Croatian Journal of Forest Engineering (Apr 2011)
Features of Spatial Snag Distribution in a Beech-Fir Forest
Abstract
With the participation of about 65% in the total growing stock of conifers, silver fir is the most important and the best represented conifer species in Croatia. The increasing share of sanitary felling in the annual felling plans demands that special attention be paid to the health status of fir forests and the quantity of unplanned felling. This is the reason why dieback of silver fir in beech-fir forests in the Dinaric Alps is an acute and topical management-ecological problem. In the past 10 years, the share of total unplanned felling in the overall annual felling for fir has amounted to 57%. Such a high share of unplanned yield (mainly snags) has had a negative effect on sustainable forest management. Determining the spatial distribution of damaged trees and snags, as well as their monitoring is one of the priorities of sustainable management. Hence, it is of utmost importance to detect less healthy stands and apply timely measures for the purpose of maintaining their vitality and productivity at an optimal level. Due to the seriously disturbed stand stability caused by forced canopy opening, which in turn affects the health condition and natural regeneration of the stand, it is necessary not only to detect snags but also determine the causes of dieback. According to the results of past research, increased tree dieback is associated with the impact of different site and stand characteristics (altitude, exposition, slope, soil, structure, etc.), as well as abiotic and biotic factors (fir needle moth, mistletoe); however, the real causes are yet to be discovered. It is difficult to make efficient snag inventories with standard field methods. For large areas, a remote sensing method (the application of CIR aerial photographs) is much more practical, more cost effective and more reliable. All research activities so far have proved that, in terms of accuracy, this method is equal to field working methods, while in terms of speed and objectivity it is far more efficient. Snag inventories by means of CIR aerial photography interpretation and GIS enable efficient enumeration (detection) and spatial distribution (mapping) of snags. The proposed method is highly important not only for the detection and monitoring of the occurrence, distribution and direction of dieback progress, but also for the detection of possible dieback causes, on the basis of which experts will prescribe adequate measures and treatments. The objective of CIR aerial photography interpretation is to determine the current situation and acquire an objective picture of forest condition (database) in as short a period as possible, while mapping on a color infrared digital orthophoto (DOP) is intended to provide the spatial distribution of snags for the surveyed area. An aerial survey of the Lika area with the CIR film (Forest Administration Gospić) was conducted in strips (Fig. 3). The coordinates of the start and finish points of aerial photogrammetric series in the associated zones of the Gauss-Krüger cartographic projection, as well as planned strip lengths, are given in the following survey (Table 1). A total of 88 aerial photographs were taken at an average scale of 1:6000 and longitudinal overlap of 60%, which allows for the interpretation of individual trees in the stereomodel. As the total length of all the three strips amounted to 40,243 m (40.24 km), an area of 5,548 ha was surveyed. Since aerial photography provided an insight into the condition of the stands in the investigated area, and since the results obtained by photointerpretation of CIR aerial photographs indicated acute fir damage (the average damaged tree is 61.70%), snag inventory (mapping) was undertaken. In the localities of fir forests, dieback was observed in the form of smaller or bigger groups of trees. Hence, areas of 1 ha (sample plots) were delineated on a random sample in the color infrared DOP (1:5000). All visible snags were then mapped in each area (Fig. 1). The central parts of the images were used to construct a DOP. The pixel size of each orthophoto is 0.5 m. According to the spatial snag distribution, the number of snags (density) per strip ranged between 23 and 105 per ha (Fig. 2). Since the values were not similar for all the plots, all the identified snags were mapped in the color infrared DOP. A total of 17,445 snags were mapped in the study area (Fig. 4). GIS was used to produce a cartographic presentation of spatial snag distribution along the survey strips. In order to produce a digital orthophoto, it was necessary to construct a digital relief model. Data for the DRM were obtained by vectorizing the contour lines from the associated sheets of the Basic Map of Croatia (BMC) of 1:5000 scale and topographic maps (TM25) at a scale of 1:25000. Since individual tree mortality is a stochastic, rare and irregular event (Eid and Tuhus 2001) and since this is a karst terrain, slope variability in the study area is very high; therefore, it is not advisable to take into account the average slope in an area. As the same was also confirmed by the research of Antonić and Legović (1999), Pernar (1997), and Božić et al. (2004), the layers were generated from the DRM. Raster-GIS modeling provided the layers of slope and exposition. In this way, the values of altitude (Fig. 5), slope (Fig. 6) and exposition (Fig. 7) obtained by means of the DRM were assigned to every snag, in dependence of the location (position in space). An orthophoto and a digital relief model (DRM) were also entered into GIS as one of the layers, which enabled rapid spatial analysis and an efficient access to data (Fig. 8). Next, the obtained results were statistically analyzed and the data of spatial snag distribution and geomorphological features were processed. The most important result ensuing from the presented method of rapid snag inventory is the thematic maps, which present spatial snag distribution for the investigated area. Not only is field work very expensive and long-lasting (Boyle et al., 1998), but it is also extremely difficult in barely accessible terrains or on steep slopes. A method based on remote sensing (aerial or satellite images) considerably contributes to the solution of the afore mentioned problems and provides a more practical, less expensive and more reliable manner of enumerating dry trees and stands over large areas. In combination with GIS technology, the production of a map of spatial snag distribution takes 5 – 10 minutes per hectare of the inventoried area. Such maps provide an insight into the field condition and spatial snag distribution in a very short time. They can also be used as a basis for the calculation of mean snag density and for conducting a variety of analyses intended to predict efficient measures of mitigating dieback effects. Based on the results of statistical analysis, it can be concluded that the spatial distribution of snags is influenced by geomorphological features (terrain orientation, slope and altitude). Variance analysis (F=278.79; p<0.001) showed that there was a statistically significant difference between the spatial distribution of snags with regard to terrain orientation (exposition) and slope. According to the results, the number of snags increases in accordance with an increase in the slope on the north-eastern and northern exposition (Fig. 9). The Tukey HSD test proved that, in terms of slope, all terrain orientations are clearly differentiated in relation to the occurrence of snags, except the S and SE. This was expected, since the participation of fir in these expositions is lower. As for terrain orientation, there is also a statistically significant difference in terms of altitude (F=102.6; p<0.001). The results show that the number of snags in the E and NE exposition rises with an increase in altitude (Fig. 10). The results of statistical analysis should not be used to make one-sided conclusions based on the correlation between fir snag distribution and mezorelief features. Undoubtedly, however, the spatial distribution (location) of each single snag is conditioned by edaphic and micro-relief site characteristics (soil depth, skeletalness), i.e. total soil capacity for water retention in the rhizosphere zone. Pernar (2001) directly associates the productivity of fir stands with soil depth, while Bigler et al. (2004) use the increment of mature fir trees as a predictor of fir dieback. Research by Certini et al. (2000) also confirms soil depth as a direct and indirect (distribution and virulence of pathogenic fungus Heterobasidion annosum (Fries.) Bref.)) factor of fir tree mortality. Taking into account the obtained results, a rapid and economic snag inventory is a must both for the needs of forest management and protection and for the needs of other scientific disciplines. At the Fourth Ministerial Conference on the Protection of Forests in Europe, held on 2 October 2003, it was proposed that dry trees should be used as a new indicator of biodiversity, or in other words, that snags should have special ecological significance for biodiversity within forest ecosystems. It is, therefore, predicted that snag inventories will be undertaken with increasing frequency. In accordance with this, the proposed method could become an irreplaceable tool and a complement to standard field method. In addition, it has versatile applications, including national, regional or local forests taxation, certification processes, studies of site quality for animal species which depend on dry trees; classification of older stands, etc. The obtained results not only provided the current forest condition (inventory), but also indicated potential (focal points) and purpose of aerial surveys. Aerial photographs provide a survey of the field condition in a short time period, as well as allow for the possibility of planning efficient measures of mitigating dieback effects and evaluating dieback probability and rate. The results of the analyses enable the study of particular environmental features and their effects on forest dieback, as well as indicate the direction of future multidisciplinary research involving complex analyses of all environmental features. The results of the research are applicable primarily in the field of forest management, protection and harvesting, but they can also provide a basis for interpretations of events occurring in other fields.
