Graduate Project

Forecasting harvesting production rates using GIS : a case study

Public Deposited

Downloadable Content

Download PDF


Attribute NameValues
  • Operational forest planning is characterized by a lack of formal planning often using only the intuition and experience of the forest planners. There are a number of sources of variability found in operational planning. Like most businesses there is significant variability in the demand forecasts obtained from customers. Forestry differs from many other businesses with significant variability in the supply due to the high sampling errors commonly used in forestry to in the statistical prediction of the volume, the non uniform distribution of the trees in the stand and finally the variability regarding the harvesting production. The aim of this project is to develop some the tools and processes to improve the forecasting of logging production at both the unit and daily levels. Data was collected from the Oregon State University student logging crew operating on the McDonald-Dunn forests. For the unit forecasting model, an existing production model from the literature was used, since no unit model was developed in this study due to the small data set and the limited timeframe. This statistical model was selected because it uses a prediction variable that can be estimated prior the operation is executed. For the daily production forecasting level, the following data was collected: location map, number of pieces, crew size, hours worked and weather conditions. This data was then analyzed using linear regression analysis The statistically significant variables found in this study to predict daily production were skidding distance, average temperature and hours worked. The model was able to explain 76 % of the variation in the daily production from the sampled area. One of the significant variables found from this study was skidding distance. A decision support system was developed using GIS techniques that easily measure this variable. The GIS system found the path from the landing to the stump by applying a shortest path algorithm to minimize the total cost incurred in the operation. Both the unit and daily level forecasts were applied to two scenarios to demonstrate the system. The first scenario forecasts the production using the actual stream pattern, and the second scenario uses a high density stream pattern that in some cases affects the path distance. The decision support system was able to capture these differences on the forecasted productivity.
Resource Type
Date Available
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Academic Affiliation
Rights Statement



This work has no parents.