Overview of the properties of 85 soil samples in Dak Lak province
Descriptive statistics of soil characteristics are presented in Table 2. The relative difference between mean and median values is smaller than 10% for silt, clay, pHKCl, OM, Ntot, Ptot, Ktot, CEC, it is in the 10 - 15% range for sand content and bulk density; this relative difference is the highest for sand content and CEC. The average clay content of the 85-soil collection is less than 27% and the average organic matter content is higher than 3%. Consequently, the cation exchange capacity of these soils is rather high (average value > 10 cmolc kg
-1). Some characteristics of soils can be considered as intrinsic properties whereas others are direct or indirect consequences of these ones. So, texture, organic matter content (as inferred from oxidizable C) and pH govern or have deep influence on many other characteristics such as CEC, extractable acidity, organic N and P (which are usually close to total N and P) (Table 2). Based on this broad analysis of the data, it seems that some samples cannot be classified as clay soils even if they were taken in the province’s mountainous region. In fact, we can see that the majority of samples are clustering in the classes of loamy sand, loamy clay and clay soils.
Yegna et al., (2024) also found that cultivated lands in Ethiopia contained a considerable amount of clay and very little sand. Most of soil samples in this study are acidic (pH 4.26), high OM (3.58%), average N (0.13%) and CEC (10.50 cmolc kg
-1), poor total P (0.02%) and K (0.30%) (
MONDRE, 2024). A significant amount of arable land is currently covered by acidic soils, which hinder crop growth by lowering the availability of vital nutrients like P and containing high levels of harmful elements like manganese (Mn) and aluminium (Al) (
Getaneh and Kidanemariam, 2021). The result of this study agrees with the findings of
Nguyen et al. (2025);
Nguyen et al. (2024);
Molla et al., (2022) who reported that soil characteristics were influenced by soil forming and practice management (bare land or deforestation).
Correlation among soil characteristics
The highest correlation was found for sand and silt contents (r = -0.90) (Table 3); it means that the variation of sand content is closely linked to an opposite variation of silt content; the increase of sand content is also accompanied by a decrease in clay content (r = -0,82), but with less systematic variation. The two sand and silt content are also well correlated with CEC (r = -0.59 and 0.57), the high sand content with low CEC values but the high silt content with high CEC values, which is not to surprise in a collection of soils with limited range of textural variation and similar pedo-climatic context. There were correlations between soil texture parameters and physicochemical properties. CEC had significant correlation with most of soil properties such as pH, BD, total N, P. Similar result was found by
Mishra et al., (2022) noting that CEC in all land uses with highest correlation in desert land use (
r = 0.94;
p<0.05). Conversely, CEC was significantly and negatively correlated with sand in all land uses, with highest negative correlation obtained in desert land use (
r =-0.84;
p<0.05). OM exhibited no significant correlation with other characteristics, this is inverse with other studies from
Ye et al., (2024); Pham and Vo (2023) who indicated that soil microbes play a vital role in the mineralization of C and N nutrients and all of them were aggregated with clay, which suggests that among the different textured soils, clay was the most relevant to soil C and N nutrient status. OM and other soil characteristics show low correlation coefficient (r <0.15); the mean C/N value of the 85 soils is around 31, which is high for not well humified organic matter; it means that the studied soils still fix nutrients by microorganisms, which can be related to low biological activity in acid sandy soils; Besides, soil acidity is caused by a variety of factors, including rainfall, leaching, the presence of acidic parent material, the breakdown of organic matter and intensive farming
(Dogiso et al., 2025).
Structure of soil properties variation through principal component analysis (PCA)
The multivariate statistical method of PCA is a very useful tool for reducing the number of variables in a data set and for obtaining useful two-dimensional views of a multiple dimensional data set (
Abdel-Fattah et al., 2021). Our set of data contains 10 measured variables which will be reduced to a smaller number of more synthetic variables called principal component (PC); the aim is to explain the maximum amount of variance with the fewest number of independent components. Accordingly, the first principal component (PC1) in our data set explains 41% of the variation, while the subsequent components (PC with eigen values > 1) (Table 4) explained 15% and 12% of the variance, respectively. This indicates that PC1 and PC2 account for 55% of the variance, while PC1-PC2-PC3 accounts for 67%. The circles of correlation between the initial variables and the computed principal components PC1, PC2 and PC3 are displayed in Fig 1. The two parameters that have the highest correlation coefficients with PC1 are sand content (0.91) and silt content (-0.85); it should come as no surprise that other characteristics including pHKCl, N, P, K, CEC correlate with PC1 (0.21<r| <0.72) (Table 4).
Notably, the bulk density and sand vectors are nearly perpendicular to the PC1 axis, indicating that bulk density and sand are almost entirely independent of the first principal component and related to the original variables. The highest correlations for the second axis, PC2, are obtained for OM (r = 0.66) and N (r = 0.55). Simple variables and PC3 do not show any particularly strong link; the best values are r = 0.56 for K and r = 0.49 for bulk density. Reducing our data to three independent and synthetic variables allows us to summarise its structure as follows: PC2 is the axis of OM, PC3 is the axis of K (very roughly) and PC1 distinguishes the soils based on texture and P concentration (and related variables). Because of its synthetic physico-chemical meaning and the fact that, as was previously said, the association between CEC and clay was less than anticipated, it was also taken into consideration. Similar results were found by
Abdel-Fattah et al. (2021) who pointed out that the first PC has stronger positive relationships with EC, OM, CEC, available NPK, ESP and clay, while the second PC has a substantial association with silt and
Li et al., (2020) reported that soil silt was significantly negatively correlated with total K content but significantly positively correlated with available K. Furthermore, the CEC of tillage layer soil showed a positive correlation with the amount of silt and clay in the soil. This was because higher clay content increased the specific surface area of soil colloids and soil charge, which enhanced the soil’s ability to adsorb ions, which was consistent with findings from earlier research
(Pelster et al., 2018; Chen et al., 2021).
The distribution of the observations (85 soils) in the new space created by the synthetic variables can also be seen using PCA. The 85 soil samples are shown in Fig 2 based on their coordinates with PC1 and PC2. With a tailing tendency of a dozen points, we observe extremely good scattering of the majority of points, indicating those soils with a higher OM content and/or less sandy texture. However, we are unable to detect actual subpopulations, which is a requirement for a useful PCA. This presentation style will be used to illustrate potential connections between soil attributes and sample location, soil preparation and cropping pattern. Points will be assigned various symbols based on these criteria.
Rajput et al., (2023) indicated that the PCA method was shown to be somewhat superior for assessing soil quality in the North hill region in India. The PCA of soil physicochemical properties, soil microbial status and soil texture factors of various texture paddy soils, taking into account the overall pattern of soil physicochemical properties, also demonstrated that variations in nutrient content and physicochemical properties among soils can be reflected in soil texture
(Ye et al., 2024). Principal component analysis (PCA) was employed to establish a minimum data set (MDS) and to evaluate key physical and chemical properties affecting soil quality, along with the associated weight factor for each indicator
(Ibrahim et al., 2025).
Implications for sustainable agriculture in Dak Lak province
The results of this study provided a general picture on the physical and chemical characteristics of soil in Dak Lak province that influence the degree of soil quality in the areas under investigation. This implies that large areas of the study are degrading and have low soil quality. Assessing soil quality using key indicators through PCA analysis is essential for resilient agricultural systems, improving food security and encouraging sustainable resource management. This supports both ecological sustainability and economic growth. Soil characteristics improvement for sustainable agriculture in Dak Lak province including to increase application of more organic matter, control soil erosion by cover crops, use suitable cropping patterns, reduce tillage, enhance water and nutrient retention.
