MAPPING DISTURBANCE DYNAMICS IN WET SCLEROPHYLL FORESTS USING TIME SERIES LANDSAT
Open Access
- 23 June 2016
- journal article
- Published by Copernicus GmbH in The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
- Vol. XLI-B8, 633-641
Abstract
In this study, we characterised the temporal-spectral patterns associated with identifying acute-severity disturbances and low-severity disturbances between 1985 and 2011 with the objective to test whether different disturbance agents within these categories can be identified with annual Landsat time series data. We analysed a representative State forest within the Central Highlands which has been exposed to a range of disturbances over the last 30 years, including timber harvesting (clearfell, selective and thinning) and fire (wildfire and prescribed burning). We fitted spectral time series models to annual normal burn ratio (NBR) and Tasseled Cap Indices (TCI), from which we extracted a range of disturbance and recovery metrics. With these metrics, three hierarchical random forest models were trained to 1) distinguish acute-severity disturbances from low-severity disturbances; 2a) attribute the disturbance agents most likely within the acute-severity class; 2b) and attribute the disturbance agents most likely within the low-severity class. Disturbance types (acute severity and low-severity) were successfully mapped with an overall accuracy of 72.9 %, and the individual disturbance types were successfully attributed with overall accuracies ranging from 53.2 % to 64.3 %. Low-severity disturbance agents were successfully mapped with an overall accuracy of 80.2 %, and individual agents were successfully attributed with overall accuracies ranging from 25.5 % to 95.1. Acute-severity disturbance agents were successfully mapped with an overall accuracy of 95.4 %, and individual agents were successfully attributed with overall accuracies ranging from 94.2 % to 95.2 %. Spectral metrics describing the disturbance magnitude were more important for distinguishing the disturbance agents than the post-disturbance response slope. Spectral changes associated with planned burning disturbances had generally lower magnitudes than selective harvesting. This study demonstrates the potential of landsat time series mapping for fire and timber harvesting disturbances at the agent level and highlights the need for distinguishing between agents to fully capture their impacts on ecosystem processes.Keywords
This publication has 25 references indexed in Scilit:
- Mega-fires, inquiries and politics in the eucalypt forests of Victoria, south-eastern AustraliaForest Ecology and Management, 2013
- Forest cover trends from time series Landsat data for the Australian continentInternational Journal of Applied Earth Observation and Geoinformation, 2013
- The efficacy of fuel treatment in mitigating property loss during wildfires: Insights from analysis of the severity of the catastrophic fires in 2009 in Victoria, AustraliaJournal of Environmental Management, 2012
- Object-based cloud and cloud shadow detection in Landsat imageryRemote Sensing of Environment, 2012
- Mapping wildfire and clearcut harvest disturbances in boreal forests with Landsat time series dataRemote Sensing of Environment, 2011
- Detecting trends in forest disturbance and recovery using yearly Landsat time series: 2. TimeSync — Tools for calibration and validationRemote Sensing of Environment, 2010
- Detecting trends in forest disturbance and recovery using yearly Landsat time series: 1. LandTrendr — Temporal segmentation algorithmsRemote Sensing of Environment, 2010
- Quantification of live aboveground forest biomass dynamics with Landsat time-series and field inventory data: A comparison of empirical modeling approachesRemote Sensing of Environment, 2010
- Estimating the probability of mountain pine beetle red-attack damageRemote Sensing of Environment, 2006
- Ecological disturbance and the conservative management of eucalypt forests in AustraliaForest Ecology and Management, 1994