Geostatistical approach to spatial analysis of forest damage
; »Hrvatske šume« d. o. o. Zagreb, Ljudevita F. Vukotinovića 2, HR-10000 Zagreb, Croatia
; Faculty of Forestry, Department of Forest Inventory and Management, Svetošimunska 25, 10000 Zagreb, Croatia
APA 6th Edition KLOBUČAR, D. i PERNAR, R. (2012). Geostatistical approach to spatial analysis of forest damage. Periodicum biologorum, 114 (1), 103-110. Preuzeto s https://hrcak.srce.hr/80904
MLA 8th Edition KLOBUČAR, DAMIR i RENATA PERNAR. "Geostatistical approach to spatial analysis of forest damage." Periodicum biologorum, vol. 114, br. 1, 2012, str. 103-110. https://hrcak.srce.hr/80904. Citirano 25.09.2020.
Chicago 17th Edition KLOBUČAR, DAMIR i RENATA PERNAR. "Geostatistical approach to spatial analysis of forest damage." Periodicum biologorum 114, br. 1 (2012): 103-110. https://hrcak.srce.hr/80904
Harvard KLOBUČAR, D., i PERNAR, R. (2012). 'Geostatistical approach to spatial analysis of forest damage', Periodicum biologorum, 114(1), str. 103-110. Preuzeto s: https://hrcak.srce.hr/80904 (Datum pristupa: 25.09.2020.)
Vancouver KLOBUČAR D, PERNAR R. Geostatistical approach to spatial analysis of forest damage. Periodicum biologorum [Internet]. 2012 [pristupljeno 25.09.2020.];114(1):103-110. Dostupno na: https://hrcak.srce.hr/80904
IEEE D. KLOBUČAR i R. PERNAR, "Geostatistical approach to spatial analysis of forest damage", Periodicum biologorum, vol.114, br. 1, str. 103-110, 2012. [Online]. Dostupno na: https://hrcak.srce.hr/80904. [Citirano: 25.09.2020.]
Sažetak Background and Purpose: Significantly increased forest damage has recently been observed in the Republic of Croatia, as well as increased proportion of unplanned felling in prescribed cuts, which has negative repercussions for sustainable management. The objective of this study was is to explore the possibilities of simple and reliable detection, inventorying (mapping) and monitoring forest health condition by means of color infrared (CIR) imagery and geostatistical methods. Materials and Methods: Four trees (crowns) closest to the point of the raster (100 ´ 100 m) which was set up in the digital orthophoto for the area, were interpreted in CIR images. Forest damage indicators, mean damage and damage index were calculated for the whole area under observation. The assessment and identification of spatial distribution of these damage indicators were performed using raster point data, from which a random (966 points) and a systematic (445 points) sample were created. The results on forest damage acquired by interpreting CIR images were used for geostatistical analysis. A model of theoretical semivariograms provided parameters which were used for interpolation of both damage indicators with ordinary kriging. Continuous maps of damage degree distribution were then constructed. The results of interpolation were tested with the cross-validation method. Results and discussion: Damage indicator maps are the result of the following: data variability, sampling intensity and method, form of experimental and theoretical semivariograms which were subsequently used to compute kriging matrices, method of selecting a particular semivariogram, assessment accuracy, the choice of interpolation methods (kriging, cokriging, stochastic simulation, inverse distance, etc.). Tree damage generally does not have regular, but rather random spatial distribution. This is why the primary aim in identifying forest damage is to incorporate the whole area of interest into sampling. Sampling intensity should be adapted to the required accuracy and to the time and funds at our disposal. Conclusions: This research relies on the application of CIR aerial photographs and geostatistical tools in spatial analysis of forest damage. Continuous maps of damage indicators acquired with kriging provide a better insight into the spatial distribution of damage than do thematic maps obtained by interpreting CIR aerial imagery on the basis of a systematic sample (the raster method). Integration of interpretation results of CIR aerial images and geostatistical approach ensures a more precise distribution of damage indicators, and consequently, the possibility of better spatial analysis of the occurrence, trends and development of damage in the study area.