Forestry Commission logo

Yield modelling & forecasting

Annual Summary of Research
1 April 2003–31 March 2004


Dynamic yield models for Sitka spruce

Research and development of the dynamic model of growth and yield of Sitka spruce stands came to a resolution at the beginning of the year with the preparation of a version of the model as a computer software package. The model is capable of representing an extremely wide range of silvicultural systems and options including:

  • Continuously-variable ‘yield class’
  • Wide range of planting spacings
  • Presence of variable proportion of multistemmed trees
  • Up to two initial systematic thinning operations (row and/or across-row)
  • Almost unlimited combinations of silvicultural/economic thinning interventions defined in terms of type and intensity.

The software produces forecasts of yield in a format similar to FC Booklet 48 but in a way that can be manipulated by the user, particularly in terms of forecast intervals. The model has already been integrated with an existing assortment forecasting program (ASORT) that enables yield forecasts to be expressed in terms of particular wood products as specified by the user. A facility has also been included to enable the user to enter stand inventory information as input data, in which case the model provides projections of future growth and yield consistent with the stand information provided.

Providing some of the facilities in the final version of the model has involved some innovative research and development. A notable example of this is the way in which the model simulates row and across-row thinnings. Fundamentally, the model is an example of a ‘distance-independent’ growth simulator which means it does not represent spatial relationships between individual trees. However the model is able to simulate patterns of systematic thinning by representing stands as a collection of explicit spatial components. In principle, this feature enables the model to represent both spatial and non-spatial variation in stands very efficiently. This offers an opportunity for further development to simulate the growth of stands with diverse structure, including mixed-age and/or mixed-species stands.

The Sitka spruce dynamic model can be applied in a research context as well as commercially, and opportunities are being explored for application of the model in other research areas, notably wind stability, timber quality, stand establishment and disease impacts.

The current version of the model is being tested in preparation for publication.

Formation of Yield Model Evaluation Team
Following on from the results of a review by selected users and experts of a beta-release version of the Sitka spruce dynamic yield model software, it was decided to form a small team of industry stakeholders to carry forward a more formal evaluation. A Yield Model Evaluation Team was duly convened with representatives of model users from Forest Enterprise (Operational Support Unit and Buchan, Galloway and Kielder Forest Districts) and the private sector (UPM-Kymmene/Tilhill). In addition Prof. Barry Gardiner joined as an expert external to the yield modelling project team to advise on modelling and software development. The Team held two formal meetings and there was also a round of on-site visits in the summer to ensure that software had been installed correctly and to assist with the evaluation. As part of this exercise, Team members have been testing the dynamic models against field observations, either collated from existing records or obtained from small purpose-designed field surveys. Operational Support Unit is also carrying out FE district-level forecasts to ascertain any potential impacts or implications of introducing the new models into general use. Team members have commented on the design of the user interface to the dynamic models and so far only minor changes have been requested. The most important of these has involved including an option to specify an initial stand density as an alternative input variable to planting spacing, so as to better represent stands established by regeneration. This and other facilities are being incorporated into the software prior to release of an updated version for further evaluation by the Team. The group has also been advising on a number of key technical issues, for example how yield class should be defined in the new models.

Software version of FC Booklet 48
At the beginning of the year a software package was developed for accessing and manipulating the yield models published in FC Booklet 48. This includes a number of facilities that have been long-requested by model users, notably:

  • A calculator giving General Yield Class based on species, stand age and top height
  • Automatic interpolation of 5-year table entries to give annual estimates of yield
  • Export of yield table estimates direct to Excel spreadsheet.

Over the course of the year, the software has been tested by the Yield Model Evaluation Team. Important feedback has been received particularly with regard to the design of the user interface where some significant revisions have been requested. A revised version of the software is under development with the aim of producing a final version early in 2004/05.

The intention is to publish the software simultaneously with the Sitka spruce dynamic yield models, so that updated and consistent tables for Sitka spruce can be included. One small but important issue to be resolved concerns the brand labelling of the two software packages to clearly distinguish them and the applications for which their use is appropriate.

Improved FC Booklet 48 models
Research and development has started on a more flexible, computer-based realisation of the models published in Booklet 48. The aim is to provide software that will enable the user to generate models for a much wider range of planting spacings and thinning cycles, types and intensities than represented by the current collection of tables. The model underlying the proposed software is based on algorithms that have been informed by the experience gained in developing the Sitka spruce dynamic yield model. The intention has been to achieve ‘quick-hit’ improvements in the representation of growth and yield models for tree species other than Sitka spruce. As such, the new model offers a number of significant advances over the static tables in Booklet 48, but does not aim to achieve a level of complexity necessary to represent stands of very diverse structure. At present a number of basic functions and mathematical relationships have been characterised and development work will need to continue in the coming year, however some of the results already obtained may have immediate application. A notable example is a set of functions that have been derived for estimating target volumes, basal areas and numbers of trees to be removed at different stages in the life cycle of a stand, depending on tree species, thinning type and intensity. These functions could be used in systems for supporting thinning control and accordingly a small piece of software has been developed to illustrate their potential for this application. The software will be demonstrated at a forthcoming meeting of the FE Thinning Project Group.

Ultimately this new model may replace the tables in a future manifestation of FC Booklet 48.

Models for forecasting wood fuel resource
Significant research and development support was provided to a DTI funded project, jointly coordinated by the FCA and the Forestry Commission (Helen McKay), to quantify the potential wood fuel resource available in Britain. A key contribution was the rapid development of a new computer-based model, BSORT, which produces estimates of the biomass of different tree components in stands of different species for a range of yield classes and management regimes. This model was designed to integrate with the computer based Booklet-48 yield models and with the FE Production Forecast program, enabling the upscaling of biomass estimates to regional and national levels.

Models for forecasting production from low-impact silviculture stands
A significant contribution was made to an FC-Wales funded/coordinated project aiming to estimate the impact on the potential timber resource of implementation of a policy of low-impact silviculture (LIS) within Welsh woodlands. The key contribution was the rapid specification and development of a complete set of alternative Booklet-48 yield tables that represented growth and yield under a LIS regime as specified by the project coordinators. These models were incorporated into the FC Production forecast program enabling upscaling of forecasts for different groups of stands constituting the woodlands of Wales. Advice was also provided to the project coordinators on the specification of the LIS regime and on classification of Welsh forests for application of appropriate yield models. The development of the LIS-type models within a very short time scale was only possible due to the progress achieved as part of the ongoing development of improved FC Booklet 48 models.

The concerted efforts this year to achieve progress on development of new models limited opportunities collaborative research. However, there is growing interest among researchers in applying the new Sitka spruce dynamic yield model to address issues beyond forecasting stand growth and yield. Most notably, FR Mensuration has being working with FR Pathology Branch on forecasting the impact of Heterobasidion annosum on growth and yield of Sitka spruce in Britain. This research has formed part of an FC/EU-funded collaborative research project (MOHIEF) and has involved providing a copy of the Sitka spruce dynamic yield model to Prof. Timo Pukkala (Joensuu University, Finland), for linking to a model of Heterobasidion annosum disease spread. Prof. Pukkala has provided useful feedback to the FR yield modelling team.

Within FR, the yield modelling team has initiated discussions with researchers working on models of stand establishment, windthrow hazard and timber quality with the aim of achieving integration with the Sitka spruce dynamic yield model.

Advisory work
Over the year about 100 enquiries were received on subjects related to yield modelling and forecasting. Most of these were in the form of telephone calls direct to FR but an increasing number were received as e-mails, some forwarded from Forestry Commission Public Information Division or other FC departments. Most enquiries were from people working in the forest industries but some were received from researchers, students and members of the public. Examples of subjects included:

  • Where to obtain and how to apply yield models
  • Correct interpretation of yield model estimates
  • Expected growth rates of trees and stands in different areas of Britain
  • Potential quantities of wood fuel produced by typical stands in Britain.

Publications and significant FC/FE reports

LEE, S.J. and MATTHEWS, R.W. (in press). An indication of the likely volume gains from improved Sitka spruce planting stock. Forestry Commission Information Note. Forestry Commission, Edinburgh.

MATTHEWS, R.W. and MACKIE, E.D. (in press). Inventory: Mensuration – Growth and Yield. In: J. Burley, J. Evans, and J.A. Youngquist (eds.) Encyclopedia of Forest Sciences. Elsevier, Oxford, 4.

MATTHEWS, R.W. (2003). The development of cumulative volume production in Corsican pine stands in Britain. Final report to Forest Enterprise England. Mensuration Report 2003/06.

MATTHEWS, R.W. (2003). Selection of mensurational variables and assessment procedures for informing production forecasts in UK forest stands. Final report to Forest Enterprise Operational Support.