Development of new remote sensing methods for mapping green vegetation and exposed bedrock fractions within heterogeneous landscapes

Abstract

Karst rocky desertification is a process of land desertification associated with human disturbance of the fragile karst ecosystems. The fractional cover of photosynthetic vegetation (PV) and exposed bedrock (Rock) are the main land-surface symptoms of karst rocky desertification. In this study, we explored a new methodology for quantifying PV and Rock by remote sensing. To reduce the effects of the high heterogeneity of karst landscapes on vegetation information extraction, a whole image was segmented into relatively homogeneous subsets and then the PV was estimated using a normalized difference vegetation index spectral mixture analysis (NDVI-SMA) model. The percentage of Rock was estimated using a karst rocky desertification synthesis index (KRDSI) and lignin cellulose absorption index (LCA). The results showed that the heterogeneity of a complex landscape is a major factor in the uncertainty of PV retrievals. The fractional cover of PV can be accurately estimated by the proposed method, but might be underestimated using NDVI and overestimated using the SMA-NDVI model. The bedrock fractions can be rapidly and objectively estimated with Hyperion or simulated Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) imagery. Compared with multispectral images, hyperspectral images could be used to estimate PV and Rock more accurately. Our findings indicate that PV and Rock can be directly and efficiently quantified using remote sensing techniques within heterogeneous landscapes.