Forest edge length estimation - a case study using the Swedish National Forest Inventory (NFI)

Forest edges may affect the organisms in a forest fragment by changing biotic and abiotic conditions, and they have important effects on biodiversity. This characteristic of forest landscape is usually assessed using remotely sensed data. When doing so, forest edge length is estimated using either polygon delineation (manually or automatically) or sampling methods. This study attempts to estimate the total forest edge length (at regional level) using three estimation techniques: mapped plot, line intersects sampling, and buffer zone. Another aim is to compare the precision of the results using these techniques with field data from the Swedish National Forest Inventory (NFI). The Swedish NFI data allows reasonable precision (less than 10% SE) estimating total forest edge length. In most cases, line intersects sampling and mapped plot (at the subplot level) produce more precise estimates than buffer zone (at cluster level). All three techniques can be applied in monitoring programs aimed at measuring important biodiversity metrics from forest landscape patterns. It was also found that edge density was highest in the south of Sweden. Another future application worth investigating is the ability to determine how forest edge length metrics change over time.