Next Article in Journal
Transformation of a Degraded Pinus massoniana Plantation into a Mixed-Species Irregular Forest: Impacts on Stand Structure and Growth in Southern China
Next Article in Special Issue
Simulation of Quaking Aspen Potential Fire Behavior in Northern Utah, USA
Previous Article in Journal
Multilevel Governance for Forests and Climate Change: Learning from Southern Mexico
Previous Article in Special Issue
Effects of a Wildfire on Selected Physical, Chemical and Biochemical Soil Properties in a Pinus massoniana Forest in South China
Open AccessArticle

Evaluating an Automated Approach for Monitoring Forest Disturbances in the Pacific Northwest from Logging, Fire and Insect Outbreaks with Landsat Time Series Data

1
Biospheric Sciences Laboratory, Code 618, NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA
2
Department of Forest Resources Management, University of British Columbia, Vancouver, BC V6T 1Z4, Canada
3
Royal Botanic Gardens, Kew, Richmond, Surrey, TW9 3AE, UK
4
Department of Geographical Sciences, University of Maryland, College Park, MD 20742, USA
*
Author to whom correspondence should be addressed.
Forests 2014, 5(12), 3169-3198; https://doi.org/10.3390/f5123169
Received: 9 September 2014 / Revised: 2 December 2014 / Accepted: 3 December 2014 / Published: 12 December 2014
(This article belongs to the Special Issue Climate Change and Forest Fire)
Forests are the largest aboveground sink for atmospheric carbon (C), and understanding how they change through time is critical to reduce our C-cycle uncertainties. We investigated a strong decline in Normalized Difference Vegetation Index (NDVI) from 1982 to 1991 in Pacific Northwest forests, observed with the National Ocean and Atmospheric Administration’s (NOAA) series of Advanced Very High Resolution Radiometers (AVHRRs). To understand the causal factors of this decline, we evaluated an automated classification method developed for Landsat time series stacks (LTSS) to map forest change. This method included: (1) multiple disturbance index thresholds; and (2) a spectral trajectory-based image analysis with multiple confidence thresholds. We produced 48 maps and verified their accuracy with air photos, monitoring trends in burn severity data and insect aerial detection survey data. Area-based accuracy estimates for change in forest cover resulted in producer’s and user’s accuracies of 0.21 ± 0.06 to 0.38 ± 0.05 for insect disturbance, 0.23 ± 0.07 to 1 ± 0 for burned area and 0.74 ± 0.03 to 0.76 ± 0.03 for logging. We believe that accuracy was low for insect disturbance because air photo reference data were temporally sparse, hence missing some outbreaks, and the annual anniversary time step is not dense enough to track defoliation and progressive stand mortality. Producer’s and user’s accuracy for burned area was low due to the temporally abrupt nature of fire and harvest with a similar response of spectral indices between the disturbance index and normalized burn ratio. We conclude that the spectral trajectory approach also captures multi-year stress that could be caused by climate, acid deposition, pathogens, partial harvest, thinning, etc. Our study focused on understanding the transferability of previously successful methods to new ecosystems and found that this automated method does not perform with the same accuracy in Pacific Northwest forests. Using a robust accuracy assessment, we demonstrate the difficulty of transferring change attribution methods to other ecosystems, which has implications for the development of automated detection/attribution approaches. Widespread disturbance was found within AVHRR-negative anomalies, but identifying causal factors in LTSS with adequate mapping accuracy for fire and insects proved to be elusive. Our results provide a background framework for future studies to improve methods for the accuracy assessment of automated LTSS classifications. View Full-Text
Keywords: Pacific Northwest; Landsat; AVHRR; disturbance; time series; NDVI; logging; insect; fire; GIMMS Pacific Northwest; Landsat; AVHRR; disturbance; time series; NDVI; logging; insect; fire; GIMMS
Show Figures

Graphical abstract

MDPI and ACS Style

Neigh, C.S.R.; Bolton, D.K.; Williams, J.J.; Diabate, M. Evaluating an Automated Approach for Monitoring Forest Disturbances in the Pacific Northwest from Logging, Fire and Insect Outbreaks with Landsat Time Series Data. Forests 2014, 5, 3169-3198. https://doi.org/10.3390/f5123169

AMA Style

Neigh CSR, Bolton DK, Williams JJ, Diabate M. Evaluating an Automated Approach for Monitoring Forest Disturbances in the Pacific Northwest from Logging, Fire and Insect Outbreaks with Landsat Time Series Data. Forests. 2014; 5(12):3169-3198. https://doi.org/10.3390/f5123169

Chicago/Turabian Style

Neigh, Christopher S.R.; Bolton, Douglas K.; Williams, Jennifer J.; Diabate, Mouhamad. 2014. "Evaluating an Automated Approach for Monitoring Forest Disturbances in the Pacific Northwest from Logging, Fire and Insect Outbreaks with Landsat Time Series Data" Forests 5, no. 12: 3169-3198. https://doi.org/10.3390/f5123169

Find Other Styles

Article Access Map by Country/Region

1
Only visits after 24 November 2015 are recorded.
Search more from Scilit
 
Search
Back to TopTop