Introduction

Soil must be protected to meet the demands of food security and sovereignty to guarantee life on earth, where diligent management is a fundamental element for sustainable agriculture (United Nations Food and Agriculture Organization Agriculture (FAO) 2015). Worldwide, around 5000 million hectares (Mha), or a third of all land surface, are used for agriculture (Burbano 2016). According to the FAO, around 1660 Mha have been affected by soil degradation processes related to erosion, decreases in nutrients and increases in salinity, with 927 Mha in humid zones, classified as severe degradation (46.6%) and slight degradation (53.4%) (FAO 2021).

Increases in agricultural activities, grazing and forestry put progressive pressure on the soil, which can lead to degradation and contamination (Olivera and Avellaneda 2018). It is becoming increasingly important to carry out studies that indicate the level of soil degradation associated with physical and chemical properties, where degradation is a set of processes that reduce the potential to produce goods and services (FAO 2015).

In Latin America and the Caribbean, around 800 million hectares are suitable for agriculture (Pennock and Mckenzie 2016). In 2015, around 281 Mha were affected by degradation in South America, with 153 Mha highly degraded and 128 Mha slightly degraded (FAO 2021). In Colombia, 22 million hectares are suitable for agriculture, of which 5 million are used (Ministerio de Ambiente y Desarrollo Sostenible (MADS) 2016). The MADS, the Instituto de Hidrología, Meteorología y Estudios Ambientales (IDEAM) and the Universidad de Ciencias Aplicadas y Ambientales (U.D.C.A) (2015) studied soil degradation and reported that around 60% of agricultural social in Colombia are degraded, especially by erosion.

In the Department of Valle del Cauca, 17.5% of the total hectares (2,067,766) are dedicated to agricultural and livestock, mainly sugar cane and fruit trees (Tafur et al. 2006). Sugar cane cultivation predominates in the flat area (more than 200,000 hectares) and the soils present some problems of compaction, drainage and the presence of sodium and magnesium salts. The soils in the hills had a large area (970,702 hectares) and presented erosive processes that make management practices necessary for conservation (Tafur Reyes et al. 2006). In 2018, Valle del Cauca reported a production of 21,190,409 tons for sugar cane, 258,338 tons for plantains, 14,5162 tons for pineapple, 142,145 tons for corn, 140,682 tons for citrus, and 499,748 for other crops (Agencia de Rural Development [ADR], 2021).

The variability of physical and chemical properties must be known for a soil analysis, including the degree to which they are affected by different agricultural uses (Salgado et al. 2019). Similarly, properties must be understood as a function of different processes that are influenced by formation factors, such as parent material, climate, relief, organisms, and time (Jaramillo 2012). Multivariate analyses are an adequate statistical technique to assess relationships between soil properties (Trujillo-González et al. 2022). On the other hand, principal component analyses (PCA) have made it possible to differentiate management systems and soils by reducing the number of original variables to new potential variables (Francirlei de Oliveira et al. 2019).

The analysis of agricultural soils is a very important tool that shows potential, fertility and degradation (Lizcano et al. 2017). This research aimed to assess the quality of agricultural soils at the regional level, especially on the plains and foothills of Valle del Cauca. The physicochemical properties in 489 agricultural soils were evaluated, establishing the level of degradation (emphasis on compaction), for use in future studies on sustainable management for agriculture.

Materials and methods

Study area and sampling

The study area was in a flat, piedmont area in the Valle Department, at 3°30' and 5°10' North and 75° 42' and 77° 33' West, with bimodal precipitation behavior in two rainy seasons (April–May and October-August) and two dry seasons (January–February and July–August). The average precipitation varied between 1000 mm at the bottom of the valley to 1800 mm in the piedmont area. The average temperature was 24 °C, and the average evaporation was 4.1 mm day−1, at altitudes ranging from 925 to 1200 m above sea level (Corporación Autonoma Regional del Valle del Cauca (CVC) and Instituto de Estudios Ambientales de la Universidad Nacional sede Manizales (IDEA) 2008). The average climate included very humid, humid and dry conditions (Martínez 2004).

489 sites were selected to characterize the physicochemical properties of agricultural soils on the flat and piedmont of the Department of Valle del Cauca. Priority was given to 13 hydrographic basins (Arroyohondo, Guabias, Guachal, Guadalajara, Jamundí, Meléndez, Pescador, Roldanillo-Unión-Toro (RUT), San Pedro, Sonso, Tuluá, Yumbo and the industrial zone) of agricultural interest to the regional environmental entity (Fig. 1). Simple random sampling (SRS) was carried out according to Otzen and Manterola (2017), where any site in the study area has the same probability of being sampled (Table 1). Of all the samples, 271 corresponded to Mollisol, 128 to vertisol, and the rest to Inceptisol. Soils cultivated with sugarcane represented 44% of the total number of sites, which is the main agricultural activity in the region.

Fig. 1
figure 1

Study area

Full size image
Table 1 Number of samples by hydrographic basins and soil taxonomy
Full size table

Evaluated properties and techniques for their determination

The evaluated physical properties of the soil were: percentage of sand (A), silt (L) and clay (Ar), quantified with the Bouyoucos method, and texture, defined with the USDA textural table. Ks (cm h−1) was determined with a constant head permeameter. Da (g cm−3) was determined with the ring method using a known volume and moisture retention. The usable water table (LAA in mm) was calculated with the difference between the field capacity (CC) and the permanent wilting point (PMP), along with the Da (see Eq. (1)):

$$\mathrm{LAA}=\frac{\mathrm{\%}(\mathrm{CC}-\mathrm{PMP})\times \mathrm{Da}\times \mathrm{Prof}\times 10}{100},$$
(1)

where CC is the moisture retention at 1 bar (%); PMP is the moisture retention at 15 bar (%); Da is the apparent density (g cm−3); Prof is the root depth; 10 conversion factor.

The porosity (\(\eta\)) was quantified with the Da and real density (see Eq. (2)):

$$\eta (\mathrm{\%})=\left(1-\frac{\mathrm{Da}}{\mathrm{Dr}}\right)\times 100,$$
(2)

where Da is the apparent density (g cm−3); Dr is the real density (2.65 g cm−3).

The evaluated chemical properties included: pH, determined with the extraction method with a 1:5 ratio of soil to water; CE, determined with potentiometric quantification; OM, determined with the Walkley & Black technique; and CEC, determined with ammonium acetate 1 N pH 7.0. Soil properties were measured at the Laboratorio de la Corporación Colombiana de Investigación Agropecuaria, AGROSAVIA, following current standards (Organismo Nacional de Acreditación (ONAC) 2006). The interpretation of the degradation was made based on the proposals of the FAO (2020) and other authors.

Statistical analysis of the results

Univariate statistical analysis was performed, which included the determination of position and dispersion. The variability of the data was based on the coefficient of variation (CV), where a CV < 10% was weak variability, CV between 10 and 100% was moderate variability, and CV > 100% was high variability (Li et al. 2013). The distribution was made using the “Kolmogorov-Smirnov” Test (non-parametric test), as proposed by Romero and Saldaña (2016). The comparison between the observed correlation coefficients and the partial correlation coefficients was done with the Kaiser–Meyer–Olkin (KMO) index, which quantifies the adjustment of a sample with values between 0.5 and 1 and a factorial analysis. Bartlett’s sphericity test looks at the null hypothesis and shows variables that are not correlated in a population; if the value is less than 0.05, said hypothesis is rejected, and the analysis continues (Montoya 2007). For the relationship between properties whose data did not follow a normal distribution, the “Spearman” correlation coefficient was used. The properties with normal data distribution were analyzed with the “Pearson” correlation coefficient (Martínez et al. 2009). Finally, the principal components analysis (PCA) was performed for the correlated variables. The statistical programs SAS University Studio (Cody 2015) and RStudio version 4.1.3 were used with the readxl, corrplot, psych and factoextra libraries (Venables and Smith 2022). For the spatial representation of the sampled sites, ArcGIS-Pro was used (Environmental Systems Research Institute (ESRI) 2021).

Results

The results of the Kolmogorov–Smirnov Test indicated that Da and \(\eta\) had a normal distribution (P > 0.05). The mean of the data distribution, plus or minus the standard deviation, represented 68.8%, the mean plus or minus twice the standard deviation represented 95.5%, and the mean plus or minus the standard deviation indicated 99.7% of the distribution (Martínez and Marí 2010), showing homogeneity for these variables. The other properties did not have a normal distribution, with high dispersion with respect to the averages: % clay, % silt, % sand, Ks, hydraulic retention at 1 bar (RH1B), hydraulic retention at 15 bars (RH15B), usable water (LAA), pH, OM, CEC and EC.

The analysis of variability expressed by the coefficient of variation, as interpreted by Li et al. (2013), showed that Ks and CE were highly variable. The other variables had moderate variability, which confirmed heterogeneity of the properties on the plains and piedmont area of Valle del Cauca and degradation processes in agricultural soils. The results of the univariate statistical analysis are in Table 2.

Table 2 Univariate statistical analysis for the study area
Full size table

The distribution of textures is in Fig. 2, with a higher percentage of loamy textures associated with the presence of clay, a condition that results in plasticity, difficulty for tillage, and low permeability; however, they conserve water and nutrients thanks to microporosity and CEC (Gisbert et al. 2010). They also favor root development. 31.5% of the soils were clay-loam (Far), and 26% were clay (Ar), as reported for textural classification by the IGAC and CVC (2004) in Valle del Cauca. According to the FAO (2020), when high percentages of clay are present, soils tend to compact and to resist changes in pH, and OM levels are maintained. Da also has a relationship with root development as a function of texture. Da values less than 1.6 g cm−3 are suitable for root development; when greater than 1.8 g cm−3, there is restriction for root growth; for silt, values of Da < 1.4 g cm−3 are ideal, and Da > 1.65 g cm−3 results in a restriction. In the case of clay, Da < 1.1 g cm−3 is ideal, and there is restriction with values greater than 1.47 g cm−3 (United States Department of Agriculture (USDA) 2008).

Fig. 2
figure 2

Distribution of predominant textures in the study area. Note: AF sandy loam, Ar clay, ArA clay–sandy, ArL clay–silty, F loam, FA sandy–loam, Far clay–loam, FArA sandy–clay–loam, FArL silty–clay–loam, FL silty–loam

Full size image

The results of the Da analysis confirmed the influence of texture variability in agricultural soils in the study area. Da values > 1.6 g cm−3 were recorded, which could present soil compaction problems (Ramírez 1997), affecting soil moisture retention and root growth (Salamanca and Sadeghian 2009). 19% of the sites presented Da values above 1.6 g cm−3 (93 sites), mainly in soils from the flat zone of the Guachal basin that had sugar cane. Da values between 1 and 1.6 g cm−3 (79.4% of the soils) mean that soil processes and plant development were not affected (FAO 2020). Only 1.6% of the soils presented Da values less than 1 g cm−3, indicative of organic soils with OM contents between 1.58 and 9.40%.

\(\eta\) is an indicator variable for the volume of water reserves in the soil (González-Barrios et al. 2012), its importance is conditioned by the movement of water and air within the soil profile, mainly influenced by the presence of macropores (Graham et al. 2008). The soils from the plains and piedmont in Valle del Cauca had an average \(\eta\) of 45.89%, with high variability, a low value with a tendency to compact (Jaramillo 2002). Of the total soils analyzed (489), 21.1% (103) had \(\eta\) values less than 40% (very low), of which 72 sites (70%) were in the Guachal basin, indicators of degraded soils with possible anthropic compaction (Jaramillo 2002). On the other hand, the data obtained with field capacity moisture, (CC) − 1 bar and PMP − 15 bars had a difference of 10.08% p–v–1 in the maximum and minimum range, which is an indicator of greater water storage that can be used by crops after being saturated (Enciso et al. 2007).

Ks was highly variable in the soils, and the range of determined values went from very slow to very fast (Jaramillo 2002). The analysis of the mean and maximum values of Ks (2.77 and 36.72 cm h−1), indicated possible alterations and/or contamination of groundwater, either by agricultural inputs or other sources, substances that can reach the aquifer. 27.6% (135 sites) of the soils had high Ks values (> 3.6 cm h−1), mainly in Mollisol soils (71.11%), with 14.07% in the Inceptisol soils and 14.8% in the Vertisol soils. Of these sites, 114 were cultivated with sugar cane.

According to Quintero (1993), the flat areas of Valle del Cauca have a pH in the range between 5.5 and 7.0, which was confirmed here, where 42.12% (206 sites) were in this range, and only 3.5% (17) of the sampled sites had a very low pH (< 5.5), the rest (276) had pH values greater than 7; pH, as a soil chemical variable, is essential in processes related to the availability of nutritional elements for plants (López and Estrada 2015). The average and mode were very close to neutrality, ideal conditions for the availability of nutrients in the area, where the data had a certain homogeneity (standard deviation close to 1). Extremely acidic (4.15 units) or very alkaline (8.51 units) pH values indicate soils altered by anthropic activities (Osorio 2012). According of the Colombian Agricultural Institute (ICA 1992), some soils would be in the strongly acid category (pH < 5.5), which constitutes a risk of toxicity from aluminum and manganese. On the other hand, soils with pH > 8.0 indicate a potential deficiency of P, Ca, Mg and Mo, however, there was possibly excess exchangeable sodium with a consequent inhibition of crop growth (ICA 1992).

OM contents contribute significantly to the physical, chemical or biological properties of soil. The results indicates that the mean (2.37%) and the mode (1.20%) presented significant differences. According to the recommendations of the Colombian Agricultural Institute (ICA 1992), the OM was at a medium level, and the most frequent datum was low for the prevailing climatic conditions in the study area (warm). 48.3% (236) of the soils had a %OM < 2%, of which 26.7% (63) were in the Guachal basin. 40.1% of the soils had OM contents between 2 and 4% (196), of which 40.0% (78) were in the Guachal basin, mainly with sugar cane crops (99%). 11.7% (57) of the soils presented levels higher than 4% OM. Soils with OM contents greater than 2% in the study area have good conditions in the edaphic medium with respect to aggregate stability, porosity, aeration, infiltration, moisture storage, consistency, and nutrient supply (Sanclemente et al. 2021).

The CEC presented a mean higher than 24 cmol(+) kg−1 and was high (> 20 cmol(+) kg−1) (ICA 1992) although an appropriate value for fertility in agricultural soils is between 15 and 25 cmol(+) kg−1 (López and Estrada 2015). Only 14.3% (70) of the analyzed soils presented a CEC lower than 15 cmol(+) kg−1, that is, with a tendency towards low fertility. 48.46% (237) of the sampled sites were between 15 and 25 cmol(+) kg−1, and 37% (181) had values higher than 25 cmol(+) kg−1. The values showed optimal conditions for cation exchange, which is a good indicator of fertility. Pérez et al. (2017) stated that the cation exchange is used by plants through electrostatic forces for their development.

The average EC in soils from the plains and piedmont of Valle del Cauca was 0.49 dS m−1, i.e., non-saline soils (ICA 1992). Maximum EC values were in the analysis, classifying them as very saline soils (Fernández et al. 2006); however, they only represented 2.2% (11 sites) of the soils, with EC values > 2 dS m−1. Some agricultural soils with crops were subject to stress caused by salinity in the Guachal, Guadalajara, Tuluá, RUT and San Pedro basins. The more critical cases were in the Guachal basin (10.04 dS m−1), Tuluá basin (9.21 dS m−1) and San Pedro basin (8.26 dS m−1).

According to the KMO index, the soil properties with a communality value of less than 0.5 included %sand (0.46), %silt (0.20), %clay (0.41%), CC (0.49) and LAA (0.33). Therefore, they were not considered in the factor analysis. While the other variables had a value greater than 0.5 KMO, such as pH (0.72), OM (0.77), EC (0.85), CEC (0.86), \(\eta\) (0.68), PMP (0.54), Ks (0.96), and Da (0.68). The Bartlett sphericity test had a significance of less than 0.05 (< 2.2 × 10–16), expressing a relationship between the variables and making the principal components analysis feasible.

In this analysis, the first three components, which had an eigenvalue greater than 1, were representative and explained 71% of the variability of the data (Table 3). The first component contributed 34% to the total explained variance, while the coefficients of the first vector indicated that the variables \(\eta\) and PMP contributed more in a positive sense, while Da did so in a negative sense, showing some risk of degradation. The second component contributed 24% to the total variance, where pH had the greatest contribution to the component in a positive sense, while %OM did so in a negative sense. Similarly, the third component explained 13% of the total variance, where the variables with a greater contribution were CEC and Ks (positive).

Table 3 Principal components analysis for the eight correlated variables
Full size table

In general, apparent density and porosity in the first component explain to soil compaction process, with displacement of water and air from the porous space from the plastic distortion of the soil (González et al. 2008). The second component maintained a fundamental relationship with soil fertility, especially for the availability of nutrients, highlighting the buffering or pH-regulating effect of excessive changes in OM (Garro 2016; Sanclemente et al. 2021). A significant reduction in pH increases the solubility of metallic elements, with possible changes in the microbial population (Torrente 2017). On the other hand, the most significant effect from pH in soils is related to the reserve of nutrients for plants, and its relationship with OM is evidenced in the availability of nitrogen. For pH values between 6 and 8, there are ideal conditions for the mineralization of nitrogen in the OM and symbiotic fixation by organisms, resulting in the maximum availability of the element (Foth 1990). Since organic matter has different forms, it contributes effectively to the physical properties of the soil (Novillo et al. 2018). There was a relationship between Ks and CEC in component 3, where expansion or dispersion processes may explain a decrease in Ks by reducing the concentration of electrolytes in the soil solution (Aristizabal 2009).

Figure 3 shows a greater contribution from bulk density and porosity via color intensity, followed by PMP and % OM, a third group with pH and Ks, and a final group with CEC and EC.

Fig. 3
figure 3

Contribution of correlated soil properties to the first two components

Full size image

The correlation analysis between soil properties (Fig. 4) ranges from weak to strong (Martínez et al. 2009). This diversity of correlations may be associated with soil properties that were grouped into physical properties and chemical properties. The Da and \(\eta\) had a perfect correlation (− 1.00), as reported by Ramírez (1997), who indicated that an increase in Da reduces the total \(\eta\), affecting the penetrability of the roots, availability of oxygen and water, among other characteristics. Meanwhile, pH presented an inverse relationship with the percentage of OM and the EC (− 0.41). There was a coherent link between pH and OM with a reduction in OM that acidified the soil, possibly associated with mineralization of OM as a result of decreased microbial activity (Cruz-Macias et al. 2020). Ks had an inverse correlation, between moderate and strong, with PMP (− 0.75) because Ks can decrease significantly under water limitations (Jaramillo 2002).

Fig. 4
figure 4

Spearman’s correlation for soil properties

Full size image

Discussion

19% of the analyzed soils presented physical degradation processes associated with compaction phenomena, with bulk density (Da) values greater than 1.6 g cm−3, and porosity (η) less than 40%, possibly because of different practices in different crops, mainly in the flat zone of the Guachal basin and in sugarcane crops. Soil compaction can cause shallow rooting, which affects crop yields, pore sealing, which reduces water infiltration, and inferior air or water movement through the soil profile (USDA 2008). The flat area of Valle del Cauca could generate flooded areas, and the foothills may have some erosive processes. However, the development of activities that increase the organic matter content would prevent soil compaction (FAO 2020). In addition, tillage could reduce bulk density but it directly affects soil functions (Corsi and Muminjanov 2019).

Therefore, to avoid compaction in soils where agricultural practices are carried out, tillage management is necessary, avoiding the use of tillage implements in soil with a high moisture content and at a constant depth over a long period of time (Jaramillo 2002). At the same time, apparent density (Da) is a soil property that is related to root development as a function of texture. Sandy, silty and clayey soils with a Da greater than 1.8, 1.65 and 1.47 g cm−3, respectively, are affected in terms of root growth and have a reduce capacity to retain water for crop growth (USDA 2008). Clay textures are ideal for root development in crops and predominate in the flat area in Valle del Cauca. Narro (2004) stated that low Da values are recommended for agricultural activities, where an increase indicates a reduction in the pore space, a compression-compaction increase, resistance to penetration and thermal conductivity. However, a high bulk density is not always synonymous with compaction; an effective soil function associated with water and nutrient retention, infiltration rate, percolation, carrying capacity and drainage of the water are more important and are mainly influenced by porosity, structure and stability of soil aggregates (Corsi and Muminjanov 2019).

Soil Ks is associated with the movement of solutes within the profile (Polo et al. 2003). This soil property is highly associated with the soil structure, the mineralogical composition of the clays and the texture. In addition, it has a close relationship with drainage (Jaramillo 2002). Here, 27.6% of the analyzed soils had Ks values higher than 3.6 cm h−1, which indicated possible risks of degradation and consequent leaching of substances that can alter groundwater quality. These Ks values were highly conditioned by the soil moisture because of the precipitation, irrigation, or drainage from the sampling sites and were influenced by other soil properties, such as texture, structure, or composition of clay minerals. The maximum Ks is associated with the saturation point of soil moisture since each pore filled with water allows unrestricted conduction, without ignoring the size, organization, and type of pores in the soil (Jaramillo 2002). Areas with high Ks values favor the circulation of water in a lateral direction or at depth in the soil profile, with transport of solutes towards the lower horizons (Castiglion et al. 2018). This scenario facilitates the quantification of soil quality in relation to the displacement of water (De falco and Penon 2010).

Conclusions

The study area had degradation processes in 19% of the evaluated sites (93) that were associated with soil compaction, with high bulk densities (1.6 g cm−3) and porosity less than 40%. These soil characteristics condition the flow of air and water in the soil environment and can affect the yield and development of different agricultural crops.

Although the pH did not present extreme values (range 4.15–8.51), some sites had critical nutrient availability, crop growth or sodium salt levels. This makes it difficult to establish crops because of the variation in some physical properties, making remediation work necessary, such as liming or soil washing.

The salinity analyses did not show critical situations in the studied soils, and the agricultural practices are not impacting the soil quality in terms of the presence of salts; however, studies should be carried out in specific areas that will facilitate based on the percentage of exchangeable sodium.

Monitoring and following up on the quality of soils that have altered physical and chemical properties are recommended to reduce, control or mitigate degradation processes in the soil environment, along with implementing agricultural practices that contribute to sustainable soil management.