Influence of trees and associated variables on soil organic carbon: a review

The level of soil organic carbon (SOC) fluctuates in different types of forest stands: this variation can be attributed to differences in tree species, and the variables associated with soil, climate, and topographical features. The present review evaluates the level of SOC in different types of forest stands to determine the factors responsible for the observed variation. Mixed stands have the highest amount of SOC, while coniferous (both deciduous-coniferous and evergreen-coniferous) stands have greater SOC concentrations than deciduous (broadleaved) and evergreen (broadleaved) tree stands. There was a significant negative correlation between SOC and mean annual temperature (MAT) and sand composition, in all types of forest stands. In contrast, the silt fraction has a positive correlation with SOC, in all types of tree stands. Variation in SOC under different types of forest stands in different landscapes can be due to differences in MAT, and the sand and silt fraction of soil apart from the type of forests.


Introduction
Soil organic carbon (SOC) is an essential component of environmental quality assessment. Atmospheric CO 2 is transferred into long-lasting pools, such as soil organic matter (SOM), thus reducing the atmospheric concentration of CO 2 (McBratney et al. 2014). SOC is vital for soil fertility, plant growth, and production (Janzen 2006). The total amount of carbon (C) stored in soil globally is estimated to be 1500 Pg C, with soil containing more C than the atmosphere (800 Pg C) and vegetation (500 Pg C) combined (FAO & ITPS 2015). A small variation in SOC concentrations can significantly affect the global carbon cycle (Walter et al. 2016).
Climate, land cover, soil texture, and soil order all affect SOC storage (Batjes 2016): Entisols and Aridisols store low amounts of SOC, whereas Histosols naturally store high amounts. When CaCO 3 is present in the soil, typically at pH 6.5 or higher, soil inorganic carbon (SIC) is formed (Lal et al. 1995). Histosols, Andisols, Spodosols, Oxisols, and Ultisols do not contain SIC, but Aridosols may store high amounts of SIC. On average, Inceptisols store the lowest amount of total C (SOC and SIC) and Histosols store the highest amount of total C (mostly SOC) (Eswaran et al. 2000). Global SOC concentrations range from low to high in soils of arid and temperate regions, respectively, and extremely high in organic or peat soils (Lal 2004).
Vegetation community structure may also affect the size of the SOC pool by altering both the microenvironment and soil characteristics (You et al. 2014). Globally, forest soil is a much more important C sink than live forest biomass, with concentrations two to four times higher in the upper 30 cm, and three to six times higher in the upper 50 cm (Calvode et al. 2020). Worldwide, forests cover 4.03 billion ha, approximately 30% of the earth's surface: of the total C stock in forest biomes, 37% occurs in low latitude forests, 14% in mid latitudes and 49% in high latitudes. A large part of the total SOC occurs in soils of tundra, pre-tundra, and taiga regions (Wyse 2012). Forests with different tree species vary in litter quality and root exudates, generating a divergence in soil properties, which may influence the soil microbial community (Chandra et al. 2016). SOC dynamics also differ due to variations in local vegetation types (Saiz et al. 2012;Gruba et al. 2015).
Research into the effect of tree species on SOC is crucial for mitigating the effects of greenhouse gases (Jandl et al. 2007). Tree species are one of several factors that influence soil C and nitrogen (N) inputs and outputs. Comparative studies of tree species grown under different conditions are beneficial in determining their effect (Binkley 1995), and the influence of tree species depends on the differences in soil conditions, such as parent material or land use (Vesterdal et al. 2008). Zhou et al. (2020) observed mixed forest stands of Cunninghamia lanceolata and Phyllostachys heterocycla have 3.33% of SOM compared with pure stands of C. lanceolata (1.77%) indicating mixed forest stands are better for storing SOM. Marler et al. (2016) have also indicated mixed deciduous tree stands can store more SOC (130.0 mgg −1 ) than pure forest stands of Leucaena leucophela (73.7 mgg −1 ). Mixing of Acacia tree species in Eucalyptus plantations in sandy and nutrient-poor soils involved soil C and N accretion after 7 years in the Republic of Congo (Koutika et al. 2019). Similar observations of mixed forest stands recording more SOC than pure stands were also reported by Yao et al. (2019). While, deciduous forests, with large C pools in the forest floor, store less carbon in soil (Oostra et al. 2006), and more C has been found in soils under mixed spruce forest in central western Europe (Berger et al. 2002). According to Guedes et al. (2016), coniferous tree stands of Pinus taeda recorded 135 Mg/ha of SOC compared with deciduous tree stands of Miombo (87 Mg/ha). Adivia et al. (2016) have provided consistent evidence of greater buildup of forest floor humus leading to more SOC in coniferous forest soils. Therefore, vegetation type is the most important variable driving the spatial pattern of SOC (Shi et al. 2012). Trees may influence the properties of soils beneath them, with a number of species affecting factors such as pH (Finzi et al. 1998), C and N levels, and the composition of the microbial community (Mitchell et al. 2010).
In this review, the author sought to determine whether variation in tree species is the only factor regulating the amount of SOC, or if other variables are involved. The effect of mean annual temperature (MAT), mean annual precipitation (MAP), tree age, elevation, soil pH, and the relative composition of sand, silt, and clay were selected for analysis as these have been commonly reported in the literature. The aim of this review is to determine what effect these variables have on SOC.

Methodology
For this review, a literature survey was carried out using the following search engines and academic platforms: ResearchGate (https://www.researchgate.net), Google Scholar (www.googlescholar.com), Science Direct (https://www.sciencedirest.com), Springer (www. springer.com), and Taylor and Francis (https://www. taylorandfrancis.com). Research published prior to January 2020 was considered for the present study. The keywords used were "SOC in different tree stands" and "influence of trees on SOC." Numerous studies have reported SOC from different forest types; however, the specific tree species and the level of SOC in the forests are less reported. Therefore, literature on "litter decomposition," where soil characteristics have been reported, were also collected. In numerous studies, SOC has been given in the form of C-storage. This is difficult to convert into SOC without knowing the bulk density value and weight of soil particles ≥ 2 mm, therefore these studies were not included. The most widely reported procedure for SOC estimation is the wet digestion method of the Walkley-Black technique (Walkley 1947), which uses the heat from a sulfuric acid reaction to oxidize SOC by hot chromic acid. In recent studies, a CHN-elemental analyzer is used for SOC estimation. In many studies, a conversion factor of 1.2 has been applied for the SOC levels determined by the Walkley-Black method; in the present review, however, no conversion factor was applied as DeVos et al. (2007) have pointed out sandy soils with conifers showed 6% higher recoveries than broadleaved species on heavier textured soils. For each laboratory and type of soil, they recommended that specific recovery factors need to be determined in order to standardize the results. Studies that reported only SOM were converted to SOC by dividing by 1.72 as SOM contains 58% SOC. In most of the studies, SOC levels were given in g kg −1 , which was converted into percentage by dividing by a factor of 10.
The composition of sand, silt, and clay were determined either by the micro-pipette method (Miller and Miller 1987), the sieve method, or the hydrometer method (Day 1965). Studies that measured soil pH using a 1:1 or a 1:2.5 soil to water ratio were selected, as many studies have also reported the pH as the soil to KCl ratio. The soil sampling depths were variable (Adekunle et al. 2011;Guedes et al. 2016;Zhou et al.2020); however, the most common depth was 0-20 cm, therefore, the present review was limited to reports from this depth. In all the selected studies, soil sampling was done by removal of the organic layer prior to sampling. As many as 60 tree species were recorded. Among the coniferous trees, the number of deciduous-conifer was very less compared with evergreen-conifers; therefore, they were grouped together as coniferous trees. The broadleaved-deciduous and broadleaved-evergreen trees were simply classified as deciduous and evergreen trees. Altogether the trees were broadly classified into four categories: coniferous, deciduous, evergreen, and mixed types.    Data were summarized by the mean, maximum, minimum, and standard deviation. The Pearson's correlation coefficient was used to find associations between SOC and the following variables: MAT; MAP; sand, silt, and clay composition; age; elevation, and soil pH. A Student's t test was performed to determine the significance level (p < 0.05) for these correlations. A one-way analysis of variance was performed on the SOC levels in the different categories of trees using SPSS (IBM, version 16.0).

Results and discussion
SOC under different tree stands SOC was found in very high concentrations under specific tree stands. Table 1 shows that the maximum level of SOC (11.02%) was found in mixed stands of the coniferous trees Picea abies and Pinus cembra (Margesin et al. 2016), followed by mixed coniferous stands of Picea abies and Larix decidua The results show that, out of the top eight tree stands that recorded the highest SOC level, three were from mixed coniferous tree stands: one from mixed deciduous trees, one from mixed coniferous and deciduous tree stands. The mixed stands recorded an average SOC of 4.62% ± 2.08%, with a range of 0.9% to 11.02% (p < 0.058). The average SOC of mixed tree stands was higher than the overall average of all trees (3.70%) ( Table 2). These results indicate that mixed tree stands are common and can store the most SOC.
The analysis of variance showed variation of SOC among the four types of stands (F 3,106 = 2.56; P < 0.058). SOC levels were most commonly reported from deciduous tree stands, with 35 concentrations recorded from different locations, followed by 34 recordings from coniferous tree stands. SOC levels from mixed tree stands were reported from 27 different locations, and only 14 evergreen tree stands had SOC levels recorded. As shown in Fig. 1, mixed tree stands can store the highest concentrations of SOC. Coniferous tree stands recorded higher amounts of SOC than deciduous trees, while evergreen tree stands recorded the lowest SOC level; however, this could also be due to fewer observations. Coniferous species generally accumulate greater SOC concentrations in the forest floor layer than deciduous  species (Augusto et al. 2015). Soil C reserves in the forest floor are generally greater under conifers than under broadleaved species (Vesterdal et al. 2013). As the needles of conifers take more time to decompose subsequent buildup of the litter can lead to more SOC, moreover conifers commonly occur in colder regions contributing to the delayed rate of decomposition. Disturbance due to human activities leading to erosion of soil is also an important factor in lowering SOC, which can be relatively lower in cold regions compared with warmer regions. Another factor can be attributed to the low amount of precipitation passing through a dense and low canopy in coniferous forests that prevented nutrient loss from the soil organic horizons (Lukina et al. 2019). Combining different tree species can have a profound effect on C accumulation ratios and SOC distribution within the soil profile (Chapin 2003), and compared with monospecific stands, the establishment of mixed forests promotes soil C sequestration. In mixed Norway spruce and European beech tree stands, the Norway spruce favored SOC accumulation in the forest floor whereas C incorporation into the uppermost mineral soil was promoted by root turnover of European beech. Therefore, conversion of monospecific plantations into mixed stands enhance SOM accumulation and stabilization in the mineral layers and, hence, the long- term storage of C (Andivia et al. 2016). The mixed type of tree stands provides a different composition in the input litter, regulating growth and survival of different types of soil macro-and microorganisms.

Soil type
The main soil type was found to be Cambisol, along with different subcategories, followed by Oxisol. Cambisols have a number of important characteristics that enable them to occur in widely differing environments: they contain weatherable minerals in the silt and sand fractions, have good water holding capacity, and have a neutral to weakly acidic soil reaction that promotes chemical fertility and an active soil fauna (Driessen 2001). Oxisols occur in the hot and humid conditions of tropical regions, where the B horizon is enriched with iron, aluminum oxides, and kaolinite (Beinroth et al. 1996).

Role of MAT and MAP
For all the tree types, SOC was significantly correlated with MAT (p < 0.05) (Table 3), with a lower air temperature resulting in a higher SOC content (Fig. 2). This correlation was also significant in deciduous and mixed tree stands. Sun et al. (2019) observed an average rate of reduction in SOC was 1.87% with 1°C increase in MAT in monospecific stands and mixed tree stands. However, annual litter fall (input) and soil microbial respiration (output) increased with MAT, indicating that a decrease in SOC concentration did not result from changes in either organic matter input or output, but from the balance between them. Chandra et al. (2016) had recorded a significant higher concentration of microbial biomass C in temperate forests having a maximum MAT range of 18-30°C than dry deciduous forests having 28-42°C. Liu et al. (2016) also recorded a negative correlation of SOC and MAT in different forests at separate provinces, citing lower temperatures slow down decomposition and respiration allowing SOC to accumulate. The correlation between SOC and MAP was not significant for all tree types, as well as for different tree stands (P > 0.05). This result was in contrast to the results of Yost and Hartemink (2019) who reported that SOC generally increased with increasing rainfall. They found that, on average, SOC exceeded 2% in soils from cold and temperate zones, and increased to 4% when MAP was between 500 and 750 mm, and 11% when MAP was between 800 and 1000 mm. In the tropics, SOC was < 0.5% when MAP was between 500 and 700 mm, and increased to 2% when MAP was between 800 and 1000 mm. Calvode et al. (2020) also stated that MAP was the variable most predictive of SOC, followed by lithology, land use, and soil pH.

Age and altitude
Tree age showed a significant and positive correlation with SOC for all tree types (p < 0.05): C storage in the soil increased with tree age. Singh and Sharma (2007) reported greater SOC and available macronutrients in older plantations of Populus deltoides compared with younger plantations. In Gmelina arborea stand, the highest percentage of SOM with 4.78% was recorded in the oldest stand of 20 years while the least 0.7% was recorded in the youngest stand of 10 years (Adekunle et al. 2011). However, it was not significant for the different types of tree stands (p > 0.05). Edmondson et al. (2014) reported there was no effect of tree size on soil C storage, especially for oaks where the largest individuals were 1.6 to 2.0 m dbh. The correlation of SOC with altitude was not significant overall (p > 0.05); however, it was significant in the deciduous tree stands (p < 0.05). Shedayi et al. (2016) demonstrated that organic C has a Table 3 Correlation coefficient between soil organic carbon (SOC) and mean annual precipitation (MAP), mean annual temperature (MAT), sand, silt, clay, age, altitude, and pH for all tree species and the four categories of tree stands  strong positive correlation with elevation in different types of tree stands.

Texture
The sand fraction showed a significant negative correlation with SOC for all trees (Fig. 4), as well as for the different types of tree stands (Table 3). The silt fraction demonstrated a significant positive correlation for all trees and for evergreen and mixed types of tree stands: the average silt fraction was highest in mixed tree stands (46.08%) ( Table 4). The clay fraction did not show a significant correlation with SOC, except for in coniferous tree stands (p < 0.05). The negative correlation of SOC with the sand fraction has been reported previously (Tiessen & Stewart 1983;Liu et al. 2016;Zhong et al. 2018), however, a positive correlation with the silt fraction is less well documented, although Riestra et al. (2012) reported a positive correlation of SOC content with clay plus silt content in forest soils. Loam soils are 40% sand, 40% silt, and 20% clay, which is ideal for plant growth due to the desirable characteristics of these mineral particles (Sun et al. 2019). Therefore, silt composition of up to 40% is beneficial for maintaining a stable SOC level in forest ecosystems. Sand content is affected by soil erosion; therefore, it can be used as an indicator for evaluating soil degradation under different land-use systems (Ayele et al. 2013). There is only limited information on how physical fractions and the chemical structures of SOC relate to climate and vegetation types, especially for forest soil (Watson et al. 2000). C storage is affected by soil texture and aggregation: the highest amount of soil C is found in the silt-and claysized fractions, while the sand-sized fraction is low in soil C (Galeote et al. 2015). Vegetation characteristics may be a local modifier of clay content under similar climatic conditions. Therefore, the correlation between SOC concentrations and clay content may be localized and climate-dependent, and regulated by the large moisture difference that plays an essential role in driving the significant positive correlation of SOC and the clay fraction (Zhong et al. 2018).

pH
The average pH of all the tree stands was 5.23 (Table 5), with the lowest pH found in coniferous trees (4.84) ( Table 6). The correlation of SOC with pH was not significant; however, a negative trend was observed (Table  3). Negative relationships between SOC and soil pH have been found in all soil groups, except the soil group, which has the highest SOC level . In general, pH values in the topsoil were lower because topsoil is rich in organic matter, which decomposes, leading to the production of more organic acids and thus lowering the pH (Hong et al. 2019).

Conclusions
This review demonstrates that variation in tree species is an important factor in the amount of SOC: mixed forest stands store more SOC than simple pure forest stands. Processes associated with individual trees, such as stem  flow and litter accumulation, can have significant effects on soil chemical properties, as well as on chemical components of litter composition (Riha et al. 1986). Coniferous tree stands have a greater SOC storage capacity than deciduous tree stands. A common variable that regulates or reduces the level of SOC could not be established for the different categories of forest stands except the sand fraction. However, the silt fraction and MAT were found to have positive and inverse relationships respectively, in all forest stands, as well as in the category of mixed forest stands. As highlighted by Mayer et al. (2020), further research is necessary to tease apart the influence of species and sites. Combining a network of common garden experiments, at greater spatial scales, could identify where and how certain tree species could be beneficial to C soil sequestration, and in which forms and soil layers. The important influencing factors of spatial variations in SOC concentration in different forest stands are MAT, sand, and silt fractions of soil.