Skip to content
Frontiers
Search

Original Research

Frontiers in Sustainable Resource Management ·

Volume 5 - 2026 | https://doi.org/10.3389/fsrma.2026.1814702

Assessing a farmer-based soil classification system vis-à-vis USDA soil taxonomy suborders in Kamuli District, Uganda

  • Francis Akitwine1*corresponding author
  • Rebecca A. Wokibula1
  • Johnson G. Mtama2
  • Shillah Kwikiiriza3
  • Amber D. Anderson1
  • C. Lee Burras1*corresponding author
  1. 1Department of Agronomy, Iowa State University, Ames, Iowa, United States
  2. 2Department of Research and Innovation, Tanzania Agricultural Research Institute, Soil Science, Uyole, Mbeya, Tanzania
  3. 3Department of Horticulture, Iowa State University, Ames, Iowa, United States
Article metrics0 viewsView details
PDFCiteMetrics

Abstract

Based on 5 years of integrated pedological work in Kamuli District, this study examines how farmers in three villages understand and classify soils. The farmers' knowledge was documented through a participatory approach that included discussions, soil mapping, subsequent field reconnaissance, and 23 pedon descriptions. Based on comparable needs and landscapes, farmers across three villages classified the soils into four groups: Lirugavu, Lubalebale, Mukyanga, and Mutaala. These corresponded to six USDA Soil Taxonomy suborders, Aqualfs, Aquolls, Ustalfs, Ustults, Ustox, and Ustipsamments. Farmers demonstrated spatial awareness through the soil mapping exercise, in which soils were associated with specific landscape positions. A farmer-based ranking in terms of soil productivity aligned with Corn Suitability Rating (CSR2) values, with Lirugavu and Lubalebale as the most and least productive, respectively. The farmer-based soil groupings/ classification distinguished soils by depth of mollic colors, clay content, organic matter (SOM), total nitrogen (N), and cation exchange capacity (CEC). The ability of the soil groupings to discern vertical variation, that is, significant differences from 0–15 cm to 15–30 cm depth layers, was captured only for SOM and N in Lirugavu. Soil Taxonomy suborders captured a wider range of chemical properties and more consistent vertical differentiation. These findings highlight the differences in epistemology but with potential for complementarity between farmer-based and scientific classification systems, leading to integrated approaches to soil assessment and management.

Keywords

1 Introduction

Soil classification is essential to ecosystem management, agricultural growth, and environmental sustainability. Buol and Denton (1984) contended that the primary objective of soil classification was to transfer knowledge about soil use and management from one location to another with comparable soil characteristics. This is vital to advancing scientific understanding and effective communication about soil behavior and potential. Nevertheless, even with their accuracy and rigor, soil classification schemes fall short of providing farmers, especially those without ready access to soil scientists with functional information. However, with minimal or no soil science training, farmers possess a tacit understanding of their soils. This understanding is based on experiential knowledge rather than formalized frameworks (Braimoh, 2002; Ishida et al., 2001; Kyebogola et al., 2020; Nulkar, 2024; Payton et al., 2003; Sandor and Furbee, 1996; Toxqui-Roldan et al., 2024).

Scientific soil classification systems, though differing in methodology, categorize soils through standardized, measurable physical, chemical, and morphological criteria. Systems such as World Reference Base (IUSS Working Group, 2014), USDA Soil Taxonomy (Soil Survey Staff, 2022), the Brazilian Soil Classification System (SiBC; Santos et al., 2018), and the French Référentiel Pédologique (Baize and Girard, 2008) are highly based on experiential knowledge. Smith (1983), a principal architect of Soil Taxonomy, noted that the system's knowledge base emerged from accumulated experience in the United States and abroad. This reliance on practical observations and experience parallels the foundations of local and indigenous knowledge systems. These local systems rely on perceivable attributes such as color, texture, moisture, vegetation indicators, depth, productivity, and stoniness (Abdulrashid and Yaro, 2014; Ishida et al., 2001; Kyebogola et al., 2020; Payton et al., 2003; Tenywa et al., 2014; Winowiecki et al., 2014; Yageta et al., 2019). The reliance on these physical characteristics underscores Ingram (2008) findings that farmers have more “know-how” or “practical knowledge” than the scientific or “know-why” that the formal classification systems offer about soils.

The integration of scientific and local indigenous knowledge offers a range of benefits. These range from preserving the traditional knowledge base (Tengo et al., 2021), a hybrid extension approach (Black, 2000), knowledge co-creation (Utter et al., 2021), and adoption of innovation (Kyebogola et al., 2020; Omatayo and Musa, 1999). These benefits can build resilience and improve resource efficiency, leading to more sustainable systems. However, achieving mutual understanding between the two knowledge systems is complicated. A central challenge in soil science is reconciling farmers' holistic, situation-based soil classifications (Duvall, 2008) with the reductionist, component-based framework of formal soil science (Yageta et al., 2019). Additional barriers arise from the undocumented nature of indigenous knowledge (Rushemuka et al., 2014; Sandor and Furbee, 1996), its localized character (Yageta et al., 2019), and natural scientists' limited familiarity with social science methodologies and cross-cultural engagement (Hunnington, 2000).

Uganda, home to over 56 tribes, each with its unique culture (Achieve Global Safaris, 2025; Tulibaleka et al., 2021), offers a rich context for documenting human-environment interactions. Every community possesses unique traditions, customs and knowledge developed through sustained engagement with their specific landscapes. Kyebogola et al. (2020) and Tenywa et al. (2014) showed how the Baganda in Masaka (Central Uganda) classify and understand soil productivity. Gowing et al. (2004) and Payton et al. (2003) showed the potential of indigenous soil mapping in Katakwi District. In Northern Uganda, Muwanga (2020) and Nuwategeka and Nyeko (2019) noted how indigenous knowledge ties together soil and land suitability in Karamoja and Amuru District respectively. In Western Uganda, Hart and Mouton (2005) and Tweheyo et al. (2024) underscore the critical role of indigenous knowledge in agriculture and natural resource management, respectively. These studies across various regions of Uganda, as shown in Figure 1, demonstrate the potential of local knowledge systems. There is a need for further research in other undocumented areas to inform the potential of these systems to advance locally adapted and sustainable land management practices.

Studies (blue points) that have captured indigenous knowledge related to agriculture and environment in Uganda in relation to Kamuli District whose boundary is shown by the black polygon.
Figure 1. Studies (blue points) that have captured indigenous knowledge related to agriculture and environment in Uganda in relation to Kamuli District whose boundary is shown by the black polygon.

Kamuli District has received limited attention in formal soil research, resulting in a scarcity of detailed soil information. Veenstra et al. (2015), Akitwine (2021), and Anderson et al. (2023) are the only studies that have documented soils in the area. These contributions have been invaluable for establishing baseline pedological data. However, a critical component has remained unexplored: the deep knowledge held by local communities who interact with these soils daily. Our study builds directly on this foundation by integrating local perspectives on soil classification, use, and management .The Basoga are the predominant ethno-cultural group in Kamuli District, where Lusoga is the most widely spoken language. Farming serves as the primary livelihood, and food self-sufficiency is a key measure of a community's economic well being and prosperity (Isiko, 2018). Farming decisions are made under uncertain conditions with respect to formally measured soil properties. But agricultural production has persisted over time, indicating the presence of a locally grounded way of recognizing soil differences. This place-based understanding of soils can enrich the area's soil information and yet remains largely undocumented. As Sillitoe (1998) noted, such knowledge systems are inherently culturally embedded and imparting scientific frameworks risks decontextualization and misrepresentation. This study avoids this pitfall by utilizing a participatory pedological approach in which farmer-generated insights guided the structure of the investigation. Further quantitative analysis was employed to assess how the farmer-based categories correspond with the different soil properties.

The objectives of this paper were two-fold: (a) to document and analyze how local farmers in Kamuli District, identify and classify soils, and (b) to evaluate how the farmer-based classification compares to the Soil Taxonomy suborder level in distinguishing soil properties relevant to agricultural productivity.

2 Materials and methods

2.1 The study area

Kamuli District's landscape, as shown in Figure 2, is comprised of flat-topped ridges that lead into narrow valleys. The valleys drain into the Nile River to the west and Lake Kyoga to the north (Wortmann and Eledu, 1999). Kamuli District lies in an area bound by coordinates 0.651613, 33.137512 (latitude, longitude) and 1.222823, 33.033142 (latitude, longitude). Soils in the area are classically mapped as the Buruli catena comprising reddish-brown sandy loams often underlain by laterite (Ollier and Harrop, 1959). Geologically, Kamuli District lies on an end-tertiary erosion surface dominated by Precambrian granite (Batte et al., 2008; Ruhe, 1954). The vegetation ranges from savanna to regenerating secondary forests in the less cultivated areas (Radwanski and Ollier, 1959). The climate is tropical wet and dry, with an average annual rainfall of 1100 mm and a bimodal distribution (Komuntunga and Musitwa, 2001). Peak rains are received from March to May (Season 1) and August to November (Season 2). Drier conditions typically occur from June to July and from December through February (Kamuli District Local Government, 2020).

The location of Kamuli district. Top right; Uganda and Kamuli District; Left foreground; Kamuli District elevation showing the location of the three villages used in the study (red points).
Figure 2. The location of Kamuli district. Top right; Uganda and Kamuli District; Left foreground; Kamuli District elevation showing the location of the three villages used in the study (red points).

2.2 Soil data collection

The data was collected from three villages: Butegeere, Kawaaga, and Kikonko (Figure 2). This was part of pedological work done in the area in the past 5 years. Additional details are documented in Akitwine (2021), Anderson et al. (2023), Wokibula et al. (2024) and Akitwine et al. (2025). The three selected villages were based on two criteria to ensure both data adequacy and representativeness of soil variability. To ensure sufficient data, each village had of 5 or more fully described soil pedons traversing the village. These pedons ensured that within-village soil variability was fully characterized. The pedons in the village had to be on the major parent materials including residuum, colluvium, and alluvium to capture as much soil variability in each village. Parent material is a primary determinant of soil properties (Jenny, 1941; Soil Survey Staff, 2022). By selecting villages that encompass all three types, we ensured that each village served as a microcosm of the Kamuli's soil-landscape.

A participatory pedological approach with farmers in the three villages was utilized to identify the local soils. Semi-structured group discussions in the local Lusoga language were conducted to document soil nomenclature. The insights from the discussions were then validated through field reconnaissance and soil profile descriptions. The 200 participants chosen were active farmers (including those in whose fields pedons were dug and neighbors) in the respective villages. Male (60%) and female (40%) participants were included to capture a broad spectrum of knowledge and experiences. Zhang et al. (2021) and Christie et al. (2016) showed that men and women interact with soils differently. The discussions focused exclusively on soils, and no personal, demographic, or behavioral information was collected. They also discussed how the different soils are ranked by productivity and the criteria used for this ranking.

The farmers made soil maps of their respective villages with the help of satellite imagery from Google Earth (© 2025 CNES/ Airbus). The printed Google Earth images had a scale of 1:5000, determined using the 200 m scale bar and measured length of the bar on the printouts. A transect-based soil sampling strategy was employed to validate the boundaries of the community drawn maps. Along these transects, a preliminary assessment of horizonation, color, and texture was made using a hand probe to 43 cm. The sampling employed a profile-based replication strategy to capture variability within and between farmer-based categories such that each soil category was represented by a minimum of 3 pedons. In the end, a total of 23 soil pedons, with depths ranging from 28 cm to 254 cm across the three villages, were georeferenced and characterized according to the methods outlined in Schoeneberger et al. (2012). Pedons that were shallower than 100 cm had digging barriers, most commonly a cemented plinthite layer or a high-water table that restricted further excavation. A 500 g sample from each horizon was collected for physical and chemical analysis in the laboratory. Combining morphological, laboratory data and farmer insight, each of the 23 pedons was classified using both the farmer-based approach and USDA soil taxonomy according to Keys to Soil Taxonomy (Soil Survey Staff, 2022).

2.2.1 Validation workshops

Two additional workshops were conducted with additional 120 participants from outside the three study villages to ensure independent assessment of the soil classifications. These participants included knowledge holders from other villages, in whose fields we had explored in a large-scale study on the pedology of the region. The selection of these individuals ensured that the individuals evaluating the classifications had no social connections with the original participants, thereby providing an independent assessment of the soil categories. Also involved in the workshops were regional extension assistants, and scientists from the National Agricultural Research Organization (NARO). The extension assistants offered grounded perspectives on the documented soil classifications and their own extensive observations across multiple communities. The participation of NARO scientists enabled rigorous examination of correspondences between the farmer-based classifications and established pedological understanding of the region. This strategy followed established approaches for ensuring robustness including the multiple evidence base framework (Bentley et al., 2004; Tengö et al., 2014) and group consensus methods for countering bias (Kiptot, 2007).

2.3 Laboratory soil analysis

Air-dried samples were gently crushed to pass through a 2 mm sieve. The hydrometer method was used for particle size analysis (Bouyoucus, 1951) after dispersal in 5% sodium hexametaphosphate. Soil pH was measured potentiometrically from a supernatant solution of a 1:2.5 soil: water suspension (Mclean, 1982). Walkley and Black's wet oxidation was used to measure organic carbon (Nelson and Sommers, 1982). Organic carbon was converted to organic matter (SOM) by multiplying by a factor of 1.724. Total nitrogen (N) was determined using the micro-Kjeldahl digestion-distillation method (Bremner and Mulvaney, 1982). Available phosphorus (P) was quantified using the Bray-1 method (Bray and Kurtz, 1945) and calorimetrically quantified by measuring the blue molybdenum complex with a spectrophotometer at 884 nm (Watanabe and Olsen, 1965). Exchangeable bases and cation exchange capacity (CEC) were determined by saturating the sample with neutral 1M ammonium acetate, followed by displacement using 1M potassium chloride (Chapman, 1965). The bases (K+, Na+, and Ca2+) were quantified using atomic absorption spectrophotometry (Thomas, 1982). Base saturation was calculated by dividing the sum of exchangeable bases by the respective CEC multiplied by 100.

2.4 Statistical analysis

The data used for statistical analysis were from laboratory analysis of individual horizons from the 23 described pedons. The lab results were standardized by depth into 0–15 cm and 15–30 cm intervals for consistent pedon comparison. Ideally, soil horizons are suitable for assessing soil variation because they reflect pedogenic processes. However, farmer assessments often rely on observable and near-surface properties, typically within 30 cm of the profile. The two depths were used to evaluate the farmer-based system's robustness in grouping soils with similar properties. Additionally, the two depths enabled the assessment of the two systems to capture vertical stratifcation (shown by the differences between the two depth).

Statistical tests were performed using R (R Core Team, 2024). Data were checked for normality using the Shapiro-Wilk test. Soil properties that deviated from normality, Box-Cox transformation (Osborne, 2010; Rajarathinam and Ramji, 2023) were applied. To ensure validity of the assumptions, further analyses were conducted on the original or transformed data, as applicable. The performance of the two systems was tested using the general linear model (GLM). Equation 1 was used for physical and morphological properties, while Equation 2 was used for chemical properties.

Yij=μ+Ai+εij(1)

Where;—Yij is the measured soil property

—μ is the overall mean of the measured property

—Ai is a variable representing the major grouping of the soil property. In the farmer-based knowledge, is the local name, and suborders of the USDA Soil Taxonomy system

—εij is the error term, that is, the variability unexplained by the classification system.

Yijk=μ+Ai+Bj+(AB)ij+εijk(2)

Where;—Yijk, μ, and Ai, and εijk are similar to those as shown in Equation 1.

—Bj is the effect of the depth (0–15 cm, 15–30) considered for chemical properties.

—(AB)ij is the interaction effect between classification and depth.

The suborder level was used as a scientific basis for comparison to the farmer-based classification because the sample sizes at the subsequent taxonomic levels were insufficient for statistical comparison. For each depth interval (0–15 cm and 15–30 cm), analysis was performed separately. Between-group comparisons assessed the systems' ability to discriminate between soil groups. To investigate whether the various soil classifications consistently reflected vertical stratification, within-group comparisons across depths were also carried out. Tukey's HSD test followed significant results at α = 0.05. Means are presented with their standard deviations. The classification models' fit were also evaluated using the coefficient of determination (R2) statistic and the Root Mean Square Error (RMSE). These two parameters quantify how well each classification explains the variability in the soil property of interest (Yost and Fox, 1981, 1983). Higher R2 and lower RMSE values indicate a better classification system at accounting for the differences in properties between the defined soil groups.

3 Results

3.1 Farmer-based soil nomenclature

The farmers identified four major soils across the three villages, as shown in Supplemental Figure 1. The soil nomenclature was based mainly on epipedon characteristics, such as color, gravel/rockiness, and texture. Lubalebale is a direct translation to “stony soil”, conveying its high content of surface and subsurface concretions and quartz rocks. Lirugavu is a direct reference to a black hue that dominates the soil. Mukyanga means sand and is used for soil that is predominantly sandy. Mutaala is a well-drained upland and often deep soil. The morphology and extended descriptions of the representative soils are shown in Table 1. Additional details on the soils, including morphology and pedogenesis in the area can be found in Akitwine (2021), Anderson et al. (2023), Wokibula et al. (2024), and Akitwine et al. (2025).

Table 1. Description of a representative pedon from each of different farmer-based classification soil groups in Kamuli District.

Local soil name in Lusoga languageUSDA Soil taxonomy familyLandscape positionParent materialDepth to limiting layerEpipedon characteristicsSubsoil characteristicsDrainage
MutaalaLoamy-skeletal, parasesquic, semiactive, thermic, and Mollic PlinthustalfSummit with a 2% slopeWeathered granitic residuum152cm to a layer that is about 75% plinthic nodulesVery dark brown sandy clay loam layer with a coarse subangular blocky structure parting to granularDark brown to dark reddish brown clay loam to loam texture. Subangular blocky/ prismatic structure in deeper layers. A layer that is 75% plinthic nodules encountered at 152cm with a predominantly a single grain structureWell drained
LubalebaleClayey, Kaolinitic, shallow, thermic, and Plinthic EutrustoxStraight mid-slope (backslope) with a 4% slopeColluvium over weathered residuumShallow depth to a cemented plinthite layerDark reddish brown sandy clay loam with a fine subangular blocky structureDark reddish brown with a sandy clay texture. 8 cm thick layer with a subangular blocky over a 50cm cemented plinthic rock. Quartz stone lines commonWell drained
LirugavuFine, mixed, active, thermic, and Umbric EndoaqualfConcave lower midslope to Footslope 2% slopeAlluvium with a mix of some colluviumNone observedBlack sandy loam to sandy clay loam with medium to fine subangular blocky structureVery dark gray sandy clay layer with a subangular blocky structure over a dark gray clayey layer with a massive structure. Redoximorphic features often encountered in these soilsSomewhat poorly drained
MukyangaMixed, thermic, and Typic UstipsammentFootslope to stream valleys. 1% slopeAlluviumRounded rocks of about 5cm to 10cm diameter (quartz nature) encountered oftenVery dark gray structureless sandVery dark grayish brown to dark grayish brown loose sand with clay infillings often lodged between the bouldery rocks. Redoximorphic features encountered oftenPoorly drained

The farmers' understanding beyond the surface characteristics is shown in Figure 3. The dark topsoil, with a friable consistency and generally corresponding to the A horizon was referred to as Lirugavu. Lirugavu refers directly to its relatively darker color as compared to subsoils and B-horizons. The subsoils in the upland well-drained soils were often referred to as imyufu or enkuusi, due to their red color. Further downslope, differences in surface and subsurface soil characteristics indicated the influence of drainage. Poorly drained subsurface horizons would have color distinctions such as Kyevu (yellow) or Kivuvu (ashy appearance). The four local soil classifications fell into six USDA Soil Taxonomy suborders. Mutaala were classified as Ustalfs and Ustults. Lubalebale were classified as Ustox, with Lirugavu being classified as Aquolls and Aqualfs. Mukyanga were classified as Aquolls, Aqualfs, and Ustipsamments.

Differences in soil morphology among the four major soils according to farmers in Kamuli District.
Figure 3. Differences in soil morphology among the four major soils according to farmers in Kamuli District.

The participatory mapping exercise revealed how the local soil classifications are rooted in an understanding of landscape position (Figure 4). Mutaala are often associated with the upper slopes (summits). The farmers preferred these soils for crop growth, with this preference linked to better drainage and more favorable working conditions during the rainy season. Lubalebale, often encountered at the shoulder position, are often shallow to plinthite (ebisilokoko), a common root-limiting layer in the area. These were considered the least suitable due to the limited rooting depth and higher susceptibility to erosion.

Toposequence cross-section showing the implicit understanding of soil-landscape relationships by farmers in Butegeere village, Kamu.li District. Cross-section generated using Google Earth elevation profiles.
Figure 4. Toposequence cross-section showing the implicit understanding of soil-landscape relationships by farmers in Butegeere village, Kamu.li District. Cross-section generated using Google Earth elevation profiles.

Lirugavu, located on the foot slopes, are generally regarded as the more fertile soil. These have greater depth of mollic colors, higher moisture retention, and perceived fertility. These soils benefit from the accumulation organic matter and nutrient-rich sediment transported downslope. Despite being simple to work with, Mukyanga, situated at the valley bottom, have drawbacks like poor nutrient retention (Kogge Kome et al., 2018) and increased susceptibility to flooding. However, Mukyanga are still considered more suitable for cultivation compared to Lubalebale due to moisture availability, particularly in the dry season. This spatial relationship between soil groups influences the soil colors. The well-drained Mutaala and Lubalebale exhibit red and brown hues whereas the poorly drained Mukyanga and Lirugavu exhibit darker hues.

3.2 Comparative performance of soil classifications in grouping soils

3.2.1 Physical and morphological properties

As shown in Table 2, the farmer-based classification showed significant differences among groups regarding depth of mollic colors (p-value 0.01). The Soil Taxonomy suborders showed no significant differences (p-value 0.083). Lirugavu had the deepest mollic colors and was considered statistically different from Mutaala. In terms of explaining the variability, the farmer-based classification performed better with a higher R2 value and a lower RMSE (Table 2).

Table 2. Statistical analysis of the classification systems at grouping soils with similar depth of soils with color values and chromas less than 3/3 in Kamuli District.

Depth of mollic colors (cm)Model parameters
Farmer-based soil classificationp-valueR2RMSE (cm)
Soil nameMean ±Standard deviationStandard error0.010.6417.49
Mutaala (7)35 ± 16a4.72
Lubalebale (5)37 ± 12ab4.61
Lirugavu (5)66 ± 26b12.02
Mukyanga (6)57 ± 22ab9.03
USDA soil taxonomy suborders
Aqualf (3)64 ± 2413.850.080.4218.37
Aquoll (3)47 ± 3118.09
Psamment (3)84 ± 75.00
Ustalf (6)35 ± 228.06
Ustox (3)30 ± 127.23
Ustult (3)45 ± 43.00

Number in parentheses shows the number of pedons considered for each classification system.

Means with different letter are significantly different at p < 0.05 according to Tukey's HSD test.

Since texture was a major criterion used by farmers to differentiate soils, particle size results were evaluated (Table 3). Notably, Mukyanga references “sand”, suggesting a strong association with sandy texture. However, sand content did not show statistical significance between groups. Instead, clay content showed statistical significance. Mukyanga had the lowest clay content and the highest sand content. The lack of statistical significance for sand suggests a relatively uniform distribution of coarse textures across the different soil groups attributed to the underlying granitic parent material which strongly influences soil texture in the region. Despite the suborders not showing clear distinctions detectable by ANOVA at p = 0.05, USDA Soil Taxonomy better model performance for silt and sand. The farmer-based soil classification performed better for clay content, explaining 63% of the variation, compared with 41% for the suborders (Table 3).

Table 3. Statistical analysis of the classification systems at grouping soils with similar particle sizes in Kamuli District.

Farmer-based soil classificationSand (%)Silt (%)Clay (%)
Lirugavu (5)67 ± 512 ± 320 ± 2ab
Lubalebale (5)63 ± 1514 ± 822 ± 12ab
Mukyanga (6)79 ± 1011 ± 48 ± 6 b
Mutaala (7)64 ± 1513 ± 422 ± 9a
*p*-value0.070.890.03
R20.270.030.63
RMSE10.44.75.9
USDA soil taxonomy subordersSand (%)Silt (%)Clay (%)
Aqualf (3)70 ± 10.0115 ± 3.6014 ± 8.14
Aquoll (3)72 ± 10.0611 ± 1.5216 ± 9.84
Psamment (3)87 ± 0.018 ± 0.015 ± 0.01
Ustalf (6)72 ± 6.4711 ± 2.9716 ± 5.38
Ustox (3)69 ± 13.0010 ± 3.4621 ± 14.93
Ustult (3)59 ± 5.6511 ± 4.2430 ± 1.41
*p*-value0.090.240.11
**R2**0.430.330.41
RMSE7.12.56.8

Number in parentheses shows the number of pedons considered for each classification system.

Means with different letters are significantly different at p < 0.05 according to Tukey's HSD test.

3.2.2 Chemical properties across the classification systems

Chemical properties were assessed for both group differences at each depth and within-group comparison across the two depths. Statistical performance for chemical properties differed across the two classification systems (Tables 4 and 5). The farmer-based classification differentiated soils on three chemical properties: SOM, N, and CEC. Soil organic matter and N showed statistical significance at only the 0–15 depth between groups, while CEC showed statistical significance between groups at both 0–15 cm and 15–30 cm depths. Fitting Equation 2, the farmer-based soil groups explained 64, 57, and 72% of the variation in SOM, P, and CEC, respectively. Soil Taxonomy suborders separated soils into distinct groups for N (0–15 cm), P (both 0–15 cm and 15–30 cm), Na+ (0–15 cm), CEC (0–15 cm), and base saturation (15–30 cm). Fitting Equation 2, the suborders accounted for 63, 58, and 79% of the variation in P, Na+, and CEC respectively.

Table 4. Statistical analysis and model parameters comparing the performance of the different classification systems at grouping soils' chemical properties in Kamuli District.

Classification systemDepth (cm)pHO.M (%)Total N (%)Available P (mg/kg)
Mutaala (7)0–156.3 ± 0.62.7 ± 0.6 b0.14 ± 0.06ab4.8 ± 1.7
15–306.1 ± 0.82.4 ± 0.70.13 ± 0.054.1 ± 1.8
Lubalebale (5)0–156.2 ± 0.53.0 ± 0.7ab0.13 ± 0.07b6.5 ± 4.1
15–306.1 ± 0.62.7 ± 0.50.14 ± 0.066.1 ± 3.5
Lirugavu (5)0–156.4 ± 0.14.2 ± 1.2 a (*)0.22 ± 0.01a (*)11.0 ± 6.5
15–306.3 ± 0.32.3 ± 0.8 (*)0.16 ± 0.04 (*)8.8 ± 3.5
Mukyanga (6)0–156.4 ± 0.72.5 ± 0.7b0.12 ± 0.06b9.2 ± 3.8
15–306.6 ± 0.82.2 ± 1.10.13 ± 0.068.4 ± 5.6
Model parameters*p*-value0.940.030.110.03
**R2**0.100.640.270.57
RMSE0.350.780.043.52
Aqualf (3)0–156.4 ± 0.03.3 ± 0.70.21 ± 0.01a12.0 ± 1.2ab
15–306.2 ± 0.22.7 ± 0.70.18 ± 0.029.7 ± 2.6ab
Aquoll (3)0–156.3 ± 0.34.2 ± 2.0 (*)0.2 ± 0.06a12.6 ± 5.3a
15–306.2 ± 0.61.6 ± 0.7 (*)0.12 ± 0.057.6 ± 4.2ab
Psamment (3)0–155.9 ± 0.52.8 ± 1.50.08 ± 0.01b11.7 ± 0.3ab
15–306.5 ± 0.12.9 ± 1.90.18 ± 0.0914.5 ± 2.4a
Ustalf (6)0–156.1 ± 0.32.8 ± 0.3 (*)0.18 ± 0.02a (*)5.0 ± 3.2b
15–305.8 ± 0.42.2 ± 0.5 (*)0.14 ± 0.03 (*)3.9 ± 2.1b
Ustox (3)0–156.1 ± 0.52.7 ± 0.70.17 ± 0.04ab8.1 ± 3.0b
15–305.9 ± 0.42.7 ± 0.60.17 ± 0.045.6 ± 4.1b
Ustult (3)0–156.0 ± 0.42.3 ± 0.90.14 ± 0.04ab5.4 ± 0.2b
15–305.9 ± 0.32.7 ± 1.50.17 ± 0.085.1 ± 1.0b
Model parameters*p*-value0.630.210.080.0003
**R2**0.220.340.410.63
RMSE0.350.800.032.56

Number in parentheses shows the number of pedons considered for each classification system.

Means with different letter are significantly different at p < 0.05 according to Tukey's HSD test.

Means with * in parentheses showed significant differences within group across depths.

Table 5. Statistical analysis and model parameters of additional chemical properties across the two classification systems.

Classification systemDepth (cm)Potassium (Cmolc/kg)Sodium (Cmolc/kg)Calcium (Cmolc/kg)CEC (Cmolc/kg)Base saturation (%)
Mutaala (7)0–150.37 ± 0.200.18 ± 0.035.39 ± 1.218.17 ± 5.43b60.52 ± 26.76
15–300.24 ± 0.070.29 ± 0.046.04 ± 3.818.46 ± 6.74a56.06 ± 20.81
Lubalebale (5)0–150.54 ± 0.420.39 ± 0.215.85 ± 1.1718.42 ± 9.96ab48.23 ± 12.66
15–300.34 ± 0.130.44 ± 0.274.51 ± 0.9119.18 ± 9.16a45.3 ± 17.48
Lilugavu (5)0–150.64 ± 0.230.24 ± 0.139.46 ± 6.1928.3 ± 13.90a42.97 ± 24.74
15–300.66 ± 0.510.57 ± 0.4912.15 ± 10.8420.95 ± 9.0481.18 ± 12.84
Mukyanga (6)0–150.40 ± 0.210.46 ± 0.695.35 ± 0.9511.57 ± 6.85b55.98 ± 22.49
15–300.31 ± 0.210.47 ± 0.856.53 ± 3.7313.26 ± 8.00a63.83 ± 21.49
Model parameters*p*-value0.090.750.110.010.74
**R2**0.240.100.240.720.1
RMSE0.240.575.548.4032.20
Aqualf (3)0–150.55 ± 0.050.25 ± 0.04 ab(*)6.36 ± 0.9524.90 ± 7.61a (*)39.17 ± 9.47
15–300.48 ± 0.070.16 ± 0.01 (*)5.13 ± 1.0622.24 ± 7.49 (*)34.96 ± 5.83 b
Aquoll (3)0–150.61 ± 0.450.36 ± 0.02a11.29 ± 7.9627.54 ± 20.04a52.44 ± 32.40
15–300.72 ± 0.740.86 ± 1.1316.51 ± 19.2117.23 ± 10.7789.52 ± 17.05a
Psamment (3)0–150.32 ± 0.250.03 ± 0.01c4.97 ± 0.695.73 ± 3.50a51.63 ± 19.31
15–300.23 ± 0.140.05 ± 0.025.95 ± 0.097.89 ± 1.8575.56 ± 7.46ab
Ustalf (6)0–150.38 ± 0.200.08 ± 0.07 c(*)5.07 ± 0.618.43 ± 4.59 a(*)56.64 ± 22.52
15–300.28 ± 0.180.16 ± 0.08 (*)4.82 ± 0.477.47 ± 4.89 (*)49.40 ± 13.92ab
Ustox (3)0–150.42 ± 0.430.14 ± 0.08bc5.81 ± 1.4316.12 ± 10.82a42.39 ± 5.95
15–300.29 ± 0.130.15 ± 0.074.78 ± 0.9019.46 ± 11.2137.24 ± 8.30b
Ustult (3)0–150.52 ± 0.220.15 ± 0.10bc4.55 ± 0.494.8 ± 0.01a34.22 ± 11.94
15–300.29 ± 0.060.15 ± 0.104.55 ± 0.496.25 ± 3.7949.21 ± 20.45ab
*p*-value0.780.0010.050.0010.16
**R2**0.190.580.430.790.36
RMSE0.260.285.397.5527.14

Number in parentheses shows the number of pedons considered for each classification system.

Means with different letter are significantly different at p < 0.05 according to Tukey's HSD test.

Means with * in parentheses showed significant differences within group across depths.

Within-group comparisons across the two depths to assess whether vertical variation was captured for the farmer-based classification showed significance for SOM and N only in Lirugavu (Table 4). The Soil Taxonomy suborders showed significant vertical stratification in SOM, N, Na+, and CEC (Tables 4 and 5). Soil organic matter stratification was captured within the Aquoll and Ustalf suborders (Table 4). Total N stratification with depth was captured within Ustalfs (Table 4). Significant Na+ and CEC variation with depth was captured for Aqualfs and Ustalfs (Table 5).

4 Discussion

The farmer-based soil classification system is not a standardized scientific process but a practical-based and lived experience embedded in interaction with the land. The word itaaka, which they use to identify soil, is polysemic. It refers not only to the material substrate on the earth's surface, but also to land as a productive, managed space (Nabirye, 2009). The dual meaning tied to this word corroborates the farmers' holistic understanding that land and soil are inseparable. This aligns with other studies which noted soil and land as a multidimensional component of the environment (Barrera-Bassols et al., 2006; Chukov and Yakovlev, 2019; Postigo and Valdez, 2022). Taking this into consideration indicates that the farmers' knowledge considers the soil, water, climate, and topography as being dynamically interconnected. Such integrative thinking closely parallels Jenny's (1941) soil-forming factors model. Thus, the farmers' knowledge system is also grounded in the same environmental concepts that underpin formal pedological theory.

Four distinct soil groups emerged from the farmer-based soil classification. These four soil groups were based on functional characteristics like perceived fertility and crop performance, as well as visual cues like color, texture, and stoniness. Unlike formal soil taxonomies, the soil classification system had no hierarchical classification levels. The lack of hierarchy in these local systems was also noted by Kyebogola et al. (2020) in their study of indigenous classification system in Masaka District. Masaka is located approximately 250 km southwest of Kamuli District. To an extent, a diagnostic criterion like what is used in formal systems emerged. Lirugavu, a local term used for dark color, was used as a diagnostic surface feature for Mutaala and Mukyanga. As a diagnostic feature, it implies a dark, organic matter-rich horizon. Lirugavu was functionally similar to a mollic epipedon. Lirugavu, became a distinct soil group, and soils with a sufficiently thick, black horizon categorized under it.

Three out of the four farmer-based soils did not align with a single suborder when compared with USDA Soil Taxonomy suborders. This non-perfect relationship shows how broad and overlapping the farmer-based classification is. For the farmers, the structuring is embedded in the language. The farmers condense complex pedological realities into terms that capture several soil attributes at once, hence the broad soil groups (Kyebogola et al., 2020). The names encode a nuanced understanding of key properties such as fertility, texture, stoniness, and even landscape position. Because of this, these soil groups are functionally sufficient and broad enough to capture inter-class variation yet specific enough for the locals to make informed agricultural decisions. As a result, they may serve as locally derived equivalents to the soil series used in Soil Taxonomy. Just as soil series guide decisions about crop selection and input application, these farmer-based classes function as the primary framework through which farmers in this region make strategic land management decisions.

Noteworthy, like formal soil taxonomies, farmers' classification system extended beyond the topsoil and incorporated vertical variation, as shown in Figure 3. Farmers could differentiate Mutaala from Lubalebale with almost identical surface characteristics but differing depth to a root-limiting plinthite layer. The farmers referred to plinthite as “kisilikoko”. Mutaala typically had a deeper profile, often with an intervening murram layer before plinthite. Whereas in Lubalebale, plinthite occurred closer to the surface, often exposed due to erosion. The plinthic concretions and exposed quartz lead to the stony nature of this soil. The use of subsurface horizon features in local soil knowledge has also been documented by Sandor and Furbee (1996) in Peru and Buthelezi-Dube et al. (2018), only when the soils were shallow, as in the case of Lubalebale in this study.

Functionally, farmers' perception of soil fertility and crop performance also influenced their classification. Farmers assess soil fertility through observable phenomena such as weed vigor and biomass, crop diversity, and consistently higher yields (Kyebogola et al., 2020). This indicates an outcome-oriented understanding of soil productivity. Across all three villages, the farmers consistently ranked Lirugavu as the most productive, followed by Mutaala, Mukyanga, and Lubalebale as the least productive. Lirugavu was considered the most fertile because of its dark color, which in formal pedology is associated with organic matter richness. This finding adds to several studies that noted the use of color as an indicator of soil fertility (Barrera-Bassols et al., 2006; Kyebogola et al., 2020; Jimenez et al., 2021; Murage et al., 2000; Nuwategeka and Nyeko, 2019).

Considering that maize (corn) is a major staple and widely cultivated crop in Kamuli District, we utilized the Corn Suitability Rating (CSR2) to evaluate the productivity of the farmer-based soil groups. The CSR2 is an index initially developed in Iowa that integrates soil characteristics to assign points based on the soil's capacity to grow maize (Burras et al., 2015). The higher the points, the better suited the soil is to grow maize. The CSR 2 points aligned with the farmers' productivity ranking. Lirugavu had the highest rating, while Lubalebale had the lowest, as shown in Figure 5. The consistency between the qualitative assessment of the farmers and the quantitative assessment affirms that farmers, through their assessment, can make soil quality evaluations even without formal tools (Anderson et al., 2023; Yageta et al., 2019).

Variation in CSR2 values across the soil groups in the farmer-based soil classification in Kamuli District. The soils are arranged according to the sequential position along the landscape, from upper to lower landscape positions. Higher CSR2 values indicate better suitable for maize production.
Figure 5. Variation in CSR2 values across the soil groups in the farmer-based soil classification in Kamuli District. The soils are arranged according to the sequential position along the landscape, from upper to lower landscape positions. Higher CSR2 values indicate better suitable for maize production.

Through the soil mapping exercise, it became evident that the farmers' knowledge of soil distribution is spatially grounded, as shown in Figure 4. Farmers across all three villages associated specific soils with certain landscape positions, demonstrating a tacit awareness of topography. Soil landscape relationships have been reported in other studies. Farmers in South Africa produced local soil maps closely aligned with geomorphic units (Buthelezi-Dube et al., 2018). In Mexico, farmers linked soils to landscape positions and, as a result, used spatial knowledge to inform their farming decisions (Barrera-Bassols et al., 2006). In Masaka, Kyebogola and colleagues showed two farmer-informed toposequences (Kyebogola et al., 2020). The plausibility of soil-landscape associations in Kamuli District is further supported by the historical use of the catena concept in earlier regional mapping efforts (Borden et al., 2020; Radwanski and Ollier, 1959).

The two systems were assessed for their ability to differentiate different soil groupings based on noted physical and chemical properties. The limitation of the sample sizes, which prevented comparison to Soil Taxonomy at the subsequent classification levels, is acknowledged. The farmer-based classification separated soils by depth of mollic colors, whereas the suborders did not. Lirugavu and Mutaala were separated based on this property. This further affirms, how visually-based the farmer-based system is since Lirugavu, a black soil, had a significantly thicker mollic colored horizon. Additionally, Lirugavu are in a depositional area where organic matter rich sediment is deposited. Mutaala is located upland and susceptible to erosion. Texture has been widely documented as a key criterion used in indigenous soil classification systems (Abdulrashid and Yaro, 2014; Barrera-Bassols et al., 2006; Buthelezi-Dube et al., 2018; Ishida et al., 2001; Kyebogola et al., 2020; Nuwategeka and Nyeko, 2019). In Kamuli District, Mukyanga refers to very sandy soils. Interestingly, statistical comparisons revealed no significant results in sand across all the soil groups. However, Mukyanga had the highest average sand content, but the groups differed in clay content. Mukyanga was separated from Mutaala by clay, with the latter having the lowest clay content. This suggests that while farmers may not isolate texture components with the precision of laboratory analysis, their classifications align with functional differences, particularly those that affect soil workability. The lack of statistical significance across the suborders is attributed to the design of the USDA Soil Taxonomy. At the suborder level, group differentiations are based on diagnostic horizons, moisture regime, and drainage (Soil Survey Staff, 2022), not specifically particle size distributions. Texture becomes prominent at the family level, where particle size and mineralogical distinctions are used to refine classification. Yost and Fox (1981) noted that increasing classification detail in hierarchical systems, such as the USDA Soil Taxonomy, progressively enhances the ability to explain variation in soil properties.

Lirugavu, which the farmers considered the most fertile, consistently showed superior nutrient levels and fertility indicators. The farmer-based classifications showed the ability to statistically differentiate soil groups based on SOM, N, and CEC (Tables 4 and 5). Most of these differences were observed at 0–15 cm depth indicating that the farmer-based classification relies on surface characteristics (Hart and Mouton, 2005). The consistent significance of SOM, N, and CEC is attributed to their strong interdependence; N is closely linked to SOM through mineralization processes (Khalil et al., 2005; Wibowo and Kasno, 2020; Yang et al., 2022). Cation exchange capacity is also influenced by the quantity and quality of soil organic matter (Gruba and Mulder, 2015; Mtama et al., 2018; Solly et al., 2020). Cation exchange capacity also showed significance between groups at 15–30 cm, unlike SOM and N. Across all groups, it was generally higher at 15–30 cm depth, attributed to an increase in clay content with depth. Clay illuviation, although not reported directly in this study, is a well-established pedological process in the region. Vertical stratification of soil properties was not very well captured within the farmer-based classification. Management practices such as tillage (especially at the start of every growing season) mix the surface layers. As a result, differences in properties to 30 cm are masked. Additionally, farmer-based classifications rely heavily on surface observations with minimal or no laboratory testing; thus, subtle vertical variation may go unrecognized. In Lirugavu, which are inherently fertile, significant vertical variation was evident for SOM and N. Overall, the farmer-based approach captured more surface-level variation between groups and minimal vertical variation within groups.

The Soil Taxonomy suborders showed statistical significance for N, P, and CEC between groups at the 0–15 cm depth (Tables 4 and 5). Statistical significance was noted at 15–30 cm for P and base saturation between groups. Overall, the suborders differentiated soils on five of the nine measured chemical properties, indicating a better capacity to capture this variation than the farmer-based approach. The performance can be attributed to the system's grounding in pedogenic processes and the use of diagnostic horizons (Soil Survey Staff, 2022). Suborders often show overlap in chemical characteristics, according to Yost and Fox (1983). This suggests that although significant basic patterns are captured at the suborder level, the granularity required for management decisions is lacking. Vertical variation within soil groups was more pronounced in USDA Soil Taxonomy than in the farmer-based classification. Organic matter was considered significant in Aquolls and Ustalfs. The vertical stratification of SOM in Aquolls is attributed to humification of the surface horizons, while the poor drainage leads to a decline at depth. Suborders of Alfisols, that is, Ustalfs and Aqualfs, also showed significant variations with depth. Alfisols are characterized by clay accumulation at depth (Soil Survey Staff, 2022). This clay accumulation at depth enables the USDA Soil Taxonomy to capture the vertical variation within Alfisol suborders for properties like sodium and CEC. However, greater variation can be captured at the lower classification levels in the USDA Soil Taxonomy, as Yost and Fox (1981, 1983) noted.

The farmer-based approach and the USDA Soil Taxonomy capture different yet complementary dimensions of soil systems which have important implications for soil management. The soil group Lubalebale was consistently classified as Ustox. The convergence between the two systems means management of Ustox can be transferred to Lubalebale with confidence. The other soil groups, despite not corresponding with a specific suborder, other aspects such as landscape position can be leveraged as a bridge. Landscape position and topography influence soil erosion risk and moisture retention which have implications for soil management. Upland soils, Mutaala and Lubalebale are prone to runoff and soil erosion, requiring conservation-oriented strategies such as reduced tillage, mulching, cover cropping. Mukyanga and Lirugavu, are at risk of waterlogging and nutrient leaching. Farmers managed these areas by using raised mounds to plant their crops while others often planted sugarcane and rice. Planting sugarcane and rice in these soils reflects adaptive knowledge where rather than attempting to eliminate the excess moisture, the farmers align crop selection with hydrologic behavior.

This adaptive approach reflects an understanding of ecological matching, where farmers align crops grown with inherent landscape capability. The Food and Agricultural Organization (FAO) frameworks for land evaluation (FAO, 1976, 2007) incorporate a similar school of thought. Further complementarity was evidenced through a quantitative assessment of the farmer-based productivity rankings with CSR2. Corn Suitability rating (CSR2) is an index that integrates the whole soil profile, particle size, available water-holding capacity, field conditions (sloping, erosion risk, and flooding risk) into a single productivity score (Burras et al., 2015). Farmers' rankings aligned with CSR2 values. The coherence demonstrates that farmer-based soil classifications, like scientific indices, synthesize complex biophysical information into holistic assessments of productive potential. The alignment between local knowledge and scientifically measured indicators has been noted in other studies (Hualpa et al., 2025; Izadi and Aghamir, 2024; Kolawole and Cooper, 2022; Musinguzi et al., 2015).

Despite differences in epistemological origin, where the farmer-based classifications are rooted in experiential observation, local functionality, and place-based practice and USDA Soil Taxonomy is based in standardized methods, genetic relationships and principles. Both systems rely on diagnostic features such as color, texture, and drainage to differentiate soils. Additionally, the recognition of spatial relationships and productivity assessments represent domains of convergence between the two systems. Differences persist in how each system handles soil properties as shown in Tables 4, 5. It is precisely this combination of convergence and divergence that constitutes genuine complementarity. Areas of agreement provide a reliable basis for sharing and applying knowledge across systems. Where they diverge, each system can offer insights that the other lacks. As a result, when combined, both systems enable a more adaptive, context-specific approach to soil management with direct implications for agricultural extension, technology transfer, and participatory land use planning.

5 Conclusion

This study provides results from a farmer-based soil classification scheme in Kamuli District. The farmers in the three villages classified their soils into four major groups, that is, Lirugavu, Lubalebale, Mukyanga, and Mutaala. The farmers utilize these soil groups to navigate landscapes strategically in terms of crop allocation, timing operations, and making management decisions. The four farmer-based soil groups were associated with six soil taxonomy suborders. Practical and observable characteristics, such as texture, color, crop performance, and fertility were used to differentiate the soil groups. Embedded within the farmer-based soil classification system are insights also reported within formal systems, highlighting shared epistemology. This alignment indicates the potential for integrating the farmer-based understanding with the diagnostic rigor of the scientific taxonomies. This can start with recognition of the farmer-based categories as functional management zones while scientific rigor provides the quantitative validation and predictive tools to anticipate new crops and changes in practices.

In terms of differentiating soils based on soil physical and chemical properties, the two systems performed differently. The farmer-based soil classification groups performed better in terms of morphology (mollic colors) and clay. Chemical properties variation was better captured by the Soil Taxonomy suborders. Perhaps greater variation would have been captured by other classification levels, such as the great group or even family, which was a limitation of sample size acknowledged in this study. The differences between the two systems reflect the variation in strategies used, contextual needs, and intended purposes. Future research could extend these insights to a culturally and linguistically similar region, such as Jinja District, with more recent, updated soil maps. This would help assess the broader applicability of the relationships identified here and make a case for a hybrid approach between soil information systems.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Lejone John Ntema, University of South Africa, South Africa

Reviewed by: Weiyu Qu, University of Emergency Management, China

Teshome Yitbarek, Wolkite University, Ethiopia

Statements

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding

The author(s) declared that financial support was not received for this work and/or its publication.

Data availability

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions

FA: Conceptualization, Visualization, Formal analysis, Software, Investigation, Methodology, Writing – original draft, Writing – review & editing, Data curation. RW: Software, Conceptualization, Writing – review & editing, Writing – original draft, Investigation, Methodology. JM: Writing – original draft, Writing – review & editing, Investigation, Methodology, Software. SK: Software, Writing – original draft, Writing – review & editing, Investigation, Methodology. AA: Methodology, Investigation, Writing – review & editing, Writing – original draft. CB: Project administration, Supervision, Visualization, Methodology, Writing – review & editing, Conceptualization, Funding acquisition, Investigation, Writing – original draft, Resources.

Acknowledgements

We sincerely thank the people of Kamuli District who allowed us to visit their fields. Additionally, we thank them for their time to share their knowledge. We also appreciate the contributions of Mrs. Mbeiza Moureen, Mr. Musobya Lawrence, Mr. Bakali Waiswa, and the Local council leadership of the respective villages, who mobilized the logistics and translation that made this study possible. Special thanks go to the Center for Sustainable Rural Livelihoods, the Iowa State University- Uganda Program, and the Department of Agronomy for providing accommodation and office space during fieldwork and writing of this manuscript.

References

  1. Abdulrashid L., Yaro A. (2014). Local theories on soil formation and classifications by farming communities of semi-arid areas of Katsina State, Nigeria. J. Environ. Earth Sci. 4, 65 – 71.
  2. Achieve Global Safaris (2025). Tribes in Uganda. Achieve Global Safaris. Available online at: (Accessed October 25, 2025).
  3. Akitwine F. (2021). Pedology of Typical Soils of Kamuli District and Pedotransfer Functions to Predict Cation Exchange Capacity in Limited-Resource Settings [Master's thesis]. Ames, IA: Iowa State University. doi: doi:10.31274/td-20240329-574
  4. Akitwine F., Wokibula R. A., Mtama J. G., Burras L., Anderson A., Timbas N. L.. (2025). Cropping Patterns and Pedology in Kamuli District, Eastern Uganda. CANVAS 2025, Salt Lake City, UT. Available online at: (Accessed December 10, 2025).
  5. Anderson A. D., Akitwine F., Streeter M., Timbas N. L., Burras L. C. (2023). “Linking a farm-scale soil toposequence to farmer interpretations and societal implications in Kamuli District, Uganda,” in A. D. Anderson, ed Pedology as a Scientific, Teachable, and Artistic Discpline in the 21st Century (pp. 18–32). [PhD Dissertation]. Ames, IA: Iowa State University. Retrieved from: (Accessed December 10, 2025).
  6. Baize D., Girard M. C. (2008). Referentiel pedologique 2008. Edition Quae. Paris: Institut National de la Recherche Agronomique.
  7. Barrera-Bassols N., Zinck J. A., Van Ranst E. (2006). Local soil classification and comparison of indigenous and technical soil maps in a mesoamerican community using spatial analysis. Geoderma 135, 140 – 162. doi: doi:10.1016/j.geoderma.2005.11.010
  8. Batte A. G., Muwanga A., Sigrist P., Owor M. (2008). Vertical electrical souding as an exploration technique to improve on the certainty of groundwater yield in the fractured crystalline basement aquifers of eastern Uganda. Hydrogeol. J. 16, 1683 – 1693. doi: doi:10.1007/s10040-008-0348-4
  9. Bentley J. W., Boa E., Vilca P., Stonehouse J. (2004). “Triangulation with Técnicos: a method for rapid assessment of local knowledge,” in Investigating Local Knowledge; New Directions, New Approaches, ed. P. Sillitoe (London: Routledge), 45 – 64. doi: doi:10.4324/9780429199387
  10. Black A. W. (2000). Extension theory and practice: a review. Aust. J. Exp. Agric. 40, 493 – 502. doi: doi:10.1071/EA99083
  11. Borden R. W., Baillie I. C., Hallet S. H. (2020). The East African contribution to the formalisation of the soil catena concept. Catena 185: 104291. doi: doi:10.1016/j.catena.2019.104291
  12. Bouyoucus G. J. (1951). A recalibration of the hydrometer method for making mechanical analysis of soils. Agron. J. 43, 434 – 438. doi: doi:10.2134/agronj1951.00021962004300090005x
  13. Braimoh A. K. (2002). Integrating indigenous knowledge and soil science to develop a national soil classification system for Nigeria. Agric. Human Values 19, 75 – 80. doi: doi:10.1023/A:1015097419295
  14. Bray R. H., Kurtz L. T. (1945). Determination of total, organic, and available forms of phosphorus in soils. Soil Sci. 59, 39 – 46. doi: doi:10.1097/00010694-194501000-00006
  15. Bremner J. M., Mulvaney C. S. (1982). “Total nitrogen,” in Methods of Soil Analysis: Part 2 Chemical andMicrobiological Properties, ed. A. L. Page (Madison, WI: American Society of Agronomy), 595 – 624. doi: doi:10.2134/agronmonogr9.2.2ed.c31
  16. Buol S. W., Denton H. P. (1984). “The role of soil classification in technology transfer,” in Soil Taxonomy- Achievements and Challenges, Vol 14, eds. R. B. Grossman, R. H. Rust, and H. Eswaran (Wiley for the Soil Science Society of America), 29 – 43. doi: doi:10.2136/sssaspecpub14.c4
  17. Burras L. C., Miller G. A., Fenton T. E., Sassman A. M. (2015). Corn Suitability Rating 2 (CSR2) Equation and Component Values. Ames, IA: Iowa State University. Retrieved from: (Accessed December 10, 2025).
  18. Buthelezi-Dube N. N., Hughes J. C., Muchaonyerwa P. (2018). Indigenous soil classification in four villages of eastern South Africa. Geoderma 332, 84 – 99. doi: doi:10.1016/j.geoderma.2018.06.026
  19. Chapman H. D. (1965). “Cation-exchange capacity,” in Methods of Soil Analysis: Part 2 Chemical and Microbiological Properties, ed. A. G. Norman (Madison, WI: American Society of Agronomy), 891 – 901. doi: doi:10.2134/agronmonogr9.2.c6
  20. Christie M. E., Parks M., Mulvaney M. (2016). Gender and local soil knowledge: Linking farmers' perceptions with soil fertility in two villages in the Philippines. Singap. J. Trop. Geogr. 34, 6 – 24. doi: doi:10.1111/sjtg.12134
  21. Chukov S., Yakovlev A. (2019). Soil and land categories in the modern legislation of Russia. Eurasian Soil Sci. 52, 865 – 871. doi: doi:10.1134/S1064229319070020
  22. Duvall C. S. (2008). Classifying physical geographic features: the case of maninka farmers in southwestern mali. Geografiska Annaler: Series B, Human Geograp. 90, 327 – 348. doi: doi:10.1111/j.1468-0467.2008.00297.x
  23. FAO (1976). A Framework for Land Evaluation. Rome: FAO.
  24. FAO (2007). Land Evaluation: Onwards to a Revised Framework. Rome: FAO.
  25. Gowing J., Payton R., Tenywa M. (2004). Integrating indigenous and scientific knowledge on soils: recent experiences in Uganda and Tanzania and their relevance to participatory land use planning. Uganda J. Agricult. Sci. 184 – 191.
  26. Gruba P., Mulder J. (2015). Tree species affect cation exchnage capacity (CEC) and cation binding properties of organic matter in acid forest soils. Sci. Total Environ. 511, 655 – 662. doi: doi:10.1016/j.scitotenv.2015.01.013
  27. Hart T., Mouton J. (2005). Indigenous Knowledge and its relevance for agriculture: a case study in Uganda. Indilinga African J. Indigen. Know. Syst. 4, 249 – 263. Retrieved from: (Accessed July 25, 2025).
  28. Hualpa G., Carrión-Paladines V., Jiménez W., Capa-Mora D., Quichimbo P., Fierro N.. (2025). Farmers' indigenous knowledge of soil management in an altitudinal gradient in Southern Ecuador. Sustainability 17: 17114983. doi: doi:10.3390/su17114983
  29. Hunnington H. P. (2000). Using traditional ecological knowledge in science: methods and applications. Ecol. Appl. 10, 1270 – 1274. doi: doi:10.1890/1051-0761(2000)010[1270:UTEKIS]2.0.CO;2
  30. Ingram J. (2008). Are farmers in England equipped to meet the knowledge challenge of sustainable soil management? An analysis of farmer and advisor views. J. Environ. Manage., 86, 214 – 228. doi: doi:10.1016/j.jenvman.2006.12.036
  31. Ishida F., Tian G., Wakatsuki T. (2001). “Indigenous knowledge and soil management,” in Sustaining Soil Fertility in West Africa, Vol. 58, eds. G. Tian, F. Ishida, D. Keatinge, R. Carsky and J. Wendt (Wiley for the Soil Science Society of America). doi: doi:10.2136/sssaspecpub58.ch5
  32. Isiko A. P. (2018). Gender Roles in Traditional Healing Practices in Busoga. Leiden: Leiden University. Available online at: (Accessed December 10, 2025).
  33. IUSS Working Group (2014). World Reference Base for Soil Resources. International Soil Classification System for Naming Soils and Creating Legends for Soil Maps. Rome: Food and Agricultural Organization.
  34. Izadi N., Aghamir F. (2024). Integration of indigenous and formal knowledge in the assessment of soil quality performance assessment using multiple factor analysis in Alborz central mountains. Soil Security 17: 100172. doi: doi:10.1016/j.soisec.2024.100172
  35. Jenny H. (1941). Factors of Soil Formation: A System of Quantitative Pedology. New York, NY: Dover Publications.
  36. Jimenez L., Andrade E., Capa-Mora D., Fierro N., Quichimbo P., Jimenez W.. (2021). Traditional knowledge on soil management and conservation in the inter-Andean region, northern Ecuador. Spanish J. Soil Sci. 11: 5. doi: doi:10.3232/SJSS.2021.V11.N1.05
  37. Kamuli District Local Government (2020). District Development Plan (DDPIII) 2020/21-2024/25. Available online at: (Accessed October 25, 2025).
  38. Khalil M. I., Hossain M. B., Schmidhalter U. (2005). Carbon and nitrogen mineralization in different upland soils of the subtropics treated with organic materials. Soil Biol. Biochem. 37, 1507 – 1518. doi: doi:10.1016/j.soilbio.2005.01.014
  39. Kiptot E. (2007). Elicit indigenous knowldege on tree fodder among Maasai pastoralists via a multi-method sequencing approach. Agric. Human Val. 24, 231 – 243. doi: doi:10.1007/s10460-006-9057-6
  40. Kogge Kome G., Enang R. K., Yerima B. P. (2018). Knowledge and management of soil fertility by farmers in western Cameroon. Geoderma Reg. 13, 43 – 51. doi: doi:10.1016/j.geodrs.2018.02.001
  41. Kolawole O. D., Cooper G. W. (2022). Classifying soils: points of convergence in indigenous knowledge engagement with scientific epistemologies. Int. J. African Renaiss. Stud. 17, 125 – 145. doi: doi:10.1080/18186874.2022.2035784
  42. Komuntunga E. T., Musitwa F. (2001). “Climate,” in Agriculture in Uganda, ed. J. K. Mukibi (Kampala: Fountain Publishers), 21 – 32. Available online at: (Accessed July 25, 2025).
  43. Kyebogola S., Burras L. C., Miller B. A., Semalulu O., Yost R. S., Tenywa M. M.. (2020). Comparing Uganda's indigenous soil classification system with world reference base and USDA soil taxonomy to predict soil productivity. Geoderma Reg. 22: e00296. doi: doi:10.1016/j.geodrs.2020.e00296
  44. Mclean E. O. (1982). “Soil pH and lime requirement,” in Methods of Soil Analysis: Part 2 Chemical and Microbiological Properties, ed. A. L. Page (Madison, WI: American Society of Agronomy), 199 – 224. doi: doi:10.2134/agronmonogr9.2.2ed.c12
  45. Mtama J. G., Burras C. L., Msanya B. M. (2018). Pedotransfer functions for cation exchange capacity, available water holding capacity and soil organic carbon for representative soils of the southern highland zone of Tanzania. Int. J. Agricult. Biol. Sci. 26 – 42.
  46. Murage E. W., Karanja N. K., Smithson P. C., Woomer P. L. (2000). Diagnostic indicators of soil quality in productive and non-productive smallholders' fields of Kenya's Central Highland. Agric. Ecosyst. Environ. 79, 1 – 8. doi: doi:10.1016/S0167-8809(99)00142-5
  47. Musinguzi P., Ebanyat P., Tenywa J. S., Basamba T. A., Tenywa M. M., Mubiru D. (2015). Precision of farmer-based fertility ratings and soil organic carbon for crop production on a Ferralsol. Solid Earth 6, 1063 – 1073. doi: doi:10.5194/se-6-1063-2015
  48. Muwanga S. (2020). Influence of Selected Land Use Change and Indigenous Knowledge on Soil Quality in the Agro-Pastoral Semi-Arid Karamoja-Uganda. University of Nairobi. doi: doi:10.5539/jas.v12n4p148
  49. Nabirye M. (2009). Eiwanika ly'olusoga. Eiwanika ly'aboogezi b'Olusoga n'abo abenda okwega Olusoga. Kampala: Menha Publishers.
  50. Nelson D. W., Sommers L. E. (1982). “Total carbon, organic carbon, and organic matter,” in Methods of Soil Analysis: Part 2 Chemical and Microbiological Properties, 2nd ed, ed. A. L. Page (Madison, WI: American Society of Agronomy), 539 – 579. doi: doi:10.2134/agronmonogr9.2.2ed.c29
  51. Nulkar G. (2024). “Traditional knowledge and sustainable livelihoods,” in G. Nulkar, Ed The Economics of Sustainable Development. Singapore: Springer. doi: doi:10.1007/978-981-99-7379-8
  52. Nuwategeka E., Nyeko M. (2019). Using indigenous knowledge to model land suitability for crops in Amuru District, Northern Uganda. African J. Food Agricult. Nutrit. Develop. 19, 14805 – 14828. doi: doi:10.18697/ajfand.87.17555
  53. Ollier C. D., Harrop J. F. (1959). The Soils of the Eastern Province of Uganda. Kampala: Kawanda Research Station.
  54. Omatayo A., Musa M. W. (1999). The role of indigenous land classification and management practices in sustaining land use system in the semi-arid zone of Nigeria. J. Sustain. Agric. 14, 49 – 58. doi: doi:10.1300/J064v14n01_05
  55. Osborne J. (2010). Improving your data transformations: applying the box-cox transformation. Pract. Assess. Res. Eval. 15: 12. doi: doi:10.7275/qbpc-gk17
  56. Payton R. W., Barr J. F., Martin A., Sillitoe P., Deckers J. F., Gowing J. W., Zuberi M. I. (2003). Contrasting approaches to integrating indigenous knowledge about soils and scientific soil survey in East Africa and Bangladesh. Geoderma 3, 355 – 386. doi: doi:10.1016/S0016-7061(02)00272-0
  57. Postigo J., Valdez S. L. (2022). The Sociocultural Construction of Soil Among Communities of the Bolivian. Altiplano: Potential for supporting transitions to sustainability. Mountain Research and Development. doi: doi:10.1659/mrd.2022.00015
  58. R Core Team (2024). A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing. Available online at: (Accessed July 30, 2025).
  59. Radwanski S. A., Ollier C. D. (1959). A study of an East African catena. Eur. J. Soil Sci. 10, 149 – 168. doi: doi:10.1111/j.1365-2389.1959.tb02339.x
  60. Rajarathinam A., Ramji M. (2023). Discriminant analysis for soil attributes. Int. J. Agricult. Stat. Sci. 19, 921 – 931. doi: doi:10.59467/IJASS.2023.19.921
  61. Ruhe R. V. (1954). Erosion Surfaces of Central African Interior High Plateaus. IINEAC.
  62. Rushemuka N. P., Bizoza R. A., Mowo J. G., Bock L. (2014). Farners' soil knowledge for effective participatory integrated watershed management in Rwanda: toward soil specific fertility management and farmers' judgmental fertilizer use. Agric. Ecosyst. Environ. 183, 145 – 159. doi: doi:10.1016/j.agee.2013.10.020
  63. Sandor J., Furbee L. (1996). Indigenous knowledge and classification of soils in the Andes of southern Peru. Soil Sci. Soc. America J. 60, 1502 – 1512. doi: doi:10.2136/sssaj1996.03615995006000050031x
  64. Santos H. G., Jacmine P. K., Anjos L. H., Oliveira V. A., Lumbreras J. F., Coelho M. R.. (2018). Sistema brasileiro de classificao de solos. Brasilia, DF. Embrapa, 2018.
  65. Schoeneberger P. J., Wysocki D. A., Benham E. C. (2012). Field Book for Describing and Sampling Soils Version 3.0. Natural Resource Conservation Service, National Soil Survey Center.
  66. Sillitoe P. (1998). Knowing the land: soil and land resource evaluation and indigenous knowledge. Soil Use Manag. 14, 188 – 193. doi: doi:10.1111/j.1475-2743.1998.tb00148.x
  67. Smith G. D. (1983). “Historical development of soil taxonomy—background,” in Pedogenesis and Soil Taxonomy, I. Concepts and Interactions, ed. L. P. wilding (Washington, DC: Elsevier), 23 – 49. doi: doi:10.1016/S0166-2481(08)70597-X
  68. Soil Survey Staff (2022). Keys to Soil Taxonomy, 13th ed. Washington, DC: Natural Resources Conservation Service, United States Department of Agriculture.
  69. Solly E. F., Weber V., Zimmermann S., Walthert L., Hagedorn F., Schmidt M. W. (2020). A critical evaluation of the relationship between the effective cation exchange capacity and soil organic carbon content in Swiss forest soils. Front. Forests Glob. Change 3: 98. doi: doi:10.3389/ffgc.2020.00098
  70. Tengo M., Austin B. J., Danielsen F., Fernandez-Llamazares A. (2021). Creating synergies between citizen science and indigenous and local knowledge. BioScience 71, 503 – 518. doi: doi:10.1093/biosci/biab023
  71. Tengö M., Brondizio E. D., Elmqvist T., Malmer P., Spierenburg M. (2014). Connecting diverse knowledge systems for enhanced ecosystem governance: the multiple evidence base approach. Ambio 43, 579 – 591. doi: doi:10.1007/s13280-014-0501-3
  72. Tenywa M. M., Ssemalulu O., Luswata C. K., Kyomuhendo P., Namakula J., Abbott E.. (2014). Farmer Decision-Making Strategies for Improved Soil Fertility Management in Maize-Bean ProductionSsystems. Focus group interviews with bean farmers in Masaka and Rakai Districts, Uganda. Unpublished.
  73. Thomas G. W. (1982). “Exchangeable cations,” in Methods of Soil Analysis: Part 2 Chemical and Microbiological Properties, ed. A. L. Page (Madison, WI: American Society of Agronomy), 159 – 165. doi: doi:10.2134/agronmonogr9.2.2ed.c9
  74. Toxqui-Roldan J. L., Ortiz-Solorio C. A., Gutierrez-Castorena M. C., Licona-Vargas L. A. (2024). Agricultural terraces in Puebla, Mexico: an ethnopedological approach. Geoderma Reg. 39. doi: doi:10.1016/j.geodrs.2024.e00873
  75. Tulibaleka P. O., Tumwesigye K., Nakayima L. S. (2021). Urbanization in Africa: integrating multiculturalism in urban development in Uganda. J. African Stud. Develop. 13, 74 – 80. doi: doi:10.5897/JASD2021.0635
  76. Tweheyo R., Kiconco M., Bamwesigye D. (2024). Climate Change Mitigation in Southwestern Uganda Using Indigenous Knowledge. Epubpreprint. doi: doi:10.20944/preprints202401.0422.v1
  77. Utter A., White A., Mendez V. E. (2021). Co-creation of knowledge in agroecology. Sci. Anthrop. 9: 26. doi: doi:10.1525/elementa.2021.00026
  78. Veenstra J., Dembe D., Nsamba J., Wasko L., Burras L., Nonnecke G.. (2015). Mapping soils in Central Uganda. Soil Survey 50, 31 – 33. doi: doi:10.2136/sh2009.1.0031
  79. Watanabe F. S., Olsen S. R. (1965). Test of an ascorbic acid method for determining phosphorus in water and NaHCO3 extracts from soil. Soil Sci. Am. J. 29, 677 – 678. doi: doi:10.2136/sssaj1965.03615995002900060025x
  80. Wibowo H., Kasno A. (2020). Soil Organic Carbon and Total Nitrogen Dynamics in Paddy Soils on the Java Island, Indonesia. IOP Conference Series Earth and Environmental Science. doi: doi:10.1088/1755-1315/648/1/012192
  81. Winowiecki L. A., Whelan M. P., McDaniel P. A., Villalobos M., Somarriba E. (2014). Local soil knowledge and its use in crop allocation: implications for landscape-scale agricultural production and conservation efforts in Talamanca, Costa Rica. Agricult. Forest. Fisher. 3, 93 – 101. doi: doi:10.11648/j.aff.20140302.17
  82. Wokibula R. A., Akiwine F., Mtama J. G., Burras C. L. (2024). Cation exchange capacity within and between pedons of the Kamuli catena, Busoga sub-region, Uganda. Soil Science Society of America Conference. San Antonio, Texas. Available online at: (Accessed December 10, 2025).
  83. Wortmann C. S., Eledu C. A. (1999). Uganda's Agroecological Zones: A Guide for Planners and Policy Makers. Centro Internacional de Agricultura Tropical.
  84. Yageta Y., Osbahr H., Morimoto Y., Clark J. (2019). Comparing farmers' qualitative evaluation of soil fertility with quantitative soil fertility indicators in Kitui County, Kenya. Geoderma 344, 153 – 163. doi: doi:10.1016/j.geoderma.2019.01.019
  85. Yang S., Jansen B., Absalah S., Kalbitz K., Chunga Castro F. O., Cammeraat E. L. (2022). Soil organic carbon content and mineralization controlled by the composition, origin and molecular diversity of organic matter: a study in tropical alpine grasslands. Soil Till. Res. 215: 105203. doi: doi:10.1016/j.still.2021.105203
  86. Yost R. S., Fox R. L. (1981). Partitioning variation in soil chemical properties of some andepts using soil taxonomy. Soil Sci. Soc. Am. J. 45, 373 – 377. doi: doi:10.2136/sssaj1981.03615995004500020029x
  87. Yost R. S., Fox R. L. (1983). Partitioning variation in chemical properties of some andepts-A comparison of classification systems. Geoderma 29, 13 – 26. doi: doi:10.1016/0016-7061(83)90027-7
  88. Zhang W., Marlene E., Meinzen-Dick R., Swallow K., Calvo-Hernandez C., Nkonya E. (2021). Soil health and gender: why and how to identify the linkages. Int. J. Agricult. Sustain. 19, 269 – 287. doi: doi:10.1080/14735903.2021.1906575

Citation

Akitwine, F., et al. (2026). Assessing a farmer-based soil classification system vis-à-vis USDA soil taxonomy suborders in Kamuli District, Uganda. Frontiers in Sustainable Resource Management, Volume 5. https://doi.org/
Received
Accepted
Published
Volume
5 - 2026

Copyright

Copyright © 2026 Akitwine, Wokibula, Mtama, Kwikiiriza, Anderson and Burras.

License: https://creativecommons.org/licenses/by/4.0/

*Correspondence: Francis Akitwine; C. Lee Burras

Article metrics

Views
0
Downloads
0
Citations
0
Shares
0

Related articles

Original ResearchThe final count-down: estimating the total value of all the world's natural resources and an estimate of how long they may lastHarald Ulrik Sverdrup, Anders Nordby, Ole van Allen, Anna Hulda Olafsdottir, Kristín Vala Ragnarsdóttir, Wouter van DierenOriginal ResearchForecasts of United States oil and gas production to 2100J. H. Laherrère, R. W. BentleyPerspectiveConservation concessions or green grabbing? A critical perspective on mega conservation investmentsYamikani Harry Makwinja, Belinda Makare