Latent Profiles of Eco-Anxiety: Resilience, and Vulnerability Factors in a Portuguese-Sample

Abstract

Eco-anxiety refers to emotional and cognitive responses to environmental degradation and can manifest in both adaptive and maladaptive forms. This study aimed to identify distinct eco-anxiety profiles and examine their associations with resilience and vulnerability factors in a sample of 917 Portuguese-speaking adults. Latent profile analysis revealed five profiles: adaptive eco-anxiety, highly impaired maladaptive eco-anxiety, psychological distress independent of eco-anxiety, non-anxious/disengaged, and moderate I I have separated the addresses into different affiliations.have separated the addresses into different affiliations.eco-anxiety. These profiles differed significantly in psychological symptomatology, nature connectedness, pro-environmental attitudes, and prior exposure to cumulative social and environmental stressors. Higher-distress profiles were more likely among younger individuals, women, urban residents, unemployed participants, those without children, individuals with a prior psychiatric history, and those reporting direct exposure to drought. In contrast, stronger environmental identity and greater engagement with natural environments were associated with adaptive eco-anxiety, suggesting protective and resilience-promoting mechanisms. Overall, the findings highlight the multidimensional and heterogeneous nature of eco-anxiety and its complex relationship with psychological well-being and environmental engagement. Tailored interventions that promote adaptive coping, strengthen psychological resources, and facilitate access to natural environments may help mitigate maladaptive distress while supporting constructive environmental concern and action.

1. Introduction

Climate change poses a significant twenty-first-century threat to ecosystems and human health, driven by increasingly frequent and severe extreme weather events [1]. Awareness of these risks has generated widespread concern, with many individuals experiencing anxiety related to ecological crises, which can affect mental health both directly, through exposure to environmental hazards, and indirectly, via climate-related information [2,3,4].

Eco-anxiety is a multidimensional psychological response to climate change, environmental degradation, and ecological uncertainty, rather than a diagnosable disorder [2,5]. Although modestly associated with anxiety, depression, and distress [6,7], it is distinct from conventional psychopathology [5,8]. Eco-anxiety may serve an adaptive function within a sustainability framework, promoting reflection, environmental awareness, and constructive action, with anxiety-related processes such as fear and worry enhancing threat sensitivity and preparedness [9,10,11].

However, eco-anxiety can occasionally manifest as significant psychological distress, characterized by intense emotions such as fear, depression, helplessness, and anger [12,13]. This distress may impair functioning, necessitating clinical attention [2,13], and is often accompanied by persistent concern over ecological collapse, limited coping resources, and existential anxieties [5,11].

Eco-anxiety involves cognitive, behavioral, and physiological symptoms and is predominantly future-oriented, reflecting anticipatory responses to ongoing and projected ecological crises [6,14]. Despite increasing scholarly attention, evidence on determinants of its severity and progression remains limited, underscoring the need to identify factors that heighten vulnerability to severe eco-anxiety [5].

Although anxiety is central to eco-anxiety, the construct encompasses broader psychological responses to environmental threats [15]. Its association with psychopathology is multifaceted [10,16]. Perceived lack of control increases vulnerability by fostering maladaptive cognitions, thereby amplifying depressive affect, rumination, and pathological eco-anxiety, which may manifest as intrusive worry and maladaptive coping in susceptible individuals [17,18].

Conversely, the absence of manifest self-reported eco-anxiety does not imply emotional indifference or complete disengagement but may reflect denial or other emotion-regulation processes such as emotional suppression, distancing, or rationalization [5]. Confronting climate change can evoke distressing existential concerns, which some individuals manage through minimization, denial, or apathy, thereby reducing conscious anxiety while sustaining internal conflict, a process described as disavowal [19,20]. As a result, anxiety may remain latent or repressed rather than absent, indicating a dynamic interplay between denial and eco-anxiety [21].

In the context of climate change, the distinction between adaptive and maladaptive eco-anxiety remains conceptually and empirically underdeveloped [5,22]. Adaptive eco-anxiety may be operationally defined as a climate-related emotional response characterized by moderate levels of concern, preserved perceived self-efficacy, problem-focused coping, and engagement in sustainable and pro-environmental behaviors (PEB). In contrast, maladaptive eco-anxiety refers to a persistent and dysregulated form of distress, characterized by heightened emotional arousal, diminished perceived control, cognitive rumination, and functional impairment, often resulting in avoidance or eco-paralysis [23,24].

Importantly, this distinction should not rely solely on distress intensity but rather on integrated psychological, cognitive, and behavioral criteria [7]. Eco-anxiety may be considered adaptive when it promotes goal-directed engagement, and maladaptive when it overwhelms regulatory capacities and undermines effective functioning [2,5]. However, the psychological, contextual, and behavioral factors underlying individual differences in these responses remain insufficiently understood and warrant further investigation. Latent Profile Analysis (LPA) offers a promising person-centered approach for identifying distinct profiles of climate change anxiety.

This study addresses these gaps by extending prior LPA of eco-anxiety in several ways. First, it adopts an integrative framework that simultaneously considers vulnerability factors (e.g., exposure to climate change–related extreme events and socioeconomic disadvantage) and resilience-related processes (e.g., engagement with natural environments), thereby providing a more comprehensive understanding of the mechanisms underlying distinct eco-anxiety profiles. Second, whereas previous LPA studies have often focused on relatively homogeneous or geographically restricted samples, the present study applies this person-centered approach to a broader and underexamined population—Portuguese adults—enhancing the generalizability and contextual relevance of the findings. Finally, by examining how these profiles relate to both adaptive and maladaptive outcomes, this study contributes to clarifying the functional significance of eco-anxiety, helping to distinguish when it operates as a constructive, motivating response versus a source of psychological burden.

1.1. Identification of Individual Differences in Eco-Anxiety Using Latent Profile Analysis

LPA is a person-centered method used to identify unobserved subgroups based on patterns across multiple variables [25]. Although few studies have applied LPA, this approach is well-suited to examining social and personal heterogeneity in eco-anxiety, capturing individual differences in cognitive, behavioral, and emotional manifestations. Wullenkord et al. [26] using LPA identified four distinct manifestations of climate anxiety, comprising three different profiles of climate-anxious groups and one non-anxious group. The profiles differed primarily in degree of functional impairment: one “highly impaired” group, one “moderately impaired” group, and a prevalent “climate-anxious functioning” group showing elevated appraisals and affect but minimal functional impairment. The “non-anxious” group reported low levels across all indicators, reflecting minimal climate-related anxiety.

Ágoston et al. [27] applied LPA to eco-anxiety, eco-guilt, ecological grief, and PEB, identifying five profiles reflecting a graded increase in both eco-emotional intensity and PEB, with most participants in the average group. Chan et al. [28] used LPA in preregistered U.S. and Chinese samples to compare state-based climate anxiety with symptom-based anxiety. A four-profile solution emerged in both studies, revealing distinct patterns of affective and clinical symptom responses. Some profiles showed elevated state anxiety with low symptom levels, whereas others exhibited similar affective concern but higher symptomatology, indicating that emotional climate concern and clinical symptoms represent related yet distinct dimensions.

Collectively, these findings highlight the heterogeneity of climate-related anxiety and eco-emotions, with profiles differing in affective intensity, symptom severity, functional impairment, and pro-environmental engagement. They support a multidimensional conceptualization of eco-anxiety, in which emotional concern, clinical symptoms, and behavioral outcomes do not consistently align [3,5]. However, the use of LPA in eco-anxiety research remains limited, indicating a significant gap in person-centered approaches.

1.2. Resilience and Vulnerability in Different Manifestations of Eco-Anxiety

Grounded in appraisal-based emotion theories, Wullenkord et al. [26] conceptualize eco-anxiety as arising from cognitive appraisals of environmental stressors, with coping processes shaping whether anxiety results in maladaptive impairment or adaptive regulation [29,30]. When adaptively regulated, eco-anxiety may support individual well-being and foster engagement in pro-environmental actions, thereby contributing to broader sustainability goals [26]. The model delineates three interrelated dimensions—anxiety-related cognitive appraisals, affective experience, and functional impairment—highlighting the need for precise conceptualization and systematic measurement in order to advance theory and research [29].

Direct exposure to climate change–related environmental transformations constitutes a substantial risk factor for adverse psychological outcomes, with affected individuals demonstrating more intense and enduring emotional responses [31,32,33,34]. Such exposure is consistently linked to heightened anxiety, depressive symptoms, post-traumatic stress, and feelings of loss and helplessness, particularly among populations experiencing extreme weather events, ecosystem degradation, or climate-related displacement [35,36].

Vulnerability to eco-anxiety is unevenly distributed, with younger individuals—particularly those under 35—reporting higher levels than older cohorts, likely due to their disproportionate exposure to long-term climate risks and anticipated lifelong consequences of environmental change [14,37]. The cumulative effects of climate change and heightened risk-awareness further exacerbate eco-anxiety [2,17], which has been associated with depressive symptoms, persistent sadness, loneliness, and heightened fear, underscoring its role as a risk factor for severe psychological distress [24,38].

Evidence consistently shows that women report higher levels of eco-anxiety than men [10,39]. This pattern is usually interpreted using social role theory, which suggests that women’s gendered socialization into caregiving and nurturing roles fosters greater perceived responsibility for others’ well-being in the face of climate threats [40]. Such expectations may heighten climate-related concern, worry, and endorsement of pro-environmental worldviews among women relative to men [41]. In addition, men may be more represented in more dismissive or denial-oriented responses to climate change, which could further contribute to the observed gender gap [42,43].

Climate change disproportionately impacts socioeconomically disadvantaged populations, with poverty being associated with heightened vulnerability to actual climate risks [3,17]. Additionally, lower income and unemployment are associated with greater environmental as well as socioeconomic stressors, poorer mental health, and reduced coping resources [44,45]. On the other hand, higher educational attainment and the absence of children have been linked to elevated eco-anxiety [10,46]. Pre-existing mental health conditions further increase vulnerability, as does prior psychological distress which amplifies emotional, cognitive, and somatic responses to climate threats, increasing the risk of maladaptive coping [5,37].

In contrast, engagement with natural environments is a well-established resilience factor, which is associated with improved mental health and well-being [47]. Such engagement promotes nature connectedness—defined as an experiential sense of belonging and perceived relationship with the natural world [48]—that has been linked to lower anxiety and lower depression and greater pro-environmental engagement. These associations suggest that nature connectedness can be a potential resilience factor in coping with eco-anxiety [47,49].

However, evidence also indicates that a strong identification with nature may increase vulnerability to climate-related distress, as higher levels of nature connectedness have been associated with elevated climate worry, anxiety, and, in some cases, a greater emotional impact of perceived environmental degradation [2,13]. These findings suggest that perceived oneness with nature may paradoxically function as a resilience factor in some cases as well as a vulnerability factor in others, amplifying sensitivity to climate change impacts and hence contributing to greater eco-anxiety [10].

Evidence suggests that cumulative exposure to multiple social and environmental stressors may cumulatively erode mental health and coping capacity. Chronic, low-level stressors (e.g., daily hassles), together with environmental and socio-economic disadvantages, are consistently associated with poorer psychological outcomes and diminished coping resources [43,44]. Their co-occurrence can produce additive effects, generating greater psychological burden [50]. In the context of climate change, pre-existing social and environmental vulnerabilities may further intensify climate-related distress [51,52]. Accordingly, cumulative stress may constitute a vulnerability factor for eco-anxiety, influencing the heterogeneity of its manifestations [5].

1.3. Purpose of the Study and Hypotheses

Eco-anxiety can be considered clinically significant when it becomes difficult to regulate and disrupts an individual’s everyday functioning [2]. However, current evidence remains limited regarding whether, and to what extent, eco-anxiety influences human health, the specific health outcomes with which it may be associated, and the underlying mechanisms through which eco-anxiety interacts with health-related processes [40]. To the best of our knowledge, no previous study has employed LPA to investigate eco-anxiety within the Portuguese population, underscoring a critical gap in person-centered research that the present study seeks to address.

The objective of this study is to investigate different manifestations of eco-anxiety in a large Portuguese-sample, and to examine the roles of nature connectedness and protection of nature, as well as the roles of exposure to cumulative environmental stressors, as factors of resilience or vulnerability to psychological distress manifested in anxiety and depressive symptoms.

Based on previous studies using latent profile analysis (LPA), a four- to five-profile solution is expected, comprising (1) a non-anxious/disengaged profile, (2) an adaptive eco-anxiety profile with elevated concern but low psychological distress, (3) one (or two) moderately anxious profiles, and (4) a profile of highly psychologically distressed with maladaptive eco-anxiety. We further examined differential associations between these profiles and potential resilience and vulnerability factors, specifically nature connectedness and exposure to cumulative environmental stressors.

2. Materials and Methods

2.1. Sample and Study Design

The study included 917 native Portuguese-adults aged 18 or older (M = 35.00, SD = 16.13; range: 18–83). As can be seen in

Table 1. Sample demographic characteristics.

Variables	n	%
Sex
Female	579	63.1
Male	338	37.9
Civil status
Single	507	55.3
Married or cohabitation	321	35.0
Divorced	77	8.4
Widowed	12	1.3
Education
Less than nine years of education	17	1.9
Nine years of education	36	3.9
12 years of education	449	49.0
Higher education	415	45.2
Professional status
Student	381	41.6
Employee	438	47.8
Unemployed	62	6.8
Retired	36	3.9
Socioeconomic status
Low	76	8.3
Medium	643	70.1
Medium-high	182	19.8
High	16	1.7

. the sample was predominantly female (63%) and single (55%), most participants had completed at least 12 years of education, with many holding higher education degrees and few reporting nine or less years of schooling. As to working status, most were employed and students. Socioeconomic status was primarily reported as medium, with smaller proportions indicating medium–high or low levels and very few reporting high status (Table 1).

This study is part of a broader research project investigating psychological variables related to climate change. Its primary aim was to assess climate change–related concerns and associated emotional response variables among Portuguese-individuals. The study protocol received ethical approval from the Institutional Review Board of Universidade Europeia. Data were collected using a convenience sampling, with participants recruited through mailing lists and social media platforms (Facebook, Instagram, LinkedIn, and WhatsApp).

Eligibility criteria included being Portuguese and at least 18 years of age. Prior to participation, individuals received detailed information about the study’s aims and procedures in accordance with the Declaration of Helsinki and provided electronic informed consent. A convenience sampling method was employed for data collection. Participation was voluntary and uncompensated. Data were collected through an online survey, consisting of self-report measures of eco-anxiety, environmental identity, psychological symptoms, and sociodemographic variables. The questionnaire package was accessed through a web link and required approximately 20 min to complete.

2.2. Measures

The Hogg Eco-Anxiety Scale (HEAS) [6,53] was used to measure eco-anxiety. The HEAS is a 13-item self-report measure designed to assess eco-anxiety symptoms over the preceding two weeks. Items are rated on a 5-point Likert scale (0 = not at all to 4 = every day) and are grouped into four dimensions: emotional symptoms (“Feeling nervous, anxious or on edge”), rumination (“Unable to stop thinking about future climate change and other global environmental problems”), behavioral symptoms (“Difficulty working and/or studying”), and anxiety about personal impact (“Feeling anxious about your personal responsibility to help address environmental problems”). There are no reverse-scored items. Mean scores (range 0–4) were calculated for each dimension, with higher values indicating greater eco-anxiety. Internal consistencies were good for emotional symptoms (α = 0.85), rumination (α = 0.85), behavioral symptoms scale (α = 0.81), and anxiety about personal impact (α = 0.86) subscales.

Information was also collected on direct exposure to climate change–related events (e.g., storms, floods, droughts, wildfires) and on the presence of prior psychiatric conditions. These data were obtained through dichotomous (yes/no) self-report questions.

The Revised Environmental Identity Scale (Revised EID) [48,52]. The Revised EID is a 14-item self-report measure rated on a 7-point Likert scale (1 = not at all to 7 = true of me). Although originally conceptualized as unidimensional, cross-cultural evidence, including the Portuguese validation, supports a two-factor structure: Connectedness with Nature (“I think of myself as a part of nature, not separate from it”) and Protection of Nature (“I consider myself a steward of our natural resources”). Mean scores (range 1–7) were computed for each dimension, with higher scores indicating stronger environmental identity. Internal consistency was good (Connectedness with Nature α = 0.94; Protection of Nature α = 0.81).

The Brief Symptom Inventory-18 (BSI-18) [54,55] was used to assess psychological distress in the dimensions of anxiety and depression. The BSI-18 is a self-report measure assessing symptoms over the past week on a 5-point Likert scale (0 = not at all to 4 = extremely). Although the instrument includes three dimensions (Somatization, Depression, and Anxiety), only Anxiety (“Feeling tense or keyed up”) and Depression (“Feeling blue”) were analyzed in this study. Mean scores (range 0–4) were calculated for each dimension, with higher values indicating greater psychological distress. Internal consistency was good for both Anxiety (α = 0.88) and Depression (α = 0.89).

Sociodemographic data included sex, age, marital status, education, employment, parental status, and socioeconomic status.

2.3. Statistical Methods

LPA was conducted in R (version 4.0) using the tidyLPA package, with mean scores for four eco-anxiety dimensions and symptoms of anxiety and depression specified as continuous indicators. Models were estimated within a Gaussian finite mixture framework, assuming conditional independence of indicators within profiles and equal variances across classes with zero covariances (i.e., diagonal variance–covariance matrices). Robust maximum likelihood estimation was employed, with multiple sets of random starting values (e.g., 1000 initial starts and 250 final optimizations) to reduce the risk of local maxima. Models with two to nine profiles were estimated to identify distinct eco-anxiety patterns.

Model selection relied on the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), sample-size adjusted BIC (SABIC), bootstrapped Lo–Mendell–Rubin adjusted likelihood ratio test (BLRT), entropy, and χ² fit, with lower AIC/BIC values indicating better fit and a non-significant χ² suggesting acceptable fit [56]. The BLRT compared nested models, where non-significant results supported retention of the more parsimonious solution [57]. Classification precision was evaluated using entropy (range 0–1), with higher values indicating clearer class separation; values of 0.40, 0.60, and 0.80 reflected poor, moderate, and high classification quality, respectively. Solutions in which the smallest profile comprised <5% of the sample were interpreted cautiously alongside statistical fit, classification accuracy, and theoretical interpretability [58].

Subsequent analyses were conducted in IBM SPSS Statistics (Version 29). A multivariate analysis of variance (MANOVA) examined differences between latent profiles in Connectedness with Nature and Protection of Nature. Bonferroni-adjusted post hoc tests were used for pairwise comparisons to control the familywise error rate. These procedures aimed to identify resilience and vulnerability factors associated with Connectedness with Nature and Protection of Nature across the LPA-derived profiles.

Multinomial logistic regression analyses examined associations between latent class membership (based on posterior probabilities assigning individuals to their most likely class) and sociodemographic, exposure, and clinical variables, including sex, age, marital status, educational attainment, professional and socioeconomic status, direct exposure to storms, floods, droughts, and wildfires, and prior psychiatric conditions (all coded dichotomously). The sample size met the recommended minimum of 10 cases per predictor. Model fit was assessed using Nagelkerke R² [59], a rescaled version of Cox–Snell R² [60], and odds ratios were interpreted as the relative change in likelihood of class membership compared with the reference group.

3. Results

3.1. Latent Profile Analysis

Following Lanza et al. [56], latent profile selection considered fit indices (AIC, BIC, SABIC, BLRT, entropy), smallest profile size, and theoretical interpretability [60]. A five-profile solution provided the best balance between model fit and parsimony, as indicated by a low SABIC, relatively low AIC and BIC values, a significant BLRT, and high entropy. Although the six-profile model showed slightly lower BIC and SABIC values, the five-profile solution demonstrated higher entropy and adequate profile sizes, supporting its theoretical interpretability and its selection as the final model (see

Table 2. Fit indices for the latent profile analysis.

Model	AIC	BIC	SSABIC	BLRT	p	Entropy	Smallest Group Proportion
2 classes	8406.44	8489.73	8429.41	942.37	0.001	0.88	25.0
3 classes	8173.29	8287.26	8204.72	247.15	0.001	0.85	12.8
4 classes	7898.05	8042.71	7937.95	289.24	0.001	0.88	10.5
5 classes	7783.92	7959.26	7832.27	128.14	0.001	0.87	7.9
6 classes	7777.77	7983.80	7834.59	20.15	0.02	0.76	7.6
7 classes	7869.80	7923.80	7898.37	158.68	0.001	0.86	2.5
8 classes	7877.41	7938.47	7883.75	37.08	0.001	0.80	3.4
9 classes	7873.95	7941.35	7858.07	48.14	0.001	0.80	1.2

Note. AIC = Akaike information criterion; BIC = Bayesian information criterion; SSABIC = Sample-size adjusted bayesian information criterion; BLRT = Bootstrapped Lo-Mendell-Rubin’s adjusted likelihood ratio.

).

Figure 1 presents the mean indicator scores across the five profiles. Profile 1 (n = 108, 11.8%) showed elevated psychological symptoms and high eco-anxiety across all subscales, particularly depressive symptoms and emotional–behavioral eco-anxiety, reflecting substantial distress and maladaptive responses to environmental challenges. This profile was labeled “highly impaired maladaptive eco-anxiety”. Profile 2 (n = 72, 7.9%) exhibited elevated anxiety and depressive mean scores but low eco-anxiety across all subscales, indicating substantial psychological distress unrelated to ecological and environmental concerns. This profile was labeled “psychological distress independent of eco-anxiety”. Profile 3 (n = 75, 8.1%) exhibited low anxiety and depressive symptoms and minimal behavioral eco-anxiety but exhibited however elevated eco-anxiety in the scales of emotional symptoms, rumination, and personal-impact anxiety, indicating low psychological general distress alongside heightened cognitive–emotional distress specifically related with ecological environmental issues. This profile was labeled “adaptive eco-anxiety”. Profile 4 (n = 440, 48.0%) showed low psychological symptoms and minimal eco-anxiety across all subscales, reflecting generally low distress and limited environmental concerns. This profile was labeled “non-anxious/disengaged”. Profile 5 (n = 222, 24.2%) exhibited moderate psychological symptoms and eco-anxiety across all subscales, reflecting intermediate levels of distress and environmental concern. This profile was labeled “moderate eco-anxiety”.

3.2. Multivariate Analysis of Variance

A MANOVA was conducted to examine differences among the five profiles in Connectedness with Nature and Protection of Nature. Assumption testing for normality, linearity, outliers, homogeneity of variance–covariance, and multicollinearity revealed no substantial violations. Significant differences among profiles were observed on the combined dependent variables, F (8, 1810) = 8.28, p < 0.001, η² = 0.03. Post hoc univariate analyses, using a Bonferroni-adjusted α = 0.05, indicated significant differences on both Connectedness with Nature and Protection of Nature (

Table 3. Tests of between-subjects effects.

Dependent Variables	Mean Square	df	F	p	Partial Eta Squared	Observed Power
Connectedness with nature	6.40	4	5.29	0.001	0.02	0.97
Protection of nature	16.81	4	12.07	0.001	0.05	0.99

). For Connectedness with Nature, Profile 3 scored significantly higher than Profiles 2 and 4 (p = 0.02). For Protection of Nature, Profile 3 scored significantly higher than all other profiles (p = 0.01–0.001) (

Table 4. Means, standard deviations of profiles on environmental identity dimensions.

Profile	Connectedness with Nature			Protection of Nature
	M	SD	95% CI	M	SD	95% CI
Profile 1: Highly Impaired Maladaptive Eco-anxiety	54.11	12.80	51.56; 57.17	17.13	5.37	15.85; 18.41
Profile 2: Psychological Distress Independent of Eco-anxiety	51.27	16.00	46.41; 56.14	16.13	7.18	14.02; 18.24
Profile 3: Adaptive Eco-anxiety	59.92	12.12	56.36; 63.47	20.69	5.00	19.24; 22.14
Profile 4: Non-Anxious/Disengaged	53.43	13.95	51.80; 55.06	16.48	5.80	15.80; 17.15
Profile 5: Moderate Eco-Anxiety	54.46	12.61	52.36; 56.56	16.99	5.87	16.06; 17.92

).

3.3. Multinomial Logistic Regression

A multinomial logistic regression was conducted with profile membership as the dependent variable and sociodemographic characteristics, previous exposure to climate-related stressors, and psychiatric history as predictor variables.

Table 5. Results of the multinomial logistic regression predicting latent profile membership.

Profile 1: Highly Impaired Maladaptive Eco-Anxiety
Effect	-2 Log Likelihood	Chi-Square	df	p	OR (95% CI)
Intercept	274.17	0.00	0
Gender (female)	276.79	4.19	1	0.04	1.97 (1.12; 3.31)
Age (young)	278.34	4.23	1	0.04	2.81 (1.05; 7.54)
Education (low)	275.04	1.63	1	0.20	1.51 (0.80; 2.86)
Professional status (not employed)	284.41	10.28	1	0.001	3.06 (1.51; 6.23)
Socioeconomic status (high)	274.18	0.01	1	0.91	1.04 (0.51; 2.10)
Living environments (urban)	280.58	6.40	1	0.01	3.88 (1.19; 12.64)
Parental status (no)	280.08	6.32	1	0.01	2.43 (1.19; 6.51)
Flood (yes)	275.46	1.13	1	0.24	1.48 (0.77; 2.85)
Droughts (yes)	274.22	0.01	1	0.95	1.02 (0.58; 1.80)
Wildfire (yes)	274.73	0.37	1	0.54	1.24 (0.62; 2.48)
Storms (yes)	274.28	0.01	1	0.94	1.04 (0.41; 2.67)
Prior psychiatric conditions (yes)	300.82	22.33	1	0.001	4.97 (2.60; 9.47)
Profile 2: Psychological Distress Independent of Eco-Anxiety
Effect	-2 Log Likelihood	Chi-Square	df	p	OR (95% CI)
Intercept	228.87	0.00	0
Gender (female)	231.61	2.65	1	0.10	1.73 (0.85; 3.51)
Age (young)	229.04	0.06	1	0.81	1.13 (0.42; 3.07)
Education (low)	229.23	0.15	1	0.70	1.15 (0.57; 2.30)
Professional status (not employed)	228.97	0.01	1	0.96	1.02 (0.49; 2.13)
Socioeconomic status (high)	229.93	0.29	1	0.59	1.24 (0.57; 2.73)
Living environments (urban)	230.58	1.61	1	0.20	1.76 (0.76; 4.09)
Parental status (yes)	229.82	0.86	1	0.36	1.59 (0.60; 4.20)
Flood (yes)	229.77	0.84	1	0.36	1.39 (0.70; 2.77)
Droughts (yes)	229.48	0.76	1	0.38	1.24 (0.55; 1.74)
Wildfire (yes)	228.98	0.02	1	0.88	1.06 (0.49; 2.30)
Storms (yes)	230.78	2.44	1	0.06	3.37 (0.76; 14.95)
Prior psychiatric conditions (yes)	237.90	10.73	1	0.001	3.52 (1.72; 7.20)
Profile 3: Adaptive Eco-Anxiety
Effect	-2 Log Likelihood	Chi-Square	df	p	OR (95% CI)
Intercept	233.85	0.00	0
Gender (female)	234.06	0.21	1	0.65	1.16 (0.62; 2.17)
Age (young)	234.34	0.50	1	0.48	1.43 (0.53; 3.86)
Education (low)	235.45	1.60	1	0.21	1.57 (0.78; 3.16)
Professional status (not employed)	236.22	2.37	1	0.12	1.79 (0.85; 3.76)
Socioeconomic status (high)	234.58	0.73	1	0.39	1.42 (0.65; 3.10)
Living environments (urban)	233.85	0.01	1	0.96	1.02 (0.37; 2.83)
Parental status (yes)	280.08	6.32	1	0.01	2.43 (1.19; 6.51)
Flood (yes)	236.30	2.44	1	0.12	1.73 (0.88; 3.41)
Droughts (yes)	238.29	4.55	1	0.03	2.05 (1.04; 4.05)
Wildfire (yes)	236.89	4.04	1	0.04	2.07 (1.08; 4.66)
Storms (yes)	234.39	0.54	1	0.46	1.39 (0.59; 3.29)
Prior psychiatric conditions (yes)	235.90	2.14	1	0.14	2.13 (0.71; 6.36)
Profile 4: Non-Anxious/Disengaged
Effect	-2 Log Likelihood	Chi-Square	df	p	OR (95% CI)
Intercept	501.47	0.00	0
Gender (male)	512.45	11.19	1	0.001	2.09 (1.30; 2.75)
Age (older)	505.67	4.03	1	0.04	1.78 (1.02; 2.33)
Education (low)	502.36	0.85	1	0.36	1.20 (0.81; 1.79)
Professional status (not employed)	502.66	1.39	1	0.24	1.29 (0.84; 1.98)
Socioeconomic status (high)	501.52	0.31	1	0.86	1.05 (0.64; 1.72)
Living environments (urban)	502.55	1.09	1	0.30	1.38 (0.75; 2.52
Parental status (yes)	503.58	2.23	1	0.14	1.52 (0.88; 2.64)
Flood (yes)	504.12	2.81	1	0.10	1.43 (0.94; 2.18)
Droughts (yes)	501.96	0.29	1	0.87	1.36 (0.56; 2.96)
Wildfire (yes)	502.03	0.55	1	0.46	1.19 (0.75; 1.87)
Storms (yes)	501.49	0.07	1	0.80	1.08 (0.60; 1.95)
Prior psychiatric conditions (no)	532.42	29.90	1	0.001	6.01 (2.85; 12.65)
Profile 5: Moderate Eco-Anxiety
Effect	-2 Log Likelihood	Chi-Square	df	p	OR (95% CI)
Intercept	461.31	0.00	0
Gender (female)	466.87	4.02	1	0.04	1.88 (1.27; 2.25)
Age (young)	463.77	2.41	1	0.12	1.64 (0.88; 3.05)
Education (low)	461.58	0.25	1	0.62	1.72 (0.72; 1.72)
Professional status (not employed)	439.38	0.01	1	0.99	1.01 (0.62; 1.59)
Socioeconomic status (high)	463.15	1.79	1	0.18	1.46 (0.83; 2.57)
Living environments (urban)	461.44	0.13	1	0.74	1.51 (0.64; 2.33)
Parental status (yes)	462.67	1.38	1	0.24	1.45 (0.78; 2.66)
Flood (yes)	463.03	1.71	1	0.19	1.34 (0.86; 2.09)
Droughts (yes)	467.54	4.56	1	0.03	1.97 (1.27; 2.43)
Wildfire (yes)	462.27	0.94	1	0.33	1.97 (1.27; 2.43)
Storms (yes)	461.32	0.01	1	0.93	1.03 (0.55; 1.94)
Prior psychiatric conditions (yes)	461.34	0.02	1	0.90	1.97 (1.27; 2.43)

shows the odds ratio tests for the multinomial logistic regression.

Gender significantly predicted profile membership. Participants in the Highly Impaired Maladaptive Eco-anxiety and Moderate Eco-Anxiety groups had approximately twice the odds of being female, whereas male gender was associated with higher odds of belonging to the Non-anxious/Disengaged group. Age and professional status significantly predicted profile membership. Younger participants had nearly threefold higher odds of belonging to the Highly Impaired Maladaptive Eco-anxiety group, while older participants were twice as likely to belong to the Non-anxious/Disengaged group. Non-employed individuals also had threefold higher odds of membership in the Highly Impaired Maladaptive Eco-anxiety group.

Living in urban areas was a significant predictor of belonging to the Highly Impaired Maladaptive Eco-anxiety group, increasing the odds by nearly fourfold. Parental status also predicted membership: participants with children had twice the odds of being in the Adaptive Eco-Anxiety group, whereas those without children had 2.5-fold higher odds of belonging to the Highly Impaired Maladaptive Eco-anxiety group.

Regarding direct exposure to extreme weather events, prior exposure to drought and wildfires significantly predicted profile membership. Exposure to drought and wildfires was associated with a twofold increase in the odds of belonging to the Adaptive Eco-anxiety group, whereas prior exposure to drought alone nearly doubled the odds of belonging to the Moderate Eco-anxiety group. Prior psychiatric history significantly predicted profile membership: absence of such history was associated with sixfold higher odds of belonging to the Non-anxious/Disengaged group, whereas its presence increased the odds of membership in the Highly Impaired Maladaptive Eco-anxiety and Psychological Distress Independent of Eco-Anxiety groups by fivefold and over 3.5-fold, respectively.

4. Discussion

This study addresses a gap in research on eco-anxiety, which is clinically significant when it disrupts daily functioning [17]. Evidence on its mental-health associations and underlying mechanisms remains limited [42]. Using LPA in a large Portuguese sample, we identified distinct profiles of eco-anxiety manifestations and examined their links with resilience and vulnerability factors, including nature connectedness, engagement in nature protection, and cumulative exposure to stressful climatic events. A person-centered approach such as LPA highlighted the heterogeneity of eco-anxiety and its integration within broader psychosocial contexts.

Eco-anxiety can be understood as a set of emotional, cognitive, and behavioral responses emerging under conditions of persistent uncertainty, reflecting not only exposure to environmental threats but also the instability of knowledge and limited perceived control. From the perspective of contemporary risk and resilience frameworks, it underscores the limitations of probabilistic models in contexts of deep uncertainty, where risk assessment may be distorted and adaptive responses constrained [61]. In the case of climate change, eco-anxiety thus reflects both environmental risk and enduring epistemic uncertainty, situating it at the intersection of vulnerability and adaptive capacity.

Resilience theory shifts the focus from prediction to adaptation, encompassing both recovery and the capacity to absorb and reorganise following disturbance. Rooted in ecological systems theory, it is now conceptualised as a dynamic capability involving flexibility, learning, and transformation [62,63]. Applied to eco-anxiety, these insights suggest that responses to environmental threats reflect differing adaptive capacities. Eco-anxiety may be maladaptive when individuals feel overwhelmed and unable to cope, but, when supported by adaptive processes, it may promote constructive action and psychological adjustment. This aligns with emerging views that resilience is not merely the absence of distress, but a dynamic process through which individuals engage with and adapt to complex, uncertain threats [64].

4.1. Profiles of Eco-Anxiety Manifestations

LPA identified five distinct eco-anxiety profiles, supporting its conceptualization as a multifaceted phenomenon and highlighting the utility of person-centered approaches [27]. Profiles 1 (“highly impaired maladaptive eco-anxiety”) and 3 (“adaptive eco-anxiety”) distinguish two pronounced patterns of ecological concern: one characterised by preserved functioning and adaptive engagement, and another marked by heightened eco-anxiety accompanied by substantial general distress [10]. From a resilience perspective, these profiles may reflect different capacities to cope with and adapt to environmental uncertainty. Profile 1 may indicate reduced resilience, where uncertainty and perceived uncontrollability amplify distress and impair functioning. In contrast, Profile 3 may represent resilient adaptation, in which concern about environmental threats is integrated into meaningful engagement without compromising psychological well-being [65].

Profile 2 (“psychological distress independent of eco-anxiety”) identifies general distress (anxiety and depression) that is partially autonomous from ecological concern [5], suggesting that vulnerability factors unrelated to environmental uncertainty may dominate psychological functioning in this group. Profile 4 (“non-anxious/disengaged”) reflects previously documented segments of the population with low environmental concern and low general distress [30], which may be interpreted as either low exposure, disengagement, or limited risk perception. Profile 5 (“moderate eco-anxiety”) mirrors graded patterns of eco-emotional intensity, challenging deterministic notions of “eco-paralysis” [24], and suggesting intermediate levels of adaptive engagement.

Profiles 3 and 5 exhibit partial alignment with established indicators of psychological distress. Profile 3 reflects strong environmental engagement with minimal general distress, whereas Profile 5 shows moderate co-occurrence of psychological symptoms and ecological concern. Interpreted through a resilience lens, these findings underscore that eco-anxiety does not uniformly lead to dysfunction; rather, its outcomes depend on the interplay between perceived uncertainty, coping resources, and adaptive capacities. This is consistent with resilience-oriented approaches to risk, which emphasize the ability to anticipate, cope with, and adapt to disruption rather than attempting to eliminate uncertainty altogether [66].

While eco-anxiety can be often adaptive, some forms of its manifestations are linked to significant psychological distress. For instance, profile 1, marked by high emotional and behavioral eco-anxiety alongside elevated psychological symptoms, represents a subgroup in which environmental concern co-occurs with depressive affect, maladaptive coping, and functional impairment [3,10]. This profile is also associated with persistent ecological preoccupation, limited coping resources, and existential anxiety, underscoring the dimensional nature of eco-anxiety and the need to identify factors that increase the vulnerability to severe manifestations of ecological anxiety [11,14].

Consistent with conceptualizations of eco-anxiety as a multidimensional construct which can be associated with general anxiety and depressive features [4,15], the Highly Impaired Maladaptive Eco-anxiety profile was characterized by elevated depressive symptoms alongside high emotional and behavioral eco-anxiety, reflecting maladaptive responses that can be associated with helplessness and hopelessness [3,35]. In contrast, the Moderate Eco-anxiety profile exhibited concurrent mid-range levels of psychological distress (anxiety and depression) together with moderate eco-anxiety, indicating an intermediate degree of distress [16,18].

Profiles 1, 3, and 5 illustrate distinct moderate to severe eco-anxiety configurations within a multi-dimensional framework. Profile 3 (“adaptive eco-anxiety”) is characterized by heightened rumination and concern about the personal implications of the environmental crisis, coupled with minimal behavioral symptomatology of eco-anxiety, suggesting that eco-anxiety may function adaptively by enhancing threat sensitivity, reflection, and environmental awareness [10,11]. In contrast, profile 1 (“highly impaired maladaptive eco-anxiety”) exhibits high cognitive, emotional, and behavioral symptoms alongside general psychological distress (general anxiety and depression), indicating a maladaptive pattern associated with functional impairment and increased psychopathology risk [2,67]. Profile 5 (“moderate eco-anxiety”) shows moderate levels across all domains. These patterns suggest that the intensity and configuration of cognitive, affective, behavioral, and personal-impact symptoms, rather than any single domain, distinguish adaptive from maladaptive eco-anxiety [12].

Profiles 2 and 4 are characterized by low levels of eco-anxiety among most participants but differ in their levels of general distress. While profile 2 (“psychological distress independent of eco-anxiety”), the smallest group, showed elevated general anxiety and depression but minimal eco-anxiety, suggesting distress largely unrelated to environmental concerns, on the other hand profile 4 (“non-anxious/disengaged”), the largest subgroup, exhibited low psychological symptoms and also minimal eco-anxiety, reflecting generally low distress as well as limited environmental engagement. However, the absence of manifest eco-anxiety in profile 4 should not be interpreted as emotional indifference or full disengagement [5]. Rather, minimal eco-anxiety—particularly among individuals in the non-anxious/disengaged profile—may reflect emotion-regulation processes such as denial, suppression, distancing, or rationalization, through which climate-related distress is managed and remains latent rather than consciously experienced [19,20].

4.2. Resilience and Vulnerability Factors

The present findings indicate that demographic, socio-environmental, and clinical factors jointly shape membership in eco-anxiety and psychological distress profiles, suggesting underlying mechanisms of vulnerability and resilience. However, this interpretation, which is predominantly centered on individual-level determinants, may be further strengthened by incorporating process-oriented and system-level perspectives on resilience. From this broader standpoint, resilience is not solely an individual trait or outcome but rather emerges dynamically from interactions among individuals, social structures, environmental conditions, and informational systems that shape the perception, interpretation, and response to climate-related stressors [68].

Female participants were overrepresented in the maladaptive and moderate eco-anxiety profiles, whereas males were predominantly classified as non-eco-anxious and non-distressed. This pattern is consistent with prior research indicating higher eco-anxiety among women, often associated with caregiving roles, greater perceived responsibility for others’ well-being, and heightened sensitivity to climate-related stressors [10,18]. From a systems perspective, such differences may reflect gendered positions within social and caregiving networks, where relational demands and risk communication processes amplify sensitivity to environmental threats, underscoring the co-production of vulnerability through social roles and structural expectations rather than solely individual dispositions [69]. Conversely, male gender was associated with higher odds of belonging to the Non-anxious/Disengaged group, consistent with evidence that men show lower climate-related concern and greater denial-oriented responses compared to women [41,42].

Age also significantly predicted psychological distress and eco-anxiety, with younger individuals exhibiting higher levels of both variables. This is consistent with evidence that those under 35 tend to experience greater eco-anxiety, reflecting their heightened exposure to long-term climate risks [14,34]. Within a system-level framing, younger cohorts are embedded in socio-temporal processes characterized by prolonged uncertainty, future-oriented risk accumulation, and intensified informational exposure. These dynamics may shape cognitive appraisal processes and collective narratives about climate futures, thereby reinforcing eco-anxiety as an emergent property of socio-informational systems [24,38].

Additionally, several socio-environmental factors significantly distinguished the profiles. Participants with children were more likely to belong to the Adaptive Eco-Anxiety group, a pattern that may be understood in light of parental caregiving motivations, whereby concern for children’s future fosters a more functional, action-oriented form of eco-anxiety associated with adaptive coping and greater engagement in sustainability-oriented behaviors [2,7]. Conversely, urban residence, unemployment or retirement, and childlessness increased the likelihood of membership in profiles characterized by both elevated psychological distress and eco-anxiety, consistent with evidence that cumulative social and environmental burdens heighten vulnerability [42,70].

Although seemingly counter-intuitive, the higher likelihood of urban respondents being classified as “Highly Impaired” may result from amplified risk perception in information-rich environments [71]. From a process-oriented perspective, urban systems are characterized by dense informational flows, fragmented social cohesion, and increased exposure to mediated representations of climate risk, which may intensify cognitive load and reduce the capacity for adaptive sensemaking [72].

Unemployment was also associated with high-distress profiles, likely due to financial insecurity, social marginalization, and limited access to coping resources associated with job loss, which can exacerbate general and climate-related anxiety [43]. Similarly, a previous history of psychiatric disorders predicted membership in elevated-distress profiles, reflecting the amplifying effect of pre-existing vulnerabilities. From a systems viewpoint, these factors can be understood as disruptions in access to stabilizing resources within broader socio-economic systems, reducing adaptive capacity and constraining feedback mechanisms that would otherwise support recovery and regulation [5,35].

Direct exposure to climate extremes, particularly drought and wildfire, was associated with adaptive and moderate eco-anxiety profiles, supporting evidence that personal experience heightens climate risk perception and engagement [2,4]. However, cumulative exposure is also linked to financial strain and psychological vulnerability [69,73]. From a resilience engineering perspective, such experiences may enhance situational awareness and adaptive capacity, while repeated exposure—especially in the absence of institutional support—can overwhelm coping systems, reflecting the context-dependent nature of resilience [72]. In rural settings, strong social networks may foster collective coping despite elevated individual stress [3], although recurrent droughts in Portugal may simultaneously increase risk awareness and obscure underlying vulnerabilities [74].

Another group of variables showing statistically significant associations with the profiles concerned individuals’ relationship with nature, including reported connectedness and pro-environmental attitudes. Individuals in Profile 3 (adaptive eco-anxiety) exhibited greater connectedness to nature and stronger endorsement of its protection than other groups, suggesting that variations in eco-anxiety may reflect domain-specific engagement with sustainability-related environmental issues rather than generalized negative affectivity [47]. This combination of affective and cognitive engagement likely enhances attention to environmental information, moral concern, and pro-environmental and sustainable behavior [48], supporting resilience while potentially increasing vulnerability through heightened identification with nature and sensitivity to environmental loss within the context of sustainability challenges [2,42].

The findings indicate that vulnerability to eco-anxiety emerges from the cumulative and dynamic interplay of social, environmental, and individual stressors. Chronic conditions—such as urban residence, limited access to nature, unemployment, low income, and low educational attainment—may deplete coping resources and, when combined, amplify psychological burden and reduce resilience [43,44]. From a systems perspective, resilience within complex adaptive systems depends on both resource availability and adaptive capacities, as pre-existing vulnerabilities interact with environmental, institutional, and informational factors to shape risk perception and management, highlighting the need for integrative, multilevel resilience frameworks [70,74].

4.3. Limitations

Several limitations of this study should be acknowledged. The cross-sectional design precludes causal inference, underscoring the need for longitudinal or experimental research. Reliance on self-report measures may introduce response biases, common method variance, and potential construct overlap. In the present study, correlations among measures were low to moderate (r = 0.18–0.47), suggesting some discriminant validity; however, reliance on a single method represents a limitation. Future research should incorporate behavioral, ecological, or physiological assessments to strengthen construct validity and reduce potential biases inherent in self-report data. The convenience sampling via mailing lists and social media may have led to an overrepresentation of digitally connected or environmentally engaged individuals, potentially limiting generalizability. The predominantly female and middle-aged sample further constrains representativeness. Other relevant factors, such as social support, coping strategies, or personality traits, were not assessed. Finally, although latent profile analysis identified distinct eco-anxiety patterns, replication in independent and more diverse samples is needed to confirm profile stability and external validity.

5. Conclusions

This study underscores the heterogeneous, nuanced, and multidimensional nature of eco-anxiety within the context of sustainability challenges, revealing distinct profiles ranging from adaptive, cognitively and emotionally engaged responses to sustainability and climate change to maladaptive patterns accompanied by high general distress and functional impairment. Vulnerability and resilience to climate and personal life stressors appear to be shaped by individual, socio-environmental, and clinical factors, including nature connectedness, environmental engagement, cumulative stressors, psychiatric history, and direct exposure to ecological and climate-related catastrophes. Eco-anxiety may hypothetically function as both a psychological resource, promoting adaptive engagement with sustainability and climate change, and a source of personal distress, thereby impairing well-being.

These findings have important implications for climate adaptation and mental health interventions. Strategies should explicitly account for the heterogeneity of eco-anxiety, including its diverse manifestations, underlying determinants, and associated covariates, in order to effectively distinguish between adaptive and maladaptive responses. Interventions should prioritize the promotion of adaptive coping strategies, nature-based resilience, and psychological resources, while ensuring targeted support for individuals at risk of persistent or maladaptive distress. Furthermore, enhancing access to natural environments and strengthening environmental identity may serve as protective factors, particularly when integrated with efforts to address cumulative social and environmental stressors. Such a multidimensional approach may help mitigate the psychological impacts of climate change while fostering both individual well-being and collective environmental stewardship.