Status and trend analysis in landscape pattern through field-based sampling data

Abstract

Traditionally, calculation of landscape metrics is commonly conducted on land cover/use maps of entire landscape which is created from remotely sensed data. An interesting approach, however, is to make use of sample data, without the use of wall-to-wall mapping. In the present review and case study, it is aimed to estimate three basic landscape metrics, namely Shannon’s diversity (SH), forest edge length (E) and contagion (C) from field-based sampling data. It is also intended to estimate landscape change using time series datasets. Estimated variance (sampling error) was used to assess landscape metric estimators. For this purpose, sampling data from National Inventory in the Landscape of Sweden (NILS) is used. In this case study, the metrics are estimated with acceptable precision. In most cases, the estimated variance (sampling error) was less than 10 %. The largest sampling error was 28 % for forest edge length. We will be able to compare different landscape at a given time or a landscape over time using filed-based sampling data. Furthermore, in an ecological survey it may be possible to find a relationship between landscape pattern and ecological processes such as biodiversity. The methods applied in this study is very simple and there is no need for extra measurements.