Official website of the New Zealand Farm Forestry Association

Improving siting recommendations for farm forestry eucalypts using GIS maps

Angus Gordon, New Zealand Tree Grower February 2014.

This mapping project set out to collect growth performance data on a range of eucalypt species which were planted in the 2004 species survival trial, project SFF L03-007. All the trial sites were visited and the relevant site descriptors, growth and performance data of the surviving trees and any other site attributes were collected. This site visit and data collection was critical before the growth of any individual species at any trial site began to affect the trees in a neighbouring species row, and before dominant trees could suppress their sub-dominant neighbours.

Mean annual height growth increment in metres per year was chosen as the most accurate measurement of tree and species performance across the sites. Other performance measures were obtained, including diameter at breast height and stem form. They present a picture of each species, but they are less reliable for describing between species performance variation across a range of sites.

The ultimate aim of this project was using the collated data to generate a GIS-based mapping system to display the potential performance of any of the mapped species in the New Zealand landscape and show the results. The aim back in 2003 was to join the dots of survival and growth performance of individual eucalypt species, and if possible come up with a nationwide species performance selection process. This would help reduce the dependence that many growers have on small snippets of anecdotal information in any one region or locality.


Data collection was undertaken by Ian Nicholas and myself with the assistance of trial owners and members of the broader farm forestry community. Roger May is the GIS mapping specialist who collated all the data provided, plotted it and performed statistical analysis. All the site descriptors measured were plotted against mean annual height growth increment to produce scatter plots.

The site descriptors which had the strongest mathematical relationships were chosen for further use. They were latitude, consecutive frost-free days, February maximum temperature and July minimum temperature. A predicted growth response for each of our chosen species was produced, enabling us to produce the GIS mapping. The differences in sensitivity to our four site descriptors which each species displays cause the changes in inferred performance on the maps.

These changes in predicted performance on the maps directly link areas of the landscape that should produce similar growth results of those of our trial sites for any of the mapped species. The maps have a coarse resolution due to the scale of data available from the national data for climate and soils. Actual field performance by any species at the local level will therefore have further variability which cannot be accurately shown on these maps.

One important point to note was that not all of the 19 species which were placed in the original trial were planted at an equal number of sites. This limited the number of species which had good enough data for statistical analysis and accurate plotting. The six stringybark eucalypt species chosen for mapping were Eucalyptus globoidea, E. laevopinea, E. macrorhyncha, E. muelleriana, E. pilularis and E. youmanii. A seventh species, E. fastigata which is one of the ash group of eucalypts, was used as our control species and had a plot produced for comparison purposes. These maps are all available on the NZFFA website.

Major species planting, survival and mapping

Species baxteri blaxlandii fastigata globoidea laevopinea maculata macroryncha microcorys muelleriana pilularis youmanii
Planted sites 13 9 47 43 46 23 42 33 32 34 47
Live sites 11 5 39 30 34 16 31 17 26 25 37
Live site per cent 85 55 83 73 74 70 74 52 81 74 79
Planted sites per cent 27 19 100 92 98 49 89 70 68 72 100
GIS maps     yes yes yes   yes   yes yes yes


The GIS maps are the first attempt to produce an indicative productivity mapping tool for the six chosen stringybark eucalypt species on a nationwide basis. Importantly it is based on a large number of trial sites spread across a diverse sample of the New Zealand landscape. This is rare, as these species have not been widely adopted by the commercial forestry sector or planted collectively in an organised fashion.

From the work undertaken in this project we can now predict the generalised performance of the six mapped species in comparison to each other and our control species. For the unmapped species in the trial, manual comparisons can still be made at a trial site level using the data supplied in the spreadsheet available on the NZFFA website. Much of the data from the original trial sites which have very low or no surviving trees had good species survival performance initially until their second year following planting. A heavy late frost, minus five degrees centigrade in December 2006, destroyed the trees on those sites. This inability to withstand late out-of-season frosting has always been a limiting factor for some species in eucalypt silviculture.

Success or failure

By mapping together parts of the landscape with similar environmental attributes to the trial sites, we have been able to push the boundaries of where it could previously have been considered prudent to establish these species. Many of what were previously thought to be very tender species are surviving as far south as Southland, where appropriate siting was undertaken. There are, however, two important measures of the relative success or failure of any trial species.

The first is the survival percentage of the planted seedlings, and the second is the subsequent growth performance. It is important to note that some of our mapped species had far superior survival performance nationwide but have lower annual height growth increments. When reviewing the entire data it would appear that some species are less favourable to plant than others. This good survival, but slower initial growth, seems to be an attribute buried deep within the genome of some of the eucalypt species that were trialled.

These particular species are tough, with good survival characteristics, but can be slow starters. Others are very quick growers under favourable conditions, but can be far more fragile when it comes to climatic extremes. It is a classical tortoise and hare analogy.

A short summary of the six mapped stringybark species performance follows, but it is only brief. A full assessment of all the species will be done at a later date and included in the eucalypt section of the NZFFA website.

Eucalyptus youmanii

This species survives and performs over the widest range of latitudes of any of the trial species and on many sites to high altitudes. It confirms what has been seen anecdotally in the few small-scale plantings previously attempted in New Zealand with this species. It was generally the first stringybark species to survive in cold frosty areas in the North Island and the South Island.

Eucalyptus macrorhyncha

This species also has a wide survival range, although is more limited over the lower half of the South Island. It has superior height growth to E. youmanii which was evident on most sites. One characteristic which both these species share is their poor performance in the warm humid north of the North Island. We would recommend that neither of these species be planted north of Auckland or in the warm humid parts of Bay of Plenty without careful consideration of the siting of the potential planting.

Eucalyptus globoidea

This species survives and grows over a slightly more limited range of New Zealand, and shares some of the slow starting characteristics of the previous two species in cooler regions of the country. However it
is far happier in the warmer and higher rainfall areas of the upper North Island and has superior growth performance as the farther north you go. E. globoidea did survive in Southland and coastal Otago on three sites, but has not really performed well unless planted north of Banks Peninsula.

Eucalyptus laevopinea

This species probably produced the biggest surprises of the project. Previously held assumptions were that this was a warmth and moisture-loving species. However our results show that while this is true, it is far from the whole truth in terms of survival and to a lesser extent growth.

E. laevopinea survived very well in the North Island, with an extreme example growing well at 700 metres above sea level, and also performed well in the South Island. Like E. globoidea, it did survive on lower altitude South Island sites as far south as Southland, especially near to the coast.

Growth was compromised on many of the lower South Island sites, so in all practical terms this species is best planted from Banks Peninsula north. Its survival performance and ability to grow well in wet and dry climates indicates that it is a candidate species for further work if provenances which perform well in cold climates can be found.

Eucalyptus pilularis and E. muelleriana

The last two of the mapped species were E. pilularis and E. muelleriana. Both of these species seem to prefer warm, moist growing conditions and showed that they survive and perform well in relatively benign sites. E muelleriana can do well in the South Island north of Banks Peninsula and on the West Coast, but comes into its own in the North Island, especially on sites where large numbers of consecutive frost-free days predominate. However, it has been the most unpredictable species we have mapped. Its lack of consistent survival and growth performance at adjacent sites has remained unexplained. Where warmth and moisture are present it is a good performer.

E. pilularis is the most warm-loving of the species which we plotted. It did survive in some South Island sites, notably at a coastal site in Dunedin, but it has only really performed adequately from Blenheim north.
In the North Island it again performed adequately at low altitude coastal sites, but it was not until it was grown from the Bay of Plenty and north that its performance could be described as good. This is really a regionally important species for the upper North Island and for coastal sites further south.

Indicative site suitability for Eucalyptus laevopinea, based on mean annual height data from eight and nine-year-old trees at all 46 sites

Snapshot of a gene pool

All of the six mapped stringybark species can be compared against the map for E. fastigata, which we have plotted using the same analysis techniques. The GIS results of the mapped species are summaries of performance, and in reality there is a lot more knowledge to be gained by reviewing the spreadsheet data and the maps simultaneously. One very important point to note is that all the original plantings used a maximum of three provenances or families, and sometimes only two provenances per species. A very small subset of the available gene pool of each individual species and its potential performance characteristics was therefore available for planting.

This may well have skewed the performance of any of the planted species away from where it potentially could be. This is an important point to emphasise as this project is not the gold standard of knowledge on any of these species. It is just a snapshot of a limited gene pool over an equally limited time frame on a small sample of New Zealand sites. In comparison is a recent multi-family importation of E. globoidea by another research body, the Dryland Forests Initiative. This has shown that far superior growth rates and survival to anything that was achieved by any species in this project are possible if you are prepared to cast the genetic net wide and spend enough money.

Other species

A review of the spreadsheet data will show the performance of the other species which were trialled and analysed. The main reason for not attempting to plot all the original species was lack of data size of those species. This does not exclude these species from further work using these or other techniques, as the data for this project relied on the availability of planting stock from the previous species survival trial set up in 2003.

This is the case with two species, in particular E. baxteri and E. blaxlandii, which were planted at a very limited number of sites and have shown some promise. There are other candidate stringybark eucalypt species which may well be worthy of further research and planting, but similarly they were not included in this project due to their lack of inclusion in the previous 2003 project.

It is also worth noting that this project has not dealt with performance variations due to micro-site influences. Eucalypts can be very sensitive to any one of a range of site factors which individually may limit survival or growth performance and collectively can be the difference between success and failure at that site. Examples are position in the landscape, potential water drainage, exposure to wind and frost ponding.

Any user of this data who intends to undertake large-scale planting would be wise to familiarise themselves with the particular micro-site problems that their chosen species may be sensitive to, and whether these exist at the potential planting site. In the data collection and analysis of this project it became obvious that in some cases micro-site factors had come into play in the poor performance at some sites. However in most cases it was only a problem when a species was pushing the performance boundaries of what its particular genetics allowed.


The project has been able to benefit from the information gained from the earlier survival trial SFF (L03-007). The results from it have been turned into guidance for a grower who wants to plant in an area which previously had no history of eucalypt plantings or a site which has no history of planting with the species selected for trial. The inclusion of E. fastigata as a benchmark species was important as this species
is not very strict in its site requirements. It survives and grows well over a wide range of sites, therefore enabling it to be used as a proxy for the stringybark group of species wherever it can be successfully grown. Historically, E. fastigata has been widely planted in New Zealand, and for this project its inclusion is a particular strength.

Opportunities for growers

Now that this project has been completed, anyone wishing to become involved in the New Zealand hardwood industry has information which can guide them in species choice and the expected growth performance on any chosen site. This is for the species we have mapped and to a lesser extent the ones we have not. The importance of this is that prospective growers have the opportunity to shift the siting of any potential plantation to an area where growth performance of their chosen species can be maximised, or to choose which species to grow if changing location of the plantation is impractical.

Helping answer the question ‘what species do I choose for my site’ may well be the most prevalent use of this resource as the project has shown that there are several species options for almost every site in New Zealand, irrespective of latitude and to a lesser extent altitude.

One of the main uses of this information is the Excel spreadsheet package which has been developed to display the site data on scatter plots. It is downloadable and open in terms of data entry. We anticipate that in the future we will be able to undertake further analysis including data from other trial series. In this respect getting closer to a ideal answer is little more than a numbers game – the more data you have the closer you are likely to get to having an answer which approaches the reality in the field.

The future

Looking to the future it is obvious that some of the trialled species have the potential to play a significant role in hardwood forestry in New Zealand. To have some chance of success, however, an effort needs to be made to obtain a more diverse gene pool for any selected species. E. globoidea and E. macrorhyncha have been included in the Dryland Forests Initiative genetics collection, and E. muelleriana and E. pilularis are in historical provenance collections now managed by Scion. The hardwood growing community within New Zealand will have to seriously look at what other species may be worthy of future intensive work by way of importation of new genetic material and subsequent breeding to suit local conditions.

As a result two other species for which productivity maps were produced, E. laevopinea and E. youmanii, could both be contenders for having broadly-based provenance collections made. There are others in the list of species which showed real promise, due to their performance in this project and in other independent plantings or historical collections. Two in particular, E. baxteri and E. blaxlandii, could both be contenders. This bodes well for the future as it means that hardwood growers do have plenty of options. However it will be vital to develop enough focus on the species that we choose in order to make some real gains in productivity performance.

A lot has been achieved

This mapping project and the previous survival trial have achieved a great deal, thanks in no small part to the trial hosts. They purchased the plants for row plot trials on their land and tended them with care, and in some cases have little to show for their investment, apart from a grassy corner where the eucalypt trial once was.

There are in excess of 50 of you scattered across the landscape between North Cape and Bluff. All were equally important. A thank you must also go to the project team which included Ian Nicholas, myself, Roger May, Paul Millen, Dean Satchell, Graham Milligan, Gary Fleming and two co-opted members, Denis Hocking and Shaf van Ballekom. It is the combination of wise heads that keep projects like this on the right track.

This project had primary funding from the Sustainable Farming Fund and co-funding from the Neill Barr Foundation, Vineyard Timbers, Proseed, Ian Nicholas Consulting, the NZFFA Research Fund and the Eucalypt Action Group. Lastly, a big thank you must go to Ann Nicholas and her family who found the time and energy to help retrieve bits of information from Ian’s files that were crucial to this project during his sudden brief illness and after his death.

Parting shots

A good blueprint has been laid down for what the future of one section of the New Zealand hardwood industry might look like in the future. Vital for this future is adequate investment in planting on a large enough scale to develop some critical mass so that downstream processors and members of the value chain have confidence that there is an industry out there to supply them. In addition, an adequate supply of improved genetics is crucial and there are a number of people and organisations who are focussing on that very issue at this moment.

The biggest hurdle that we as growers face is our own motivation to change. If we are brave enough we might replant that gully in eucalypts when the pines come out in a couple of years and step from an industrial commodity species into a niche market. It has been done before. Are you brave enough to step out into the sunlight or would you rather hunker down in the shadows and let other people take the lead and make the profits instead of you? Only you can answer that question – welcome to your future.

Angus Gordon is a member of the NZFFA executive. He chairs the Eucalypt Action Group and has been involved in several Sustainable Farming Fund projects including eucalypt species survival trials, the eucalypt GIS mapping and the Trees on Farms projects.


Farm Forestry - Headlines