Comparative Analysis of Soil Biological Activity and Macroinvertebrate Diversity in Amazonian Chakra Agroforestry and Tropical Rainforests in Ecuador

Abstract

Soil biological activity and macroinvertebrate diversity are key indicators of ecosystem function in tropical landscapes. This study evaluates the effects of different land-use systems—Amazonian Chakra agroforestry (timber-based and fruit-based), cocoa monoculture, and tropical rainforest—on soil microbial respiration, enzymatic activity, and macroinvertebrate diversity in the Ecuadorian Amazon. Forest soils exhibited the highest edaphic respiration (240 ± 64.3 mg CO₂ m² ha⁻¹, p = 0.034), while agroforestry systems maintained intermediate biological activity, surpassing monocultures in microbial diversity and enzymatic function. The soil organic matter (SOM) content at a 10 cm depth was significantly higher in monocultures (19.8 ± 3.88%) than in agroforestry and forest soils (p = 0.006); however, the enzymatic activity showed greater functional responses in agroforestry and forest systems. The relationship between recorded CO₂ respiration (REC_CO₂) and basal respiration (RBC_CO₂) exhibited a non-linear trend, as revealed by LOWESS smoothing, suggesting that microbial respiration dynamics are influenced by substrate availability and enzymatic thresholds beyond simple linear predictions. These findings underscore the potential of agroforestry as a sustainable land-use strategy that enhances soil biodiversity, carbon sequestration and nutrient cycling. Implementing optimized agroforestry practices can contribute to long-term soil conservation and ecosystem resilience in tropical agroecosystems.

1. Introduction

The Ecuadorian Amazon is part of a globally significant biodiversity hotspot [1,2] and plays a key role in climate regulation through carbon storage and hydrological functions. However, it faces increasing threats from deforestation, biodiversity loss, and soil degradation due to extractive activities and unsustainable land-use changes [3,4]. This region is also home to Indigenous communities whose livelihoods rely on traditional systems, such as the Chakra, which integrates conservation and production and offers sustainable alternatives to intensive monocultures [5]. Understanding soil biological responses in these land-use systems is essential for informing strategies that support both ecological resilience and socio-economic sustainability.

Recent studies have documented a significant increase in deforestation and degradation of extensive forested areas, leading to severe alterations in soil properties, including erosion and nutrient depletion [6,7,8]. These processes have also been associated with the loss of up to 70% of soil biodiversity [8], primarily driven by intensive agricultural expansion and livestock production [9,10,11]. The conversion of these landscapes into managed forestry systems induces substantial modifications not only in the aboveground vegetation but also in the soil’s physicochemical and biological properties [12].

Several authors have considered agroforestry a sustainable and environmentally friendly practice that contributes to carbon sequestration and atmospheric carbon storage, thereby playing a role in climate change mitigation [13,14,15,16]. Agroforestry is a land management approach that integrates agricultural and forestry practices simultaneously within the same landscape [4,13,17]. According to the Food and Agriculture Organization (FAO) of the United Nations, agroforestry systems can be classified into three types: agrosilvicultural systems, which integrate trees and crops; silvopastoral systems, which combine forestry with livestock grazing; and agrosilvopastoral systems, which incorporate trees, animals, and crops within the same production system [13,18]. In this same context, traditional agroforestry systems have been developed over millennia by Indigenous communities, incorporating rich cultural, agricultural, and food security practices [19]. One of these systems in Ecuador is the “Amazonian Chakra”, which was recently recognized by the FAO as a Globally Important Agricultural Heritage System (GIAHS) in February 2022 [5]. The Amazonian Chakra is managed by Indigenous Kichwa families and is characterized by the integration of diverse crops, fruit trees, timber species, and medicinal plants in multistratum arrangements. Rooted in ancestral knowledge [20], it contributes to the conservation of agrobiodiversity and enhances soil organic matter and nutrient cycling [6]. As an example of agroforestry systems (AFSs), in the Ecuadorian Amazon region, the Chakra is a polyculture agrarian system that is characterized by the use of market-oriented crop(s) for the generation of monetary income, such as cocoa, bananas, and manioc, and livestock for the agricultural family’s consumption [19].

Agroforestry systems integrate a mix of crops and trees and serve as important carbon sinks, contributing to climate change mitigation [7,14]. In addition to their carbon sequestration potential, AFSs provide multiple ecosystem services, including soil biodiversity conservation, erosion control, nutrient cycling, and enhanced soil health and biological activity [21,22]. Studies on cacao and coffee-based agroforestry systems have demonstrated increases in microbial biomass and enzymatic activity [17], supporting sustainable land use and maintaining agricultural productivity [7,14,23]. In contrast, monoculture systems have been associated with biodiversity loss, soil fertility decline, and overall soil degradation [14]. Cacao (Theobroma cacao) has traditionally been cultivated in association with forest species, medicinal plants, and fruit trees; however, monoculture-based cacao production models have also been widely implemented [24]. However, due to the recognized soil and biodiversity benefits, there has been a growing preference for both traditional and intensified agroforestry management systems in tropical ecosystems [6,16,24].

Soil biodiversity encompasses the full spectrum of biological variability within soil ecosystems, ranging from genetic diversity to community-level interactions and from microaggregates to entire landscapes [25]. They play a fundamental role in provisioning, regulating, and supporting ecosystem functions that sustain soil health [25,26]. Soil biological activity can be assessed using multiple indicators of soil quality, including soil respiration, which reflects the carbon flux into the atmosphere and terrestrial carbon cycling [21]; soil biodiversity, which responds rapidly to anthropogenic disturbances; organic matter and litter content, which serve as key regulators of nutrient availability [24,27]; and soil enzymatic activity, which provides insights into microbial functionality under different environmental conditions [27,28]. These processes are largely influenced by vegetation residue accumulation, temperature, moisture levels, and the availability of soil organic matter (SOM) [21].

Among biological indicators, soil enzymatic activity exhibits rapid responses to environmental changes and disturbances, making it a valuable tool for assessing biochemical processes related to the carbon (C) and nitrogen (N) cycles in response to management practices and crop rotation systems [29,30,31]. However, soil enzyme activities are regulated by temperature, moisture, pH, aboveground biomass, and microbial communities [12], and in many cases, they can be significantly influenced by agronomic practices, such as fertilization [31]. Several authors have considered soil enzyme activity as a potential indicator of soil quality, nutrient balance [12,32], and the metabolic status of edaphic microbiota [17,30], while it can also serve as a biomarker for soil contamination [32]. Microbial activity, like soil enzymatic processes, is closely linked to the early decomposition of plant residues, litterfall, and transformation of SOM [21,29]. Enzymes associated with the C and N cycles play a fundamental role in nutrient availability for both soil biota and plants [29]. Microbial diversity is widely recognized as a critical indicator of soil biological quality because of its direct relationship with nutrient cycling and ecosystem stability [7]. Furthermore, despite the ecological relevance of soil invertebrate diversity, recent research has shown a declining interest in studying these organisms, even though they exhibit rapid responses to environmental changes and disturbances [9].

Management systems based on cover crops and pasture cultivation tend to reduce SOC levels, microbial biomass formation, and enzymatic activity, mainly due to the biochemical composition of plant residues, which influences the decomposition dynamics and nutrient cycling [23,33]. In contrast, forest ecosystems rely on litter accumulation as a key factor in nutrient cycling, serving as a reliable indicator of soil fertility due to its role in nutrient input and enzymatic activity enhancement [12,22].

Soil extracellular enzymes from roots, soil fauna, and microorganisms are involved in multiple biological processes [27,31]. These enzymes are responsible for converting macromolecular organic matter into bioavailable nutrients, facilitating plant nutrient acquisition and the soil carbon cycle [11,15,34]. The complexity of cellulose-degrading enzymes, such as β-glucosidase, is particularly relevant as it is involved in the decomposition of the most abundant plant polysaccharide, cellulose, leading to glucose release, which is the primary carbon source for soil microorganisms [29,33]. Dehydrogenase activity is considered a reliable indicator of microbial activity and oxidative metabolism in the soil [30]. Meanwhile, urease activity serves as a soil quality indicator because of its regulatory role in nitrogen availability for plant uptake [30]. Similarly, soil phosphatase (PHO) is an activity, that plays an essential role in the mineralization of organic phosphorus, favoring is bioavailability by plants. Arylsulfatase activity (SU) is one of the most common enzymatic activities observed in soils. In addition to satisfying microbial and plant needs for sulfate ions, it participates in the degradation of xenobiotic compounds in soils [35]. Therefore, the determination of soil enzymatic activities is a suitable indicator of soil status and can be readily applied to assess the effects of management practices on soil quality [36].

Despite the increasing interest in understanding soil biological activity in Amazonian ecosystems, research providing a comprehensive assessment of key biological parameters remains limited. Existing studies often focus on isolated aspects of soil microbiology, leaving critical gaps in our understanding of how land-use changes influence soil biodiversity, microbial functionality, and biochemical cycles. This study provides the first integrated analysis of cocoa-based agroforestry systems—including associations with timber and fruit species and cocoa monocultures—compared to Amazonian forests, allowing us to quantify the biological impact of land-use transformation on soil microbial communities, enzymatic activity, and macroinvertebrate diversity. This study addresses an important research gap by comparing soil biological activity across traditional Amazonian Chakra agroforestry systems, cocoa monocultures, and secondary forests in the Ecuadorian Amazon. Unlike previous studies that focused on single indicators, it combines microbial respiration, enzymatic activity, soil organic matter, and macroinvertebrate diversity in a comprehensive analysis. From this perspective, three main objectives were established: (a) to assess the effects of land-use change on soil biological activity using indicators such as edaphic and basal respiration, enzymatic activity, and macroinvertebrate diversity; (b) to evaluate the relationship between soil organic matter content and soil biological indicators across different land-use systems; and (c) to identify linear and non-linear trends in microbial respiration responses related to changes in soil organic matter and land-use practice. In this study, we hypothesized that soil biological processes (edaphic and basal respiration) and soil biological properties (microbial populations, organic matter, and enzyme activity) are directly influenced by land-use change. Additionally, variations in soil organic matter content due to land use can directly impact soil biological processes and activities.

2. Materials and Methods

This research was conducted in the Arosemena Tola and Tena cantons of Napo Province (Figure 1). The region’s climate ranges from humid to humid tropical, with temperatures varying between 22 °C and 25 °C. The area maintains a high relative humidity throughout the year and receives an average annual rainfall of approximately 3000 mm. The topography consists of a series of medium-sized hills that originate from the eastern slopes of the Andes Mountains [6,37].

2.1. Land Uses

Due to the characteristics of the relief, topography, climate, and diversity of flora and fauna, the Ecuadorian Amazon constitutes the optimal space to develop different types of management practices, such as associations of crops, monocultures, forage, and forestry or agroforestry systems, which are adapted according to the needs of production. This research covers 11 sampling units: 3 cocoa monocultures (CMC), 7 Amazonian Chakra under agroforestry systems with cocoa in association with timber forest species (TAFS) and fruit tree species (FAFS), and a secondary forest.

The Amazonian Chakra systems are traditional agroforestry practices developed by Indigenous Amazonian communities, particularly the Kichwa people. These systems have been cultivated for centuries as a sustainable way to manage land while maintaining the ecological balance and cultural heritage. Unlike modern agricultural methods, Chakras integrate ancestral knowledge that has been passed down through generations [5,19]. Additionally, Chakras are small-scale biodiverse agroforestry plots, where families cultivate a variety of crops, such as cacao, cassava, plantains, medicinal plants, native fruit trees (e.g., Inga edulis, Citrus sinensis, Terminalia oblonga, Citrus aurantiifolia, Bactris gasipaes), and timber trees (e.g., Cordia alliodora, Piptocoma discolor, Schefflera morototoni, Persea americana). They also include trees that enhance soil fertility and support biodiversity. These systems are maintained using organic methods, avoiding the use of chemical fertilizers and pesticides. Beyond food production, Chakras hold cultural and spiritual significance as they embody Indigenous worldviews that emphasize harmony between humans and nature [5,38]. As a result, Chakra systems serve as a model for sustainable agriculture in the Amazon, balancing food production with environmental conservation. They support Indigenous food sovereignty, enhance resilience to climate change, and offer alternatives to destructive land-use practices. By preserving traditional knowledge, they also contribute to the cultural identity and well-being of Indigenous communities [19,38], as described by Huera-Lucero et al. [6].

2.2. Field Sampling

The experimental design followed systematic sampling, which included: 4 treatments (CMC, TAFS, FAFS, and secondary forest), 3 measurement points or subplots (10 m × 10 m), and a soil sampling depth factor. During the data collection and sampling phases, necessary measures were taken to obtain accurate and representative information.

The response to the treatments was evaluated through sampling of large plots [39,40] of 1600 m² (40 m × 40 m); in order to increase the number of repetitions, they were divided into three subplots (10 m × 10 m) to represent different sections of the field, and sampling was performed in triplicate. (Figure 2).

In each sampling subplot (10 m × 10 m), different types of samples were taken. To determine the enzymatic activity of the different land uses, five subsamples were taken at a depth of 10 cm, which were homogenized to obtain a composite sample for each subplot, represented by circles in Figure 2. In the part closest to the center of each subplot, a monolith (0.25 m × 0.25 m) was extracted to a depth of 20 cm to perform an on-site count of macroinvertebrates. Likewise, a soil sample was collected up to 10 cm deep to determine soil diversity using the Berlese method, near the center of the subplot. In addition, with the help of a 0.25 m² quadrant, the material corresponding to dead plant remains (leaf litter) was collected. Soil respiration was carried out in situ at the part closest to the center of the subplot. Basal respiration was estimated in the laboratory from previously collected soil samples, which were processed and sieved according to the protocols of the laboratory of the Amazon State University [41].

2.3. Sample Analysis

Once all the information was collected in the field and organized, it was possible to estimate the abundance and diversity of macroinvertebrates existing in the selected types of land use, using some diversity indicators and indices (

Table 1. Equations used.

Parameter	Equation	Source
Soil organic carbon (SOC)	$\mathrm{S}\mathrm{O}\mathrm{C}=\mathrm{B}\mathrm{D}\times \left({\displaystyle \frac{\mathrm{T}\mathrm{O}\mathrm{C}}{100}}\right)\times \mathrm{D}\times 10000$	[37,42,43]	(1)
Litter carbon (LC)	$\mathrm{L}\mathrm{C}=\mathrm{D}\mathrm{M}\times \mathrm{C}\mathrm{F}\times 10000$	[44]	(2)
Simpson index (S)	$\mathrm{S}={\displaystyle \frac{\sum \mathrm{n}(\mathrm{n}-1)}{\mathrm{N}(\mathrm{N}-1)}}$	[44,45]	(3)
Shannon Wiener index (H’)	$\mathrm{H}\mathrm{’}=-\sum \left(\mathrm{p}\mathrm{i}\mathrm{x}\left(\mathrm{l}\mathrm{o}\mathrm{g}2\mathrm{p}\mathrm{i}\right)\right)$ $\mathrm{p}\mathrm{i}=\mathrm{n}\mathrm{i}/\mathrm{N}$	[46]	(4)

BD: bulk density (Mg m⁻³); TOC: total organic carbon (Mg ha⁻¹); D: depth (m); DM: dry matter (Mg); CF: IPCC carbon fraction, which is 0.5; n: number of total organisms per spp. N: total number of organisms for all spp.; pi: relative abundance; ni: number of ind. spp.

).

2.3.1. Berlese Extraction Method

This method is based on the negative phototropism and thermotropism of the soil mesofauna, with a source of light and heat that falls on the soil and leaf litter sample for 14 days, during which the mesofauna present in the sample gradually move in the opposite direction to the source. Once the required time has elapsed, the containers are covered (70% alcohol) for subsequent analysis and taxonomic identification in the laboratory under a stereomicroscope [25,47], using the taxonomic identification keys described by [25,48] and based on the taxonomic identification described by [49,50]. Taxonomic identification was only carried out down to the family and, in some cases, subfamily level.

2.3.2. Manual Techniques in the Field

Sampling was carried out within each subplot, for which a 25 cm × 25 cm monolith was extracted up to 20 cm deep, where all visible macroinvertebrates were collected according to manual sampling techniques [24,26,50]. Macroinvertebrates were stored in 70% ethyl alcohol until identification in the laboratory using the taxonomic identification keys described by [25,48] and based on the taxonomic identification described by [49,50]. Taxonomic identification was only carried out down to the family and, in some cases, subfamily level.

2.4. Biological Parameters

Total organic carbon (TOC) was determined using the Walkley–Black wet digestion method [51,52], and bulk density (BD) was measured using the cylinder method at a depth of 10 cm, in which the samples were weighed and dried at 105 °C for 24 h [53,54]. The determination of this parameter allows chemical and biological variables to be expressed in terms of the surface area (kg ha⁻¹ or Mg ha⁻¹) [6]. Litter carbon (LC) was obtained after drying the material collected with the help of the 0.25 m² quadrant [44] at 105 °C for 24 h, obtaining dry matter (DM) per hectare, according to the equation in Table 1.

Soil edaphic respiration (ER) was estimated in the field according to a modification of the static chamber method, which consists of the use of plastic containers buried at a depth of 1 cm that enclose a known area of soil [41]. The CO₂ produced over 24 h was captured in the alkali, which was placed in a container inside the trap. To stop the reaction after this time, BaCl₂ is used, and the excess NaOH is titrated with HCl in the presence of a phenolphthalein indicator. It is recommended to use covered containers with alkali as blanks, and the results are expressed in mg CO₂ m⁻²h⁻¹ or mg C-CO₂ m⁻²h⁻¹ [41]. Soil basal respiration (BR) was estimated in the laboratory based on the physiological response of microorganisms under minimal conditions of labile substrate sources. The CO₂ released was measured using alkali traps at level, using root-free samples sieved at <2 mm. The CO₂ production was determined for 24 h at 25 °C when microbial respiratory processes were carried out inside plastic containers, in which a vial containing 20 mL of 0.5 M NaOH was suspended and hermetically sealed [24,41].

2.5. Enzymatic Activity

Soil samples were air-dried at room temperature and sieved (<2 mm) according to laboratory protocols. Soil enzyme activities were determined according to López-Piñeiro et al. [55] as follows:

▪

Dehydrogenase (DH) activity: Here, 1 g of soil was used, which was incubated in the shade at 20 °C for 20 h with 0.20 mL of 0.4% 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-phenyl-2H-tetrazolium (INT) chloride as substrate. After incubation, the iodonitrotetrazolium formazan produced was extracted with 10 mL of methanol, and the absorbance was measured at 490 nm.

▪

β-glucosidase (GL) activity: For this determination, 1 g of soil was used, which was incubated at 37 °C for 1 h, with 4 mL of 25 mM 4-nitrophenyl-β-d-glucopyranoside in 0.1 M modified universal buffer (MUB) pH 6.0. Then, to stop the reaction, the samples were refrigerated at 2 °C for 15 min, and the p-nitrophenol produced in the enzymatic reactions was determined at 400 nm.

▪

Urease (UR) activity: Here, 0.50 g of soil was used and incubated at 30 °C for 1.5 h with 2 mL of 0.1 M pH 7.0 phosphate buffer and 0.50 mL of 1.07 M urea. The ammonia released during the hydrolytic reaction was measured spectrophotometrically at 636 nm.

▪

Phosphatase (PHO) activity: Here, 1 g of soil was used, which was incubated at 37 °C for 1 h with 4 mL of 25 mM 4-nitrophenyl phosphate MUB pH 6.5. Then, to stop the reaction, the samples were refrigerated at 2 °C for 15 min, and the p-nitrophenol produced in the enzymatic reactions was determined at 398 nm.

▪

Arylsulfatase (SU) activity: For this determination, 1 g of soil was used, which was incubated at 37 °C for 1 h with 4 mL of 5 mM 4-nitrophenyl sulfate in 0.5 M acetate buffer pH 5.8. Then, to stop the reaction, the samples were refrigerated at 2 °C for 15 min, and the p-nitrophenol produced in the enzymatic reactions was determined at 410 nm.

All enzymatic activities were determined in triplicate [55,56]. Blank assays without soil or substrate were performed at the same time as controls.

2.6. Data Analysis

All statistical analyses were conducted using R Studio version 4.4.0 and SPSS IBM version 25. Descriptive statistics and normality testing (Shapiro–Wilk test) were performed using SPSS, while ANOVA, PCA, linear regression, and LOWESS smoothing were conducted using R Studio. Macroinvertebrate abundance and diversity data were processed using Excel databases. PCA was used to identify the main components explaining the variability in soil biological and physicochemical variables among land-use systems. Prior to the PCA, all variables were standardized using z-score normalization (mean = 0, standard deviation = 1) to ensure equal contribution of variables with different units and magnitudes. This multivariate approach allowed us to reduce dimensionality, explore overall system-level trends, and visualize groupings based on biological activity. This analysis directly supports objective three by revealing broader patterns of soil functioning that are not captured through univariate methods. Prior to applying ANOVA, the assumptions of normality and homogeneity of variances were tested using Shapiro–Wilk and Levene’s tests, respectively. In this study, since the values of the Shapiro−Wilk statistic for the variables evaluated ranged from 0.80 to 1.0 and the probability was greater than 0.05, it was considered that the data were adjusted or consistent with a normal distribution; therefore, the relationship between the different variables was determined using Pearson’s correlation. Pearson’s correlation coefficient (r) was used to assess the strength and direction of linear relationships among soil biological and physicochemical variables. An exploratory data analysis was also conducted to validate the statistical assumptions and detect potential outliers or deviations.

The LOWESS (Locally Weighted Scatterplot Smoothing) regression was employed to capture non-linear trends in microbial respiration dynamics, providing a more flexible and robust visualization of relationships that may not be adequately described by linear models. Unlike conventional regression approaches, LOWESS does not assume a predefined functional form, making it particularly useful for exploring gradual variations in biological activity in response to environmental gradients. This method has been widely applied in ecological and environmental sciences for trend detection and complex data structure interpretation [57].

3. Results

3.1. Functional Groups and Taxonomic Composition of Soil Macroinvertebrates

3.1.1. Soil Macroinvertebrate Composition Based on the Manual Method

The analysis of soil macroinvertebrate functional groups (

Table 2. Macroinvertebrate taxonomic identification using manual sampling techniques.

Functional Group	Taxa	Order	Family	Subfamily	Number Individuals
Soil engineers	Oligochaeta	Crassiclitellata	Lumbricidae	---	361
	Hexapoda	Blattodea	Termitidae	---	216
		Hymenoptera	Formicidae	Amblyoponinae	77
				Dolichodeninae	23
				Dorilynae	10
				Formicinae	66
				Myrmicinae	310
				Ponerinae	43
Detritivores	Crustacea	Isopoda	Armadillidae	---	16
	Gastropoda	Ellobiida	Ellobiidae	Pedipedinae	9
	Hexapoda	Coleoptera (larva)	---	---	61
Depredators	Myriapoda	Chilopoda	Geophilomorpha	Geophilidae	14
	Chelicerata	Araneae	---	---	2
Herbivores	Hexapoda	Orthoptera	Gryllotalpidae	---	5
		Hemiptera	Cicadidae	---	1
Omnivores	Hexapoda	Blattodea	Blattidae	---	6
		Dermaptera	Carcinophoridae	---	11
Total					1231

) identified 1231 individuals classified into five functional categories: soil engineers, detritivores, predators, herbivores, and omnivores. Soil engineers represented the most abundant group, dominated by Oligochaeta (Lumbricidae, 361 individuals) and Blattodea (Termitidae, 216 individuals). Hymenoptera (Formicidae) was also prevalent, with Myrmicinae (310 individuals) and Formicinae (66 individuals) being the most common subfamilies.

Detritivores were represented by Crustacea (Isopoda, Armadillidae, 16 individuals), Gastropoda (Ellobiidae, nine individuals), and Hexapoda larvae (Coleoptera, 61 individuals). Predators included Geophilomorpha (Geophilidae, 14 individuals) and Araneae (two individuals), indicating a relatively low abundance of carnivorous taxa. Herbivores and omnivores were less abundant, with Orthoptera (Gryllotalpidae, five individuals) and Hemiptera (Cicadidae, one individual) representing herbivores, and Blattodea (Blattidae, six individuals) and Dermaptera (Carcinophoridae, 11 individuals) classified as omnivores.

3.1.2. Soil Macroinvertebrate Composition Based on the Berlese Extraction Method

The Berlese extraction method identified 5219 individuals distributed across six functional groups: soil engineers, detritivores, predators, saprophages, omnivores, and herbivores (

Table 3. Soil macroinvertebrate composition based on the Berlese extraction method.

Functional Group	Taxa	Order	Family	Subfamily	Number Individuals
Soil engineers	Hexapoda	Hymenoptera	Formicidae	Ectatominae	43
				Amblyoponinae	56
				Dolichodeninae	14
				Dorilynae	42
				Formicinae	101
				Myrmicinae	272
		Blattodea	Termitidae	---	76
Detritivores	Crustacea	Isopoda	Armadillidae	---	59
	Hexapoda	Collembola	Entomobryomorpha	---	1285
			Poduromorpha	---	1354
		Coleoptera (larva)	---	---	118
Depredators	Chelicerata	Araneae	---	---	39
		Pseudoescorpion	---	---	21
Saprophages	Myriapoda	Simphyla	Scolopendrellidae	---	33
	Hexapoda	Diplura	Endognatha	---	47
		Protura	Eosentomidae	---	21
	Chelicerata	Acari	---	---	1576
Omnivores	Hexapoda	Dermaptera	Carcinophoridae	---	38
Herbivores		Thysanoptera	Aeolothripidae	---	24
Total					5219

). Soil engineers were primarily represented by Hymenoptera (Formicidae, 528 individuals), with Myrmicinae (272 individuals) and Formicinae (101 individuals) as the dominant subfamilies. Termitidae (Blattodea, 76 individuals) was also present, reinforcing the structural role of these taxa in soil modification.

Detritivores constituted a significant portion of the macrofauna community, dominated by Collembola (Entomobryomorpha: 1285 individuals, Poduromorpha: 1354 individuals), Isopoda (Armadillidae: 59 individuals), and Coleoptera larvae (118 individuals). Their high abundance highlights their key roles in organic matter decomposition. Predators included Araneae (39 individuals) and Pseudoscorpions (21 individuals), suggesting a moderate presence of predatory arthropods. Saprophages were mainly represented by Acari (1576 individuals), Diplura (47 individuals), Symphyla (33 individuals), and Protura (21 individuals), indicating their importance in fungal decomposition and soil nutrient cycling. Omnivores and herbivores were less abundant, with Dermaptera (38 individuals) and Thysanoptera (24 individuals), respectively.

3.1.3. Soil Macroinvertebrate Abundance and Diversity Across Land-Use Systems

The results indicate that soil macroinvertebrate abundance (Ab) and diversity indices (H’, S) varied across land-use systems, with significant differences primarily observed in soil samples analyzed using the Berlese method (

Table 4. Macroinvertebrate diversity, using manual and Berlese sampling techniques.

Method	Index	Types of Land Use				ANOVA 1 p-Value
		CMC	FAFS	TAFS	Forest
Manual techniques	Ab10cm	31.0 (±21.4)	26.0 (±18.6)	18.5 (±7.37)	30.0 (±21.0)	n/s
	Ab20cm	10.9 (±10.7)	11.2 (±7.93)	14.9 (±21.6)	10.0 (±8.89)	n/s
	H’10cm	0.95 (±0.43)	0.94 (±0.39)	0.92 (±0.40)	0.99 (±0.22)	n/s
	H’20cm	0.87 (±0.38)	0.93 (±0.24)	0.63 (±0.30)	0.86 (±0.58)	n/s
	S10cm	0.53 (±0.20)	0.52 (±0.18)	0.49 (±0.18)	0.53 (±0.17)	n/s
	S20cm	0.50 (±0.19)	0.54 (±0.10)	0.39 (±0.19)	0.49 (±0.34)	n/s
Berlese	AbLitter	95.0 (±38.4)	81.7 (±34.5)	93.8 (±33.5)	99.0 (±54.7)	n/s
	AbSoil	67.3 b (±33.6)	69.4 b (±21.1)	76.6 b (±14.7)	143 a (±24.0)	***
	H’Litter	1.57 b (±0.23)	1.50 b (±0.34)	1.58 b (±0.22)	2.01 a (±0.12)	*
	H’Soil	1.64 ab (±0.10)	1.56 b (±0.31)	1.56 b (±0.09)	1.87 a (±0.10)	*
	SLitter	0.74 (±0.08)	0.72 (±0.08)	0.74 (±0.06)	0.84 (±0.02)	n/s
	SSoil	0.75 (±0.03)	0.75 (±0.06)	0.74 (±0.03)	0.80 (±0.03)	n/s

¹ p-value: * p < 0.05; *** p < 0.001; n/s = not significantly different; different letters indicate significant differences between groups. Values ± in parentheses are the standard deviations of the mean. CMC: cocoa monoculture; FAFS: fruit agroforestry system; TAFS: timber agroforestry system; Ab: abundance; H’: Shannon index; S: Simpson index.

). For manual sampling techniques, no significant differences (n/s) were detected in macroinvertebrate abundance at 10 cm and 20 cm soil depths (Ab10cm, Ab20cm), with values ranging from 18.5 to 31.0 individuals per sample at 10 cm depth and 10.0 to 14.9 individuals per sample at 20 cm depth across the different land-use types. Similarly, Shannon (H’) and Simpson (S) diversity indices at both depths showed no statistical differences, suggesting that land use does not strongly influence macroinvertebrate diversity when assessed using manual extraction.

Conversely, the Berlese-extracted samples revealed significant differences in macroinvertebrate abundance and diversity. Soil macroinvertebrate abundance (AbSoil) was significantly higher in forest soils (143 ± 24.0) than in agroforestry (FAFS, TAFS) and monoculture (CMC) systems, where the values ranged from 67.3 to 76.6 (p < 0.001). However, litter macroinvertebrate abundance (AbLitter) showed no significant differences across systems. Regarding diversity indices, Shannon’s index for litter (H’Litter) and soil (H’Soil) was significantly higher in the forest (H’Litter = 2.01 ± 0.12, H’Soil = 1.87 ± 0.10) than in the other land-use systems (p < 0.05). Likewise, Simpson’s index for litter (SLitter) and soil (SSoil) was highest in the forest, but the differences were not statistically significant.

3.2. Soil Quality Indicators

3.2.1. Soil Quality Indicators and Biological Parameters

Soil biological activity varied significantly across different land-use systems. Edaphic respiration (ER) was highest in the forest (240 ± 64.3 mg CO₂ m² ha⁻¹), showing a statistically significant difference (p < 0.05) compared to the agroforestry and monoculture systems, which exhibited lower values ranging from 170 to 174 mg CO₂ m² ha⁻¹. Similarly, carbon respiration (C-CO₂) was significantly higher in the forest (65.5 ± 17.5 mg C-CO₂ m² h⁻¹) than in the managed systems (p < 0.05). No significant differences were observed in basal respiration (BR) across land uses, with values ranging between 389 and 711 mg CO₂ kg⁻¹ soil d⁻¹ (

Table 5. Summary of biological parameters and soil health quality indicators.

Parameters	Types of Land Use				ANOVA 1 p-Value
	CMC	FAFS	TAFS	Forest
ER (mg CO2 m2 ha−1)	173 b (±39.2)	170 b (±29.8)	174 b (±28.3)	240 a (±64.3)	*
ER (mg C-CO2 m2 ha−1)	47.2 b (±10.7)	46.3 b (±8.10)	47.5 b (±7.73)	65.5 a (±17.5)	*
BR (mg CO2 kg−1 soil d−1)	711 (±398)	508 (±502)	499 (±343)	389 (±371)	n/s
BR (mg C-CO2 kg−1 soil d−1)	194 (±108)	139 (±137)	136 (±93.6)	106 (±101)	n/s
SOM10cm (%)	19.8 a (±3.88)	9.10 b (±6.71)	14.0 ab (±7.81)	8.49 b (±0.38)	**
SOM20cm (%)	12.0 a (±1.22)	5.33 b (±2.90)	8.73 ab (±4.80)	3.94 b (±0.26)	***
TOC10cm (%)	11.5 a (±2.25)	5.28 b (±3.89)	8.13 ab (±4.53)	4.92 b (±0.22)	**
TOC20cm (%)	6.99 a (±0.71)	3.09 b (±1.68)	5.07 ab (±2.79)	2.28 b (±0.15)	***
Litter (Mg ha−1)	7.33 (±5.73)	6.37 (±2.38)	10.1 (±4.75)	10.4 (±0.74)	n/s
UR (µg NH4+/g−1 ha−1)	298 a (±65.1)	190 a (±94.0)	293 a (±109)	280 a (±52.2)	*
SU (µg pNP/g−1 ha−1)	159 a (±39.4)	114 b (±17.2)	153 ab (±21.9)	114 b (±9.81)	***
PHO (µmol pNP/g−1 ha−1)	3.53 a (±1.10)	3.62 a (±1.46)	4.66 a (±1.17)	5.27 a (±1.03)	*
GL (µmol pNP/g−1 ha−1)	0.58 (±0.08)	0.54 (±0.13)	0.53 (±0.14)	0.54 (±0.05)	n/s
DH (µg INTF/g−1 ha−1)	0.47 (±0.14)	0.45 (±0.14)	0.49 (±0.31)	0.17 (±0.26)	n/s

¹ p-Value: * p < 0.05; ** p < 0.01; *** p < 0.001; n/s = not significantly different; different letters indicate significant differences between groups. Values presented as ± in parentheses are the standard deviations from the mean. CMC: cocoa monoculture; FAFS: fruit agroforestry system; TAFS: timber agroforestry system; ER: edaphic respiration; BR: basal respiration; SOM: soil organic matter; TOC: total organic carbon; UR: urease activity; SU: arylsulfatase activity; PHO: phosphatase activity; GL: β-glucosidase activity; DH: dehydrogenase activity.

). Likewise, the total organic carbon (TOC) at a 10 cm depth was significantly higher in the monoculture system (11.5 ± 2.25%) than in agroforestry and forest soils (p < 0.01), whereas at a 20 cm depth, the TOC was highest in monoculture (6.99 ± 0.71%) and lowest in forest soils (2.28 ± 0.15%) (p < 0.001). However, enzymatic activity varied among land use types. Urease activity (UR) showed marginal significance (p < 0.05), with higher values in the monoculture system (298 ± 65.1 µg NH₄⁺ g⁻¹ ha⁻¹) than in agroforestry and forest soils. Arylsulfatase activity (SU) was significantly lower in the forest and CMC systems (114 ± 9.81 and 114 ± 17.2 µg pNP g⁻¹ ha⁻¹, respectively) than in the TAFS and FAFS (p < 0.001). Phosphatase (PHO) activity showed significant differences, with values ranging between 3.53 and 5.27 µmol pNP g⁻¹ ha⁻¹ (p < 0.05). β-Glucosidase (GL) and dehydrogenase (DH) activities exhibited no statistically significant differences across land-use types (Table 5).

3.2.2. Correlation Between Soil Properties and Biological Activity

Pearson’s correlation analysis (

Table 6. Pearson correlations showing functional relationships between soil properties and biological activity.

Variables	ER	BR	SOM10cm	SOM20cm	Litter	UR	SU	PHO	GL	DH	H’10cm	S10cm	H’Litter	H’Soil	SLitter	SSoil
ER	1	−0.11	−0.08	−0.10	0.08	0.10	−0.08	0.19	−0.03	−0.208	−0.12	−0.16	0.26	0.31	0.19	0.345
BR		1	−0.08	−0.01	−0.02	0.06	0.16	0.003	−0.17	−0.134	0.15	0.09	−0.39 *	−0.34 *	−0.34	−0.30
SOM10cm			1	0.927 **	−0.24	0.69 **	0.37 *	0.18	0.53 **	0.383 *	−0.004	−0.03	0.08	0.30	0.08	0.18
SOM20cm				1	−0.27	0.77 **	0.52 **	0.21	0.55 **	0.487 **	−0.08	−0.08	0.02	0.15	0.06	0.07
Litter					1	−0.17	0.10	−0.22	−0.48 **	−0.39 *	−0.15	−0.18	0.10	−0.06	0.083	−0.04
UR						1	0.53 **	0.63 **	0.57 **	0.366 *	−0.06	−0.07	0.16	0.33	0.18	0.24
SU							1	0.31	0.38 *	0.331	−0.15	−0.13	−0.04	0.04	0.10	−0.06
PHO								1	0.51 **	0.246	−0.02	−0.06	0.16	0.39 *	0.20	0.30
GL									1	0.51 **	−0.002	0.03	0.02	0.31	0.09	0.22
DH										1	0.09	0.14	−0.02	−0.003	0.01	0.11
H’10cm											1	0.96 **	0.03	0.18	−0.10	0.12
S10cm												1	0.02	0.14	−0.11	0.12
H’Litter													1	0.49 **	0.93 **	0.46 **
H’Soil														1	0.39 *	0.87 **
SLitter															1	0.37 *
SSoil																1

* The correlation is significant at the 0.05 level (bilateral); ** the correlation is significant at the 0.01 level (bilateral); ER = ERC_CO₂: edaphic respiration; BR = BRC_CO₂: basal respiration; SOM: organic matter; UR: urease activity; SU: arylsulfatase activity; PHO: phosphatase activity; GL: β-glucosidase activity; DH: dehydrogenase activity; H’_10cm: Shannon index (in situ); S_10cm: Simpson index (in situ); H’_Litter: Shannon index (Berlese); H’_Soil: Shannon index (Berlese); S_litter: Simpson index (Berlese); S_Soil: Simpson index (Berlese).

) revealed significant associations between soil properties and biological activity indicators. SOM10cm and SOM20cm exhibited a strong positive correlation (r = 0.927, p < 0.01), suggesting consistency in SOM distribution across depths. SOM10cm and SOM20cm were also positively correlated with UR (r = 0.69 and r = 0.77, p < 0.01, respectively), SU (r = 0.37 and r = 0.52, p < 0.01), and PHO (r = 0.18 and r = 0.21, p < 0.05), indicating that SOM plays a key role in enzyme-mediated nutrient cycling. Among the enzymatic activities, UR showed significant positive correlations with SU (r = 0.53, p < 0.01), PHO (r = 0.63, p < 0.01), and GL (r = 0.57, p < 0.01), indicating that nitrogen and sulfur transformations in the soil are functionally linked to carbon degradation processes. Similarly, DH was positively correlated with SOM20cm (r = 0.487, p < 0.01) and GL (r = 0.51, p < 0.01), suggesting that microbial oxidative metabolism is influenced by organic matter availability.

In terms of soil biodiversity, the Shannon index in situ (H’10cm) showed a strong correlation with the Simpson index in situ (S10cm) (r = 0.96, p < 0.01), highlighting the consistency between the diversity metrics. Additionally, the Shannon and Simpson indices obtained from the Berlese extractions (H’Litter, H’Soil, SLitter, and SSoil) exhibited moderate to strong correlations (ranging from 0.39 to 0.93, p < 0.01), reinforcing the reliability of these metrics in evaluating soil biodiversity across different land-use systems. No significant correlations were found between ER or BR and enzymatic activities, suggesting that microbial respiration in these systems may be influenced by additional environmental factors beyond enzyme activity alone.

3.2.3. Principal Component Analysis (PCA) of Soil Biological Activity

Principal component analysis (PCA) showed that the first two axes explained 45.3% of the total variance (PC1: 26.1%, PC2: 19.2%). PC1 was mainly associated with basal respiration (BRC_CO2), litter accumulation, the Shannon diversity index (H’10cm), and soil invertebrate richness (S10cm), suggesting a strong influence of macrofaunal activity and organic inputs on this axis. PC2 was more closely related to soil organic matter (SOM10cm, SOM20cm), urease, phosphatase, and dehydrogenase activities, reflecting microbial enzymatic functions (Figure 3).

The positioning of land-use systems in the ordination space indicates partial overlap but some separation. Forest and Chakra agroforestry systems (FAFS, TAFS) clustered near variables associated with SOM and enzymatic activity, while cocoa monocultures (CMC) were located closer to the vectors BRC_CO₂ and litter. This suggests that agroforestry systems share functional similarities with forest soils, whereas monocultures exhibit a different soil biochemical profile dominated by microbial respiration and residue accumulation.

3.2.4. Regression Analysis of CO₂ Respiration

The linear regression model between the recorded ER and BR showed a negative relationship, with a regression equation of Y = 207.19 − 1.18X. This indicates a slight decreasing trend in BR as ER increases (Figure 4). However, the dispersion of data points and large residuals suggest high variability in basal respiration across different land-use types.

The residual plot reveals considerable deviation from the regression line, indicating that the model does not fully explain the variation in basal respiration. The presence of large residuals suggests that other soil factors, such as organic matter content, microbial composition, and enzyme activity, may play a significant role in determining basal respiration levels.

Although the regression line shows a negative trend, the relatively small slope (−1.18) suggests that the decline in BR with increasing ER is minimal, requiring further analysis to determine the significance of this association.

3.2.5. Locally Weighted Scatterplot (LOWESS)

Our analysis suggested that greater organic matter availability enhanced the microbial processes involved in nutrient cycling, particularly through its strong association with soil organic matter content and enzymatic activities, especially urease and phosphatase. The relationship between recorded CO₂ respiration (ERC_CO₂) and basal respiration (BRC_CO₂) was modeled using linear regression, which provided a general trend. However, the residuals displayed high variability, indicating that a simple linear approach may not fully explain the complexity of microbial respiration dynamics.

To better understand the relationship between organic matter and biological indicators, we applied a Locally Weighted Scatterplot Smoothing (LOWESS) curve (Figure 5), which revealed a non-linear trend in the data.

The curve suggests that microbial respiration fluctuates with changes in CO₂ levels, likely due to the interplay of substrate availability, microbial community composition, and enzymatic activity thresholds. Specifically, the initial decline in RBC_CO₂ at lower REC_CO₂ levels, followed by an increase and subsequent decrease, may indicate metabolic constraints or shifts in microbial efficiency as organic matter undergoes different stages of decomposition.

4. Discussion

4.1. Effects of Land-Use Systems on Soil Biological Activity

Our study revealed that forest soils exhibited higher edaphic respiration and microbial enzyme activities than agroforestry and monoculture systems, confirming their role as reservoirs of soil biological functionality. This aligns with the findings of [58], who reported enhanced biological activity in forest ecosystems due to greater organic matter inputs and diverse plant—microbe interactions. Conversely, monoculture systems were clearly differentiated and dispersed across the PCA space, reflecting reduced and inconsistent biological activity, possibly due to intensive management and low microbial diversity [58,59]. Agroforestry systems demonstrated intermediate levels of biological activity, suggesting that integrating trees with crops can mitigate some of the negative impacts of monocultures on soil health [7,14], clustering closer to forest conditions in the PCA ordination (Figure 3).

4.2. Soil Macroinvertebrate Diversity and Functional Groups Across Land-Use Types

The diversity and abundance of soil macroinvertebrate were significantly influenced by land-use changes. Forest soils supported a higher diversity of functional groups, including soil engineers, such as Lumbricidae and Formicidae, which play crucial roles in soil structure and nutrient cycling. Agroforestry systems maintained a moderate diversity of macroinvertebrates, while monocultures exhibited the lowest diversity. These observations are consistent with the findings of [60], who noted that increased land-use intensity leads to a reduction in the density and diversity of edaphic macroinvertebrates due to habitat alteration and decreased resource availability.

4.3. Functional Linkages Between Soil Organic Matter and Biological Indicator

These findings are consistent with previous research that has reported non-linear soil respiration responses to organic inputs and environmental changes [58,61,62]. For example, studies examining soil management practices have demonstrated that variations in enzyme activity and microbial biomass—critical factors influencing respiration—are closely tied to nutrient availability [63]. This suggests that at low REC_CO₂ levels, substrate limitation or suboptimal enzyme activity may restrict respiration, while changes in substrate quality at higher CO₂ concentrations could temporarily enhance microbial activity before metabolic thresholds are reached. Similarly, long-term fertilization studies have shown that microbial biomass phosphorus and related soil chemical properties exhibit complex, non-linear relationships with soil respiration. These results imply that shifts in nutrient dynamics and microbial efficiency, driven by long-term organic inputs, may underlie the observed oscillatory behavior of respiration rates. Furthermore, research on the Loess Plateau has demonstrated that integrated organic and inorganic fertilization can modify soil aggregate stability and nutrient dynamics, thereby influencing microbial access to organic matter and altering respiration patterns [64]. This supports the idea that structural changes in soil, in addition to biochemical factors, play a role in the non-linear responses observed in our study. These studies in our discussion suggest that microbial respiration is not solely a linear response to substrate availability but is modulated by a range of interacting factors, including nutrient status, enzyme activity, and soil structural properties. Future research should aim to delineate the threshold levels of substrate availability and enzymatic activity that drive these metabolic shifts, ultimately refining the predictive models of soil respiration under variable environmental conditions.

The inclusion of both linear regression and LOWESS analyses is essential as the linear model provides a simplified relationship, while LOWESS uncovers more complex underlying trends that would otherwise be overlooked. In heterogeneous tropical soils, microbial respiration patterns are influenced by microclimatic fluctuations, SOM decomposition rates, and spatial variations in microbial activity, making non-parametric approaches such as LOWESS particularly relevant [65].

The observed non-linearity in microbial respiration underscores the need for integrating flexible, non-parametric techniques alongside traditional regression models when analyzing soil biological processes. This approach allows for a more nuanced interpretation of soil carbon dynamics, which is crucial for optimizing microbial activity and carbon sequestration in tropical agroforestry systems. Understanding these patterns is vital for designing land-use strategies that enhance soil functionality and resilience under changing environmental conditions.

4.4. Implications for Sustainable Soil Management and Agroforestry Practices

The intermediate levels of soil biological activity and macroinvertebrate diversity observed in agroforestry systems emphasize their potential as a sustainable land-use strategy that integrates ecological functionality and agricultural productivity. Agroforestry enhances soil structure, organic matter accumulation, and microbial diversity, thereby improving nutrient cycling and carbon sequestration while mitigating the soil degradation risks associated with intensive agriculture [66,67].

Evidence from a meta-analysis by [58] demonstrates that agroforestry significantly enhances biological activity, soil biodiversity, and key ecosystem functions in Latin American mountain agroecosystems. Similar findings have been reported in tropical landscapes, where the presence of trees in agricultural matrices increases soil microbial biomass, enzymatic activity, and macroinvertebrate diversity compared to conventional monocultures [22,59,68]. These effects are largely attributed to higher organic matter inputs from litterfall, root exudates, and rhizosphere interactions, which promote beneficial microbial communities and soil fauna that drive decomposition and nutrient cycling processes [24].

Given the accelerating trend of agricultural intensification in tropical regions, agroforestry has emerged as a viable solution for maintaining soil health and ensuring long-term productivity. However, its effectiveness depends on site-specific management strategies, tree species selection, and integration with local agroecological practices [69]. Future research should prioritize optimizing agroforestry configurations to promote belowground biodiversity and improve ecosystem services, ensuring economic viability and ecological resilience in the face of climate change.

5. Conclusions

This study provides novel evidence that land-use systems significantly shape soil biological activity, macroinvertebrate diversity, and enzymatic processes in the Ecuadorian Amazon. Forest soils exhibited the highest microbial respiration and enzymatic activity, while agroforestry systems maintained intermediate levels of soil biological function, surpassing monocultures in microbial diversity and macroinvertebrate abundance. The strong correlations between soil organic matter and key enzymatic activities (urease, phosphatase, and arylsulfatase) highlight the vital role of organic inputs in sustaining microbial-driven nutrient cycling. Furthermore, the non-linear relationship between CO₂ respiration (ERC_CO₂) and basal respiration (BRC_CO₂), as revealed by the LOWESS curve, underscores the complexity of microbial responses to soil conditions, emphasizing the need for multi-model analytical approaches in soil ecological studies.

Agroforestry has emerged as a viable and sustainable alternative to intensive monocultures, offering a balance between soil fertility conservation and agricultural productivity. By integrating tree species into agricultural landscapes, agroforestry enhances carbon sequestration, microbial diversity, and soil structure, thereby mitigating the degradation risks associated with conventional farming. However, its long-term effectiveness depends on species selection, management strategies, and soil conservation practices. Future research should focus on optimizing agroforestry systems to maximize belowground biodiversity and carbon storage, thereby reinforcing their role in climate change mitigation and sustainable land management in tropical ecosystems.