The study was conducted in a managed pine forest in Ahuacapán Ejido (19° 36′–19° 41′ N and 104° 16′–104° 21′ W) of Autlán de Navarro, Jalisco, Mexico (Figure 1). This area is located within the upper zone of the Sierra de Manantlán Biosphere Reserve, at 1,900 m a.s.l., where the climate is temperate sub-humid Ca (w2) (w) (e) (g) (Martínez et al., 1991). The mean annual temperature, precipitation, and humidity, from 2011 to 2020, are 15.7°C ± 0.17°C, 1,771 ± 173 mm, and 80.7% ± 20.8%, respectively (Zuloaga-Aguilar, 2021). The soils are classified as Distric Cambisols, developed on Tertiary (Paleocene-Oligocene) intermediate extrusive igneous rocks (Jardel and Moreno, 2000). The forest vegetation is dominated by Pinus douglasiana under silvicultural practices, together with Arbutus xalapensis, Fraxinus uhdei, and Carpinus tropicalis in the underground (Jardel and Moreno, 2000). Because of human pressure and climatic conditions, this forest is considered a fire-prone ecosystem, subjected to moderate and low-intensity fires every 4.2–7.3 years (Rubio, 2007; Cerano-Paredes et al., 2015). Therefore, previous fires might mask part of the results of the present analysis.

The prescribed burning event was carried out on 21 March 2017, to reduce hazardous fuels and improve pine forest regeneration. Forests firefighters of the National Forestry Commission (CONAFOR), in conjunction with a local forest brigade of the Ahuacapán Ejido, conducted the burning supervised by fire management specialists from the University of Guadalajara (DERN-IMECBIO) and from the Sierra de Manantlán Biosphere Reserve.

The size of the burned area was 7.3 ha (Figure 1). The weather conditions were a relative humidity of 38%–76%, a temperature of 19°C–22°C, and a monthly accumulated precipitation of 3.81 mm from 1 to 21 March. The prescribed burning had a duration of 10 h, with the following observed characteristics: a flame height of 1.01–3.27 m; a rate of spread between 1.4–2.5 m min⁻¹; flame temperature between 439°C and 593°C; and total vertical consumption of 9 cm of fuel (81% of the pre-burn material, see Figure 2) (Pérez-Salicrup, 2018).

A transect randomly oriented from SE to NW 175 m long (Figure 1B) was established within the burned area. Before and after the prescribed burn, a sampling area of 1 m² every 25 m was delimited. After the fire event, on 22 March 2017, in the center of each square, the soil burn severity (SBS) was evaluated by a visual estimation using a 30 cm × 30 cm sampling metallic frame, following the procedure described by Ryan and Noste. (1985) modified by Vega et al. (2013). This SBS index considered the condition of the remaining soil organic layer, whether it was unburned, partially or wholly charred, or if it was completely consumed remaining only ash. Also, this post-fire assessment considered the visual characteristics of the mineral soil: if this was undisturbed, if the soil structure and the soil organic matter (SOM) were affected, and if there were soil color changes. Then, with these descriptions, a value from 0 to 5 was assigned for defining the SBS level in each sampling point, according to Vega et al. (2013) classification.

Before and after the fire, three types of samples were collected inside the 1 m² area from the first 5 cm of the mineral soil at each sampling point: 1) five samples integrated into one bulk sample for chemical analyses, stored in a plastic bag, 2) one unaltered sample was taken with a 118.8 cm³ cylindrical steel core (5 cm depth), which was used for bulk density and moisture determinations, and 3) one unaltered sample was taken, from the organic layer and the upper 5 cm of the mineral soil, for micromorphological and morphometric analyses using cylindrical PVC cores.

A correlation analysis (significance of p ≤ 0.05) was performed among physical, chemical, and morphometric variables to avoid redundancy in the principal component analysis (PCA). The PCA was done to explain the changes due to the fire effect considering samples before and after the prescribed burn. The selection of the variables for the PCA considered the highest significance among them, demonstrated by their highest correlation. A JMP trial software was used for statistical analyses (JMP Statistical Discovery LLC, Cary, NC, United States).

The prescribed burn event in the Ahuacapán Ejido (Figure 1) left a heterogeneous mosaic of areas with different SBS (Figure 2). Although most of the area was affected by low SBS (levels 2 and 1), high SBS levels (4 and 5) were only found in smaller areas (only in two sampling sites of eight, Table 1; Figures 3C, D). In level 1, the organic layer was partially burned; in the case of level 2, the Oa layer was completely charred, and gray ash was present. In these low SBS, the temperature did not affect the mineral soil and remained undisturbed (Figures 3A, B). On the contrary, in the highest SBS areas, the organic layer was entirely consumed (levels 4 and 5). In SBS 4 level, high temperatures also affected the uppermost mineral soil horizon in the 0–2 upper cm, showed a slightly reddish color, lack of fine roots, and degraded structure. Besides, in SBS 4, a thick layer of white ash covered the mineral soil; while in SBS 5 most of the surface soil was bare, with black and reddish colors (Figures 3C, D).

Soil color measurements and mean moisture values for samples collected in the uppermost 5 cm showed changes after the prescribed burn (Figure 4). The mean values of the color parameters (L*, a*, and b*) decreased after soil burning, no matter the severity level. The reduction of luminosity (L*) mean values evidenced a soil darkening after the burn (Figure 4A). The decrease in the mean value of the a* and b* parameters showed fewer reddish and yellowish colors than the unburned condition (Figures 4B, C, respectively). Regarding soil moisture, this property tended to decrease slightly after the fire (Figure 4D). On the other hand, the soil bulk density increased 21.4% in the burned soils affected by low SBS and 22.6% by high SBS (Figure 4E).

Concerning the chemical soil properties, the pH values slightly increased after burning, especially in the high SBS (Figure 5A). The total organic carbon (TOC), no clear trend could be distinguished; it was reduced in soils affected by low SBS and increased in high SBS (Figure 5B). The TN concentrations did not show evident changes either (Figure 5C). Different trends in C/N ratios were evident because this ratio decreased in the low SBS. However, remarkable increases were found in the high SBS (Figure 5D). The TP values increased from unburned to burned condition 0.7 mg g⁻¹ (Figure 5E).

Images from thin sections with the organic layer and the upper 5 cm mineral soil from unburned and burned samples were morphometrically analyzed (Table 2; Figure 6). As logical, the lowest percentage of charred features was found in soils unaffected by fire (Figure 6A). All the scanned thin sections from the burn samples exhibited a darker layer on top (Figure 6). Samples with SBS level 1 showed the lowest number of unburned elements (e.g., aggregates, seeds, and plant tissues), coinciding with the highest porosity (46.3%, Table 2; Figure 6B) and followed by samples from SBS level 2 (37.5%, Figure 6D); the lowest porosity was found in the soil with SBS level 5 (14.5%). On the other hand, the morphometric analysis demonstrated that the highest number of elements with evidence of burning (e.g., charred and reddened aggregates, burned plant remains, and charcoal) was related to level 5 of the SBS index, reaching 50% of the considered area (Figure 6E; Table 2). Although level 4 showed fewer percentages of burning elements, only here mineral ash was detected (Figure 6E; Table 2).

In general, we identified a loss of structure and a reduction of the number of unburned elements and porosity with the increasing SBS. Additionally, ashes percentages and reddened aggregates were more significant in the highest SBS levels (4 and 5, respectively, Table 2). The morphometric data showed that the prescribed burn affected the soil organic layer and the first centimeter of the mineral soil for low SBS levels (Figures 6B, C) and the first 2 cm for high SBS levels (Figures 6D, E). The main proportion of burned remains corresponded to charred elements from the organic layer and the upper Ah horizon (Figure 6). It is important to mention that this examination might be including burning residues from previous fire events preserved in the soil. However, this previous evidence was discriminated by micromorphology analysis.

The following descriptions will relate to the main micromorphological characteristics of unburned and burned samples from the organic layer, the top 5 cm of the mineral soil, and the burned features associated with each SBS level.

In the organic layer of unburned samples, fresh plant and animal residues, such as wood fragments (Figure 7A), seeds (Figure 7B), and shells (Figure 7C), were frequent and could be distinguished. The mineral topsoil showed an organic soil structure (sub-rounded aggregates pigmented by organic matter) (Figure 7B). Although these thin sections belong to unburned soils, a few rounded charcoal fragments were identified, possibly from previous fire events. A zoogenic structure (dark rounded peds) integrated by coprolites was also common (Figure 7D). Brown subangular aggregates were frequent from a depth of 3 cm and onward (Figure 7C).

At this microscopic scale, samples from low SBS level 1 areas showed an organic layer with partially charred plant tissues and some burned vegetal remains, such as charred wood, roots, and spikes. The mineral soil exhibited an organic structure, where some aggregates get darker after burning in all samples affected by this low SBS level (Figures 8A–D).

In the thin sections that showed a low SBS level 2, the prescribed burning heat induced a color change in the first centimeters of the soil, commonly covered by a black organic layer of 1 cm thick (Figure 6C). This layer comprised charcoal fragments from the charred organic layer remains, such as seeds, leaves, needles, stems, roots, and spikes in all samples (Figures 8E–H). These charred organic materials were frequently complete and preserved their cellular structure (Figure 8F). The surface soil aggregates for SBS level 2 was charred, fractured, and partially or totally disaggregated; a loss in the soil structure was also identified (Figures 8G, H, 9B–D).

The organic materials were reduced to ash in the samples showing SBS level 4, (gray material in Figure 6D, yellowish material associated with charcoal in Figures 9B, C and small shiny crystals in Figure 9D). Some ash particles fill the pores of the surface soil (Figures 6E, 9C, D). At this level, the combustion process of the organic layer was completed. In the mineral soil, the organic matter in aggregates, fine roots, and seeds were charred or turned to ash; rounded reddened aggregates appeared (Figure 9A); the porosity decreased; and the surface soil lost its structure (Figure 6A). The organic layer was absent in the SBS level 5, while the soil organic matter was charred in the first centimeter (black aggregates) (Figure 6E, 9E–H).

In the PCA model, the first two principal components (PC) together explained 63% of the variance in the data (Figure 10A). The *L and pH are the variables that explain the most variance in PC1, while TOC and TN are in PC 2 (Figure 10B). On PC1, unburned samples are far from burned samples, mainly from those at SBS level 5 (Figure 10A). In contrast, on PC2 is possible to identify the overlaying by SBS level; in the center of the plot, the unburned samples are together to SBS level 1. The SBS 2 samples are in lower values related to an increment of TN concentration and bulk density (Figure 10), while SBS levels 4 and 5 group are in higher values related to the increment of TOC concentration and reddened elements (Figure 10).

The effect of fire on soil physical-chemical properties has been described in several studies. Soil organic layer amount and composition can be affected (Nave et al., 2011; Merino et al., 2015; Santín et al., 2016). With respect to mineral soil, important changes in morphological, physical, and chemical properties can happen (Neary et al., 1999; Mataix-Solera et al., 2011), particularly in the uppermost 5 cm, where the highest temperature is reached (DeBano et al., 1979; Neary et al., 1999; Mataix-Solera et al., 2011; Badía et al., 2017). All these perturbations depend on the heating and are reflected as SBS levels. In the case of low-intensity prescribed burns, the heating caused minimal adverse changes in mineral soils (Debano, 2000; Mataix-Solera et al., 2011). The present study’s PCA analysis separates the unburned from the burned samples by burn severity (Figure 10A), which agrees with the findings of Vega et al. (2013). The burned samples generally showed a darker color (L* value decrease), higher bulk density, higher pH values, and a higher percentage of charred and reddened elements than the unburned samples. Mainly, these properties explained 63% of the variance in the PCA analysis (Figure 10A). In addition, we observed modifications in other soil properties, such as moisture, TOC, TN, and TP.

Changes in the mean values of the chromatic parameters (L*, a*, and b*) indicate a darkening and a decrease in the redness and yellowness colors in the burned bulk samples (Figure 4). These results coincide with those obtained by Badía and Martí (2003a), where both value and chroma decrease as temperature increases, suggesting a darkening of the soil color of the charred samples. On the contrary, in the experiment conducted by Cancelo-González et al. (2014), the color parameters increased with the temperature rise in soils with different moisture contents (0%, 25%, and 50%). On the one hand, our unburned samples showed a mean value of soil moisture around 27%, and after the prescribed burn, a slight reduction of moisture was detected due to the water loss by vaporization in the uppermost 5 cm (Badía et al., 2017). On the other hand, Cancelo-González et al. (2014) experiment was done in soils derived from granitic rocks, whose mineralogy differs from that of our study (intermediate extrusive igneous rocks). Another difference was that we evaluated the color in sieved and homogenized samples, while Cancelo-González et al. (2014) measured the post-fire color on the undisturbed surface, which proved to be the best way for the evaluation. Nevertheless, the decrease in the L* parameter coincided with the morphometric and micromorphological analyses of the burn soils, where a dark black layer on the topsoils was identified. The darker color measured on burn samples with the colorimetric method corresponded to the abundant charred elements from the organic layer and the upper Ah horizon. Ketterings and Bigham. (2000) showed that soil became darker after fires at low temperatures (<250°C) of short duration because charred organic materials are dominant in the soil. Also, Cancelo-González et al. (2014) showed that organic matter content, temperature, and fire duration affected the color of burn soils. Concerning the reddened elements identified by micro-scale analyses of the samples with high SBS (Figure 9A), the colorimetric measurements did not detect them because this reddening was punctual, and the colorimeter performed an average of soil bulk samples color.

The observed increase of bulk density with heating (Figure 5) coincides with several studies where prescribed burn has been applied, independently of the kind of vegetation, brushes (Hubbert et al., 2006) or oak forest (Phillips et al., 2000), and for different soils heated under laboratory conditions (Badía and Martí, 2003a). Morphometric and micromorphological results showed the collapse of soil aggregates and the ash presence into the voids (Figures 9G, H), which can be other reasons to elevate the bulk density in the Ahuacapán Ejido. Likewise, Badía and Martí. (2003a) also found a moderate diminution of aggregate stability at 250°C, and Agbeshie et al. (2022) exhibited the breaking of organo-mineral aggregates and the clogging of porous by the ash-like causes of the bulk density increase. Moreover, this increase would indicate soil organic matter composition changes affecting the soil structure. Different studies have revealed carbohydrate decreases with heating, implying lower aggregate stability (Gregorich et al., 1996; Kavdir et al., 2005; Merino et al., 2018). Another process involved in structure stability is the loss of binders, such as fine roots, hyphae, and microorganisms (Mataix-Solera et al., 2011). Additionally, decreases in biological properties in the soil surface after controlled burn have been reported both in the lab (Badía and Martí, 2003b) and in the field (Armas-Herrera et al., 2018; Girona-García et al., 2018; Alfaro-Leranoz et al., 2023).

We observed a decrease in TOC in the low SBS samples but an increase in the soils with a high-level SBS (Figure 5). The critical consumption of organic carbon starts between 200ºC and 250°C and is completed approximately at 460°C (Giovannini et al., 1988; Vega et al., 2013); however, when the temperature is higher than 500°C, a total loss of organic matter occurs (Terefe et al., 2008). Likewise, Badía and Martí. (2003a) found a coefficient of correlation of R = −0.84 (p < 0.01) between the SOM and heat, due to a significant decrease of the organic matter with temperature increase at 250°C and almost the completed consumption at 500°C. Therefore, no conclusive trend could be distinguished with the TOC, probably because of mixed effects: the unfinished combustion process of the organic matter and the integration of charred materials into the first centimeter of the mineral soil (Soto and Díaz-Fierros, 1993; Úbeda et al., 2005; Afif and Oliveira, 2006; Scharenbroch et al., 2012). Thus, it is possible that the light increases in TOC in the area showing high SBS were due to the presence of charcoal, as the high C/N ratios suggest it. This is supported by the important percentage of charred organic fragments integrated into the mineral topsoil detected in thin sections from the high SBS (Figures 6D, E).

Regarding another chemical property, the slight rise in pH registered after the prescribed burn is described by the whole oxidation of organic matter producing ash with high Ca, Mg, K, and carbonate content (Ulery et al., 1993; Arocena and Opio, 2003; Certini, 2005), as well as by the formation of Fe-hydroxides (Schwertmann and Fisher, 1973). Under the microscope, calcite crystals were detected in high-level SBS (Figures 9F–H). Higher pH values in burned samples are often reported (e.g., Alcañiz et al., 2018).

Little change in TN concentrations was noted after the prescribed burn (Figure 5). Similarly, Arocena and Opio. (2003) found no important modifications in the TN; the authors considered that TN loss was mainly restricted to the forest floor, or TN losses were compensated by atmospheric N deposition. However, this last possibility can be considered in the long term. Another explanation for this slight change in TN is that the maximum temperature on the soil surface of the prescribed burn was not enough to allow the organic nitrogen volatilization from the soil (Neary et al., 1999), and only a minor portion of the nitrogen was thermally mineralized and integrated into the soil from the forest floor as NH₄ ⁺ (Dannenmann et al., 2011; Bird et al., 2015). On the other hand, Badía and Martí. (2003a) found a significant (R = −0.70; p < 0.01) reduction of TN at 500°C for two different soils under laboratory conditions. According to a review of several studies (Alcañiz et al., 2018), the behavior of TN after burning remains unclear. Due to the insignificant TN variations, the mean values of the C/N ratio showed the same trend as TOC concentration, with a reduction of the C/N ratio from the unburned to the burned samples at low SBS and an increment for the burned samples at high SBS.

The increase in the TP content after the prescribed burn (Figure 5) coincides with the results reported by other studies (Wienhold and Klemmedson, 1992; Úbeda et al., 2005; Afif and Oliveira, 2006; Merino et al., 2018). Besides, the available P (Olsen extractable P) showed a significant increase by heating (R = 0.48; p = 0.01) in lab conditions (Badía and Martí, 2003a). This TP enrichment is mainly favored by the mineralization of organic matter (García-Oliva et al., 2018), together with the increase in pH, due to the liberation of basic cations during the organic matter combustion and the ash production (Ulery et al., 1993; Kennard and Gholz, 2001; Arocena and Opio, 2003; Certini, 2005).

As mentioned before, one of the properties affected by fire is the soil structure. However, changes are not completely visible to the naked eye; consequently, the different SBS degrees are not identified. Micromorphological characterization is a complementary tool to evaluate the effect of the prescribed burns, as shown by Badía et al. (2020). The classification proposed by Vega et al. (2013) considers a macromorphological analysis, where the SBS 1 is associated with partial changes of the organic layer, while the mineral horizon remains unaltered. Our microscopic observations document some charred elements and burned features in the organic and mineral horizons (charred aggregates, partially charred plant tissues, burned roots, wood, and spikes). Although part of these charred materials can be formed during the present burn, it is important to notice that some features might also originated in previous fires (Figure 6B), as they were also identified in the unburned samples (Figure 6A). However, charcoals from prior fire events are easier to identify with micromorphological methods because they are more rounded than the fresh ones due to mechanical breaking during time (Ponomarenko, 1997; Ponomarenko and Anderson, 2001). Hobley et al. (2017) documented that charcoal fragments are quickly integrated into the soil profile after a fire event and can persist in the soil environment for a long time. In many archaeological sites, burnt features, identified at micro-scale, are well preserved, providing information about human impact on the ecosystem (Mallol et al., 2007; Huisman et al., 2019).

Similarly, in the SBS 2 Vega et al. (2013) index, the mineral soil is also undisturbed. However, the thin sections showed that the soil was covered by a 1 cm-thick black organic layer (Figures 3B, 6C). Under the microscope, we observed a loss of the structure while charred aggregates are fractured and partially or completed disaggregated, in the first centimeter of the mineral soil (Figures 8E–G). This observation might explain the increase of the soil bulk density (Figure 4E). Greene et al. (1990) showed that frequent prescribed burnings can break down aggregates surface. These results have been also confirmed by Arocena and Opio. (2003), who demonstrated the presence of cracks on soil aggregates through SEM-EDS observations. They interpreted these cracks as a thermal shock related to the temperature growth. Other studies, however, have not found immediate effects on mineral topsoil aggregates after the low-intensity sand rapid prescribed burn in shrubland in Central Pyrenees, Spain (Badía et al., 2020). In agreement to our findings, DeBano et al. (1998) and Mataix-Solera et al. (2011) mentioned that even low-intensity fires can affect soil structure and aggregate stability due to the combustion and/or transformation of the organic matter compounds that binds soil aggregates; this generates a loss of soil structure, a porosity reduction, and a bulk density increase (DeBano, 1981). Moreover, Badía et al. (2017) revealed that controlled burning only affects the physical and chemical properties of the first centimeter of dry soils. Another explanation for the loss of aggregate stability after prescribed burnings, without significant loss of organic matter, was exposed by Albalasmeh et al. (2013) and Chief et al. (2012), who found that steam pressure can break down aggregates at temperatures of 150ºC and 175°C in soils with high initial moisture. Nevertheless, it is important to mention when the fire temperature rises, an increment in the ped stability is noted because of the low contents of clay and Fe oxides in the soils (Thomaz, 2021). With these findings, we can complement the SBS Vega et al. (2013) index adding a slight perturbation of the first centimeter of the mineral soil at level 2.

For the SBS level 4, the micromorphological and morphometric descriptions agree with the general properties proposed by Vega et al. (2013), where the organic layer was completely consumed and transformed into a thick ash layer. The fire also affected the uppermost mineral soil structure and the surface fine roots and seeds (Figures 3C, 6D, 9A–D). The ash layer resulted from the oxidation of vegetal materials beyond charring during the fire, where the calcium oxalate was transformed into calcium carbonate by loss of TOC (Canti and Brochier, 2017). Calcite crystals were identified in the thin section of the SBS level 4 (Figure 9D). Reddened aggregates were also detected under the microscope. However, they were not recognized in the SBS index of Vega et al. (2013) because it is hard to identify them at a macroscopic scale. Our micro-scale study demonstrated that these mineralogical transformations expressed by reddened aggregates could start even before SBS level 5.

Regarding the Vega et al. (2013) index for the SBS 5, the micro-scale analyses showed that the soil organic layer was absent because it was completely consumed (Figure 6E). This characteristic can result from a higher temperature during the prescribed burn in this specific zone. The forest floor may have reached a temperature of around 460°C, in which the organic carbon consumption is completed (Giovannini et al., 1988), or the organic layer was absent in this area before de fire. In our study, the micromorphological and morphometric results showed that the aggregates and soil structure were slightly affected in the first 2 cm of the mineral soil (Figure 6E). Additionally, soil organic matter was charred and some aggregates were reddened in the surface soil (Figure 6E). The incipient soil reddening is a consequence of Fe oxides transformation due to heating; however, a continuous reddened layer covered by ash is expected only for severe fires and for SBS level 5 (Wells et al., 1979; Ulery and Graham, 1993; Ketterings and Bigham, 2000; Vega et al., 2013).