Comparison of Remote Sensing Estimation Models for Leaf Area Index of Rubber Plantation in Hainan Island
Leaf area index (LAI) is an important index to describe the growth status and canopy structure of vegetation, is of great theoretical and practical significance to quickly obtain LAI of large area vegetation and crops for ecosystem science research and agricultural & forestry production guidance...
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | article |
| Lenguaje: | EN ZH |
| Publicado: |
Editorial Office of Smart Agriculture
2021
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/d63fc6aa0ac445d9809a311f58402d63 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | Leaf area index (LAI) is an important index to describe the growth status and canopy structure of vegetation, is of great theoretical and practical significance to quickly obtain LAI of large area vegetation and crops for ecosystem science research and agricultural & forestry production guidance. In this study, the typical tropical crop rubber tree in Hainan Island was selected as the research area, the LAI estimation model of rubber plantation based on satellite remote sensing vegetation indices was constructed, and its spatiotemporal variation was analyzed. The results showed that, compared with correlations between LAI and the indices of normalized difference vegetation index (NDVI), green NDVI (GNDVI), ratio vegetation index (RVI) and wide dynamic range vegetation index (WDRVI), correlations were higher between LAI and the indices of enhanced vegetation index (EVI), soil adjusted vegetation index (SAVI), difference vegetation index (DVI) and modified soil adjusted vegetation index (MSAVI). Among the LAI estimation models based on different vegetation indices (linear, exponential and logarithmic models), the linear estimation model based on EVI index was the best, and its coefficient of determination (R2) was 0.69. The accuracy of LAI estimation model was high. The linear fitting R2 of observed and simulated LAI was 0.67, the root mean square error (RMSE) was 0.16, and the average relative error (RE) was -0.25%. However, there was underestimation in the middle value and overestimation in the high and low value area of LAI. The high LAI values (4.40－6.23) were mainly distributed in Danzhou and Baisha in the west of Hainan Island, the middle LAI values (3.80－4.40) were mainly distributed in Chengmai, Tunchang and Qiongzhong in the middle of Hainan Island, and the low LAI values (2.69－3.80) were mainly distributed in Ding'an, Qionghai, Wanning, Ledong and Sanya in the east and south of Hainan Island. In summary, the linear estimation model for rubber plantation LAI based on EVI index obtained high accuracy, and has good values of popularization and appliance. |
|---|