Risk assessment of forest decline by application of geostatistics and multi-criteria analysis

Abstract

In this paper, the risk zone mapping of declining lowland forests belonging to "Morovia", management unit "Varadin-Zupanja" (northern Serbia) is performed using geostatistics analysis. Based on the monitoring of groundwater level, the Kriging method has been performed for the spatial distribution of groundwater level for a multiyear period (2010-2013) - reference level and characteristic levels for the wettest and the driest year during the analyzed period. Risk assessment was determined by the variance of characteristics compared to reference levels. Then, multi-criteria decision analysis methods (AHP, PROMETHEE II) were applied to define the rank of each department (smaller forest management units) located in the research area. These analyses are very important because they enable to locate of the area with a high risk of forest decline and to rank departments using criteria: deviation from water level recorded during dry periods, species demand for water, conservation status and purpose of the unit (seed stands or technical wood). The proposed methodology is usable for the determination of the primary localities for the application of management measures conducting on the level of lower planned units (departments) and thus lead to the successful planning and more efficient forest management. Obtained results at the researched area showed that a negative influence on the watering regime has groundwater level decreasing compared to the reference level because it directly affects available water for the plants. Based on multicriteria analysis methods, it was deduced that the most endangered parts are located at the edge, while this risk is much lower in the central part of the management unit. A combination of applied methods (geostatistics and multicriteria analysis) is of great importance for forestry management.