Maruia, Okarito, Saltwater forest, beech forest, logging,
the beech scheme, rimu, Timberlands, Kit Richards, sustainable management of forests,
conservation, preservation, ancient rainforest,
all buzz-words of a vital and passionately debated topic in New Zealand in 1999 and 2000.

Some more are; sustainability, Resource Management Act, the West Coast Accord,
Labour, Minister of Forests, Hodgson, Clark, election mandate, city greenies, nfa, Native Forest Action,
F&B, Royal Forest and Bird, twigs and twitters, destruction, clearfelling, helicopter logging, Efford, modelling; and they apply to

Summary of critical points made against the Efford growth model.

Timberlands West Coast Ltd. carefully planned for 10 years, at a cost of $3 million NZ, with the guarantee of the West Coast Accord as a firm and contractually binding basis, to begin innovative and world-leading forest management processes and techniques for perpetually sustainable production of fine beech timber for furniture, and for enhanced conservation of threatened plants and wildlife in the process. The newly-elected Labour government in December 1999, halted the environmental planning process, claiming that the support they obtained (before the November 1999 election) from northern city electorates for a new policy to 'put a stop to all logging in state indigenous forests', was a valid mandate to break the 1990 contract with the West Coast region. Thus began a vigorous debate in New Zealand on indigenous forest management.

During 1999, the Royal Forest and Bird Preservation Society contracted with Dr Murray Efford of Landcare Crown Research Institute to write a computer-based growth model in which to run the basic data set used by Timberlands West Coast Limited (TWC) to develop their beech silviculture system and predict forest responses to sustainable harvesting. They wanted him to test the TWC predictions using his expertise.

He wrote a model and it has been available as a down-loadable program from the Landcare site. Based on what he found upon using it, Dr Efford gave a presentation on 1st September 1999 to the Transport & Environment Select Committee of Parliament. The results of running this model were drawn upon by various MPs, seemingly none with any background of forestry or modelling. The results of all these experiences and Dr Efford's presentation have been quite damaging, and many people, including skilled forest modellers, have been and still are, critical of many aspects of this model and its presentation for public use. Following initial heavy criticism, Dr Efford modified the model and this modified version has also been available from the Landcare site in quite another place from the first version. He made some changes to remove some of the most obvious faults, but many remain.

The following are some of the major faults that people have found with the models created by Dr Efford, and how they have been used.

For those not familiar with computer terminology, the word 'default' means the functioning state that a computer, or computer program, is set to work at as a matter of course - unless the operator chooses otherwise and changes control values. The word 'simulation' means to mimic or predict through calculations copying natural processes, the course of growth pattern of - in this case - an average stand of beech trees.

  1. The model assumes that all trees harvested are in addition to natural mortality. Dr Efford does not believe the statements of the TWC planners that they can and will predict which trees will die within a stated period and harvest among those. Hence he set a default harvest value of 1.0 and not 0.5. This action has a severe effect on the outcome of the model predictions, since it doubles the effective rate of stem removal (harvest + death instead of (all natural death includes harvested stems)).
  2. The model has a default period of 400 years. That is, once started, the simulation runs for a practice period of 400 years without any operator intervention or correction to the calculations at any stage between year 0 and year 400.

    The matter with this is:-
    1. The period of Resource consents sought was 35 years. So it is misleading to assume that the silvicultural regime that commenced at year 0 would continue for 400 years.
    2. It is made clear within the Consents application and supporting documents, that the TWC model would be used with continual correction based on feedback from researchers and monitoring. To put this in another perspective, the Efford model runs as if, when driving to work from home, your car is carefully pointed to work , and the controls are not touched again until work is reached. The TWC model was written in the expectation of intervention and continued correction along the way so as to stay on the road.
    3. It was made clear by TWC that their current model was the best one they had at the moment, and they expected to have increasingly better models based on better inventory and experience data.
  3. The model assumes that the greater a proportion of mature stems are cut, (i.e. the fewer remain standing) the lesser the amount of natural regeneration. Note that Dr Efford's work experience is primarily with modelling animal populations. That is what often happens with animals; reduce the numbers of breeding adults, and the numbers of offspring per unit of land area in smaller.
    With trees, especially beech, it does not work like this. Provided there is a seed source within seed flight distance of seed trees, the more 'vacant' land space there is, the more seedlings will grow. Land not occupied by adult trees will be colonised by new juveniles.
    This is potentially a major fault in the model. In practice, however, the effect of this defect is slight since none of the realistic simulation would cut so much for this to take effect. However, untutored members of the public, and MPs, playing with the program and 'trying this action out', could soon draw a very wrong, and potentially damaging conclusion.
  4. The model made no allowance for the fact that if mature trees are removed from beech forest, the trees remaining around the gap created by the removal will respond to the availability of increased light, water and nutrients by growing faster. This is a well-established principle in managing plant crops ranging from leeks and cabbages, to beech trees. Note; it doesn't happen with animals though, except in extreme circumstances.
  5. There is a more technical fault in Dr Efford's model in that he assumed that the beech stands contained trees always in a continuous size gradient from small to large. In fact, because many stands of the forest have been affected by some catastrophe such as earthquake or wind-storm, there are significant places where the stands of trees are even-aged and even-sized due to simultaneous regeneration. Dr Efford therefore claimed that the TWC model has a hidden bias due to this effect, whereas the TWC scientists state that it is minor.
  6. The models available on the internet produce an attractive graphic display which pretends to show how the stands of trees "before" and "after" each run of the model would appear in aerial view. In fact, the models have no mathematical provision to calculate where each tree would germinate and grow on any putative area of forest, so these "pictures" are not truthful representations of what might happen in real life.
  7. TWC believes that the extraordinarily low rates of timber removal and all of the safety margins built into its proposal will make imposed changes on the forests safe and minor.
  8. In the presentation to the Transport and Environment Select Committee, Dr Efford made statements and assertions that, although they reflected his apparent inexperience in forest science, would probably carry weight and be believed by people who knew no better.

    It seems that much of the timing and content were designed to deliver political damage to Timberlands and its plans.

    There has been considerable criticism by forest science experts including modellers, of the Landcare model and the presentation, because these experts consider the model and presentation were inaccurate to a damaging degree. As a qualified and experienced forester I can clearly see that assumptions and modelling results portrayed by Dr Efford are wrong and misleading. I have had correspondence with him on these matters, as have others. Dr Euan Mason, a scientist doing forest modelling on a regular basis and teaching the topic, has written a peer -reviewed scientific paper (published in "New Zealand Jurnal of Forestry" Vol44/4, February 2000) dealing with the model deficiencies; it is available on his web-site, at here. He mentions the problems in the page before that at just here.
Dr Graham Whyte also wrote a scientific paper in the above issue of the Journal of Forestry, and his observations are quoted here.
"Matrix models (see Vanclay, 1994) assume a steady state approach, which is inappropriate for a system of adaptive management and unrealistic irrespective of whether or not any wood harvesting is envisaged. Forest ecosytems are dynamic, not static entities. The claims made by Dr. Efford using this matrix model approach, therefore, have no scientific credibility, given that the input data refer to only one species, a notional average size class distribution for the whole forest, growth by size classes based largely on a relatively small sample size using increment cores, the employment of moving averages for size-class growth tables, a mortality function as yet unconfirmed, a constant rate of recruitment for all forest types, incorrect interpretation of harvesting rules, and forecasts with these constant inputs for hundreds of years ahead."

Contact me here, at bj@caverock.net.nz

.