Farm Forestry New Zealand

Alternative Species Site Mapping Review and Analysis

By Dean Meason, Pierre Bellè, Barbara Höck, Stefanie Lumnitz, and Christine Todoroki , June 2016.

Download SWP-T004 (pdf)

Executive summary

Background

Commercial forestry in New Zealand is dominated by the site insensitive species Pinus radiata (radiata pine). The commercial planting of non-radiata pine (alternative) forestry species over the last century has largely been small and fragmented. Although alternative species could potentially be commercially successful at larger scales, the tools for predicting where to grow such species and the potential yield of different sites are not at their full potential. The dominance of radiata pine in the last 40 years in the forestry sector and the stop-start support for a number of species has at times created a piecemeal approach in the development of productivity models for alternative species. Robust and accurate productivity model tools are now required to give certainty for landowners and investors to grow alternative species at a large scale.

This report investigates the state of knowledge of seven alternative species: Eucalyptus fastigata, E. nitens, E. regnans, Sequoia sempervirens (coast redwood), Pseudotsuga menziesii (Douglas- fir), Cupressus lusitanica, and C. macrocarpa. The report recommends what can be done to improve current models to more accurately predict species productivity at the national, regional, and local scale.

Project objective

The objectives of this report are four-fold. The first objective is to review the permanent sample plots (PSPs) for each species as the primary source of data for predicting productivity. Second is a literature review on the productivity models of each species and evaluate their ability to predict productivity throughout New Zealand. Third is to evaluate existing national spatial information in regard to each species’ modelling basis (the PSPs for the species) to determine the locationswhere the models are robust predictors of growth. The fourth objective is to provide recommendations to the SWP program on the steps needed to improve the accuracy and precision of the models.

Key results

The literature review revealed that model developments for alternative species were largely sporadic with no overall strategic direction until the Future Forests Research program was started.

Empirical productivity models use statistical relationships to predict growth and yield for a stand and/or geographic region. For these models to produce robust results, data from a large number of plots over a wide environmental range are required. Process-based productivity models require fewer plots than empirical models to be robust as they are based on a species’ physiologicalprocesses and limitations. However, this type of modelling does require data across the environment range. For the species in this report, the amount of PSPs available numbered in the low hundreds – at best. There was also an uneven geographical distribution of the species’ PSPs with the majority of them concentrated in central North Island. Only the P. menziesii PSPs had a wide geographic spread, however even they were not located in every region. Growth and yield models generally performed well in the central North Island and less so in other regions. Arguably, the regions outside the central North Island may be where alternative species could have the best productivity rates.

The literature review of the spatial productivity models showed that on a nationwide scale there were strong statistical relationships of species productivity with mean annual temperature and to a lesser extent other environmental factors. However, mean annual temperature only gives an indication of regions in New Zealand that are more or less suitable for each species. It did not give a finer scale prediction of productivity below the regional scale. Indeed, the spatial maps only showed that productivity was higher when a site was closer to the coast. These productivity maps are inadequate for locating optimal sites for alternative species.

The current productivity models were used in previous projects to produce spatial productivity surfaces for E. fastigata, E. regnans, S. sempervirens, and C. lusitanica. These surfaces were used in this report were used for objective three to analyse the robustness of the models’ prediction of productivity and to provide an indication of the strength of the alternative species productivity models in general. This analysis used environmental factors that exist at each species PSPs to give a measure of confidence of the productivity spatial maps. The range of two climatic (mean annual temperature and rainfall) and soil data (soil group and soil C:N ratio) were extracted from PSP locations and two confidence intervals were developed per species. Areas between the 25th and 75th percentile of each of the four environmental factors were classified as having high confidence. Areas between the 5th and 95th percentile were classified as having medium confidence. Overlaying these confidence intervals revealed for the C. lusitanica surface that only 9.3% fitted the high confidence category, S. sempervirens only 4.6%, and only 1.1% for both Eucalyptus species. For the moderate confidence interval, the fits were 49.6% of the original surface for C. lusitanica, 25.5% for S. sempervirens, and E. fastigata and E. regnans were 16.8% and 8%, respectively.

This report highlights that the models developed for the alternative species were based on a limited number of PSPs unevenly distributed throughout the country and with differing ages, stocking, and silviculture. Several of the empirical models heavily relied on PSP data from the central North Island region – which is not ideal environments for all the species. Spatial productivity maps are a good approach for identifying areas of New Zealand most suitable for an alternative species and comparing different species. These should be developed for all alternative species. However, the productivity models for all alternative species need to be improved for the spatial precision of the models to robustly predict productivity at the regional and local level. Recommendations to improve the productivity models for alternative species are below...