Mortality and recruitment of tagged stems grouped by palatability class
Background
This measure relates to indicator 1.5.1 – Species composition and diversity.
Before human occupation, Aotearoa New Zealand was 80% forest, and forest remains the dominant landcover on public conservation land (PCL). Measuring tree survival and recruitment (demographic rates) helps predict how forests will change. Change is driven by natural and anthropogenic factors such as climate change and invasive species including browsing mammals like possums, goats and deer. Evaluating demographic rates for common trees grouped by browsing mammals’ food preferences may help direct DOC’s management and provide a baseline to compare with future trends and/or management outcomes.
National tree populations are stable but their composition may be changing.
What did we measure?
DOC has a national monitoring programme to assess status and trends of biodiversity by regularly measuring approximately 1,400 sites evenly spaced on an 8km grid across PCL (Figure 1). Approximately 280 randomly selected sites are measured each field season (September - May) so that each site is measured on a 5-year rotation.
At each site, DOC staff measure all tree stems rooted in a 20x20m plot. Stems are permanently tagged so they can be tracked over time. In addition, saplings and seedlings are counted, and every plant species (mosses, liverworts and all) in the plot is identified and recorded. Bats, birds, possums and ungulates including goats and deer are also monitored at each site.
The vegetation monitoring part of DOC’s programme builds on the Ministry for the Environment’s Land use and carbon analysis system to monitor carbon stored in natural forests, which was first measured from 2002-2007. The second measurement was from 2009-2014. The third measurement was from 2015-2019. This factsheet uses data collected up to the end of the 2017/2018 austral summer, so two measurements are available for most sites and three measurements for some (Table 1).
Tagged stem data were analysed with a Bayesian individual-based response model to estimate annual rates of recruitment and mortality for species grouped by their palatability to different browsing mammals, allowing for underlying differences between sites. Because mammals selectively browse some plants and not others, they can alter recruitment or mortality rates for particular species. Similar rates indicate a stable population, higher recruitment indicates population growth, and higher mortality indicates population decline.
What did we find?
The national population of indigenous trees on PCL has been stable or slightly increased. Over all sites and species, stem mortality was 1.6 - 1.66% p.a (95% CrI) and recruitment was 2.41 - 2.48% p.a.between the first and second measurements. Between the second and third measurements, higher mortality (1.97 - 2.06% p.a) was compensated by higher recruitment (2.26 - 2.36% p.a.).
There may be a shift in forest composition as populations of deer-avoided species increase (Figure 2). Deer-preferred trees have similar rates of recruitment (1.4 - 1.7% p.a, 95% CrI) and mortality (1.4 - 1.6% p.a). However, trees of species deer avoid eating have higher recruitment (1.6 - 2% p.a.) than mortality (1.2 - 1.5% p.a.).
Goats appear to have a similar impact to deer. Goat preferred trees have lower recruitment (1.2 - 1.3% p.a, 95% CrI) than mortality (1.3 - 1.5% p.a). However, trees of species goats avoid eating have higher recruitment (1.7 - 2% p.a.) than mortality (1.5 - 1.8% p.a.). This is of concern because ungulates have increased from occupying 63% of PCL in 2012/2013 to occupying 82% of PCL in 2019/2020.
Populations of possum-avoided tree species may also be increasing relative to those they prefer. The estimated rate of mortality for possum-preferred trees was relatively high 1.7 - 2% p.a. (95% CrI), although balanced by a high rate of recruitment (1.7 - 2.1% p.a.). However, for trees possums avoid eating, mortality (0.9 - 1% p.a.) is lower than recruitment (1.2 - 1.3% p.a.).
Demographic rates are highly variable across the sites sampled, as expected given underlying range in environmental factors and forest histories. This means local patterns may differ from national summaries reported here (Figure 1).
Within overall palatability group responses, rates of mortality and recruitment for individual species are variable (Figure 3).
Figure 1: Location of plots on PCL with tree data showing rates of stem turnover, mortality and recruitment. Click on an individual site to see the number of live stems recorded at each measurement.
Figure 2: Estimated annual rates of mortality and recruitment for stems grouped by their palatability to three widespread exotic mammals. Rates are shown for the period between the first and second measurements (1_2), the period between the second and third measurements (2_3) and the whole time period (w) from the first to the most recent measurement for each site
Figure 3: Balance between stem mortality and recruitment for common or widespread tree species, accounting for their palatability to deer, goats and possums. Hover over a point to see the species code and sample size.
| Measurement interval | Years between measures (median, range) | Number of plots | Stems that survived | Stems that died | Stems that recruited | Total stems recorded |
|---|---|---|---|---|---|---|
| period 1 to 2 | 8 ( 6 - 9 ) | 1,005 | 118,271 | 15,949 | 24,912 | 165,299 |
| period 2 to 3 | 5 ( 3 - 9 ) | 694 | 88,759 | 9,100 | 11,685 | 129,748 |
| overall | 12 ( 3 - 16 ) | 1,103 | 119,269 | 25,516 | 37,743 | 196,465 |
Data quality
This measure is classified as a partial measure and complies with data quality guidelines used in the Environmental Reporting framework.
Stem demographic rates are key metrics of forest ecosystem function and census of marked individuals in a defined area is a standard technique for measuring this. The analysis method applied here was published in an international review which showed the models used should not unduly bias results.
Glossary of terms
95% credible interval (CrI) indicates that the true mean lies inside the interval with 95% probability given the posterior probability distribution.
Annual recruitment is the rate at which trees in a population surpass 2.5cm diameter at breast height (dbh). Rates reported here use eq.7 from Kohyama et al. (2018), that is the proportion of stems in a population that recruit relative to the final size of the population, estimated with the Bayesian model described above.
Annual mortality is the rate at which a population of trees die. Rates reported here use eq. 6 from Kohyama et al. (2018), that is the proportion of stems in a population that died relative to the initial size of the population, estimated with the Bayesian model described above.
Annual turnover is the rate at which a population of trees changes. Rates reported here compare the proportion of the population that are die or are recruited relative to the whole population using a metric described by Mason and Bellingham (2018) calculated for each plot.
Palatability reflects the dietary preference of a target animal species or group. Categories reported here are from Allen et al. (2013) and were compiled from published research about which plants the animals prefer and actively select to eat, and those that they deliberately avoid eating. Methods to determine this include comparisons of abundance inside / outside exclosures, direct observations of feeding animals, gut samples, etc.
Stems are trunks of trees or treeferns over 2.5cm diameter at breast height (1.35m). If trees fork below breast height, each stem is individually tagged and monitored.
Additional resources
Allen, R.B., Wright, E.F., MacLeod, C.J., Bellingham, P.J., Forsyth, D.M., Mason, N.W., Gormley, A.M., Marburg, A.E., MacKenzie, D.I., McKay, M., 2013. Designing an inventory and monitoring programme for the department of conservation’s natural heritage management system, Landcare research contract report. Department of Conservation, Wellington.
Kohyama, T.S., Kohyama, T.I., Sheil, D., 2018. Definition and estimation of vital rates from repeated censuses: Choices, comparisons and bias corrections focusing on trees. Methods in Ecology and Evolution 9, 809–821.
Lee, W., McGlone, M., Wright, E., 2005. Biodiversity inventory and monitoring: A review of national and international systems and a proposed framework for future biodiversity monitoring by the Department of Conservation. Landcare Research Contract Report LC0405/122 (unpublished) for the Department of Conservation, Wellington.
Mason, N.W., Bellingham, P.J., 2018. Evaluating optimum measurement of biodiversity indicators, Contract report. Manaaki Whenua Landcare Research, Lincoln.
McGlone, M.S., McNutt, K., Richardson, S.J., Bellingham, P.J., Wright, E.F., 2020. Biodiversity monitoring, ecological integrity, and the design of the New Zealand biodiversity assessment framework. New Zealand Journal of Ecology 44, 3411.