Mapping understory invasive plant species with field and remotely sensed data in Chitwan, Nepal

Authors: Jie Dai*, San Diego State University/University of Califronia, Santa Barbara, Dar Roberts, University of California, Santa Barbara, Doug Stow, San Diego State University, Li An, San Diego State University, Phaedon Kyriakidis, Cyprus University of Technology
Topics: Remote Sensing
Keywords: Invasive species, Understory vegetation, Spectral mixture analysis, Maxent, Landsat, Chitwan National Park
Session Type: Paper
Presentation File: No File Uploaded


We utilized Landsat imagery, spectral mixture analysis (SMA), image texture metrics and a maximum entropy (Maxent) modeling framework to map the extent of an understory invasive vine (Mikania micrantha) in Chitwan National Park, Nepal. We collected reference spectra from both online spectral libraries and in the field using ASD FieldSpec 4 spectroradiometer. Combined with image spectra, we developed a spectral library and applied multiple endmember SMA (MESMA) to selected Landsat images. Incorporating the resultant green vegetation and shade fractions into Maxent, we successfully mapped the spatial distribution of understory M. micrantha in the study area (training and testing AUC values around 0.80). In vegetated places, especially mature forests, increase of green vegetation fraction and decrease of shade fraction were associated with higher likelihood of M. micrantha presence. In addition, the inclusion of elevation as a model input further improved map accuracy (AUC around 0.95).

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login