Jacek Kozak

Mapping Spatial Patterns with Morphological Image Processing

We use morphological image processing for classifying spatial patterns at the pixel level on binary land-cover maps. Land-cover pattern is classified as ‘perforated,’ ‘edge,’ ‘patch,’ and ‘core’ with higher spatial precision and thematic accuracy compared to a previous approach based on image convolution, while retaining the capability to label these features at the pixel level for any scale of observation. The implementation of morphological image processing is explained and then demonstrated, with comparisons to results from image convolution, for a forest map of the Val Grande National Park in North Italy.

Forest cover changes in the northern Carpathians in the 20th century: a slow transition

The reversal from deforestation to forest increase is named forest transition. It combines factors related to social and economic development or policies attempting to overcome the possible scarcity of timber. This study focuses on forest transition in the northern Carpathians, stretching across four countries (Czech Republic, Poland, Slovakia and Ukraine) with complex history and variability with respect to present-day economical development. Forest maps were computed for the 1930s and 1990s on the basis of available topographical and satellite data. Forest cover changes were then analysed and discussed in the context of population change and environmental conditions of the area. Between the 1930s and the 1990s net forest cover change was between 5% and 31%. The highest values (>20%) were recorded in areas which suffered post-war re-settlement and depopulation. Medium values (10–15%) were found in areas affected by the post-war gradual marginalization and depopulation. The lowest values (<10%) were noted in areas where population has increased with respect to the 1930s. In this latter case, gradual transformation of mountain agriculture related to overall social and economic changes has driven a slow forest cover increase. The forest cover increase was selective and occurred on areas less suitable to agriculture (higher and steeper). This selectiveness was less obvious in areas affected by re-settlement actions.

Forest Cover Change in the Western Carpathians in the Past 180 Years A Case Study in the Orawa Region in Poland

Abstract Since the 19th century a slow expansion of forests into previous agricultural areas has been recorded in the Carpathians. The present article analyzes forest cover change in the Orawa region of Poland, using historical maps and contemporary satellite data. Forest cover change was analyzed with reference to elevation, under the assumption that it reflects a transformation of the vertical land use system developed in the 17th and 18th centuries. For the past 180 years, the proportion of forest in the study area has increased from 25% to 40%. Forest expansion largely affected pastures cleared within the forest belt and areas located immediately above and below this belt. Changes in forest area were largely related to a decline in agriculture and have occurred along with population growth. As a result, grazing has been replaced by forestry, nature conservation, and tourism.

Impact of scale on morphological spatial pattern of forest

Assessing and monitoring landscape pattern structure from multi-scale land-cover maps can utilize morphological spatial pattern analysis (MSPA), only if various influences of scale are known and taken into account. This paper lays part of the foundation for applying MSPA analysis in landscape monitoring by quantifying scale effects on six classes of spatial patterns called: core, edge, perforation, branch, connector and islet. Four forest maps were selected with different forest composition and configuration. The sensitivity of MSPA to scale was studied by comparing frequencies of pattern classes in total forest area for various combinations of pixel size (P) and size parameter (S). It was found that the quantification of forest pattern with MSPA is sensitive to scale. Differences in initial composition and configuration influence the amount but not the general tendencies of the variations of morphological spatial pattern (MSP) class proportions with scale. Increase of P led to data generalization resulting in either a removal of the small size features or their potential transformation into other non-core MSP classes, while an increase of S decreases the MSP core area and this process may transform small core areas into the MSP class islet. We established that the behavior of the MSPA classes with changing scale can be categorized as consistent and robust scaling relations in the forms of linear, power, or logarithmic functions over a range of scales.

Global Change Research in the Carpathian Mountain Region

Abstract The Carpathian Mountains in Europe are a biodiversity hot spot; harbor many relatively undisturbed ecosystems; and are still rich in seminatural, traditional landscapes. Since the fall of the Iron Curtain, the Carpathians have experienced widespread land use change, affecting biodiversity and ecosystem services. Climate change, as an additional driver, may increase the effect of such changes in the future. Based on a workshop organized by the Science for the Carpathians network, this paper reviews the current status of global change research in the Carpathians, identifies knowledge gaps, and suggests avenues for future research.

European forest cover mapping with high resolution satellite data: The Carpathians case study

Abstract The aim of the study was to elaborate a methodology for forest mapping based on high resolution satellite data, relevant for reporting on forest cover and spatial pattern changes in Europe. The Carpathians were selected as a case study area and mapped using 24 Landsat scenes, processed independently with a supervised approach combining image segmentation, knowledge-based rules to extract a training set and the maximum likelihood decision rule. Validation was done with available very high resolution imagery. Overall accuracies per scene ranged from 93 to 96%. The labelling disagreement in overlapping areas of adjacent scenes was 6.8% on average. The proposed methodology is easily applicable to other regions in Europe. It allows an accurate forest–non-forest map to be obtained in a quick and cost effective manner, spatially more detailed than the currently available European data sets.

Neutral model analysis of landscape patterns from mathematical morphology

Mathematical morphology encompasses methods for characterizing land-cover patterns in ecological research and biodiversity assessments. This paper reports a neutral model analysis of patterns in the absence of a structuring ecological process, to help set standards for comparing and interpreting patterns identified by mathematical morphology on real land-cover maps. We considered six structural classes (core, perforated, edge, connector, branch, and patch) on randomly generated binary (forest, non-forest) maps in which the percent occupancy (P) of forest varied from 1% to 99%. The maps were dominated by the patch class for low P, by the branch and connector classes for intermediate P, and by the edge, perforated, and core classes for high P. Two types of pattern phase changes were signaled by abrupt transitions among the six structural classes, at critical P thresholds that were indicated by increased variance among maps for the same P. A phase change from maps dominated by the patch class to maps dominated by the branch and connector classes was related to the existence of a percolating cluster of forest, and the P threshold varied depending on the co-existence of the core class. A second phase change from the edge class to the perforated class was related to the existence of a percolating cluster of non-core (including non-forest) and represents a change of context from exterior to interior. Our results appear to be the first demonstration of multiple phase changes controlling different aspects of landscape pattern on random neutral maps. Potential applications of the results are illustrated by an analysis of ten real forest maps.

Modeling and Mapping of Soil Salinity with Reflectance Spectroscopy and Landsat Data Using Two Quantitative Methods (PLSR and MARS)

The monitoring of soil salinity levels is necessary for the prevention and mitigation of land degradation in arid environments. To assess the potential of remote sensing in estimating and mapping soil salinity in the El-Tina Plain, Sinai, Egypt, two predictive models were constructed based on the measured soil electrical conductivity (ECe) and laboratory soil reflectance spectra resampled to Landsat sensor’s resolution. The models used were partial least squares regression (PLSR) and multivariate adaptive regression splines (MARS). The results indicated that a good prediction of the soil salinity can be made based on the MARS model (R2 = 0.73, RMSE = 6.53, and ratio of performance to deviation (RPD) = 1.96), which performed better than the PLSR model (R2 = 0.70, RMSE = 6.95, and RPD = 1.82). The models were subsequently applied on a pixel-by-pixel basis to the reflectance values derived from two Landsat images (2006 and 2012) to generate quantitative maps of the soil salinity. The resulting maps were validated successfully for 37 and 26 sampling points for 2006 and 2012, respectively, with R2 = 0.72 and 0.74 for 2006 and 2012, respectively, for the MARS model, and R2 = 0.71 and 0.73 for 2006 and 2012, respectively, for the PLSR model. The results indicated that MARS is a more suitable technique than PLSR for the estimation and mapping of soil salinity, especially in areas with high levels of salinity. The method developed in this paper can be used for other satellite data, like those provided by Landsat 8, and can be applied in other arid and semi-arid environments.

Uncertainty in Historical Land-Use Reconstructions with Topographic Maps

Abstract The paper presents the outcomes of the uncertainty investigation of a long-term forest cover change analysis in the Polish Carpathians (nearly 20,000 km2) and Swiss Alps (nearly 10,000 km2) based on topographic maps. Following Leyk et al. (2005) all possible uncertainties are grouped into three domains - production-oriented, transformation- oriented and application-oriented. We show typical examples for each uncertainty domain, encountered during the forest cover change analysis and discuss consequences for change detection. Finally, a proposal for reliability assessment is presented.

Forest Cover Changes and Their Drivers in the Polish Carpathian Mountains Since 1800

In many developed countries, forest area has been increasing slowly since the nineteenth century, an increase that has been a clear reversal of forest cover trends in the previous period of dominant deforestation, referred to as the ‘forest transition’ (Mather 1992; Mather and Needle 1998). Agricultural land was abandoned or purposefully converted to forests, reflecting the economic and social development of societies, changing population pressures, and national policies attempting to overcome the scarcity of forest products (Mather 1992; Grainger 1995; Mather et al. 1999; Rudel et al. 2005; Chapters 2 and 3).