Continuous Detection of Forest Loss in Vietnam, Laos, and Cambodia Using Sentinel-1 Data

In this study, we demonstrate the ability of a new operational system to detect forest loss at a large scale accurately and in a timely manner. We produced forest loss maps every week over Vietnam, Cambodia, and Laos (>750,000 km2 in total) using Sentinel-1 data. To do so, we used the forest loss detection method based on shadow detection. The main advantage of this method is the ability to avoid false alarms, which is relevant in Southeast Asia where the areas of forest disturbance may be very small and scattered and detection is used for alert purposes. The estimated user accuracy of the forest loss map was 0.95 for forest disturbances and 0.99 for intact forest, and the estimated producer’s accuracy was 0.90 for forest disturbances and 0.99 for intact forest, with a minimum mapping unit of 0.1 ha. This represents an important step forward compared to the values achieved by previous studies. We also found that approximately half of forest disturbances in Cambodia from 2018 to 2020 occurred in protected areas, which emphasizes the lack of efficiency in the protection and conservation of natural resources in protected areas. On an annual basis, the forest loss areas detected using our method are found to be similar to the estimations from Global Forest Watch. These results highlight the fact that this method provides not only quick alerts but also reliable detections that can be used to calculate weekly, monthly, or annual forest loss statistics at a national scale.

Operational Large-Scale Segmentation of Imagery Based on Iterative Elimination

Image classification and interpretation are greatly aided through the use of image segmentation. Within the field of environmental remote sensing, image segmentation aims to identify regions of unique or dominant ground cover from their attributes such as spectral signature, texture and context. However, many approaches are not scalable for national mapping programmes due to limits in the size of images that can be processed. Therefore, we present a scalable segmentation algorithm, which is seeded using k-means and provides support for a minimum mapping unit through an innovative iterative elimination process. The algorithm has also been demonstrated for the segmentation of time series datasets capturing both the intra-image variation and change regions. The quality of the segmentation results was assessed by comparison with reference segments along with statistics on the inter- and intra-segment spectral variation. The technique is computationally scalable and is being actively used within the national land cover mapping programme for New Zealand. Additionally, 30-m continental mosaics of Landsat and ALOS-PALSAR have been segmented for Australia in support of national forest height and cover mapping. The algorithm has also been made freely available within the open source Remote Sensing and GIS software Library (RSGISLib).

Multitemporales, flugzeuggetragenes Laserscanning für die Forstwirtschaft

Multi-temporal airborne laserscanning data for forestry applications During the last decade airborne laserscanning (ALS) data has been established as a suitable data source for three-dimensional description of forests and for deriving forest parameters. For the study area Vorarlberg, Austria, the potential of multi-temporal ALS data (data from 2004 and 2011) was analyzed for deriving the amount of harvests in terms of area and stem volume. The data were used as well to assess the site index with the aid of the changes in top heights. The analyses have shown that harvested areas could be detected with a minimum mapping unit of 20 m2, which corresponds to harvested single trees from the dominant canopy layer. The average amount of harvested stem volume could be estimated with an overall accuracy of 96.4%. The derived changes in top heights clearly reflect the local growth conditions. For the study area Liechtenstein two ALS data sets (leaf-on and leaf-off) were used for the differentiation between coniferous and deciduous forests. Here, an overall accuracy of 92% could be achieved.

Fire frequency analysis in Portugal (1975 - 2005), using Landsat-based burnt area maps

Fire frequency in 21 forest planning regions of Portugal during the period 1975–2005 was estimated from historical burnt area maps generated with semi-automatic classification of Landsat Thematic Mapper (TM) satellite imagery. Fire return interval distributions were modelled with the Weibull function and the estimated parameters were used to calculate regional mean, median and modal fire return intervals, as well as regional hazard functions. Arrangement of the available data into three different time series allowed for assessment of the effects of minimum mapping unit, time series length and use of censored data on the Weibull function parameter estimates. Varying the minimum mapping unit between 5 and 35 ha had a negligible effect on parameter estimates, whereas changing the time series length from 22 to 31 years substantially affected the estimates. However, the strongest effect was caused by censored data. Its exclusion led to substantial overestimation of fire frequency and of burning probability dependence on fuel age. We estimated a country-wide mean fire interval of 36 years and an annual burnt area of 1.2%. Regional variations in fire frequency descriptors were interpreted in terms of land cover and land use practices that affect the contemporary fire regime in Portugal.

Detection of post-fire residuals using high- and medium-resolution satellite imagery

Forest fires can burn across large forested areas over short time periods, but they rarely consume all the trees in their path. Fires leave live irregularly shaped patches or rows of mature trees known as "residuals" within the fire perimeter. IKONOS and Landsat Enhanced Thematic Mapper Plus satellite imagery were acquired over two forested areas affected by fire in the northern boreal forest of Alberta. Each image was classified and residuals were detected with greater than 88% accuracy. Residual patches were grouped into nine minimum mapping unit (MMU) classes and area, patch, and shape level metrics were calculated for each group. Analysis of metric results highlighted how the choice of satellite imagery used to characterize and quantify residuals, the size of the MMU used to define the residuals, and human induced land use cover change (LUCC) processes occurring within fire perimeters were interrelated factors that impacted estimates of residual numbers and sizes. Residual metrics calculated in one fire perimeter should therefore be carefully assessed according to local land use and land cover change dynamics before suggesting that residual information captured in any fire perimeter can typify residual patterns elsewhere. Key words: remote sensing, high resolution, medium resolution, satellite imagery, forest fires, wildfire, residual forest islands, geographic information systems (GIS), minimum mapping unit