The review utilised the preferred reporting Items for systematic review and meta-analysis (PRISMA) framework, as described by Page et al. (2021), for collecting and reporting information on S in agricultural soils in sub-Saharan Africa (SSA). This method is widely recognised for evaluating published systematic reviews through critical analysis (Hutton et al., 2015). To find and download journal articles and published reports, a literature search was conducted in four academic databases and search engines: Web of Science (http://apps.webofknowledge.com/), ScienceDirect (https://www.sciencedirect.com/), PubMed (https://pubmed.ncbi.nlm.nih.gov/), and Google Scholar (https://scholar.google.com/) The systematic search aimed at selecting fully published journal articles by carefully crafting search strings, ensuring that the majority of relevant articles for the review objective were included.

All search terms were based on predefined keywords related to the title, keywords, and abstract of the articles. To refine and hone the search results, specific keywords included (“Sulphur” OR “Sulfur”) AND (“fertiliz*” OR “fertilis*” OR “nutrition” OR “deficiency” OR “status”) AND (“crop production” OR “yield” OR “growth”) AND (“sub-Saharan Africa” OR names of specific countries like “Tanzania” OR “Kenya” OR “Nigeria”). The search was limited to English-language articles in the fields of agriculture, biological sciences, and plant sciences.

A total of 577 journal articles were identified through a search conducted to identify articles published between January 1980 to June 30, 2024. Of these, 156 articles were removed as duplicates, and the remaining 419 articles were further screened. The inclusion criteria used included (i) original research studies, (ii) studies that investigated soil S status and management in crop production, (iii) studies conducted within Sub-Saharan Africa (SSA) (iv) studies published within the time frame of 1980 – 2024, and (v) studies written exclusively in the English language. During the initial screening of titles and abstracts, 267 articles out of 419 articles were excluded for not meeting the inclusion criteria. Following a thorough filtering, 97 out of the 152 remaining articles were eliminated because the main text did not contain relevant information directly related to the review’s scope. In the end, 55 articles were included after a thorough full-text screening. The flowchart of the entire process from identifying the relevant literatures to making the final inclusion decision is illustrated in Figure 1.

The included articles focused on various parameters, including the year of publication, the S extraction method used, the study country, as well as study target. Additionally, we assessed the target crop, S materials, and rates used to supply S for the crop production. The summary of the evaluated parameters is presented in Supplementary Table 1.

Soil sulphur (S) levels are interpreted relative to the extraction method used, as each extractant varies in its ability to extract different S fractions, and associated crop responses to applied S. In Tanzania, for example, Amuri et al. (2023) reported a critical concentration range of 4.1–4.8 mg S/kg for S extracted using 0.01 M CaCl₂ with the SoilDoc method, providing a localised benchmark for deficiency assessment. In Ethiopia, a critical range of 20–80 mg/kg following extraction by Mehlich-III has been proposed by the Ethiopia Soil Information System [EthioSIS], 2014, as reported by (Lelago et al., 2016). For the widely used 0.01 M Ca(H₂PO₄)₂ method, Horneck et al. (2011) proposed the following classification: < 2 mg S/kg as very low, 2–5 mg/kg as low, 5–20 mg/kg as medium, and >20 mg/kg as high. In tropical contexts, Landon (1991) recommended slightly higher threshold values, classifying a critical range of 6–12 mg S/kg as adequate for various extraction methods. Furthermore, Peverill et al. (1999) suggested a general sufficiency range of 5–10 mg S/kg that could apply to multiple methods, including both 0.01 M Ca(H₂PO₄)₂ and KCl extractions. These ratings provide essential context for interpreting soil S data and identifying potential deficiencies, based on the specific extraction method used. However, for efficient and accurate interpretation, there is a pressing need to align critical values with both the extraction method and geographic context.

Sulphur (S) status in soils varies with soil depth. Thus, to determine general S fertility status in soils, this study categorised sampling depth into two groups (shallow topsoil sampling: 0–15 cm, and deeper sampling: 0 – 20/30 cm). A total of 14 studies were conducted at the 0–15 cm depth, with the majority located in West Africa, notably in Nigeria and Ghana, between 1987 and 2022. Studies with shallow-depth sampling are comparatively rare in East and Southern Africa, with only one example from Malawi (Weil and Mughogho, 2000). In contrast, the 0–20/30 cm category dominates the literature, comprising 34 studies spread across East, Southern, and West Africa. This deeper sampling approach has been prevalent from the early 1990s through to 2024, reflecting a widespread emphasis on deeper sampling S dynamics.

Of the 52 reviewed articles, 31 reported soil S status, covering 38 experimental sites across Sub-Saharan Africa (SSA). A regional assessment of soil S status revealed distinct regional variability (Table 1). In East Africa (n = 19), the majority of sites (63.2%, n = 12) were classified as having low S levels, with a mean extractable S concentration of 4.38 mg kg^-¹. Medium and high S levels accounted for 26.3% (n = 5; mean = 8.23 mg kg^-¹) and 10.5% (n = 2; mean = 74.15 mg kg^-¹) of sites, respectively. In West Africa (n = 17), low S was observed at 47.1% of sites (n = 8; mean = 3.14 mg kg^-¹), followed by medium S (41.2%, n = 7; mean = 9.28 mg kg^-¹) and high S (11.8%, n = 2; mean = 31.95 mg kg^-¹). In Southern Africa (n = 2), both sites fell within the medium S category, with a mean S concentration of 8.69 mg kg^-¹. Overall, low S status was the most prevalent across the SSA fields, representing 55.3% of all observations. Ratings were determined using interpretation ranges for most commonly used extraction reagents based on Landon (1991) and for the Mehlich-III thresholds by Ethiosis (2014), as cited in (Lelago et al., 2016).

S concentrations in soil profiles varied significantly with depth, showing distinct vertical distribution patterns across sites and soil types (Table 2). S levels in the topsoil (0–20 cm) ranged broadly from 1.0 to 51.0 mg/kg. Low values were observed in Typic Paleudalf profiles in semi-deciduous rain forest of Ofin, Ghana, with 1.0–4.0 mg/kg, while high concentrations were recorded in Typic Haplorthox (38.0 mg/kg) and Oxic Tropaquent (51.0 mg/kg) of Ankasa and Kwabeng profiles, respectively (Acquaye and Beringer, 1989a). Uloro and Mengel (1994) reported 26.4 mg/kg in the 0–25 cm layer in Loam soils of Kalisha, Ethiopia, while Kalonga et al. (2024) reported 14.2 mg/kg at 0–20 cm and 13.95 mg/kg at 20–50 cm at the Maasai landscape of Arusha, Tanzania, indicating low Mehlich-III S levels. In the Ethiopian rift valley, Haplic Nitisols had low S levels; at 25–60 cm depth was 1.4 mg/kg, while at 60–102-cm depth was 1.0 mg/kg (Itanna, 2005). In contrast, Calcaric Fluvisols recorded elevated values: 33.6 mg/kg at 0–29 cm and 46.4 mg/kg at 29–75 cm. Similarly, Haplic Solonetz profiles showed 7.6 mg/kg in the Ah horizon (0–24 cm) and 33.2 mg/kg in Bm (24–56 cm), with progressively higher values in deeper clay horizons. S concentrations were highly variable in deeper layers, ranging from 0.2 to 288.0 mg/kg. In profiles like the Vitric Andosol and Haplic Nitisol, very low S levels were recorded (e.g., 0.2–1.0 mg/kg between 60 and 140 cm) (Itanna, 2005). However, Haplic Solonetz showed extraordinarily high values in Bt and Cm horizons: 288.0 mg/kg (56–118 cm), 107.2 mg/kg (118–122 cm), and 203.4 mg/kg (122–136 cm). Loam soil in Kalisha, Ghana, exhibited high S levels from 25 to 93 cm, ranging between 23.6 and 32.0 mg/kg (Uloro and Mengel, 1994).

The increased focus on deeper sampling aligns with agronomic research prioritising cereals (maize, sorghum) and legumes (soybean, common beans), whose root systems often extend >15 cm deep (Supplementary Table 1). Determining S in deeper soil is critical for these crops, as grain filling and nitrogen fixation depend on sustained S availability beyond surface layers. The observed regional differences in soil S status reflect underlying variations in agroecological conditions, soil parent material, land use intensity, and fertiliser management across SSA. The predominance of low S concentrations in East and West Africa soils is consistent with the inherently low S reserves of many weathered tropical soils and may be further exacerbated by continuous cropping, low use of S-containing fertilisers, and nutrient depletion (Tabatabai, 1982).

The low mean S levels remain below agronomically sufficient thresholds in East and West Africa regions (4.07 mg/kg in East Africa and 3.87 mg/kg in West Africa), highlighting a widespread risk of S deficiency that could limit crop productivity (Haneklaus et al., 2007; Landon, 1991). Conversely, the high S values of 74.15 mg/kg and 31.95 mg/kg occurred in a few East and West African sites, respectively. These observed high S values suggest localised enrichment, potentially due to site-specific factors, for instance, deposition from volcanic ash such as in parts of Kenya (Omenda, 2011), manure or compost applications, or mineralogical contributions from the parent rock (Haneklaus et al., 2007; Weil and Brady, 2016). The medium S levels in Southern Africa, though based on limited data, may reflect inherent soil characteristics or more effective nutrient management practices (Ngezimana and Agenbag, 2014; Ngezimana and Agenbag, 2015). However, these findings highlight the need for more comprehensive representative data to enable more robust conclusions. Overall, the spatial heterogeneity in soil S levels underscores the need for site-specific diagnostics and the integration of S into balanced fertilisation strategies. These findings also emphasise the importance of providing extraction methods used for accurate compilation or comparisons to scale to support informed nutrient management recommendations. Soil S stratification with depth is primarily shaped by pedogenic processes, texture, and leaching intensity (Weil and Brady, 2016). Surface horizons typically exhibit high soil S concentrations (Table 2), which may be attributed to organic matter inputs and fertiliser additions (Acquaye and Beringer, 1989a; de Mello Prado, 2021a). Subsurface layers often retain moderate S levels, as seen in Calcaric Fluvisols and Haplic Solonetz (Table 2), where limited leaching or higher clay content restricts vertical S movement. In contrast, sharply reduced S levels (to values less than 1.0 mg/kg) in Haplic Nitisols reflect the influence of organic matter accumulation in surface horizons of highly weathered profiles (Itanna, 2005). Exceptionally high S accumulation (up to 288 mg/kg) in the Bt horizon of clay-rich Haplic Solonetz is likely due to SO₄²^- illuviation and restricted drainage. Data from Tanzania show similarly low Mehlich-III S in both topsoil (14.2 mg/kg) and subsoil (13.95 mg/kg), potentially due to S depletion from continuous cultivation and limited organic inputs (Kalonga et al., 2024).

Organic matter stabilises S through microbial mineralisation–immobilisation cycles, which decline with depth (Das et al., 2024). Moreover, routine soil tests may underestimate subsoil S in clay-rich profiles due to strong adsorption onto mineral surfaces, rendering some S non-extractable (Bankole et al., 2022; Tabatabai, 1982; Weil and Brady, 2016). Such conditions may explain the widespread low S reported in subsoils (Acquaye and Beringer, 1989a; Bankole et al., 2022). Soil texture and organic matter further mediate S retention by influencing leaching and adsorption: loam soils, for instance, maintained moderate to high S (21.6–32.0 mg/kg) throughout the upper 90 cm due to favourable nutrient-holding capacity (Uloro and Mengel, 1994). In contrast, coarse-textured, weakly developed soils readily lose SO₄²^-, while highly weathered tropical soils with abundant Fe and Al oxides can strongly retain it (Weil and Brady, 2016). These findings underscore the importance of depth-specific S assessments in developing tailored nutrient management strategies for diverse agroecosystems across SSA.

The relationship between soil S concentration and yield response was evaluated across six crops (maize, rice, wheat, soybean, canola, and soybean) under varying S application rates (Figure 6). Yield responses were most reported at soil S concentrations below 2.0 mg/kg, where treatments resulted in yield increases exceeding 50% in many cases. For example, in soils with <1.5 mg S/kg, maize achieved yield increases up to 125% at 0–10 kg S/ha, while rice responded with yield increases as high as 62% at 11–20 kg S/ha. Linear regression lines fitted for each S application rate and crop combination further confirmed that yield responses diminished as soil S levels increased. In contrast, crops grown in soils with medium S levels (>6 mg S/kg) exhibited minimal response, with per cent yield increases ranging from 17% to 60% for canola and wheat, respectively.

These results indicate that yield benefits from S fertilisation are crop- and site-specific, with the greatest yield response achieved under low soil S conditions. Excessive application in soils with S levels above 6.0 mg/kg yielded no significant advantage in some crops, reinforcing the need for site-specific fertiliser recommendations based on soil test results (Amuri et al., 2023; Bekele et al., 2022).

The use of S-containing amendments and fertilisers such as Sub-Saharan Africa (SA) and elemental S is critical for improving soil fertility and crop yields. However, these amendments may have environmental side effects that must be carefully managed to prevent undesirable consequences. Among primary concerns associated with S amendments is soil and water acidification (Gerson and Hinckley, 2023). When elemental S is applied to the soil, it undergoes microbial oxidation to form sulfuric acid (Tabatabai, 1982), which can lower soil pH (Marschner, 2012; Mengel and Kirkby, 2001), potentially harming crops not adapted to acidic conditions and disrupting soil microbial communities (Gerson and Hinckley, 2023). In Tanzania, soil acidification is a substantial concern, particularly in cashew-growing regions where S is used to control powdery mildew disease (Dondeyne et al., 2001). This practice leads to reduced crop yields and soil fertility, especially in sandy soils with low buffering capacity (Dondeyne et al., 2001; Majule and Omollo, 2009). Acidification can lead to nutrient imbalances, affecting plant health and growth (Gerson and Hinckley, 2023; Mengel and Kirkby, 2001; Weil and Brady, 2016). According to Gilliam et al. (2020), extreme soil acidity can raise the solubility of heavy metals such as aluminium to a toxic level, and render the availability of phosphorus, further hampering plant development. Runoff from acidic soils can carry S material into water bodies, leading to the acidification of water ecosystems (Munodawafa, 2007). This can be observed in poorly managed soils where water erosion can easily carry away soil to nearby water bodies.

Studies have primarily examined the effects of nitrogen fertilisation on greenhouse gases (GHG) emissions like CO₂, CH₄, and N₂O (Mosongo et al., 2022; Pelster et al., 2017). However, the particular influence of S fertilisation on GHG emissions in SSA has been less explored. In other areas, S fertilisation is recognised for its ability to regulate methane (CH₄) emissions from wetland environments like flooded rice fields (Gauci et al., 2004). However, using these materials can lead to the release of S gases like SO₂ upon oxidation, which play a significant role in climate change (Ward, 2009). The use of elemental S increases the solubility and availability of heavy metals such as iron (Fe) and cadmium (Cd), aiding plants in absorbing other metals like zinc (Zn), manganese (Mn), lead (Pb), and cadmium in maize (Safaa et al., 2013). This is attributed to the decrease in soil pH, which facilitates the solubility of these metals in the soil. This situation threatens food safety and human health, as toxic metals can accumulate in crops and infiltrate the food chain.

In the long run, applying elemental S and SA regularly without effective management can increase soil and environmental susceptibility to degradation, thereby reducing soil productivity. The use of these amendments should be tailored to agro-ecological zones and crops, based on the results of soil analysis, with application right fertiliser source, at the right rate, the right time, and in the right place (4R fertiliser stewardship) (Nalivata et al., 2017). S amendments should be sourced appropriately, ensuring high-quality sources free from impurities, and applied at the correct rates, timings, and placements in line with soil nutrient recommendations. Integrated soil fertility management, combining S with other nutrients and organic matter, can enhance long-term environmental and soil health by promoting soil biodiversity and structure, improving nutrient retention, and ultimately boosting agricultural production.