Organic farming enhances soil carbon and nitrogen dynamics in oil palm crops from Southeast Amazon

ABSTRACT

The expansion of oil palm (Elaeis guineensis Jacq, Arecaceae) crops threatens tropical rainforests. It negatively impacts a series of ecosystem services and functions, including carbon (C) sequestration and dynamics, as well as nutrient cycling. Such negative impacts have pressured companies to adopt conservationist practices in palm oil production. And yet the conversion from conventional to organic farming has gained space in the last decade, studies assessing the effects of organic oil palm crops on ecosystem functioning are still scarce. Here, we assessed how alternative farming practices affect organic matter dynamics in oil palm crops. We compared oil palm crops under conventional (CP) and organic (OP) farming in Southeast Amazon. We also sampled lowland dense ombrophilous forest (floresta densa de terra firme, FT) as reference. Soils were sampled at 0–10, 10–20, and 20–30 cm depth intervals to determine soil physical–chemical properties and C and nitrogen (N) concentrations and stocks. The highest soil C and N concentrations were found at 0–10 cm interval in CP and OP. We detected no variation in soil C and N stocks within depth intervals in FT and CP, while OP had higher soil C and N stocks at the 0–10 cm interval. When comparing OP and CP crops, soil C concentrations and stocks did not vary within zones or depth intervals. All OP zones had higher soil N concentrations and stocks than their conventional counterparts, and we found a variation within depth intervals. Our results show that organic farming has positively influenced organic matter dynamics. Organic oil palm crops preserved and even increased C and N sequestration.

KEYWORDS:

• Carbon
• ecosystem services
• land-use change
• nitrogen
• tropical forest

1. Introduction

Palm oil comes from the fruit of oil palm trees (Elaeis guineensis Jacq, Arecaceae) and is the most widely used vegetable oil in the world; United States Department of Agriculture 2021). It is used by several industrial sectors (Dislich et al. 2017), besides being an alternative biodiesel and bioenergy feedstock also an alternative biodiesel and bioenergy feedstock (Kaniapan et al. 2021).

The rising global demand, coupled with tree high productivity, easy establishment, and low costs, has led this industry to its rapid expansion (Sabine, Martin Persson, and Kastner 2015; Ordway, Asner, and Lambin 2017; Vijay et al. 2016). Globally, oil palm crops more than doubled between 1990 and 2020, going from 4.20 million to 9.34 million ha (United States Department of Agriculture 2021). Oil palm expansion became one of the most important drivers of deforestation in the humid tropics (Houghton 2012; Vijay et al. 2016; Ordway et al. 2019), especially in Southeast Asia (United States Department of Agriculture 2021).

More recently, the palm oil industry has also advanced in Equatorial Africa and South America (Furumo and Mitchell Aide 2017; Laurance, Sayer, and Cassman 2014; Ordway, Asner, and Lambin 2017; Ordway et al. 2019; Pacheco 2012; Vijay et al. 2016). In the latter, the combination of climate and soil makes the Amazon region the suitable site for oil palm crops (Butler and Laurance 2009; Frazão et al. 2013). Brazilian palm oil production is concentrated in Northeast Pará (Silva et al. 2018), state nestled in one of the most threatened areas of the Amazon (Instituto do Homem e Meio Ambiente da Amazônia 2021; Instituto Nacional de Pesquisas Espaciais 2021).

The replacement of forested areas by oil palm crops threatens not only the embodied species (Lenore et al. 2013; Fitzherbert et al. 2008; Philips, Newbold, and Purvis 2017; Teuscher et al. 2016; Wilcove and Pin Koh 2010) but also leads to severe soil degradation and erosion (Afandi, A et al. 2017; Guillaume et al. 2016; Silva et al. 2018). It negatively impacts tropical landscapes and several other ecosystem services and functions and the well-being of the traditional populations (Ivan et al. 2018; Dislich et al. 2017; Duru et al. 2015).

One of the functions affected is the maintenance of biogeochemical cycles, including carbon (C) sequestration and dynamics and nutrient cycling (Dislich et al. 2017). There is a drastic reduction of aboveground biomass and C storage (Guillaume, Damris, and Kuzyakov 2015; Quezada et al. 2019), coupled with increased greenhouse gas (GHG) emissions, the primary cause of global warming (Cooper et al. 2020; Germer and Sauerborn 2008).

Short- and long-term modifications in soil nutrient status are also observed. Tropical forests, hitherto N-rich systems with high N availability and cycling rates (Hedin et al. 2009), undergo a cycling slowdown (Allen et al. 2015), with decreases in N-pools and transformation rates (Arnold, Corre, and Veldkamp 2009; Corre, Dechert, and Veldkamp 2006; Kurniawan et al. 2018; Sahner et al. 2015). Additionally, to meet crop demand, oil palm producers generally intensify the addition of N fertilizers (Allen et al. 2015). More reactive N in the ecosystem entails a cascade of environmental effects, ranging from GHG emissions (Cassman, Dobermann, and Walters 2002; Hewitt, Hayward, and Tani 2003; Meijide et al. 2020) to terrestrial acidification and eutrophication of aquatic systems (Galloway et al. 2003).

Such negative impacts and increased awareness in food safety (Reeve et al. 2016) have pressured companies to abolish/substitute palm oil use or ensure palm oil producers adopt conservation practices (Azhar et al. 2015; Brandão et al. 2021). One of them is converting conventional oil palm crops to organic farming, an approach gaining space in the last decade (Nordin et al. 2004; Reeve et al. 2016). Studies comparing the effects of farming system conversion showed decreased nitrate loss by leaching, reducing water and stream pollution (Delate et al. 2003), and improving soil health (Reeve et al. 2016; Tully and McAskill 2020). Despite the potential benefits, studies assessing the effects of organic oil palm crops on ecosystem functioning are still scarce (Foster et al. 2011). To our knowledge, this is the first study comparing these two farming systems in the Amazon.

Here, we assessed how alternative farming management practices affect soil organic matter dynamics in oil palm crops in Southeast Amazon. Accordingly, we compared oil palm crops under conventional and organic farming systems. We also sampled areas of lowland dense ombrophilous forest (floresta densa de terra firme) as reference. Our main question was: Does management changes in organic farming influence soil C and N dynamics in oil palm crops? To this end, we determined in each land-use type: (1) soil physical–chemical properties, C and N concentrations, and their elemental ratio and (2) C and N stocks. We expected that changes in management practices would promote and enhance soil quality, improving its fertility and transformation processes in organic farming. Thus, oil palm organic crops would have higher soil C and N concentrations than conventionally farmed ones, reflecting higher stocks.

2. Materials and methods

Study sites. The study was carried out at the AGROPALMA Group agro-industrial complex, located in the municipality of Tailândia, Pará, Brazil (02°36’–03°24’ S; 48°58’–48°33’ W). The Agro-industrial complex covers 107,000 ha and comprises five mills, two refineries and fractionation plants, and two fats and shortening production units. It also includes areas for the cultivation of oil palm – Elaeis guineensis Jacq. (Arecaceae), and environmental conservation (Figure S1). The AGROPALMA Group, including the refinery and other downstream operations, are certified under the applicable Roundtable on Sustainable Palm Oil (RSPO) standards (Roundtable on Sustainabile Palm Oil 2021).

Oil palm crops are spread over 39,000 hectares and were grown on a combination of pasture, former oil palm crops, and pristine forest. The last forest clearing for plantation establishment was in 2001, and soil burning to prepare the land before crop establishment has been abolished since 2002 (Rodrigues et al. 2005). Besides conventional farming, nearly 10% of these crops (4,153 ha) are certified organic, and another 3,965 ha are in the conversion process to organic farming. Field has been managed under organic farming since 2002 (AGROPALMA Group, (personal communication, 3 January 2021).

Organic crops differ from their conventional counterparts concerning weed control and soil fertilization. In conventional farming, chemical weeding is carried out twice a year by applying glyphosate (N-(phosphonomethyl) glycine). Industrial fertilizers are used. Conversely, brush cutters are used seven times a year to clean up weeds in the organic crops, and reactive natural phosphate (phosphate rock, P₂O₅) and empty oil palm fruit bunches are used as mulch to promote soil fertility. Empty fruit bunches might have been previously grounded or not (AGROPALMA 2019). The biological control of diseases and pests and palm tree pruning are similar in both crops.

Other 64,000 ha consists of natural forest reserves (legal reserve, LR). LR represents the native vegetation that needs to be maintained in rural properties to guarantee the sustainable economic use of natural resources, assist in rehabilitating ecological processes, promote biodiversity conservation, and protect the native fauna and flora (Santiago, de Rezende, and Antônioa Coimbra Borges 2017). LR areas are distributed in fragments that vary in size (from 200 to 3,000 ha) and conservation status (Cunha, Fogaça De Assis Montag, and Juen 2015). These fragments mainly consist of lowland dense ombrophilous forests (florestas densa de terra firme). Tree height averages 20 m, and the understory is relatively open due to former selective logging. The fragments placed within palm oil crops consist of alluvial ombrophilous dense forest (mata de galeria) removed for farming and abandoned for regeneration (AGROPALMA 2019).

The soil type in the area is a dystrophic yellow oxisol (Rodrigues et al. 2005; Santos et al. 2018; Soil Survey Staff 2014). These soil types have low natural fertility, are highly weathered, with low exchangeable base saturation and high exchangeable aluminum saturation (Table 1).

Table 1. Soil (0–30 cm) physical–chemical properties^ǂ (mean ± SE) of the three land-use types assessed in the AGROPALMA Group agro-industrial complex

Land-use type	Sand	Silt	Clay	pH	C/N ratio	P	H+ Al	CEC^†	V%^§	m%^¥
	(g kg^−1)			(H2O)		(mg dm^−3)	(cmolc dm^−3)
Lowland dense ombrophilous forest (FT)	668.67 (29.88)	82.44 (7.77)	248.89 (24.75)	4.01 (0.08)	14.62 (0.66)	2.11 (0.20)	4.78 (0.22)	5.06 (0.22)	5.67 (0.81)	82.00 (2.71)
Conventional oil palm crop (CP)	620.78 (20.12)	119.22 (10.84)	260.00 (13.07)	4.31 (0.04)	14.18 (0.39)	9.85 (2.04)	6.19 (0.33)	7.10 (0.43)	11.69 (1.33)	59.27 (4.51)
Organic oil palm crop (OP)	603.85 (32.17)	171.70 (20.59)	224.44 (19.81)	4.85 (0.09)	12.75 (0.38)	11.07 (2.30)	4.97 (0.34)	6.58 (0.40)	24.34 (2.03)	28.95 (3.79)

^†Analyses carried out at the Laboratório de Solos – Embrapa Amazônia Oriental (Belém, PA, Brazil); values indicate mean (±SD)

^‡CEC = total cation exchange capacity [sum of bases + potential acidity (H⁺ + Al³⁺) at pH 7];

^§V% = base saturation percentage [(100 × sum of bases)/potential CEC at pH 7,0];

^¥m% = Al saturation percentage [(Al × 100)/effective CEC].

Climate is humid tropical (Af, Köppen classification), with annual rainfall between 1750 and 2500 mm (Vanda Maria Sales de et al. 2017). The highest precipitation levels occur from January to May. A well-defined dry season is absent, as the precipitation level in the driest month (November) exceeds 60 mm (Bergson Cavalcanti de et al. 2005). The region is subject to phenomena such as the El Niño-Southern Oscillation (ENOS), characterized by a reduction in the incidence of rainfall in the rainy season (Vanda Maria Sales de et al. 2017).

Field sampling. Samplings were carried out between September and October 2017. We sampled soils from three different land-use types within the agro-industrial complex: lowland dense ombrophilous forest (FT), conventional (CP), and organic (OP) oil palm crops (Figure2).

In FT, we randomly performed three samplings within a 10 × 10 m plot. In CP and OP samplings were performed in the different zones that comprise the crops: (a) palm rows, (b) inter rows, and (c) frond piles. Palm rows are the oil palm cultivation line. Oil palm planting uses an equilateral triangle model, where the spacing between rows is 9 m. Palm trees are planted 7.8 m apart, at a density varying from 128 to 160 plants ha⁻¹ (AGROPALMA Group, (personal communication, 14 June 2021). Inter rows are the zone between palm rows, where harvest products, employees, and machinery pass. Frond piles are the zone where oil palm and weeding residues are mounded. A detailed scheme of the oil palm planting is available in Figure 1. As we carried out three samplings per zone, we had nine soil samples per plantation area. Soil samplings were equidistant from one another.

Figure 1. Schematic illustration of the three subsamples collected in the oil palm plantations. Created with BioRender.com.

Please substitute Figure 1 for its new version attached.

We collected soil samples with a Dutch auger from the layers: 0–10, 10–20, and 20–30 cm depths to determine soil physicochemical properties, C and N concentrations, and their elemental ratio. In these layers, we also collected undeformed soil samples using stainless steel rings (5 × 5 cm) to determine soil bulk density (BD) at each depth interval.

Analytical methods. Soil samples were air-dried, sieved using a 2-mm mesh, and homogenized. They were then handpicked to remove fine roots and other debris and sent to the Laboratório de Solos – Embrapa Amazônia Oriental (Belém, PA, Brazil) to determine physical-chemical properties. Soil undeformed samples were dried at 105°C for 24 h, and BD (g cm⁻³) was determined as the ratio between soil dry mass and ring volume. We used soil BD for soil mass calculation, which is the product of the multiplication between the thickness of each depth interval sampled and BD. We calculated soil C and N stocks for all sites as proposed by Ellert and Bettany (1995), as follows (Equation 1(1) $ES=E_{conc}BDe$(1) ):

(1) $ES=E_{conc}BDe$(1)

where $ES$ is element stock (Mg ha⁻¹), $E_{conc}$ is soil element concentration (g kg⁻¹), $BD$ is bulk density (g cm⁻³), and $e$ is the thickness of each soil layer (cm).

Because C and N concentrations are closely related to soil texture, we corrected their stocks by the clay content of the reference soil (from lowland dense ombrophilous forest), as suggested by J. F. de Moraes et al. (1996) (Equation 2(2) $E{S}_{corr}=ES\frac{cla{y}_{ref}}{cla{y}_{treat}}$(2) ):

(2) $E{S}_{corr}=ES\frac{cla{y}_{ref}}{cla{y}_{treat}}$(2)

where $E{S}_{corr}$ is the corrected element stock (Mg ha⁻¹), $ES$ is the element stock calculated according to eq. 1(1) $ES=E_{conc}BDe$(1) , $cla{y}_{ref}$ is the clay content in the soil reference (g kg⁻¹), and $cla{y}_{treat}$ is the clay content in each land-use type assessed (g kg⁻¹).

Statistical Analyses. We assessed land use type (FT, CP, and OP) variation concerning soil physical–chemical properties, BD, C and N concentrations and stocks according to depth interval (0–10, 10–20, and 20–30 cm). The same procedure was carried out to assess variation among each depth interval from the zones that comprise conventional and organic oil palm crops (palm row, inter row, and frond pile).

Analyses were performed in R version 4.1.1 (R Core Team 2021). A principal component analysis (PCA) was performed to allow an integrated view of soil physicochemical properties according to land-use type. For comparisons of BD, C and N concentrations and stocks among land-use types at different depths (0–10, 10–20, and 20–30 cm), we ran an analysis of variance (ANOVA) followed by post hoc Tukey tests after testing normal data distribution and homoscedasticity. We carried out the same procedures to detect differences among oil palm crop zones and their depth intervals.

3. Results

Land-use type and soil physicochemical properties. We considered the first two principal components in the PCA. These two components reduced data variation to more than 60%; the first component accounted for 35.45%, and the second for 25.27% (Table 2). We noticed the separation of FT soils from CP and especially OP soils in the first component. A partial split of CP and OP soils in the second component was also noticed (Figure 2). The first component was directly related to the variables H+ Al, m%, and clay content. The second component was directly associated with CECt, V%, and H+ Al (Table 3).

Table 2. Eigenvalues, variance explained % and cumulative proportion of total variance from principal component analysis (PCA) components for soil physicochemical properties across land-use types

PCA	Eigenvalue	% of variance	Cumulative %
1	3.55	35.45	35.45
2	2.53	25.27	60.72
3	1.39	13.95	74.67
4	1.05	10.51	85.18
5	0.72	7.17	92.35

Table 3. Variable correlations with principal component analysis (PCA) components for soil physicochemical properties across land-use types

Variables	PCA1	PCA2
CN	−0.27	0.52
P	−4.27	0.60
pH	−16.87	6.19
H+ Al	10.02	13.03
CECt	2.29	26.45
V%	−15.40	14.17
m%	14.13	−17.36
Silt	3.90	11.46
Clay	17.31	1.49
Sand	−15.54	−8.72

Figure 2. Principal component analysis (PCA) based on soil physicochemical properties of each land-use type assessed in the AGROPALMA Group agro-industrial complex. FT = lowland dense ombrophilous forest, CP = conventional oil palm crop, and OP = organic oil palm crop.

Comparisons among land-use types. We detected no variation in BD among land-use types and within depth intervals. Soil C concentration (0–30 cm) was 25% higher in both oil palm crops than in FT. Soil N concentration (0–30 cm) was around 0.55 g kg⁻¹ in FT, nearly 20% and 30% less than in CP and OP, respectively. The highest soil C and N concentrations were found at 0–10 cm interval in organic and conventional crops (p < 0.05). For both elements, we noticed a decrease in concentration with depth interval increment (p < 0.05) (Table 4). Soil C stock (0–30 cm) was close to 47 Mg ha⁻¹ in OP, followed by CP and FT, with values around 39 and 34 Mg ha⁻¹, respectively. We detected no soil C stock variation within depth intervals in FT and CP. In OP, C stock in the 0–10 cm interval was twice the value observed in the 10–20 and 20–30 intervals (p < 0.05) (Figure 3a). OP soil N stock (0–30 cm) was nearly 4 Mg ha⁻¹, while CP and FT values were around 3 and 2 Mg ha⁻¹. When comparing the depth intervals from each land-use type, we observed the highest soil C stock at 0–10 cm in OP. At this depth interval, OP stored 50% more C than CP and 75% than FT (p < 0.05). We found no variation in soil N stocks in CP and FT within depth intervals. In OP soils, we detected a significant variation among depth intervals (p < 0.05). Soil N stock at 0–10 cm was around 30% and 50% higher than in 10–20 and 20–30 cm intervals, respectively (Figure 3b).

Table 4. Variation (mean ± SE) of soil bulk density (BD), carbon (C) and nitrogen (N) concentrations in the three land-use types assessed in the AGROPALMA Group agro-industrial complex

Land-use type	Depth interval	n	BD	C	N
	(cm)		(g cm^−3)	(g kg^−1)
Lowland dense ombrophilous forest (FT)	0–10	9	1.36 (0.03)	9.10 (1.08) ab	0.63 (0.04) bc
	10–20	9	1.39 (0.05)	8.63 (1.21) ab	0.55 (0.05) bc
	20–30	9	1.44 (0.06)	6.74 (0.79) bc	0.47 (0.02) c
Conventional oil palm crop (CP)	0–10	27	1.35 (0.03)	12.49 (1.58) ac	0.80 (0.06) ab
	10–20	27	1.39 (0.03)	9.06 (0.73) ab	0.61 (0.04) bc
	20–30	27	1.33 (0.03)	8.15 (0.75) b	0.61 (0.04) bc
Organic oil palm crop (OP)	0–10	27	1.38 (0.02)	13.32 (1.20) a	0.98 (0.07) a
	10–20	27	1.35 (0.03)	8.79 (0.72) bc	0.71 (0.04) bc
	20–30	26	1.39 (0.04)	7.67 (0.79) bc	0.60 (0.03) c

^ǂChemical analyses carried out at the Laboratório de Solos – Embrapa Amazônia Oriental (Belém, PA, Brazil).

Different letters indicate differences among each land-use (p < 0.05).

Figure 3. Soil C (A) and N (B) stocks (Mg ha¹) under each land-use type assessed in the AGROPALMA Group agro-industrial complex at 0–10 cm, 10–20 cm and 20–30 cm depth. FT = lowland dense ombrophilous forest, CP = conventional oil palm crop, and OP = organic oil palm crop. Different letters indicate differences among land-use types and depth intervals (p < 0.05).

Comparisons between and within farming systems. When we compared oil palm crops concerning their zones (palm row, inter row, and frond pile), soil BD and C concentration varied neither within zones nor depth intervals (p < 0.05). CP and OP had similar soil N concentrations within each depth interval in palm rows and inter rows. On the other hand, OP frond pile at 0–10 cm interval had almost 35% more N than its conventional counterpart (p < 0.05) (Table 5). Soil C stocks were similar for each CP and OP zones and their depth intervals. Soil N stocks in CP and OP palm rows were similar at 0–10 cm intervals and higher at 10–20 and 20–30 cm intervals in OP (p < 0.05). OP inter row had higher N stocks than CP in all depth intervals (p < 0.05). In the frond piles, CP and OP had similar soil N stocks at 0–10 and 10–20 cm intervals, while the 20–30 cm interval N stock was lower in OP (p < 0.05) (Figure 4).

Table 5. Variation (mean ± SE) of soil bulk density (BD), carbon (C) and nitrogen (N) concentrations in the three oil palm crop zones of conventional and organic farming assessed in the AGROPALMA Group agro-industrial complex

Zone	Depth interval	n	BD	C	N
	(cm)		(g cm^−3)	(g kg^−1)
Conventional oil palm crop (CP)
Palm row	0–10	9	1.30 (0.05)	12.58 (2.29) abcd	0.89 (0.10) ac
	10–20	9	1.33 (0.04)	8.19 (0.83) cd	0.60 (0.07) bc
	20–30	9	1.35 (0.05)	7.55 (1.05) cd	0.55 (0.04) bc
Inter row	0–10	9	1.48 (0.05)	8.97 (1.36) bcd	0.66 (0.04) bc
	10–20	9	1.43 (0.07)	6.80 (0.74) cd	0.51 (0.04) b
	20–30	9	1.32 (0.06)	6.01 (0.67) cd	0.56 (0.05) abc
Frond pile	0–10	9	1.27 (0.06)	15.92 (3.78) abc	0.84 (0.14) bc
	10–20	9	1.40 (0.05)	12.18 (1.44) abcd	0.71 (0.07) bc
	20–30	9	1.32 (0.04)	10.89 (1.53) abcd	0.71 (0.08) abc
Organic oil palm crop (OP)
Palm row	0–10	9	1.40 (0.03)	14.74 (1.83) abc	0.91 (0.10) ac
	10–20	9	1.38 (0.05)	8.47 (0.87) bcd	0.65 (0.02) bc
	20–30	9	1.29 (0.06)	7.03 (1.23) cd	0.57 (0.06) bc
Inter row	0–10	9	1.38 (0.05)	8.01 (1.35) cd	0.78 (0.09) bc
	10–20	9	1.38 (0.04)	5.96 (0.72) cd	0.60 (0.05) bc
	20–30	8	1.41 (0.05)	6.86 (0.93) cd	0.64 (0.07) bc
Frond pile	0–10	9	1.37 (0.04)	17.21 (1.78) a	1.24 (0.10) a
	10–20	9	1.30 (0.07)	11.93 (1.25) abcd	0.89 (0.07) ac
	20–30	9	1.48 (0.07)	9.04 (1.74) bcd	0.60 (0.06) bc

^ǂChemical analyses carried out at the Laboratório de Solos – Embrapa Amazônia Oriental (Belém, PA, Brazil).

Different letters indicate differences among oil crop zones and depth intervals (p < 0.05).

Figure 4. Soil C and N stocks (Mg ha⁻¹) at each zone in conventional (A, C) and organic (B, D) oil palm crops assessed in the AGROPALMA Group agro-industrial complex. Different letters indicate differences between zones and depth intervals (p < 0.05).

4. Discussion

Compared to conventional farming systems, organic farming relies on management practices that promote and enhance biodiversity, biological cycles, and soil biological activity (United States Department of Agriculture 2019). It seeks to augment ecological processes that foster plant nutrition yet conserve soil and water resources (Gomiero, Pimentel, and Paoletti 2011; Tully and McAskill 2020). Our results confirm our expectations indicating that organic farming has positively influenced soil quality, especially organic matter dynamics.

Soils from both oil palm crops differed from FT concerning fertility and Al⁺ saturation. Amazon soils are generally old and highly acidic, Al⁺-rich, and nutrient-poor (Baillie 1996). Despite these, the oxisols sampled are suitable for agriculture if inputs and good management practices are employed (Vanda Maria Sales de et al. 2017). It was interesting to note the subtle separation between CP and OP soils, with similar establishment ages. It was mainly due to the higher pH and fertility in OP soils compared to their counterparts, confirming the influence of management practices on soil physical-chemical properties in the long term.

From comparing soil physical properties among land-use types, we found no variation in BD. Some studies indicated lower BD values in soils under organic farming (Das et al. 2017), as organic materials present in SOM have lower densities than inorganic particles (Humberto, Hergert, and Nielsen 2015). On the other hand, the conversion to organic farming may not influence soil BD or lead to its increment in short time (Papadopoulos et al. 2014; Williams et al. 2017; Autret et al. 2016; Michael et al. 2021). Such responses may be a consequence of the similarity in soil management practices adopted in both farmings (Reeve et al. 2016; Colla et al. 2000). Besides, it is noteworthy that oil palm is a perennial crop, and such systems protect soil from erosion and/or compaction of the superficial layers, as well as improve its structure.

The influences of soil management practices on nutrient availability and losses in oil palm crops are poorly understood (Dislich et al. 2017). To our knowledge, this is the first study comparing SOM dynamics in oil palm farming systems in the Amazon. The similarity among land-use types may indicate that forest conversion to perennial crops slightly impacts soil organic C (Frazão et al. 2013; Teng, Sorensen, and Einvind Olesen 2018; Leifeld and Fuhrer 2010). Despite a significant reduction in the aboveground biomass, almost all C previously found under native vegetation is maintained (Maia et al. 2010), as long as crops be managed with soil conservation practices. It is important to remember, though, that our study only assessed the first 30 cm in depth – the interval that comprises more than 90% of oil palm fibrous roots (Jourdan and Rey 1997). Conversely, in native forests, taproots from trees may exceed 8 m in depth, contributing to a greater C stock along with soil profile (Nepstad et al. 1994).

For soil N concentration, the similarity between CP and OP was restricted to palm rows and inter rows. Our results showed that OP could maintain nutrient concentrations without the use of industrial fertilizers. The lack of variation within the other intervals reinforces the importance of quantifying the impacts of palm oil cultivation on soil properties over longer timescales (Quezada et al. 2019; Dislich et al. 2017). The highest N observed in OP frond pile at 0–10 cm interval is also consequential to the presence of the exotic vine Pueraria phaseoloides (Roxb.) Benth. (Fabaceae). This legume is cultivated as plant cover in oil palm crops, contributing to increased N availability and nutrient cycling via N-rich residues and N biological fixation (Viégas et al. 2021). Such findings suggest that the inputs of organic amendments to maintain OP enhance soil nutrient availability (Zulkifli and Khalid 2008), as well as nutrient transformations and biologically active SOM turnover (Leifeld and Fuhrer 2010; Reeve et al. 2016).

Even though not statistically significant, it is essential to consider that soils under organic farming tend to store more C and N than their conventional counterparts. The similarity in soil C stock identified between conventional and organic crops may also be influenced by soil sandy texture, which eases SOM fine fractions percolation to depths below 30 cm. Forest to oil palm conversion leads to less soil C stored, and discontinuing industrial fertilizers reduces N₂O emissions to the atmosphere (Dislich et al. 2017; Colin et al. 2019). That is, organic systems of perennial crops have the potential to preserve and even increase C and N sequestration after land misuse, and they help mitigate the emission of two major greenhouse gases that contribute to climate change.

Organic farming also has economic benefits, yet some negative aspects should be considered. OP is onerous: the costs for maintenance are high; soil cover hampers fruit harvest, and crop yield is reduced compared to conventional farming (AGROPALMA Group (personal communication, 30 October 2017). On the other hand, the added value of palm oil certified as organic leads to a more profitable commodity (Cranfield, Henson, and Holliday 2010). Adding to the savings with industrial fertilizers, this income shows that organic farming is a viable alternative for the environment and oil palm producers in the Amazon (Brandão et al. 2021). Given the results above, it is notorious that further long-term studies assessing SOM dynamics are needed in agricultural areas in the Amazon. Especially in monocultures focused on commodity production in farming systems that support organic management practices – such as the oil palm crops studied here.

5. Conclusions

Our results show that organic farming has positively influenced soil properties and organic matter dynamics in oil palm crops. Organic oil palm crops preserved and even increased C and N sequestration. We acknowledge that the value of ecosystem services provided by standing forests is incomparable. Moreover, further studies are needed to understand how this change in oil palm cultivation affects ecosystem processes in the long term. Still, the replacement of conventional crops by organic ones or their implementation in degraded areas may be a glimpse of how to reconcile commercial demands while preventing environmental injuries and improving the life quality of people involved in the palm oil supply chain in the Amazon.

Author contributions

SF Mardegan conceived, designed, and performed the analyses, collected the data, and wrote the manuscript.

AF de Castro collected the data, performed the analyses, and wrote the manuscript.

SSNF Chaves conceived and designed the analysis and wrote the manuscript.

RSS Freitas, MS Avelar and FAO Teixeira Filho collected the data.

Acknowledgments

We thank the AGROPALMA Group and Conservation International Brazil for their logistics support. We are also grateful to Francisco Joaquim for his valuable help during soil samplings, Jailson Silva Sousa, from the AGROPALMA Group, for the figure and information provided during manuscript preparation, and Neusa Maria da Silva Ferreira, from Embrapa Amazonia Oriental – CPATU, for support during sampling preparation.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed here

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This work was supported by the Pró-Reitoria de Pesquisa e Pós-Graduação - PROPESP, Universidade Federal do Pará also granted SF Mardegan and AF de Castro (Programa de Apoio ao Doutor Pesquisador – PRODOUTOR, Edital 02/2016) during this study.