From Contamination to Impact: Cadmium Levels in Cacao Soil and Beans and Their Effect on Economic Sustainability Along the Coast of Ecuador

Abstract

This study assessed, in two coastal locations of Ecuador (Cerecita, Guayas; Bajada de Chanduy, Santa Elena), cadmium (Cd) occurrence in cacao cultivated soils, its transfer to plant tissues (leaves and cotyledon/beans), and its implications for producers’ economic sustainability. Twelve cacao-producing sites in Cerecita and eleven in Bajada de Chanduy were georeferenced, and thematic GIS maps were generated to identify potential Cd hotspots. Sampling comprised topsoil (0–10 cm), leaves (fourth fully expanded leaf), and dried/fermented beans, followed by laboratory Cd quantification. In addition, producer surveys were conducted to characterize productive and economic structure, the economic sustainability index (IK) was calculated using Sarandón’s methodology, and interviews with collectors and agri-export companies were performed. Soil Cd levels were comparable between locations (0.24–1.55 mg kg⁻¹), whereas higher concentrations were detected in cotyledons/beans (0.53–5.01 mg kg⁻¹) and leaves (1.13–11.07 mg kg⁻¹), following the pattern leaves > cotyledon > soil. From an economic perspective, all farms exhibited IK < 2, with a marked territorial gap (≈1.6 in Cerecita vs. ≈0.5 in Chanduy). Cadmium in cocoa beans poses a long-term risk to marketing; in addition, total cadmium in the soil did not consistently predict cadmium in the cotyledons, and adverse impacts are amplified in territories with limited economic capacity to respond.

1. Introduction

Cadmium (Cd) in cacao has become a major challenge for natural resource management and agri-export competitiveness due to its toxicity and the regulatory limits applied to cocoa-derived products. From a food-safety and market-access perspective, Gramlich et al. [1] showed that Cd in beans may exceed European standards even in soils without apparent contamination, supporting the need for risk-based monitoring. However, other authors argue that the discussion should not be reduced to commercial thresholds: Chavez et al. [2] showed in Southern Ecuador that bean Cd concentrations are related to extractable soil fractions, suggesting that management and bioavailability control Cd transfer. In contrast, Barraza et al. [3] documented settings influenced by petroleum-related activities in which Cd bioaccumulation reinforces that the problem integrates environmental and public-health dimensions, and that its origin and dynamics are not always attributable to a single cause. The European Union has progressively tightened its regulatory framework on cadmium in cocoa and chocolate products, driven by growing concerns over consumer health risks associated with chronic cadmium exposure. The most recent framework, Regulation (EU) 2023/915 [4], which replaced the former Regulation (EC) No. 1881/2006, establishes the following maximum cadmium limits for cocoa-derived products: 0.10 mg kg⁻¹ for milk chocolate with less than 30% total cocoa content; 0.30 mg kg⁻¹ for chocolate with 30–49% cocoa content; 0.80 mg kg⁻¹ for chocolate with more than 50% cocoa content; and 0.60 mg kg⁻¹ for cocoa powder sold to the final consumer. These thresholds have direct implications for cocoa-producing countries in Latin America, particularly Ecuador, Peru, and Colombia, where volcanic and geogenic soil conditions contribute to naturally elevated Cd concentrations [4]. Compliance with these limits is increasingly critical for market access, as non-conforming shipments are subject to rejection at EU borders, placing smallholder farmers at significant economic risk. A persistent tension in the literature concerns attributing Cd to geogenic versus anthropogenic sources. On the one hand, Guarín et al. [5] in Piura (Peru) reported that lithology and alluvial sediments may impose an elevated baseline, supporting the interpretation of a partially natural origin. Along similar lines, Bravo et al. [6] identified wide variability of Cd in cacao soils at the national scale in Colombia, suggesting landscape heterogeneity conditions risk even before management interventions are considered. Conversely, Scaccabarozzi et al. [7] demonstrated that site and management factors are associated with Cd concentrations in cacao soils, indicating that agricultural practices and land-use decisions can amplify or attenuate effective Cd exposure. This duality implies that territorial diagnoses should avoid monocausal explanations and instead evaluate sources and processes within local contexts.

The methodological debate deepens when considering which measurement better predicts Cd in beans: total Cd or bioavailable fractions. Gramlich et al. [1] argued that bioavailable Cd measured with diffusive gradients in thin-films (DGT) better predicts Cd in beans and leaves than total Cd, supporting availability-oriented management approaches. Consistently, Chavez et al. [2] found associations between bean Cd and soil-extractable Cd (e.g., Mehlich-3 or HCl), reinforcing that soil chemistry governs transfer. Nevertheless, other studies suggest that the soil–plant relationship may be unstable due to interactions among edaphic properties, microenvironments, and spatial variability: Scaccabarozzi et al. [7] emphasized combinations of site and management factors that hinder simple extrapolations, whereas Bravo et al. [6] highlighted the need to map variability to avoid inferences based on averages. Consequently, the evidence suggests that prediction requires integrating matrices and scales rather than relying on a single indicator.

In Ecuador, the available evidence is concentrated in specific territories and leaves gaps for coastal zones with contrasting environments. On the one hand, Chavez et al. [2] provided key evidence on bean Cd and its relationship with soil in the south of the country; on the other, Barraza et al. [3] showed that sources associated with petroleum activities can modify exposure pathways and risk in cacao. However, these frameworks do not allow robust inference for areas such as Cerecita (Guayas) and Bajada de Chanduy (Santa Elena), where soil variability and land-use history may produce patterns distinct from those reported elsewhere [5,7]. Complementarily, a national study indicated that a substantial proportion of bean samples exceeds regulatory thresholds in cocoa products, suggesting a risk of rejection in regulated markets and underscoring the need for localized diagnoses [8,9]. Therefore, generating territorial evidence for Ecuador’s coastal region is critical to inform mitigation measures and risk management.

Cadmium (Cd) accumulation in cacao has emerged as a critical constraint for the sustainability of tropical production systems because bean contamination depends not only on total soil Cd but also on the edaphic controls that regulate its bioavailability and transfer to the plant. In Ecuador, Argüello et al. [8] showed that bean Cd increases with increasing total soil Cd and with decreasing soil pH and organic carbon, supporting the view that acid conditions and low organic matter favor Cd mobility and soil-to-bean transfer. Likewise, ref. [2] reported in Southern Ecuador that bean Cd was closely related to extractable soil Cd, reinforcing that available rather than total Cd may better explain cacao uptake under certain field conditions. In contrast, ref. [1] found in Honduras that DGT-available Cd was the best predictor of bean Cd, and that when DGT was excluded, bean Cd was best explained by total soil Cd, pH, and geology, highlighting the importance of both soil reactivity and parent material. A complementary perspective was provided by [10], who found that management and genotype explained only a limited fraction of bean Cd variability, suggesting that soil-related processes may outweigh agronomic effects in many cacao systems. Therefore, Cd transfer in cacao should be understood as the result of a dynamic soil chemical balance in which pH, organic matter, and exchange related retention processes interact to either buffer or intensify the movement of Cd from soil to economically relevant tissues such as cacao beans.

Beyond diagnosis, Cd is linked to economic sustainability by increasing the probability of non-compliance with international regulations and reducing access to premium markets, with effects on prices and expected profitability [7,11,12,13]. From an economic evaluation standpoint, Vázquez-deCastro et al. [12] estimated, using surveys and contingent valuation, relevant costs under non-compliance scenarios associated with the European Union regulatory framework (0.8 mg kg⁻¹ in cocoa products), suggesting potentially high aggregate impacts if non-adaptation strategies prevail. However, other authors emphasize that impacts are not inevitable: technical options exist to reduce Cd bioavailability or accumulation and thus mitigate economic risks, including amendments that increase pH and organic matter and microorganisms that modify Cd availability [7,8,9,13,14]. In contrast, effectiveness may depend on genotype, site conditions, and post-harvest practices; therefore, responses require local evidence and combined strategies [15,16].

In this context, the present study integrates an environmental and socioeconomic approach to bridge the gap between cadmium measurement and evidence-based decision-making for sustainable cacao production. On the one hand, spatial diagnosis of Cd in soils and beans can help prioritize areas and dominant transfer processes [1,2]; on the other hand, interpreting Cd exclusively through compliance may underestimate implications for livelihoods and production decisions [11,12]. Accordingly, the objectives of this research were to (i) diagnose the presence and magnitude of Cd in cacao soils and beans in Cerecita and Bajada de Chanduy (Guayas), and (ii) estimate the potential impact of these levels on the economic sustainability of producers in Guayas and Santa Elena considering market-access risk.

2. Materials and Methods

2.1. Study Area

For the present study, sampling plots were selected through a participatory process, resulting in the inclusion of farmers from the northern area of Guayas Province and the southern area of Santa Elena Province, in the sectors locally known as Cerecita and Bajada de Chanduy, respectively. Producers’ farms were georeferenced, and the information was incorporated into a Geographic Information System (GIS) to generate thematic maps aimed at identifying the presence of cadmium (Cd²⁺) in soils and plant tissues within this cacao-producing area. Cerecita’s community is located in the Juan Gómez Rendón (Progreso) parish, Guayaquil canton, Guayas Province, approximately 48 km west of Guayaquil, near km 51 of the Guayaquil–Salinas road, km 51 on the coastal highway to the Pacific. Bajada de Chanduy is located very close to Cerecita, at a distance of 3.8 km; however, this community falls under the jurisdiction of Santa Elena canton, which borders Guayas Province.

This area has a steppe or semi-arid climate according to the Köppen classification, characterized by annual rainfall ranging between 200 and 400 mm and a mean annual temperature of 23.3 °C recorded at the Progreso meteorological station, with August as the warmest month and February as the coolest month [17]. The climate is marked by two well-defined seasons: a rainy season from December to May, which concentrates approximately 90% of the annual precipitation, and a dry season from June to November [18]. Temperatures in the region range between 21 °C and 35 °C. These conditions are generally favorable for agriculture; however, the combination of high temperatures and water scarcity poses significant challenges for agricultural production [19]. From an edaphological perspective, Guayaquil canton is characterized by alluvial soils with variable textures, including clays, gravels, silts, and fine to medium grained sands, formed on alluvial and colluvial alluvial deposits typical of Ecuador’s coastal plains, with slopes ranging from 0 to 5% and a relative relief of less than 5 m [20].

Soils in the area exhibit undulating topography with low slopes, directly influenced by river dynamics and susceptible to erosion and sedimentation processes that respond to climatic variability [18]. Agriculture in this arid zone is constrained by semi-arid climatic conditions and high evapotranspiration rates, making the development of technically managed irrigation systems necessary to ensure the viability of crops, including cacao [21].

2.2. Producers in Cerecita and Bajada de Chanduy

Once the farms were identified, they were georeferenced to produce a location map of cacao-producing farms. Soil, leaf, and pod samples were collected to determine cadmium occurrence indices, using the geographic coordinates of each sampling point recorded with a high-precision GPS (Figure 1). A total of eight cacao farms were sampled in Cerecita; in four of these farms, two samples were collected because their cultivated area exceeded 100 ha, and 11 in Bajada de Chanduy; all farms were in production at the time of the study.

Producers were interviewed to document crop management practices, particularly fertilization, irrigation, litter management, and harvesting. In addition, semi-structured interviews were conducted in 2025 with a total of seven to ten key informants, including cacao collectors and representatives of agroexport companies operating in the study area. Interviewees were selected through a combination of criteria: direct contact established during fieldwork and snowball sampling based on referrals from producers. Each interview lasted approximately 30 to 60 min. The interviews explored topics related to price dynamics, commercialization conditions, incentive structures, and risk distribution along the cacao value chain. Responses were analyzed through thematic categorization, identifying recurring patterns across informants. The resulting insights were used as contextual evidence to support, nuance, or challenge interpretations of the economic sustainability of cacao producers in the study area. Anonymity was guaranteed to all participants.

2.3. Sampling Techniques for Soil, Leaf, and Cacao Pod Collection

2.3.1. Soil Sampling

Composite soil samples of approximately 1 kg were collected around representative, productive cacao plants within each farm. Using a soil auger and sampling within the fertilization zone, the mineral soil layer was removed from the surface down to 10 cm depth. Each soil sample was placed in a new plastic bag and labeled by plot and site to prevent cross-contamination; labels were kept separate from direct contact with the soil.

Samples were air-dried under greenhouse conditions for five days, ground, sieved through a 2 mm mesh, and transferred to transparent plastic bags previously labeled for subsequent analysis at the Heavy Metals Laboratory (LMPesados) of the Department of Soil and Water Management (DMSA), Tropical Experimental Station Pichilingue (EETP), National Institute of Agricultural Research (INIAP), to determine pseudo-total Cd concentrations.

2.3.2. Leaf Sampling

Leaf samples were collected using pruning shears from the same cacao plant where the soil sample was obtained. The fourth fully expanded leaf was cut from the mid-canopy branch at the four cardinal points of the selected plant (four leaves per plant), following the methodology proposed by Lainez [22]. Leaves were placed in labeled paper bags and transported to the laboratory for tissue analysis, including washing with distilled water and acidified HCl (0.1 M), oven-drying for 72 h under forced-air circulation, and grinding using a Wiley-type mill, thereby preparing the material for digestion and Cd determination.

2.3.3. Cacao Bean Sampling

A composite sample consisting of at least two mature pods was collected from the same cacao plant used for soil and leaf sampling. Under greenhouse conditions, beans were extracted and fermented for five days in Rohan-type wooden micro-boxes. Fermented beans were placed in labeled paper bags (as for leaf samples) and oven-dried under forced-air circulation at 70 °C for 72 h.

After drying, cotyledons (beans without testa) were ground using a Wiley-type tissue mill (IKA) and stored in labeled plastic bags prior to laboratory analysis for Cd determination.

2.4. Analytical Methods for Cd Determination in Soils, Leaves, and Beans

Pseudo-total Cd in soils was determined using aqua regia digestion. For Cd determination in leaf and bean samples, nitric–perchloric mineralization (4:1 ratio) was applied following the methodology described by Carrillo Zenteno et al. [23]. Cadmium quantification in the resulting extracts (soil and plant tissues) was performed using a PerkinElmer Analyst 400 atomic absorption spectrophotometer (Perkin Elmer, model AAnalyst 800, Yokohama, Japan) equipped with a graphite furnace. To ensure analytical accuracy and quality control throughout the procedure, certified reference materials (CRMs) were included: NIST 2709a (San Joaquin Soil) for the soil matrix, and NIST 1575a (Pine Needles) together with ERM-BB512 (Dark Chocolate) for the validation of plant tissues. Recovery percentages ranged from 99% to 107% for both soil and plant tissue matrices, confirming the reliability of the methodology employed. Results are reported in mg kg⁻¹.

2.5. Characterization of Cacao-Producing Farms

Producer surveys were conducted to document farm conditions and management practices relevant to the study objectives. Survey results were processed using the statistical software R (R version 4.3.2). Based on a prior characterization of the study area, indicators were developed to estimate the sustainability of cacao producers using the methodology proposed by Sarandón and Flores [24]. This preliminary characterization made it possible to identify the main productive, economic, social, and ecological features of the evaluated farms and served as the basis for selecting the indicators and sub-indicators included in the sustainability assessment.

An initial proposal of indicators and their relative weighting coefficients was then established according to their relevance within the local production context. Subsequently, this proposal was presented to a group of experts in sustainability, who evaluated the pertinence of the selected indicators and their assigned weights. As a result of this review, some indicators were retained, others were modified or removed, and the weighting coefficients were adjusted to obtain the final structure of the sustainability assessment.

Social, ecological, and economic dimensions were assessed through 11 indicators and 29 sub-indicators (see Appendix A). To enable comparison across dimensions, indicators were standardized on a 0 (least sustainable) to 4 (most sustainable) scale. The standardized values were then weighted according to their relative importance and consolidated in a spreadsheet for visualization using radial (spider) charts.

The analyzed dimensions were computed using the following equations:

Economic Index IK

$$\mathrm{IK}={\displaystyle \frac{A_1+B_1+2\left(\frac{C_1+C_2+C_3+C_4+C_5+C_6}{6}\right)}{4}}$$

(1)

Ecological Index IE

$$\mathrm{IE}={\displaystyle \frac{\left(\frac{A_1+A_2}{2}\right)+\left(\frac{B_1+B_2+B_3}{3}\right)+C_1+\left(\frac{D_1+D_2+D_3+D_4}{4}\right)+2\left(\frac{E_1+E_2+E_3}{3}\right)}{6}}$$

(2)

Socio Cultural Index ISC

$$\mathrm{ISC}={\displaystyle \frac{\left(\frac{A_1+A_2+A_3+A_4}{4}\right)+B_1+\left(\frac{C_1+C_2}{2}\right)}{3}}$$

(3)

Overall sustainability index (ISG)

$$\mathrm{ISG}={\displaystyle \frac{\mathrm{IK}+\mathrm{IE}+\mathrm{ISC}}{3}}$$

(4)

2.6. Geostatistical Analysis

Spatial analysis was conducted in QGIS to generate thematic maps of Cd distribution in soil and cacao beans. The interpolation dataset consisted of georeferenced sampling points and their corresponding Cd concentration values. Spatial interpolation was performed using the inverse distance weighting (IDW) method in QGIS. This deterministic interpolation method was used as an exploratory approach to represent spatial patterns of Cd distribution within the sampled study area. The resulting interpolated surfaces were used to generate thematic maps of Cd distribution. Because interpolation was based exclusively on sampled cacao producing sites, the resulting spatial patterns are representative only of the sampled production area and should not be extrapolated beyond that domain.

3. Results and Discussion

3.1. Characterization of Producers

Producers in Bajada de Chanduy belong to a farmers’ association that was formed after the Ministry of Agriculture promoted the production of Nacional cacao in the area, with production surfaces ranging from 0.5 to 2 ha. They do not use technology for crop management; however, because the zone is dry, they rely on sprinkler, drip, and micro-sprinkler irrigation systems that they installed themselves in an artisanal manner, based on recommendations from equipment vendors. They are not organic producers, but their agricultural practices are also not intensive conventional agriculture: they fertilize twice per year with a complete fertilizer, weed manually, leave leaf litter in the field, and dispose of diseased pods and branches by burying or burning them. They harvest every fifteen days, obtaining approximately 45 to 90 kg of dry cacao, which they transport to collectors in Durán (Guayas), receiving the price available at that time. They are aware that the average price per quintal of cacao (45 kg) in Ecuador has shown a strong recovery over the last two years. According to the National Association of Cocoa Exporters of Ecuador (ANECACAO) and the Central Bank of Ecuador (BCE), the producer price rose from approximately USD 100 per quintal in 2023 to an annual average of USD 380–400 in 2024, with a historic peak exceeding USD 460 in late 2024, driven by a global supply deficit attributed to reduced production in Côte d’Ivoire and Ghana [25]. This price surge positioned cacao as Ecuador’s third-largest non-petroleum export product in 2024, generating USD 3.618 billion according to BCE data, representing a 159% increase compared to the previous year. In 2025, prices fluctuated between USD 265 and USD 413 per quintal, reflecting ongoing volatility in international futures markets [26,27,28,29,30,31]. Nevertheless, to meet their economic needs and food security, they cultivate other crops such as maize, cassava, plantain, and papaya, and they also produce charcoal for sale.

In contrast to Chanduy producers, Cerecita producers are more technified. Fifty percent of the interviewees cultivate areas of around 10 ha, and the other fifty percent between 100 and 150 ha, with micro-sprinkler irrigation systems installed by specialized companies. Fertilization management is based on soil analysis information, and they use amendments to improve soil pH. Leaf litter is left on-site, and at the beginning of the plantation they intercrop with plantain, which is removed once production begins; for the CCN-51 clone, production starts at 18 months [32]. All producers sell dry cacao to obtain higher profits, to an exporting company or collector who picks it up in situ, including Nestlé.

3.2. Cadmium Content in Soils and Cacao Tissues in Cerecita and Bajada de Chanduy

Cadmium (Cd) concentrations were analyzed in soil, cotyledon, and cacao leaves at two locations in Guayas Province: Cerecita (12 sites) and Bajada de Chanduy (11 sites) (see Figure 2). Argüello et al. [8] reported an average soil Cd concentration of 0.44 mg kg⁻¹; whereas Ecuadorian legislation establishes a maximum Cd level of 0.5 mg kg⁻¹ [33]. In Cerecita, only farmers 4, 6, 10, and 11, and in Bajada de Chanduy only farmers 4, 8, and 9, were below this reference value; the remaining farms exceeded it.

In the European Union, cadmium maximum levels are established for cocoa- and chocolate-derived products rather than for raw cocoa beans. Depending on product type and cocoa content, these limits range from 0.10 to 0.80 mg kg⁻¹ [8,13,34,35]. In practice, a technical threshold of 0.60–0.80 mg kg⁻¹ is often considered to allow industry compliance with these limits [8,34]. Therefore, the cotyledon Cd concentrations observed in this study should not be interpreted as direct regulatory non-compliance. Instead, they indicate a potentially relevant commercial concern, particularly because many sampled farms showed bean concentrations within or above ranges that could complicate compliance in downstream cocoa products. This pattern should be interpreted in the context of the study area, which was selected as a known Cd hotspot and does not represent cacao-growing conditions across Ecuador as a whole.

In both locations, a consistent magnitude pattern was observed across leaves, cotyledon, and soil suggesting differential Cd accumulation across plant matrices. Intra-location variability was relevant, particularly in Bajada de Chanduy, where an extreme case (farmer 3) showed high values in all three matrices (soil = 1.45 mg kg⁻¹; cotyledon = 5.01 mg kg⁻¹; leaves = 11.07 mg kg⁻¹), markedly increasing the mean and dispersion of the dataset. During the interview with the farm owner, charcoal production was reported within the production unit. This activity is mentioned here as a contextual observation that could potentially contribute to local Cd inputs; however, it was not directly evaluated in this study and therefore cannot be identified as a confirmed source of contamination. Source attribution was beyond the scope of the present work, and this point should be interpreted as a possible explanation requiring further investigation.

In Guayas, the high foliar Cd concentrations (Cerecita: 3.17 ± 1.24 mg kg⁻¹; Bajada de Chanduy: 3.22 ± 2.69 mg kg⁻¹) indicate that leaves could be used as an early monitoring tool to identify farms under higher Cd pressure. However, this should not replace bean analysis, as the leaf–bean relationship may be non-linear and vary across environments [1,3] (see Figure 3).

At a descriptive level (see

Table 1. Descriptive statistics of cadmium concentrations in soil, cotyledon, and leaves by locality.

Locality	Matrix	n	Mean (SD)	Median [Q1–Q3]	IQR	Min–Max	CV (%)
Cerecita	Soil	12	0.70 (0.32)	0.57 [0.48–0.84]	0.36	0.34–1.55	46.61
Cerecita	Cotyledon	12	1.40 (0.59)	1.25 [1.14–1.53]	0.38	0.65–2.78	42.16
Cerecita	Leaves	12	3.17 (1.24)	3.63 [2.67–3.94]	1.27	1.13–4.73	39.03
Bajada de Chanduy	Soil	11	0.69 (0.31)	0.68 [0.51–0.78]	0.28	0.24–1.45	44.67
Bajada de Chanduy	Cotyledon	11	1.74 (1.24)	1.43 [1.04–2.04]	1.00	0.53–5.01	71.23
Bajada de Chanduy	Leaves	11	3.22 (2.69)	2.51 [2.08–3.00]	0.92	1.13–11.07	83.46

Note: Cd is expressed as mg kg⁻¹. SD: standard deviation; Q1 = 25th percentile; Q3 = 75th percentile; IQR = interquartile range; CV = coefficient of variation. Outliers were identified using the 1.5×IQR rule: Cerecita (soil: Producer 2; cotyledon: Producers 3 and 8) and Bajada de Chanduy (Producer 3 in soil, cotyledon, and leaves).

), Cerecita showed a median soil Cd concentration of 0.57 mg kg⁻¹ (IQR = 0.36) and a median leaf Cd concentration of 3.63 mg kg⁻¹ (IQR = 1.28). Bajada de Chanduy showed a median soil Cd concentration of 0.68 mg kg⁻¹ (IQR = 0.28) and a median leaf Cd concentration of 2.51 mg kg⁻¹ (IQR = 0.93). Because several distributions did not meet normality (Shapiro–Wilk test; $p<0.05$ for soil and leaves in Cerecita, and for cotyledon and leaves in Bajada de Chanduy), nonparametric tests were prioritized for inference.

To assess within-locality differences among matrices under a producer-related (repeated-measures) design, the Friedman test was applied. In Cerecita, results indicated ${\chi }^2\left(2\right)=20.17$, $p<0.001$, with Kendall’s $W=0.84$ (very large effect). In Bajada de Chanduy, ${\chi }^2\left(2\right)=18.18$, $p<0.001$, with Kendall’s $W=0.83$ (very large effect). Post hoc pairwise comparisons (Wilcoxon tests with Holm adjustment) confirmed significant differences for all pairs (

Table 2. Within-locality differences in Cd concentrations among soil, cotyledon, and leaf samples.

Locality	Global Test	Pairwise Comparison	W	p (Holm)	${\mathit{r}}_{\mathit{rb}}$
Cerecita	Friedman: ${\chi }^2\left(2\right)=20.17$, $p<0.001$, Kendall’s $W=0.84$	Soil vs. cotyledon	1.0	0.001	0.97
		Soil vs. leaves	0.0	0.001	1.00
		Cotyledon vs. leaves	1.0	0.001	0.97
Bajada de Chanduy	Friedman: ${\chi }^2\left(2\right)=18.18$, $p<0.001$, Kendall’s $W=0.83$	Soil vs. cotyledon	1.0	0.002	0.97
		Soil vs. leaves	0.0	0.002	1.00
		Cotyledon vs. leaves	1.0	0.002	0.97

Note: W = Wilcoxon statistic; $r_{rb}$ = rank-biserial correlation (effect size).

), empirically supporting that foliar tissue concentrates Cd to a greater extent than cotyledon and soil.

For the between-locality comparison (Cerecita vs. Bajada de Chanduy), Mann–Whitney U tests were applied separately for each matrix. No statistically significant differences were detected for soil ($U=64.0$, $p=0.926$), cotyledon ($U=59.5$, $p=0.710$), or leaves ($U=88.0$, $p=0.185$). Although the leaf median was higher in Cerecita, the effect size was moderate (Cliff’s $\delta =0.33$), and the statistical evidence was insufficient to conclude a systematic difference at this sample size (

Table 3. Comparison of Cd concentrations between Cerecita and Bajada de Chanduy by sample matrix.

Matrix	U	p	Median (Cerecita)	Median (Bajada de Chanduy)	Cliff’s $\mathit{\delta }$ (C–Ch)
Soil	64.0	0.926	0.57	0.68	−0.03
Cotyledon	59.5	0.710	1.25	1.43	−0.10
Leaves	88.0	0.185	3.63	2.51	0.33

Note: U = Mann–Whitney test statistic. Cliff’s $\delta >0$ indicates a tendency toward higher values in Cerecita, whereas $\delta <0$ indicates a tendency toward higher values in Bajada de Chanduy.

).

Finally, the association among matrices was examined using Spearman correlations (

Table 4. Spearman correlations among soil, cotyledon, and leaf Cd concentrations.

Locality	Pairs	$\mathit{\rho }$	p
Cerecita	Soil—cotyledon	0.56	0.061
Cerecita	Soil—leaves	0.53	0.075
Cerecita	Cotyledon—leaves	0.09	0.791
Bajada de Chanduy	Soil—cotyledon	0.66	0.028
Bajada de Chanduy	Soil—leaves	0.42	0.193
Bajada de Chanduy	Cotyledon—leaves	0.57	0.067

Note: In Bajada de Chanduy, an extreme case was identified (Farmer 3). In a sensitivity analysis excluding Farmer 3, correlations decreased and became non-significant ($p>0.10$).

). In Cerecita, soil–cotyledon ($\rho =0.56$, $p=0.061$) and soil–leaves ($\rho =0.53$, $p=0.075$) were moderate but did not reach significance at the 5% level. In Bajada de Chanduy, the soil–cotyledon correlation was significant ($\rho =0.66$, $p=0.028$), whereas soil–leaves was not. Given the presence of an extreme case (farmer 3), a sensitivity analysis was conducted by excluding it: correlations decreased and became non-significant ($p>0.10$), suggesting that part of the statistical signal is concentrated in a localized hotspot rather than reflecting a stable relationship across the whole locality.

3.3. Thematic Maps of Cadmium Concentration in Soils and Plant Tissues

In Figure 4, the sampling sites and Cd concentrations in beans are shown for Cerecita and Bajada de Chanduy. Figure 5 shows the spatial distribution of Cd in soils across the studied areas.

Maps may reflect the influence of a small number of farms with exceptional conditions rather than a stable mechanism representative of an entire locality. The literature agrees that total soil Cd is an imperfect predictor of Cd in beans, and that the bioavailable fraction for example, estimated using chemical extractants or by Diffusive Gradients in Thin Films (DGT) often better explains actual uptake [1,2]. Therefore, the practical contribution of the statistical analysis is twofold: it confirms that a soil bean component exists, but it also warns that efficient interventions should prioritize hotspot identification and bioavailability assessment, rather than relying exclusively on averages of total soil Cd.

The presence of extreme values, as observed in Bajada de Chanduy (Farmer 3: soil 1.45; cotyledon 5.01; leaves 11.07 mg kg⁻¹), suggests local processes that increase Cd load and/or availability. The reported background of activities associated with charcoal production allows proposing, as a hypothesis, a point source or local deposition that intensifies exposure within a reduced radius; however, this study does not allow causal attribution and requires verification through targeted sampling around the potential source and complementary analyses. Beyond point sources, the literature suggests several plausible and potentially coexisting routes: (i) geogenic contributions linked to parent materials and alluvial sediments which, based on our investigation, would not be the case in this area, and (ii) diffuse anthropogenic inputs associated with agricultural inputs and practices, which were evidenced here. In Peru and Honduras, for example, environments with alluvial sediments and specific soil conditions have been associated with higher concentrations even without clear evidence of industrial contamination [1,7]. In Ecuador, it has also been emphasized that, in addition to fertilizers, transfer to the cacao tree is modulated by soil properties and agricultural practices [3]. In this context, additional evidence from agricultural environments indicates that legacy contamination can persist in soils and be redistributed through fine particles and dust, especially where agricultural disturbance, bare soil, and erosion-prone conditions facilitate the movement and deposition of metal-bearing material, thereby creating localized hotspots and heterogeneous exposure patterns [36]. Likewise, nationwide evidence for cacao systems in Ecuador shows that cacao has a high affinity for Cd uptake, with soil-to-bean transfer factors ranging from 0.13 to 12.5 and a median of 1.60, supporting that even moderate soil Cd concentrations can result in comparatively high Cd accumulation in beans [8]. This transfer is not controlled by total soil Cd alone: bean Cd increases with increasing total soil Cd and with decreasing soil pH, organic carbon, and oxalate-extractable manganese, indicating that Cd mobility and sorption processes are critical determinants of soil-to-plant and especially soil-to-bean transfer. In particular, lower pH favors Cd solubility, whereas higher organic carbon and Mn oxyhydroxides can reduce Cd bioavailability through adsorption and retention [8]. In Guayas, the combination of high variability and the presence of outliers indicates that mechanisms are likely not unique; different farms may be dominated by geogenic factors, land-use history, local deposition processes, or soil chemical properties that favor Cd transfer, as well as diffuse anthropogenic inputs, such as those related to phosphate-based fertilization.

Table 5. Economic sustainability indicator scores used to calculate IK by locality.

Economic Sustainability Indicators	Cerecita	Bajada de Chanduy
Production diversification	0.38	0.27
Monthly income	3.75	0.91
Sales diversification	1.38	1.27
Marketing channels	0.00	0.00
Dependence on external inputs	0.00	0.00
Area under cacao	3.00	0.36
Annual production (kg ha−1)	1.63	1.45
Financing sources	0.38	0.00
IK	1.56	0.55

presents the economic sustainability indicators disaggregated by zone, because the two areas differ in their production systems. Producers in Cerecita manage larger farms and are almost exclusively dedicated to CCN-51 cacao production. In contrast, producers in Bajada de Chanduy emerged from a Ministry of Agriculture program six years ago, which provided fine flavor cacao plants to be established in household orchards as a long-term investment. Nevertheless, during the last two years, the increase in prices for dried or wet cacao has represented a significant income for smallholders and has stimulated renewed investments, including pressurized irrigation systems, improved fertilizer sources, and increased fertilization frequency, among the most notable changes in medium and large producers in the area.

Of the 19 cocoa producers surveyed, all exhibited $IK<2$; therefore, according to the methodology of Sarandón and Flores [24] applied in this study, economic sustainability is insufficient in both localities. However, severity differed markedly: Cerecita ($n=8$) showed a mean $IK$ of 1.56, while Bajada de Chanduy ($n=11$) dropped to a mean $IK$ of 0.55, indicating critical economic vulnerability and limited capacity to buffer shocks, invest and sustain income. In parallel, Cd emerged as a cross-cutting environmental pressure, with soil Cd ranging from 0.24 to 1.45 mg kg⁻¹ and bean Cd from 0.53 to 5.01 mg kg⁻¹. Mean soil Cd was practically identical between areas (∼0.69 mg kg⁻¹), but bean Cd showed higher load and variability in Bajada (mean ∼1.74 mg kg⁻¹; maximum 5.01 mg kg⁻¹). This suggests that Cd is not an isolated issue affecting only a few farms but rather a territorial condition that increases market risk and compliance and/or mitigation costs. Thus, the expected economic impact depends not only on Cd levels but also on the economic capacity to respond (captured by $IK$), with Bajada de Chanduy showing the lowest resilience to face commercial restrictions, price discounts, rejections, or the investments needed to reduce soil-to-bean Cd transfer.

The most critical factors, as shown in Figure 6, were the near-null diversification of production. In Cerecita, farmers mainly cultivate cacao, with only some adding one additional crop. In Bajada de Chanduy, cacao cultivation originated from government support; however, because the area is close to Guayaquil, household income largely comes from off-farm jobs in the city, Ecuador’s main economic hub. Consequently, low diversification implies fewer alternatives to offset price penalties or rejections, particularly under elevated Cd levels. Previous assessments using the same framework have shown that the economic dimension often conditions ecological and sociocultural performance due to strong interdependence among dimensions: when the economy is insufficient, management options and technological improvements become constrained [37].

On the other hand, limited access to financing constrains the adoption of mitigation measures, including soil amendments, soil management practices, plantation renovation, varietal selection, and post-harvest improvements. When monthly income is low and unstable, as has occurred during the last two years due to fluctuations in cocoa prices, the impact of Cd becomes even more severe, because any additional cost or commercial discount can push the production system below its economic viability threshold.

Cerecita shows a marked economic gap relative to Bajada de Chanduy, with an IK of 1.6 versus 0.6, indicating greater relative economic capacity and representing the main differentiating factor between the two locations. This contrast is explained primarily by monthly net income (0.9 vs. 0.2) and, to a lesser extent, by economic risk (0.5 vs. 0.3), whereas food self-sufficiency is low or absent in both areas. In the ecological dimension, Cerecita shows a slightly higher index (2.26) than Bajada de Chanduy (1.98), with fertilization standing out as the main difference (0.6 vs. 0.4), while other indicators, such as soil conservation, biodiversity, and phytosanitary management, are similar. In the sociocultural dimension, Bajada de Chanduy slightly exceeds Cerecita (SC of 2.29 vs. 2.15), mainly because of greater integration (0.5 vs. 0.2), although this advantage does not offset its economic constraints (see

Table 6. Ecological, sociocultural, and economic sustainability indicator scores by locality.

Indicator	Cerecita	Bajada de Chanduy
Soil-life conservation (E)	0.34	0.33
Erosion risk (E)	0.56	0.46
Biodiversity management (E)	0.23	0.21
Phytosanitary management (E)	0.54	0.54
Fertilization (E)	0.58	0.42
Basic needs satisfaction (SC)	0.98	0.67
Acceptability (SC)	1.00	1.12
Integration (SC)	0.17	0.50
Food self-sufficiency (K)	0.09	0.07
Net monthly income (K)	0.94	0.23
Economic risk (K)	0.53	0.26

Note: (E) = ecological; (SC) = sociocultural; (K) = economic. Values closer to 4 indicate higher sustainability.

and

Table 7. Sustainability indices by zone and overall mean.

Index	Bajada de Chanduy	Cerecita	Overall
$IK$ (economic)	0.55	1.56	0.96
$IE$ (ecological)	1.98	2.26	2.10
$ISC$ (sociocultural)	2.29	2.15	2.23
$ISG$ (overall)	1.61	1.99	1.76

).

In both localities, ecological indicators ($IE$) showed low and fairly homogeneous levels, such as soil-life conservation (≈$0.3$), biodiversity management (≈$0.2$), and phytosanitary management (≈$0.5$), with similar erosion-risk values (0.6 in Cerecita vs. 0.5 in Bajada de Chanduy) and a more marked difference in fertilization (0.6 vs. 0.4). Within an indicator-based assessment framework, these ecological performances suggest a limited biophysical base to sustain productivity and stability, which tends to translate into greater economic vulnerability [24,38,39]. Consistently, the economic components (K) are very low, especially in Bajada de Chanduy, where net monthly income (0.2) and economic risk (0.3) are lower than in Cerecita (0.9 and 0.5, respectively), and food self-sufficiency is nearly null (0.0 vs. 0.1). Thus, even though Cd constitutes a relevant external pressure for commercialization, the results show that ecological constraints (soil, biodiversity, and nutrition) operate as a “floor” that conditions income generation, input-use efficiency, and the ability to buffer shocks, ultimately affecting $IK$ performance.

Sociocultural indicators ($ISC$) also help explain differences in $IK$: Cerecita shows higher basic-needs satisfaction (1.0) than Bajada de Chanduy (0.7), while acceptability is similar (1.0–1.1) and integration is higher in Bajada (0.5) than in Cerecita (0.2). From sustainability and socio-ecological systems perspectives, better satisfaction of basic needs is often associated with greater room to invest, plan, and sustain management practices, thereby supporting economic viability [24,38]. Conversely, higher integration may reflect social capital and collective-action capacity (e.g., learning, coordination, and access to networks), with potential to reduce risks and transaction costs; however, in Bajada de Chanduy this social asset is not reflected in better economic values, likely because very low monthly income and a weak ecological base constrain the translation of social organization into productive and financial improvements [39,40]. Overall, the results suggest that economic sustainability depends not only on monetary variables but on the interaction between ecological conditions (which sustain production) and sociocultural conditions (which enable or constrain the adoption and persistence of strategies), consistent with the integrative logic of indicator-based methodologies. Finally, the overall sustainability index favors Cerecita (1.99) over Bajada de Chanduy (1.61), consistent with the greater weight of the economic component in the global comparison (see Table 7 and Figure 7).

In conceptual terms, these findings reinforce the relevance of indicator-based frameworks for decision making and intervention prioritization as proposed by Sarandón and also by participatory evaluation approaches such as MESMIS, which recommend integrating attributes of productivity, stability, resilience, adaptability, and equity through indicators tailored to the local context [38]. Moreover, in the sustainability-assessment literature, indicators are described as “information-structuring tools” that make sustainability operational for management and public policy, particularly when territories must be compared and improvement pathways designed [41].

Interviews were conducted anonymously with key actors in the commercialization chain (collectors) and with agroexport companies linked to the National Association of Cacao Exporters and Industrialists of Ecuador (ANECACAO). This framing is pertinent because ANECACAO defines itself as a trade association that groups and represents around 80% of exporters of cacao beans and derivatives and provides technical and statistical information services for the sector [42]. Operationally, collectors reported receiving cacao from multiple areas and classifying it by origin; a recurrent preference was purchasing wet cacao to carry out post-harvest processing under their control, with fermentation commonly performed in jute sacks or concrete boxes. While these post-harvest decisions respond to quality standards and logistical efficiency, they concentrate process control and limit producers’ ability to capture value through differentiation a critical issue for economic sustainability when quality and sanitary risk become market attributes.

Regarding cadmium (Cd), interviewees reported a high level of general awareness of the problem, but also acknowledged irregular analytical management: verification of Cd in incoming cacao is infrequent and, when performed, is often limited to random sampling funded privately and analyzed in external laboratories. This qualitative evidence aligns with regional diagnoses indicating that, among producers and intermediaries, explicit Cd requirements are not always transferred as a purchase condition, whereas the costs of mapping, traceability, blending, and laboratory testing tend to concentrate in cooperatives and exporters [43]. In parallel, Ecuadorian scientific literature has documented that Cd in beans can be high and spatially heterogeneous, associated with both soil properties and agronomic practices, reinforcing the risk of commercial decisions based on incomplete information [3,8].

In terms of foreign trade, considering that Ecuador exports globally, including to the United States, Italy, Japan, Indonesia, Malaysia, and Russia, and that ANECACAO statistics report 471 thousand tonnes exported in 2024 (about 30% to the European Union) [42], interviewees emphasized that for cacao-bean exports there is no systematic requirement for Cd certificates at ports of departure. This perception is consistent with the European regulatory design, where maximum limits apply mainly to finished products (e.g., chocolate and cocoa powder) rather than directly to beans as a raw material, despite the bean Cd conditions compliance in the final product [35,44]. Exporters nevertheless reported origin-based risk-management strategies: they identify areas more likely to have elevated Cd and blend lots to adjust the shipment profile to the target market; this practice is recognized in technical guidance as an industrial response to meet Cd limits, especially when the final destination includes cocoa powder [44,45]. Even so, interviewees did not report recent sanctions for Cd nor explicit incentives for low Cd, suggesting that the price signal remains diffuse and that mitigation costs tend to be absorbed by the segment with greater bargaining power (export/exporting firms) rather than by producers.

From the indicator-based sustainability evaluation framework, this pattern of commercial governance has direct implications for the economic dimension ($IK$) proposed by Sarandón and Flores [24]: Cd functions as an exogenous risk factor that increases uncertainty of market access (economic risk), weakens net income stability, and reduces producers’ planning capacity, especially when control and traceability mechanisms are outside their reach. Looking ahead for Guayas and Santa Elena, the combination of a more demanding international environment in terms of food safety and due diligence, blending as a short term solution, and the absence of systematic national verification mechanisms could deepen territorial segmentation: “suitable” zones capturing premium destinations, and zones with elevated Cd facing implicit discounts, greater dependence on intermediaries, and pressure to adjust practices or genetic materials [43,45]. Critically, the interviews suggest that the main driver of this scenario is not lack of awareness but an incomplete incentive structure: while final markets set limits on processed products, coordination to manage Cd at origin (sampling, traceability, soil management, and quality differentiation) remains fragmented, shifting part of the risk toward producers and straining their economic sustainability.

In this context, it is relevant to note that the chair of ANECACAO’s board (Iván Ontaneda) highlighted that cacao exports would inject “about $5 billion” into Ecuador’s economy, with benefits expected to reach rural areas and producers [42].

4. Conclusions

While this study provides useful insights into Cd distribution and soil–plant relationships in cacao systems, its findings should be interpreted with caution due to the limited sample size (19 farms in total and 23 sampling points across the two localities). Under these conditions, both interpolation-based spatial representations and correlation analyses may be sensitive to outliers, random variation, and local heterogeneity, which limits the extent to which the observed patterns can be generalized beyond the sampled farms. Accordingly, the spatial maps should be regarded as exploratory visualizations of local Cd distribution, and the correlations as preliminary evidence of association rather than definitive proof of underlying processes. Considering the high spatial variability typical of soil contamination, future research based on larger datasets and spatially explicit analytical approaches will be necessary to improve statistical robustness and support broader-scale inference.

The measured Cd concentrations confirm a relevant presence of this element in the cacao production systems of Cerecita and Bajada de Chanduy. Although soil Cd levels were similar between localities, cotyledon Cd reached clearly higher values, indicating substantial transfer to the organ of commercial interest. In this context, most sampled farms showed cotyledon Cd concentrations within or above ranges that could represent a potential commercial concern for compliance in downstream cocoa products destined for demanding international markets. The matrix-specific accumulation pattern was consistent within each locality, supporting the interpretation that Cd accumulation differed among soil, leaves, and cotyledons. The soil–cotyledon relationship was moderate, indicating that total soil Cd alone does not provide a strong prediction of bean Cd. Part of the observed variability may depend on local factors or processes affecting Cd transfer, bioavailability, and mobility, which supports the interpretation of a multi-causal dynamic rather than a simple linear relationship between soil and beans. It should be noted that soil properties known to influence Cd bioavailability and plant uptake, such as pH, organic matter content, and cation exchange capacity, were not measured in this study. Their inclusion in future research would allow a more mechanistic understanding of the soil–plant transfer dynamics observed here and would complement the bibliographic context presented in the Introduction and Discussion.

Cd presence did not appear to be restricted to a single site but rather represented a shared concern at the territorial scale, despite differences in varieties and crop management practices. The fact that 100% of farms had IK < 2 indicates that, under the conditions evaluated, none of the farms reached the threshold of economic sustainability, with Bajada de Chanduy showing the greatest limitations, mainly related to income generation, income stability, and access to investment resources. The ecological (IE) and sociocultural (ISC) dimensions were interconnected with the economic dimension. In the sociocultural domain, the higher degree of integration observed in Bajada de Chanduy did not offset economic constraints, but it may represent an opportunity for collective interventions, such as traceability initiatives, joint purchasing, and shared technical assistance. In addition, awareness of the Cd problem in the study area remains limited and, even where the risk is recognized, knowledge of practical remediation or mitigation alternatives is still insufficient. For this reason, in zones already identified as vulnerable, timely and periodic diagnosis should be encouraged as a basis for early detection and decision-making. This should be accompanied by the implementation of site-specific mitigation strategies, particularly those focused on soil management, and by the promotion of economically accessible technologies that can be realistically adopted by local producers. Overall, the results suggest that Cd should be understood not as an isolated issue, but as part of a broader territorial vulnerability affecting cacao production in the study area.