Potential for the wider application of national forest inventories to estimate the contagion metric for landscapes

Abstract

National forest inventories (NFIs) have traditionally been designed to assess the production value of forests as well as forest biodiversity. However, in this study, the aim is to show a new application of NFIs, namely the estimation of the landscape metric contagion. This metric is commonly calculated on raster-based land cover/use maps. In this study, a sample-based dataset from the Swedish NFI was used. The estimated contagion metric is based on a distance-dependent function so that the value of the metric is small for longer distances, whereas the corresponding estimated variance is large for longer distances. With this procedure, comparisons can be made for different landscapes at a given time and or to compare any given landscape over time. The main advantages are that the approach can be applied where raster-based land cover/use maps of the landscape are not available and that the data obtained from NFIs (e.g., land cover type) typically are of high quality in comparison with remotely sensed data due to being based on direct observation in the field survey. The procedure applied here accommodates both the patch-mosaic and the gradient-based model approach to landscape structure.

Semi-Automated Sample-Based Forest Degradation Monitoring with Photointerpretation of High-Resolution Imagery

“…PAI, as we calculated it for MD and PG, is an index calculated from the set of between-subplot and between-plot values collected on a subpopulation of our PI plots. Because the nature of the sampling distribution of this statistic is unknown, and for illustrative purposes, we chose to use a delete-one jackknife variance estimation approach, described in Thompson [52] and used by Kleinn [34] and Ramezani and Ramezani [38]. In this approach, one cluster plot is deleted from the sample, and PAI is calculated.…”

“…Forests 2019, 10, x FOR PEER REVIEW 14 of 19 variance estimator, therefore, appears to be a reasonable approach to addressing this situation, as it does not require distributional assumptions. Both Kleinn [34] and Ramezani and Ramezani [38] used the jackknife variance estimator when calculating landscape metrics based on forest inventory plots. It must be noted, however, that procedures for comparing means using this method are not straightforward.…”

“…This research was motivated by the need for information on forest degradation using definitions that correspond precisely with what a human interpreter would choose, as well as interest in leveraging the abundance of georeferenced national forest inventory data [34,38,60] and PI-based forest monitoring information [28] that has become readily available. Potential applications include multi-scale comparisons of changes in metric values or differences between analysis units (e.g., county groups, Figure 5, and related discussion), or use as a covariate in assessments of forest condition and health, as has been done with other fragmentation metrics [61,62].…”

Forest fragmentation assessment using field-based sampling data from forest inventories

“…A landscape has various aspects (composition and configuration) and thus it is not practical to measure and quantify all aspects of it using only a single index like AI. An alternative is to combine AI with other landscape indices such as contagion (Ramezani and Ramezani 2015) and forest edge length (Kleinn et al 2011, Ramezani 2017), which can be estimated from the same dataset.…”

“…For instance, Hunsaker et al (1994), Hassett et al (2011), and Ramezani and Grafström (2014 applied hexagon and square plots on satellite images; Ramezani et al (2010) and Lister et al (2019) used point grid on aerial photos; and Corona et al (2004) and Ramezani and Holm (2011) used line interest sampling (LIS) method on aerial photos. It is also possible to estimate some currently used landscape indices from field-based inventories such as National Forest Inventories (NFIs; Kleinn 2000;Ramezani and Ramezani 2015), but this possibility has received less attention.…”