1 Introduction

Forest managers face many climate-related risks such as storms, wildfires and pest and disease outbreaks. However, most climate change research in forestry has focused on understanding the climate change impacts on ecological processes (Seidl et al. 2016). In New Zealand, very little attention has been paid to climate change adaptation in managed production forests (Villamor et al. 2022). As of 2019, around 1.7 million ha of plantation forests with an annual export value of $6.3 billion are exposed to climate change risks (MPI 2021b). The NZ Superannuation Fund Climate Change report in 2020 ranked timber investments 1st of the five investments with the greatest physical climate-related risk to the fund’s real asset (NZSF 2020). Moreover, New Zealand has not developed a Climate Change Adaptation Policy, particularly for the forestry sector or a roadmap for forest adaptation to guide forest growers. Thus, adaptation is crucial to reduce the risks to forest plantations and also to protect the financial investments in forestry businesses as risk drivers from chronic and acute hazards such as fire, pests and wind damage are projected to increase (Watt et al. 2019).

Adaptation to climate change is defined as “an adjustment in natural or human systems in response to actual or expected climatic stimuli or their effects, which moderates harm or exploits beneficial opportunities” (Smit and Pilifosova 2001, p.881). Schoene and Bernier (2012) identified adaptation to climate change as the parent activity of mitigation. Adaptation is a social process in human systems and highly contextual (Burton et al. 2006; Inderberg et al. 2015; Pielke 1998) and is framed by uncertainties and constraints. For this reason, there is a growing interest in adaptation behaviour; how and why individuals adapt to climate change (Blennow 2012; Käyhkö 2019; Mostegl et al. 2019; Ontl et al. 2018; van Valkengoed and Steg 2019). A growing body of literature argues that subjective factors such as personal belief in climate change and perceived risk responses are better at explaining climate change adaptation action than individual income, education or gender (Feng et al. 2017; Grothmann and Patt 2005; Mead et al. 2012; Vulturius et al. 2018).

To the best of our knowledge, there is a lack of understanding of what influences owners and managers of plantation forests and woodlots in New Zealand, (hereafter referred to as forest growers) to adapt to climate change. Understanding their adaptation motivation and behaviour would not only help to reduce the impacts of climate change but also strengthen support for adaptation policy and risk communication for the forestry sector (Blennow et al. 2014). This paper applies the protection motivation theory (PMT) to identify the socio-psychological factors influencing forest growers’ adaptation behaviour to climate change.

1.1 New Zealand’s forestry context

Forests play an important role in the global carbon cycle. Maintaining and increasing the carbon stock in forests are recognised as important tools for climate change mitigation (Griscom et al. 2017). Forests contribute in three ways (IPCC 2014): they act as sinks, sources and reservoirs of carbon; they provide wood products that can extend the time that carbon is stored for; and they can substitute for fossil fuels—directly as a solid or liquid biofuel, or indirectly by substituting for more fossil fuel-intensive products. New Zealand has a total of 10.1 million ha of forest (38% of the total land area), of which 8 million are native forests and 2.1 million ha are plantation forests (including unplanted areas such as roads, skid sites and wetlands) (MPI 2022). Most of the native forests are owned and managed by the Crown (through the Department of Conservation). These forests are part of the network of national parks, scenic reserves and other conservation areas managed for biodiversity conservation, heritage and recreation values.

On the other hand, 96% of the plantation forests area are privately owned and used for commercial timber production. According to the 2021 National Exotic Forest Description report, around 90% of plantation forests are Pinus radiata with a rotation age of 28–30 years (MPI 2021a). Figure 1 shows the proportion of the total production forest area according to the scale of owners. Around 70% of the total exotic plantation forests are owned by forest owners or companies with more than 1000 ha. In contrast, the vast majority (estimated to be 13,000 to 14,000 forest owners) own less than 1000 ha each, making up almost 30% of total production forests (NZFOA 2018; West and Satchell 2017). In terms of Māori-owned land, around 20,500 ha are plantation forest as of 2022 (MPI 2022).

Fig. 1
figure 1

Proportion of different forest owners and companies owning plantation forests by size (Source: MPI 2021a; West and Satchell 2017)

Full size image

The carbon uptake by these plantation forests has offset about one-third of gross greenhouse gas emissions since 1990, while at the same time, plantation forests make a significant contribution to the New Zealand economy. Goods make up about 70% of New Zealand goods and services exports, and forest products from planted forests are the third largest export merchandise sector after dairy and meat products. More than 99% of forestry production in New Zealand is from plantations and over 70% of this is exported. While the main objective of plantation forest management is timber production, a range of other ecosystem services are provided including recreation and biodiversity (Yao et al. 2013) and almost 70% of plantation forests are certified under the Forest Stewardship Council Certification scheme (Durand 2017).

With regard to climate change projections, the mid-range estimate for projected New Zealand temperature increases provides an expected increase of about 0.8 degree Celsius by 2040, 1.4 degree Celsius by 2090, and 1.6 degree Celsius by 2110 (relative to 1986–2005); and by 2040 (2031–50, relative to 1986–2005), temperatures are projected to increase by between 0.7 °C (RCP2.6) and 1.0 °C (RCP8.5) nationally (MFE 2018). Given these projections, and climate change-related risks (Watt et al. 2019), it is clear that there is an opportunity to future proof economically important plantation forests through adaptation approaches. Managers and owners of native forests on public and private lands also face climate change challenges, but these owners are not addressed in this study.

1.2 Adaptation behaviour and protection motivation theory

In the context of this paper, we refer to adaptation as a process, action or outcome in a system (households community, group, sector) to better cope with, manage or adjust to some changing condition, stress, hazard, risk or opportunity (Smit and Wandel 2006). In order to predict adaptation actions and provide input to adaptation policies, there is a need for improved knowledge about processes involved in adaptation decisions, such as choices of adaptation type and timing, and conditions that stimulate or dampen adaptation (Smit and Pilifosova 2001). These include the characteristics (e.g. sensitivity, vulnerability and adaptive capacity) that influence the ability or propensity of the target sector (in this case the forestry sector) to adapt (Adger et al. 2004; Smit and Pilifosova 2001; Smit and Wandel 2006).

However, there is growing evidence that the forestry sector is considering adaptation actions, with the response determined by socio-cultural, cognitive, experiential and other subjective factors (Blennow 2012; Fischer et al. 2022; Seidl et al. 2016; Vulturius et al. 2018). These subjective factors explain how individuals perceive and respond to climate change. One theoretical framework that is widely applied for understanding the individual process of adaptation and adaptative behaviours is the protection motivation theory.

1.2.1 Determinants of adaptation behaviour

Protection motivation theory (PMT) has been extensively used for some time now to explain adaptation behaviours and is used to explain the effects of fear on health-related behaviours (Rogers 1975, 1983). It has recently been extended to natural environmental hazards such as droughts (Truelove et al. 2015), flood risks (Babcicky and Seebauer 2019; Le Dang et al. 2014) and climate change adaptation and mitigation among farmers and smallholder land managers (Bostrom et al. 2019; Dang et al. 2012; Fischer 2019; Ghanian et al. 2020; Grothmann and Patt 2005; Regasa and Akirso 2019). A PMT socio-cognitive model is more suitable for predicting proactive adaptation to climate change risk such as flooding than a model using socioeconomic variables (Grothmann and Patt 2005) as PMT assumes that an individual’s motivation to protect themselves from any risks is the main reason to direct behaviour against existing threats (Fig. 2).

Fig. 2
figure 2

Protection Motivation Theory in climate change adaptation behaviour of forest growers

Full size image

PMT as a theoretical framework explicitly addresses both risk and adaptation. This is because PMT postulates two appraisals or cognitive processes: (1) risk or threat appraisal and (2) coping or adaptive appraisal. The risk appraisal focuses on the source of the threat and factors that increase or decrease the likelihood of making non-protective or maladaptive responses. According to Floyd et al. (2000), maladaptive responses are responses that do not protect oneself or properties. This appraisal captures the individual perception of how severe the consequences of the threat are (perceived severity) and how vulnerable to suffering from those consequences individuals perceive themselves to be (perceived vulnerability). PMT assumes that the probability of engaging in protective responses is a positive function of the amount of perceived severity and perceived vulnerability, whereas maladaptive response rewards refer to the intrinsic and extrinsic benefits of neglecting a given protective behaviour and has a negative relationship with motivation to act protectively (Prentice-Dunn and Rogers 1986). Empirical consensus shows that non-protective behaviours such as doing nothing or avoiding the risks are significantly related to threat appraisal components (see Section 1.2.2) (Babcicky and Seebauer 2019; Rippetoe and Rogers 1987). According to Kothe et al. (2019), there is a limited number of studies that examined the relationship between maladaptive response reward and the motivation to protective response, even in the context of personal health where maladaptive responses are often omitted from the tests (Floyd et al. 2000; Milne et al. 2000).

Coping or adaptive appraisal refers to individuals’ cognitive processes when evaluating their ability to avoid a particular risk. Within the adaptive appraisal, three separate components are used to evaluate response measures: self-efficacy, response efficacy and response cost. Self-efficacy refers to one’s perception of how competent they are in organising and executing actions needed to manage a risky situation—that is, whether a person feels able to implement a specific measure. According to Blennow et al. (2012), self-efficacy or adaptive capacity is an individual phenomenon that either promotes or hinders adaptive response. Perceived self-efficacy directly affects a person’s motivation to change behaviour (Zimmerman 2000). Response efficacy refers to one’s belief that the actions or measures will effectively reduce the threat or certain risk. According to PMT, the greater the self-efficacy and response efficacy, the greater the motivation to engage in a protective response or behaviour. The other component of coping appraisal is the response costs, which refers to the perceived costs associated with protection actions such as financial, time, effort and emotional costs. The theory assumes that larger response costs are associated with lower motivation to engage in adaptive response.

These two cognitive processes influence an individual’s protection motivation and may lead to omitting non-protective (maladaptive) behaviour or adopting protective behaviour. According to PMT, the risk-appraisal process is addressed first because a threat must be perceived or identified before there can be an evaluation of the coping options (Floyd et al. 2000). Overall, adopting the assumptions of this theory suggests that forest growers are motivated to engage in adaptive responses that protect their assets and investments from climate change risks when their perceived vulnerability, self-efficacy and response efficacy are high, while their perceived response costs and maladaptive response rewards are low.

Several studies have used these subjective perceptions to understand the adaptation behaviour of individuals as predictors for behaviour. However, the question is, do some of the variables relate more strongly to adaptation behaviour than others? For example, Mead et al. (2012) and Bostrom et al. (2019) suggested that response efficacy is more associated with behaviours relating to climate change. Burnham and Ma (2017) found that self-efficacy beliefs are a strong and positive predictor of adaptation intent. Among forest owners in Sweden, Eriksson (2017) found that response costs were associated with the intention to engage in site-specific forestry practices in response to climate change. In contrast, Vulturius et al. (2018) suggested that personal risk appraisal and belief about the connection between personal experience and climate change can explain the adaptation of private non-industrial forest owners in Sweden.

In this study, our main objective is to investigate which components of the appraisal, along with other socioeconomic factors, are predictors of the protective and non-protective measures identified by forest growers in a survey. Among the socioeconomic and demographic factors considered are age, gender, education (Le Dang et al. 2014; Vulturius et al. 2018) and experience (Vulturius et al. 2018). We hypothesise that forest growers are motivated to implement adaptive responses when perceived risks, self-efficacy and response efficacy are high while response costs and maladaptive rewards are low.

1.2.2 Climate change adaptation measures/responses in forestry sector

In PMT, actions that reduce the threat or risk can be labelled protective measures, adaptation actions or risk response measures (Babcicky and Seebauer 2019; Dang et al. 2012; Ghanian et al. 2020; Grothmann and Patt 2005; Le Dang et al. 2014; Truelove et al. 2015). On the other hand, there are non-protective responses associated with reduced protection motivation (Babcicky and Seebauer 2019). In the original study of PMT in human health, the two cognitive processes may either lead to threat-intensifying behaviour or adopting protective behaviour (both called “adapting coping”), whereas retaining non-protective behaviour or avoiding the risk cognitively is called “maladaptive coping” (Rogers 1983). Rippetoe and Rogers (1987) provided specific examples of maladaptive coping such as wishful thinking, denial, hopelessness and other avoidance behaviour, which should not be confused with maladaptation in climate change research based on the characterisation of Barnett and O’neill (2010). Similar examples of maladaptive coping include non-protective behaviours applied in other studies such as “avoid or suppress the negative emotions associated with it” (Babcicky and Seebauer 2019; Bubeck et al. 2012) and “do nothing” (Koerth et al. 2013).

For this study, the protective and non-protective adaptation strategies identified are further classified into the risk response types (Table 1) for forest management under climate change by Birot and Gollier (2001) and Hanewinkel et al. (2011). This is also consistent with the distinct avenues to address risks by Peterson (2010), namely risk acceptance, risk avoidance, risk reduction, risk spreading and risk transfers.

Table 1 Risk response or measures in forest management
Full size table

We further characterise these measures using the adaptation behaviour typologies in terms of timing and scope (Fischer 2019; Smit and Pilifosova 2001). For timing, it can either be reactive or proactive. According to Fischer (2019), reactive response or action refers to an immediate behavioural response to regain stability, such as spreading risk and securing resources. On the other hand, proactive response or action entails reorienting practices in anticipation of new conditions to reduce future damage, vulnerability or risk through planning, monitoring, increasing awareness, building partnership and learning. In Table 1, these examples are similar to the risk reduction examples.

Responses can be incremental or transformational. Incremental responses are actions that make small changes within current contexts to avoid disruptions while continuing the same objectives. In contrast, transformational actions fundamentally alter the entire system’s ecological and/or social properties and functions (Fedele et al. 2019). For example, a forest grower may switch from radiata pine monoculture to mixed forest species, classified as risk spreading and risk avoidance (Table 1).

2 Methods

2.1 Data collection and analysis

We used a survey of forest managers to investigate the three adaptive appraisal components: perceived self-efficacy, response efficacy and response costs. We used both online and paper-based questionnaires with 26 questions and face-to-face interviews. The questions were formulated based on the PMT framework (Annex 1) and then pre-tested for refinements and adjustments with the help of representatives from NZ Farm Forestry Association, a private forest owner, and a forest consultant. The final questionnaire was distributed online through the SurveyMonkey (Momentive Inc.) from 3 March to 3 May 2021. We targeted the forest managers and/or owners of the 24 large-scale forest companies reported in the 2020 Forestry facts and figures (NZFOA 2021), which manage a total of 1.2 million hectares of productive plantation forests on both private and public lands. The paper-based survey was distributed to participants of the New Zealand Farm Forestry Conference on 11 March 2021, where 15 questionnaires were completed. This conference was targeted so as to reach respondents from small- to medium-scale forest owners (who might have limited access to internet to participate in the online survey). During the conference, we interviewed 5 of the 15 willing respondents (who completed the paper-based survey) due to lack of time and in order of availability (usually during coffee breaks and after sessions). These interviews focused on exploring interviewee’s perceptions of the climate change and adaptation strategies they are currently practicing for the purpose of triangulation. In total, 60 respondents participated in the survey, of which 45 were online and 15 completed the survey at the conference. The response rate is 28% resulting from an existing list of 218 forest growers and practitioners who had signed-up for information updates on forestry related research activities.

Of the total of 26 questions, we used nine to address the objective of this study (Table 2). The identified adaptative measures (from Q12) were grouped into the five risk response measures (see Table 1). The data were analysed with the statistical software STATA 17 (StataCorp 2021). First, a descriptive analysis of single variables was carried out, characterising their distributions (Table 3). Then, a binary logistic regression (or logit model) was used to examine the relationships of the risk response measures among explanatory variables. The logistic regression was applied to take account of the categorical nature of the dependent variables with results presented in terms of an odds ratio as a measure of association between a variable of interest and the outcome (Bubeck et al. 2013; Villamor et al. 2014). If the value of the odds is below 1.0, then, it is considered low, whereas, if the value of the odds is greater than 1.0, then, it is considered high. A correlation matrix of all explanatory variables was made to check for problems associated with multicollinearity (Annex 2).

Table 2 Survey items used to derive explanatory variables
Full size table
Table 3 Descriptive statistics of respondents (n = 60)
Full size table

3 Results and discussion

3.1 Characteristics of respondents

The descriptive statistics of the respondents are shown in Table 3. Both skewness and kurtosis values are less than ± 1.0 suggesting that the values are normally distributed. Of the total of 60 respondents, 13% are females and 87% are males. The respondents represent 90% of all the country’s regions. Around 33% of the respondents are from the age bracket of 55–66, 23% are from 45–55 and 22% percent are from 68 + of age. Around 30% of the total respondents are forest managers managing on average approximately 86,000 ha of plantation forests; 30% of the respondents are sole forest owners, which own an average of 186 ha of managed forests; 28% of the respondents are forest consultants working for small- and large-scale forest companies; 12% of the respondents are both forest owners and managers and consultants; and the remainder are farmers (6%) who planted trees on their farms and or recently harvested their trees. In terms of the scale of forestry operations, 18 of the respondents are forest managers of large-scale companies (ranging from 6000 ha to more than 200,000 ha) (see Annex 3 for detailed distribution). In total, respondents are managing more than 70% (1.2 million hectares) of the New Zealand plantation forest area.

3.2 Climate change risk or threats to forest growers

The ranking of climate change risks or threats as perceived by the respondents in the next 5 years are shown in Fig. 3. Market disruption due to climate change was ranked as the most important risk by most of the forestry businesses surveyed. Windthrow and wind damage were ranked second, and forest fires ranked third. According to the respondents, market disruption is perceived as being worrisome due to their reliance on only a few exotic species in their plantations and the limited number of current export markets.

Fig. 3
figure 3

Major climate change related risks as perceived by the respondents (n = 60)

Full size image

3.3 Response measures to climate change risks

This section details the response measures identified by survey respondents to manage risks. Fifteen measures were identified by respondents and grouped into major categories: acceptance, avoidance, reduction, spreading and transfer. Among these groups, the top three response measures are categorised into reduction (39%), spreading (27%) and acceptance (17%) (Fig. 4). Table 4 details the specific measures according to each group with frequencies. Among the frequently reported measures are diversification of species, do nothing and measures related to silvicultural activities. Some of the respondents reported a combination of different measures such as risk spreading (e.g. diversification of tree species and market) and risk reduction (e.g. silvicultural activities).

Fig. 4
figure 4

Response measures according to risk categories (in percent)

Full size image
Table 4 Protective measures to reduce/manage/avoid climate risks (n = 55)
Full size table

In terms of specific perceived risks, the most recorded measure for addressing market disruption is diversifying tree species, which belongs to the risk spreading category. For windthrow and wind damage, the most reported measure is silvicultural activities, which belong to the risk reduction group. For forest fires, response fire plan, fire breaks and maintaining fire equipment are the most reported measures belonging to risk reduction group. For pests and diseases outbreak, biosecurity surveillance and monitoring (belonging to the risk reduction group) and planting fire-resistant species and shelterbelts (belonging to the risk spreading group) are preferred by respondents. A risk transfer measure was only recorded for forest fire risk, with only 3% of respondents reporting fire insurance. Although insurance is available to forest growers to protect their assets particularly against fires and catastrophic wind, forest owners in New Zealand have traditionally rarely insured their standing timber (Manley and Watt 2009). This is most likely because of a belief that premiums are too expensive and the risk of fire is not large enough to warrant the purchase of fire insurance (Manley and Watt 2009). A survey of NZ forestry companies found that only 36% of 1.1 million ha of plantation in 2008–2009 was insured for fire (Manley and Watt 2009).

“Do nothing” and “no plan of action” are considered non-protective responses and were reported across market disruption, windthrow, heavy rains and forest fire risks. Although, risk acceptance measures are not effective in reducing the physical risk of climate change, this approach does help to avoid or suppress the negative emotions associated with it temporarily until a better option comes along (Bubeck et al. 2012). Some respondents view the risk as beyond their control, whereas ceasing operation and retiring areas (as examples of risk avoidance) were measures reported for windthrow and landslides due to heavy rains. Others who reported market risk as the most concerning threat felt that it is beyond their expertise to deal with this risk. Although we observed that most of the forest managers belonging to large-scale forest companies identified risk reduction and spreading measures, we could not be certain for other forest consultants and small growers. Nevertheless, the New Zealand government has passed recently about the Financial Sector (Climate-related Disclosures and Other Matter) Amendment Act 2021, mandating large businesses to start making climate-related disclosures in 2023 (MFE 2022). For large businesses (which may include large forestry companies, in the future if not now), the protection of assets and investments must be explicitly addressed in the mandatory disclosure of climate-related financial risks.

3.4 Adaptive appraisal

This section qualitatively assesses the three components of adaptive appraisal according to each group of response measures. Also, the explanatory variables’ influence is quantitatively assessed accordingly to four groups of response measures.

3.4.1 Perceived response efficacy

Response efficacy refers to the respondent’s belief that a proposed adaptive action or protective measure will effectively reduce the expected damage from climate change risk. Figure 5 illustrates each measure type in relation to the respondent’s perceived response efficacy. Avoidance, reduction and spreading were the risk measures that respondents were extremely confident about their effectiveness. Of the respondents who selected risk acceptance measures, 83% were not confident that this measure will effectively manage the risk.

Fig. 5
figure 5

Perceived response efficacy (effectiveness of the adaptation actions) (in percentage)

Full size image

3.4.2 Perceived self-efficacy

Self-efficacy refers to whether the respondent feels capable of carrying out the selected protective action. Figure 6 illustrates each of the measure types in relation to the respondent’s perceived self-efficacy. All the respondents who selected risk acceptance measures felt that they are not confident in protecting themselves and their properties. The measure types which respondents selected with extreme confidence are avoidance, spreading and reduction. Of the respondents who selected risk avoidance, 25% felt extremely confident, whereas the majority of the respondents (between 54 and 67%) were somewhat confident to carry out risk spreading, reduction and avoidance.

Fig. 6
figure 6

Perceived self-efficacy of the adaptation actions (in percentage)

Full size image

3.4.3 Response cost

Perceived response cost of each protective measure or adaptation action is presented in Fig. 7. Of the respondents who selected risk acceptance, contrary to our expectations, only 13% considered this measure not at all expensive. This suggests that they were perhaps responding based on their expected cost of the outcome of their (in)action, rather than the direct cost of the measure itself, which should be zero. Of the respondents who selected risk avoidance, reduction and spreading, between 20 and 25% considered these measures moderately expensive. None of the respondents perceived their measure as extremely expensive (or perhaps they did not perceive the measure as more expensive than expected benefits).

Fig. 7
figure 7

Perceived response cost of the adaptation actions (in percentage)

Full size image

3.5 Predictors of adaptive measures and non-protective response

Tables 5, 6, 7, and 8 present the result of the logistic regression for each measure. As hypothesised, respondents are more likely to implement adaptive responses to protect themselves from climate change risks when their perceived vulnerability, self-efficacy and response efficacy are high (i.e. value of the odds greater than 1.0), while their perceived response costs and maladaptive responses are low (i.e. value of the odds lower than 1.0).

Table 5 Logistic regression for risk acceptance
Full size table
Table 6 Logistic regression for risk avoidance
Full size table
Table 7 Logistic regression for risk reduction
Full size table
Table 8 Logistic regression for risk spreading
Full size table

For respondents who are likely to implement risk acceptance (Table 5), they had a low level of perceived vulnerability (p < 0.10) but a greater level of evading behaviour (as a maladaptive response) (p < 0.05), and had a low level of perceived self-efficacy (p < 0.05). Based on PMT, these predictors suggest that risk acceptance could be categorised as non-protected response or behaviour. If a 10-percent criterion of statistical significance is used, age is positively and significantly associated with risk acceptance, suggesting that the older the forest grower, the greater the likelihood of risk acceptance.

Respondents who are likely to implement risk avoidance had a greater level of evading behaviour (p < 0.05) but a low level of perceived resource efficacy (p < 0.05) (Table 6). If a 10-percent criterion of statistical significance is used, postponing behaviour is negatively associated with risk avoidance. Following the PMT assumption, these predictors suggest that risk avoidance could be categorised as a non-protective response or behaviour despite being proactive (Table 1). This observation is in line with the description of Fennelly and Perry (2017) that risk avoidance is sometimes considered an unsatisfactory approach or impractical for dealing with many risks. One reason for this is that if a forest grower uses avoidance extensively, the ability to accomplish objectives (e.g. timber production) is reduced and opportunities to deal with risks are missed.

For respondents who are likely to implement risk reduction (Table 7), they have both large levels of perceived response cost (p < 0.10) and perceived resource efficacy (p < 0.05). Risk reduction is the only response measure for which perceived response cost is positive and significant at 10-percent criterion, suggesting that the more expensive the response measure is, the less likely it will be implemented. Although perceived response cost is a relevant predictor for risk reduction, perceived resource efficacy becomes more significantly associated with this response measure when additional explanatory variable, i.e. forestry experience of respondent was added (from p = 0.06 to p < 0.05). This observation suggests that risk reduction could be categorised as a protective response or behaviour.

Respondents who are likely to implement risk spreading had both greater levels of perceived severity (p < 0.10) and perceived self-efficacy (p < 0.10), and a low level of evading behaviour (p < 0.05) (Table 8). Based on PMT, these predictors suggest that risk spreading could be categorised as a protective response or behaviour.

Overall, respondents who have greater ( +) perceived self-efficacy (or perceived adaptive capacity), resource efficacy and severity and low (-) maladaptive response (e.g. evading behaviour) are more likely to implement protective responses such as risk reduction and spreading (Fig. 8a). On the other hand, respondents who have low (-) perceived vulnerability, self-efficacy and resource efficacy and greater ( +) perceived maladaptive behaviour are likely to implement non-protective responses such as risk acceptance and risk avoidance (Fig. 8b). Our findings provide empirical evidence that the components of PMT are associated with the likelihood of implementing measures to reduce risks while explaining the forest growers’ behaviour or motivation behind selecting specific risk management response. Moreover, our study provided insights into maladaptive responses such as evading and postponing behaviour, and the association of this type of response with protective and non-protective responses. This type of response is considered a research gap in the application PMT (Kothe et al. 2019).

Fig. 8
figure 8

Adaptation of forest growers is predicted by PMT components for a risk reduction and avoidance; and b for risk acceptance and avoidance

Full size image

These observations corroborate the findings of Eriksson (2017), Sousa-Silva et al. (2016) and Vulturius et al. (2018) that response cost, personal risk assessment and personal experience are determinants of forest growers adaptation behaviour. A similar application of PMT in Ethiopia found that perceived severity, self-efficacy and response efficacy are predictors of an individual’s motivation to practice climate change adaptation and mitigation (Regasa and Akirso 2019). In relation to climate change mitigation study, our finding on self-efficacy is consistent with the study of Bostrom et al. (2019) that personal self-efficacy supports a response favouring climate change mitigation. However, Bostrom et al. (2019) pointed out that the perception and judgement of self-efficacy and response efficacy may differ between personal and government/collective mitigation actions. Accordingly, government and collective actions are both seen as much more challenging to perceive than personal actions. In the New Zealand’s forest plantation context, where the majority of plantations are privately owned, the perceptions of self-efficacy, response efficacy and response costs may differ with the scale of operations and management, e.g. small growers versus corporate growers (see Section 3.5 Limitations).

3.6 Limitations

The primary limitation of the research described here is the cross-sectional nature of the data. The small number of sample cases limits the assessment of explanatory variables as well as the ability to make comparison between types of forest growers (e.g. large-scale forest managers vs. small-scale forest owners). These results primarily reflect the perspective of forest growers managing medium- and large-scale managed forests. Small-scale forest owners (NZFOA 2018) with less than 40 hectares of forest as well as Māori forest owners are not well represented in this study. Future research should explore this issue of forest owner representation further.

One reason for the lack of consistent findings of risks and coping appraisals is due to the lack of consistent methodological designs. Some studies use structural modelling (Babcicky and Seebauer 2019; Le Dang et al. 2014), others have used principal component analysis (Bostrom et al. 2019), whereas others have used qualitative analysis (i.e. focus group interviews) (Fischer et al. 2022). Due to the small number of samples, we applied regression analysis, which allows the assessment of a single dependent variable (i.e. response measure) at a time to be explained and in addition we have used the commonly applied method of PMT (Bubeck et al. 2013; Dang et al. 2012; Regasa and Akirso 2019). However, the PMT structure consists of two endogenous variables, protection motivation and non-protective responses, which would be better explained using structural equation modelling to analyse the interrelationships among the PMT components. Other methodological weaknesses of existing PMT applications are described by Babcicky and Seebauer (2019).

In our results, we have combined responses from forest managers working for forest companies with forest or landowners managing their own trees, despite small sample size. In practice, salaried professionals faced with certain short-term costs to mitigate uncertain longer-term risks may behave differently, especially as benefits may not be seen during their working lives. Perception of response costs is also likely to be different among forest owners. For example, private forest owners must bear all the costs and are not compensated until after the trees are harvested. In contrast, forest company managers may have clearly defined responsibilities within larger teams that limit their ability to make decisions and implement them. They may also have greater exposure to and knowledge of potential response measures than private landowners. A larger survey would be required to determine if differences in perceptions of self-efficacy, response efficacy and response costs are significant.

4 Conclusions and recommendation

This paper establishes a baseline of information on the role of socio-psychological factors in assessing adaptation behaviour which describes the behavioural response to climate change at the level of individual forest managers in New Zealand. Such an assessment remains limited in New Zealand’s forestry sector. Thus, this baseline will allow investigation of how changes in responses occur as climate change adaptation policy evolves. Our results demonstrate that components of PMT such as risk perception, efficacy perceptions and maladaptive behaviour are significantly associated with forest growers’ responses to manage climate-related risks. The result suggests a low self-efficacy and response efficacy belief among forest growers that makes them less motivated to enact protective or adaptation measures, leading them to do nothing (i.e. risk acceptance) or stop operation (i.e. risk avoidance), which can be characterised as more maladaptive coping.

Understanding their responses has important implications for developing the forestry adaptation roadmap for New Zealand as well as measuring its success. These lessons are also of relevance to the understanding of how forest managers in different parts of the world will behave in response to climate change, given that the nature and scale of impacts could be vastly different to those experienced by survey respondents in New Zealand. A global understanding of this nature may well be important in terms of influencing global trends in forest investment and future global initiatives around the role of forests in responding to global climate change.

Our findings indicate a need for strategies to boost the self-efficacy or adaptive capacity of forest growers to implement adaptation action. Marshall et al. (2010) suggested strategies to enhance adaptive capacity at various scales which include adaptive governance approaches that encourage climate learning and adaptive management. Regarding forest growers, their adaptive capacity can be enhanced by developing their capacity for individual evaluation of climate change adaptation through access to climate adaptation information, expertise and technology.