The Role of Grassland Land Use in Enhancing Soil Resilience and Climate Adaptation in Periurban Landscapes

Abstract

Urbanisation and land-use change are among the main pressures on soil health in periurban areas, but the multifunctionality of grassland soils is still not sufficiently recognised. In this study, the physical and chemical properties of soils under grassland, forest and croplands in the periurban area of Zagreb were investigated in a two-year period. Grasslands consistently exhibited multifunctional benefits, including high organic matter content (4.68% vs. 2.24% in cropland), improved bulk density (1.14 vs. 1.24 g cm⁻³) and an active carbon cycle indicated by increased CO₂ emissions (up to 1403 kg ha⁻¹ day⁻¹ in 2021). Forest soils showed the highest aggregate stability (91.4%) and infiltration (0.0006 cm s⁻¹), while croplands showed signs of structural degradation with the highest bulk density and lowest water retention (39.9%). Temporal variation showed that grassland was particularly responsive to favourable climatic conditions, with soil porosity and water content improving yearly. Principal component analysis showed that soil structure, biological activity and moisture regulation were linked, with grassland plots favourably positioned along the axes of resilience. The absence of tillage and the presence of permanent vegetation cover contributed to their high capacity for climate and water regulation and carbon sequestration. These results emphasise the importance of protecting and managing grasslands as an important component of urban green areas. Practices such as mulching, minimal disturbance and continuous cover can maximise the ecosystem services of grassland soils. In addition, the results highlight the potential risk of trace metal accumulation in cropland and grassland soils located near urban and farming infrastructure, underlining the need for regular monitoring in periurban environments. Integrating grassland functions into urban planning and policy is essential for improving the sustainability and resilience of periurban landscapes.

1. Introduction

Soils support ecosystems and provide essential services, including carbon sequestration, water filtration, and food production [1]. However, urbanisation and land-use change can significantly impact soil properties and ecosystem services, often leading to degradation, compaction and loss of fertility [2]. Understanding these dynamics is crucial to ensure sustainable soil management and mitigate the negative impacts of urban expansion.

Like many other fast-growing cities, the Croatian capital, Zagreb, faces major challenges due to increasing urban sprawl, which affects soil health and ecosystem stability. Converting agricultural land and natural habitats into residential and industrial areas impairs soil functions such as water retention, organic matter accumulation and erosion control. Grasslands and forests provide important ecosystem services, including climate and flood regulation [3]. Previous studies confirm that agricultural land has a significantly higher degree of compaction than land with natural vegetation in arid [4] and humid conditions [5]. In an urban environment, the comparison of land use confirms a similar compaction behaviour [6,7,8]. In addition, the resilience of forest and grassland soils to extreme precipitation events is enormous compared to bare soils, mainly due to frequent tillage and herbicide use [9]. Bare cultivated soils significantly reduce the organic carbon content, deteriorate soil structure and increase the susceptibility of cultivated land to water erosion [10].

In addition to the deterioration of the physical properties, the chemical properties of the soil also change considerably under anthropogenic influences. Various factors are responsible for the differences in soil chemical properties between land-use types in urban and periurban areas. Agricultural soils often receive a high nutrient input through fertilisation and pesticide use, which often leads to increased phosphorus and nitrogen levels, while at the same time, the organic carbon content of the soil decreases due to frequent tillage [11,12]. In contrast, forest soils typically accumulate organic matter via natural litterfall processes, maintaining a higher carbon-to-nitrogen ratio and a slightly more acidic state due to slower decomposition rates [13]. Grasslands often retain high levels of organic matter, which supports nutrient cycling, but this level fluctuates depending on the frequency of mowing and mulching. In addition, proximity to cities poses contamination risks. Heavy metals, such as Pb, Zn and Cu, are often more prevalent in urban soils due to traffic emissions, atmospheric deposition and wastewater use, and particularly affect croplands near industrial areas [14,15,16]. Soil compaction and sealing in urban areas further disrupt chemical gradients by limiting the input of organic matter and altering mineralisation processes [17]. Thus, soil chemical responses in urban and periurban contexts are land-use dependent and spatially heterogeneous, reflecting both natural ecosystem functions and the intensity of human-induced pressures. Sustainable land-use planning must take these considerations into account in order to improve urban resilience and environmental sustainability.

The main objective of this study is to evaluate and compare the soil physical, chemical, and biological properties under grassland, cropland, and forest land uses in the periurban area of Zagreb over a two-year period. Special attention is given to grassland systems due to their potential role in enhancing soil resilience and providing ecosystem services. The study does not aim to propose land-use conversions (e.g., grassland to cropland), but rather to assess how different land uses contribute to climate adaptation and soil multifunctionality in urban planning contexts. By examining these land types, the main trends will be identified. The resilience of the different land-use categories will be assessed, and their contribution to ecosystem services will be evaluated. The results will inform urban planning policy and enable strategies to maintain ecological balance while adapting to urban growth.

There are several challenges to soil health in Zagreb soils: overuse [18], low biodiversity [19], structural degradation [20], low soil quality [21], soil compaction [22], low soil organic matter content [7], abandonment of land [23], risk of flooding [24] and chemical contamination [25]. Heterogeneous land-use patterns and the lack of systematic soil condition monitoring across different land types exacerbate these challenges. With its diverse land uses and types, Zagreb significantly impacts soil, environment, and ecosystem services due to land-use change. The choice of the city of Zagreb as a study area was motivated by the coexistence of native forests, agricultural land and urbanised zones in a compact spatial framework. This mosaic-like landscape provides a unique opportunity to compare land uses with different intensities of human intervention against a shared climate and topographic background.

However, despite the obvious pressure on soil resources, there is a lack of long-term data and integrated assessments comparing the resilience of different land uses, especially grassland, in urban and periurban areas. Deforestation and cultivation have rapidly expanded around Zagreb in recent decades, but the effects on the soil’s physical, chemical and biological properties remain poorly understood. This study fills this gap by conducting a two-year comparison of soil conditions in forests, grasslands and croplands in eastern Zagreb. The novelty of this study lies in the systematic and temporal assessment of soil health indicators in a periurban context, with a particular focus on grassland, a land use often overlooked in climate adaptation and flood mitigation strategies.

2. Materials and Methods

2.1. Study Area, Environmental Conditions, and Soils

Zagreb is Croatia’s historical capital, consisting of densely populated urban areas and surrounding periurban regions with mixed land use. Zagreb’s geographic setting includes a combination of lowland and hilly terrain, with a temperate continental (Dfb) climate [26] that influences land-use dynamics. The mean annual (1961–2023) precipitation is 887 mm, with June (95 mm) and October (90.4 mm) as the most humid and January (50.9 mm) and February (46.4 mm) as the most arid part of the year. The mean annual (1961–2023) temperature is 11.7 °C, ranging from 0.6 °C (January) to 22 °C (July) [27]. The Zagreb metropolitan area has diverse conditions. Air temperature, precipitation, and potential evapotranspiration are higher in the city centre than in the suburban part of the city [28], indicating the urban heat island effect [29]. The study area is in Zagreb’s eastern part (Figure 1) (45.8527778 N; 16.1797222 E, 118 m a.s.l.). Soils are dominantly composed of silty clay, Stagnosols, and Fluvisols [30]. General soil characteristics are presented in

Table 1. Soil physical, chemical and mechanical properties on the upper horizon of studied land uses.

Soil Properties	Forest	Cropland	Grassland
Total organic C [g/kg]	28.5	13.8	22.7
pH in KCl [w w−1 1:5]	4.43	7.35	8.30
EC [dS/m]	0.050	0.071	0.151
P2O5 [mg/kg]	45.2	41.8	45.4
Exchangeable K [mg/kg]	43.7	88.7	118.3
Ntotal [g/kg]	2.25	1.55	2.79
Bulk density [g/cm3]	0.92	1.32	1.14
Sand [%]	9.5	5.8	16.2
Silt [%]	80.0	81.5	72.8
Clay [%]	10.5	12.8	11.0

.

2.2. Sampling Design and Laboratory Procedures

The study focuses on three primary land-use types: cropland, grassland, and forest, representing distinct ecological and management characteristics. The forest (native vegetation, dominated mainly by Quercus petraea and Robinia pseudoacacia) is >200 years old under the same land use. Sustainable management practices that permit high soil health in forest land use are the absence of soil disturbance and mineral fertilisation. At the same time, the natural decay of trees and leaves accumulates organic matter on topsoil. Grassland land use has been established since 1996, and grass–clover mixtures dominate. The annual management consists of mechanised mowing 3–4 times per year, while the mowed biomass remains as mulch and a source of nutrients. Sustainable management practices that permit high soil health in grassland land use are mulching, no soil disturbance, and permanent vegetation cover. Conversely, cropland land use is a nearby control area with low soil health, similarly conventionally managed since 2007. Management practices performed in croplands that lead to low soil health are conventional tillage (annual plowing followed by two disking and roto-harrowing before sowing), intensive mineral fertilisation, lack of organic fertilisation and the absence of cover cropping and mulching practices. The main crops grown on cropland are winter wheat, barley, oats, soybeans, and maize. Herbicides and insecticides are regularly used annually. The specific type and rate of mineral fertilisers and herbicides applied varied depending on the crop in rotation and followed standard conventional practices typical for the region. As these inputs were not the focus of this study, they are not detailed further. All land-use types in this work are nearby and have similar topographic and geomorphological characteristics.

The sampling strategy consists of two time samples, in 2021 and 2023, to avoid possibly wrong conclusions from one-time observations, a common shortcoming in agricultural and environmental studies. Sampling points are randomly chosen in a straight direction and identified using a GPS.

Eight sampling points were selected in each land-use area (Figure 1), making 24 sampling points per measurement, for a total of 42. Water infiltration measurements were conducted near sampling points with a minidisc infiltrometer (Decagon Devices, Pullman, WA, USA). Undisturbed core samples were conducted at 0–10 soil depths using 100 cm³ cylinders. Soil cores were wetted to measure water-holding capacity (WHC) and dried in an oven at 105 °C for 24 h to obtain the bulk density (BD) and soil water content (SWC) [31]. Near soil core sampling, undisturbed samples (8 per land-use class, 24 in total) were sampled by shovel at a 0–10 cm depth and stored in plastic boxes to preserve the aggregation. In current moisture conditions, aggregates were broken and gently manipulated by hand into small pieces before dry sieving [32]. At the same time, near sampling points, in situ measurements of soil CO₂ emissions and infiltration were carried out. Soil CO₂ emissions were measured on bare soil in a closed static chamber during 30 min of incubation using an infrared CO₂ detector (GasAlertMicro5 IR, BW Technologies Honeywell, Calgary, AB, Canada, 2011). Air temperature was measured with Testo 610 (Testo SE & Co KGaA, Lenzkirch, Germany, 2011) and air pressure with Testo 511 (Testo SE&Co KGaA, Lenzkirch, Germany, 2011). CO₂ emissions (kg ha⁻¹ day⁻¹) were calculated according to Widén & Lindroth [33].

The percentage of water-stable aggregates (WSA) was determined using a wet sieving apparatus by soaking 4 g of aggregates (diameter 0.4–0.5 mm) in distilled water for 3 min [34]; sieving was continued until only the sand particles were left on the sieves after replacing cans with a dispersing solution (2 g NaOH/L). Afterwards, cans were dried at 105 °C and weighed, and the percentage of WSA was obtained with the following equation:

$$\mathrm{WSA}={\displaystyle \frac{Wds}{Wds+Wdw}},$$

(1)

WSA is the percentage of stable water aggregates, Wds is the weight of aggregates dispersed in dispersing solution (g), and Wdw is the weight of aggregates dispersed in distilled water (g). The remaining aggregate sizes were milled, sieved (2 mm mesh), and analysed to determine soil chemical properties, clay, silt, and sand content.

Soil particle size distribution was determined by the pipette method with wet sieving (sand fractions) and sedimentation (silt and clay fractions) after soil dispersion with sodium pyrophosphate. Soil pH was determined in a 1:5 soil–water suspension, using the electrometric method with a Beckman pH meter Φ72 in a 1 mol dm⁻³ KCl solution. Soil organic matter (SOM) concentration was measured using the oxidation of organic matter by a potassium dichromate–sulfuric acid mixture, followed by back titration of the excessive dichromate by ferrous ammonium sulphate solution [35]. Available phosphorus was determined using the ammonium-lactate extraction method, and exchangeable potassium was measured in the same extract by flame photometry. Finally, the Cr, Mn, Fe, Ni, Cu, and Zn contents were determined using a portable X-ray fluorescence (pXRF) spectrometer (VantaTM XRF analyser C Series) (Olympus, Waltham, MA, USA, 2019).

2.3. Statistical Analysis

Data normality and homogeneity of variances were assessed using the Shapiro–Wilk and Lilliefors tests (p > 0.05). Normality and homogeneity of variances were only observed in soil water content, bulk density, and infiltration data. The mean weight diameter dataset has normality requirements after the Box–Cox transformation. At the same time, the other variables did not follow the Gaussian distribution and homogeneity of variances, even after logarithm, square root, and Box–Cox transformations. A parametric two-way ANOVA was applied to the soil water content, bulk density, infiltration, and Box–Cox transformed mean weight diameter data to identify differences between sampling years and land-use treatments. A nonparametric Friedman ANOVA test was applied to observe differences between sample years in the other soil properties. The differences between land-use treatments were carried out using the nonparametric Kruskal–Wallis ANOVA test (K–W). If significant differences were observed at p < 0.05, Fisher’s post hoc (for MWD Box–Cox, SWC, BD, and infiltration rate) and multiple comparisons rank tests were applied [36,37]. Principal component analysis (PCA) was performed, based on the correlation matrix, to identify the relationships among all soil properties. The Box–Cox transformed data were used for the PCA since it was the closest to normality when observing all datasets. All analyses were performed using Statistica 12.0 for Windows.

3. Results

3.1. Soil Physical Properties

Significant differences in physical soil properties were found between land-use types and sampling years (

Table 2. Land-use effect on soil physical properties during 2021 and 2023. Different letters indicate significant differences between sampling years (capital letters) and land use (lower-case letters). Kruskal–Wallis (K–W) tests are shown for each comparison between land uses and Friedman ANOVA between sampling years at water-stable aggregates (WSA), sand, silt, and clay content. Two-way ANOVA is shown for each land use-sampling year comparison at bulk density (BD), soil water content (SWC), water holding capacity (WHC), infiltration, and temperature.

Property	Year	Land Use	Mean		Med	Min	Max	S.D.	C.V.	SE
BD (g cm−3)	2021	Forest	1.164	b	1.137	1.028	1.337	0.112	9.6	0.040
		Cropland	1.245	a	1.242	1.077	1.398	0.096	7.7	0.034
		Grassland	1.236	abA	1.241	1.153	1.294	0.049	4.0	0.017
	2023	Forest	1.107	b	1.096	0.976	1.215	0.075	6.8	0.027
		Cropland	1.232	a	1.235	1.203	1.269	0.022	1.8	0.008
		Grassland	1.048	bB	1.055	0.918	1.125	0.069	6.6	0.024
SWC (%)	2021	Forest	36.1		35.6	28.2	44.4	5.97	16.5	2.11
		Cropland	36.7	B	37.7	23.9	45.0	6.26	17.1	2.21
		Grassland	30.5	B	29.7	13.3	46.1	11.72	38.5	4.14
	2023	Forest	39.1	b	36.7	30.6	54.8	7.84	20.1	2.77
		Cropland	49.0	aA	49.0	45.4	53.1	2.81	5.7	0.99
		Grassland	49.6	aA	50.9	39.0	55.3	5.06	10.2	1.79
Temp. (°C)	2021	Forest	16.5	aB	16.4	16.2	17.5	0.41	2.5	0.14
		Cropland	14.5	abB	14.6	14.0	14.7	0.21	1.5	0.08
		Grassland	11.4	bB	11.4	11.2	11.5	0.11	1.0	0.04
	2023	Forest	23.3	bA	23.3	23.2	23.5	0.09	0.4	0.03
		Cropland	25.6	abA	25.6	24.4	26.8	0.94	3.7	0.33
		Grassland	27.9	aA	28.0	27.3	28.4	0.49	1.7	0.17
WHC (%)	2021	Forest	51.9	a	51.8	43.0	59.4	6.46	12.5	2.28
		Cropland	39.9	b	37.4	27.5	58.6	12.44	31.2	4.40
		Grassland	49.2	ab	51.2	39.0	55.6	5.85	11.9	2.07
	2023	Forest	51.7	ab	51.3	49.1	54.5	1.95	3.8	0.69
		Cropland	45.1	b	41.9	35.7	59.7	7.94	17.6	2.81
		Grassland	54.6	a	55.2	51.0	57.1	1.96	3.6	0.69
WSA (%)	2021	Forest	91.5	a	90.8	89.6	94.1	1.65	1.8	0.58
		Cropland	74.7	b	75.1	67.9	83.0	4.66	6.2	1.65
		Grassland	84.1	ab	83.2	79.3	91.1	4.29	5.1	1.52
	2023	Forest	91.2	a	91.0	88.2	94.3	1.82	2.0	0.64
		Cropland	81.0	b	81.1	72.5	88.6	4.44	5.5	1.57
		Grassland	81.3	b	81.3	63.9	90.5	8.81	10.8	3.11
Infiltration (cm/s)	2021	Forest	0.00062		0.00062	0.00044	0.00088	0.00013	21.8	0.00005
		Cropland	0.00050		0.00042	0.00034	0.00074	0.00016	32.5	0.00006
		Grassland	0.00061		0.00058	0.00029	0.00105	0.00029	47.4	0.00010
	2023	Forest	0.00057		0.00061	0.00040	0.00071	0.00011	20.0	0.00004
		Cropland	0.00067		0.00064	0.00057	0.00078	0.00008	12.1	0.00003
		Grassland	0.00056		0.00056	0.00036	0.00072	0.00012	20.6	0.00004
Sand (%)	2021	Forest	7.4	a	7.2	3.3	11.0	3.11	42.0	1.10
		Cropland	5.2	ab	3.9	2.7	10.0	3.10	59.7	1.10
		Grassland	3.8	b	3.6	2.9	4.9	0.80	21.2	0.28
	2023	Forest	7.5	a	7.2	3.5	11.2	3.09	41.1	1.09
		Cropland	5.2	ab	4.0	2.9	10.0	2.95	56.6	1.04
		Grassland	4.4	b	4.2	3.1	7.7	1.46	32.6	0.52
Silt (%)	2021	Forest	76.1		76.4	72.2	79.1	2.04	2.7	0.72
		Cropland	75.8		75.0	73.1	83.4	3.46	4.6	1.22
		Grassland	77.0		76.4	72.8	83.3	4.11	5.3	1.45
	2023	Forest	75.9		76.2	72.0	78.9	2.04	2.7	0.72
		Cropland	75.7		74.8	72.9	83.2	3.48	4.6	1.23
		Grassland	76.9		76.2	72.6	83.1	4.07	5.3	1.44
Clay (%)	2021	Forest	16.5		15.8	11.8	23.0	4.68	28.3	1.65
		Cropland	19.0		19.8	13.2	24.2	4.98	26.2	1.76
		Grassland	19.2		18.8	13.3	24.1	4.37	22.7	1.54
	2023	Forest	16.6		15.8	11.8	23.0	4.65	28.0	1.64
		Cropland	19.1		20.0	13.4	24.2	4.90	25.6	1.73
		Grassland	18.7		18.5	13.3	24.1	4.04	21.7	1.43

). In both years, the bulk density was significantly higher in croplands than in forests (p < 0.05), while grassland values were intermediate. In 2023, grassland BD decreased significantly compared to 2021. Soil water content (SWC) increased significantly in 2023 across croplands and grasslands compared to 2021. In 2021, cropland and forest had higher SWC than grassland (p > 0.05). In contrast, cropland and grassland had a significantly higher SWC than forest in 2023 (p < 0.05). Water-holding capacity was highest in forest and grassland soils. In 2021, the WHC in croplands was significantly lower (p < 0.05) than in forests. In 2023, grassland areas had a significantly higher WHC than croplands (p < 0.05).

The WSA showed clear differences between the land uses (Table 2). Forest soils showed significantly higher WSA than cropland soils in both years (p < 0.05). A significant year effect was only observed for grassland, where WSA decreased in 2023 (p < 0.05). Infiltration rates did not show statistically significant differences among land-use types (p > 0.05) or between sampling years (p > 0.05), although slight trends indicated higher values for grassland and forests compared to croplands. Soil texture (sand, silt and clay content) showed no significant effect by year but showed slight differences between land uses. Grassland had a significantly higher sand content than croplands in both years (p < 0.05), while forest and cropland soils had a higher proportion of silt. Clay content did not differ significantly across land uses or years (p > 0.05).

3.2. Soil Chemical Properties

In 2021, CO₂ emissions were significantly higher on grassland than on cropland, while the forest had average values (

Table 3. Land-use effect on soil chemical properties during 2021 and 2023. Different letters indicate significant differences between sampling years (capital letters) and land use (lower-case letters). Kruskal–Wallis (K–W) tests are shown for each comparison between land uses and Friedman ANOVA between sampling years at CO₂ emissions, pH, Cr, Fe, Ni, Cu, and Zn. Two-way ANOVA is shown for each land use-sampling year comparison at soil organic matter (SOM) and Mn.

Property	Year	Land Use	Mean		Med	Min	Max	S.D.	C.V.	SE
CO2 (kg ha−1 day−1)	2021	Forest	66.8	abB	67.3	54.7	79.9	8.8	13.2	3.1
		Cropland	5.6	bB	5.5	4.2	8.4	1.4	26.0	0.5
		Grassland	1403.0	aA	1059.8	687.4	3492.3	940.1	67.0	332.4
	2023	Forest	122.0	bA	116.3	95.5	157.8	22.7	18.6	8.0
		Cropland	123.0	bA	116.3	70.6	232.6	47.9	38.9	16.9
		Grassland	200.9	aB	199.3	149.5	249.2	38.7	19.3	13.7
pH	2021	Forest	4.43	bA	4.04	3.69	7.45	1.25	28.2	0.44
		Cropland	7.36	aA	7.30	7.16	7.75	0.20	2.7	0.07
		Grassland	7.25	abB	7.71	3.93	7.86	1.35	18.6	0.48
	2023	Forest	3.66	bB	3.60	3.40	4.15	0.23	6.2	0.08
		Cropland	7.06	abB	7.08	6.86	7.27	0.12	1.7	0.04
		Grassland	7.36	aA	7.37	7.18	7.45	0.08	1.1	0.03
SOM (%)	2021	Forest	7.97	a	5.78	2.78	17.66	4.82	60.5	1.70
		Cropland	2.15	c	2.18	1.41	2.69	0.44	20.5	0.16
		Grassland	4.76	b	4.49	3.78	6.89	0.93	19.5	0.33
	2023	Forest	6.63	a	5.53	4.32	13.31	2.97	44.8	1.05
		Cropland	2.33	b	2.29	1.95	2.75	0.28	12.2	0.10
		Grassland	4.59	a	4.63	2.98	6.10	0.91	19.9	0.32
Cr (mg/kg)	2021	Forest	110	b	96	85	195	36	32.8	12.80
		Cropland	197	a	197	184	211	9	4.4	3.07
		Grassland	180	abB	184	128	201	23	12.9	8.16
	2023	Forest	103	b	104	70	125	19	18.3	6.65
		Cropland	203	b	198	189	228	14	6.9	4.96
		Grassland	205	aA	205	180	223	17	8.1	5.84
Mn (mg/kg)	2021	Forest	531	b	563	256	663	141.1	26.6	49.9
		Cropland	617	b	618	545	704	50.9	8.3	18.0
		Grassland	682	a	740	360	793	139.9	20.5	49.5
	2023	Forest	536	b	553	382	684	109.1	20.4	38.6
		Cropland	677	a	683	647	701	20.8	3.1	7.3
		Grassland	686	a	681	620	748	45.7	6.7	16.2
Fe (mg/kg)	2021	Forest	30,362	b	29,589	27,986	38,176	3302	10.9	1168
		Cropland	40,216	aA	40,149	39,588	41,147	545	1.4	193
		Grassland	37,167	ab	37,826	29,419	39,884	3278	8.8	1159
	2023	Forest	29,217	b	29,833	25,490	30,953	1854	6.3	655
		Cropland	39,282	aB	39,022	38,595	40,161	601	1.5	213
		Grassland	37,775	ab	37,515	36,443	40,027	1070	2.8	378
Ni (mg/kg)	2021	Forest	41.0	b	32.0	21.0	109.0	28.0	68.4	9.9
		Cropland	102.3	a	103.5	95.0	105.0	3.3	3.2	1.2
		Grassland	97.3	ab	106.0	27.0	120.0	29.1	29.9	10.3
	2023	Forest	30.8	b	31.5	26.0	36.0	4.0	12.9	1.4
		Cropland	106.0	a	108.0	94.0	113.0	6.3	5.9	2.2
		Grassland	108.3	a	108.5	100.0	117.0	6.0	5.6	2.1
Cu (mg/kg)	2021	Forest	19.1	b	16.5	15.0	36.0	6.9	36.2	2.5
		Cropland	31.8	ab	32.0	27.0	35.0	2.5	7.9	0.9
		Grassland	34.0	a	37.5	15.0	40.0	8.3	24.5	2.9
	2023	Forest	16.3	b	17.0	12.0	19.0	2.4	14.6	0.8
		Cropland	30.4	a	29.5	28.0	34.0	2.3	7.7	0.8
		Grassland	35.3	a	36.5	29.0	40.0	4.0	11.4	1.4
Zn (mg/kg)	2021	Forest	75.1	b	73.1	67.0	95.0	9.1	12.2	3.2
		Cropland	97.3	a	97.0	88.0	105.0	4.9	5.0	1.7
		Grassland	90.3	abB	91.5	74.0	96.0	7.0	7.7	2.5
	2023	Forest	71.0	b	70.5	66.0	77.0	3.5	5.0	1.3
		Cropland	96.1	a	96.5	91.0	102.0	3.3	3.4	1.2
		Grassland	95.6	aA	96.0	88.0	101.0	4.0	4.1	1.4

). In 2023, emissions decreased in grassland and increased in forest and cropland, resulting in statistically higher CO₂ emissions than in forest and cropland land. The annual effect was significant for all land uses. Grassland had significantly higher emissions in 2021 than in 2023, while the opposite effect was observed for forest and cropland land use. Soil pH was significantly lower in forest soils than in cropland (in 2021) and grassland (in 2023) (p < 0.05). A significant decrease in pH was observed in forest soils in 2023 compared to 2021, while the pH of cropland and grassland soils was significantly higher in 2021 than in 2023. Soil organic matter content differed significantly between land uses in both years (p < 0.05). In 2021, forest soils had the highest SOM values, followed by grassland, while croplands had the lowest. In 2023, forest and grassland soils had a significantly higher SOM content than cropland soil.

All trace elements (Table 3) showed lower values in forest soils than in grassland and cropland soils. In 2021, cropland soils had significantly higher Cr concentrations than forest soils, while in 2023, the Cr concentrations in grassland soils were significantly higher than in cropland and forest soils. Temporal changes occur in grassland, where significantly higher Cr concentrations were found in 2023 than in 2021. In 2021, Mn concentrations were significantly higher in grassland, while in 2023, they were significantly lower in forest soils than in other land-use types. Fe concentrations were significantly lower in forest soils than in grassland and croplands in both years. A significant temporal decrease in Fe concentrations was observed in cropland soils. The Ni and Zn concentrations in the soil follow a similar trend. In 2021, significantly higher concentrations were measured in croplands than in forest soils, while grassland soils were in between. In 2023, forests had significantly lower Ni and Zn concentrations than grassland and croplands. In grassland soils, Zn concentrations were significantly higher in 2023 than in 2021. In addition, soil Cu concentrations were significantly lower in forest soils than in grassland soils in 2021. In 2023, soil Cu concentrations were significantly higher in cropland and grassland soils than in forest soils.

3.3. Principal Component Analysis

Factor 1 explained 40.16% of all variance, and Factors 2, 3, 4 and 5 explained 13.70%, 12.10%, 9.89% and 5.71%, respectively. Factor 1 had high positive loadings in soil pH, Cr, Mn, Fe, Ni, Cu, and Zn concentrations and high negative loadings in WSA and SOM (

Table 4. Principal component analysis results. In bold, the eigenvalues are retained in each factor.

Property	Factor 1	Factor 2	Factor 3	Factor 4	Factor 5
Water holding capacity (%)	−0.23	0.57	0.41	0.17	0.29
Soil water content (%)	0.27	0.49	0.35	−0.50	−0.10
Bulk density (g cm−3)	0.27	−0.73	−0.25	0.14	−0.21
Water stable aggregates (%)	−0.74	0.18	0.10	0.21	−0.08
Infiltration rate (cm/s)	−0.16	0.11	0.19	0.06	−0.86
Temperature (°C)	−0.01	0.56	0.40	−0.55	−0.13
CO2 (kg ha−1 day−1)	−0.06	0.36	0.19	0.73	−0.24
pH	0.95	0.11	−0.12	0.17	0.05
Soil organic matter (%)	−0.66	0.29	0.16	0.47	0.18
Cr (mg/kg)	0.94	0.15	0.05	−0.09	0.00
Mn (mg/kg)	0.65	0.34	−0.19	0.40	−0.13
Fe (mg/kg)	0.95	−0.09	−0.07	−0.10	−0.06
Ni (mg/kg)	0.96	0.19	−0.01	0.08	0.02
Cu (mg/kg)	0.89	0.20	0.00	0.25	0.07
Zn (mg/kg)	0.94	0.06	0.09	−0.05	0.02
Sand (%)	−0.49	0.24	−0.54	−0.25	−0.09
Silt (%)	0.01	0.53	−0.73	0.05	0.03
Clay (%)	0.25	−0.47	0.81	0.14	0.06

). Factor 2 had high positive loadings in WHC and soil temperature and high negative loadings in BD. Factor 3 had positive loading in clay and negative in silt and sand content, while factor 4 had positive loading in CO₂ emissions and negative in SWC. Factor 5 was mainly defined by infiltration rate (negative loading), representing a hydrological property gradient.

4. Discussion

4.1. Soil Properties

This study highlights the significant influence of land-use type and temporal variation on soil’s physical and chemical properties in an urban environment. The results show that land-use decisions directly impact soil structure, water regulation and carbon cycling, all of which are important ecosystem services.

Bulk density was consistently higher on cropland, confirming the long-term compacting effect of conventional tillage and machine passes [4]. When ploughed regularly over the years, BD tends to increase, which was demonstrated in clay loam [38], silty clay loam [21], and silty [39] and sandy [40] soils. In addition, conventional farming methods usually involve frequent machine passes. In conventional cultivation of cereals or root crops alone, every centimetre of soil is worked 3 to 8 times with wheels [41]. The reduction in BD observed in grassland in 2023 suggests that the structure of such systems can recover over time without tillage. Similar results were reported by Franzluebbers et al. [42] and Six et al. [43]. Furthermore, these results support the protective role of herbaceous vegetation in maintaining soil porosity and infiltration.

Soil water content and WHC showed effects on land use and time course. In 2023, there was a significant increase in SWC on grassland and cropland, likely influenced by seasonal weather patterns, including higher precipitation and lower evapotranspiration. In contrast, the SWC of forest soils remained relatively stable or slightly decreased. This reduction can be attributed to increased water uptake by the root systems of mature trees, which have a deeper and more extensive root profile than grasses or crops. Tree roots can extract water from deeper soil layers and deplete the upper soil water table, especially during the growing season. A similar trend was noted by Wang et al. [44], who observed that dense forest vegetation significantly impairs soil water dynamics through high transpiration rates and interception losses, resulting in lower near-surface soil moisture despite a higher total storage potential. In addition, Evrendilek et al. [45] found that greater biomass and a denser canopy in forest ecosystems can increase evapotranspiration, decreasing surface SWC even when deeper layers retain substantial moisture. This phenomenon emphasises the dual role of forests in water storage and the active water cycle in vegetation.

In contrast, permanent herbaceous cover in grasslands has likely enhanced infiltration and near-surface moisture retention, as Gajić [46] and Ajayi et al. [47] confirmed. These systems benefit from high porosity and dense root networks in the topsoil, which improve soil structure and facilitate water infiltration and retention. Consequently, the grasslands in our study showed an improved WHC in 2023, suggesting that extensively managed grasslands can act as a buffer against climate variability and support a stable hydrological function. Forest soils exhibited consistent WHC in both years, likely due to their stable structure, deeper root activity and higher SOM, which buffers water availability over time. A significantly higher WSA characterises forest soils as evidence of superior structural stability. This is consistent with previous findings that minimal disturbance and continuous organic inputs from litterfall promote aggregate formation and reduce susceptibility to erosion and compaction [45,48]. The high WSA values indicate a strong resistance to degradation under forest cover. In contrast, croplands showed the poorest structural integrity due to intensive tillage and lack of organic inputs [49,50]. Despite little disturbance, the decline in WSA in grasslands between 2021 and 2023 could be due to environmental stress or insufficient biomass return, warranting further monitoring.

Interestingly, infiltration rates did not show statistically significant differences between land uses, possibly due to the high variability within each land-use type. However, the observed trends support the notion that grassland and forest soils are more permeable than cropland, consistent with previous studies [5,8,47]. Soil organic matter content was significantly higher in forest and grassland soils than in cropland, highlighting the detrimental effects of intensive agriculture on carbon stocks [11]. Over-oxidation of ploughed soils increased inter- and intra-aggregate deterioration and loss of SOM [4]. The decline in SOM in forest soils between 2021 and 2023 is unexpected and likely related to environmental conditions or reduced litter input, emphasising the importance of long-term monitoring. Stable SOM concentrations in grasslands indicate high resilience and the potential for long-term carbon sequestration, especially under current mulch management.

CO₂ emissions were highest in grassland in 2021 due to increased microbial activity and carbon turnover in high organic input systems [12]. The decline in emissions in 2023 could be due to temperature, humidity, or substrate availability changes. The increasing emissions in forest and cropland soils in 2023 indicate increased microbial respiration under favourable conditions. However, despite this temporary increase, CO₂ emissions from cropland soils remained relatively low overall, likely a consequence of long-term intensive tillage, reduced organic matter input and microbial depletion. Continuous plowing accelerates the depletion of labile carbon pools at the beginning of land use, leading to a progressive decline in soil organic carbon and microbial biomass over time [43,51]. Without a regular supply of crop residues, manure or cover crops, the microbial community becomes carbon limited, reducing soil respiration [52]. In addition, the deteriorated soil structure caused by tillage restricts gas diffusion and microbial habitat, further limiting microbial activity [53].

The acidic pH observed in forest soils is a typical feature of ecosystems with high organic matter input, slow decomposition and limited base cation cycling [13]. In such systems, organic acids released during litter decomposition, combined with minimal disturbance, lead to gradual acidification. In addition, tree roots preferentially take up basic cations (e.g., Ca²⁺, Mg²⁺) and leave hydrogen (H⁺) and aluminium ions (Al³⁺) in the soil solution, which further contributes to pH reduction [54,55]. In contrast, cropland soils showed near-neutral pH values. Although nitrogen fertilisation, especially with ammonium-containing fertilisers, such as urea or ammonium sulfate, is known to acidify soils [56], regular liming and tillage can mitigate these effects. Farmers in the region occasionally apply lime (CaCO₃) to correct acidification, which may explain the current pH balance despite decades of synthetic fertilisation and the absence of organic amendments or cover crops. Grassland soils, on the other hand, showed an alkaline pH, which gives a more differentiated picture. One explanation could lie in the mulch residues on the soil surface. Decomposing mulch can increase the pH of the soil depending on its composition, especially if the residues come from species with a high content of base cations (e.g., Ca²⁺, K⁺, Mg²⁺). Studies have shown that organic mulches and crop residues can temporarily increase soil pH by releasing base cations and buffering hydrogen ions during decomposition [57,58]. Furthermore, the grassland soils in this study were not acidified by fertiliser inputs, which could explain their higher initial pH. Interestingly, a decline in pH in forest soils between 2021 and 2023 could be due to increased decomposition rates, increased acid leaching due to higher precipitation or root activity, or decreased buffering capacity due to changes in the quality of litter input. Such changes underline the dynamic nature of forest soil chemistry and the importance of long-term monitoring.

Trace element concentrations (e.g., Cr, Ni, Cu, Zn) were higher in arable and grassland soils than in forest soils. Although these patterns are still within permissible environmental limits, they probably reflect a combination of anthropogenic influences. The spatial arrangement of land-use types, in particular, supports this interpretation. Both croplands and grassland are located close to the farm’s access road and operational infrastructure, with the grassland only 20 m away from mechanisation zones, agrochemical storage areas and an intensively managed orchard (Figure 1). The arable land is also relatively close (within 100 m) to the farm and equipment zones. In contrast, the forest floor, which is located at a greater distance from anthropogenic sources, showed the lowest trace element levels. This spatial gradient suggests that localised sources of contamination, such as mechanical equipment, liquid spills, fertilizer-handling areas, and road dust, likely contribute to the elevated metal concentrations in grassland and cropland soils. Atmospheric deposition of particulate-bound metals (especially Zn and Cu from tyres and brake abrasion and Cr/Ni from fuel and machinery combustion) is a known phenomenon in soils adjacent to agricultural land [59,60]. In addition, the intensive and long-term use of phosphate-based fertilisers on croplands is another key factor, as these fertilisers often contain trace concentrations of heavy metals, such as Cd, Cr and Zn, as contaminants [61,62]. The proximity of grassland to orchards and farms could explain the comparatively high Cu concentrations. Cu-based fungicides are often used in fruit cultivation and can remain in the surrounding soils through drift or runoff [63]. These results underline the susceptibility of urban soils and soils adjacent to farms to diffuse contamination, even with low-use or near-natural land-use forms such as grassland. Therefore, they underline the need to regularly monitor trace elements on croplands and adjacent uncultivated land exposed to indirect agricultural activities.

4.2. Relation Among Soil Properties

The principal component analysis helped to clarify the underlying structure and relationships between the measured soil properties. Factor 1 showed a strong association between elevated soil pH and trace element concentrations (Cr, Mn, Fe, Ni, Cu, Zn) and a significant negative reduction in SOM and WSA. Although alkaline conditions generally reduce the solubility of trace metals, the co-occurrence observed here reflects common anthropogenic influences. Both grassland and cropland, especially near agrochemical storage areas, access roads, and intensively managed orchards, showed higher pH values due to liming or base-rich organic residues and also received long-term inputs of trace metals via fertilisers, pesticides, and atmospheric deposition [15,63]. At the same time, low SOM content has been associated with higher metal concentrations, supporting the idea that lower organic input and microbial activity limit the soil’s ability to complex and immobilise trace metals [64,65]. In soils with lower SOM, metals remain in more available or weakly bound forms, which increases their environmental risk and long-term persistence. This interplay of chemical enrichment and depletion of organic matter is typical of disturbed, management-intensive systems and underlines the importance of organic inputs for buffering contamination.

Water-stable aggregates and SOM were positively associated in Factor 1, reflecting their degradation under long-term anthropogenic pressure. As an important binder, SOM promotes the formation and stability of aggregates, while the loss of SOM weakens the soil structure and makes aggregates more susceptible to degradation [48].

Factor 2, which explained 13.70% of the total variance, showed strong positive loadings for WHC and soil temperature and a strong negative loading for BD. This pattern reflects a dimension primarily related to soil porosity and its ability to retain moisture, which is strongly influenced by vegetation cover and SOM. High WHC and low BD are characteristics of soils with well-developed structure, low compaction, and sufficient organic input. In this study, these conditions will likely exist in grasslands managed with mulch residues and in forested areas where minimal disturbance and continuous litter inputs contribute to better soil aggregation and porosity [48]. The negative relationship between BD and WHC is well known; as BD decreases, soil pore space increases, improving the soil’s ability to store available water [66,67]. Soils with lower BD also provide a more favourable environment for root and microbial activity, further supporting hydrological function. The positive influence of soil temperature on this factor may reflect the insulating effect of vegetation cover and surface residues that buffer diurnal and seasonal fluctuations. While soils with lower BD and higher porosity may exhibit more rapid surface heating due to lower thermal mass, especially under dry conditions, the presence of SOM and vegetation can mitigate these effects by stabilising moisture and shading the soil. Therefore, with perennial vegetation and mulch management, such soils maintain a more stable microclimate [68]. In addition, the presence of mulch and vegetation can reduce evaporative losses, contributing to a more favourable heat and moisture regime, especially in grasslands. This component represents an axis of soil physical quality, where improved structure (low BD), improved moisture retention (high WHC), and moderate thermal dynamics (soil temperature) together reflect better soil function under perennial vegetation.

Factor 4, accounting for 9.89% of the total variance, was characterised by positive loadings for CO₂ emissions and negative loadings for SWC, indicating a trade-off between soil respiration and SWC availability. At first glance, this negative correlation may seem counterintuitive, as soil respiration is generally enhanced by higher moisture, which promotes microbial activity and substrate diffusion. However, this relationship is non-linear, temperature dependent, and even more complex under field conditions. At moderate soil temperatures, such as those observed during sampling in 2023, higher SWC generally promotes CO₂ emissions by supporting microbial growth, enzymatic activity and decomposition of SOM [69]. However, when soils become excessively wet, oxygen diffusion is limited and microbial respiration shifts to anaerobic pathways, which are less efficient and result in lower CO₂ emissions [70,71]. In our study, SWC was relatively high in 2023 for all land-use types, which may have suppressed aerobic microbial activity in some plots, especially in dense vegetation or compacted conditions. Conversely, higher CO₂ emissions were recorded in 2021, when soils were drier and possibly better aerated, especially in grasslands. This likely reflects increased microbial turnover in well-drained surface layers and/or increased respiration following rewetting events, commonly called the “Birch effect” [72]. Drying and rewetting cycles stimulate microbial degradation of previously protected SOM, leading to short-term pulses of CO₂ [73,74]. This could explain the decoupling of SWC and CO₂ emissions observed in the PCA. In addition, soil temperature moderates the SWC–CO₂ relationship. Studies have shown that the effect of SWC on microbial respiration decreases at low or high-temperature extremes [75]. Under the current field conditions, where soil temperatures were in a favourable range, SWC became the dominant controlling factor, but with the caveat that very high SWC may have dampened microbial respiration due to suboptimal aeration, especially in finer textured or compacted soils. Thus, factor 4 likely represents a seasonal axis of soil biological activity, where the interplay of SWC, temperature and aeration determines microbial CO₂ emissions across all land uses.

5. Conclusions

This study demonstrates that land use significantly governs soil physical, chemical, and biological properties in periurban environments, with grasslands emerging as a high-functioning and resilient land-use type. Compared to croplands, grasslands had lower bulk density, higher water retention, organic matter content, and aggregate stability, revealing the benefits of permanent vegetation cover, minimal disturbance, and organic residue cycling. Temporal trends highlighted the ability of grassland soils to enhance water regulation and reduce compaction under favourable climatic conditions. However, observed shifts in soil structure and carbon content over time also point to the dynamic nature of grassland systems and the need for active monitoring to sustain their multifunctionality. Insights from Principal Component Analysis reinforced the central role of grasslands in maintaining soil porosity, biological activity, and structural integrity, especially when compared to degraded croplands. Grasslands show a balance between microbial CO₂ activity and water conservation, a combination critical for climate adaptation, carbon cycling, and flood mitigation in periurban zones. These findings directly contribute to understanding the multifunctionality of grassland soils. They show that well-managed grasslands deliver agronomic and ecological benefits and provide nature-based solutions for urban resilience. Promoting mulching, permanent cover, and low-intensity use can maximise the delivery of soil-based ecosystem services across grassland landscapes. Future research should prioritise long-term monitoring of grassland soil functions, especially under changing land-use intensity and climate variability. Moreover, stronger integration of grassland soil value into urban planning, agri-environmental policy, and green infrastructure investment is essential to secure their multifunctional benefits for decades.