Mountain forests and their soils provide multiple benefits to the inhabitants of both high and lowlands, including the sustainability of the forest itself. However, today these areas are exposed to the impact of climate warming and selective logging. These, influence the mountain forests capacity to climate warming mitigation by C capture and storage, and these impacts might depend on the elevation at which these forest occur. In this study, SOC stocks in topsoil under P. hartwegii forests increases with elevation, while in the lower elevations the explanation of lowest SOC stocks appears to be confounding, due to environmental and logging factors. Nonetheless, despite we did not measure logging as such and the fact that we do not know the intensity and cutting cycles, results in this study indicate an effect of long-term selective logging not only on the forest structure but also on SOC stocks. Thus, logging has generated not only a greater structural change in P. hatwegii stands between 3,400-3,800 m, which show to be younger tree stands at earlier successional stages (higher density of trees, herbs and shrubs, as well as lower tree diameter and height), but also the SOC stocks decrease in an elevation where the species climax should result in higher accumulation of SOC.
Effect of elevation on soil traits
SOC stocks in topsoil under P. hartwegii forests increase with elevation along the 600 m gradient at the Nevado de Toluca; consistent with this, SOM content also raised as elevation increased, indicating, therefore, that this variable as the most significant in describing SOC stock trends along the elevation gradient. This association is expected because SOM is the main pool of SOC, which depends on two main drivers. One of them is precisely the input of OM to the soil through litterfall production, which is the dominant way of OM transfer from aboveground to the soil (Six et al. 2004; de la Cruz-Amo et al. 2020). The second one is the SOM accumulation given the low temperature at the highest elevations, limiting SOM output through a lower decomposition rate (Garten and Hanson 2006; Davidson and Janssens 2006; Tashi et al. 2016).
Temperature, on the other hand, decreases with elevation and has been reported to significantly influence OM decomposition (Davidson and Janssens 2006; Garten and Hanson 2006; Tashi et al. 2016), explaining up to 95 % of the variability of OM decomposition in high elevation forests (Salinas et al. 2010). In this study, the fact that MAT decreased significantly along the sampled elevation gradient while SOM and SOC stocks increased indicates that the last two are not driven by elevation as such on this mountain forest, but by the lower temperature associated with a higher elevation and its effect on vegetation productivity and OM decomposition (Salinas et al. 2010; Shedayi et al. 2016; Tashi et al. 2016). Temperature reduction has a positive impact on SOC stocks due to a lower rate of litter decomposition, lower turnover rate, and higher SOC residence time at high elevations (Du et al. 2014; Tashi et al. 2016; Becker et al. 2019). For example, Garten and Hanson (2006) found a declining SOM decomposition rate with increasing elevation (335-1,670 m) in a forest at the southern Appalachian Mountains, reporting a longer residence time from 11 to 31 years, which resulted in an increase of SOC stocks with elevation. Conversely, based on a soil incubation experiment, Tian et al. (2016) concluded that SOC residence time (stability of SOC) does not depend on temperature only but on the labile size of SOC. Therefore, the SOC residence time in P. hartwegii mountain forests depends not only on temperature increase by climate change but also on the proportion of labile C in this mountain soils rather than the total SOC.
BD, the second variable explaining most of the SOC stocks variability along the elevation gradient in this study, increased significantly with elevation. BD is typically low (>0.9 g cm-3) in volcanic soils such as Andosols due to the dominance of largely weathered volcanic glass in the sand and silt fractions (Delmelle et al. 2015). However, BD is not only related to the density and arrangement of mineral soil (sand, silt, and clay), but also to the OM particles and SOC (Neall et al. 2006; Delmelle et al. 2015). Thus, the role of soil texture in both SOC content and BD is widely known; for example, sandy soils have shown higher BD and lower SOC, while low BD soils tend to have higher OC (Lukac and Godbold 2011). In this study, sand and silt fraction were positively related to elevation, and negatively related to MAT (Table 2). An increase in sand content as elevation increases is expected in mountain volcanic soils (resulting from volcanic ash) where rock weathering and soil formation processes are less intense, resulting in shallower, coarser, and sandier soils at higher elevations (Simon et al. 2018).
Andosols, equivalent to Andisols in the World Reference Base for Soil Resources (WRB), are volcanic soils characterized by their great storage capacity of OC (from 42 to 207 Mg OC ha-1) and their stabilization, due to the presence of aluminum-humus complexes in the soil surface horizons (Msanya et al. 2007; Covaleda et al. 2011). However, according to Msanya et al. (2007) and Covaleda et al. (2011), volcanic ash soils also stabilize OC by forming pseudo-sands and silts. Pseudo-sands are derived from short-range ordered materials, resulting from the weathering of volcanic glass and are present in the sand fraction of soils (Msanya et al. 2007). Under the climatic change context, sand content in soils might play an important role on SOC stabilization in high mountain forests because sand content is strongly related to the intermediate SOC pool from SOM decomposition, which contributes significantly to the cumulative C-CO2 release (Tian et al. 2016). Therefore, it would be important to explore the role of the sand fraction on potential mechanisms of SOC stabilization in these relatively young volcanic ash soils.
Main drivers of SOC stocks on mountain forest soils from the Nevado de Toluca
In this study, the fact that the largest trees were found in the highest elevation plot, consistent with the largest amount of SOM and SOC, suggests litter as the main reason why there were greater SOC stocks recorded at the highest elevation, i.e, there is greater litter production at the highest elevation. However, it is well known that plant growth is reduced at higher elevations due to environmental limitations, such as low partial CO2 pressure, short growing seasons, soil nutrients limitations, extremely low temperatures, etc., as reported for P. hartwegii (Harsh and Bader 2011; Körner 2015; Alfaro-Ramírez et al. 2017). For this pine species, the climax has been reported at 3,700 m (Perry 1991), and a decrease of tree height and diameter as elevation increases towards the treeline ecotone (from 3,980 to 4,130 m) has been shown at the northwest and east-southeast side of the volcanic cone of the Nevado de Toluca, as per the findings of Alfaro-Ramírez et al. (2017). Although this last study was performed on a different side of the Nevado de Toluca (northeastern side) where the ecotone occurs at a much higher elevation, what was reported by Alfaro-Ramírez and collaborators, is an indicator of the limiting conditions to the P. hartwegii growth as elevation increases. Thus, the fact that we found the largest trees and highest SOC stocks at the high elevation plots, and not as expected at 3,700 m, indicates that other external factors have modified the P. hartwegii forest structure as well as the SOM input and the SOC residence time. Change in forest structure is also supported by the highest herbs and shrub density, and the lowest grass cover, not typical of P. hartwegii forests at low elevations. According to Challenger (1998), typical P. hartwegii forests have a relatively simple structure, with only one tree layer and few or no herbs and shrubs in the understory. Jafari et al. (2013), report similar results with respect to that factors such as harvesting and selective logging, that modified vegetation structure, have a stronger effect in lower elevation forestlands than in higher.
In general, changes in the plant community due to natural or anthropogenic causes have a major impact on SOC stocks (de la Cruz-Amo et al. 2020). These mountain forests have been under big anthropogenic pressure for a long time. Timber from P. hartwegii has been in great demand for commercialization of furniture’s and construction wood (most of them unauthorized), the largest and highest trees (>35 cm and >20 m, respectively) being the most utilized (Franco et al. 2006). On the other hand, illegal logging occurs in a much lesser degree on the upper area, the buffer zone of the PAFF of the Nevado de Toluca. The logging restrictions allow the permanence of large trees, contrary to the low elevations (up to 3,700 m) where we found the smaller trees and where the forest is more degraded because of the different human activities. Thus, the lack of protection and implementation of the law means that there are no mechanisms to control illegal logging and its impact on the capacity of these forests and their soils to store OC. SOC losses of about 8 % due to logging have been reported in temperate forests (Nave et al. 2010). Logging effects on SOC stocks topsoil can be direct and indirect; tree removal directly reduces OM input to the soil, and indirectly affects decomposition rate, modifying litter quality and microsite conditions (Pérez-Suárez et al. 2012). Logging modifies litter quality input to the soil by creating canopy gaps that promote secondary succession in the understory, i.e., it increases the presence of easily decomposable herbs and shrubs (less recalcitrance), limiting SOM pools (Cepáková and Frouz 2015; Bomfim et al. 2020). Microsite conditions are also altered when canopy gaps increase the light incidence and temperature on the forest floor (Coletta et al. 2017; Bomfim et al. 2020). This increases litter decomposition rate and, hence, the heterotrophic respiration and the continuous C-CO2 outputs from the soil (Six et al. 2004; Coletta et al. 2017) for a long time (up to 45 years) after logging (Chiti et al. 2016).
In summary, SOC stocks along the elevation gradient are controlled through the elevation effect on MAT and soil texture. The results suggest that SOC stocks in high elevation soils occur due to lower temperatures that allow long-term SOC accumulation which is simultaneously promoted through microsite conservation by restrictions on forest exploitation; conversely, at lower altitudes, changes in forest structure strongly reduce SOC stocks. Based on this study, soils under mountain forests not only should be considered in the carbon inventories but also in conservation and differential reforestation programs according to the elevation gradient. Elevation and forest management activities, as important drivers of SOC stocks, could make it possible to keep the greatest SOC reservoirs and long-term mountain forest sustainability.