Modern forest machines are highly effective but their availability is reduced through frequent relocation. Relocation has been estimated to constitute between 6-20% of the delivered roadside cost of cut-to-length (CTL) timber. Machine utilisation is increased when relocation frequency is reduced (larger stands), and when relocation distances are shorter. The geo-spatial structure of forests at a stand and landscape level is therefore assumed to play a role in setting the efficiency threshold of modern harvesting systems. It is further assumed that this effect varies between regions and forest ownership patterns, and that the extent of the effect is quantifiable.

Testing this assumption, the size and mutual distance between 29 000 coniferous stands constituting some 70,000 ha and divided into 4 machinery management regions in Denmark was analysed using single-linkage cluster analysis. Furthermore, benefits of using the shortest path algorithm to schedule machine deployment in an optimal way were compared with a fully randomised (customer-oriented) deployment in a simulated environment. Finally, a comparison of the advantage of sandwiching multiple (3) years of scheduled thinnings into 1 package were compared with the re-deploying of machines across the region every year.

Results showed that the geo-spatial structure at landscape level mean distances between clusters ranging from 49 km in region East, to 90 km in region North. Weighting clusters with stand size reversed this ranking, where the mean distance in North was reduced from 90 km to 17 km. This highlights the importance of using the correct statistic in planning. Furthermore, when comparing a fully randomised relocation with shortest path scheduling, the mean relocation distance in region East was reduced from approximately 49 km to under 5 km, increasing productivity for a single machine set by 900 m3 a-1. This increase was slightly larger when 3 years of thinnings were grouped into one planning parcel as compared to deploying across the whole region every year. Finally, proper scheduling of relocation was shown to be of increasing importance with increasing machine productivity.

Findings are considered to have important connotations for both the layout of administrative forest areas, and the manner in which machines are deployed. The clustering method used proved a powerful tool for generating packages of stands, for e.g. a tendering process, for finding an appropriate number of machine systems to cover a region, and for using as a method to evaluate the performance of harvesting systems, and the effectiveness of machinery managers and machine operators within these regions.

Keywords: Machinery management, spatial planning, forest topology, cluster analysis, simulation