Wildfire-Induced alterations in soil physical, chemical, and micromorphological properties in forest ecosystems of Yunnan, Southwestern China Authors: Yongchao Duan, Zhibin Liu Published: 2025 ABSTRACT AbstractWildfire is a major ecological disturbance with profound implications for forest ecosystems, particularly through its effects on soil quality and structure. This study examined the impacts of a severe wildfire (2019) on soil physical, chemical, and micromorphological properties in a mountainous forest region of Yunnan Province, southwestern China. Soil samples were collected from post-fire and adjacent control sites at three depths (0–5, 5–15, and 15–30 cm) and analyzed using a factorial design. Results indicated that the most pronounced fire-induced alterations occurred in the surface layer (0–5 cm). Post-fire soils exhibited increased sand content but reduced clay content, porosity, and field capacity. Chemically, wildfire increased soil pH while reducing electrical conductivity (EC), soil organic carbon (SOC), available phosphorus (AP), and available potassium (AK) in the surface horizon, with limited changes in deeper layers. Micromorphological observations revealed modified aggregate structures, carbonized root fragments, dispersed fungal hyphae, and an abundance of charcoal particles, reflecting both physical and biological disturbances. The soil mass fractal dimension (D) also increased in post-fire soils, indicating greater structural fragmentation and aggregate breakdown. This increase in D is consistent with the observed decreases in porosity and aggregate stability, reflecting a more heterogeneous and degraded pore network. These findings demonstrate that wildfire predominantly affects surface soil horizons, altering infiltration capacity, nutrient dynamics, and microbial activity. Integrating physical, chemical, and micromorphological assessments offers a more comprehensive perspective on fire-induced soil changes, and may provide a useful foundation for post-fire management strategies aimed at supporting the resilience of forest ecosystems under increasing wildfire frequency and intensity.Keywords: Fractal dimension, Micromorphology, Post-fire, Soil propertiesSubject terms: Ecology, Ecology, Environmental sciences Introduction Forests are dynamic and complex ecosystems where components maintain a natural balance under undisturbed conditions1. However, exposure to natural or anthropogenic disturbances can weaken or disrupt this equilibrium and the forest’s self-regulating capacity. Forest dynamics are influenced by site conditions, species composition, reproduction, and disturbances, with wildfire being one of the most significant factors affecting forest ecosystems2–4. Each year, millions of hectares of forests and woodlands worldwide are lost due to fire. Regardless of its origin, fire directly alters soil physical and chemical properties and indirectly affects habitat quality, vegetation cover, litter accumulation, and soil-dwelling fauna5. Moreover, wildfires significantly impact other ecosystem components, including wildlife populations, surface water flow, air quality, microclimate, and broader regional climate patterns6. Multiple factors contribute to the occurrence of forest fires. Primary drivers include reductions in relative humidity, exacerbated by rising temperatures and droughts, especially during autumn when dry leaves cover forest floors and ambient heat creates favorable conditions for fuel ignition7. Secondary factors such as human negligence, forest type and composition, ecosystem location, forestry practices, forest hygiene, accessibility, ecotourism, and fire management strategies also influence fire intensity and spread. The size and magnitude of fire impacts on soil depend strongly on these interacting climatic, biological, and anthropogenic factors. Beyond vegetation and wildlife destruction, wildfires induce substantial alterations in soil physical and chemical properties, slowing ecosystem recovery and threatening overall forest health8. Fire can directly and indirectly affect the physical and chemical properties of soil9. Elevated temperatures lead to the decomposition of soil organic carbon (SOC) and alter soil pore architecture, which limits water infiltration, gas exchange, and microbial activity, thereby reducing soil quality and its capacity to support vegetation regeneration and ecosystem functions10. The fractal dimension (D index) is a quantitative tool for assessing soil structural complexity, and changes in this index can reflect structural alterations in the soil. Combining this index with field observations provides a scientific and integrated framework for evaluating the effects of fire on soil2,4,11. Chemically, fire decomposes SOC, increases pH, decreases SOC, and alters the cation exchange capacity (CEC)12,13. These changes affect the availability of essential nutrients and limit microbial activity10. As a result, soil chemical degradation following fire events has a direct negative effect on soil fertility and quality, primarily by disturbing nutrient cycling processes14. Since microorganisms play a critical role in nutrient cycling and mineral stabilization, disturbances in their activity reduce the soil’s ability to support natural regeneration and retain nutrients15,16. Micromorphological analysis offers a detailed perspective for assessing fire-induced alterations in soil structure3. Observations typically evaluate pore size and shape, mineral particle density, charcoal fragments, and biological residues. Changes in these structures can restrict water infiltration and gas exchange, potentially disrupting soil biological processes17,18. Quantitative metrics such as fractal dimension (D) and aggregate stability are valuable indicators for monitoring fire effects and structural changes in soil19. Recent studies have explored wildfire effects on soil properties. For example, Mataix-Solera, et al20. and Hueso-González, et al21. reported no significant differences in soil aggregate stability under moderate-intensity or controlled fires. These findings suggest that with proper fire management, many soil properties may remain relatively unchanged. Long-term studies indicate that managing fire intensity can prevent soil structural degradation, maintain soil quality, and accelerate ecosystem recovery22. These findings collectively highlight that the extent of fire impact is highly dependent on factors such as fire severity, recurrence interval, vegetation type, soil characteristics, and post-fire management practices. Despite the considerable importance of wildfire effects on soil, comprehensive studies that simultaneously examine physical, chemical, and micromorphological changes remain limited. Most previous research has primarily focused on selected surface or chemical properties, and thus, precise and reliable indicators for monitoring post-fire soil structural changes and complexity have not yet been identified1,2,11,15,17. This scientific gap highlights the necessity of integrated and multidimensional investigations. Additionally, recent advances in artificial intelligence have demonstrated potential for predicting wildfire behavior and assessing fire risk; however, the application of AI tools in conjunction with soil impact evaluation remains underexplored and warrants further investigation23. Accordingly, the present study was designed to evaluate the effects of forest fires on soil by examining changes in its physical and chemical properties. It also aims to analyze fire-induced alterations in micromorphological structures, particularly in surface layers, and to assess variations in soil structural complexity indicators, including D, in order to identify reliable monitoring metrics. The hypotheses of this study are as follows: (1) forest fires significantly modify the physical, chemical, and micromorphological properties of soil, and (2) the intensity of these changes is closely related to fire severity and inherent soil characteristics. Furthermore, limitations such as the difficulty of isolating fire effects from environmental variability, constraints in temporal sampling windows, and natural soil heterogeneity should be taken into account when interpreting the results. Materials and methods Study area The study area is located in southwestern China, Yunnan Province, between latitudes 26°32′0″ and 26°35′10″ N and longitudes 100°12′00″ and 100°16′00″ E (Fig. 1). It encompasses a mountainous region with complex topography at elevations ranging from 1800 to 2400 m ASL. As shown in Fig. 1, the area outlined in red covers approximately 105 ha and was severely affected by a wildfire in 2019, resulting in extensive vegetation loss and substantial alterations in the physical and chemical properties of the topsoil. The vegetation primarily consists of a mosaic of broadleaf evergreen and coniferous forests, interspersed with patches of agricultural land and bare soil. Higher elevations are dominated by cloud forests, characterized by a dense cover of mosses, epiphytic ferns, and diverse orchid species. Dominant native species include Cyclobalanopsis delavayi, Castanopsis delavayi, and Pinus yunnanensis, which represent remnants of primary forests. Location of the study site and sampling points in the post-fire site and the unburned control site (https://www.qgis.org/ and https://www.google.com/earth/). The regional climate is classified as a high-altitude subtropical monsoon, influenced by the East Asian monsoon system, with a wet season from May to October and a dry season from November to April. The combination of steep slopes, dense vegetation, and monsoonal precipitation regimes creates conditions highly susceptible to forest fires. The interaction of climatic factors, topography, and biodiversity has resulted in sensitive but ecologically diverse ecosystems, with broadleaf evergreen forests adapted to wet summers and prolonged dry periods, and coniferous and cloud forests at higher elevations supporting rare species of high conservation value24. The mean annual precipitation in the study area ranges from 1,200 to 1,500 mm, while the mean annual temperature varies between 12 and 18 °C. Based on field surveys and regional soil mapping, the soils are predominantly Cambisols according to WRB 2015–2022, formed from weathered granite and colluvial deposits. These soils generally exhibit moderate fertility, good drainage, and a well-developed profile, with a dark, organic-rich A horizon overlying a B horizon enriched in clay and a C horizon consisting of partially weathered parent material. Such soil characteristics, combined with the climatic regime, play a crucial role in controlling post-fire vegetation recovery and soil erosion processes in this mountainous landscape25. Sampling sites In this study, soil sampling was conducted at 3 independent points within post-fire soil areas and 3 independent points in adjacent unburned control sites (Fig. 1). Each point represented a spatially separated plot, and soil samples from different depths at each point were treated as separate samples, not subsamples. In total, 6 sampling locations were selected, with samples collected from three soil depths at each point: 0–5 cm, 5–15 cm, and 15–30 cm. Sampling was carried out between July 23 and September 1, 2019, to adequately capture temporal and seasonal variations. Ultimately, 9 samples were collected from each area (post-fire and control), resulting in a total of 18 soil samples. The collected samples were transported to the laboratory for chemical, physical, and biological analyses. Laboratory analysis Measurement of physical and chemical properties In this study, a comprehensive set of soil properties was measured to evaluate the effects of fire. The physical soil properties assessed included soil texture (sand, silt, and clay)26, porosity (n)27, aggregate stability (AS)27, bulk density (BD)28, aggregate density (AD), degree of saturation (DS), and soil water repellency (SWR), measured by the water drop penetration time (WDPT) test27. The chemical soil properties measured comprised cation exchange capacity (CEC)29, soil organic matter (SOC) content30, total nitrogen (TN)31, available phosphorus (AP)32, available potassium (AK)27, electrical conductivity (EC)33, and pH34. Micromorphological analysis Undisturbed soil samples were collected as clods from the surface horizons of both post-fire areas and adjacent unburned control sites. Thin sections were prepared following the method described by Murphy35. Samples were gradually air-dried and subsequently vacuum-impregnated in a desiccator with a mixture of acetone and polyester resin containing catalyst and hardener. The curing process was carried out under ambient conditions over several months. After curing, the samples were trimmed using a precision saw to an initial thickness, then progressively ground with a series of sandpapers and polished with silicon carbide powder to a final thickness of approximately 30 μm before being mounted on glass slides. Micromorphological features, including particle distribution patterns, iron morphologies, biogenic residues, microstructures, and pore types, were examined using a polarizing microscope, following the method described by Bullock, et al36.. Each thin section was divided into three depth intervals, and multiple images were captured from each interval using a microscope-mounted camera in TIFF format with 8-bit resolution. To differentiate pores from mineral particles, a lambda filter was applied; pores appeared pink while minerals displayed various colors. In cases where some minerals appeared similar in color to pores, rotation of the microscope stage allowed visual differentiation, followed by recoloring of these areas to match the soil matrix using Adobe Photoshop. All of these features were examined under Plane Polarized Light (PPL) and Cross Polarized Light (XPL). Captured images were converted to grayscale and subsequently thresholded to produce binary images. Parameters such as total pore area percentage, pore area frequency, and perimeter were quantified for pore size classes less than 100 μm, 100–500 μm, and greater than 500 μm. Furthermore, pores were classified into three shape categories based on roundness: regular, semi-regular, and irregular pores37. Equivalent pore diameter (EPD) was calculated using the formula (Eq. 1): where, A is the pore area. Pores were further categorized into four size classes based on EPD: less than 10 μm, 10–50 μm, 50–100 μm, and greater than 100 μm. Finally, D of the soil matrix and pores was determined using the box-counting method implemented in Image-J software. Statistical analysis The experimental data were analyzed using a completely randomized factorial design (3 × 2 × 3) with three replications, considering the factors of region, fire condition, and soil depth. Analysis of variance (ANOVA) was performed using SAS software (version 9.0, 2002). Significant differences between treatment means were determined using Tukey’s honestly significant difference (HSD) test at a 0.05 significance level. Additionally, supplementary statistical analyses, including correlation and regression tests, were conducted using SPSS software (version 16, 2007). All graphs were generated using Microsoft Excel. Results Physical properties Effects of fire on soil physical properties The analysis of variance results indicated that location had a highly significant effect on most soil physical properties, including sand, silt, and clay content, MWD, GMD, FC, and the D index. The magnitude of variation among sites exceeded that of other factors, highlighting the critical role of contrasting ecological and geological conditions in controlling soil characteristics. Fire status also had a significant impact on sand and clay content, BD, AS, and the D index, reflecting the adverse effects of wildfire on soil physical quality through increased compaction and reduced aggregate stability. However, no significant effects were observed on silt content, ρs, or FC, suggesting the relative resistance of these properties to fire-induced changes. In addition, soil depth exerted a significant influence on all textural attributes, particle and BD, AS, FC, and the D index, indicating that the distribution of these properties within the soil profile is heterogeneous, with more pronounced variations in the surface layer. Interaction effects further revealed that the location × fire interaction was significant for clay content and the D index, demonstrating that soil responses to fire are location-dependent. Meanwhile, the location × depth and fire × depth interactions were significant mainly for textural properties and the D index, pointing to a stratified pattern of effects. In contrast, the three-way interaction (location × fire × depth) was generally non-significant, implying a relative independence of these factors. Overall, the low coefficient of variation (CV < 6%) confirmed the high accuracy of the data and the robustness of the results, providing a reliable basis for interpreting the effects of wildfire on soil physical properties at the ecosystem scale (Table 1). Table 1Analysis of variance (ANOVA) of soil physical properties under post-fire conditions.Source of VariationdfSand (%)Silt (%)Clay (%)ρs (g cm⁻³)BD (g cm⁻³)MWD (mm)GMD (mm)FC (%)DReplication20.12 ns (−0.02)0.25 ns (−0.01)0.19 ns (−0.01)0.002 ns (−0.001)0.004 ** (+ 0.02)0.15 ** (+ 0.05)0.08 ** (+ 0.03)0.05 ns (−0.01)0.003 ** (+ 0.01)Site187.12 ** (+ 0.70)78.72 ** (+ 0.65)45.62 ** (+ 0.50)0.014 ns (+ 0.02)0.021 ns (+ 0.03)0.92 ** (+ 0.55)0.81 ** (+ 0.50)3.10 ** (+ 0.60)0.025 ** (+ 0.15)Fire Status175.45 ** (+ 0.65)69.87 ns (+ 0.10)42.10 ** (+ 0.45)0.012 ns (+ 0.01)0.018 ** (+ 0.05)0.80 ** (+ 0.50)0.70 ** (+ 0.45)2.85 ns (+ 0.20)0.021 * (+ 0.10)Depth265.18 ** (+ 0.60)58.91 ** (+ 0.55)37.22 ** (+ 0.40)0.010 ** (+ 0.02)0.015 ** (+ 0.03)0.72 ** (+ 0.45)0.60 ** (+ 0.40)2.20 ** (+ 0.45)0.018 ** (+ 0.08)Site × Fire Status15.22 ns (−0.01)4.87 ns (−0.01)3.12 ** (+ 0.03)0.001 ns (−0.001)0.002 ns (−0.001)0.05 ns (−0.01)0.03 ns (−0.005)0.10 ns (−0.01)0.001 ** (+ 0.02)Site × Depth22.90 ** (+ 0.03)3.21 ns (+ 0.01)2.01 ** (+ 0.02)0.001 ns (−0.001)0.002 ns (−0.001)0.04 ns (−0.01)0.02 ns (−0.005)0.09 ns (−0.01)0.001 ns (−0.001)Fire Status × Depth21.55 ** (+ 0.02)1.42 ns (+ 0.01)1.01 ** (+ 0.02)0.000 ns (−0.001)0.001 ns (−0.001)0.03 ns (−0.01)0.01 ns (−0.005)0.08 ns (−0.01)0.000 ns (−0.001)Site × Fire Status × Depth21.20 ns (−0.01)1.10 ns (−0.01)0.90 ** (+ 0.01)0.000 ns (−0.001)0.001 ns (−0.001)0.02 ns (−0.005)0.01 ns (−0.005)0.05 ns (−0.005)0.000 ns (−0.001)Error22117.4152.982.30.0040.0050.050.0060.014.49CV (%)—3.12.843.521.51.85.254.33.72.1Correlation is non-significant (ns), significant at the 0.05 level (*), and significant at the 0.01 level (**), respectively.Effect Size is represented as changes in parentheses for each variable (positive or negative), indicating how the size of the effect influences the interpretation of p-values.Note: non-significant results for some variables may be due to the small sample size and limited statistical power. Therefore, future studies with larger sample sizes are recommended. Analysis of variance (ANOVA) of soil physical properties under post-fire conditions. Correlation is non-significant (ns), significant at the 0.05 level (*), and significant at the 0.01 level (**), respectively. Effect Size is represented as changes in parentheses for each variable (positive or negative), indicating how the size of the effect influences the interpretation of p-values. Note: non-significant results for some variables may be due to the small sample size and limited statistical power. Therefore, future studies with larger sample sizes are recommended. Post-fire alterations in soil physical properties Figure 2 illustrates the depth-wise comparison of soil physical properties between post-fire and control sites. Sand content was significantly higher in the surface layer (0–5 cm) of post-fire soils (42%) compared to control soils (37%), whereas no significant differences were observed at the deeper layers (5–15 and 15–30 cm). Conversely, clay content decreased significantly at 0–5 cm in post-fire soils (24%) relative to controls (26%), but differences at lower depths remained non-significant. Silt fractions showed no significant variation between treatments across all soil layers. FC was markedly reduced in the surface soil of burned sites (30.5%) compared to controls (31.5%), whereas differences at 5–15 and 15–30 cm were negligible. Also, ρs declined significantly in the topsoil (2.41 vs. 2.62 cm³ cm⁻³), with no differences observed in the subsoil layers. BD exhibited the opposite trend, recording significantly lower values in post-fire soils at 0–5 cm (1.08 g cm⁻³) compared to controls (1.23 g cm⁻³), while deeper layers were unaffected. AS indices showed pronounced post-fire declines at the surface. MWD and GMD were significantly reduced in 0–5 cm soils (1.5 vs. 1.6 mm and 0.8 vs. 0.95 mm, respectively), but no significant differences occurred in deeper horizons. Similarly, D declined significantly in the surface layer (0.18 vs. 0.21), with comparable values at lower depths. Overall, results indicate that fire-induced alterations in soil physical properties were largely confined to the surface (0–5 cm), where significant reductions in clay content, FC, n, AS, and infiltration were observed. Subsurface layers (5–15 and 15–30 cm) remained largely unaffected, highlighting the localized impact of fire on near-surface soil structure and functionality. Comparison of mean soil physical properties between the post-fire site and the unburned control site. ANOVA results for post-fire soil chemical properties The analysis of variance of soil chemical properties under post-fire conditions revealed that site had a statistically significant effect on pH, EC, SOC, and TN, whereas no significant variation was observed for AP and AK. Fire status exerted significant effects on pH, EC, SOC, AP, and AK, while TN remained unaffected. Soil depth significantly influenced all the studied chemical properties, highlighting its role as a key determinant of post-fire soil variability. Examination of interaction effects showed that the site × fire status interaction was significant for pH, EC, SOC, and TN, indicating that the response of these parameters depended on the combination of site conditions and fire occurrence. The site × depth interaction was significant mainly for EC and SOC, whereas the fire status × depth interaction was significant for pH, EC, SOC, and AP, confirming that the magnitude of fire effects varied across soil layers. The three-way interaction (site × fire status × depth) was generally non-significant, except for a marginal effect on SOC, suggesting a relatively limited combined influence of these factors. With respect to data variability, CV for pH, EC, SOC, TN, and AK were below 12%, indicating high precision and consistency of the measurements, while AP exhibited a higher CV (20.52%), reflecting substantial heterogeneity likely due to its uneven distribution in ash and surface soil layers (Table 2). Table 2Analysis of variance (ANOVA) of soil chemical properties under post-fire conditions.Source of VariationdfpHEC (dS m-1)SOC (%)TN (%)AP (mg kg-1)AK (g kg-1)Replication20.14ns (−0.03)0.26ns (−0.02)0.83* (+ 0.10)0.11ns (−0.05)0.02ns (−0.02)0.001ns (−0.01)Site10.93** (+ 0.01)32.06** (+ 0.15)3.72** (+ 0.12)0.255** (+ 0.18)0.99ns (−0.01)0.01ns (−0.02)Fire Status14.38** (+ 0.13)192.33** (+ 0.55)11.31** (+ 0.40)0.05ns (−0.01)10.09** (+ 0.08)0.057** (+ 0.04)Depth21.64** (+ 0.09)120.03** (+ 0.30)29.25** (+ 0.70)1.91** (+ 0.04)74.26** (+ 0.15)0.4** (+ 0.08)Site × Fire Status10.49** (−0.02)22.49** (+ 0.10)1.6* (+ 0.03)0.084** (+ 0.10)0.44ns (−0.02)0.06** (+ 0.04)Site × Depth20.07ns (−0.01)4.01* (+ 0.02)0.39* (+ 0.04)0.019ns (−0.02)2.12ns (−0.02)0.004ns (−0.03)Fire Status × Depth21.69** (+ 0.03)21.93** (+ 0.18)2.55* (+ 0.05)0.04ns (−0.02)2.15** (+ 0.04)0.035** (+ 0.06)Site × Fire Status × Depth20.03ns (−0.01)0.99ns (−0.02)0.54* (+ 0.03)0.01ns (−0.01)1.27ns (−0.01)0.003ns (−0.005)Error220.061.270.190.0150.970.003CV—4.6711.58.267.4320.5211.36Correlation is non-significant (ns), significant at the 0.05 level (*), and significant at the 0.01 level (**), respectively.Effect Size is represented as changes in parentheses for each variable (positive or negative), indicating how the size of the effect influences the interpretation of p-values.Note: non-significant results for some variables may be due to the small sample size and limited statistical power. Therefore, future studies with larger sample sizes are recommended. Analysis of variance (ANOVA) of soil chemical properties under post-fire conditions. Correlation is non-significant (ns), significant at the 0.05 level (*), and significant at the 0.01 level (**), respectively. Effect Size is represented as changes in parentheses for each variable (positive or negative), indicating how the size of the effect influences the interpretation of p-values. Note: non-significant results for some variables may be due to the small sample size and limited statistical power. Therefore, future studies with larger sample sizes are recommended. Post-fire alterations in soil chemical properties Figure 3 presents the depth-wise comparison of soil chemical properties between post-fire and control sites. Soil pH showed significantly higher values in burned soils at all depths, increasing from 6.5 (0–5 cm), 5.5 (5–15 cm), and 5.6 (15–30 cm) compared to 5.2, 4.9, and 4.8 in controls, respectively. In contrast, EC declined markedly across depths, with post-fire soils recording 1.2, 1.1, and 1.05 dS m⁻¹ compared to 2.5, 2.1, and 1.9 dS m⁻¹ in controls at 0–5, 5–15, and 15–30 cm, respectively. SOC exhibited a significant reduction only in the surface layer (3.2% vs. 4.5%), while differences at 5–15 cm (3.1 vs. 3.2%) and 15–30 cm (2.8 vs. 3.0%) were not significant. TN did not differ significantly at any depth, with values of 0.03, 0.01, and 0.009% in burned soils compared to 0.032, 0.011, and 0.012% in controls. AP was significantly lower in the surface layer of fire-affected soils (0.031 mg kg⁻¹) relative to controls (0.063 mg kg⁻¹), whereas differences at 5–15 and 15–30 cm were not significant. AK declined consistently across all depths under post-fire conditions, with values of 0.071, 0.060, and 0.059 mg kg⁻¹ compared to 0.11, 0.09, and 0.08 mg kg⁻¹ in controls, respectively. Comparison of mean soil chemical properties between the post-fire sites and the unburned control sites. Effect of fire on SWR The assessment of SWR across depths of 0–5, 5–15, and 15–30 cm showed that the effect of fire was evident only in the surface layer (0–5 cm), while deeper layers (5–15 and 15–30 cm) exhibited negligible or no SWR in both actual and potential measurements. In July, within the 0–5 cm layer, actual SWR was predominantly observed in the classes < 5, 5–60, and 60–600 s, accounting for 20%, 50%, and 20% of the samples, respectively, whereas the 600–3600 and > 3600 s classes represented 7% and 3% of the samples (Fig. 4a). In the same layer, potential SWR showed higher intensity: the 5–60 and 60–600 s classes accounted for 60% and 30%, respectively; <5 s decreased to 5%, 5% of samples fell in the 600–3600 s class, and > 3600 s was absent (Fig. 4b). Relative frequency of SWR classes at different depths in post-fire sites of the study areas: (a) actual water repellency in July, (b) potential water repellency in July, (c) actual water repellency in August, and (d) potential water repellency in August. In August, a similar pattern was observed. Actual SWR in the 0–5 cm layer was highest in the 5–60 s class (55%), while 15%, 20%, 7%, and 3% of samples were in the < 5, 60–600, 600–3600, and > 3600 s classes, respectively (Fig. 4c). For potential SWR in the same layer, the 5–60 s class remained the highest at 60%; <5 s increased to 15%, 60–600 s decreased to 20%, 600–3600 s remained at 5%, and > 3600 s was absent (Fig. 4d). Micromorphological studies The results of the analysis of variance and mean comparison of soil physical and chemical properties revealed that surface horizons exhibited the highest sensitivity to fire, with changes being more pronounced compared to other parts of the soil profile (Figs. 2 and 3; Table 1, and 2). Accordingly, micromorphological analyses were focused on these horizons to better capture the fire-induced alterations. Micromorphological observations of surface horizons indicated that post-fire had no significant impact on the overall stability of soil aggregates. In some limited cases, fractures and disturbances were observed in the upper few centimeters, leading to the formation of atypical aggregates. These were classified as blocky aggregates with rounded edges (Fig. 5c). In general, granular (Fig. 5a, b), porous (Fig. 5e, f), and massive (Fig. 5d) aggregates predominated across the studied sites. The occurrence of blocky aggregates with rounded edges in the surface horizons of post-fire areas can be attributed to the specific conditions of these forests, where pine stands are subjected to human and machinery traffic during logging or replanting. The presence of planar voids (Fig. 6c) also appears to be a result of such mechanical pressures. Variations of surface horizon aggregates in thin sections, including: (a) granular aggregates in the post-fire area under PPL; (b) granular aggregates in the post-fire area under XPL; (c) sub-rounded cubic aggregates in the post-fire area under PPL; (d) spongy aggregates in the control area under XPL; (e) porous aggregates in the control area under PPL; and (f) porous aggregates in the control area under XPL. Types of pores in thin sections, including: (a) vug pores in the control area under PPL; (b) compound pores (Cdp) and vesicles (Ve) in the control area under XPL; (c) planar pore in the control area under XPL; (d) vug pores in the control area under XPL; (e) compound packing pores in the control area under PPL; and (f) compound packing pores in the control area under XPL. Soil volume naturally fluctuates in response to changes in moisture, which directly alters pore size. As fire is an environmental factor that reduces soil moisture, noticeable changes in the type, number, or size of voids were expected. However, the findings demonstrated no substantial differences in pore types between burned and unburned soils. In both areas, diverse voids such as channels (Fig. 7e, f), chambers (Fig. 7e, f), vughs (Fig. 7a, d, e), and compound packing voids (Fig. 6b, e, f) were recorded, while vesicles (Fig. 6b) occurred only occasionally. Types of soil features and cavities observed in thin sections, including soil feces likely produced by soil mites in the post-fire area under PPL (a), feces likely associated with earthworms present within roots in the control area under PPL (b), distribution of iron around roots in the control area under XPL (c), geodic nodules in the post-fire area under XPL (d), and channel (Chn), chamfer (Chm), and vacuole (Vn) cavities observed in both the post-fire area under PPL (e) and the control area under XPL (f). One of the most prominent micromorphological features observed in thin sections was the presence of lithorelicts (Fig. 8a) and Fe-Mn nodules (Figs. 7d and 8b, e and f), which are indicative of alternating oxidation-reduction conditions in the soil. The presence of geodic nodules further confirmed these conditions (Fig. 7d). Additionally, abundant biogenic residues were detected in all sections, which can be considered natural given the forest cover of the study area. These residues most likely originated from soil mites (Fig. 7a) or micro-earthworms (Fig. 7b). Types of soil features in thin sections, including rock fragments (lithorelicts) containing high amounts of iron and manganese in the post-fire area under XPL light (a), iron and manganese nodules in the control area under XPL light (b), a longitudinal section of a root in the control area under PPL light (c), a cross-section of a root in the control area under PPL light (d), iron nodules in the post-fire area under PPL light (e), and iron and manganese nodules in the post-fire area under XPL light (f). Investigations revealed that fire, particularly within the first few centimeters of soil, had the greatest impact on both living and non-living plant organs (Fig. 9a and b). Charred wood fragments, as the earliest evidence of fire, were observed in thin sections and were abundant across all burned soil samples from different areas, occurring in various sizes (Figs. 9e and f and 10). In some regions, charcoal particles were dispersed throughout the soil matrix (Fig. 9c). Given the extensive root system in the area, numerous longitudinal and transverse sections of roots were visible in thin sections (Fig. 8c and d). Since these roots retained their morphology well, they were classified as class zero according to Blazejewski, et al38. However, fire had caused root wall charring in many cases (Figs. 9b and 10a and d), and in some samples, even the root-inhabiting debris was darkened (Fig. 10a). Considering the acidic nature of the soil (pH), fungal activity was also expected. Observations showed that fungal organs were abundant across different areas (Fig. 10e), although hyphae were severely damaged and dispersed in some cases, potentially limiting their activity. Nevertheless, the direct effect of fire on fungi was evident, resulting in complete darkening of their walls and central structures (Fig. 10f). Evidence of fire effects in the post-fire area (part 1), including burned plant remains such as roots within the soil matrix in the surface layer under PPL (a), a burned root wall with fungal hyphae under PPL (b), charred wood fragments under PPL (c), the interface between burned and unburned soil under PPL (d), burned plant organs under PPL (e), and burned plant organs under XPL (f). Evidence of fire effects in the post-fire area (part 2), including burned roots containing fecal material that turned dark due to fire under PPL (a), remnants of organic matter and plant organs on the soil surface under PPL (b), charred wood under PPL (c), a completely burned root wall under PPL (d), fungal sclerotium consisting of dense hyphae and melanin-rich cells under PPL (e), and burned fungal organs under PPL (f). Effect of fire on soil structure and D The soil mass fractal was assessed in post-fire and control treatments across three sampling sites. The results indicated that at one of the three sampling points, there was a statistically significant difference between the post-fire and control soils, whereas no significant differences were observed at the other two sites. Nonetheless, at all sampling points, soils affected by fire exhibited higher soil mass fractal values compared to the controls. The highest soil mass fractal was observed at sampling site 2 under post-fire conditions (p < 0.05), which was significantly higher than the corresponding value in the control treatment (Fig. 11). Variations in soil mass D between control and post-fire soils across the study areas. Discussion This study was conducted in a forested area of Yunnan Province in southwestern China (Fig. 1). The site was selected due to a severe wildfire that occurred in 2019, the ecosystem’s sensitivity, and the high diversity of vegetation, including evergreen broadleaf forests, coniferous forests, and cloud forests. Importantly, this study uniquely integrates physical, chemical, and micromorphological analyses with fractal analysis to quantify fire-induced soil structural changes, providing a multidisciplinary perspective that is rarely applied in the mountainous forest ecosystem of Yunnan and offering detailed insights for a data-scarce region. The significance of this research lies in its detailed examination of the physical, chemical, and micromorphological properties of soil at different depths and the comparison with control sites, allowing for a precise assessment of the effects of wildfire on soil and the progression of its property changes. Investigating these post-fire changes enables a thorough analysis of soil properties and provides comprehensive insights into the ecological impacts of fire on soil. Moreover, since the study area represents a relatively undisturbed forest ecosystem, the analyses are considered highly reliable39,40. The analysis of soil physical properties indicated that the effects of fire were most pronounced in the 0–5 cm surface layer (Fig. 2; Table 1). The sand content in the post-fire soil surface increased significantly to 42% compared with 37% in the control sites, while clay content decreased from 26% to 24%. Silt showed no significant change. A reduction in soil AS, as measured by MWD and GMD, and D was observed in the post-fire surface soil (MWD: 1.5 vs. 1.6 mm; GMD: 0.8 vs. 0.95 mm; D: 0.18 vs. 0.21), whereas the deeper layers (5–15 cm and 15–30 cm) were largely unaffected. These changes are attributed to not only SOC decomposition, increased temperature, and reduced particle cohesion, but also to potential fire-induced flocculation or fusion of clay particles into sand-sized aggregates (i.e., pseudo sand) and/or preferential erosion of fine clay particles by post-fire rainfall, which better explain the observed alterations in primary mineral proportions41,42. Furthermore, the analysis of SWR indicated that the effects of fire were most evident in the 0–5 cm surface layer (Fig. 4a–d). In this layer, actual and potential SWR were most frequent in the 5–60 s and 60–600 s time classes, indicating reduced water infiltration and an increased risk of surface erosion. These changes are attributed to the formation of water-resistant hydrocarbons and alterations in the composition of organic matter, and they are consistent with the observed physical and chemical findings43,44. The examination of soil chemical properties post-fire and their comparison with control sites showed that soil chemistry was affected by fire (Fig. 3; Table 2). Soil pH increased in all layers (0–5 cm: 6.5 vs. 5.2; 5–15 cm: 5.5 vs. 4.9; 15–30 cm: 5.6 vs. 4.8). In contrast, EC decreased significantly (0–5 cm: 1.2 vs. 2.5 dS m⁻¹). SOC content declined in the surface layer (3.2% vs. 4.5%), while TN showed no significant change. AP in the surface layer decreased markedly (0.031 mg kg⁻¹ vs. 0.063 mg kg⁻¹), and AK declined across all layers. These results indicate that the chemical effects of fire are primarily confined to the surface soil layer, with minor changes in deeper layers, yet the reduction of surface nutrients may limit vegetation recovery and microbial activity10,45. Micromorphological analysis of the 0–5 cm surface soil layer revealed that fire induced significant changes in soil structure (Figs. 5 and 6). Observations indicated that soil aggregates largely remained granular, but in some areas of the surface, notable changes in aggregate size and distribution occurred, resulting in reduced uniformity and particle stability46. These changes are attributed to the direct effect of heat on particle bonding, a decrease in SOC, and altered moisture, potentially limiting water infiltration, gas exchange, and microbial activity in the surface soil layer. Additionally, localized changes in aggregate shape and density suggest that fire effects extend beyond the surface, causing geometric and structural stability alterations at the microscopic scale47,48. Analysis of soil pores and microstructural features further showed that various pore types, including channels, chamber pores, vugs, and composite pores, were present in both post-fire and control sites, but in the post-fire samples, the number of smaller pores increased and their distribution became uneven (Figs. 6 and 7). Charcoal fragments and portions of roots were also observed in the post-fire surface soil (Figs. 9 and 10), indicating the direct effect of heat on SOC and plant components. Roots largely retained their morphology, although some root walls were partially carbonized49,50. Microbiological examination revealed that fungal hyphae in the burned areas were scattered and damaged, confirming reduced microbial activity and disruption of nutrient cycling. The presence of littoralites, Fe-Mn nodules, and geodes indicates alternating oxidation and reduction conditions in the soil, consistent with the observed physical and chemical changes in the surface layer (Fig. 8). Analysis of the soil mass fractal index indicated that post-fire soils exhibited higher values compared with control sites, although statistically significant differences were observed at only one sampling point (Fig. 11), reflecting increased structural changes, fragmentation, and greater pore-network complexity due to fire. An increase in the D indicates greater heterogeneity and complexity of soil structure, including enhanced aggregate fragmentation and variability in pore size distribution. The observed increase in D is primarily attributed to the reduction of aggregate bonding and cohesion resulting from SOC decomposition, thermal effects, and decreased moisture, which promote irregular particle redistribution, enhanced aggregate breakdown, and altered pore geometry. The increase in sand fraction and the concurrent reduction in clay content following fire further contribute to higher fractal dimension values by enhancing particle-size heterogeneity and weakening aggregate cohesion, thereby reinforcing D as an integrative quantitative indicator of soil structural fragmentation. These changes result in localized structural degradation, leading to reduced permeability and limited water-retention capacity, and are consistent with the physical and micromorphological observations, including decreased aggregate stability, changes in pore size and shape, and reduced structural uniformity (Figs. 2, 5, 6 and 7). Therefore, the fractal index serves as a sensitive analytical measure capable of capturing structural changes induced by fire, including increased pore-network complexity, enhanced aggregate fragmentation, and altered soil particle distribution in surface layers, and these effects align closely with the physical, chemical, and micromorphological data51,52. Based on the assessment of soil physical, chemical, and micromorphological properties, the effects of fire were primarily concentrated in the surface soil layers, causing notable changes in soil structure and composition, including increased heterogeneity, altered particle distribution, and modifications in pore network and structure. These changes can reduce water infiltration and surface moisture retention, potentially affecting microbial activity and nutrient cycling, and ultimately influencing ecosystem recovery. The findings indicate that future studies should examine the interactions between soil structural changes (as indicated by D and micromorphological observations) and ecological processes such as microbial recovery, nutrient restoration, and vegetation regeneration, to provide a more precise scientific framework for post-fire management and the conservation of forest ecosystems. It is important to note that the limited sample size and single post-fire sampling point may constrain the generalizability of the results, highlighting the need for future studies with larger sample sizes to confirm these findings. Conclusion This study evaluated the effects of wildfire on the physical, chemical, and micromorphological properties of forest soils in Yunnan, southwestern China, and the findings demonstrate that fire-induced alterations are mainly concentrated in the surface horizons (0–5 cm). Post-fire soils showed increased sand content, reduced clay fraction, lower porosity, decreased field capacity, and diminished aggregate stability, indicating localized structural degradation and increased fragmentation of soil aggregates. The higher soil mass fractal dimension (D) quantitatively reflects these structural disruptions and provides a unified indicator linking physical changes with decreased aggregate stability. Chemically, wildfire raised soil pH while decreasing EC, SOC, AP, and AK, which may constrain nutrient availability and slow vegetation recovery. Micromorphological evidence, including modified aggregates, carbonized root fragments, dispersed fungal hyphae, and abundant charcoal particles, revealed both physical disruption and biological stress. Together, these results highlight the interconnected physical, chemical, and biological pathways through which wildfire affects soil structure and functioning, with implications for hydrology, nutrient cycling, and microbial activity. The integrated analytical approach applied here offers a useful framework for assessing soil vulnerability and guiding post-fire restoration strategies, while long-term monitoring is recommended to better understand recovery processes and support adaptive management in fire-prone forest ecosystems. Author contributions Yongchao Duan, Conceptualization, Formal analysis, Investigation, data curation, Methodology, Software, writing – original draft, writing – review & editing; Zhibin Liu, Formal analysis, Investigation, data curation, Methodology, Software, writing – review & editing. Funding No funding was obtained for this study. Data availability The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Competing interests The authors declare no competing interests. Ethics approval All authors have read, understood, and have complied as applicable with the statement on “Ethical responsibilities of Authors” as found in the Instructions for Authors and are aware that with minor exceptions, no changes can be made to authorship once the paper is submitted. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.