Modelling Viability of Biowood Production from Private Woodlands

Fiona Simpson

Background

The demand for biowood, or wood fuel, in Scotland is growing as there use to help reach Climate Change targets is encouraged. This growing demand has created a new market for locally produced biowood products (FC 2014) and with increasing biowood product prices (FC 2014) landowners can look to this new market as a fresh form of income (Goor et al 2000 and Hall 2005), by harvesting persisting woodlands. Currently it is estimated that there is 330 million ODT of biomass held in private sector woodlands (FC 2014) and it is these pre-existing, mainly under-managed, woodlands which could be a major source of biowood products (FC 2002). By utilising this stock the land owner would generate income from producing much needed biowood without the risks associated with energy crops (Styles et al 2008). There are few models currently available that look to determine the economic viability of pre-existing woodlands, the few that do are implemented through Microsoft Excel accounting for aspatial factors not spatial factors such as distance to market and slope. By not having a comprehensive model, which takes account of both spatial and aspatial factors, landowners will not be supported in the decision whether to produce biowood and they will be unsure of the potential economic gains they could see by producing biowood (Clancy et al 2012). This is then leaving a gap between the demand and the production of biowood.

Research Question

To overcome this this project looked to answer the following Research Question:

Can a valid model be built to calculate the economic viability of woodlands in terms of biowood production,

which accounts for both spatial and aspatial factors?

Creation and Testing the Model

The model calculates the economic viability of woodlands for creating biowood using Net Present Value (NPV). NPV is described as the present value of revenues subtracted from the present value of costs (Holopainen & Talvitie 2006). If NPV is positive the profits are greater than the investment (Toivonen & Tahvanainen 1998) and as such the woodland is economically viable for the creation of biowood. The NPV will table account of the factors shown in Table 1 which have been shown to affect the economic viability of woodlands from the current literature and dissections with experts in the field. Table 1 shows that many factors govern viability; a landowner has many choices to make in terms type of harvesting and type of product created. As such there are many options which can be taken to produce biowood, these are summarised in the form of a decision tree, Figure 1. The NPV of each option is calculated by the model to see which one is best.

Figure 1 Decision Tree Table 1 Factors Included in the Model

The model was implemented through ArcGIS’s ModelBuilder. ModelBuilder was chosen as it allows for the same spatial geoproccesing operations and Python scripts to be rerun any number of times. As such the model can be run on different woodlands by only changing the woodland being investigated. Along with the ease of running the model for different woodlands. The model created is made up of a main model and a number of smaller sub models. Figure 2 describes the model constructed.

Figure 2 Model Created

The model creates two outputs, firstly a map with ArcGIS’s ‘Layout View’ and secondly a Text File showing the NPV’s for each of the eight options and all the Capital and income values calculated. How these appear are shown in Figures 4 and 5 for a test case woodland.

Figure 3 Model Out Put 1: Map

Figure 4 Model Out Put 2: Text File

Model Validation

To validate the two forms of Sensitivity Analysis and expert Open Valuation were undertaken.

Monte Carlo Sensitivity Analysis was undertaken to determine the sensitivity and uncertainties associated with the model as a whole. The results of this analysis showed that the relative uncertainty of the NPV’s were low but due to the very high values being calculated the actual uncertainties averaged £18,000. In a number of cases this difference could account for a woodland being deemed viable and unviable for the creation of biowood.

Sobol Sensitivity Analysis was undertaken to determine the sensitivity and uncertainties associated with each of the parameters of the model. The results of this analysis showed that the three parameters uncertainty affected the overall models uncertainty the most are, the density of trees to clearfell and the selling price of logs and chips. This shows that the potential difference in the values for these parameters greatly affects the woodland economic viability.

Expert Open Valuation Analysis was undertaken to determine whether experts in the forestry and biowood domain found the model created useful and valid against the research question it set out to answer. The results of which showed that while there were a number of factors which are important which are not considered by the model (such as windthrow, varying prices for machinery and assuming all trees will be most profitable to create biowood rather than other products) they found it useful and valid. The experts found the model very good at incorporating spatial factors which in their current models cannot be undertaken. The large uncertainties associated with the model meant that the excerpts felt that the model created in its current form does not surpass the models they currently use and as such if the model created here was to show a woodland as being viable to harvest and create biowood then the models and field work currently used should then be completed to demine a more exact value which a woodland has. Overall the experts found that the model is partially good at showing quickly and efficiently showing landowners the potential value a woodland has.

Conclusion

This project looked to answer the question ‘can a valid model be built to calculate the economic viability of woodlands in terms of biowood production, which accounts for both spatial and aspatial factors? By creating a model in ArcGIS ModelBuilder incorporating Python scripts and using SA and OV to validate it, it has been proved that such a model can be built. While the uncertainties within the model are high, the OV shows that when the model is used to show the value of a woodland before undertaking the expensive standard methods of calculating economic viability, then it is valid and very useful. OV also showed that this first version of such a model is a very useful building block to add to the complex topic in the future.

The model created here is consequently very useful in quickly and efficiently showing landowners the potential value woodland on their property has. It is hoped that from this model that landowners will take more interest in their undermanaged woodland and make use of the woodland management opportunities from the BMR, including in depth field work to estimate the value of a woodland’s timber both for the biowood and also for other appropriate markets.

References

Clancy.D Breen.J.P, Thorne.F & Wallace.M (2012) A Stochastic Analysis of the Decision to Produce Biowood Crops in Ireland, Biowood and Bioengery, 46, 353-365

Forestry Commission (2002) Woodfuel Production from Small, Undermanaged Woodlands, Forest Research

Forestry Commission (2014) UK Wood Production and Trade: 2013 Provisional Figures

Holopainene.M & Talvitie.M (2006) Effect of Data Acquisition Accuracy on Timing of Strand Harvests and Expected Net Present Value, Silva Fennica, 40 (3), pp 531-543

Styles.D, Thorne.F & Jones.M.B (2008) Energy Crops in Ireland: An Economic Comparison of Willow and Miscanthus Production with Conventional Farming Systems, Biowood and Bioenergy, 32, 407-421

Toivonen.R.M & Tahvanainen.L.J (1998) Profitability of Willow Cultivation for Energy Production in Finland, Biowood and Bioenergy, 15(1), pp 27-37