-224-

EVALUATION OF LANDSCAPE STRUCTURE CHANGES IN TATRA MOUNTAINS (POLAND) BASED ON THE 4D CIS ANALYSIS Piotr Wezyk1, Wiktor Traczb, Marcin Guzik' a

Laboratory of GIS and Remote Sensing, Faculty of Forestry, Agricultural University of Cracow, Poland; [email protected] b Department of Forest Management, Geomatics and Forest Economics, Faculty of Forestry, Warsaw Agricultural University, Poland; [email protected] e Tatra National Park, ul. Chalubinskiego 42, Zakopane, Poland; [email protected] ABSTRACT

New economic conditions and new approach to natural environment and the management of national parks cause changes of landscape and in particular its spatial structure. Spatial structure is related to spatial distribution of landscape elements (called patches) and their interaction. Spatial structure is determined by spatial (configuration) and aspatial (composition) features of patches. Defined structure of landscape supports specified functions or processes in the landscape. The main goal of presented research was to evaluate changes in the landscape, which had taken place during the 34 years-long period (1965-1999) in Dolina Sucha Stawianska and Dolina Bystrej (Tatra Mountains, Poland). Geoinformatic techniques such as: photogrammetric workout of archival aerial photos (1965 and 1999), DGPS measurement and 4D GIS analysis allow us to describe the changes in spatial structure. Several metrics (indexes) were used to quantify the landscape structure and to evaluate changes in the landscape. Key words: indexes of landscape structure, digital photogrammetry, spatial-temporal changes (4D), GIS, Tatra Mountains. 1 INTRODUCTION Increasing knowledge about forests, social needs and environmental concerns have been extending the scope of forest management [1]. This also changes landscape and landscape structure in particular [1], [2]. There are many definitions of landscape. Forman and Gordon [3] defined landscape as heterogeneous area of land composed of a cluster of interacting ecosystems, repeated in a similar form throughout. A forest landscape is a spatial mosaic of arbitrary boundaries containing distinct areas (patches) that functionally interact [4], The landscape is not necessary defined by its size, rather it is defined by an interacting mosaic of patches, relevant to the phenomenon under consideration. Landscape structure, or spatial structure, refers to the relative spatial arrangement of patches and interconnection among them [5], Jt represents both spatial (i.e. geographic arrangement) and aspatial (i.e. composition) character of landscape elements (i.e. patches). Landscape structure supports a number of functions or processes [3]. The disturbance effect on the structure and dynamics of a forest ecosystem depends on their spatio-temporal scale, physical environment and regeneration traits of the participating species in the communities [6]. In the Tatra Mountains, natural species characteristics and tree composition of forest stands as well as the vertical arrangement of the plant layers were considerably altered as a result of these disadvantageous processes.

-225-

Human activities in the Tatra Mountains, the highest part of the Carpathian range, have a long history. Though initially limited to shepherding on mountain pastures and clearings created by forest felling, it transformed into iron-ore mining and paper-milling industry and, consequently, into timber harvesting on a large scale. The upper tree-line of the subalpine forest (forest zone below tree-line) usually lies in the Polish part of Tatra Mountains at the altitude of 1550 m a.s.l some subalpine Norway spruce trees may be found at the altitude of 1650 m a.s.l. [7], In many places in Tatra the shepherds have pushed down the tree line by about 200 m or more [8]. With the establishment of the Tatra National Park (TPN) in 1955 with the area of 21,164 ha, nature was given a chance to return to its original condition. Today, recreational activities together with industrial imissions and climate changes make main factors influencing the landscape changes in upper parts of Tatra Mountains. Tatra NP is visited both in summer and in winter season by the high number of tourists. This number, however decreased in the last years from 3,3 million (1999) to 2,6 million in 2003. On the lower heights of the upper forest zone, the biotical factors (e.g. Ips typographus, fungal diseases) are the most important in affecting the dynamics of forest changes. Projects in natural sciences face growing demand for rapid data generation, which results in the increasing application of integrated geoinformatics technologies such as: Digital Photogrammetry, Geographical Information Systems (GIS), Digital Elevation Model (DEM); Global Positioning System (GPS) or Remote Sensing [9], [10]. These geoinformatic techniques make 3D spatial analyses possible, but if supplemented by the time factor (4D analyses), they allow determining the dynamics of changes within the natural environment [11], [12], [13]. The aim of the present paper was to determine the spatial-temporal (4D) changes of the landscape of the Tatra Mountains from 1965 to 1999, using geoinformation technology. 2 METHODOLOGY 2.1 RESEARCH AREA The investigation site was the area of 1797 ha situated in the Sucha Stawianska Valley (Dolina Sucha Stawianska) - a part of Sucha Woda Valley (Dolina Suchej Wody) and in the Bystra Valley (Dolina Bystrej) in the Polish Tatra Mountains (Figure 1). Altitude gradient in the study area was over 1000 m, therefore it was possible to analyze the plant cover at its different layers: from mountain mixed forest, through subalpine forest and dwarf mountain-pine belt to alpine vegetation. The study area is characterized by the prevalence of northern (N, NW, ME) and eastern aspects (E, SE), which total 70% of area under research. The southern aspects (S) and the southwestern ones (SW) occupy less than 10 % of the area [14]. 2.2 MATERIALS AND METHODS The first stage of the analyses was the data gathering and data processing, conducted with the use of geomformation technologies, e.g. Digital Photogrammetry, Geographical Information System (GIS) and Global Positioning System (GPS). Photointerpretation workout was based on 6 stereo models of the B&W aerial photos from the year 1965 (scale 1:16000) and 4 stereogrammes of color photos from 1999 (scale

-226-

1:26000). Slide scanning was done with the 1000 dpi resolution. The digital photogrammetrical station Video Stereo Digitizer (VSD-AGH) was used for the stereodigitizing of the 3D objects. Following the field work aimed at obtaining a sufficient number of ground control points (GCP) and establishing a photointerpretation key, a stereogram orientation was made and 23 object categories were defined. Finally, polygon vectors obtained in the course of photointerpretation and 3D digitalisation were assigned [15]. The VSD-AGH station made it possible to display simultaneously 3D vector (e.g. digitized from the 1965 stereo-images) in a stereomodel, which resulted in improved correctness of the photogrammetric workout. Data obtained in a vector format were exported to GIS software (Arclnfo ver 8.1 and ArcView ver. 3.2a ESRI) where their topological correctness was verified

Figure 1. Research area (red polygon) in the Tatra National Park, Poland Two cartographic GPS Pathfinder ProXRS (Trimble) receivers equipped with TSC1 data logger were used for the field survey. The DGPS measurements (WGS84) were done on the basis of the differential mode in real-time (OmniStar) or in the post-processing mode (base station Trimble in TPN-Zakopane). To carry out the 3D spatial analyses, the Digital Elevation Model (DEM) was generated (TIN model) based on the contour lines and mass points [14]. TIN was transformed into a form of regular raster (GRID; ESRI) with the pixel size 5,0 m. Next stage of the analyses was evaluation of landscape changes from 1965 to 1999. The analyses consisted of the comparison of the structure and composition of the landscape in two moments of time. Many metrics (indexes) of landscape structure were defined and used

-227in order to quantify landscape patterns, measure and evaluate changes in it. Several adequately chosen metrics are usually used, depending on a need. Metrics fall into two general categories: those that quantify the composition of a landscape without reference to spatial attributes, and those that quantify the configuration of a landscape requiring spatial information for their calculation [16], [17].

s

The basic measures of composition are: • proportional abundance of each patch type, • patch richness, • patch evenness, • patch diversity. The basic aspects of the configuration and the simplest of representative metrics are: • patch size distribution and density, • patch shape complexity, • core area, • isolation/proximity, • contrast, • dispersion, • contagion and interspersion, • subdivision, • connectivity. Individual patches posses relatively few fundamental spatial characteristics, e.g. size, perimeter and shape. Collection of patches may have a variety of aggregate properties, depending on whether the aggregation is over a single patch type or multiple classes, and whether the aggregation is within a specified subregton of a landscape or across the entire landscape. Commonly, landscape metrics can be defined at three levels: patch level, class level, and landscape level. The presented paper contains the results of spatial-temporal (4D) GIS analysis done for only 2 from 23 categories of objects (i.e. Norway spruce and dwarf mountain pine) chosen from the photogramrnetric workout. Their choice was dictated by meaning and size of changes, which happened in the area of investigation over 34 years (1965-1999). Indexes were calculated for class level and quantify spatial characteristics of the landscape. Patch Analyst software, the extension for ArcView 3.2, was used to calculate several indexes [18]. The following indexes were chosen for the analyses: • • • • •

•

Class Area (CA) - sum of areas of all patches belonging to a given class; Number of Patches (NumP) - number of patches for each individual class; Total Edge (TE) - perimeter of patches; Mean Patch Size (MPS) - average patch size. MPS>0; Mean Shape Index (MSI) - shape complexity. MSI>1. MSI=1 when all patches are circular (vector) or square (grid), and its value increase when complexity of patches increase; Area Weighted Mean Patch Fractal Dimension (AWMPFD) - shape complexity adjusted for shape size. 1