estimating boreal forest height from space-borne In SAR data without temporal decorrelation under different terrain conditions.The study presented improved method for regional forest height monitoring using space-borne In SAR system,which will provide technique support for forest resources monitoring and forestry management activities.