Mapping

We assessed the forest disturbance and recovery history in the Southern Rocky Mountains Ecoregion using a 13-year time series of Landsat image stacks. An automated classification workflow that integrates temporal segmentation techniques and a random forest classifier was used to examine disturbance patterns. To enhance efficiency in selecting representative samples at the ecoregion scale, a new sampling strategy that takes advantage of the scene-overlap among adjacent Landsat images was designed. The segment-based assessment revealed that the overall accuracy for all 14 scenes varied from 73.6% to 92.5%, with a mean of 83.1%. A design-based inference indicated the average producer’s and user’s accuracies for MPB mortality were 85.4% and 82.5% respectively.

A time series Landsat stack covering the period from 1995 to 2015 was employed to detect the urban dynamics in Northwest Arkansas via a two-stage classification approach. A set of spectral indices that have been proven to be useful in urban area extraction together with the original Landsat spectral bands were used in the maximum likelihood classifier and random forest classifier to distinguish urban from non-urban pixels for each year. A temporal trajectory polishing method, involving temporal filtering and heuristic reasoning, was then applied to the sequence of classified urban maps for further improvement.

We proposed an integrative approach to classifying leafless and full-canopy trees on fall imagery of 0.46-m WorldView-2 (WV-2). We adopted a two-step classification approach by first removing the non-forest area from the WV-2 imagery with the mask layer generated from the publicly available summer imagery of 1-m National Agriculture Imagery Program imagery via object-based image analysis. Two classification methods, namely decision tree (DT) and nearest neighborhood (NN), were employed to classify full canopy and leafless trees in the masked WV-2 imagery.