Example of Reproducible Research from Egmont National Park
New Zealand Forest Surveys undertakes unfunded public-good projects. Our unpaid voluntary work is in liaison with Universities, Government-funded research organisations and the Department of Conservation.
North Island Axial Alpine Plots
From 1958 through to 1981 the New Zealand Forest Service established over three hundred permanently-marked alpine vegetation monitoring plots on the North Island mountain tops. Plots are located along the axial ranges in naturally occurring alpine tussock grasslands: Tararua, Ruahine, Kaweka and Kaimanawa Forest Parks. Each plot has one hundred 6″∅ intercepts, spaced at two link intervals on a 40 metre long string-line. Recording the presence of plants rooted within and overhanging the intercepts provides frequency of occurrence data. Plots were marked with three metal pegs hammered into the ground at 20 meter intervals. Some of these plots are still present after six decades and can be found after hours of searching with a metal detector. These potentially valuable plots have been ignored since the demise of the NZ Forest Service in 1987. Data from re-measurements would provide long-term insights into the effects of deer and hares, vegetation responses to climate change and plant succession.
Below are the plot locations, and an example plot photo taken in 1975 towards Ruahine Corner airstrip of plot G 59.
Over the past few years, NZ Forest Surveys staff have digitised field sheets and photographs of these plots, and relocated some plots in the field. We think that these plots – established in what were to become forest parks in alpine grasslands and shrublands – deserve re-measurement. Besides which they are located in the most beautiful parts of New Zealand, and have an excellent hut and tramping track network for access and accommodation. Some plots have mountain bike access and airstrips nearby. We intend to re-measure some of these plots in the coming summer using a modified Wraight protocol (Mick Wraight. 1960. The alpine grasslands of the Hokitika River catchment, Westland. New Zealand Journal of Science 3: 306-32.)
We are seeking committed, long-term volunteers who are happy to help, and who are prepared to live in rough mountain conditions for weeks at a time.
We’ve scanned aerial and plot photographs into one pdf file for each plot. This is essential for re-locating the plots in the field Example from the Ngaruroro 1960 survey in Kaweka Forest Park. We have also prepared field sheets for any future field work we do. Example fieldsheet.
We used Google Earth to geo-reference aerial photos used to establish plots. That was really easy to use (but time consuming), and has yielded fairly accurate NZTM co-ordinates for each plot. They have been saved as large Google Earth files.
For example the Tararua scrub plots are incomplete, so the digitised aerial photos aren’t too big:
Locations for all of the plots we have digitised:
Access the latest Google Earth files at the KML Downloads page.
We’d like to make a start on South Island Wraight plots one day. Hokitika, Molesworth, Haast/Arawhata, Taramakau, Wairau, Waitaki, Hurunui, Aspiring, Takatimu, Fiordland, and Stewart Island spring to mind. These plots were all established in the 1970s or earlier by NZ Forest Service. Data, photographs and fieldsheets for South Island plots are stored on the National Vegetations Survey Databank and will require thousands of hours to digitise and geo-reference.
Aerial photographs used by the New Zealand Forest Service in the 1950-1960s were scanned, digitised, and geo-referenced. Location marks were then converted in GPS location fixes. This figure shows locations of plots established in the Tutaekuri catchment in 1981 on the Kaweka Range:
Field trials in Ruahine Forest Park located some plot markers (aluminium pegs), and showed that plots were generally within 60 m of our GPS estimates using our geo-referencing locations. More field work is required including these sites in Tararua Forest Park. Tararua Aerial Photo Example from the 1959 survey.
Depending on the enthusiasm of field workers it might be possible to find and re-measure more than one plot a day (allowing for 20% bad weather days). Initial plot measurement suggests that tussock stature has declined. It wasn’t measured in the Wraight plots, but height tiers in the impending surveys will provide stature data. If it is that clear then it might be possible to estimate tussock height from photos for 50s, 60s and 80s (a few plots were done in 80s), and once again for the remeasure and calibrate both from the height measues.
We could consider establishing additional plots at random or systematic locations. They could be used for comparison to counter criticism that the “haphazard” plot establishment in the 50s and 60s was not representative. Systematic sampling could provide additional spatial monitoring potential. We may also sample plants and soils for nutrients and traits. Harvest at some sites to provide a relationship between biomass and plot measurements. We may select some species of interest and map their densities. Many of the species, which might be exotic, would need to be mapped in belt transects. We could also measure Dickenson and Marks Scott PHI plots in Tongariro Naty Park.
Animal monitoring could be considered. Counting feacal pellets should take about 20-30 mins per site. For the really serious we might establish random monitoring points. Say 100 spaced over 200-300 km. Fly at 80-100 kmh and <200 m from ground to all points counting deer. Say 20 -30 flights of 2-3 hours. Should count several hundred deer that way, maybe even 1000. GPS locations. That could be used to model density and compare to pellet counts. Maybe do some ground based work too though you will be seeing dozens over a period of weeks and not 100s of deer. These data could be used to test the currently used method of counting deer poo. We might sample rumens from deer, hares and possums.
Axial alpine plant identification guide
We are preparing a field flora database for production of plant identification guides. The approach is to include descriptions of plants, drawings and photographs in the database. This information is then combined to produce Latex documents as required.
New Zealand scrub monitoring plots
New Zealand Forest Surveys remeasured 204 LUCAS monitoring plots for the Ministry for the Environment between 2011 and 2013 (established 2003-2006). Data (Report) have been used to determine changes in carbon by MFE. Manuals
Three methods were used to measure scrub plots to determine woody plant volume (from which biomass was calculated). Firstly, measurement of diameter of all trees >=2.5 cm DBH and measurement of height of a subset of trees in 20 x 20 m plot. Also measuement of large tree diameter and height in a 20m horizontal radius plot. Horizontal measurements of plot boundaries were also painstakingly made. Audits were undertaken on 10% of plots, with excellent results for our work. Note that results for two providers to MFE were published in an Audit Report for MFE for the 2011-12 season, where NZFS’s work was commended in comparison to the poorer work of the other provider. NZFS also passed all individual audits in the 2012-13 season, and the other provider did not. New Zealand Forest Surveys has never failed an audit. Audit Reports zipped.
Where diameter measurements did not work well other methods were used (note that protocols changed between the establishment and re-measure surveys – see manuals):
Discrete Shrubs in 2011 manual: Measure the maximum vertical natural height (i.e. not perpendicular to the ground when sloping) and two orthogonal widths in the horizontal plane (one of which is the maximum width) of each individual shrub and shrub-clump (>30 cm height) within the 20 x 20 m plot. At each plot 4 shrubs were harvested and dry weighted.
Continuous shrub cover: Cover was estimated to 10% in each 5 x 5 m subplot. In 25 grid intersects (0.2 × 0.2-m), the maximum vertical height of the woody vegetation was measured to the nearest 10 cm. At each site a 4 x 1 m plot was harvested and woody ver dry weighed.
The 2004 CMS manual was used for the establishment survey. The main difference being that attempts were made to estimate cover using linear tapes, whereas in the 2011 manual cover was guessed.
New Zealand’s Forest paired exclosure plot network
Data were collected from paired deer exclosure plots (20 m × 20 m, n= 50 pairs, one fenced to exclude deer and one not) between August 2008 and November 2010. Plots were established between 1969 and 1998. Dead dog exclosure in Ruahine Forest Park.
In each of the 100 plots, live vegetation was comprehensively measured using standardised LUCAS and permanent plot (Hurst 2007) monitoring techniques. In addition heights of all trees were measured, and dead coarse woody tree material was measured standing and as debris. At rocky plots soil sampling pits were used to sample soil (three 0.1 m 2 × 0.3 m deep), and at sites with <25% rock content sixteen soil cores (65 mm ∅) were extracted. Humus and litter samples were collected from sub-plots. At rocky sites sub-plots three 0.1 m 2 were systematically located and at sites where soil cores were extracted litter and humus was collected from sixteen sub-plots (100 mm ∅). Forest overstorey was assessed in each plot by by identifying the species of all tree stems (>2.5 cm DBH, diameter at breast height: 135 cm) and measuring diameter at breast height (135 cm, DBH). All trees were permanently tagged, so tree growth (productivity surrogate), mortality and recruitment can be calculated. When a treeferns cautex was sufficiently solid at 135 cm they were also tagged using a aluminium tree tag and 75 mm galvanised nail. Because diameters are impossible to consistently measure on tree ferns (Dicksonia spp. and Cyathea spp.) heights from the ground to the highest point of the ferns cautex were measured. All saplings (>135 cm high, <2.5 cm DBH) were counted and identified. Twenty-four circular understorey sub-plots (49 cm radius) were systematically located within each plot and established woody seedlings (15–135 cm high) were identified and counted in each of the sub-plots. The presence of seedlings < 15 cm high were also recorded in each sub-plots. Herbaceous species were recorded in the highest tier in which they occurred.
Some plots had over 100 soil assays done on them including all the trace elements, cation work as well as the more common elements. The thing that stands out is Phosphorous. Particularly soil organic P. There seems to be a complex response where some sites (where P is high) have much higher P inside the fence than at sites where P is low. Quite fascinating.
Plant assay work at around 18 of the sites didn’t show much. Again all the trace elements, foliar fiber, SLA. We also did some glasshouse work with litter bags and germination work.