The right tree in the right place? A major economic tree species poses major ecological threats

Tree species in the Pinaceae are some of the most widely introduced non-native tree species globally, especially in the southern hemisphere. In New Zealand, plantations of radiata pine (Pinus radiata D. Don) occupy c. 1.6 million ha and form 90% of planted forests. Although radiata pine has naturalized since 1904, there is a general view in New Zealand that this species has not invaded widely. We comprehensively review where radiata pine has invaded throughout New Zealand. We used a combination of observational data and climate niche modelling to reveal that invasion has occurred nationally. Climate niche modelling demonstrates that while current occurrences are patchy, up to 76% of the land area (i.e. 211,388 km2) is climatically capable of supporting populations. Radiata pine has mainly invaded grasslands and shrublands, but also some forests. Notably, it has invaded lower-statured vegetation, including three classes of naturally uncommon ecosystems, primary successions and secondary successions. Overall, our findings demonstrate pervasive and ongoing invasion of radiata pine outside plantations. The relatively high growth rates and per individual effects of radiata pine may result in strong effects on naturally uncommon ecosystems and may alter successional trajectories. Local and central government currently manage radiata pine invasions while propagule pressure from existing and new plantations grows, hence greater emphasis is warranted both on managing current invasions and proactively preventing future radiata pine invasions. We therefore recommend a levy on new non-native conifer plantations to offset costs of managing invasions, and stricter regulations to protect vulnerable ecosystems. A levy on economic uses of invasive species to offset costs of managing invasions alongside stricter regulations to protect vulnerable ecosystems could be a widely adopted measure to avert future negative impacts.

combination of observational data and climate niche modelling to reveal that invasion has occurred nationally. Climate niche modelling demonstrates that while current occurrences are patchy, up to 76% of the land area (i.e. 211,388 km 2 ) is climatically capable of supporting populations. Radiata pine has mainly invaded grasslands and shrublands, but also some forests. Notably, it has invaded lower-statured vegetation, including three classes of naturally uncommon ecosystems, primary successions and secondary successions. Overall, our findings demonstrate pervasive and ongoing invasion of radiata pine outside plantations. The relatively high growth rates and per individual effects of radiata pine may result in strong effects on naturally uncommon ecosystems and may alter successional trajectories. Local and central government currently manage radiata pine invasions while propagule pressure from existing and new plantations grows, hence greater emphasis is warranted both on managing current invasions and proactively preventing future radiata pine invasions. We therefore recommend a levy on new non-native conifer plantations to offset costs of managing invasions, and stricter regulations to protect vulnerable ecosystems. A levy on economic uses of invasive species to offset costs of managing invasions alongside stricter regulations to protect vulnerable ecosystems could be a widely adopted measure to avert future negative impacts.
Abstract Tree species in the Pinaceae are some of the most widely introduced non-native tree species globally, especially in the southern hemisphere. In New Zealand, plantations of radiata pine (Pinus radiata D. Don) occupy c. 1.6 million ha and form 90% of planted forests. Although radiata pine has naturalized since 1904, there is a general view in New Zealand that this species has not invaded widely. We comprehensively review where radiata pine has invaded throughout New Zealand. We used a

Introduction
Some of the most widely introduced tree species globally are Pinaceae, deliberately established for a range of services including timber, fibre, amenity value and soil conservation (Richardson et al. 1994;Essl et al. 2010;McGregor et al. 2012). Nearly 4% of the world's forests are plantations, established primarily to provide timber and wood products (Pawson et al. 2013). In the southern hemisphere, non-native Pinus species are the most widely grown in plantations (Richardson et al. 1994). Radiata pine (Pinus radiata, or Monterey pine) is native to North America, where it is narrowly distributed along the Californian coast, but is one of the most widely planted tree species worldwide (> 4 million ha), with the largest plantations in New Zealand and Chile (> 1.5 million ha each) and Australia (0.77 million ha; Mead 2013).
In New Zealand, radiata pine is by far the largest contributor to its forestry industry, comprising 90% of the total plantation area, supplies most domestic wood products, and is the third largest export earner, contributing c. 3% of GDP (Roche 1990(Roche , 2017NZFOA 2020). The near-complete reliance of New Zealand's forest industry on radiata pine is remarkable internationally (Castro-Díez et al. 2019). Its economic benefits are well documented, but there is a view that this species is relatively unimportant as a biological invader despite its widespread introduction effort (Froude 2011;Wyse and Hulme 2021a, b; see also Williams and Wardle 2007).
The potential negative environmental and ecosystem impacts of invasive non-native pines are well documented, resulting from a combination of high introduction effort, subsequent widespread naturalization and invasion, and relatively high growth rates and early reproductive maturity (Richardson et al. 1994;Simberloff et al. 2010;McGregor et al. 2012;Rejmánek and Richardson 2013). However, invasive pines vary widely in both biomass and potential effects across different ecosystems and scales (Rouget et al. 2001;Williams and Wardle 2005;Curtis et al. 2019). This has prompted major efforts to better understand the causes and consequences of pine invasion, and how to integrate this knowledge into management (e.g., Nuñez et al. 2017;Sapsford et al. 2020). Because of widespread increases in the distribution, abundance and ecological impacts of nonnative pines, there are increasing efforts to contain or remove these invaders at the landscape scale (Froude 2011;Rundel et al. 2014;Nuñez et al. 2017). What is unresolved is how landscapes are managed for species that can simultaneously have benefits and negative impacts depending on their location (Sapsford et al. 2020); radiata pine is an excellent species for evaluating these complexities.
Radiata pine has naturalized from commercial plantations and amenity plantings and invaded native ecosystems across the southern hemisphere including Chile (Bustamante and Simonetti 2005;Becerra and Montenegro 2013), Argentina (Franzese et al. 2020), Australia Wardle 2005, 2007;Baker and Murray 2010;Calviño-Cancela and van Etten 2018), and South Africa (Richardson and Brown 1986;van Wilgen et al. 2016). In many of these systems, positive feedbacks between radiata pine invasion and fire are common because pines contribute more flammable litter than native species (Baker et al. 2007;Franzese et al. 2020) and can promote crown fires (Richardson and Brown 1986;Burrows et al. 1989). Many invasive non-native Pinaceae in New Zealand, notably Pinus contorta, P. nigra, and Pseudotsuga menziesii, are widely controlled to prevent or minimize their effects on ecosystems and biodiversity (e.g., Ledgard 2001;Dickie et al. 2014a;Hulme 2020). In contrast, radiata pine naturalized by at least 1904 but has been considered "only a minor escape" that is only "occasionally extensively naturalized in scrub and herbaceous vegetation downwind from plantations" (Webb et al. 1988), and this view of radiata pine as a minor invasive species has prevailed for over three decades. For example, its characterisation as an invasive weed (as it is in Australia; Baker and Murray 2010) does not feature in policy decisions in New Zealand. A major afforestation initiative in New Zealand ("One Billion Trees", https:// www. teuru rakau. govt. nz/ fundi ng-and-progr ammes/ fores try/ one-billi on-trees-progr amme/) will likely increase the area of radiata pine plantation and thus propagule pressure, yet its potential as an invasive species and its effects in natural ecosystems has not been considered. We are now at the stage of requiring evidence for invasion risk and potential impacts at the local to regional scales needed to guide policy and management of both plantation forestry and environmental weeds (Hulme 2020).
More generally, New Zealand has developed approaches to conservation and biosecurity management relatively rapidly at a national scale because it is a small island nation, some of which, such as eradication of invasive mammals from islands and species translocations, have been applied globally (Simberloff 2019). In the case of invasive plants in New Zealand, high numbers of species from across major biogeographic regions coupled with long naturalization histories have prompted concerted national management responses, providing general insights into the largescale management of invasive species and prioritisation of plant biosecurity risks worldwide Hulme 2020;Sapsford et al. 2020;Brandt et al. 2021).
In this paper, we reassess Webb et al.'s (1988) view of radiata pine as only a minor invasive species in New Zealand, more than 30 years later. We present the first comprehensive review of where invasive radiata pine occurs in New Zealand, and evaluate the potential climatic suitability for radiata pine to invade nationally. We assess radiata pine invasions in three classes of natural ecosystems that are naturally uncommon and geographically confined ), many of which are threatened by increasing dominance by invasive, non-native plants (Holdaway et al. 2012). We also evaluate the extent to which radiata pine is a feature of successional ecosystems, including primary successions and secondary successions. An overarching goal of this work is to provide the evidence needed to improve management of non-native species at the landscape-scale that simultaneously provide benefits and negative impacts that are location-or context-dependent (Hulme 2020;Sapsford et al. 2020).

Current distribution of invasive radiata pine in New Zealand
To identify occurrences of invasive radiata pine in New Zealand, we used: (i) plot records from the National Vegetation Survey (NVS) databank (Wiser et al. 2001; https:// nvs. landc arere search. co. nz/ on 1 May 2020); (ii) herbarium records from the New Zealand Virtual Herbarium (https:// www. nzher baria. org. nz/ on 13 June 2019) and physical records in individual herbaria; (iii) records from the citizen science initiative i-Naturalist (https:// inatu ralist. nz/ on 17 July 2019) where locations and photographs were clearly marked as invasive, with additional observations from Global Biodiversity Information Facility (GBIF; https:// doi. org/ 10. 15468/ dl. lowakv on 9 November 2018). Our approach was conservative, scrutinising all potential records to ensure that we admitted only documented invasions. We determined invasions on offshore islands from floras of 264 northern New Zealand islands (spanning 34-38°S and 172-179°E; Mologni et al. 2021), and records from other islands (Atkinson 1997). To establish the current environmental range of invasive radiata pine occurrences, we extracted the 0.5% and 99.5% percentiles from 100 m resolution environmental layers of elevation, mean annual temperature, and annual precipitation for each of these records (McCarthy et al. 2021).

Potential distribution of radiata pine in New Zealand
A global set of reliably georeferenced radiata pine occurrences was obtained from GBIF (GBIF. org 2020). Global climatologies for the period 1979-2013 with a 30 arc second (ca. 1 km) grid resolution (Karger et al. 2017) were used to define climate space. We used mean annual temperature and annual precipitation because these variables are the strongest predictors of species distributions (Barbet-Massin and Jetz 2014) and easily interpreted. To maximize the suitability of the radiata pine occurrences for niche analysis, GBIF occurrences were retained if they were: (1) temporally consistent with climatologies (i.e., recorded during 1972-2020); (2) recorded with a known spatial uncertainty ≤ 1 km (i.e., to ensure spatial consistency with the climatologies); (3) located within 0.1 degrees of the country to which they were recorded, located inside the climatology data range, not located within a 1 km grid for another observed occurrence, and not recorded as managed or cultivated.
Because we were interested in where radiata pine could potentially occur, we modelled its fundamental niche as a multivariate ellipse (Soberón and Peterson 2020) derived using Mahalanobis distances . To meet assumptions of multivariate normality annual precipitation values were log-transformed, and then both mean annual temperature and annual precipitation data were rescaled to z-scores such that the units of climatic space were the same in all dimensions. Because GBIF data are known to be highly biased (Meyer et al. 2016), a climatic space filter was applied to reduce the effects of sampling bias in climatic space by ensuring all occurrences were at least 0.1 z-scores away from any other occurrence (Varela et al. 2014).
After cleaning the data for errors and bias, 1281 observations were available for niche analysis. The fundamental niche was defined as the climatic space with a Mahalanobis distance chi-square probability ≤ 0.99, using a minimum covariance determinant approach (with k = 0.95), which is robust to any remaining occurrence errors or bias (Etherington 2021). Fundamental niche models were generated as the mean of results from 1000 bootstrap resamples (Efron 1979;Diaconis and Efron 1983;Verbyla and Litvaitis 1989;Etherington and Lieske 2019). The fundamental niche of radiata pine was then applied to New Zealand's current climate to map the potential distribution of radiata pine. The area of New Zealand that is climatically suitable to support a population of radiata pine was calculated as a weighted sum of the potential distribution values, but excluding those 1 km 2 grid cells that consisted of ≥ 50% of inland water (LINZ 2021a, LINZ 2021b). All analyses were done using R software (R Core Team 2019) with the MASS (Venables and Ripley 2002), virtualNicheR (Etherington and Omondiagbe 2019), rnaturalearth (South 2017), raster (Hijmans 2020), sf (Pebesma 2018), and rgbif (Chamberlain and Boettiger 2017) packages.

Determining ecosystems invaded by radiata pine
We extracted data from vegetation plots archived in the NVS databank where radiata pine had been recorded and data were collected following consistent relevé-based methods (Hurst and Allen 2007). Many of these plots have been assigned to a national classification that quantitatively defined woody and nonwoody vegetation types De Cáceres 2011, 2013;Wiser et al. 2016). Previously unclassified plots were assigned to a vegetation type following the procedures defined in these earlier publications, where possible. All plots were further categorized into a physiognomic class (e.g. grassland, fernland, herbfield) based on the summed cover of different species groups, following Atkinson (1985). We also reviewed published literature from across studies of New Zealand ecosystems invaded by radiata pine, and, for some of these ecosystems, we present new data.

Current distribution of invasive radiata pine in New Zealand
Observations of invasive radiata pine extend nearly the full latitudinal range of New Zealand (Fig. 1a), and were present between 1 and 1060 m elevation, 6-16 °C mean annual temperature, and 548-2813 mm annual precipitation. Invasive radiata pine occurs on 34 (12.8%) of 266 northern New Zealand islands for which there are comprehensive floras (Atkinson 1997;Mologni et al. 2021). The 34 islands are, on average, nearly 7 times greater in area than the overall mean of 266 islands (1631 ± 854 [SEM] ha vs. 242 ha), nearly half as close, on average, to the mainland (5.4 ± 1.3 km vs. 9.6 km), and more often with permanent human settlement (62% of islands are inhabited vs. 15.4%). Radiata pine has also naturalized on another 8 islands off the northern South Island coast (40.5-41.5°S); an island (45.7°S) close to Dunedin, southern South Island (Atkinson 1997); and on Rekohu/Chatham Island, c. 800 km east of the South Island of New Zealand (Fig. 1b), which is permanently inhabited. The higher incidence of radiata pine invasion on islands with permanent human settlement is likely to coincide with planting radiata pine for shelter, fuelwood, or timber, so that invasive radiata pines on many islands are most likely derived from local seed sources (e.g. Cameron and Davies 2013) rather than long-distance dispersal from mainland populations (see also Wyse and Hulme 2021b).

Potential distribution of radiata pine in New Zealand
The climatically suitable areas for radiata pine were in the warmer and drier regions of New Zealand, which is consistent with radiata pine's native range in coastal southern California, with only the coldest and wettest climates of New Zealand predicted to be outside the 43 The right tree in the right place? A major economic tree species poses major ecological threats fundamental niche (Fig. 1b). The occurrence data conformed well to the Mahalanobis distance method's assumption of multi-variate normality, and the results from the bootstrap resampling were highly consistent, indicating the fundamental niche model is reliable (Fig. 1b). Therefore, the potential distribution of radiata pine is likely to include most of New Zealand aside from relatively high elevation (> 1000 m) and high rainfall (> 2000 mm year −1 ) regions of the centre and west, especially southwest, of the South Island (Fig. 1a). While confirmed presences of radiata pine in New Zealand are currently patchy and localized, the total area that is climatically capable of supporting populations is estimated as 76% of New Zealand's land area (or 211,388 km 2 ; Fig. 1a).

Ecosystems invaded by radiata pine in New Zealand
Of 101 relevé plots nationally that we were confident contained invasive radiata pine, 56 had been collected in a manner consistent with those supporting previous quantitative vegetation classifications. Of the 56 plots with suitable data, 31 (55%) could be classified to a previously defined vegetation type, whereas the composition of the remainder was not consistent with previously defined vegetation types. Instead, these were assigned to physiognomic types (Table 1). There were an equal number of plots in grasslands, fernland and herbfields versus successional shrublands in which radiata pine had invaded (21 plots in each of the groups, both 38% of the Fig. 1 Current invasive range and ecological niche modelling of the fundamental niche and potential distribution of nonnative Pinus radiata in New Zealand. a The national 25 km grid squares containing confirmed occurrences of invasive P. radiata; potential distribution of P. radiata (blue shading) is predicted by the species' fundamental climatic niche. b The estimated fundamental niche (blue shading) of P. radiata cal-culated from global native and non-native occurrence data using Mahalanobis niche modelling. For context, contours show the climate space containing 55, 75, and 95% percentiles for the current climatic conditions of New Zealand, and the histograms show the distribution of occurrences for each climate variable total) and fewer plots in forests were invaded (14 plots, 25% of the total).

Naturally uncommon ecosystems
Radiata pine has invaded at least three classes of naturally uncommon ecosystems (Table 2 (1) geothermal, (2) gumlands, and (3) inland cliffs, scarps, and tors. Geothermal ecosystems include heated ground (dry), hydrothermally altered ground (now cool), acid rain systems, fumeroles and geothermal streamsides. Vegetation structure ranges from nearly bare ground to moss and herbfields to shrublands to forest. Radiata pine invades the woody vegetation (Burns 1997;Boothroyd 2009;Beadel et al. 2018), where it can become structurally dominant, occupying areas > 10 ha (Table 2). Increasing ground temperatures in these ecosystems can kill mature radiata pine trees; conversely, cooling substrate temperatures can allow more radiata pine to establish. Radiata pine and Pinus pinaster are the principal invaders in geothermal areas (Burns 1997;Beadel et al. 2018).
Gumlands are mosaics of dry and wet heathlands in warm temperate northern New Zealand on infertile, podzolized soils that formerly supported forests dominated by Agathis australis (Enright 1989), and are locally prone to waterlogging, often because of silica clay or ironstone pans (Clarkson et al. 2011). These ecosystems are fire-prone, with fire disturbance likely to have increased through human ignition (Clarkson et al. 2011). Radiata pine invades gumlands throughout their range (Table 2; also Wilcox 2014), along with other non-native pyrogenic species such as Hakea sericea and Ulex europaeus. Radiata pine grows relatively rapidly and taller than Table 1 Summary of different vegetation types in New Zealand in which invasive P. radiata has been recorded as present in relevé plots Plots invaded were either assigned to non-woody (Wiser et al. 2016) or woody De Cáceres 2011, 2013) vegetation types (number of plots per type) or, where this was not possible, categorized into physiognomic classes, following Atkinson (1985) Vegetation Plots ( The right tree in the right place? A major economic tree species poses major ecological threats the resident native woody gumland vegetation (c. 1 m tall or less; Esler and Rumball 1975;Clarkson et al. 2011), becoming locally structurally dominant, and potentially hastening succession from short heaths to forest that occurs on some gumlands (Enright 1989).
Inland cliffs, scarps, and tors formed from bedrock of varying chemistry occur throughout New Zealand ). Radiata pine invades at least two of these kinds of ecosystems, on basalt and quartz sandstone (Table 2), usually as isolated individuals but, when established, attains far greater height and biomass than resident native plant species.

Primary successions
Radiata pine often invades with other Pinaceae species during primary succession (Table 3). On volcanic substrates, it is more common than Pinus pinaster as an invader of lava and scoria on Rangitoto (Brown and Wilcox 2007), and mainly invades scoria at lower elevations on Mount Tarawera ( Table 2). Much of New Zealand is mountainous and landslides are common because of unstable underlying bedrock, tectonism, and high rainfall events and storms including cyclones (e.g., Marden and Rowan 1993). Radiata pine is often planted to reduce the likelihood of landslides in erosion-prone terrain, yet it can invade landslides that form within extensive areas of natural forest (> 2000 km 2 ; Table 3). Radiata pine has been widely planted on mobile coastal sand dunes, initially to stabilize them to prevent sand movement into agriculture, and later, once pine growth rates were found to be rapid, as a timber crop (McKelvey 1999). Seedlings of radiata pine establish on stable sand, and this study supports Wardle's (1991) prediction that, wherever there is a seed source in areas with 500-6000 mm annual rainfall, radiata pine will increasingly dominate primary succession on coastal sand dunes (Table 3).

Secondary successions
Radiata pine invasions are also widespread in secondary successions (Table 4), with radiata pines rapidly growing in height and becoming emergent over the resident successional communities. Invaded ecosystems are mostly low elevation grasslands and shrublands (< 400 m a.s.l., Table 4) including those dominated by native as well as by non-native species (Table 1). Secondary successions invaded by radiata pine are often in areas previously used for agriculture and then either abandoned or farmed less intensively (Standish et al. 2008;Wilson 2008 (Gardner 1993).

Discussion
The evidence we present challenges previous (Webb et al. 1988) and current views that radiata pine is a minor invader in New Zealand (see also Wyse and Hulme 2021a, b). Rather, radiata pine is a nationally invasive tree (Fig. 1), which occurs across primary and secondary successions (Tables 3, 4) and naturally uncommon ecosystems dominated by low statured vegetation ( Table 2). The discrepancy between Webb et al. 's (1988) view and our findings could, in part, be a function of residence time: a major expansion of radiata pine plantation area in the 1970s-1980s including into new regions (Roche 1990) would scarcely have reached reproductive maturity or dispersed seed until our analyses > 30 years after Webb et al's (1988) findings. Additionally, previous work on invasion of non-native conifers in New Zealand has focussed on montane grasslands invaded primarily by Pinus contorta, P. nigra and Pseudotsuga menziesii (e.g., Ledgard 1994;2006) whereas radiata pine invasions are most prevalent in lowland regions. Several attributes or traits of radiata pine explain both its invasiveness and absolute abundance in the ecosystems it invades (Rejmánek and Richardson 1996;McGregor et al. 2012;Wyse and Hulme 2021a, b). For example, rapid growth rates of radiata pine, in height and biomass, far outstrip any native woody species: plantations of radiata pine have annual drymatter production approximately double most native Nothofagaceae forests, tree radial diameter increments an order of magnitude greater than most native trees, and within 22 years have biomass equivalent to most old-growth native podocarp-broadleaf and Nothofagaceae forests (Wardle 1991;McGlone et al. 2022). The effects of invasive radiata pine on ecosystem productivity and function, whether measured by biomass (Grime 1998), litterfall, or asymmetric competition resulting from its greater height and diameter growth, are likely to be large in the immediate neighbourhood of any individual that establishes (Sapsford et al. 2020; see also Franzese et al. 2017). Moreover, radiata pine is functionally distinct from native-and many non-native-dominated communities it invades because of its novel ecotomycorrhizal associations (Dickie et al. 2010) and relatively high flammability (Perry et al. 2014;Gómez-González et al. 2022).

Invasion current range and potential
Our ecological niche modelling demonstrates that most climatic conditions across New Zealand occur within radiata pine's estimated fundamental niche and are thus climatically suitable for its establishment (Fig. 1). Relatively few observations of invasive radiata pine were recorded for some colder or wetter regions, but absence of occurrence data is not strong evidence for climatic unsuitability because of observer biases (Atwater et al. 2018); therefore, our results should be interpreted as minimum estimates of the fundamental and potential niche. Wide national distribution was also described by Howell (2019) based on herbarium records within broad ecological regions (McEwen 1987), where radiata pine occurred in 56 regions-more than twice as many as other invasive Pinaceae species that currently attract greater concern and management (Pinus contorta and P. nigra, 23 regions each; Pseudotsuga menziesii, 26 regions; Larix decidua, 17 regions). In addition, our estimates are based on current climate conditions but if New Zealand's climate warms and dries, as future climate scenarios predict, then radiata pine's potential distribution is likely to increase ). This example adds to a growing recognition internationally that rates of naturalization and invasion are dynamic, either dampening or accelerating under future climates, thus requiring better understanding of how climate and invasions interact to predict future invasion risk (Caplat et al. 2013).
In addition to presence-only observational data used to estimate species distributions, we used both plot-and site-based data to better understand the scale and importance of radiata pine invasion. We observed relatively low incidence of radiata pine invasion from plot-based data at the 0.04 ha scale but much higher incidence for site-based data at the > 1 ha scale (Table 2). These differences between data sources reflect both different sampling objectives and ability to detect early invasions or any sparsely distributed population. For example, plot-based data typically quantify absences or the abundance of species in communities, but are thus sensitive to replication, stratification by vegetation or land use type, spatial bias and representativeness (e.g., Holdaway et al. 2014). In contrast, site-based surveys typically generate presence-only qualitative data about species but have larger spatial grain and higher detection probability for populations compared with plot-based data but are uninformative for some processes such as habitat suitability (Garrard et al. 2008). Both approaches can be combined to scale up low probability distributions or occurrence of plant populations (e.g., Peltzer et al 2008; see also Isaac et al. 2020), or alternatively, hierarchically nested plots can be used to evaluate the scale-dependence of observations (e.g., Stohlgren et al. 1995). Overall, a combination of plot-and sitebased data can be used in combination to understand current and future distributions of biological invaders, and to better understand changes in invasion rates or ecological effects. This approach is generalisable across species and different regions (see also Wilson et al. 2018;Etherington et al. 2022).
There is little information on abundance of radiata pine in New Zealand, aside from plot-based assessments within a few sites. Such information is required to confidently estimate impacts at larger spatial scales using, for example, EICAT or SEICAT systems (Blackburn et al. 2014;Bacher et al 2018). Even without this information, studies on invasive pines in New Zealand and internationally show that high ecological impacts of radiata pine are likely, based on at least large range and high per-unit effects (Parker et al. 1999;Sapsford et al. 2020;Latombe et al. 2022). However, studies of the per-unit effects of radiata pine in New Zealand ecosystems are needed to inform programmes to manage its invasion.
Invasive radiata pine as a transformer of naturally uncommon ecosystems Invasion by trees into largely treeless habitats is a global concern (Rundel et al. 2014), and the spread of non-native radiata pine into naturally rare ecosystems exemplifies this issue (Table 2). Radiata pine invasion in gumland heaths is analogous to its invasion of mesic fynbos in South Africa (Richardson and Brown 1986); in both systems, initial invasion by radiata pines was closest to plantations. As in invasions of fynbos, even young radiata pines grow rapidly in height to exceed the c. 1 m tall gumlands and achieve far greater biomass and height than the resident flora in < 10 years, and can positively feedback to increase fire frequency, thus maintaining low-statured and invasible vegetation (Clarkson et al. 2011;Tepley et al. 2018;Wyse et al. 2018). Invasion by radiata pine on inland cliffs, scarps, and tors, likely pre-empts native woody plants from re-establishing following deforestation of these systems through a combination of relatively high propagule pressure, growth and occupancy of restricted habitat or microsites for establishment (Wiser and Buxton 2009). We documented radiata pine invasion into three broad classes of naturally uncommon ecosystems but, since unpublished data shows it invades many other rare ecosystems (e.g., coastal cliffs, gravel beaches), data for the extent of radiata pine invasion and ecosystem effects throughout the invasive range of radiata pine worldwide is warranted to understand effects on rare ecosystems, and for prioritising management (e.g., by assessing its invasion against the IUCN Red List of rare ecosystems; Holdaway et al. 2012).

Invasive radiata pine in successions
Multiple lines of evidence point to increasing potential effects of radiata pine in successions. Its rapid accrual of biomass compared with co-occurring native and most other non-native species suggests that the per-individual effects of this species are high (Sapsford et al. 2020), but population-level impacts are currently low because many sites are in the early stages of the invasion process. This suggests that for many sites or successions, early detection and containment remain the most effective management strategies for radiata pine Hulme 2020).
Radiata pine may also fill a 'functional gap' in the New Zealand flora (Dansereau 1964;Wilson and Lee 2012), exerting relatively large potential impacts through functional dissimilarity (Wardle et al. 2011). This supports a view that deflected trajectories in primary succession can be driven by "pre-emptive invasion of unusual species" that alter soil nutrient concentrations and microclimates (Walker and del Moral 2003). For example, Pinus species colonize recent surfaces generated by volcanic eruptions globally (e.g., Tagawa 1964;del Moral and Wood 1993;Ohtsuka et al. 2013) and, in New Zealand, radiata pine colonises lava flows (Brown and Wilcox 2007) and ejecta (Clarkson and Clarkson 1991) on which there is no functional analogue in the native flora. Potentially indicative of its ecosystem effects, Pinus nigra planted on lava flows in Sicily caused weathering of primary minerals throughout the soil profile, depletion of base cations, and release of aluminium in the topsoil (Certini et al. 2001). Radiata pine invasions can also alter secondary successions through co-invasions with mutualists; alteration of plant community composition; and feedbacks with fire regime (Tepley et al. 2018;Sapsford et al. 2020). For example, radiata pine can exacerbate fire intensity on islands thus shifting secondary successions toward a subset of more fire-tolerant species than would otherwise occupy these islands (e.g., Kawau and Great Barrier (Aotea) Islands; Gardner 1993;Wyse et al. 2018). Together, these observations suggest that invasive pines have strong potential to alter both primary and secondary successions, but empirical tests of these impacts both in New Zealand and more generally are scarce, suggesting that the mechanisms and magnitude of these effects warrant renewed investigation.
Both plot-and site-based data revealed that radiata pine rarely invades alone to generate monodominant vegetation, but often co-occurs with other nonnative species. What is largely unknown is how novel combinations of non-native species interact to either amplify or dampen their effects in communities or ecosystem processes. For example, co-invasion by radiata pine and nitrogen-fixing, woody non-native species such as Scotch broom (Cytisus scoparius) and gorse (Ulex europaeus) might have larger biogeochemical impacts than either species alone because of complementary effects on P and N respectively. Similarly, many invasive plant species in New Zealand that co-occur with pines can alter fire behaviour or may exacerbate fire behaviour . Moreover, these same species may benefit from increased fire intensity or frequency through increased post-fire establishment (e.g., Hakea sericea in gumlands) or reductions in abundance of less fire-resistant native species. How radiata pine alters successions either directly, or through indirect effects including with co-invading species, is largely unknown. However, given both the global occurrence of pine invasions into successional systems, a greater understanding of the role that invasive pines like radiata pine have in ecosystems is required for shaping ecosystem restoration and conservation efforts (e.g., Becerra & Montenegro 2013;Taylor et al. 2017;Nuñez et al. 2021). Greater emphases are needed on building the evidence base for impacts however, including primary data for invader distribution and abundance (Wilson et al. 2018) and their impacts on social and ecological systems (Blackburn et al. 2014;Bacher et al 2018). Such evidence is required both for multiple purposes including prioritisation of landscape-scale management and developing cost-benefit analyses for supporting investment in invader management (e.g., van Wilgen et al. 2014van Wilgen et al. , 2016Nuñez et al 2021).

Dispersal and invasions of radiata pine
Despite international evidence of high potential invasion risk by radiata pine (e.g., its z-score of Rejmánek & Richardson 1996; see also Grotkopp et al. 2002;McGregor et al. 2012), a commonly held view in New Zealand is that it mostly invades close to established populations because it has poorer dispersal ability and slower maturity than some other invasive pine species (Wyse and Hulme 2021a). Several lines of evidence suggest that radiata pine can invade further than previously expected. Observational data from known point sources to invasive populations suggest that long-distance dispersal is possible. For example, downwind colonisation of landslides throughout a large tract of native forest, Te Urewera, is most likely to originate from propagule sources at Kāingaroa > 20 km away (Shaw 1990); this may be due in part to sheer spatial mass effects, where the source plantation is ca. 190 000 ha. The role of birds or mammals in radiata pine seed dispersal in New Zealand is unknown, but long-distance dispersal of radiata pine is mediated by parrots in Australia (Batianoff 2005) and pigs in Argentina . Recent work on seed traits of radiata pine validated against empirical observations of dispersal from known sources demonstrates that some propagules within individual trees have large wings that allow them to travel at least 1.5 km at wind speeds of only 2 m s −1 (Wyse and Hulme 2021b;Wyse et al. 2022), and higher windspeeds are common in New Zealand, and cause long distance dispersal of radiata pine from plantations in the South Island (Hunter & Douglas 1984). Long-distance dispersal of radiata pine over several km should thus be anticipated rather than dispersal over a few hundred metres as is currently assumed and used by managers (e.g., in the 'wildings spread risk calculator' to predict spread risk from plantations; see also Froude 2011;Strang et al. 2015;Wyse and Hulme 2021a,b).

Managing invasions of an economically important tree
A major consideration of managing some invasive plant species is that, depending on the site considered, the same species can be an environmental weed in one situation and economically beneficial in another (Dickie et al. 2014b;Nuñez et al. 2021). This is exemplified by goals within New Zealand's national strategies of "the right tree in the right place" (MPI 2014, national wilding conifer control strategy) and "right tree, right place, right purpose" (One Billion Trees afforestation initiative: https:// www. mpi. govt. nz/ fores try/ fundi ng-tree-plant ing-resea rch/ onebilli on-trees-progr amme/ about-the-one-billi on-treesprogr amme/). More generally, because non-native species can be viewed as either beneficial or harmful depending on environmental and social context, this can generate conflicting views on their management (Dickie et al. 2014b;Edwards et al. 2020;Yletyinen et al. 2021). As such, the costs and benefits of invasive species' impacts or their management require information on both ecological effects alongside information on social and economic values.
Rigorous cost-benefit analyses demonstrate accelerating costs of managing species as invasion proceeds, and some benefits of species for economic activities or carbon sequestration that are readily quantified (e.g., Wise et al. 2012;Wyatt 2018; see also Castro-Díez et al. 2019). However, the value of some processes, services or impacts of invasive species cannot easily be quantified, or have complex responses to invader abundance (e.g., Mason et al. 2017;Panetta and Gooden 2017;Bartz and Kowarik 2019). For example, a cost-benefit analysis for non-native Pinaceae in New Zealand demonstrated a 38:1 benefit:cost (Wyatt 2018). However, this analysis was highly conservative because it only considered costs to primary industry through invasions, water availability, fire risk, and future savings to management costs over 50 years, and it did not include costs to biodiversity or carbon because of difficulties in valuing damage and uncertainties around markets respectively. Economic justifications for invader management have been well developed elsewhere, for example, for non-native tree invasions in South Africa (see de Wit et al 2001;Wise et al 2012). Cost-benefit analyses could be more widely applied both to understand the net ecological, social and economic effects of invasions and their management, but also aid to understand how short-term management or investments can have longer-term benefits (van Wilgen et al. 2014).
National cost-benefit analyses are likely to oversimplify the complex nature of effective management at regional or site scales. For example, many invasive plant management efforts either fail to meet their stated management goals or do not carry out rigorous evaluations of success (e.g., Wilson et al. 2014) including the wider values or benefits beyond invader removal that typically occur at the site scale (Reid et al. 2009). Implicit goals of invasive pine management can range from conservation of sites and native species to sustained water yield in catchments and maintenance of landscape values (MPI 2014;Gawith et al. 2020;Edwards et al. 2020). As a consequence, there are opportunities to more closely link ecological knowledge with management to avoid or mitigate the impacts of major biological invaders such as radiata pine include resolving: (1) when do low-density invasions matter? Eradication is rarely feasible and ongoing invasion requires knowledge of abundance-impact relationships to avoid negative effects or legacies of invasions (Sofaer et al. 2018;Sapsford et al. 2020). Similarly, evidence is essential to determine how much control is needed relative to the distribution and abundance of the invader, and how much it would cost to meet management goals (e.g., van Wilgen et al. 2016); (2) what evidence is required to evaluate management success? Internationally there is a move to collect a minimum set of core data to better understand and manage tree invasions (Wilson et al. 2014(Wilson et al. , 2018Brundu et al. 2020). Because the outcomes of invasive species removal are variable (Kettenring and Adams 2011), it is also essential to evaluate the performance of different management strategies (Funk et al. 2020); and (3) understand the potential future risks of invasion. Given predicted changes to climate, land use, and disturbance regimes, we should anticipate that invasion risk is dynamic over decades and planning land use or management strategies should consider this dynamism (Caplat et al. 2013;Macinnis-Ng et al. 2021;Etherington et al. 2022). As an example, management of 'wilding conifers' in New Zealand has a clearly stated national goal of stopping or containing invasion by non-native conifers by 2030, estimated to cost > $110 million USD (Wyatt 2018) and, in some areas (e.g. Northland Region and coastal parts of Marlborough District), most of these costs focus on controlling invasive radiata pine. Actual costs of controlling radiata pine invasions are currently hard to determine because agencies often concentrate on multiple species of Pinaceae, amongst which radiata pine is sometimes the most abundant invader. More explicit costs of its management will be needed, including costs resulting from re-invasions, both from incomplete management, and ongoing propagule pressure from existing and new plantations. How long-term management of invasive pines in New Zealand is unresolved. Understanding the costs and benefits associated with invasions and their management over the longer-term is essential to developing approaches to both value and invest in such large-scale and long-term efforts (see also Graham et al. 2019;Peltzer et al. 2019).
In light of our findings, we recommend the following changes to afforestation and biosecurity policies and legislation in New Zealand: (1) a levy on new non-native conifer forests; and (2) stricter afforestation regulations, incorporating recommended changes and addressing gaps recently identified (Wyse et al. 2021a(Wyse et al. , 2021b. The levy could be used for surveillance and clearance of new invasions, and stricter afforestation regulations should mitigate the risk of future conifer spread including into areas that climate change renders more suitable for colonization . Since radiata pine is the most likely species to be planted in new areas and has the largest propagule pressure, these changes to policies will be critical to curb ongoing invasion. A more detailed assessment of current investments in management, both direct and in-kind, is needed to determine whether such a levy alongside other investments, is sufficient to achieve the national goal of stopping or containing invasions. Similar approaches have also been proposed in South Africa to offset the impacts of plantations on water availability, but were unsuccessful because the levy was insufficient to compensate for the negative impact on water or ongoing invasion (Tewari 2005).
Recent syntheses demostrate the growing direct and indirect economic costs of invasive species globally (e.g., Diagne et al. 2021;Welsh et al. 2021;Zenni et al. 2021). However, the costs and benefits of nonnative species vary among locations, and along the invasion pathway. For example, economic costs of invader management are often borne later, and in different locations, from original sources (Lovett et al. 2016). Here we provide an example to help overcome this issue for invasive conifers, which are economically valuable species but also major invaders globally (Richardson and Rejmánek 2004;Nuñez et al. 2017). In New Zealand, most costs of managing invasive conifers are currently met by by short-term central government funding and long-term stable mechanisms to fund management are lacking (Hulme 2020). Levies would address the large commercial plantations that provide propagule pressure but other measures will also be needed to fund long-term management. Propagule pressure from invasive trees, including radiata pine, derives from widespread noncommercial plantings for soil conservation, amenity and shelter, thus management requires multiple incentives, regulatory or economic tools for their management (e.g., Hulme 2020; Yletyinen et al. 2021). The findings and solutions presented here are globally relevant and provide critical information to improve management in other countries at earlier stages of managing these invaders ; see also Brundu et al. 2020).

Conclusion
We used plot-and site-based data and ecological niche models to demonstrate that invasive radiata pine occurs far more widely across New Zealand than previously appreciated. Niche modelling based on climatic suitability suggests that its distribution could be much wider than current plot or observational data indicate. The wide distribution of invasive radiata pine is partly explained by its wide introduction as a plantation and amenity species, but also reflects far greater spread risk than previously assumed because of long-distance (at least 1.5 km) seed dispersal. Radiata pine has potentially large, but largely undocumented, effects in many primary and secondary successions and across multiple naturally uncommon ecosystems. Although management of invasive radiata pine is already widespread in New Zealand and other southern hemisphere countries, more proactive planning is needed given ongoing large-scale afforestation efforts and potential future increases in invasion. A levy on economic uses of invasive species to offset costs of managing invasions alongside stricter regulations to protect vulnerable ecosystems could thus be more widely adopted to prevent or avert future negative impacts.
Acknowledgements We acknowledge the use of data drawn from the National Vegetation Survey Databank (https:// nvs. landc arere search. co. nz/), including the use of data drawn from the Natural Forest plot data collected between 2002 and 2014 by the LUCAS programme for the Ministry for the Environment. We thank knowledgeable practitioners for information about invasive conifer observations and management: Brad Lett (formerly Department of Conservation, Murupara) for western Te Urewera; Ceres Sharp and Don McKenzie (Northland Regional Council) for Northland; and Andrew Macalister for the Marlborough Sounds. Hamish Maule assisted with determining the invasion status of occurrence records. Fabio Mologni provided calculations of island invasions. We thank three anonymous reviewers for thoughtful and constructive comments on this manuscript.
Author contributions PJB and DAP developed initial ideas for this paper, all authors made conceptual and written contributions to the manuscript, and interpretation of results.
Funding Open Access funding enabled and organized by CAUL and its Member Institutions. We thank the New Zealand Ministry of Business, Innovation and Employment for financial support through the Endeavour research programme Winning Against Wildings (Grant Number C09X1611).
Availability of data and material Data will be made publicly available upon manuscript acceptance. (as a time-stamped collated dataset in DataStore: https:// www. landc arere search. co. nz/ tools-and-resou rces/ datab ases/ datas tore/). Code availability Custom code for the niche models will be made publicly available upon manuscript acceptance.

Conflicts of interest
We declare no conflict of interest.
Consent for publication All authors consent for this manuscript to be published in Biological Invasions.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The right tree in the right place? A major economic tree species poses major ecological threats The right tree in the right place? A major economic tree species poses major ecological threats