Effects of Bioturbation by Earthworms on Litter Flammability in Young and Mature Afforested Stands

Abstract

The quantity, quality, and accumulation rate of plant litter play a key role in forest floor flammability and, by extension, fire regimes. The varying foliage properties of different tree species also determine litter’s decomposition and its accumulation on the forest floor. The removal of litter by soil fauna, i.e., bioturbation, depends on both the dominant tree species and the successional stage of the forest stand. This research involved laboratory mesocosm experiments aiming to determine the effects of litter quality and earthworm activity on the flammability of the forest floor material at different successional ages. The mesocosms simulated the planting of four tree species (the broadleaf species Alnus glutinosa (L.) Gaertn. (Black alder) and Quercus robur L. (English oak) and the conifers Picea omorika (Pančić) Purk. (Serbian spruce) and Pinus nigra J.F. Arnold (Austrian pine)) at a reclamation site near Sokolov (NW Czechia). The mesocosms contained litter from these different tree species, placed directly on overburden soil (immature soil) or on well-developed Oe and A layers (mature soil), inoculated or not inoculated with earthworms, and incubated for 4 months. The surface material in the mesocosms was then subjected to simulated burn events, and the fire path and soil temperature changes were recorded. Burn testing showed that litter type (tree species) and soil maturity significantly influenced flammability. Pine had longer burning times and burning paths and higher post-burn temperatures than those of the other tree species. The immature soil with earthworms had significantly shorter burning times, whereas in the mature soil, earthworms had no effect. We conclude that earthworms have a significant, immediate effect on the litter flammability of immature soils.

1. Introduction

Fire represents a critical ecological mechanism at the local to global scale [1], influencing successional cycles, recovery rates [2], habitat structure, biodiversity [1], carbon cycle dynamics [3], nutrient redistribution [4], and many other ecosystem processes. In addition, fire maintains the vegetation composition of many ecosystems [5] and of several entire biomes [6,7]. The frequency of forest fires is likely to increase with ongoing climate change [8,9,10], and given their impact, understanding the drivers of forest fires is crucial.

Fire occurrence is affected by many external drivers, such as climate conditions and human impacts, including land use and management [11,12]. The occurrence of fires also depends strongly on vegetation and plant traits [13,14,15,16]. In addition to other factors, such as climate or terrain configuration, the properties of plant litter are a crucial factor in surface and ground fires in particular [13,14,15,16]. Litter composition and quality influence the fire regime [17] and the parameters of fires’ frequency and severity [7,15].

Flammability depends on many factors, such as the litter’s chemical composition, morphology, moisture content, and packing [18,19,20,21,22]. Substances such as waxes, resins, and various volatile organic compounds in general promote fires [19,20,21,23], while compounds such as proteins or protein–cellulose complexes tend to retard fires [18]. In addition to its chemical composition, the flammability of the forest floor depends on the shapes of individual components of the litter (in terms of length, thickness, curvature etc.), which cumulatively contribute to the textural character of the litter layer (such as its bulk density), with small needles forming denser, less flammable litter layers and larger needles less compacted and more flammable layers [22].

Decomposition also affects the amount of litter on the soil surface and its flammability [24,25]. Cellulose decomposes faster than lignin, so lignin accumulation presents a less flammable material [19]. Decomposition also changes the litter’s material structure and water retention ability and, by extension, its flammability [24]. Complete decomposition removes litter from the soil surface, reducing forest floor flammability [15,26,27]. The leaf traits that influence litter decomposition may differ from the traits described above as affecting flammability [22]. The energy spectrum of the leaves drives litter decomposition [28,29]. Litter originating from thinner leaves with lower C:N ratios (typical of fast-growing plants) usually decomposes rapidly and promotes nutrient turnover. In contrast, slower-growing plants produce thicker leaves with high C:N ratios and more recalcitrant litter [29]. Their varied rates of accumulation on the soil surface can thus influence the forest floor’s flammability.

Soil fauna can also critically influence litter decomposition [30,31]. Soil fauna consume more than half of the annual litter production globally, but in some biomes, such as temperate forests, they may consume volumes equivalent to the annual litter fall [32]. Cameron et al. [33] described model estimates indicating that earthworm invasions in North American forests could remove as much litter as wildfires at similar spatiotemporal scales. Litter not removed by soil fauna from the soil’s surface becomes incorporated into mineral soil through a process called bioturbation [31,34]. Although soil fauna bioturbation influences litter accumulation and humus formation, which in turn influence flammability [35,36], the impact of the soil fauna on forest floor flammability has not been studied systematically.

The effects of soil fauna and bioturbation depend in large part on litter chemistry. Interactions are clear during humus formation. In soil supplied by litter with a high C:N ratio, bioturbation by the soil fauna takes place slowly. Litter accumulates on the soil surface and gradually forms an organic layer of partly decomposed litter. This accumulation leads to the formation of mor [30,31], a type of humus characterized by slower microbial decay due to its high C:N ratios and low pH [25]. Mor is often associated with coniferous litter, characterized by high concentrations of flammable terpenes [20]. In contrast, soil supplied by litter with a low C:N ratio undergoes intensive bioturbation because most of this litter is quickly incorporated into the soil. This process leads to the mull type of humus, with little or no litter on the soil surface [31,34]. In areas involving litter with intermediate C:N ratios, bioturbation occurs at a moderate rate, and the moder type of humus forms and develops under conditions of moderate litter availability on the soil surface [31,34]. Bioturbation increases the infiltration and water-holding capacity of soil, which may increase litter moisture and reduce its flammability [37].

The present study addresses the effect of the soil fauna on the flammability of different forest substrates. This work involved an extensively studied common horticultural installation in which various tree species were planted side by side in the same parent material at a mine reclamation site [38,39]. The synchronous and proximal planting of different species offers unique opportunities to study how different tree species affect soil formation. Planting within the same substrate (mine overburden) without prior vegetation isolates the contributions of litter to the soil without effects from variations in the parent material or previous vegetation. Numerous studies have relied on the unique features of this particular common garden experiment, focusing on soil carbon, water storage, and the soil fauna’s relationship with tree species [37,38,39,40]. The findings of these previous studies give context to the results presented here.

For the current work, we chose four tree species that produce litter of varying qualities and support various levels of bioturbation and humus formation (from mull to moor). The laboratory mesocosms used in this study mimicked the forest floor environments associated with these four species in the presence or absence of earthworms. The effect of fauna soil modifications, including those caused by fauna bioturbation, is assumed to be more important in less developed soils [41], so we used immature soil that naturally lacked earthworms, and we used layers of more mature soil that had formed under the influence of the soil fauna over several decades. Three hypotheses were evaluated: (i) the flammability of the forest floor for conifers is higher than that for broadleaf species; (ii) earthworm bioturbation reduces the flammability of the forest floor; and (iii) earthworms’ effect is stronger in sites supplied with litter with a lower C:N ratio and in immature soils.

2. Materials and Methods

2.1. The Material

All of the plant and soil material used in the mesocosm experiments was collected at the extensive common garden experiment of the LTER site of the former Sokolov surface coal mine, located in the northwest of the Czech Republic (50.2417503 N, 12.7029267 E). In this experiment, various tree species were planted at the same time, on the same substrate, and under the same climatic conditions. For all forest stands in this experiment, the parent material was clay-rich mine overburden (a pH~8). The material was excavated during open-cast lignite mining [39]. More details about the sites, soil chemistry, and soil fauna communities can be found in Frouz et al. [39].

For the current experiment, forest stands of four tree species were chosen: alder (Alnus glutinosa), oak (Quercus robur), spruce (Picea omorika), and pine (Pinus nigra). These sites differ in terms of earthworm abundance and bioturbation intensity [39].

Table 1. The basic parameters of the individual older forest stands used to sample material in this study, based on Frouz et al. [39].

Parameter	Alder	Oak	Spruce	Pine
Oe depth, mm	0	0	10	10
A + AC depth, mm	85	55	18	2
C content, %	7.1	5.9	4.8	4.6
Earthworm density, ind.m−2	288	68	32	50

summarizes the conditions for c. 40-year-old sites (mature soils). The younger sites (immature soils) had developed for about 10 years but showed neither Oe nor A horizons and consisted only of litter above the overburden.

Litter was collected from the mature sites with individual tree species using litter traps, as described in Frouz et al. [38,39]. At the mature sites, we collected the soil horizons below the litter and corresponding to the Oe layer for the conifer stands and those corresponding to the organo-mineral A layer for the broadleaf stands. We also collected the layer below, at the interface between the developing Oe horizon and the original overburden or the bottom part of the A horizon, which gradually transitions to overburden. This transition between organic and mineral soil is referred to here as the AC layer. Only overburden was collected in a bare area of soil in the immature 10-year-old sites at a 0–10-cm depth. These materials were used to construct the laboratory mesocosms mimicking the developed sites. To mimic immature sites, we deposited litter collected from the mature site on top of fresh overburden collected at the immature (10-year-old) sites. The young sites lacked earthworms, whereas the older sites had a high earthworm density (Table 1). From the collected material, we removed earthworms, if present, and any large pieces of material, such as stones, large roots, and branches. The material was mixed by hand, with consolidated clay crushed into pieces about 3 mm in diameter. The litter was air-dried and kept dry until use, and the other soil material was sampled moist one or two days after rain and kept moist.

The earthworms used in the experiment were Aporrectodea caliginosa and Lumbricus rubellus. Both species were collected from a mixed forest in Praha-Modřany. Both are dominant species in post-coal-mining sites near Sokolov.

2.2. Mesocosm Establishment

The mesocosms were constructed in 40 × 210 × 297 mm iron boxes. The mesocosms mimicking the initial soil consisted of a 300 mm thick layer of overburden topped with 10 mm of litter. The mesocosms mimicking mature soil consisted of 15 mm of an AC layer overlaid with 15 mm of A (broadleaf) or 15 mm of Oe (conifer) material. The A layer was almost absent for conifers, whereas the Oe layer was absent for broadleaves. On top of the A or Oe layer, we spread a layer of litter 1 cm deep. The primary goal of this study was to compare the effect of earthworms on the forest floor’s flammability, so we chose to use the same thickness of the corresponding layers of forest floor for all tree species. Consequently, the layer thickness used in the experiments differed from the average thickness of these layers in the field (Figure 1 and Table 1). However, because of the variation in the field layer’s thickness, the mesocosm layer’s thickness fell within the range of thicknesses encountered in the field.

Once constructed, the mesocosms were frozen at −24 °C for 24 h to eradicate any earthworms that may have remained in the substrate. The mesocosms were then allowed to melt and reach the lab temperature for another 48 h. After this period, we added nine earthworms (seven specimens of A. caliginosa and two specimens of L. rubellus) to half of the mesocosms for each tree species and soil age treatment. The earthworm density used in the experiments was in the range of the earthworm densities observed in the field (Table 1).

2.3. The Burning Test

The mesocosms were topped with a metal lid and kept in the dark at 15 °C for 133 days. The mesocosms were opened weekly and sprayed with water to keep them moist. After the 133 days, the mesocosms were subjected to the burning test. On the day before the burning test, the lids were opened, and the mesocosms were allowed to air-dry for one day. On the test day, each box was lighted with a gas burner for 5 s, with the flame following a 2–3 cm wide line along the (shorter) side of the box near its edge. After 5 s, the flame of the gas burner was extinguished, and the following parameters were measured: burning time (s), taken as the time during which visible flames were observed (from the end of ignition to the point at which the last flame disappeared); fire path length (cm), taken as the perpendicular distance from the line affected by the burner flame to the point of the most distant burning litter; and the temperature (°C) in the zone affected by the fire, taken immediately after the flames disappeared. This last value was compared to the temperature of the box before burning and expressed as the temperature increase due to the passage of fire. Using the same procedure, the other side of the box also was lit and measured, followed by its application to the centers of the boxes. On the burning line at the edges of the box, we obtained one value for each of the parameters described above, and on the central burning line, the fire spread along both sides, so we obtained two values for each of the parameters.

2.4. The Statistical Analysis

A three-way analysis of variance (ANOVA) was used to evaluate the effect of tree species, the legacy of soil development, and the presence of earthworms (and their interactions) on the burning properties described above. In the three-way ANOVA, all mesocosms were evaluated together. In addition, a two-way ANOVA was applied separately to the mesocosms mimicking young and mature soils to explore the effects of tree species and earthworms only. If the effect of trees was significant, the difference between tree species was evaluated using the least significant difference post hoc test. All computations were performed using STATICA 13.0. (TIBCO Software Inc., Palo Alto, CA, USA)

3. Results

The temperature difference after fire showed the strong effect of tree species (Table 1), with the pine mesocosms experiencing a significantly higher temperature increase after the burning test than that for the other species (Figure 2). There also was a significant effect of soil legacy, with a higher temperature difference in developed soil compared with that in young soil. In the developed soil, the temperature increased for all tree species. For young and developed soil separately, tree species was the only significant factor affecting the temperature difference after the fire (Table 1), with pine showing a significantly higher temperature increase than that for the other species (Figure 2).

Tree species also significantly affected burning time (Table 1, Figure 2b). In contrast to the other parameters examined, there was no significant main effect of legacy, but there was a significant two-factor interaction between the presence of earthworms and legacy and a three-factor interaction among all tested factors (Table 1). In analyzing young and mature soils separately, we identified a significant effect of tree species in both (Table 1). In both cases, the burning times were significantly longer for pine than those for the other species. In young soil, the presence of earthworms had a significant effect, and there was a marginally significant interaction between tree species and the presence of earthworms. In mature soil, the effect of both earthworms’ presence and its interaction with tree species was marginally significant. The effects of earthworms’ presence tended to differ between the young and mature soils. In young soil, the presence of earthworms significantly decreased the burning times, whereas the effect of their presence in the mature soil was more complex, even increasing the burning times in the cases of oak and pine (Figure 2b).

The length of the fire path was significantly affected by legacy and by tree species, which significantly interacted. The fire paths were longer in young soil, but this effect varied among species. In young soil, the fire paths for pine were significantly longer than those for the other tree species, which did not differ from one another (Figure 2c). In mature soil, the fire paths were also significantly longer for pine than for the other species, and the fire paths for oak were significantly longer than those for the remaining two species. In young soil, there also was a marginally significant effect of earthworms’ presence (Table 1). Earthworms tended to decrease the length of the fire paths in young soil, but there was no significant effect of earthworms in the developed soil.

4. Discussion

We identified tree species as a major factor affecting the flammability of the litter, with pine burning for longer and having longer burning paths than the other species. In support of our hypothesis and in agreement with previous studies, we can conclude that the forest floor of the conifers was more flammable than the floor of broadleaves. This difference may be the result of the chemical composition of conifer needles, which supports litter flammability [19,20,21,23]. Needles are also likely to have higher surface volume ratios, compared to those in leaves, which supports their flammability [42]. In the older sites under the conifers, there was a well-developed Oe layer of partly decomposed litter, which was likely to burn better than the organo-mineral A layer under the litter in the broadleaf forest sites and may have contributed to the longer burning times and burning paths. Between the two conifers, pine appeared to be more burnable than spruce, possibly because of the longer needles on the pines, as longer needles contribute to the lower bulk density of the litter layer, which may support its flammability [22].

The effect of tree species was significantly modified by earthworms and by the legacy of previous ecosystem development. Earthworms’ presence decreased the flammability of the forest floor in the initial soils but not in the developed soils (Figure 2b). Previous research in the same plots has shown that earthworms support the removal of litter from the soil surface and its incorporation into mineral soil [39,40,43], which also may support C storage in mineral soil. We expect that the organic matter buried in the soil by fauna bioturbation was less flammable than the litter on the surface, consequently reducing the flammability of the forest floor. The bioturbation not only decreased the amount of litter on the soil surface but also likely reduced the connectivity of the litter and thus its flammability by removing litter fragments and depositing earthworm cast onto the soil surface. In agreement with the prediction of Frouz [41], the effect of such a modification was more pronounced in the initial soils where no such structures were present. Later, together with development of the organic soil horizon on the soil surface, bioturbation directly and indirectly determined flammability. Previous research [39] has shown that under pine, most of the organic matter accumulates on the soil surface, which may increase the forest floor’s flammability. In contrast, the development of the A layer is stronger for alder [39], supporting the water-holding capacity [37] and promoting microbial activity [39], which possibly causes faster decomposition, but both (faster decomposition and a higher water content) may reduce the flammability of the litter layer. Earthworms may contribute to these changes in the long term. Previous research on the same plots has shown that earthworms are much more abundant in broadleaf-related soil, particularly for alder, and less abundant in conifer-associated soils [38,39]. The presence of earthworms in the broadleaf habitat determines the incorporation of organic matter into the soil and A layer formation. In other words, earthworm legacy contributes to differences between young and developed soil, which appear to be significant in terms of the fire paths and temperature differences during burning.

As noted, these effects of earthworms are more pronounced when the Oe or A layer is not developed. In our data, this is reflected in the significant interaction between the earthworms’ presence and tree species and legacy (site age) in affecting burning time. Earthworms tended to reduce the burning times in young soil, but in mature soil, their effects were mixed. This pattern is in agreement with the prediction of Frouz [41] that earthworm-driven changes in soil structure will have a greater effect on young soil than on mature soil, where structures have already formed. Considering the substantial role of the soil fauna in the accumulation of litter on the soil surface [32], a better understanding of the relationship between soil fauna and forest floor flammability may support a better understanding of the fire dynamics in a changing world. This is even more important given that the bioturbation activity of the soil fauna may be affected by ongoing global climate change or the invasion of key ecosystem engineers into new areas [44], which could affect the fire regime as well.

5. Conclusions

This study shows that bioturbation by earthworms significantly shortens burning time as an important indicator of the flammability of litter, which is probably due to the removal of litter from the soil surface. This effect is more pronounced in young soils than that in mature soils. This could be due to the long-term development of the surface soil layers. For conifers, where the bioturbation is low, organic fermentation layers form underneath, which can also burn and promote flammability. In deciduous trees, where the bioturbation is high, an organo-mineral layer forms that does not burn and retains water, which can reduce the flammability. These long-term legacy effects may outweigh the immediate effects of bioturbation on mature soils.