Work

A remote sensing approach for mapping teak (Tectona grandis) using vegetation phenology in Ghana's forest-savannah transition zone

Year: 2025

Type: article

Abstract: This study explores the use of satellite-derived phenological traits to identify and map teak ( Tectona grandis ) plantations from other vegetation types in Ghana's forest-savanna transitional zone. Landsat 8 images were obtained for two dates (January 21st and May 13th) representing the leaf fall and full foliage phases of teak. Three hundred and sixteen field coordinates were collected for supervised classification (70 %) and accuracy assessment (30 %). Three distinct image band composites were created: two using image band difference and image band ratio vegetation phenology approaches and one using the dry season image bands (representing the conventional remote sensing classification approach). The Kruskal-Wallis test and Dunn's pairwise test with Bonferroni adjustment to analyse the separability of teak vegetation from other vegetation classes. The Random Forest algorithm was used to classify the three image composites. Overall accuracy, kappa, and McNemar's test were used to evaluate classification accuracies. The image difference technique yielded the highest accuracy at 71.28 %, outperforming both the image ratio method (61.7 %) and the conventional mapping approach (58.51 %) in distinguishing teak from other vegetation classes. Furthermore, comparing the land cover maps generated from image ratio and image difference methods revealed significant differences (McNemar's χ2 = 4.348, p-value = 0.04), indicating that the image difference approach is superior for mapping teak. However, the binary teak versus non-teak maps showed a similar representation of teak with equal class area (≈69 km 2 ) and no statistical difference in spatial extent (McNemar's χ 2 = 9.841, p-value = 0.06) for the two phenological approaches. The study demonstrates the effectiveness of vegetation phenology in distinguishing teak from other vegetation classes. This approach can support silvicultural efforts in managing teak plantations in Ghana's high forest zone.

Source: Heliyon

DOI: 10.1016/j.heliyon.2025.e44180

Authors: George Ashiagbor, Ophelia Ehornam Ahorlu, Isaac Stanisluv Essah, Michael Ofori Gyamfi, Stephen Boahen Asabere

Institutions: Kwame Nkrumah University of Science and Technology, University of Göttingen

Language: English

Topic: Remote Sensing in Agriculture

Subfield: Ecology

Field: Environmental Science

Domain: Physical Sciences

SDG: Life on land

Open Access status: gold

Estimated APC paid: $2,270