Study to Determine Levels of Cadmium in Cocoa Crops Applied to Inland Areas of Peru: “The Case of the Campo Verde-Honoria Tournavista Corridor”

: The presence of cadmium (Cd) in cocoa crops is currently a serious problem for farmers and producers in various regions of South America. Because its exports of cocoa and derivatives to European markets are threatened by possible signs of contamination in cocoa beans for export. Territories with a low organic component predated and exploited by illegal logging, burning and the intensity of unsustainable land use is common in large Amazonian areas in countries of the region. These factors were incorporated in statistical analysis in order to relate them to the contents of Cd in soil, leaves and beans in the study areas located in Peru. Such as the Campo Verde-Honoria-Tournavista corridor (Ucayali Region and Huanuco Region). Cadmium concentrations were determined using an atomic absorption spectrophotometer. As a consequence of this study, we determined and concluded that the observed difference in distribution of Cd contents by sectors can be explained by previous land use and age of cocoa crop. Indeed, the average content of Cd in soil in all cocoa growing areas is higher than the standard established by the Peruvian Ministry of the Environment (MINAM). However, when the measurements obtained in previously predated and exploited sectors are not considered, the Hotelling’s T 2 simultaneous 90% conﬁdence interval contains the value of the Peruvian standard 1.4 mg/kg. Therefore, with this information we prepare a geochemical Cd map in soils for the study area, which will help cocoa producers to identify areas that exceed the allowed Cd values. In this way, we can carry out in the future a mitigation plan for areas with Cd problems, which allows to reduce their content with major challenges to sustainable agriculture and rural development.


Introduction
Currently there is great concern throughout the cocoa commercial chain worldwide (producers, exporters, importers and consumers). There is due to the presence of Cd in cocoa beans and direct transmission to its derivatives for human consumption. Cadmium is part of the list of heavy metals that can accumulate in food and that reach people through their consumption, appearing in food naturally or due to contamination.
The cocoa bean is a product of great commercial relevance for Peru, it is the fifth most important product at the national level, both by area and by number of producers. In 2017, cocoa exports and its preparations in Peru exceeded 235 million dollars, which represent around 4% of total agricultural exports [29].
Then in [30], they proposed a list of factors that regulate the processes of bioaccumulation of Cd by plants, adapted to our country. This is shown in Table 1. Table 1. Factors related to cadmium absorption processes [30].

Factors
Effect of Absorption of Cadmium by Plants

Edaphic Factors
pH The absorption process increases at a lower pH, that is, it is greater in acidic soils.
Soil Salinity Cadmium absorption increases with soil salinity.
Cadmium Concentration Absorption increases with increasing cadmium concentration.

Micronutrients
The absorption process is variable according to the nature and quantity of micronutrients.

Macronutrients
The absorption process is variable according to the nature and quantity of macronutrients.

Temperature
The absorption increases with increasing temperature.

Arable Species
The absorption process of heavy metals decreases according to the following trend: Vegetables> beans> cereals> fruits.
Plant Tissue Leaf> bean> edible fruits and roots.

Sheet Age
Old leaves> young leaves.
Despite there being several relevant publications available on As, Cd, Pb and Hg in cocoa beans and cocoa-based foods, there are still gaps that encourage research on these issues [31].
In this work, considering that there is no information about Cd in the area located in the Campo Verde-Honoria-Tournavista corridor, which is located between the Ucayali and Huanuco regions of Peru, we need to carry out the analysis of the soil, leaves and cocoa beans in this corridor, to determine Cd values. In order to carry out this analysis, we collect the samples from the study area and then these were sent to the laboratory of the Faculty of Geological Mining and Metallurgical located at the National University of Engineering (FIGMM-UNI) in the city of Lima, for chemical analysis. An atomic absorption spectrophotometer was used to obtain the concentrations of this metal in soil, leaves and beans to later perform a descriptive, comparative and exploratory data analysis. Finally, the geochemical map was prepared that shows the distribution of Cd on a local scale, based on the analysis performed for Cd from the samples that were collected from the study area.
We believe that the study provided by this work can be applicable to any region or country that wishes to analyze Cd levels in their crops to reduce their content with major challenges to sustainable agriculture and rural development.

Study Area Location
The work area is located in Honoria district, approximately 63.3 km away from Pucallpa district, leaving the city on a paved road that connects to the Central Highway for 45 min, you reach the district of Campo Verde along the track 27 km from Pucallpa, in which there is a detour of the trail to Honoria district. The tour of the trail takes approximately 1 hour and 45 min, reaching the Fundo Cristina.

Study Area
The study was carried out in cocoa plantations located in The Honoria Corridor of 12 hectares installed, but currently, 8 are in production. This delimitation was made with the aid of the PHANTOM 4 PRO drone. Figure 1 shows the limits of the area where all our work will be carried out.
The work area is made up of 6 zones, divided according to the age of the cocoa plant, in that area there are the following varieties of cocoa: CCN51 (cocoa clone) and Creole-aromatic (ICS1, ICS6, ICS95, ICS39, TSH565, Huallaga 59 and Catongo). In addition, maize exploitation areas were identified, prior to the cocoa plantations.

Soil and Cocoa Fruit Sampling
For soil sampling, a grid was made with a separation of 50 m east and 50 m north, the grid plane is shown in Figure 1, where a sample for analysis is collected for each point generated. The soil samples used for our study were collected at depths of 30 cm, as indicated in the soil sampling guide, for agricultural soil [32]. In our work, 72 samples were collected equally distributed between soils, beans and leaves of approximately 1 kg of each. The samples were conditioned in Ziploc bags and sent to FIGMM-UNI analysis laboratory and were: air-dried, ground, sieved and stored at room temperature. To achieve this, approximately 50 cm deep test pits were made on average for each point of the mesh that has been designed for the work area, this is to get an easy sample taken, we have to leave a space major to required (30 cm) to not contaminate the sample with the instrument used or some waste or superficial rubbish (branches, stones, trash, among others) and then take the sample from the soil wall (fresh sample) and free of added contaminants during the sample taken, generally, when exist this waste, it is suggested to remove the first centimeters of surface outcrop in a circular area of 15 cm radius [33].
From the test pit, a portion of soil is separated from a wall of the same test pit, considering that the sample should be a portion of soil that has not been in contact with the tool that was used to extract this sample. In addition, electrical conductivity and pH in soil were measured using the norm NOM-021-RECNAT-2000 [34], the sample must be passed through a 2 mm sieve then in a 2:1 solution with distilled water, these physico-chemical parameters were measured using a Thermo Scientific Orion Star A320 series multiparameter.
To collect the fruit, a machete was used to cut the union of this with plant with great care, in order not to contaminate the sample. Then, the fruit was taken with sterile polyethylene gloves to store it in Ziploc bags, assign its respective coding and finally, be sent to the FIGMM analysis laboratory at UNI located in Lima-Peru.

Sample Processing
The main operations to be carried out with the soil and cocoa fruit samples were: milling, quartering and sieving. In these operations, we have used various electronic equipment that minimize contamination of cocoa with other components.
With the soil samples arrived at the laboratory, the samples were first dried, which was carried out in a drying oven at a temperature of 105 • C. After the sample is dried, the grinding operation was carried out, and a Cole-Parmer brand jar mill was used, which contains burundum material as the grinding medium, ideal for wet or dry grinding. This equipment provided us with a homogeneous grinding, which allowed us to bring the sample to a desired grain size of 75 microns. In the case of cocoa fruits, the operations are quite similar, first the bean was separated manually and to reduce the size for chemical treatment, a grinding was used similar to the soils, using a IKA brand small mill, model A11 b. Then, sieving and quartering were carried out with the portable 1-column electric sieve. This equipment allowed us to bring the fruits up to a desired granulometry of 212 microns.

Chemical Analysis of Samples
Prior to the chemical analysis of the samples obtained, the samples were digested with acids and then solutions were prepared from the soil, bean and cocoa leaves samples, to achieve this, independently, the following was carried out: • Digestion of soil samples. Heavy metal content was determined by the Environmental Protection Agency [35]. • Digestion of samples of cocoa beans and leaves. Heavy metal content was determined using the method mentioned by Isaac and Johnson [36].

Atomic Absorption Analysis
The analysis method for the identification of heavy metals used was flame Atomic Absorption Spectrophotometer (AAS), brand GBC Scientific Equipment, model XplorAA. Regarding the calibration and optimization of the burner height and the gas flow with each element, the standards recommended by the manufacturer were used, such as the 3.5 mg/L copper standard. Subsequently, the calibration curves for each of the three heavy metals were evaluated, observing linearity in the responses and high coefficients of determination 99%. Then, the measurement of the heavy metals contained in the samples was carried out.
Regarding the precision and reliability criteria, the repeatability was evaluated, by means of dual replications of each one of the samples and every 10 measurements were introduced: blank samples, standards and samples enriched with standards. Observing reproducibility, repeatability and reliability in the data obtained.

Results and Discussion
Results of these concentrations of metals in soils such as Cd, Pb, and Zn are shown in Table 2. These data were complemented with the hydrogen potential (pH) and electrical conductivity (CE) measurement. With these obtained results, the statistical treatment of the data is carried out. With the measurements taken at the Fundo Cristina, the geochemical map of Cd soils for the study area was generated using the interpolation method called kriging, for this the ArcGIS software [37] was used, this map is shown in Figure 2. The data obtained confirmed that in the case of soils, the average Cd concentration of 1.63 mg/kg is above the Environmental Quality Standards (ECA) for the soil established by MINAM, the value of which is 1.4 mg/kg [38]. The map shows the distribution of the element Cd based on its concentration in the area. To carry out the statistical analysis, the R software [39] was used. The work zone was divided into five sectors (zone 6 does not contain crops) based on an administrative criteria of field management that includes homogeneous geographical characteristics, accessibility and type of crops. Maize exploitation zones were identified prior to cocoa plantations, sandy or clayey zones, as well as CCN51 and Creole-aromatic cultivation zones as observed in Figure 3. It is important to highlight that zone E1 is a predated and exploited zone with soils with a low organic component where there is only pasture. This type of territories predated by illegal logging, burning and the intensity of unsustainable land use is common in large Amazonian areas in countries of the region. These factors will be incorporated in statistical analysis in order to relate them to the contents of heavy metals in soils, leaves and beans.  Table 3 shows the statistical summary on content of heavy metals (Cd, Pb and Zn) according to the variety of cocoa in areas with cocoa crops. The acidity (pH) and electrical conductivity (CE) levels of these same soils are also included. The data confirm that distribution of Cd in plant tissues and in plantation soils shows the following trend, both at the level of the entire productive unit and at the level of variety: C beans < C soils < C leaves These results confirm the findings of [20,25,30], they conclude that average Cd concentration varies in different parts of the plant. In [34], mentioned that between 70 % and 80 % of the Cd absorbed by the cocoa plantations is retained by roots and the rest moves to other parts of plant tissues, being bioaccumulated, depending on the location, on stems, leaves and lower proportion in shell and beans of cocoa.
Regarding the Cd content in cocoa beans, the data confirm that the mean Cd concentration is 0.96 mg/kg. These values are above the international standards of the European Union, which as of 1 January 2019, regulated that the cocoa powder content has a maximum concentration of 0.6 mg/kg [40].
Research by [41,42], confirms this hypothesis that plant tissues bioaccumulate higher concentrations than storage tissues, such as cocoa shells and cocoa beans, which it allows to infer that presence of physiological barriers prevents the Cd bioaccumulation process in cocoa shells and in cocoa beans.
It is observed that highest levels are registered in beans of Creole-aromatic varieties, which is directly related to highest levels of Cd concentration in soil. On the contrary, although Cd levels are lower in CCN51, the variability with respect to the average is higher in it. Therefore, these results will be complemented with an analysis of distribution of data. Figure 4 shows distribution of contents of Cd in soil, in areas with cocoa crops of productive unit (a), according to sectors (b) and previous exploitation of field in maize crops (c).  The empirical distribution of measurements of Cd in soil in cocoa cultivated area, corresponding to Fundo Cristina (FC), shows a symmetric distribution but with extreme values, which can explain the increase in average content of Cd in varieties Creole-aromatic. It should be noted that more than 50% of values are below the maximum allowed of 1.4 mg/kg established in sampling guidelines for determination of Cd levels in soils, leaves, beans and products derived from cocoa [38, 40,43]. As a result, it is observed that average value is less than median value (Figure 4a).
By blocking the Cd measurements according to sectors (Figure 4b), it is observed that the distribution of Cd in soil in these cocoa growing areas is different. The great difference between sectors E1 and E2 is notable. The first one shows a high average Cd content that exceeds the 1.4 mg/kg allowable and a slightly asymmetric distribution that affects the average value, placing it above the median. On the contrary, in the second case, although the average content of cadmium is very acceptable, the dispersion is high, observing an extreme value close to 3 mg/kg. Sectors E3, E4 and E5 show very homogeneous data distributions and centered within permissible limits.
It is possible that the observed difference in distribution of Cd contents by sectors can be explained by previous land use and age of cocoa crop in the aforementioned sectors. To obtain evidence, we use graph (Figure 4c) that shows the distribution of contents of Cd according to previous exploitation of the land. The graph presents three cases: first, land that has been planted with maize for 10 years since 1998 (MAIZE_10); second, land that since 2001 maize was sown for 7 years (MAIZE_7) and those that were pasture for 10 years (PASTURE_10) before 2008. In this regard, it is observed that the content of Cd is higher in land since about 1998, which were pastures land. These territories were poor in organic matter, because they were previously exploited.
These lands correspond to sector E1. A similar result is observed in lands that had maize planting since 1998 and 2001 that correspond to sector E3, E4 and E5. Although the average is low, there are extreme values which correspond to the high values interpolated in geochemical map.
Indeed, as shown in Table 4, the average content of Cd in soil in all cocoa growing areas is higher than the (ECA) for the soil established by MINAM [38]. However, when the measurements obtained in E1 are not considered, the simultaneous 90% confidence interval contains the value of the ECA 1.4 mg/kg. A similar result is observed in recovered areas with cocoa crops, where the simultaneous confidence intervals also contain the Peruvian standard [38]. Table 4. Average content of Cd in soil according to total productive unit and intensity of land use (mg/kg). It should be noted that these lands that were previously cultivated with maize, have already been recovered due to fertilizer (island guano) applied to cocoa crops that have existed for 8 years, which has not happened yet with the pasture that is still in that process.

Sample Mean Cd Soils Sample Size
In Figure 5, the hierarchical correlation matrix between the 11 variables analyzed is shown, considering data from the entire productive unit (Figure 5a), blocking variety CCN51 (Figure 5b) and Creole-aromatic (Figure 5c). A priori, it was determined that the correlation matrix should be constructed on the basis of three hierarchical groupings because three heavy metals were analyzed.
When analyzing the data corresponding to the entire productive unit (Figure 5a), it is verified that the three measurements of Cd (soils, leaves and beans) are significantly correlated in the same group, although these contribute less to variability than Zn and Pb respectively. Furthermore, of the three possible combinations (two to two) it is observed that the lowest is Cd soil with Cd bean. At the other extreme, we have Cd with Zn, where in all cases, they are negatively correlated. Especially, Cd soil with Zn bean where correlation is extreme negative.
Unlike what was observed in the correlation at the entire farm (Figure 5a), the correlation structure in the variety CCN51 shows that the contribution of Pb in the total variation of the data is greater than Zn. In both cases, the variability observed in the Cd contents is the one that least contributes to the total variability of the data. The correlations within groups are positive, which indicates the positive correlation in each one of the heavy metals analyzed in relation to the soil, bean and leaves. However, the same does not happen in case of aromatic and creole varieties, as shown in Figure 5c. Unlike Figure 5a, Cd is added to the Pb contribution observed in Figure 5b. In this case, amounts of Cd are the ones that explain the highest proportion of total variability and maintain positive correlation within groups. Additionally, the negative correlation of grain Zn with the content of Cd in soil, bean and leaves is highlighted in Creole and aromatic variety. Grain lead now correlates positively and significantly with Cd contents, which in the general case did not. It is important to mention that CCN51 variety is found in clay soils unlike the Creole and aromatic varieties where soils are not only clayey but also sandy. In the latter case, land has been pasture, which indicates extreme exploitation of land. Figure 6 shows a representation of observations and variables analyzed by a Biplot, of main components, at 51.7% of total inertia [44]. Variable targeting and position of sample observations were validated by a SQ Biplot for diagnosis [45,46]. Horizontal axis is mainly represented by Cd variables in soil, leaves and bean, therefore it will represent Cd dimension. Vertical axis that is represented by electrical conductivity (CE_aprox) and the pH (pH_aprox) will be the CE_pH axis. Regarding observations, it is verified that they are contained in a Gaussian interval of 90% with the exception of record 4, which is completely defined by the CE_pH axis. In Figure 6, it is possible to distinguish definition of two extreme groupings along the horizontal axis, which would be indicating that the cadmium dimension is an important discrimination factor. At the extreme left are observations collected in sector E1, which represents the highest values in soil Cd and were pasture areas, currently, with Creole-aromatic crops in the process of recovery. In the extreme right, formed by observations of sectors E2 and E3, with lower levels in Cd, formed by Creole-aromatic crops and CCN51 where maize was cultivated for more than 7 years, but currently they are recovered territories.
Regarding the CE_pH dimension, it can be seen that the observations corresponding to sector E5, whose trees are 10 years old, are well represented on this axis. The pH values are above the average of 6.49 and 26.11 uS/cm in electrical conductivity (seen in Table 2). The observations of the other sectors are distributed in the vicinity of the mean value of the CE_pH axis with a slight bias towards values below the average. This bias is more evident in the observations taken in sectors E2 and E3 where there are Creole-aromatic varieties and CCN51. This dispersion of the data around the mean pH and CE values indicates a predisposition of the soil to respond positively in the medium term to any remediation measure aimed at reducing the uptake of Cd in cocoa crops.

Conclusions
The data collected allowed obtaining statistical information on the contents of Cd, Pb, and Zn in soil, leaves and bean, in addition to CE and pH in the study area. Factors such as previous exploitation of soil, age and variety of cocoa trees have also been included. Although the research is descriptive, it was necessary to design data collection according to the updated progress of the research. The study has confirmed the relationship and correlation established between factors that regulate the processes of Cd bioaccumulation by plants. However, do not confirm that variety or age of tree clarify about Cd uptake.
Although it is evident that over-exploited soils until their exhaustion tend to establish undesirable Cd absorption conditions, and, the fertilization of the soil with island guano for a period of not less than 5 years seems to reverse the phenomenon. Hence, statistical information confirms that the oldest sector is not defined in Cd dimension, on the contrary, it strengthens the conditions of CE and pH of soil for the balance of nutrients in the soil.
The data obtained in this study confirmed that in the case of beans, the average Cd concentration is 0.96 mg/kg, this value is above of the international standards of the European Union whose maximum concentration is 0.6 mg/kg. In the case of soils, the average Cd concentration of 1.63 mg/kg, is above the ECA whose value is 1.4 mg/kg, therefore, it is necessary to adopt measures to mitigate the aforementioned values, in order to comply with the requirements. In this work, the previous knowledge of the aspects of field management, as well as the current knowledge of the cocoa crops in the "Fundo Cristina", was essential to guide the research, and was useful to take actions that allow compliance with the requirements imposed.
Therefore, we can conclude that it is essential to carry out future research aimed at proposing measures that make it possible to remedy the current situation in the study area, such as conducting studies to approach Cd mobility and avoid the uptake of Cd in the cacao tree.