The Batéké plateaus, located in the southeast of the Republic of Congo, are characterized by a savannah landscape with sandy soils low in organic matter. This region is known for its acidic, low-fertility soils, which are suitable for grasslands and open forests but vulnerable to degradation. The climate is tropical, with an extended dry season that affects soil formation processes and limits natural plant biodiversity [1] [6] . Average annual temperatures range between 24˚C and 28˚C, while annual precipitation varies from 1200 to 1500 mm, with irregular distribution. The soils of the Batéké plateaus, primarily composed of Quaternary sands, have low organic matter and essential nutrient levels, necessitating specific agricultural management practices to maintain productivity [7] .

Three types of plantations were examined: acacia, eucalyptus, and cassava. The plots, selected from 58 within the Bambou-Mingali state reserve, include acacia plantations aged 2, 4, 5, and 6 years; eucalyptus plantations aged 4, 5, 6, and 7 years; and cassava plantations aged 5 to 15 months. These plots were chosen based on criteria such as age, current condition, planting history, and specific species [7] .

In each plot, soil was collected from three points at two different locations to create a composite sample. At least two composite samples were taken per plot, resulting in a total of 28 samples for MIR spectroscopy analysis across all plantations. Soil samples were collected at a depth of 0 to 30 cm, air-dried, and sieved at 2 mm to homogenize the samples and minimize variations due to particle size and residual moisture [17] .

Analyses were conducted using mid-infrared (MIR) spectroscopy with a BRUKER Vertex70 FTIR spectrometer, equipped with an MCT detector, in diffuse reflection mode. Each spectrum represents the average of 32 scans, with a spectral resolution of 4 cm⁻¹, allowing for the non-destructive analysis of soil chemical composition. As noted by [8] , this method is effective for soil analysis, and additional studies, such as [24] , have demonstrated its utility in predicting soil fertility indices.

The raw spectral data were improved through several preprocessing methods: absorbance correction, normalization to reduce variations in sample quantity, and application of derivatives to accentuate spectral variations and reduce noise [16] [18] ). The preprocessing and spectral processing methodology aligns with FAO MIR analysis recommendations, particularly for managing the effects of particle size and residual moisture in soil spectral data [25] .

Statistical analysis of the spectral data was conducted using Principal Component Analysis (PCA), a dimensionality reduction technique that simplifies the data and extracts the primary sources of variation between soil samples [26] . PCA aids in distinguishing chemical profiles by grouping samples according to their spectral characteristics, which is particularly useful for analyzing soils under different types of plant cover, such as acacia and eucalyptus plantations [10] [12] . This statistical approach enables the capture and interpretation of structural and chemical variations present in the MIR spectra of the soils studied [8] .

The MIR spectra of soils reveal distinctive spectral signatures for each plantation type studied ( Figure 1 ). Soils under acacia plantations exhibit specific peaks associated with organic nitrogen absorption bands in the region of 1510 - 1650 cm⁻¹ ( Figure 1(a) ). For soils under eucalyptus plantations, the spectra display characteristic absorption bands of lignin and phenolic compounds around 1320 - 1400 cm⁻¹ and 1600 - 1700 cm⁻¹ ( Figure 1(b) ). In contrast, the MIR spectra of soils under cassava plantations show bands in the region of 1000 - 1200 cm⁻¹ ( Figure 1(c) ), indicative of a high content of silicates and other labile mineral components. These spectral values reveal the distinctive effects of each plantation type on the soil’s chemical composition and quality, underscoring the relevance of MIR spectroscopy for rapid, non-destructive assessment of soil properties based on plant cover.

The results ( Figure 2 ) indicate that each plantation type influences the soil’s chemical properties uniquely, with varied impacts depending on the plant species. Soils under acacia exhibit spectral signatures rich in organic nitrogen, with peaks in the 1510 - 1650 cm⁻¹ region. Soils under eucalyptus display absorption bands in the 1320 - 1400 cm⁻¹ and 1600 - 1700 cm⁻¹ regions, characteristic of lignin and phenolic compounds. Lastly, soils under cassava show spectral signatures in the 1000 - 1200 cm⁻¹ region, associated with silicates and other labile minerals.

The factorial map of individuals ( Figure 3 ) presents the projection of soil samples from acacia, eucalyptus, and cassava plantations in the space defined by the main dimensions, Dim1 and Dim3, explaining 52.7% and 12.9% of the total variance, respectively. The ellipses indicate groupings for each plantation type, with soils under cassava distinctly positioned along Dim1 with predominantly negative values. This suggests a unique spectral signature and chemical composition, characterized by a high content of silicates and other labile minerals, indicative of an extractive profile. Soils under acacia and eucalyptus are closer to the center of the factor space, with some overlap and relatively similar characteristics compared to cassava, though their impacts on soil chemical composition remain distinct. Soils under acacia and eucalyptus are primarily located in the positive zone of Dim1, showing less variability. The differentiation of soil samples is further enhanced by Dim3, capturing an additional 12.9% of variance and allowing further isolation of vegetation groups, with centroids indicating proximity between acacia and eucalyptus, while cassava remains more distant.

The results of this study indicate that each plantation type specifically influences the chemical composition of soils in the Batéké plateaus. Soils under acacia plantations show a notable increase in organic nitrogen, with absorption bands in the 1510 - 1650 cm⁻¹ region, illustrating this legume’s ability to fix atmospheric nitrogen through root symbioses [9] [10] . This nitrogen enrichment enhances soil fertility, which is particularly beneficial in this nutrient-poor region. Acacia also contributes to an increase in labile organic matter, improving soil structure and water retention—two key factors for long-term productivity [6] .

In contrast, soils under eucalyptus show absorption bands associated with lignin and phenolic compounds in the 1320 - 1400 cm⁻¹ and 1600 - 1700 cm⁻¹ regions, indicating stabilized organic matter that is less biodegradable [8] [12] . Although this composition limits rapid nutrient availability, it promotes carbon storage, contributing to long-term carbon sequestration but necessitating compensatory practices to maintain soil fertility [3] . Finally, soils under cassava are characterized by a spectral signature in the 1000 - 1200 cm⁻¹ region, revealing a high content of silicates and other labile minerals. This is attributed to cassava’s deep root system, which mobilizes nutrients from lower soil layers [4] .

However, cassava’s high nutritional requirements can lead to rapid depletion of surface nutrients without sustainable management practices, such as crop rotation or the addition of amendments [1] [6] . These observations underscore the distinct influence of each plantation type on soil chemical properties, emphasizing the importance of selecting species based on sustainable soil management objectives—whether to enrich fertility, conserve nutrients, or store carbon.

Mid-infrared (MIR) spectroscopy has proven to be a powerful tool for tropical soil management due to its ability to provide precise, non-destructive information on soil chemical composition. In this study, MIR spectroscopy identified distinctive spectral signatures associated with acacia, eucalyptus, and cassava plantations, underscoring the specific impact of each vegetation cover on soil properties. By measuring specific absorption bands—particularly in the regions of 1510 - 1650 cm⁻¹ for organic nitrogen, 1320 - 1400 cm⁻¹ and 1600 - 1700 cm⁻¹ for lignified compounds, and 1000 - 1200 cm⁻¹ for silicates—this technique enables rapid characterization of the effects of various plantations on soil [8] [15] .

The use of MIR spectroscopy for tropical soil management offers several advantages. It facilitates rapid assessment of soil fertility, carbon storage potential, and chemical structure—essential parameters for adapting agricultural and forestry practices in vulnerable ecosystems. Additionally, it provides a method for long-term soil quality monitoring without requiring destructive sampling or complex laboratory analyses, which is particularly beneficial in tropical regions where logistical resources may be limited [2] . In previous studies, [24] demonstrated the effectiveness of near-infrared (NIR) spectroscopy for predicting site indices in sandy soils under eucalyptus plantations in the Republic of Congo, suggesting that combined MIR and NIR approaches could further enhance soil characterization in similar tropical environments.

In conclusion, MIR spectroscopy holds significant potential for optimizing tropical soil management by enabling monitoring and anticipation of the effects of planting practices on soil chemical properties. It is thus a valuable tool for sustainable management strategies, promoting productivity and preservation of tropical soils, and providing a robust database for informed decision-making in agriculture and forestry