Estimation of Leaf Area Index Over Heterogeneous Regions Using the Vegetation Type Information and PROSAIL Model

The leaf area index (LAI) is a parameter that can indicate the vegetation canopy structure and accurately reflect the growth state of vegetation. Most studies estimate the LAI of single vegetation in homogeneous areas, but only few studies have explored the LAI inversion of heterogeneous areas. Canopy heterogeneity in heterogeneous regions may increase the uncertainty and difficulty of LAI quantitative inversion. Therefore, LAI retrieval in heterogeneous areas needs to be studied to obtain LAI distribution maps in a large spatial range. In this study, an LAI inversion model considering vegetation types was proposed based on the look-up table (LUT) method and the PROSAIL model for estimating the LAI of heterogeneous surfaces. First, the LUTs of different vegetation types were generated by using PROSAIL with a priori information of multispecies. Second, the corresponding LUT for LAI estimation was selected according to the determined vegetation types. Finally, a parametric sensitivity analysis was conducted based on the PROSAIL model to recognize the key parameters of the algorithm's efficiency. Results show that the approach considering vegetation types (
R
²
= 0.63, RMSE = 0.75 /
R
²
= 0.64, RMSE = 0.50) is superior to the traditional approach that does not consider vegetation types (
R
²
= 0.50, RMSE = 1.32 /
R
²
= 0.17, RMSE = 1.81). Therefore, the former approach can greatly improve the accuracy of multispecies LAI estimation, especially for areas with high canopy heterogeneity. The proposed approach for multispecies vegetation LAI retrieval can provide new insights for studying the ecological status of complex land surface regions and exhibit an excellent potential for the extended application of the PROSAIL model in heterogeneous areas.