Abstract:
Harvesting costs were determined for commercial thinning of young stands to
achieve vegetation and wildlife objectives. This included replicated comparisons of
thinning treatments. Treatments were defined based on residual tree stocking after
thinning. Study procedures were developed and evaluated to improve statistical
relevance. Multiple linear regression models were used to compare cycle times of thinning
treatments. Indicator variables were found to be effective in evaluating the treatment
effects. Extraction costs of harvesting and 95% confidence intervals were determined for
skyline yarding, tractor skidding and mechanized systems. There is not a marked difference in costs between treatments except the heavy thinning treatment of the tractor site. Also the light with openings treatment was more expensive in the mechanized forwarding. Skyline yarding costs are approximately double tractor skidding costs. Skyline costs are more sensitive to yarding distance than tractor skidding. Mechanized harvester costs are higher than manual felling costs. When extraction distance increases the harvester-forwarder system becomes less expensive than the felling-skidding system. Harvesting delay analysis for small and large delays shows that mean delay frequency and duration are not significantly different between treatments. Kolmogorov-Smirnov test results show that delay frequency and delay duration distribution of both small and large delays fit the Poisson and exponential curves respectively. The length of study needed to achieve reasonable delay percentage estimates are roughly equivalent to those needed to achieve delay free cycle time estimates. Simulation of delay demonstrated the need for long study periods to correctly estimate the large delays, which are tracked in shift level studies. A study length of 334 cycles for small delays and 140 days for large delays were predicted with 90% confidence intervals of ±2% and ±3.68% respectively. Comparison of detailed time study versus shift level study shows that although there is no significant difference in predicted means, the 95% confidence interval in shift level studies is much wider than the detailed. To attain an interval width of ±0.20 cycles per hour a detailed time study would be run for 4 days and a shift level could achieve this level only after 171 days.
