A comparison between satellite and airborne multispectral data for the assessment of mangrove areas in the eastern Caribbean

ABSTRACT

Satellite (SPOT XS and Landsat TM) and airborne multispectral (CASI) imagery of mangroves was acquired from the Turks and Caicos Islands, British West Indies. The descriptive resolution and accuracy of each image type is compared for two applications: mangrove habitat mapping and the measurement of mangrove canopy characteristics (leaf area index and canopy closure).

Mangroves could be separated from non-mangrove vegetation to an accuracy of only 57% with SPOT XS data but better discrimination could be achieved with either Landsat TM or CASI (in both cases accuracy was >90%).

CASI data permitted a more accurate classification of different mangrove habitats than was possible using Landsat TM. Nine mangrove habitats could be mapped to an accuracy of 85% with the high-resolution airborne data compared to 31% obtained with TM. A maximum of three mangrove habitats were separable with Landsat TM: the accuracy of this classification was 83%.

Measurement of mangrove canopy characteristics is achieved more accurately with CASI than with either satellite sensor, but high costs probably make it a less cost-effective option. The cost-effectiveness of each sensor is discussed for each application.