Invasive plants have profoundly reshaped the structure and function of ecosystems globally (Vilà et al. 2011). Once established, these species can modify ecosystem functions, alter community structure, and reduce biodiversity (Hejda et al. 2009; Vilà et al. 2011; Le Maitre et al. 2015; Dickie et al. 2022). Among the most impactful invaders are woody species, often originally introduced into a variety of landscapes for horticulture, agriculture, or forestry (Richardson and Rejmánek 2011). Conifers belonging to the Pinaceae family are recognized as particularly important woody invaders, exerting significant ecological and economic impacts, especially in Southern Hemisphere countries such as South Africa, Chile, Argentina, and New Zealand (Richardson and Rejmánek 2004; Nuñez et al. 2017; Torres et al. 2018; Martin et al. 2025). Disturbance, such as grazing, mechanical clearing, and fire, has been a key factor influencing the global spread of invasive conifers, primarily by creating recruitment opportunities and weakening biotic resistance (van Wilgen and Richardson 2012; Nuñez et al. 2017; Gioria et al. 2023; Green et al. 2025).

In New Zealand, non-native conifer species, such as Pinus contorta Loudon and Pseudotsuga menziesii (Mirb.) Franco, pose significant environmental and land-use challenges, invading a wide range of natural and semi-natural ecosystems (Howell 2019; Sapere Research Group 2022). It is estimated that over 2 million hectares of New Zealand are invaded by non-native conifers, or ‘wilding’ conifers, with ~ 80% of affected land classified as sparsely infested (1–15% cover), 15% as intermediate (15–75% cover), and 1.5% as dense (75–100% cover) (Sapere Research Group 2022). New Zealand’s wilding-invaded landscapes have undergone extensive ecological transformation since Polynesian and later European settlement, driven by widespread and repeated burning and extensive forest clearance (McGlone 1989). Pastoral land use, especially in the inland South Island, further accelerated change through sustained burning, intensive grazing, fertilization, and the introduction of exotic grasses and herbs, pest animals, and weeds (McGlone 2004; McIntyre 2007). In the mid-20^th century, government-led afforestation initiatives introduced several exotic conifers to stabilize erodible soils and provide shelter for livestock. Many of these plantings have since naturalized and now contribute to the widespread conifer invasion, particularly in high-country semi-natural grasslands (Ledgard 2001). Management to contain or eliminate non-native conifers has been ongoing in New Zealand since the late 20^th century. These efforts were expanded substantially following the establishment of the National Wilding Conifer Control Programme in 2016, and more than 1 million hectares have since received at least one round of control (Biosecurity New Zealand 2025). However, to date, the effectiveness of these operations in controlling the existing pine population and the subsequent reinvasion dynamics has not been quantified.

Comprehensive, nationally coordinated programs to manage or control invasive alien plant species are relatively rare (e.g., Hoffmann et al. 1998; Dew et al. 2017; Nunez et al. 2017; Dickie et al. 2022), with smaller and more targeted regional programs more frequently reported (e.g., Constán-Nava et al. 2010; Lombard et al. 2012; Lindenmayer et al. 2015; Pauchard et al. 2015; Ernst-Brock et al. 2019). South Africa and New Zealand stand out for their national responses to invasive pines, with significant efforts also undertaken in Chile and Argentina (van Wilgen et al. 2012; Ministry for Primary Industries 2014; Nunez et al. 2017; Martin et al. 2025). South Africa’s Working for Water Programme, launched in 1995, was one of the largest and most costly government-funded initiatives globally for managing invasive alien plants, though outcomes have been mixed (van Wilgen et al. 2012; Dew et al. 2017). South Africa has also implemented a suite of legislative and strategic tools under the National Environmental Management: Biodiversity Act, including Alien and Invasive Species Regulations and the National Invasive Species Strategy (Department of Environmental Affairs 2014), which guide prioritization, permitting, and compliance.

New Zealand is similarly recognized for its integrated and science-led approach to invasive species management, spanning border biosecurity, national eradication programs, and landscape-scale pest control (Peltzer et al. 2019; Simberloff 2019). In response to the extensive spread of non-native conifers, the government established the National Wilding Conifer Control Programme (NWCCP) in 2016 to fund and coordinate management efforts across central and regional government agencies, landowners, and industry stakeholders [https://www.wildingpines.nz/national-control-programme/]. Since then, an estimated 70% of the known infestation area has received at least one round of control, an effort that has cost more than NZD 150 million in the 9 years since its inception. Accepted control methods include a range of manual, mechanical, and chemical treatments typically informed by the infestation density (Edwards et al. 2021; Dickie et al. 2022). Land rehabilitation or restoration post-control has not been a major focus to date.

It is difficult to gauge the success of large invasive species management programs without including formal and rigorous monitoring over the long term. However, with a predominant focus on funding as many control operations as possible, long-term monitoring is often lacking, leaving several key questions unanswered. This situation was described by Lindenmayer et al. (2015), who stated that outstanding questions often include whether the target species has been effectively controlled; whether its control is likely to lead to ecosystem recovery; whether the methods used for control affect other fauna or flora, either positively or negatively; and whether invasive species management is cost-effective in terms of its benefits relative to levels of expenditure. Multiple factors affect the success of management interventions, including initial environmental conditions (abiotic and biotic) and population structure (Dew et al. 2017; Ernst-Brock et al. 2019); the method and timing of control (Lombard et al. 2012; Lindenmayer et al. 2015; Dew et al. 2017); insufficient long-term monitoring and follow-up control (Lombard et al. 2012; van Wilgen et al. 2012; Ernst-Brock et al. 2019); and uncertainties in invader response to control efforts. Success has often been attributed to prioritizing resources toward high-risk areas, clear alignment with management objectives (e.g., Dew et al. 2017), and long-term monitoring and control, which are fundamental for preventing reinvasion and maximizing the success of past clearing efforts (Maxwell et al. 2009). Assessments of the effectiveness of control are needed not only to guide ongoing control efforts but also to evaluate whether the management approach is achieving the targeted outcomes. Managers need to know whether their efforts are keeping the spread and density of invasive plants within acceptable limits or whether scarce resources should be reprioritized.

The objective of the NWCCP has been to control invading conifer populations at the landscape scale using a variety of methods. Dense, mature infestations of conifers are typically treated using broadcast aerial application of herbicides, while emerging or scattered infestations are managed through ground-based methods such as manual pulling, felling, and cut-stump herbicide application (Richardson et al. 2020; NWCCP 2021, 2023; Rolando et al. 2024). These are the two most widely applied control approaches within the program. Other methods include aerial spot application of herbicides for isolated trees and very sparse infestations, while fire, drill-and-fill, mechanical clearing, and mulching are used less frequently (WCIS 2023). Broadcast aerial herbicide treatments were developed from a series of trials (Gous et al. 2014a, 2015a; Rolando et al. 2025), which demonstrated that herbicide mixtures containing triclopyr, dicamba, and picloram achieved >85% mortality for pines 2–12 m tall, although effectiveness declined for larger individuals. Fewer studies have evaluated the effectiveness of ground-based herbicide applications. Experimental studies have shown that pines <1 m tall are highly susceptible to low herbicide doses and that individuals up to 5–6 m tall can be effectively controlled (>85% mortality) using relatively low application rates (0.5–1 L) (Gous et al. 2014b, 2015b; Rolando et al. 2021). However, no studies have assessed the post-treatment population dynamics for either aerial or ground-based methods under operational conditions.

New Zealand is unusual in its large-scale use of aerial herbicide application for the management of dense invasive pine stands, providing a valuable opportunity to evaluate operational outcomes. The use of aerially applied herbicides is restricted in many regions of the world, particularly in the Northern Hemisphere; for example, the European Union generally prohibits aerial spraying of plant protection products (including herbicides), allowing use only under strict exceptions (Directive 2009/128/EC). In New Zealand, aerial herbicide application is generally permitted but subject to substantial regulatory constraints (New Zealand Government 1991; Standards New Zealand 2021). Given this context, control carried out by the NWCCP offers a unique opportunity to quantify the effectiveness of operational treatments applied to different infestation levels of a woody invader.

In this study, the effectiveness of operational ground-based control (GND) and aerial boom spray (ABS) for the management of invasive Pinus contorta was assessed. Specifically, the study quantified (1) treatment-induced mortality; (2) the abundance and size structure of residual populations; and (3) the relative contributions of survivors and recruits to post-treatment population recovery. It was hypothesized that both ground-based control and aerial boom spray would achieve high mortality across a range of invaded environments, but that treatment effectiveness would vary with pine height. The objective was to quantify mortality, survival, and recruitment following operational control.

A total of 146 plots (59 ABS; 87 GND) were assessed up to 4 years post-treatment (Table 2), with a total of 10,095 P. contorta pines recorded. P. contorta was the most abundant exotic conifer recorded, with an average of > 10,000 pines ha^-1 in ABS treatments and approximately 3,000 pines ha^-1 in GND treatments (considering all height classes at the time of control). Larger height classes were dominant in ABS treatments, while the opposite occurred in GND treatments (Suppl. material 3). There were 4,992 ± 4,588 individuals per hectare (mean ± SD) above 2 m tall in the plots at the time of application of ABS, compared with 222 ± 510 individuals per hectare (mean ± SD) above 2 m where GND had been applied (Suppl. material 3). In other words, there was high variability in pine density between plots even prior to control. Some plots, notably at Flock Hill and the Hanmer Range, also contained Douglas fir (Pseudotsuga menziesii), European larch (Larix decidua), and other pines (e.g., Pinus nigra, Pinus mugo, and Pinus sylvestris); in total, fewer than 200 individuals of these species were observed across all plots and were not analyzed further. Mortality varied widely between plots, ranging from all live to all dead pines in any one height class for both ABS and GND plots (mortality at the plot level is summarized in Suppl. material 3).

This is one of the first studies to critically evaluate the effectiveness of operational treatments applied to manage non-native conifer invasions. This study assessed 146 plots spanning the main invaded regions within the key management areas of national wilding control, covering a range of invasion stages and the two main control methodologies. The results demonstrate that conifer mortality following control is strongly influenced by control method, height, and time since treatment. The results did not provide evidence that the broadscale environmental variables examined affected initial control treatment success for either ABS or GND. Instead, the results suggest that climate variables influence growth rates and therefore influence population dynamics following control. Aerially applied herbicide treatments were highly effective on average irrespective of area conditions, with mortality increasing over time and reaching a maximum at 3 years post-treatment (average mortality of 95%) and being particularly high for medium-sized pines (1.35–2 m). This 3-year timeframe coincides with improving health status of the remaining pines. In contrast, GND showed mortality patterns driven by pine height and was only effective for larger individuals (>0.6 m tall). While there were similar numbers of recruits and active cone bearers 4 years after control, population structures were significantly different depending on treatment type, with post-treatment ABS populations dominated by surviving larger pines and post-GND populations dominated by surviving seedlings and saplings, with numbers of new recruits starting to match the survivors after 4 years. These outcomes reflect a combination of the pre-treatment population structure and control treatment effects and indicate that further intervention for both control approaches is critical for long-term local eradication.

Despite the high post-ABS mortality, substantial numbers of pines persisted 3 years after treatment – an estimated 291 pines ha^-1 (asymptotic 95% confidence interval: 134–632) in the 2–5 m height category and 113 pines ha^-1 (asymptotic 95% confidence interval: 50–255) in the >5 m category – due to the very high initial density of these stands (>10,000 pines ha^-1, Suppl. material 3) and the possible sheltering effect of neighboring trees in very dense stands. Delayed mortality, especially among larger trees (Figs 2, 4), is consistent with prior observations and cautions against premature evaluation of ABS effectiveness (Rolando et al. 2025). At 4 years post-treatment, approximately 72 actively coning pines ha^-1 remained – representing a substantial source of propagules. These coning individuals were primarily survivors that had recovered or escaped the spray, with very few of the latter continuing to cone in the 1–2 years immediately post-ABS (Fig. 7), suggesting that ABS may only limit coning of survivors for 3–4 years. As P. contorta cones develop over two years (Owens 2006), this pattern suggests that herbicide likely eliminates cones already in development (resulting in an initial 1-year delay), after which cone initiation can resume in the following season (with an additional 2-year lag to maturity). In addition, because P. contorta retains cones on the tree for decades, mature dense stands targeted by ABS have the potential to maintain substantial aerial seed banks on both surviving and standing dead trees (Carlin et al. 2025a, 2025b, 2025c). Carlin et al. (2025a, 2025b) showed that P. contorta cones in New Zealand produce, on average, 74 seeds per cone, with up to 19% of open cones retaining viable seed after coning. These authors found that the oldest cone retaining viable seed was 15 years old. Further, in a separate study, it was found that the number of viable seeds per cone did not differ between herbicide-treated and untreated populations, notably for serotinous cones (Carlin et al. 2025c). However, despite this potentially high propagule pressure, seedling emergence remained low beneath treated canopies during the first 2 years post-ABS despite the likely presence of abundant viable seed. There are multiple plausible mechanisms to explain this pattern. First, residual herbicide effects may contribute: field measurements indicate that, while soil concentrations are typically below detection 1 year after aerial treatment, residues persist in cast needles for up to 2 years and gradually leach out, suppressing regeneration (Rolando et al. 2023, 2024). Second, microbiome-mediated legacy effects may contribute: pot trials show that, in controlled P. contorta stands, the soil microbiome may impose negative conspecific feedback on early P. contorta growth, and the litter microbiome may impose negative conspecific feedback on seedling establishment, with some nuance arising from interacting biotic and abiotic impacts (Dudenhöffer and Hulme 2025a). Finally, there may be high predation on any seed exposed from the aerial seed bank (Carlin et al. 2024). Together, these legacy effects appear to dampen recruitment even where seed supply is high.

Ground-based control achieved high mortality for pines >0.6 m, but many smaller seedlings persisted (<0.6 m), likely due to imperfect detection during large-scale operations across mixed age classes, often in tall grass and shrubland. These residual seedlings can rapidly transition into taller classes following control (Ledgard 1999). Apparent low effectiveness of GND in the 0.6–1.35 m and 1.35–2 m height classes in warmer areas (Fig. 3) likely reflects post-treatment growth rather than low mortality of these pine sizes. Surveys occurred 1–3 years following treatment; therefore, most surviving individuals would have been shorter and harder to detect at the time of control. Moreover, ground-controlled populations are typically dominated by seedlings and saplings rather than mature trees. Faster growth at warmer temperatures accelerates upward transitions of survivors, while dead trees remain fixed in their pre-treatment height classes. Consequently, mortality estimates at cooler sites (e.g., mean annual temperature ~ 5 °C), where growth is slower, provide a clearer proxy for initial mortality under GND, indicating high kill rates for pines >0.6 m. The rapid upward transitions observed at warmer locations underscore the risk that a residual population can attain reproductive maturity within a few years in the absence of timely follow-up control.

Estimating the initial effectiveness of GND is complicated by the rapid growth of juvenile pines unless measurements are taken close to the treatment dates. Mortality is also harder to quantify because small dead seedlings and saplings can decay or evade detection under vegetation, in contrast to ABS areas where dead stems generally remain standing. Conservatively, counts in GND-treated areas may underestimate mortality in the smaller height classes, but even so, >125 surviving pines ha^-1 were still recorded in the 0.15–0.6 m height class 1 year after treatment, supporting the inference that GND effectiveness is relatively low for smaller individuals. Unlike GND, the surviving pines at ABS plots were not likely to have been growing rapidly into taller height classes during the 1–3 years after treatment due to the residual effects of the herbicides on tree health and may have even shrunk due to the loss of dead tops. However, pine densities show substantial numbers of survivors >2 m high (Fig. 5), and there is no indication that 3-year ABS mortality is materially underestimated or overestimated. Since age estimates were used to classify live pines as survivors or new recruits, it is possible that there was some misclassification. It is more likely that age was underestimated since seedlings can remain small for multiple years due to resource competition or herbivory in both GND and ABS plots or specifically at ABS plots due to plant health recovery from the herbicide treatment. This might lead to an overestimation of mortality and an overestimation of new recruits compared with survivors. Regardless, this reinforces the need for follow-up control treatments to remove the surviving pines. While pre-control treatment measurements would have strengthened the inference, the chronosequence approach practically precluded the collection of pre-treatment data, considering the scale of operations and resources available for assessment. Further, the study design avoided any risk of influencing operations.

This study provides the first quantification of the effectiveness of ground-based and aerially applied herbicide treatments for invasive P. contorta when applied at an operational scale across different environments. The data show that although both treatment types result in high mortality across environments, the density of the surviving individuals and subsequent reinvasion suggest the need for either follow-up control or a strategic shift in the control regime to achieve effective eradication or long-term containment. The residual populations in both ABS and GND cases differ in structure, with tall, mature pines being the largest cohort of individuals present following aerial application of herbicides and young, rapidly maturing seedlings in the case of GND. Managing these residual populations to optimize initial gains from control will be critical to the long-term success of the program. Beyond that, the data provide valuable knowledge for managers planning and scheduling operations and for building an understanding of the cost and impact of deferred maintenance when lack of funding prevents scheduled maintenance from occurring. This alone, however, will not be enough to restore an environment to one that is resilient to reinvasion, given the often highly disturbed and altered states of the large areas that have been and will continue to be invaded. Restoration and a plan to rehabilitate the land are critical and much-overlooked components of the management approach for the NWCCP and are ultimately needed for the long-term sustainability of the control approach.