The current study was carried out in farmers’ fields of Pauri district and Rudraprayag district 30.28°N 78.98°E and 30.15°N to 78.78°E with elevations ranging from 800m to more than 1800m. The experiment site is situated in the mid-hills subtropical and sub-temperate zone of Uttarakhand, India (Fig. 1).In Rudraprayag, the mean annual rainfall is 637mm, and the average temperature rises to 29℃ in summers and falls up to -3℃ in the winters. In Pauri district, the average rainfall is 172.88mm, and temperature ranges from 26℃ to 14℃.

Three elevation ranges were selected: lower (800- 1300 m), mid (1300-1800 m), and upper (above 1800 m). At each elevation, there are three agroforestry systems with different models: Agrihorticulture (Bhimal-based, Ficus+Bhimal, and Oak-based), Agrisilviculture (Malta + Bhimal + Oak + Gmelina, Walnut + Malta + Bhimal + Napati, and Oak + Malta + Peach), Agrisilvihorticulture (Malta based, Malta- based and Apple + Malta) with two depth levels (0-15 cm) and (15-30 cm) were categorized (Table 1).

The agroforestry fields were selected at three altitudes viz. low (800-1300 m), mid (1300-1800 m), and upper elevation (>1800 m) (Table 1). At each altitude, based on the level of the system, agroforestry was categorized as Agri-silviculture (Tree on bunds), Agri- horticulture (Scattered trees of fruit crops), and Agri- silvi-horticulture (Homegarden), (Table 2). Selected models are Bhimal-based, Ficus+Bhimal, Oak-based, Malta+Bhimal+Oak+Gmelina, Walnut+Malta+ Bhimal+Napati, Oak+Malta+Peach, Malta based, Malta-based and Apple+Malta (Table 2).

Table 1. Selection of study sites

AS= Agri-silviculture system, AH=Agri-horticulture system and ASH=Agri-silvi-horticulture system

Table 2. Description of selected agroforestry systems and their components

Low=Lower elevation (800-1300m), Mid= Middle elevation (1300-1800m) and Upp=Upper elevation (>1800m).

The soil samples were collected during April–June 2024. Soil sampling was done randomly collected from each system at two depths, 0–15 cm and 15–30 cm, using a soil corer. Fifty-four samples (3 altitude × 3 Agroforestry system × 2 depth × 3 replicates (Salve et al., 2018) were collected from the site. Then, the samples was air-dried at room temperature, and any visible organic material, stones, plant roots, and other debris were carefully removed. The Physical properties of soil include Moisture content, Water holding capacity, Soil texture, Texture class, Soil bulk density, chemical properties pH, Soil organic carbon, Soil carbon stock, and Soil organic matter were studied and followed all standard procedures to explore the impact of trees on soil properties under selected agroforestry systems (Table 3).

Table 3. Parameters used to evaluate the physio-chemical properties of soil

The soil analysis data was collected and organized for statistical analysis using Microsoft Excel 2007. The relationship between soil physio-chemical properties and altitude, agroforestry system, and soil depth was determined through Pearson's correlation matrix using the OPSTAT statistical analysis tool (www.hau. ernet.in). The data in the tables are presented as the mean of three replicates ± standard error (SE). Principal Component Analysis (PCA) was further applied to identify which soil properties influence the outcome using OriginLab software (Version 10.2).

At lower elevations (800-1300m), maximum sand percentage was found in the Malta based model, where minimum was observed in the Bhimal based model, while clay percentage was observed more in the Bhimal model and less in the Malta model. The silt was present more in Malta + Bhimal + Oak + Gmelina at 15-30cm depth, and at 0-15cm Malta-based have a high amount of silt. Silt percentage increased with depth, clay percentage decreased with increasing depth, and sand percentage increased with depth in Bhimal-based and Malta + Bhimal + Oak + Gmelina, while it had shownshowed a decreasing pattern in Malta-based agrofrestry. There was no change in soil texture class with a depth of Bhimal-based (clay) and Malta-based (sandy clay loam) whereas Malta + Bhimal + Oak + Gmelina agroforestry model showed changes with depth at 0-15 cm clay and 15-30cm it was loam.

At midddle elevations (1300-1800m), Ficus + Bhimal had maximum sand and minimum clay percentage while Walnut+Malta+Bhimal+Naspati hadmaximum clay and minimum sand percentage. Silt percentage was high in Malta-based at 15-30cm and low in Walnut + Malta + Bhimal + Naspati at 0-15cm. No variation was found in soil texture class and soil percentage with depth in all systems. The soil texture observed in Walnut + Malta + Bhimal + Naspati was clay, and the sandy clay was in Ficus+Bhimal and Malta-based agroforestry models.

Above 1800 m, the maximum percentage of sand was found in Oak+Malta+Peach and the minimum in the Oak+Malta agroforestry model. Silt had a low rate in Oak+Malta+Peach and high in Apple+Malta at 0-15 cm and 15-30 cm Oak-based agrofroestry. The sand and silt percentage increased with depth in the Oak+Malta+Peach and Oak model but decreased in the Malta based model, whereas the clay percentage decreased in the Oak+Malta+Peach and Oak model but increased in the Apple+Malta model. The soil texture class remained consistent across all systems. Over time, the physio-chemical properties of soil are influenced by topography, climate, and vegetation. The result confirmend that and content increased in the ASH system at deeper depths, contributing to better drainage and supporting root development. The sand, silt, and clay proportions change across the altitudinal gradient (Table 4). Sand content exhibited a strong positive relationship with altitude, whereas silt and clay content negatively correlated with altitude. Similar findings were reported by Kamal et al. (2023), who noted that sand content is lower at the bottom and mid elevations due to reduced soil erosion.

Table 4: Soil texture and soil texture class as influenced by soil depth, agroforestry system and altitude.

AS=Agri-silviculture system, AH= Agri-horticulture system and ASH= Agri-silvi-horticulture system

At 800-1300m, the Malta+Bhimal+Oak+Gmelina- based agroforestry system showed the highest pH, while the Bhimal-based system showed the minimum. Malta-based had maximum moisture content and minimum Bhimal-based. Malta+Bhimal+Oak+ Gmelina model showed the highest WHC% at 0-15 cm (49.86%) and the lowest at 15-30cm (27.33%). Malta- based at 15-30 cm (38%) showed maximum and minimum at 0-15 cm (36.34%). Malta-based exhibited the highest bulk density and the lowest in the Malta+Bhimal+Oak+Gmelina. pH, moisture content, and Bulk density increase with depth while Water holding capacity decreased with depth.

At 1300-1800m, Malta-based showed the highest pH and minimum in the Ficus+Bhimal agroforestry system. The Malta-based system had maximum moisture content and minimum in Ficus+Bhimal. Maximum WHC% showed in Ficus+Bhimal at 0- 15cm (42.44%) and minimum in Malta-based at 15-30 cm (26.56%). Bulk density was high in Malta-based at 15-30 cm (1.83 g/cm3) and low in Walnut+Malta+ Bhimal+Naspati at 0-15cm (1.45 g/cm3). pH and bulk density increase with depth while Water-holding capacity and moisture content decrease.

Above 1800m, Oak+Malta+Peach had high pH and is low in Oak-based. Moisture content was maximum in Oak+Malta+Peach and minimum in Apple + Malta. Water holding capacity was high in Oak+Malta+ Peach at 0-15 cm (54.98%) and minimum at 15-30 cm ( 34 . 49 %). Bulk density was maximum in Apple+Malta at 0-15cm (1.98 g/cm3) and minimum at 15-30 cm (1.44 g/cm3. pH showed an increasing pattern with depth and moisture content, and Water holding capacity showed a decreasing pattern except for Apple+Malta model. In contrast, Bulk density showed an opposite pattern compared to Water holding capacity and moisture content (Table 5). Soil moisture content was high in higher elevations and the Agri-silvi-horticulture system. Soil moisture content was influenced by the availability of irrigation. The surface layer (0–15 cm) had higher moisture levels than the sub-surface layer (15–30 cm), likely due to a greater accumulation of crop residue and organic matter in the topsoil. These results support with the results of Dahiya et al. (2022) who explained the effect of saturated hydraulic conductivity under different agroforestry systems, soil depth, and their interaction was significant.

The water holding capacity (WHC) was greater in Agri- silvi-horticulture system, likely due to the thick litter layer on the ground and the more significant accumulation of organic matter due to higher species richness and density. WHC lowered with increasing soil depth, with the surface (0-15 cm) layer showing the highest values across all home gardens. The percentage of clay was maximum in all Agri-silvi-horticulture systems. (Gomez et al., 2002) study, the water-holding capacity of clay and the high permeability of sand may explain the specific relationship between clay, sand, and moisture content in soil.

Bulk density decreased with increasing elevation. The lower bulk density at higher altitudes may be attributed to the combined use of organic and inorganic fertilizers, which help improve soil bulk density. Bulk density (g/cm3) was observed highest in the subsurface layer, followed by the top surface layer. These findings are supported by Singh et al. (2011). Bulk density was maximum in the Agri-horticulture system in lower elevations. The plausible reason for higher bulk density in the Agri-horticulture system may be more tillage operations for cultivating crops and low litterfall input. Meanwhile, Kundan (2024) recorded the highest bulk density in the Agrisilviculture system due to minimum litterfall accumulation.

Among three different systems, the Agri + silvi + horticulture system recorded the highest and lowest in the Agri-horticulture system. The slightly acidic pH in the Agri-horticulture system can be attributed to the high organic content, its accumulation, and the gradual decomposition process, which releases organic acids that lower the pH. Similar findings were observed by Mehraj et al. (2022), who reported lower pH under high canopy cover because of the release of organic substances during litter deterioration. Also, soil pH showed a higher value at the low layer layer than the top layer in all agroforestry systems because tree root abundance is high at the surface layer (Maqbool et al., 2017).

Table 5. Soil pH, Moisture content (MC), Water holding capacity (WHC) and Bulk density (BD)

AS- Agri-silviculture system, AH- Agri-horticulture system and ASH- Agri-silvi-horticulture system

Soil organic matter and organic carbon were maximum in Malta + Bhimal + Oak + Gmelina and minimum in the Bhimal based. The Malta-based model had high soil carbon stock at 0-15 cm (76.61) and at 15-30 cm Bhimal based (42.75), where soil carbon stock was high in Bhimal based at 15-30 cm (36.60) and (35.88) at 0-15 cm in Malta + Bhimal + Oak + Gmelina. Soil organic matter, organic carbon, and carbon stock decrease with the rise in soil dept at 1300-1800m, the Ficus+bhimal model showed minimum soil organic matter, soil organic carbon, and soil carbon stock. Malta-based model at 15-30 cm (2.75%) and Walnut + Malta + Bhimal + Naspati at 0-15 cm (2.17%) have maximum soil organic matter. Malta-based at 15-30 cm (2.11%) and Malta + Bhimal + Naspati at 0-15 cm (1.74%) showed high Soil organic carbon. Charan et al. (2012) found a positive correlation between SOC and altitude. They observed a decline in SOC content with increasing soil depth, with the highest SOC recorded at the 0-15 cm depth across all agroforestry systems. Some researchers have noted a similar trend of SOC decreasing with increasing soil depth (Singh et al., 2021).

The highest SCS was recorded in Malta-based at 15-30 cm (60 t ha-1) and Malta+Bhimal+Naspati at 0-15 cm (36.65 t ha-1). SOM, SOC, and SCS increased with depth in all systems except the Bhimal+Ficus agroforestry model. The increase of SCS stocks in topsoil in Agroforestry systems is due to higher broadleaf tree species than coniferous species (Mayer et al., 2022). Across the different systems, the maximum value of SCS (64.38 t ha-1) was observed in the Agri-silvi-horticulture system, while the minimum value (15.67 t ha-1) was observed in Agri+silvi. This may be due to adding more inorganic fertilizers than organic fertilizers (Bargali & Bargali, 2020). The decline in soil carbon with elevating soil depth could be due to slow carbon flow and soil compaction at deeper levels (Dar & Somaiah, 2015). Due to continuous addition of partially decomposed plant and animal matter in the surface layer the SCS decreases with soil depth.

Over 1800 m of SOM and SOC were observed high in Apple+Malta models at 0-15 cm and low at 15 -30 cm. Oak+Malta+Peach showed maximum soil carbon stock and minimum in the Oak-based agroforestry model. With increasing depth, the SOM, SOC and SCS decrease. Soil organic matter is essential for influencing soil's physical, chemical, and biological characteristics, creating an environment that supports biological activity and life. Once depleted, it generally takes a long time to recover. The differences in organic matter levels across the study sites may be due to variations in plant species composition and the amount of organic material on the soil surface (Sahoo, 2020). As altitude increased, temperatures drop, which reduces microbial activity. This leads to a slower rate of SOM decomposition, causing plant-derived SOM to be more effectively stored in the soil, especially in tropical and subtropical regions (Sundqvist et al., 2013). Organic matter varies with the study sites, likely due to differences in vegetation composition and the amount of biomass on the top layer. For example, the Agri-silvi-horticulture system, which had a greater variety of plant species, produced more litter, leading to higher organic matter at the site. Additionally, the accumulation of humus, other waste materials, and the incorporation of plant debris may have contributed to the increased organic matter (Pinho et al., 2010). They are also associated with It has also been linked to higher concentrations of SOM as Dori et al. (2022) found that home garden (AHS) systems positively impact soil properties.

Clay and Sand were strongly inversely related, meaning higher sand content typically led to lower clay content, and vice versa, with silt being positively correlated to Sand but negatively correlated to clay content. SOC and SOM were strongly correlated, indicating that more organic matter led to more organic carbon in the soil. Soil moisture was positively correlated with clay, showing that soils with more clay retained more moisture. Conversely, Sand tended to reduce moisture content. The pH was only weakly correlated with most variables, but it was positively related to SOC and SOM. The SOC and SOM showed a positive significant association (p < 0.01) with pH. Changes in soil pH can influence the availability of soil nutrients and the nutritional health of vegetation. The percentage of silt showed a positive correlation (p < 0.01) with sand, and the percentage of clay showed a negative relationship (p<0.01) with the percentage of sand and silt. However, Hu et al. (2019) found that the percentage of sand had a positive relationship with silt. Clay (%) showed a positive connection (p>0.05) with moisture content (%). The positive relationship between clay and soil moisture may be due to decreased in the size of soil particles as earlier mentioned by Rong (2017). SOM had a perfect positive correlation (p<0.01) with SOC. Xiangrong et al. (2022) also found the same results, with SOM positively associated with SOC (p<0.01).

Based on correlation matrix, Principal component analysis was conducted (Table 6). It reveals the relationship between different variables including, sand, silt, clay, BD, pH, WHC, SOM, SOC and SCS and their sites (Agri-silviculture system, Agri-silvi- horticulture system, Agri-horticulture system at LAAS, LAASH, LAAH: at low altitude, MAAS, MAASH, MAAH at mid-altitude, HAAS, HAASH, HAAH at high altitude, 1; 0–15 surface layer and 2; 15–30 subsurface layer) (Fig. 2). It displayed the loading values of the first two principal components, which indicate the contribution of each variable to the principal components. These two principal components is 58.59% of the variance (33.63% from PC1 and 24.96% from PC2). The two principal components with eigenvalues more than 1 were selected, according to method given by Kaiser (1960). The analysis helps identify patterns and interrelationships among the soil properties, highlighting the most influential variables in the dataset. PC1 (33.63%) is mainly associated with sand, silt, SOC %, and SOM, with clay having a substantial negative contribution. Soils that are sandy or have higher organic carbon content have higher scores on PC1. So, PC1 represents a general soil texture and organic matter component, with sand and organic content being the key contributors. PC2 (24.96%) is influenced by pH, SOC%, SOM, and clay, positively correlated to clay and SOC%. This suggests that PC2 represents a component related to soil acidity, organic carbon content, and soil structure (with clay contributing positively).

Table 6. Soil organic matter (%), Soil organic carbon (%) and soil carbon stock (t ha-1) of different agroforestry systems along altitudinal gradient

AS- Agri-ilviculture system, AH- Agri-horticulture system and ASH- Agri-silvi-horticulture system

Table 7. Pearson correlation analysis between soil physiochemical properties

MC (%) =Moisture content, SOC (%) = Soil organic carbon, SOC (t ha-1) = Soil organic stock, WHC= Water holding capacity, SOM (%) = Soil organic matter. *Significant at p > 0.05, **significant at p < 0.01.