Snow-covered Landsat time series stacks improve automated disturbance mapping accuracy in forested landscapes

Stueve, Kirk M.; Housman, Ian W.; Zimmerman, Patrick L.; Nelson, Mark D.; Webb, Jeremy; Perry, Charles H.; Chastain, Robert A.; Gormanson, Dale D.; Huang, Chengquan; Healey, Sean P.; Cohen, Warren B.

doi:10.1016/j.rse.2011.07.005

Cited by 32 publications

(27 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, forest clearcuts in temperate forests can be more easily detected in winter due to the high radiance contrast in the image data (Püssa et al 2005;Liira et al 2006;Kardakov et al 2009). For example, Pinder et al (1999) successfully mapped pine forest with winter images, and Stueve et al (2011) improved overall accuracy from 86.3% to 91.2% when including winter imagery in order to map mixed forests of the Great Lake basin. In China, bamboo understory in panda habitats is difficult to map with summer images alone, because both understory and canopy provide similar radiance in summertime, but winter images improved the classification accuracy up to 89% (Wang et al 2009).…”

Section: Discussionmentioning

confidence: 99%

Monitoring selective logging with Landsat satellite imagery reveals that protected forests in Western Siberia experience greater harvest than non-protected forests

et al. 2017

View full text Add to dashboard Cite

SUMMARYWhen timber harvesting is an important source of local income and forest resources are declining, even forests that are designated as protected areas may become vulnerable. Therefore, regular monitoring of forest disturbance is necessary to enforce the protection of forest ecosystems. However, mapping forest disturbance with satellite imagery can be complicated if the majority of the harvesting is selective logging and not clearcuts. Our goal was to map both selective logging and clearcuts within and outside of protected areas in Western Siberia, a region with a highly developed timber industry. Combining summer and winter imagery allowed us to accurately estimate not only clearcuts, but also selective logging. Winter Landsat images substantially improved our classification and resulted in a highly accurate forest disturbance map (97.5% overall accuracy and 86% user accuracy for the rarest class, clearcuts). Selective logging and stripcuts were the dominant disturbance types, accounting for 96.3% of all forest disturbances, versus 3.7% for clearcuts. The total annual forest disturbance rate (i.e. disturbance rate for clearcuts, stripcuts and selective logging together) was 0.53%, but total forest disturbance within protected areas was greater than in non-protected forest (0.66% versus 0.50%, respectively), and so was the annual rate of selective logging (i.e. without clearcuts, 0.37% versus 0.25%, respectively). Our results highlight that monitoring only clearcuts without assessing selective logging might result in significant underestimation of forest disturbance. Also, when timber harvesting is important for the local economy and when protected areas have valuable timber resources that have already been depleted elsewhere, then additional protection may be necessary in order to maintain natural forests within protected areas. We suggest that this is the * Correspondence: Alexander Shchur e-mail: alexander.shchur. 1985@gmail.ru situation in our study area in Western Siberia right now and is likely the situation in many other parts of the globe as well.

show abstract

Section: Discussionmentioning

confidence: 99%

Monitoring selective logging with Landsat satellite imagery reveals that protected forests in Western Siberia experience greater harvest than non-protected forests

et al. 2017

View full text Add to dashboard Cite

show abstract

“…For example, an algorithm developed by Huang et al 2010 [29] has been implemented wall-to-wall for the continental US and has been shown to be very effective at mapping clear cut harvest, but has also been noted to have difficulty in complex heterogeneous landscapes [40]. This algorithm uses the concept of normalized values for pixels based on the mean and standard deviation of known forest pixels from the same scene and uses multiple images to improve the forest/non-forest classification.…”

Section: Introductionmentioning

confidence: 99%

An Automated Approach to Map the History of Forest Disturbance from Insect Mortality and Harvest with Landsat Time-Series Data

et al. 2014

View full text Add to dashboard Cite

Forests contain a majority of the aboveground carbon (C) found in ecosystems, and understanding biomass lost from disturbance is essential to improve our C-cycle knowledge. Our study region in the Wisconsin and Minnesota Laurentian Forest had a strong decline in Normalized Difference Vegetation Index (NDVI) from 1982 to 2007, observed with the National Ocean and Atmospheric Administration's (NOAA) series of Advanced Very High Resolution Radiometer (AVHRR). To understand the potential role of disturbances in the terrestrial C-cycle, we developed an algorithm to map forest disturbances from either harvest or insect outbreak for Landsat time-series stacks. We merged two image analysis approaches into one algorithm to monitor forest change that included: (1) multiple disturbance index thresholds to capture clear-cut harvest; and (2) a OPEN ACCESS Remote Sens. 2014, 6 2783 spectral trajectory-based image analysis with multiple confidence interval thresholds to map insect outbreak. We produced 20 maps and evaluated classification accuracy with air-photos and insect air-survey data to understand the performance of our algorithm. We achieved overall accuracies ranging from 65% to 75%, with an average accuracy of 72%. The producer's and user's accuracy ranged from a maximum of 32% to 70% for insect disturbance, 60% to 76% for insect mortality and 82% to 88% for harvested forest, which was the dominant disturbance agent. Forest disturbances accounted for 22% of total forested area (7349 km 2 ). Our algorithm provides a basic approach to map disturbance history where large impacts to forest stands have occurred and highlights the limited spectral sensitivity of Landsat time-series to outbreaks of defoliating insects. We found that only harvest and insect mortality events can be mapped with adequate accuracy with a non-annual Landsat time-series. This limited our land cover understanding of NDVI decline drivers. We demonstrate that to capture more subtle disturbances with spectral trajectories, future observations must be temporally dense to distinguish between type and frequency in heterogeneous landscapes.

show abstract

“…Mathieu et al 2007;Myint et al 2011), some widely used moderate-resolution products such as NLCD and Vegetation Change Tracker (e.g. Stueve et al 2011;Wickham et al 2013), and established thresholds for acceptable accuracy in land cover products (Shao and Wu 2008). The comparatively narrow adjusted area ranges of several individual high-resolution land cover classes demonstrate a higher degree of confidence in these respectively mapped areas.…”

Section: Discussion/conclusionmentioning

confidence: 99%

High-resolution maps of forest-urban watersheds present an opportunity for ecologists and managers

et al. 2014

Self Cite

View full text Add to dashboard Cite

Context Green infrastructure may improve water quality and mitigate flooding in forest-urban watersheds, but reliably quantifying all benefits is challenging because most land cover maps depend on moderate-to low-resolution data. Complex and spatially heterogeneous landscapes that typify foresturban watersheds are not fully represented with these types of data. Hence important questions concerning how green infrastructure influences water quality and quantity at different spatial scales remain unanswered. Objectives Demonstrate the feasibility of creating novel high-resolution land cover maps across entire watersheds and highlight deficiencies of standard land cover products. Methods We used object-based image analysis (OBIA) to create high-resolution (0.5 m) land cover maps and detect tree canopy overlapping impervious surfaces for a representative forest-urban watershed in Duluth, MN, USA. Unbiased estimates of accuracy and area were calculated and compared with similar metrics for the 30-m National Land Cover Database (NLCD). Results Mapping accuracies for the high-resolution land cover and canopy overlap maps were *90 %. Error-adjusted estimates of area indicated that impervious surfaces comprised *21 % of the watershed, tree canopy overlapped *10 % of impervious surfaces, and that three high-resolution land cover classes differed significantly from similar NLCD classes. Conclusions OBIA can efficiently generate highresolution land cover products of entire watersheds that are appropriate for research and inclusion in the decision-making process of managers. Metrics derived from these products will likely differ from standard land cover maps and may produce new insights, especially when considering the unprecedented opportunity to evaluate fine-scale spatial heterogeneity across watersheds.

show abstract

Snow-covered Landsat time series stacks improve automated disturbance mapping accuracy in forested landscapes

Cited by 32 publications

References 58 publications

Monitoring selective logging with Landsat satellite imagery reveals that protected forests in Western Siberia experience greater harvest than non-protected forests

Monitoring selective logging with Landsat satellite imagery reveals that protected forests in Western Siberia experience greater harvest than non-protected forests

An Automated Approach to Map the History of Forest Disturbance from Insect Mortality and Harvest with Landsat Time-Series Data

High-resolution maps of forest-urban watersheds present an opportunity for ecologists and managers

Contact Info

Product

Resources

About