Nitrogen and Sulfur Cycling in Diverse Farm Ages and Ecological Zones Under Agricultural Expansion

Abstract

Background: Agriculture degrades soils, affects the delivery of ecosystem services, and contributes to climate change. Methods: This research examined nitrogen and sulfur recycling in soils under cropland expansion in Ghana at (a) reconnaissance scale in northern Guinea savannah (NGS), southern Guinea savannah (SGS), forest–savannah transition (FST), and semi-deciduous forest (SDF) agro-ecological zones (AEZs), and (b) farm level in rain Forest and the FST AEZs based on “duration of cultivation”. Fresh soils (20 cm depth) were incubated for 28 days at 28 °C, followed by the determination of mineralized nitrogen and sulfur at 14 and 28 days using standard methods. Results: Low nitrogen and sulfur contents led to predominant nitrogen and minor sulfur immobilizations, particularly in FST and savannah AEZs. Microbial biomass and pedogenic Fe controlled much of the nitrogen immobilization. At the farm level, dithionite Al and soil pH controlled nitrogen immobilization, particularly in relatively older farms, being pronounced in forest-related AEZs. Conclusions: Although the study is laboratory-based, it highlights the severe nature of soil degradation (SD) under cropland expansion in regions prone to poor nutrient budgets. Therefore, it calls for drastic measures to halt SD by adopting ecozone- and climate-driven sustainable soil management and agricultural systems.

1. Introduction

Growing populations around the world require unprecedented amounts of food, yet widespread soil degradation significantly threatens food security in many regions [1,2]. Agriculture, as a major anthropogenic stressor, has global multifaceted impacts on ecosystem services [3,4]. This challenge is acute in Sub-Saharan Africa (SSA), where low-input farming systems and rapid nutrient depletion dominate [5,6]. These perturbations risk undermining soil ecosystem resilience and impairing the provision of supporting, provisioning, regulating, and cultural services. As a result, SSA is witnessing enormous cropland expansion into natural habitats as communities seek fertile soils to address food insecurity [7,8]. In Ghana, cropland expansion averaged 2.7% per year over the past three decades [7]. Environmental degradation driven by illegal mining and land mismanagement now poses a pressing existential challenge [7,8,9]. Changes in land use and management directly affect soil organic carbon (SOC) stocks and nutrient cycles, influencing crop productivity [10,11]. Therefore, Ghana faces significant challenges in achieving Sustainable Development Goals (SDGs) 1, 2, and 3 amidst the changing climatic conditions.

Research has demonstrated significant impacts of agricultural or cropland expansion on below- and above-ground biodiversity, soil properties [12,13,14,15], ecosystem productivity [16], and soil and nutrient exports via surface runoff [17]. However, many of them have been skewed towards global and regional scales. Local-scale and plot-level studies on soil dynamics across agro-ecological zones (AEZs) are largely limited. Recent cropland expansion studies in two AEZs reported that over 10% SOC declined along with microbial properties during the first three years of cultivation in Ghana’s forest AEZ [12,13]. These findings highlighted the immense effects of soil organic matter (SOM) on other physical, chemical, and biological soil properties. It is unclear how these effects could translate to nutrient cycling and release, ecosystem functions, and soil ecosystem resilience. Understanding the degree, direction, and changing aspects of these drivers of soil ecosystem services is necessary for sustainable soil management in the AEZs. This resonates with the fact that distinct ecological processes govern biological dynamics under changing conditions [18].

Soil nutrients are either from geogenic or biogenic sources, but those of biogenic origin must be transformed into plant-available forms for plant uptake, thus making soil nutrient cycles the core of supporting and provisioning services, without which human survival is impossible. It is estimated that plants require about seventeen nutrient elements for growth and productivity, with 82% obtained directly from the soil, depending on the quantities required [19]. These include primary essential or macronutrients, secondary nutrients required in moderate quantities, and micronutrients [20,21]. Nutrient cycling is an indicator of soil health [22,23,24], providing most of the elements needed for crop productivity, e.g., carbon (C), nitrogen (N), phosphorus (P), and sulfur (S), and the other nutrients from SOM. Soil nutrient cycling, or biogeochemical cycles, are interconnected processes necessary for ecological stability and productivity [25,26].

Over the past decades, N and P mineralization and fertilization have dominated nutrient cycling research [27]. This overlooked and overshadowed research on S (a secondary nutrient) and its cycling until recently [28,29], when S nutrition and mineralization research became relevant owing to (a) low inputs and widespread deficiencies [30] and (b) key roles S plays in plant nutrition and productivity [31]. Sulfur is critical and important, such that the Sulphur Institute named it the “fourth element” or “an advantaged element” [32], while Rasheed et al. [33] named it the second crucial nutrient after N for the excellent growth of oil crops. Nitrogen and S have become key drivers of crop yields worldwide due to the close N-S functions in cellular processes and secondary metabolism [34,35,36]. A previous review reported that the combined application of micronutrients along with S improved crop yields by an average of 25% across SSA [37]. Other studies in temperate regions revealed that S increased wheat and rice yields by 26% and 10%, respectively [29]. In SSA, S application increased cereal and legume yields by margins dictated by crop type and application rates [38]. For instance, maize yield increased by 20–125% (0–10 kg S ha⁻¹) and 261% (21–30 kg S ha⁻¹), wheat by 60–80% (11–20 kg S ha⁻¹), rice by 26–62% and 214% (31–40 kg S ha⁻¹), sorghum and millet by 24–31% and 14–15% (0–10 kg S ha⁻¹), respectively, soybean by 25% (21–30 kg S ha⁻¹), canola by 17%, sesame by 13%, and common beans by 2% (31–50 kg S ha⁻¹) [38]. An average of 16% was obtained from 58 harvests of No-Till fields in Brazil [39]. Variable effects of S on yield increases have been attributed to factors such as climate (particularly precipitation) and initial soil properties (e.g., pH, SOM, and oxide contents) [31,38,40].

Recognizing the key roles of N and S in food systems makes it imperative to understand their cycling under cropland expansion with different cultivation periods. Therefore, this study focuses on N and S mineralization for the following reasons: (i) N is an indispensable nutrient required in larger quantities [41] due to its key role in the plant pigment chlorophyll and other biochemical processes; these include the synthesis of amino acids, nucleotides, and enzymes, as well as hormonal regulations in plant growth and development [10]; (ii) over 90% of total soil S comes from soil organic matter (SOM) [40,42,43], which must be mineralized for plant uptake; (iii) there are similarities in N and S cycling processes [40] and requirements in biochemical processes [33,44]; and (iv) S research, particularly its dynamics in SSA, is still very limited, although S depletion in highly weathered SSA soils is pronounced and widespread [38]. The similarity in biological requirements makes S the second crucial nutrient after N in the growth, development, and protein contents of grains [33]. The objective of this study was to investigate the patterns of N and S mineralization at AEZs and farm levels under prevalent cropland expansion in Ghana through the following questions: (a) What are the N and S cycling patterns at the AEZs and farm levels? What factors drive the patterns? (b) What mechanisms may explain the patterns? The findings of this study are expected to provide demand-driven recommendations of appropriate management practices to meet crop demand while minimizing environmental consequences. The results of this study are expected to provide essential information to guide soil nutrient management under various cropping systems and patterns.

2. Materials and Methods

2.1. Study Design, Site Description, and Soil Sampling

The study was conducted at two levels: (i) reconnaissance survey level in May 2019 from north to south of Ghana in four agro-ecological zones (AEZs) and (ii) farm levels involving two AEZs in May and June 2021 (Figure 1). The four AEZs considered for the reconnaissance survey level include the northern Guinea savannah (NGS), southern Guinea savannah (SGS), forest–savannah transition (FST), and semi-deciduous forest (SDF) zones of Ghana (Figure 1). The AEZs and specific sites were selected as part of the collaborative UK Research and Innovation (UKRI)-funded Sentinel (Social and Environmental Trade-offs in African Agriculture) research project (https://www.sentinel-gcrf.org/). In summary, the key criteria used to identify study sites are as follows:

• Proximity to natural habitats that still provide a range of ecosystem services and a potential for agricultural expansion;
• Evidence of agricultural expansion, partially driven by food crop production, and largely dominated by smallholder farmers;
• Agro-ecological contrast and natural habitat lost to agriculture within a 10–20 km² area;
• Proximity to road networks or settlements, combined with communities’ readiness to work with the project team.

The Guinea savannah has a unimodal mean annual rainfall of 900–1200 mm with a single growing season. Rainfall increases towards the south in the transition and forest zones (1500–2200 mm per annum) with bimodal peaks that enhance major and minor growing seasons. Conversely, temperatures are higher (mean range of 28 to 29 °C) in the savannah than in the southern AEZs (mean of 27.3 °C). These are based on analyses of data from the Ghana Meteorological Agency between 1981 and 2010 and 1989 and 2015 as reported by Yamba et al. [45,46]. The relatively heavy rainfall in the southern forest-related AEZs yields denser vegetation cover comprising native forests with deciduous and evergreen tree species such as Cynometra ananta Hutch. and Dalziel, Lophira alata Banks ex Gaertn., Tarrietia utilis Sprague, Ceiba pentandra (L.) Gaertn., Celtis mildbraedii Engl., Triplochiton scleroxylon K. Schum., and Ficus spp. [47,48]. The forest vegetation co-exists with tree crop plantations such as Theobroma cacao, Hevea brasiliensis Müll. Arg., Elaeis guineensis Jacq., Musa spp., etc. In the savannah AEZs, the vegetation is open and dominated by tall grass species such as Megathyrsus maximus (Jacq.) B.K. Simon & S.W.L. Jacobs (formerly Panicum maximum), Andropogon gayanus, Kunth., and Pennisetum purpureum (Schumach.) Morrone. These coexist with high-value trees such as Parkia biglobosa (Jacq.) R.Br. ex G. Don, Vitellaria paradoxa C.F. Gaertn., Adansonia digitata L., and Ceiba pentandra (L.) Gaertn.

Figure 1. Study sites showing the AEZs and farm-level locations (Credit: Open Street Map under the Open Database License).

Figure 1. Study sites showing the AEZs and farm-level locations (Credit: Open Street Map under the Open Database License).

The second-level study was conducted at the farm level in the rain forest (RF) AEZ (Dompem: 5°9′33.7″ N; 2°4′29.4″ W) and FST (Adansam: 7°50′35.9″ N; 1°45′59.9″ W) (Figure 2), which was a product of the reconnaissance survey. We aimed to investigate the influence of farmers’ resources (wealth categories) and the “duration of cultivation” on soil health. Wealth categories were identified using local criteria, including farm size, the number of farm plots for different crops, the diversity of agrochemicals used, the type of house built, the ability to purchase vehicles such as bicycles, motorcycles, or tricycles, pay school fees, and establish alternative livelihoods, such as owning a retail shop. These include (a) Wealthy, (b) Moderately wealthy, and (c) Poor, each having twenty (20) farms for each AEZ. Under each category, there were four (4) farms each under the “duration of cultivation” with their designations such as (i) native vegetation cultivated for one year (“Forest farms”), (ii) fallows cultivated for one year (Fallows), (iii) farms cultivated for three years since clearing (Three years), (iv) farms cultivated for five (Five years), and (v) farms cultivated for ten years (Ten years) (Figure 2). The FST AEZ farms (Adansam farms) were dominantly cultivated fallows and were designated as “One-year” farms, while the RF AEZ (Dompem) had both newly cleared forest and fallow farms. Sixty (60) farmers from each AEZ (twenty (20) farmers per wealth category) were identified before sampling.

The Ethics Committee for the College of Basic and Applied Science (ECBAS) of the University of Ghana issued ethical clearance for the study. At the community levels, permissions and/or consent were granted by the district Agricultural Extension Officers and the individual farmers whose farms were sampled. Survey-level farms were randomly sampled without any consideration for the duration of site cultivation. In all farms, soil was sampled randomly at five locations per farm up to a depth of 20 cm, then composited into one sample per farm. Samples for physicochemical analysis were composited, while samples for biogeochemical analyses were stored at 4 °C until use.

2.2. Laboratory Procedures

2.2.1. Soil Analyses

Standard laboratory analysis of basic soil properties was conducted in triplicate at the Department of Soil Science, University of Ghana, and the Soil Research Institute of the Council for Scientific and Industrial Research, Kumasi. Analytical procedures for basic properties of the soils can be found in the works of Neina and Agyarko-Mintah [12,13] and Neina and Adolph [43]. Some of the properties of relevance to this study that were previously measured [13] include labile C or permanganate oxidizable C (POXC) measured using the Blair et al.’s [49] method modified by Weil et al. [50]; microbial biomass measured by the chloroform fumigation–incubation method [51] (because the fumigation–extraction method did not work for the soils); and free aluminum and iron oxides extracted using dithionite–citrate solution (Al_d, Fe_d) and acid oxalate solution (Al_ox, Fe_ox) [52]. Both categories of oxides are described here as “pedogenic minerals” or “pedogenic” Al and Fe. The water-holding capacity was determined using a laboratory gravimetric method that involved weighing, wetting by capillarity, draining, oven-drying, and final weighing before calculation.

2.2.2. Nitrogen and Sulfur Mineralization

The SOC contents of soils sampled varied widely (Adansam = 0.3–1.6%, Dompeme = 0.9–6.8%), with many farms (farm level study) having <0.5%. These constitute 42 Adansam farms and 47 Dompem farms out of 57 and 49 farms obtained from the field. The SOC contents affected total S detection by the CNS analyzer (CNS-2000 Analyzer, LECO Corporation TruMac Series, St. Joseph, MI, USA) in the soils. The detection limit of S had a narrow range of 0.02–50 mg, compared to those of C (0.02–200 mg) and N (0.02–300 mg). Additionally, previous studies on the soils did not show any significant effects of wealth categories on soil quality and health [12]. Considering that total S content in healthy plants ranges from 0.1 to 0.4% (dry weight basis) [40], S content can be negligible in soils with very low SOC contents, particularly those of SSA [53]. To avoid analytical challenges, only soils with over 0.5% SOC were used for this study. For incubation, two sets (for the two sampling times) of fresh soils (100 g) per farm were wetted to 60% water-holding capacity, placed in glass jars (1.5 L) and incubated in the dark at a mean temperature of 28 °C for 28 days. To reduce extreme moisture loss, glass vials containing distilled water were placed in the glass jars for the entire period. The temperature was chosen to capture the mean ambient temperature for all study sites and the optimum temperature range (25–35 °C) for nitrification [54,55]. The duration was used to target nutrient supply during the early growth period of most crops. Additionally, 62–74% of S is found to mineralize within the first 14 days [56]. Samples were taken at 14 and 28 days to extract mineral N (NH₄⁺ and NO₃⁻) using 40 mL 0.5 M K₂SO₄ at an extract-to-soil ratio of 4 to 1 after 30 min of shaking [57,58]. The extracts, together with those of day zero (initial N), were analyzed for NH₄⁺ and NO₃⁻ using the Kjeldahl distillation method. To extract sulfate, all receptacles and filter papers were rinsed with 0.1 M HCl to prevent contamination with traces of S that could be present. Sulfate was extracted using Ca(H₂PO₄)₂ solution at an extractant-to-soil ratio of 2.5:1 and shaking for 30 min at ≥200 rpm [59]. The sulfate contents of the extracts were determined by Ion Chromatography (Shimadzu HPLC IC 20A System, Kyoto, Japan).

2.3. Statistical Analysis

The exclusion of soils with low SOC resulted in ten to eleven replicate farms for FST AEZ (Adansam), while RF AEZ (Dompem) had 9 to 10 used for statistical analysis. In this case, parametric analyses were not possible. Therefore, non-parametric Kruskal–Wallis tests were used to examine whether the ecological zones and duration of cultivation had significant effects on N and S mineralization. The Mann–Whitney U test was used for mean comparisons at 5% level of significance. Non-parametric Spearman correlations were used to examine relationships between soil properties and mineralization for normal and non-normal data. All statistical analyses were carried out in SPSS Version 20 (IBM^® SPSS^® Statistics, New York, NY, USA), while Sigma Plot 12 (Systat Software Inc., San Jose, CA, USA) was used to plot the graphs.

3. Results

3.1. Physical and Chemical Properties of the Soil

Key soil properties of the study sites (except total N) in

Table 1. Summary of relevant soil properties of the AEZs and the farm types used in the study.

Site/Farm	Silt + Clay	pHwater	POXC	Ald	Alox	Fed	Feox
	(%)		(mg kg−1)	(mg kg−1)
Agro-ecological zone
Northern Guinea Savannah (NGS)	6.33	6.28	28.0	25.7	17.6	55.4	3500
Southern Guinea Savannah (SGS)	7.50	5.63	33.1	24.7	17.0	45.3	960.3
Forest–Savannah Transition (FST)	11.7	6.28	73.7	24.6	16.9	9.27	621.9
Semi-Deciduous Forest (SDF)	9.25	6.21	46.7	24.8	16.7	28.5	454.8
Adansam (FST AEZ)
One year	10.1	6.28	47.3	39.4	13.9	44.9	498.1
Three years	10.4	6.23	38.0	28.7	15.8	103.8	507.5
Five years	8.64	6.40	36.6	24.9	16.5	16.0	563.4
Ten years	10.9	6.35	38.9	25.0	16.3	31.5	510.8
Dompem (RF AEZ)
Forest (One year)	26.6	4.28	64.8	27.2	22.9	450.2	2477
Fallow (One year)	28.8	4.25	67.0	26.7	19.7	423.4	1427
Three years	28.7	4.36	53.9	34.6	18.0	229.5	2249
Five years	23.6	4.53	65.1	28.8	22.3	258.4	1293
Ten years	19.4	4.53	52.0	34.1	20.0	309.6	1955

Data source: Adopted from Neina and Agyarko-Mintah [12,13]. POXC: permanganate oxidizable C, Al_d: dithionite extractable aluminum, Al_ox: oxalate extractable Al, Fe_d: dithionite extractable iron, Fe_ox: oxalate extractable Fe.

, total C (

Table 2. Means (standard error of means, SE) of total C and N, and N mineralization in soils from different AEZs and specific farm types of Adansam and Dompem. Data in columns followed by different letters depict significant differences (5% significance level).

Site/Farm	Total C	Total N	Mineral N Day 0	Nmin Day 14	Nmin Day 28
	(g kg−1)		(mg kg−1)	(mg kg−1)
Agro-ecological zone
Northern Guinea Savannah (NGS)	10.8 (3.42) b	1.13 (0.14) a	53.5 (5.45)	47.6 (1.19) d	41.6 (9.73) c
Southern Guinea Savannah (SGS)	8.90 (1.06) c	1.21 (0.06) a	56.2 (8.78)	54.4 (1.71) c	52.6 (3.68) b
Forest–Savannah Transition (FST)	17.03 (2.48) a	0.76 (0.06) b	73.6 (8.04)	62.7 (4.03) b	56.3 (4.57) b
Semi-Deciduous Forest (SDF)	5.86 (0.66) d	0.65 (0.16) b	62.9 (9.79)	74.0 (3.70) a	72.8 (1.24) a
p-value	0.020	0.033	0.209	0.015	0.022
Adansam (Forest–Savannah Transition)
One year	12.2 (0.63) a	1.02 (0.07) a	57.8 (1.08) c	52.3 (1.08) b	60.8 (1.33)
Three years	10.3 (0.70) a	0.74 (0.06) a	141.6 (1.14) a	56.5 (1.05) b	63.8 (2.69)
Five years	8.79 (0.69) b	0.49 (0.05) b	79.8 (1.16) b	61.0 (1.08) b	60.1 (2.67)
Ten years	10.9 (0.73) a	0.81 (0.07) a	66.4 (1.11) b	80.4 (1.06) a	62.3 (1.22)
p-value	0.012	0.012	0.001	0.001	0.388
Dompem (Rain Forest)
Forest (One year)	22.14 (1.15)	0.85 (0.33) b	65.82 (3.26)	94.57 (0.06) b	97.23 (7.43) a
Fallow (One year)	21.09 (1.09)	2.02 (0.13) a	66.89 (3.53)	110.36 (0.18) a	81.35 (5.32) b
Three years	19.25 (1.17)	1.67 (0.29) b	85.44 (9.50)	88.43 (0.18) b	94.39 (9.05) a
Five years	16.52 (1.07)	1.36 (0.09) b	88.54 (10.47)	72.21 (0.04) c	88.28 (8.36) b
Ten years	15.64 (1.13)	1.56 (0.22) b	83.29 (8.36)	88.55 (0.34) b	68.98 (4.55) c
p-value	0.103	0.003	0.195	0.006	0.058

), and sulfate (Table 4) have been reported in previous studies [12,43]. However, a summarized version of relevant properties is presented in Table 1. Generally, soils of the northern sector are coarser than those of the south. Consequently, soils of the savannah AEZs are coarse, slightly acidic, have low SOC, N, permanganate oxidizable C (POXC), and pedogenic Al, but relatively high pedogenic Fe levels (Table 1 and Table 2). An exception is the total N of the savannahs, which had 1.5–1.9 times (p = 0.033) the N contents of the SDF and FST AEZs (Table 2). For all the AEZs, the mean pH values were below the critical level of 5.5.

At the farm level, the total C contents were high in the forest and fallow farms of the Dompem (RF zone) soils, showing a decreasing trend (from forest and fallow farms) of 5 to 14% with no significant differences among them. These values were 1.4 to 1.9 times the total C contents of Adansam (FST zone) soils. The mean total N for both sites differed significantly (p < 0.05), where the Dompem soils had 1.4 to 2.8 times more N. In the Dompem soils, mean N decreased by 17 to 19% with the duration of cultivation from fallow to five-year farms, and increased again in ten-year farms (Table 2). This was contrasted by a steep decrease of 27 to 34% (p = 0.003) in Adansam soils up to the five-year mark, and increased again in year ten.

3.2. Nitrogen Mineralization

3.2.1. Mineralization Trends and Mineral N Species Across the Zones

Before incubation, NH₄⁺ contents, indicative of ammonification, were almost twice those of NO₃⁻ (Figure 3). These accounted for 6–11% of the total N, with relatively higher mineral N contents observed in the wet AEZs (i.e., FST and SDF) (Table 2). After 14 days of incubation, NH₄⁺ content decreased by 0.9 to 15 mg kg⁻¹ in all AEZs except for the NGS zone, with the largest reduction occurring in the FST AEZ. At this stage, both mineral N species represented 5–9% of total N, with no differences (p > 0.05) across the AEZs. After 28 days, NH₄⁺ content decreased further by 4.8 to 9.9 mg kg⁻¹ in all AEZs, with a more pronounced reduction in the wet zones than in the NGS AEZ. This reduction resulted in only minor nitrification in the wet zones (i.e., FST and SDF) and no reduction in the savannah AEZs. By the 28th day, NO₃⁻ content decreased by 1.2 mg kg⁻¹ only in the NGS zone (Figure 3). The N mineralization (N_min) trend across the AEZs differed from total N contents (Table 2), with the N_min of the relatively wet zones exceeding (p < 0.05) those of the savannahs (Table 2). By the 14th day of incubation, 15% net N mineralization (difference between the mineral N content before and after incubation) occurred only in the SDF zone, whereas net N immobilization of 3%, 12%, and 17% occurred in the NGS, SGS, and FST zones, respectively. By the 28th day, N immobilization of 12%, 10%, 3%, and 2% occurred in NGS, FST, SGS, and SDF zones, respectively (Table 2). Notably, the highest absolute values of mineral N were consistently immobilized in the soils of the NGS and FST zones. These trends corresponded with strong positive correlations between day 14 N_min and POXC, sulfate, the sum of exchangeable bases (SEB), and effective cation exchange capacity (ECEC), and between N_min day 28 and POXC, SEB, ECEC, and sulfate 28 (

Table 3. Spearman correlation coefficients for N mineralization in soils of the AEZs and farms of Adansam and Dompem.

Soil Properties	Agroecological Zones		Adansam (FST)		Dompem (RF)
	Day 14	Day 28	Day 14	Day 28	Day 14	Day 28
Permanganate oxidizable C	0.70 **	0.64 *	-	-	0.35 *	0.42 **
Total N	−0.71 **	-	-	-	0.38 **	-
Smin day 14	0.78 **	0.60 *	-	-	-	-
Smin day 28	-	0.77 **	-	-	-	-
Microbial biomass	−0.67 **	−0.57 *	-	-	0.32 *	-
Basal respiration	-	-	0.49 **	-	-	-
Soil pH	-	-	-	-	−0.33 *	-
Dithionite extractable Al	−0.51 ns	−0.22 ns	−0.45 **	-	−0.41 **	-
Dithionite extractable Fe	−0.81 ***	−0.72 **	-	-	-	-
Oxalate extractable Fe	−0.49 ns	-	-	-	-	-
Sum of basic cations	0.59 *	0.66 *	-	-	-	0.44 **
Effective CEC	0.58 *	0.65 *	-	-	0.43 **	0.40 **
Silt + clay	0.43 ns	-	-	-	0.29 *	0.49 **
Clay	0.42 ns	-	-	-	-	0.37 *

Microbial indices were measured in a previous study [13]; *** p < 0.001; ** p < 0.01; * p < 0.05; ns: not significant (only for moderate r values).

). Conversely, N_min day 14 correlated negatively with total N, microbial biomass (C_mic), dithionite extractable Fe (Fe_d), and oxalate extractable Fe (Fe_ox), whereas N_min day 28 correlated negatively with C_mic and Fe_d (Table 3). Overall, more soil properties correlated with AEZ-level N_min than those at the farm level.

3.2.2. Mineralization Trends in the Farms

Before incubation, the Adansam soils had 12.3 mg kg⁻¹ more mineral N than the Dompem soils. The nitrate content was 5.7 to 16 mg kg⁻¹ higher than NH₄⁺, showing significant differences (p ≤ 0.002) among the farms except for the three-year farms. After 14 days of incubation, NH₄⁺ content exceeded that of NO₃⁻ by 8.3 and 10 mg kg⁻¹ in the one- and three-year farms, respectively (Figure 4). Ammonium immobilization of 129% and 14% occurred in the three- and five-year farms, respectively. Nitrate immobilization of 43%, 212%, and 57% occurred in the one-, three-, and five-year farms. By the 28th day, all farms experienced NH₄⁺ immobilization except for the five-year farms, while NO₃⁻ immobilization occurred only in the relatively older farms (i.e., five and ten years). At day 14, the mean N_min in Adansam soils was 28.3 mg kg⁻¹ less than in the Dompem soils, but reduced to 22 mg kg⁻¹ by day 28. After 14 days of incubation, N_min was highest in the ten-year farms and differed (p = 0.001) among farm types (Table 2). About 10, 150, and 31% N was immobilized in the one-, three-, and five-year farms, respectively. The N_min positively correlated with basal respiration, but negatively with dithionite extractable Al (Al_d) (Table 3). By the 28th day, net N_min of 14 and 11% (p = 0.355) occurred in one- and three-year farms, respectively. Surprisingly, the 14-day net N_min of the five- and ten-year farms resulted in net immobilization of 1 and 29%, respectively, within this period. The N_min of Adansam soils had the least number of correlations with other soil properties.

The initial mineral N contents of Dompem soils were about 20 mg kg⁻¹ higher in the three-to-ten-year farms. The fallow farms had more NO₃⁻ than their forest counterparts. The mean NH₄⁺ in all farms was consistently higher than NO₃⁻ by 9.5, 19.6, and 16.7 mg kg⁻¹ before incubation, at 14 and 28 days, respectively (Figure 5). There were no clear trends (either increasing or decreasing) or significant differences (p < 0.05) in both mineral N species of the farm types, except for the initial NO₃⁻ (p = 0.002) and 28th-day NH₄⁺ (p = 0.037).

Ammonification of 23 to 36% occurred in all farms after 14 days, except immobilization of 4% NH₄⁺ in the five-year farms. By the 28th day, NH₄⁺ immobilization of 7%, 24%, and 46% occurred in the fallow, three- and ten-year farms. Before incubation, soils of the relatively older Dompem farms had more NO₃⁻ contents. These were immobilized by day 14, followed by minor nitrification at 28 days, except for the fallow farms. After 14 days of incubation, net N_min occurred in all soils of Dompem farms except for the five-year farms, where 18% N was immobilized (Table 2). The highest N_min (p = 0.006) occurred in the fallow farms, followed by forest, ten-year, three-year, and five-year farms, constituting 0.7 to 3.4% of total soil N. The N_min had a positive correlation with total N, C_mic, and ECEC, but a negative correlation with soil pH and Al_d. By day 28, 26% and 22% N were immobilized in the fallow and ten-year farms, respectively, whereas net N_min of 2.7, 6, and 16 mg kg⁻¹ occurred in the forest, three, and five-year farms, respectively. Day 28 N_min positively correlated with POXC, SEB, ECEC, clay content, and silt + clay contents (Table 3).

3.3. Sulfur Mineralization

In the soils of the AEZs, the mean initial sulfate contents [43] were very low (7.2 to 13 mg kg⁻¹) with no differences among the AEZs. The SDF had about 3.6 to 5.8 mg kg⁻¹ more sulfate than the other AEZs. After 14 days of incubation, the net S mineralization, S_min (difference between initial and final sulfate contents), constituting four- to 13-fold of the initial sulfate contents, occurred in all AEZs, showing significant differences (p = 0.014) among the AEZs (

Table 4. Initial sulfate contents in soils of the four agro-ecological zones, Adansam and Dompem, on 14 days and at 28 days of incubation. Data in columns followed by different letters depict significant differences (5% significance level).

Zone/Farm	Initial	Day 14	Day 28
	(mg kg−1)
Agro-ecological zone
Northern Guinea Savannah (NGS)	8.42 (0.40)	34.4 (5.97) d	17.4 (1.06) c
Southern Guinea Savannah (SGS)	9.33 (1.34)	51.9 (4.49) c	42.6 (1.46) b
Forest–Savannah Transition (FST)	7.19 (0.13)	93.5 (14.54) b	41.4 (1.09) b
Semi-Deciduous Forest (SDF)	13.0 (2.14)	114.4 (16.56) a	122.1 (1.35) a
p-value	0.092	0.014	0.044
Adansam (FST AEZ)
One year	5.39 (1.10) b	61.8 (2.98)	66.1 (0.01) a
Three years	9.24 (1.18) a	62.5 (3.10)	68.9 (0.05) a
Five years	12.3 (1.14) a	63.2 (3.01)	60.2 (0.01) b
Ten years	9.51 (1.11) a	64.8 (3.99)	58.9 (0.02) b
p-value	0.001	0.723	0.017
Dompem (RF AEZ)
Forest (One year)	11.13 (0.04) a	112.3 (1.10) a	105.1 (1.12) a
Fallow (One year)	11.17 (0.05) a	53.0 (1.19) c	98.7 (1.10) a
Three years	12.24 (0.06) a	64.8 (1.15) b	70.9 (1.15) b
Five years	11.58 (0.06) a	55.9 (1.08) b	64.9 (1.05) b
Ten years	5.68 (0.04) b	59.4 (1.05) b	66.6 (1.06) b
p-value	0.019	0.002	0.001

). The S_min of the wet zones was about two- to 3.9-fold that of the savannahs. By day 28, 56%, 18%, and 49% of the sulfate content were immobilized in all zones (p = 0.044) except for net S_min of 7% in the SDF AEZ (Table 4). The 14-day S_min had positive correlations with POXC and 14-day N_min, while the reverse was observed with total N, MBC, and pedogenic Fe (

Table 5. Spearman correlation coefficients for S mineralization in soils of the AEZs and farms of Adansam and Dompem.

Soil Properties	Agroecological Zones		Adansam (FST)		Dompem (RF)
	Day 14	Day 28	Day 14	Day 28	Day 14	Day 28
Permanganate oxidizable C	0.78 **	0.71 **	-	0.47 **	-	0.41 **
Total C					0.38 **	0.50 ***
Total N	−0.75 **	-	-	-	-	-
Nmin day 14	0.78 **	0.70 **	-	-	-	0.45 **
Nmin day 28	-	0.77 **	-	-	-	-
Microbial biomass	−0.55 *	−0.41 ns	-	-	-	0.30 *
Basal respiration	-	-	-	-	-	-
Soil pH	-	-	-	-	-	-
Dithionite extractable Al	-	-	-	-	-	−0.56 ***
Dithionite extractable Fe	−0.94 ***	−0.65 *	-	0.40 **	-	0.29 *
Oxalate extractable Fe	−0.67 **	-	-	-	−0.26 ns	0.30 *
Sum of basic cations	-	0.57 *	-	-	0.41 **	0.27 ns
Effective CEC	-	0.56 *	-	-	0.36 *	0.42 **
Silt + clay	-	-	-	-	-	0.30 *
Clay	-	-	-	-	-	-

*** p < 0.001; ** p < 0.01; * p < 0.05; ns: not significant (only for moderate r values).

). The 28-day S_min correlated positively with POXC, N_min, SEB, and ECEC, but negatively with Fe_d (Table 5). Generally, the correlation coefficients were stronger here than those encountered at the farm level. The initial sulfate contents of the Adansam soils ranged from 5.4 to 12.3 mg kg⁻¹, with the lowest content in the one-year farms (p = 0.001). By the 14th day of incubation, net S_min was observed, amounting to 4.2- to 10.5-fold the initial contents with no relationships with other soil properties. By the 28th day of incubation, net S_min of 7% and 10% occurred in the fallow and three-year farms (p = 0.017), respectively, whereas S immobilization of 5% and 9% occurred in the five- and ten-year farms, respectively. These only correlated positively with POXC and Fe_d. Before incubation, the Dompem soil sulfate contents were also very low (5.7 to 12.2 mg kg⁻¹), differing (p = 0.004) among the farm types, with the lowest contents in the ten-year farms (Table 4). On day 14, net S_min rose sharply, representing about 3.7- to 9.5-fold the initial contents. The forest farms had about twice the sulfate contents (p < 0.001) of other farms. Total C, SEB, and ECEC had positive correlations with S_min at this point. By the 28th day, S immobilization occurred in the forest farms (p = 0.001), whereas net S_min of 86, 9, 16, and 12% occurred in the fallow, three, five, and ten-year farms, respectively. Sulfur mineralization correlated positively with total C, POXC, N_min, MBC, pedogenic Fe, SEB, and ECEC, but negatively with Al_d (Table 5). Generally, the sulfate contents of the Dompem soils were 1.3, 6, and 17.7 mg kg⁻¹ higher than the Adansam soils before incubation, at 14 and 28 days, respectively.

4. Discussion

The study revealed unique nutrient cycling trends at AEZ and farm levels within two zones, presenting specific drivers involved. For instance, the findings resonate with those of previous studies on the same sites [12,13] and elsewhere [14,60], suggesting that agricultural expansion is a key anthropogenic stressor. Generally, agricultural expansion destabilizes substrate quantities and qualities and disrupts nutrient cycling [26,61], with varying extents of impacts. From the AEZ perspective in this study, net N immobilization occurred in all AEZs within 14 days of incubation except for net N mineralization in the forest–savannah zone. By the 28th day, immobilization dominated all the soils with substantial amounts in the forest–savannah zone. The net N_min in the SDF zone also resonates with those of the Dompem farms, which had higher mineral N contents (day 14 = 28 mg kg⁻¹; day 28 = 22 mg kg⁻¹) compared to Adansam farms. Conversely, net N immobilization in the forest–savannah zone corroborates those of relatively older Adansam farms (in the same zone). Unlike N_min, net S_min occurred in all AEZs by the 14th day, with higher values in the relatively wet AEZs (FST and SDF), followed by substantial immobilization in all zones by the 28th day, except net S_min in the SDF AEZ (C:S = 207). At the farm scale, net S_min occurred in all Dompem soils by day 14, and the forest farms had twice the amounts in other farms, followed by substantial net S_min in all farms (particularly the fallows) except for newly cleared forests. Furthermore, net S_min occurred in all Adansam farms by the 14th day of incubation, followed by net S_min only in one- and three-year-old farms compared to S immobilization in relatively older farms (five- and ten-year-old farms). These patterns are similar to those observed in previous studies [13,60], where SOC strongly correlated with POXC (stronger in Dompem soils) and followed a similar trend with total N as in this study.

The observed N_min trends were either positively or negatively influenced directly and indirectly by soil and environmental factors. Higher contents of C, N, and S, microbial biomass, and their cycles were generally observed in wet zones, indicating the impact of climatic conditions such as rainfall and temperature. At a broader scale (AEZ level), N_min was enhanced by S_min, although this effect decreased over time, as well as by POXC, SEB, and ECEC. Conversely, total N content, microbial biomass, and pedogenic iron negatively affected N_min. At a smaller scale (farm level), net N_min, which dominated the Dompem (rainforest AEZ) soils with high values in younger farms (one to three years), was enhanced by total N, microbial biomass, POXC, fine soil fractions, and charge properties, but was reduced by soil pH and Al_d. In contrast, factors such as POXC, N_min, SEB, and ECEC enhanced S_min, while total N, microbial biomass, and pedogenic Fe reduced it at the AEZ level. Sulfur mineralization in Adansam farms was enhanced by POXC and Fe_d. Additionally, S_min in Dompem farms was enhanced by total C, POXC, N_min, microbial biomass, SEB, and ECEC, but was negatively affected by pedogenic minerals.

Among the factors, POXC (labile SOC) appeared to have an overarching effect on N and S mineralization at both levels. This reinforces the role of climate parameters and conforms with previous studies [62,63,64], confirming the key role of labile C as a substrate for biological reactions. Consequently, net mineralization is high in soils containing a substantial amount of labile SOC [13,65,66]. This explains the predominant net N_min in SDF at the large scale and in the RF AEZs at the farm level, where a sharp contrast is observed between the N_min of Adansam and Dompem farms. Generally, the study sites ranged from semi-arid conditions with coarse-textured soils in the north to moist conditions with fine-textured soils in the south. Thus, the observed mineralization patterns could be attributed to ecological factors such as rainfall amount and temperature, which determine biomass production and SOC accumulation, and corroborate the meta-analysis by Du et al. [67], and the findings of Yuan et al. [68] and Zhang et al. [69].

Carbon and nitrogen mineralization are stimulated by microbial energy demand for their activities as needed for growth, leading to phosphorus (P) and S mineralization [70,71]. This explains the strong correlations observed between S_min and N_min throughout incubation (Table 3) and is supported by the findings of Zhou et al. [72]. Soil bacterial demand for energy in the form of C necessitates N_min [70], which begins with ammonification, followed by nitrification and immobilization, as indicated by high NH₄⁺ levels in the soils before incubation (Figure 3, Figure 4 and Figure 5), similar to observations made by Li and Li [73]. Interestingly, the microbial biomass of fine-textured Dompem soils showed a similar trend, reflecting the factors that dominated at the AEZ level, where the negative effects of pH and free Al were observed only during the first 14 days of incubation (Table 3). During the later days of incubation, silt and clay became more significant in the cycling process. In the coarse-textured Adansam soils, microbial biomass and free Al played opposing roles, especially in the positive ∆pH of the soils [12]. The notable N immobilization observed at the end of the incubation further supports the occurrence of both abiotic and biotic immobilization [74,75] in these soils. This is evidenced by negative correlations between N_min day 14 and total N, microbial biomass, and pedogenic Fe, as well as N_min day 28 involving microbial biomass and Fe_d (Table 3). The correlations suggest that limited N content and soil pH may have stimulated biotic immobilization, where inorganic N and S forms were reincorporated into microbial tissues, and abiotic immobilization by adsorption onto soil minerals [74,75,76]. Previous findings indicated that the farms had very low SOC levels (mostly mean values of 1.8% with a few cases of 3.2%), along with low total N, total S, and POXC contents [12,13]. These values were much lower than those found in oil palm agroforestry and monocultures of Amazonia [77], and in different land uses in the USA [78]. Additionally, microbial biomass exceeded C mineralization by more than twofold in these soils [13], implying nutrient constraints [79], as well as microbial needs and preferences for ionic species. Research shows that microorganisms prefer NH₄⁺-N to NO₃⁻-N for their growth and will only utilize NO₃⁻-N when NH₄⁺-N content is below 1 µg N g⁻¹ soil [73,80]. This is because NH₄⁺-N can be directly incorporated into biomolecules, while NO₃⁻ must be converted to NH₄⁺ before assimilation—an energy-intensive process [74]. Moreover, bacteria and fungi differ in their preferences for each N species; bacteria prefer simpler forms than fungi do [74,75]. The role of soil pH cannot be overlooked, as it determines which microbial and N species are present. Acidic conditions promote biotic immobilization of NO₃⁻ through the ferrous wheel hypothesis [81], while slightly acidic conditions favor abiotic immobilization [75], and neutral to slightly alkaline conditions enhance microbial immobilization of NH₄⁺ [66,75].

Secondly, minerals that form during soil development or pedogenic soil minerals, such as Al and Fe oxides, charged soil properties [12,82], and fine soil fractions played key roles in the nutrient cycling process. These pedogenic compounds have been shown to negatively impact N and S mineralization, which is consistent with previous findings on their adverse effects on microbial indicators in these same soils, especially those associated with iron oxides [13]. This observation is supported by numerous studies [83,84] that emphasize the significant role of soil minerals, particularly Fe-Al oxides, in controlling SOC dynamics [83,85]. This influence occurs through the formation of mineral-associated organic carbon (MAOC) or organo-mineral complexes [83,84], which helps stabilize SOC and reduce its mineralization. MAOCs are created via physicochemical sorption, chemical bonding, and physical aggregation, involving processes like co-precipitation, chelation, and aggregation [83,84]. Research estimates that Fe-Al oxides account for about 3% to 72% of SOC storage [86,87], while poorly crystalline minerals contribute over 25% to the SOC pool due to their larger specific surface area and higher chelation capacity [88]. Consequently, Fe_ox, with its larger specific surface area and more reactive sites, exerts a stronger effect on SOC dynamics [83,89,90]. In addition to the properties of Fe-Al oxides, the distribution of soil particle sizes, particularly fine particles with large surface areas, further enhances the binding strength of Fe-Al oxides [83,91]. This highlights the importance of soil texture, a point reinforced by Matus et al. [92], who found that fine fractions of silt and clay accounted for over 95% of maximum SOC accumulation. Environmental factors such as climate, precipitation, and ecological conditions strongly affect Fe-Al oxide concentrations [83]. These complexes are especially prevalent in forested, wet tropical regions, where intense weathering of primary minerals leads to greater retention of Fe-Al oxides within the soil profile [83,90].

Microbial demand for growth and reproduction drives S_min [70,93], whereas their demand for energy in the form of C necessitates N_min [70]. Therefore, net S_min occurs when SOM provides sufficient S to fully satisfy microbial requirements and is likely to be closely associated with N_min [40]. This is said to occur within the first 14 days, releasing about 62–74% of S, which is found to mineralize [56,94]. Sulfur mineralization and immobilization occur concurrently, exhibiting a strong relationship with the soluble sulfate pool in soil, and the direction is dictated by the aforementioned biotic and abiotic driving forces. In limited S conditions, mineralized S is immobilized by microbes, but at varying rates, as seen in the high proportions of S sequestered by fungi (14–45%) compared to bacteria (6–23%) [71,95].

Again, it is obvious that both biotic and abiotic immobilization occurred because (a) C:S ratios of the farms varied widely ranging from 110 to 2190 for the AEZ sites, 59–6729 for Adansam, and 14–1900 for Dompem soils, which were mostly above the critical limit of 200–400 [71,95], where microbial incorporation of S can easily occur, and because (b) pedogenic Fe and Al contents promote abiotic immobilization through sorption [12,71]. Sulfur mineralization is influenced by energy and nutrient supply, with a critical C:S ratio of 200–400 [71,95], an abundance of organic sulfur, water availability, pH, temperature, and redox potential [71]. The total S and sulfate contents had been investigated earlier, suggesting critical levels [43] that require attention. The total N contents were equally low and similar to those found by Alam et al. [96]. Widespread immobilization in this study suggests two major issues: (a) nutrient depletion caused by agricultural expansion, and (b) the low-nutrient input and high-nutrient export nature of agriculture in the AEZs. It further confirms that soil fertility has been a key driver of agricultural expansion [97] to closing the gaps between production and consumption in the tropics [98,99]. This is evidence of interruptions in soil resilience for the delivery of ecosystem services under a rapidly changing climate [100], and also reiterates the persistently low-input agricultural systems of SSA [101,102].

The findings indicate that N and S cycling varied significantly across the AEZs and farm ages, reflecting diverse environmental, soil, and soil management factors influencing these processes [75]. Nutrient cycling constitutes a biogeochemical cycle mediated by microbes and shaped by soil and environmental conditions. These cycles are interconnected, with each influencing and being influenced by others [26]. In this study, factors such as fine soil fraction, clay content, pedogenic minerals, charge properties, soil pH, POXC, total N and S, basal respiration, and microbial biomass predominantly governed nutrient cycling processes at both scales. Previous research suggests that specific factors regulate distinct stages of mineralization [103] and operate at varying rates and extents [64,104]. These factors determine whether net mineralization, immobilization, or denitrification occurs. Key determinants include soil texture, pH, soil minerals, soil organic matter (SOM) quantity and quality, aeration, microbial biomass and community composition, metabolic activity, soil management practices, soil–microbe–plant interactions, seasonal variations, and parent material. The influence of each factor varies spatially according to the AEZ, which affects mean annual temperature, precipitation, SOC, and N contents.

A general pattern observed at the farm scale was the variation or fluctuation of nutrient cycling among the farms, with fallow farms occasionally outperforming forest farms. These patterns can be attributed to specific cropping systems, soil management practices, and soil textures, which result in different nutrient balances due to varying levels of inputs and outputs (such as crop exports) [105,106]. Additionally, the slight differences observed between fallow and forest farms of Dompem (RF) reflect the nature of soil-vegetation interactions. This is because most of the tree roots in the zone form turf covering the soil surface, thereby restricting effective organic matter cycles. Furthermore, thick crown cover also restricts rainfall through fall to the soil, potentially limiting effective biological activity. It is a common occurrence in some forest vegetation in Ghana and was persistently observed during the fieldwork. As a result, when the trees are cleared, SOC decomposition may lag behind that of fallow land, which has already been exposed to such processes during previous cultivation cycles.

It is acknowledged that this study was laboratory-based and precludes the effects of plant uptake, rhizosphere interactions, continuous deposition of organic materials, and the external environmental conditions that play out under field conditions. Notwithstanding, the findings and measured properties provide some indications of the aforementioned stresses. This calls for more intentional, intensive, widespread, persistent actions that allow for the diagnosis of mechanisms for restoration and possible prevention [4]. Such efforts align with the World Soil Charter of the Global Partnership [107], the international 4 per 1000 initiative: soils for food security and climate, https://4p1000.org/?lang=en (accessed on 5 December 2025), FAO Climate-Smart Agriculture, and possibly the Climate-Vulnerable Agriculture [108]. The findings of the study suggest that although agricultural expansion is an ecological stressor, its effects have disparities linked to factors of soil formation, differing from place to place. Consequently, we encounter findings that observed the most overriding factors in different ecological zones [67,74,92,109,110]. This implies that for future soil management and sustainability, soil-forming factors and socio-cultural factors need to be considered in the design and interpretation of nutrient cycling for site-specific applications. For reliable and sustainable soil management, a widespread adoption of conservation agriculture, agroforestry, integrated crop–livestock systems, and crop intensification with integrated soil management is needed to achieve variable nutrient crop uptake and mandatory target SOC levels for good soil health [111]. Moreover, Amazonian terra preta philosophy can be adopted [112] to promote the closure of nutrient loops, such as the application of ecological sanitation, tailor-made biochar technologies, and enhancement of farmers’ capacity to integrate soil and environmental management.

5. Conclusions

The research uncovered distinct patterns of N and S mineralization at AEZ and farm levels. It was noted that N immobilization was predominant at the AEZ level, with significant N immobilization occurring primarily in the relatively older farms (five and ten-year farms) of Adansam. Additionally, net N_min was observed in the Dompem soils, while net S_min was recorded at all levels, albeit with a few instances of minor S immobilization. The correlations identified ecological factors, including climate and vegetation, alongside geogenic factors such as soil type and its inherent properties, as well as the quantity and quality of SOM as indicated by POXC, which played a crucial role in the mineralization processes at both levels. In general, biotic immobilization, resulting from limited N and S availability, led to microbial incorporation at later stages, while abiotic immobilization was attributed to soil characteristics, pedogenic minerals, and particle sizes. Further observations indicate that the expansion of cropland results in the depletion of SOC and soil nutrients. This decline tends to occur mostly during the initial years, although the extent of this decline varies based on site characteristics and the duration of cultivation. This investigation was conducted in a laboratory setting, which limited the influence of environmental and real-world factors. Consequently, the results may have either overstated or understated the actual field conditions and scenarios. Nevertheless, the study offers critical insights to inform soil nutrient management across various cropping systems and patterns, as well as site characteristics and AEZs. Therefore, future research could focus on comparing laboratory and field studies to establish relationships both with and without crops, thereby elucidating the synchrony between nutrient release during cropland expansion and crop productivity. However, these findings underscore the necessity for customized soil and land management strategies across different AEZs to maximize the benefits derived from ecosystem services associated with soil resources. Achieving this can be facilitated through policy discussions and implementation.