The study area covers an estimated land-take of 419 acres and is approx. 17 km northwest of Twifo Praso in the Twifo Heman/Lower Denkyira District, Central Region, Ghana and 1.2 km west of Wamaso (latitude: 5.708333 and longitude: −1.658333) (Figure 1). The dominant vegetations are thickets consisting of an impenetrable mass of shrubs, climbers, coppice shoots and young trees and soft-stemmed leafy herbs e.g. Ageratum conyzoid, Bambusavulgaris that appear in abandoned farms and cultivated lands grown to cocoa, maize, plantain, cassava and rice. The topography generally shows a gentle slope that varies from 1% -

10% whilst the elevation ranges from 40 - 70 m above sea level (asl). Minimum plateaued summits were between 87.18 - 91.77 m with the highest summits ranging from 118.40 - 123.56 m. A few seasonal streams border the study site and tend to flood during the rainy seasons. The temperature and rainfall pattern across the study area is shown in (Figure 2). The geology of the area is generally undulating with gently rolling, steeped slope topography at few places, including a Cape Coast granite and Wamaso albitized rock type (Gyamera, 2014).

The free survey method of soil investigation was adopted where the use of aeriall photography obtained from the drone were used to make interpretation coupled with massive field observations to aid in characterization of the toposequence (Beckett & Burrough, 1971). This method is conducive for the field setting as there were significant open spaces to allow for field walkovers and tracing of footpaths for broadened observations and judgement of the soil features. Current landuse features were assigned unique identification numbers. To further obtain a widened aerial view of the study area, an unmanned drone (DJI Phantom 3 Professional) was used to obtain photographs of the approximate landmass. The aerial photographs were processed in Pix 4D Mapper software together with Mission planning software (Mappilot) (Burnham, 2019) and topographic maps generated in ArcGIS environment (ArcGIS 10.7). The topographic maps were used to develop respective landuse maps. Overall, five (5) topographic maps based on observed land uses were developed. Soil profile pits (PP’s) were established randomly in each of the identified land uses along the toposequence. Five PP’s using a soil auger were sunk to obtain representative soil samples along the toposequence (Figure 3). Three discrete samples were collected vertically at varying depths across respective PP’s. The PP’s were then described per Soil Survey Manual (Soil Science Division Staff, 2017). The field-moist soil samples were collected in sealable plastic bags, with stones, roots, and large organic residues removed. Sample labels codes/IDs corresponded uniquely to the soil profile

test pits. Overall 270 samples (i.e. 18 samples obtained from 5 profile pits at 3 samples at respective depths each). They were transported to the laboratory in an ice cooler to maintain the temperature at 4˚C. The remaining organic materials were removed, soil samples air-dried at room temperature for 15 days, pulverized and sieved (2 mm mesh-size) for storage.

A summary of the various laboratory analytical methods employed are shown in Table 1. Before any analytical analysis, all glassware was washed in distilled water and Hydrochloric acid (HCl) solution and oven-dried at 105˚C. Fresh soil samples were used for each run of batch analysis to avoid cross-contamination of chemical parameters. Detailed description of analytical methods is provided as in the supplementary information.

Ata were tested for normality of distributions and homogeneity of variances before analysis (Kolmogorov-Smirnov, p > 0.05). Standard skewness and kurtosis were determined to test whether the data originated from a normal distribution. Statistical values which were outside the range of −2 to +2 were considered as departures from normality. Differences in soil physicochemical properties across the toposequnce were compared using one-way analysis of variance (one-way ANOVA) with Tukey post hoc test. Subsequently, all relationships between soil properties were evaluated using Pearson’s correlation tests and correlation coefficient (R²). Principal component analysis (PCA) was applied to examine the

degree of contribution of the variables towards overall variation within the soil’s physicochemical properties. All statistical significance was tested at (p > 0.05). The study area map was created using the ArcGIS 10.7 software (ESRI, Inc., Redlands, CA, USA software version 10.7). All statistical analysis and graphing were performed using OriginPro v2021 software (OriginLab Corporation, Northampton, MA, USA).

The descriptive statistics of soil properties across profiles pits are shown in (Table 2). Overall the data demonstrated a platykurtic distribution indicating significant variation across samples. The maximum coefficient of variations was recorded for P_av (1.00), Zn (0.98), EC (0.92) which could signify as major source and cause of variation in the soil datasets. The mean values of soil particle size decreased from sand (58.46) > clay (32.8) > silt (8.64) while maximum pH across the profile pits was recorded at 5.89.

All the pedons had moderately deep profiles (≈120 cm). PP1 and PP2 contained four generic soil horizons, PP3 was made up of three, PP4 had four whereas PP5 had three generic horizons. The pedons of PP1, PP2, and PP3 representing the summit, shoulder and mid slopes respectively were well-drained whilst the lower (PP4) and valley bottom slopes (PP5) were poorly drained. This was observed by small water pools possibly caused by intermittent heavy rains. The pedons of PP1 comprised secondary forest with thicket vegetation lying on a soil potentially

*Available Phosphorus (P_av), Cation Exchange Capacity (CEC), Electrical Conductivity (EC); Total Nitrogen (TN), Organic Carbon (OC), Exchangeable Acidity (EA), Bulk Density (BD), Coefficient of Variation (CV), Standard Deviation (SD), Standard Error of Mean (SE).

evolved from the current geologic formation. Pedons PP2, PP3, PP4 and PP5 were established in an abandoned cocoa plantation surrounded by a secondary forest with thicket vegetation while pedons.

The topsoils for all five pits had weak fine to moderate granular structures whereas the subsurface soils had moderate and strong, medium and coarse, angular and subangular blocky structures. Non-sticky and slightly sticky (wet) soil consistencies were obtained in the surface soils while sticky (wet) as well as firm (moist) consistencies were mainly obtained in the subsurface soils.

Layer boundaries were distinctively visible from the topsoil through to the last horizon of each of the pits. Burrows of ants and worm casts were also detected within the pits which is an indication of faunal pedoturbation. Lateral roots, tip roots and root hairs were present in the soil column, predominantly within the first three identified layers of the various profile pits. The soils exhibited variable colouration, both in surface and subsurface soils (Figure 4). The first layer in all the profile pits revealed significant distinction in soil colour. The colouration

varied from very dark greyish brown 10Yr 3/2 for PP5, 10Yr 3/3 dark brown for PP4, reddish-brown 5Yr 3/4 for PP3, reddish-brown 5Yr 3/2 for PP2, and 5GY 2.5/1PP greenish-black surface soils for PP1. The pedons were in constant exposure to erosion. The uplands and mid soils revealed light-coloured soils caused by erosion strips. The greatest erodibility was observed from the upper slope which exhibited coarse texture and very low SOM whereas the lower and valley bottom slopes experienced minimal erosion with mostly deep, organic-rich leached soils

The soil texture varied from sandy loam to sandy clay loam in the surface horizon for all five profiles (Table 3). The bulk densities increased proportionally to the respective profile pit depths (Figure 5).

Physicochemical properties across the various soil profile pits significant across the profile pits as determined by the corresponding slope gradients of the toposequence. pH across the PP’s were strongly acidic and ranged from 4.48 in PP1 to 5.89 in PP5 although there were no significant differences (p > 0.05) in the pH (Figure 6(A)). The SOC was unevenly distributed across the PP’s and decreased from the summit profile pits (PP1) to the valley bottom profile pit (PP5) (Figure 6(B)). In the summit pedon’s SOC ranged from 1.14% - 0.34%, 1.05% - 0.28% within the shoulder, 1.59% - 0.47% in the middle slope, 1.67% - 0.29% in the base slope and finally 1.83% - 0.07% at the toe slope. There was a significant difference (p < 0.05) in SOC concentration from PP1 to PP2 whilst no statistical significance was detected from PP3 to PP5 (>0.05). Soil available phosphorus (P_av) was averagely low from across layers of PP1 to PP4, with PP5 recording the highest value (Figure 6(C)). From the layers of PP1 to PP4, total nitrogen (TN)

was averagely high in PP1 and PP2 as concentration decreased from PP3 to PP4 (Figure 6(D)). There were significant differences in TN concentrations from PP1, PP2 and PP4 (p < 0.05) but no significant difference was observed in PP3 and PP5 (p > 0.05).

The exchangeable bases recorded in the study were plotted to obtain a piper plot in (Figure 7). It is observed the samples are classified into three types, representing the Mg-Ca-K (plane 1), Fe-Zn-Cu (plane 2) and Fe+Cu-Ca+Mg (plane 3).

A comparison of the exchangeable acidity (EA) and cation-ion exchange capacities (CEC) across the various soil PP’s is shown in (Figure 8). From PP1 EA and CEC were evenly distributed across the layers, with Layer 1 and Layer 3 recording the highest values. A similar trend was observed across PP2; however, the EA values were very low in Layer 1 but evenly distributed from Layer 2 and 3 with CEC values in Layer 1. The CEC values from layer 2 to Layer 4 were equally distributed. A similar trend of the distribution in PP2 was observed in PP3 and PP4. Meanwhile, in PP5, Layer 2 recorded values were almost half that of Layer 1 whilst in Layer 3, CEC was higher than EA values.

Pearson’s correlation analysis revealed significant correlations between most of the soil physicochemical properties, and that the correlation between soil Physico-chemical properties was consistent across the profile pits (Figure 8). The analysis showed more positive correlations amongst chemical parameters such as between Cu, Zn, Ca, Mg, Na, and similar positive correlations observed between available phosphorus, Cu, Na, and Fe. However, there was a negative correlation between TN and pH while TN showed a positive correlation with Ca, Na, Zn and OC. pH also negatively correlated with OC, EA, Ca and Na. Also, EA exhibited a negative correlation across all cations in the soil profiles. CEC on the other hand yield a positive correlation with cations in general and showed a negative correlation association with EA. There was however a difference in the correlation of soil textural properties with the exchangeable bases. Sand exhibited a positive correlation with the major ions while the opposite was observed with silt and clay soils. Similarly, bulk densities across the profile pits correlated negatively with exchangeable bases.

The weights for the variables are along the principal components are shown in

(Figure 9). According to the results of the PCA ordination, a higher variation was observed along the ordinate axes (49.9%) of principal component 1 (PC1) while the second principal component (PC2) was explained (18.4%).

The key determinants in soil characteristics are mainly the soils parent material and the degree slope. Soil properties such as particle size and pore space distribution, hydraulic conductivity among others are all inherent properties derived from the parent materials. Overall, the pedons exhibited differences in slope, water percolation and degree of water erosion. Hydrological information of the study area was previously reported by (Gyamera, 2014).

Physicochemical properties of soils across the various ecological zones in Ghana have been documented (Bationo et al., 2018). Our study area represented a typical semi-deciduous forest agro-ecological zone (Issaka, Buri, Tobita, Nakamura, & Adjei, 2012) which is characterized by moderately acidic soils to low Nitrogen and Phosphorus levels (Table 4) and results obtained were consistent with reports by Bationo (Bationo et al., 2018). The descriptive statistics revealed that most of the variation occurred at the upper and base slopes of the toposequence with relatively little change occurring in the middle-shoulder portions. Across all layers of the PPs, clay percentages increased from surface soils down the subsurface soils revealing typical pedogenic eluviation-illuviation processes particularly within the upper to middle slope profiles while the opposite was observed for soil and silt compositions (Mulugeta & Sheleme, 2010). Increasing

clay amount with depth could have also resulted in the moderate and strong, coarse, subangular and angular blocky structures as well as the firmness that increased with soil depth.

Generally, soil pH in the surface and subsurface layers of the pedons were rated as medium to low per rating scores by Landon (2014). Soil acidity assumed a descend across the pedons with corresponding depths. This could be related to the relatively higher content of basic cations observed at the subsurface layers. The highest pH, which was observed in PP5, could be due to the high solubility of ions e.g. Na²⁺ Ca²⁺ and Mg²⁺. The results are consistent with observations by (Ogbodo, 2011) that basic cations are capable of naturally displacing H⁺ from exchangeable complexes into soil solution where they are leached. The leaching process and accumulation of ions at particular soil pedons trigger changes in pH. Strongly acidic soils are likely to contain a significant amount of exchangeable Al³⁺ and H⁺ which affect plant growth (Schoeneberger, Wysocki, & Benham, 2012) whiles they become deficient in some bases such as Ca and Mg (MM & Ahmad, 2020). Such soil conditions can induce phosphate fixation and reduce the ability of micro-organisms to fix atmospheric nitrogen. There was no significant difference in pH across the pedons which could further be attributed to soils buffering and self-regulating abilities which can balance the changes of the external acid and alkali environment to some extent.

The significant colour variations across could be attributed to the impact of topography and patterns coupled with the influence of water infiltration rates, surface runoffs, erosion and also deposition of the eroded soils. The dark, brown and dark yellowish-brown surface soil colouration indicates either uniformity in SOM distribution or mineralization occurred at the same rate while dominant reddish and yellowish subsurface colours indicate the presence of hematite and goethite as forms of Fe oxides, defining ferrugination. Similar results were reported by (Townsend, Vitousek, & Trumbore, 1995). The reddish and brownish colouration indicated that soils on slopes become well-drained with adequate aeration and less saturated even under frequent rainfall. Foth (1990) attributed this phenomenon of soils to the availability of metal compounds in oxidized states (e.g. CO²⁺, Fe³⁺). The varied colours associated with the horizons in the pedons within the region of depression can also be attributed to the reduction reaction caused by water saturation, such pedons mostly are located in poorly drained and waterlogged regions where soils are water-saturated most of the time and thus tend to have the greyish colour (Foth, 1990). This scenario was observed down the profile of the fifth profile pit with surface soil colour N5/5 grey.

SOC decreased with increasing depth across all five (5) pedons with the highest values recorded in the surface soils of each. The observation was consistent with reports by (Gaudinski et al., 2000) who indicated that such properties are typical of SOC content in moist soils. Previous environmental conditions, in the study area, could also influence SOC content due to “soil’s memory” (Janzen, 2005). Overall, the soils were classified as medium in SOC as values obtained were below 2.0% (Enwezor et al., 1988). They were rated as low in accordance per (Landon, 2014) which indicates that the soils are vulnerable to water erosion. The influence of climatic variables according to (Hobley et al., 2015) is eminent mostly in surface soils and diminishes moving down a pedon making site and landuse factors more important in soil organic carbon storage down a pedon. Earlier reports have indicated that SOC less than 1.16% for tropical soils is an indication of soil degradation involving a highly raised risk of soil erosion (Barrow, 1991). This effect could be attributed to loss of TN across the pedons as explained by (Sheleme & Singh, 2011) concerning the transport of organic matter and variation in clay content of soils. Further, (Bürkle, 1989; Jobbágy & Jackson, 2000) linked site factors such as soil type, texture and soil mineralogy as key determinants in the distribution of soil organic matter along pedon. Organic matter correlated positively with clay content and negatively with sand. This is consistent with the findings of this research where soil organic carbon decreased down the profiles and this could be attributed to climate, landuse, soil type and mineralogical properties of the parent material.

The bulk density increased down gradient the pedons possibly attributed to high SOC in the surface layers. This effect reported by (Worku & Bedadi, 2016) where SOC is found lighter than mineral matter tend to improve soil porosity but reduces soil bulk density.

Available phosphorus (P_av) was observed to decrease down the profiles pits due to increases (P_av) in clay contents down the profiles (Brady & Weil, 2008). The relatively higher in surface horizons, as compared to sub-surface horizons could also be attributed to the difference in organic matter contents. Available P experienced some inconsistencies with depth but generally decreased with increasing depth. This trend however was vice-versa along with the various landscape positions. The correlation between organic carbon and phosphorus was positive indicating proportionality between SOC and (P_av). Also, the high occurrence of mineralization could have facilitated the liberation of phosphate ions outside the soil matrix and available to crops. These phosphorus releases also react with Al³⁺ and Fe²⁺ in acid soils to form aluminium phosphate ( AlPO 4 − 2 ) which are not readily available for plants uptake.

A trilinear plot is an important tool to visualize the interaction of major ions across the layers of the soil profiles pits as shown in (Figure 8). Much of the ions were clustered along the Fe²⁺+Cu²⁺, Mg²⁺ and Cu²⁺ axes of the planes. This may infer that the exchangeable bases across the study area were borne from a Fe²⁺, Cu²⁺, and Mg²⁺ laden rock weathering process. Generally, the exchangeable bases were of lower values and fall within the threshold for critical limits for deficiency and hence the likelihood for fertilization as proposed by (Musinguzi et al., 2015). Sodium (Na²⁺) in many circumstances is utilized by plants as a partial substitute for potassium. It is not an essential nutrient. Its absence or presence in very limited quantities is therefore not usually detrimental to plant nutrition. However, Na²⁺ in large quantities particularly in proportion to other cations present can affect crops and the physical conditions of the soil. The CEC of the soils was reasonably high as per (Landon, 2014) and the exchange complex of the soils was dominated by Ca followed by Mg²⁺, K²⁺ and Na²⁺. The content of exchangeable cations however experienced fluctuations with high and low values with increasing soil depth, which could be attributed to their leaching from the surface horizon down to the sub-surface according to (Gebrekidan & Negassa, 2006). The Ca²⁺:Mg²⁺ ratio of the soils was in the range of 3.2 - 2.13 from the surface soil down to the subsurface soils in the summit soil, 2.99 - 6.09 in the upper slope, 1.94 - 2.78 in the middle slope, 3.18 - 1.11 in the lower slope and then 1.20 - 0.56 in the toe slope. The exchangeable K²⁺, Ca²⁺ and Mg²⁺ of the soils had some being above the critical values and some also below the critical values (Landon, 2014). K²⁺ and Na²⁺ values both across and along the various landscape positions fell below 0.6 and 1.0 respectively and therefore according to (Landon, 2014) are low.

Determining soil health by assessing disparities in soil physicochemical properties is important as it reveals necessary interactions, associations and behaviour amongst the various properties (Havlin & Heiniger, 2020). Our results reflected these features as there were significant associations by correlation effect amongst most of the soil properties. The results have shown that there are positive and negative associations amongst soil physicochemical properties in the semi-deciduous forest. pH for instance is negatively correlated to Ca²⁺, Na²⁺, EA and TN while it is positively correlated to other ions such as Mg²⁺, Fe²⁺, Cu²⁺, K²⁺ (p < 0.05). The impact of pH is critical. Excessive pH tends to hinder soil microbial and enzyme processes hence suitable pH ranges are beneficial to support the entire pedon. Generally, pH showed positive correlations with soil textures (sand and silt) but a negative association with clay. This could be due to high mineralization (ions) which exist in clay samples, causing a inbalance across the pedons (Yu et al., 2017). A similar negative correlation was observed between pH and EA, TN and Cu²⁺ Similar, observations were made by (Ulakpa, Ulakpa, & Eyankware, 2021).

The overall association between SOC and soil mineralization is described by (Six, 2002). Near the surface, SOC content is highest and so fine minerals are most likely to be saturated with SOC, limiting their retention capacity. Below the surface SOC content generally decreases, so that saturation of fine minerals is less likely to be an issue and mineralogy becomes more relevant to SOC storage (Grimm, Behrens, Märker, & Elsenbeer, 2008). The association between phosphorus, SOC and other mineral ions were positively correlated but not significant (p > 0.05). This could be related to low which was more significant in profile pits at the toe of the slope. At the toe and below the pedons, is easily released from the SOC matrix undergoing mineralization and accumulation. The high can also be determined by high percentage of clay fractions as the pedons descend and high leaching effects due to pH variations which causes low binding of minerals. Similarly, correlated positively to TN and K²⁺ and the results conform to other reports by (Jiang et al., 2021).

The results of the study further revealed that the TN, SOC and P_av exhibited positive correlation, while the association between TN and SOC were not significantly related (p > 0.05) unlike TN and P_av (p < 0.05). A similar association is described in (Agboola & Corey, 1973). The association can be explained by the fact that SOC as part of soil organic matter (SOM) in the study area influences a series of soil properties and nutrient cycling, while soil enzymes are mainly adsorbed on the soil organic matter particles or combined with humus (Liu et al., 2020). High SOC is also reported to cause an increase in TN and K²⁺ levels (Wang & Huang, 2001) as observed from the results of this study.

In this study, we showed that there are significant variations with soil physicochemical parameters, contributed by variations in profile pits across the toposequence. The eigenvalues from the PCA of pH, Ca²⁺ and Mg²⁺ which were greater than 1 could have together contributed the total variance of 49.9%. The position of the variables in the multivariate space further confirms the results of Pearson correlation analysis and association significant positive correlation amongst soil exchangeable ions in the study area. The second principal component explained 18.4% of the total variation with a positive correlation to EA and clay contents while exhibiting a negative correlation with BD. The findings are consistent with reported studies in (Sadiq et al., 2021).