Study area and plant material

The study was conducted with samples of oil palm plants collected in 1) Tapachula de Córdoba (14°45´29.31” N, 92°32´16.35” W; altitude 17 m), 2) Huehuetán (14°58´52.20” N, 92°28´30.35” W; altitude 18 m), and 3) Villa Comaltitlán (15°06´54.70” N, 92°37´42.94” W; altitude 10 m) Soconusco region, Chiapas.

Shoot sampling

The aboveground biomass was collected from 12 plants of more than 12 years old. Plants were cut with a chainsaw (Truper, Mod. Mot-4518; China) where the stipe begins, and the stem separates from the leaves. The stems were measured for total height (Ht) and diameter and breast height (DBH: 1.30 m) with a metric measuring tape (Truper, model Tp20ME de 20 m; China). Each stem was divided into three or more sections to determine fresh weight in situ with an electronic platform scale (OHAUS; Mod. ES50L AM, 50 kg capacity; USA). Volume of biomass was calculated with the truncated cone geometric formula. The lower and upper diameters of each section were measured. Sections 10 cm thick were cut with the chainsaw and dry weight was determined in three subsamples. All the sections were obtained from the height where the lower stem diameter was determined. Leaf biomass was determined in four leaves, one per cardinal point, with an electronic platform scale. The stem and leaf subsamples were weighed immediately after cutting to determine fresh weight, placed in plastic bags and taken to the biotechnology laboratory of the Facultad de Ciencias Agrícolas, Campus IV, Huehuetán, Chiapas. The samples were dried to constant weight in a circulating air oven (VWR international, Mod. 1390FM, Sheldon Manufacturing, Inc. USA) for 96 h at 60 ºC. The ratio dry weight:fresh weight was used as the estimator of the amount of biomass of the stem and branches in samples and subsamples using the technique described by ^{Díaz-Franco et al. (2007}) and ^{Acosta et al. (2002}).

Root sampling

Carbon in the palm roots and C stored in the soil was quantified by humification of the biomass. The latter was obtained in 20 samples representative of the approximately 4.5 m diameter drip area at the stem base with a stainless steel drill (no brand name, 10 cm diameter, 50 cm long). Carbon stored in the soil (CSS; Mg ha^-1) was quantified in 20 soil subsamples per plant using soil content (% CS), bulk density (BD) and depth (D) of sampling (1 to 50 cm), with the following relation: CSS = % CS*BD*D (^{Ávila et al., 2001}). For this estimation, the technique used was that reported by ^{Leblanc et al. (2006)}, modified because soil BD was determined at 3,927 cm³ in a sample (extracted with a stainless steel drill). Of each sample, roots were separated; CSS and C in roots were determined with the modified wet combustion method of Walkley and Black (^{Kaurichev et al., 1984}; ^{Schlichting et al., 1995}).

Total biomass (Y) per plant was the sum of biomass from stem, leaves and root. In stem, it was determined with the formula Y = Volume (V) x Density (D) because the stem shape is conical and irregular, and the diameter of the base is larger than the upper diameter. The volume of each section and total volume was calculated with the above equation. ^{Brown (2001)} indicated that the biomass of the leaves accounts for 65 % of the stem weight. This relationship was used to estimate sample biomass and the total leaf biomass. To corroborate this value, V and D of the five leaves of each plant were determined after they were completely shredded and weighed. With this information, values of the leaves per plant were calculated.

Determination of specific density

Specific density (g cm^-3) of stems, roots, soil, subsamples of stem sections and leaves was determined with the ratio dry weight:volume of fresh tissue. Dry weight was obtained with the methodology described above, and the volume of fresh tissue was determined by the method of water displacement. Specific density of sections, leaves and roots were compared with the Tukey test (^{SAS Institute Inc., 2009}). The total sum of stem (Y_f), canopy (Y_c), and root (Y_r) biomass was the total biomass (Y), Y = Y_f + Y_c +Y_r of each plant (^{Acosta et al., 2002}).

Determination of the allometric model

Average biomass of plants of the sites were analyzed together with an equation that would permit interpretation of plant development. To this end, a simple regression test was conducted following ^{Aquino-Ramírez et al. (2015)}. A linear regression analysis was included with indicator variables, in this, case biomass, amount of C and DBH. With the data of total biomass (Y) and total height (X), the linear equation was fit from the potential expression. The model was fit by minimum squares, and the order to origin (b) and slope (k) were obtained according to ^{Avendaño et al. (2009)} and ^{Acosta et al. (2002)}. The resulting model was Y = bX ^k (potential equation), where Y is total biomass, X is the diameter at breast height, and b and k are the function parameters. The equation was fit to a linear expression by transformation to natural logarithm of Y and X; thus, the expression of the equation in its linear form was ln (Y) = ln (b) + k ln (X) (Eq. 2) where b and k are the parameters to be estimated.

Determination of carbon content

The amount of C in stem, leaf and root biomass was determined by the modified Walkley and Black wet combustion method, based on incomplete oxidation of C, with an oxidizing mixture of K₂Cr₂O₇, H₂SO₄ and heat by aqueous acid dilution. Boiling between 110 and 130 ºC depends on initial temperature, organic matter content and size of the reaction recipient. The amount of oxidizing agent consumed in the reaction was determined by spectrophotometry (spectrophotometer METASH, Mod. UV-6000, China), by measuring initial intensity of green Cr⁺³ and yellow K₂Cr₂O₇ ions, and a sucrose calibration curve. In the reaction, the Cr⁺³ ions are proportional to the amount of oxidized C, according to the reaction: C₆H₁₂O₆ + 4 K₂Cr₂O₇ + 16 H₂SO₄ → 4 K₂SO₄ + 4 Cr₂ (SO₄)₃ + 6 CO₂ + 22 H₂O.

Five mL of K₂Cr₂O₇ and 10 mL sulfuric acid were added to 0.02 g dry sample; it was left to cool for 30 min and 35 mL distilled water was added. The mixture was shaken after each addition and left to stand overnight. Absorbance was measured at 590 nm. The total content of each sample was multiplied by the percentage of C obtained by spectrophotometry and by the sample biomass; the total C content per plant would be the sum of the total C content in stem, branches and roots. To obtain the allometric equation and estimate C, the allometric model that relates the content of C to plant height was used (Swift, 2000; ^{Avendaño et al., 2009}). Descriptive statistical analysis and correlation tests (α = 0.05) were applied to the data.

Biomass estimation

The general average of stem, leaf and root biomass was 1104.34 ± 87.85 kg. Plants from Huehuetán had more biomass than plants from Tapachula and Villa Comaltitlán. The plants of the two latter sites had higher density (Table 1). These values were not reflected in total biomass, since average DBH and height were larger in Huehuetán than in other sites. According to the information provided by growers, variability of the field data is due to differences in: 1) agronomic management of the crop, such as the application of essential macro and micro minerals; 2) water supply: drought in high areas and flooding in low areas and in areas close to the sea; and 3) environmental conditions, such as temperature and precipitation. This coincides with ^{Corley and Gray (1982)}. All the factors mentioned are determinant for the crop since maximum development is achieved after 6 to 12 years, although the plant can continue growing at a rate of 35 to 75 cm per year, depending on environmental conditions (^{Corley and Gray, 1982}; ^{Cayón, 1999}). In our study, the plants evaluated were about 12-years-old.

The linear allometric model was Y = 3.571X +816.86 (R² = 0.695) and Y = 98.349X + 737.41 (R² = 0.577) to determine biomass based on DBH and plant height, respectively (Figures 1A and 1B). The coefficients of determination for both variables reduce reliability of the equations to determine biomass in 12-year-old palm plants, but they are not ruled out. We suggest increasing the number of replications and older plants to estimate applicable allometric equations (^{Thenkabail et al., 2004}: ^{Leblanc et al., 2006}). However, for growers, it is difficult to provide plants for sampling. For this reason, the number of plants for destructive sampling was reduced in the Soconusco region. The proportion of root biomass and shoot biomass in this study was 0.06, which is notably low relative to other studies.

Figure 1 Dispersion of the measured values and regression line generated with the linear model to estimate biomass based on diameter at breast height (DAP) and total height of E. guineensis Jacq. 

The allometric equations and similar transformations were reported in Abies religiosa (Kunth) Schltdl. & Cham. (^{Avendaño et al., 2009}) and other tree species, such as Cupania dentata DC., Alchornea latifolia Sw. and Inga punctata Willd (^{Aquino-Ramírez et al., 2015}). Therefore, the results of our study still do not corroborate the fit of the allometric model to determine biomass and sequestered C if DBH is known. However, by using plant height the model could be more precise. For this reason, it is necessary to conduct more studies with more samples to approach the acceptable models used in the literature with other species.

The relationship between DBH and plant height was positive (Figure 2). The linear equation was height (m) = 0.029(DBH) + 1.439 and R² = 0.762. This equation can thus be used to estimate plant height (m) with data on DBH (cm) of oil palm in the wet tropical region of Chiapas.

Figure 2 Relationship between palm height (m) and diameter at breast height (DAP, cm). Dispersion of mean values for estimation oil palm (E. guineensis Jacq.) and regression line generated with the model. 

Allometric equation to determine carbon stored in oil palm plants

The amount of C can depend on multiple factors. ^{Leblanc et al. (2006)} pointed out that among the main factors are nutrient management, plant age, and climatic factors such as precipitation, and type of soil. In our study, regardless of the variability of the determined values, the quantities of C were not different among sites nor among plants (p > 0.05). Oil palms about 12 years old at a density of 143 plants ha^-1 in the Soconusco region has the potential to store an average of 76.18 ± 0.047 Mg C ha^-1 and the annual rate of fixation is 6.53 Mg C ha^-1. The values per hectare and the fixation rate were higher than the 42.87 Mg C ha^-1 in plants reported by ^{Khalid et al. (2000)} and 31.21 Mg C ha^-1 in 12-year-old plants reported by Rodríguez et al. (2000)⁶. In contrast, ^{Thenkabail et al. (2004)} found 14.75 and 14.94 Mg C ha^-1 using satellite images (IKONOS Satellite Images), but this technique differs from that applied in our study. ^{Pulhin et al. (2014)} used a system of destructive sampling like that of our study and reported 55.0 Mg C ha^-1, a C fixing potential of 6.1 Mg C ha^-1 year^-1, similar to that found in our study.

With the total C content (stem, leaves and root) of the 12 sampled plants, allometric equations were generated to estimate the amount of stored C relative to plant height since, when DBH was used, the coefficient of determination (R²) was not acceptable. The equation with the best fit was the linear model (y = β₀ + β₁x) (Figure 3). The small number of field samples could have determined the low coefficient of determination for the linear equation between the amount of C and plant height; this improved when the evaluation was between the amount of C and DBH.

Figure 3 Dispersion of observed values for estimation of C in oil palm plants (E. guineensis Jacq.) and the generalized regression line with the linearized model. 

In Central America and Asia, cultivation of oil palm has the potential to sequester C, with values of 96.02 and 103.35 Mg C ha^-1, in the 0 to 30 cm and 0 to 50 cm soil profiles (^{Leblanc et al., 2006}; ^{Ibrahim et al., 2007}). These quantities surpass those we registered in the root fraction and soil in our study. ^{Leblanc et al. (2006)} and ^{Ibrahim et al. (2007)} reported sequester of 22.68 Mg C ha^-1 in shoot biomass, lower than the average of C determined in the thee municipalities in the Soconusco region. Thus, the C stock in oil palm could increase with the interactions of factors, such as favorable geographic location for this crop, annual precipitation of 1800 to 2200 mm, temperatures of 25 to 28 ºC, suitable soils (site 1, Cambisols and Acrisols; site 2, Cambisols and Solonchak; site 3, Feozen, Cambison and Solonchak), which are deeper and salty because of proximity to the Pacific coast (INEGI, 2010). Specifically, agronomic management of the plantation with applications of mixtures of micro and macro nutrients necessary for adequate crop development could be determinant in carbon sequester.

In the literature, quantities from 18.2 to 57.3 Mg C ha^-1 sequestered by sugarcane agricultural production systems were reported in Saccharum oficinarum L. Variability is determined by agroclimatic factors and soil quality (^{Aryal et al., 2014}). In other agricultural systems in Chiapas C sequester (Mg C ha^-1) was estimated at 116.8 in secondary shrub, 46.7 in plantations of live fences, 116.8 in livestock grazing pastures, 130.2 in coffee plantations and fences and 180.8 in milpas with paths. The highest sequester was recorded in the trees established in each of these systems and recycling was through wood formation (^{Alegre et al., 2004}; ^{Lapeyre et al., 2004}; ^{Díaz-Franco et al., 2007}). In annual crops, the sequester rate is 0.1 to 0.8 Mg C ha^-1 for maize-bean systems. These systems generate higher emissions of greenhouse effect gases (GEG) when agrochemicals are used and when harvest residues are burned. In grasslands used for livestock grazing C sequester is 0.1 to 0.2 Mg de C ha^-1; it is 4 to 8 Mg C ha^-1 in reforested areas, and 0.2 to 3.1 Mg C ha^-1 in agrosilvopastoral systems (crops plus poaceae).

Although methods, protocols and management systems do exist, it is necessary to standardize the processes to determine and calculate emissions and C sequester to have real, reproducible and comparable results and to follow the C footprint in the crop. The method in this study can be reliable for those estimations in E. guineensis in Chiapas, as it has been applied in other forest species (^{Castilla, 2004}; ^{Avendaño et al., 2009}).