Academic journal article Geopolitics, History and International Relations

Evaluation of Digital Data Layers for Establishing Natural Landscape Types in Slovenia

Academic journal article Geopolitics, History and International Relations

Evaluation of Digital Data Layers for Establishing Natural Landscape Types in Slovenia

Article excerpt

ABSTRACT.

We have evaluated the digital data layers according to their relevancy in establishing the natural landscape types of Slovenia. To find out which digital data layer (e.g. slope, temperature regime) can sufficiently reveal differences among potential Slovenian landscape types, each layer was compared with the map of landscape-ecological types that was produced by a group of geographers. We used different methods (e.g. gain ratio, Mann-Whitney U test, Spearmann correlation coefficient, decision tree) to select and to test the most appropriate data layers.

Keywords: geography, landscape types, quantitative methods, evaluation of data, geographic information systems, Slovenia

Introduction

The classification of Slovenia into spatial units has had a long tradition in the Slovene geography. Many classifications are the result of human attempt to categorize and organize life (Haggett 2001). This is one of the reasons why there are several landscape classifications of Slovenia regarding the physico-geographic and socio-geographic factors (Melik 1935; Melik 1946; Ilesic 1957/58; Ilesic 1972; Gams 1986; Natek 1993; Gams, Kladnik and Orozen Adamic 1995; Ogrin 1996; Perko 1998; Plut 1999; Spes et al. 2002). The most predominant among classifications are regionalizations, while typifications occur on fewer occasions. Typification is based on the principle of similarity, while régionalisation is based on aspects of individualism and particularity (Geografija 2001). In case of régionalisation, we deal with regions, which are unique and are written with a capital letter, whereas types are spatially separated units with similar characteristics and are written without capital letters.

Computers enable the use of various quantitative methods and a great number of data. Additionally, many researchers (Kladnik 1996; Hargrove and Hoffman 2005) focus on the fact that the field of classifications should be based on computer analyses, statistical methods, etc. Several examples of the use of geographic information systems (GIS) in landscape classifications are known in Slovenia (Perko 1989; Perko 1998; Perko 2001; Repe 2006; Ferreira 2006; Breskvar Zucer and Marusic 2006) as well as across the world (Soto and Pintó 2010; Jongman et al. 2006; Metzger et al. 2005; Hargrove and Hoffman 2005; Wolock, Winter and McMahon 2004).

Scope

The article includes the review of digital data layers for Slovenia, as well as the evaluation of their usefulness in establishing the natural landscape types via the GIS technology. In order to evaluate the suitability of data layers, a comparison was made with the current landscape-ecological types of Slovenia (Spes et al. 2002). We assume that the geographical findings related to this topic contributed to a suitable typification of Slovenia. Therefore, the existing classifications (without GIS) may serve as a reference for the evaluation of relevancy of different data layers regarding their use for new GIS-based classifications. For establishing and mapping the landscape nature, human experience and the knowledge of landscape play a vital part (Groom 2005). Zhou et al. (2003, 458) pointed out the suitability of factors and their importance regarding the delineation of eco-regions.

It should be pointed out that the best way of collecting data is based on scientific knowledge of the issue at hand. In this respect, automatic methods can be as useful as any other. The choice of right data improves the functioning of various algorithms (Witten and Frank 2005, 289). Examples of determining the importance of individual natural factors have already existed from the land use research field; Perko (1987), for example, centered his research on discovering the factors that affect the land use the most. Leathwick, Overton and McLeod (2003), in the classification of New Zealand, selected the data on the basis of statistical correlation with the distribution of tree species as well as their relevancy for the physiological processes of trees. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.