Introduction

The soils of mangroves store 5 to 10.4 Pg carbon (C) (Atwood et al. 2017) and are an important component of the global C cycle (Duarte et al. 2005; McLeod et al. 2011). Soil C pools are the result of the balance between C inputs via primary production and outputs via decomposition, volatilization and leaching of organic compounds (Amundson 2001). However, there are considerable uncertainties in the factors that influence the size of soil C stocks in mangroves, due to the inherent variability among and within mangrove systems (Bouillon et al. 2008; Breithaupt et al. 2012). Mangroves are amongst the most productive ecosystems on earth (Alongi 2014), but C accumulation rates vary locally, as below-ground productivity, which is a major contributor to these soil C stocks, is determined by a number of site specific biotic and abiotic factors (Fujimoto et al. 1999; McKee et al. 2007). Biotic factors include species composition, herbivory and the composition of the microbial community, while abiotic factors include salinity, soil bulk density (BD), soil water and nutrients (Krauss et al. 2014).

Detritus from mangrove below-ground biomass is known for its slow decomposition rates compared to detritus from terrestrial trees (Huxham et al. 2010). Slow decomposition rates of mangrove root tissue have been attributed to the predominantly anoxic conditions in mangrove soils, resulting in the build-up of C stocks over time (Middleton and McKee 2001). Additionally, fine root decomposition is slow with turnover rates of 1.7 to 4.4 years observed in mixed species mangrove forests in Florida (Castañeda-Moya et al. 2011). In oligotrophic mangrove habitats in Belize even slower turnover rates of approximately 10 years have been estimated (Middleton and McKee 2001) with varying decomposition rates among root fractions (Castañeda-Moya et al. 2011; McKee et al. 2007). Surprisingly, the decomposition of fine roots is about 1.5 times slower than that of coarse roots (McKee et al. 2007). Since mangrove species differ in root morphology (Ola et al. 2020; Pi et al. 2009) decomposition rates are likely to vary among species (Huxham et al. 2010; Middleton and McKee 2001). Variation in root decomposition rates has also been observed among sites, which was associated with variation in soil conditions such as BD (McKee and Faulkner 2000; Middleton and McKee 2001; Poret et al. 2007).

Variation in BD triggers responses in tree root growth, which may influence the rate of decomposition of root tissues (e.g. Day and Bassuk 1994; Greacen and Sands 1980; Omi 1985). In mangroves, root growth was reduced and root diameters were larger in high soil BD conditions (Ola et al. 2020). Root morphological changes are often accompanied by anatomical changes, which are linked to the decomposability of tissues. Anatomical changes in response to increasing BD in terrestrial plants have included increases in the diameter of the xylem vessels (Alameda and Villar 2012), size of the pith (Chimungu et al. 2015), size of cortical cells (Bengough et al. 2011; Wilson et al. 1977) and the size and number of endodermal cells (Wilson and Robards 1977). In the mangrove species Rhizophora mangle, the type of growth substrate (mud vs. sand) affected the volume of root gas spaces (Gill and Tomlinson 1977). A recent study showed that high soil BD had negative effects on the development of aerenchyma, and initiated the production of fibrous strands as part of the vascular tissue in Rhizophora stylosa (Ola et al. 2019). These anatomical changes may be reflected in root tissue chemistry such as increasing C to nitrogen (N) ratios and the presence of structural compounds such as lignin and cellulose (Bingham et al. 2010; Terzaghi et al. 2016), which may reduce rates of decomposition. Additionally, high soil BD has been found to enhance root tissue N content (Ola et al. 2019), which may increase rates of decomposition (Huxham et al. 2010). Thus, variation in root tissue anatomy and chemistry, due to varying soil BD, may affect root decomposition rates and therefore soil C stocks (Huxham et al. 2010; Middleton and McKee 2001).

Root chemistry is a major factor controlling root decomposition rates (Silver and Miya 2001). In mangroves, high concentrations of phosphorus (P) and N, and low C:N and lignin contents have been indicators of relatively high decomposition rates of both above- (Robertson 1988; Twilley et al. 1986) and below-ground tissue (Albright 1976; Friesen et al. 2018; Huxham et al. 2010; Inoue et al. 2011). The structural compound lignin occurs in cell walls of all vascular plants and is often associated with slow decomposition rates, due to its heteropolymeric structure and phenyl-propanoid subunits, which result in a stable biopolymer resistant to microbial attack (Sarkanen and Ludwig 1971). In R. mangle, lignin was decomposed much more slowly than other components leading to lignin-enriched soils (Benner and Hodson 1985; Lallier-Vergès et al. 2008). For example, mineralization rates of the readily leachable fraction of leaf litter were 10 times faster than those of the lignocellulosic fraction, of which the polysaccharides were mineralized at rates twice as fast as the lignin (Benner and Hodson 1985; Lallier-Vergès et al. 2008). Mangrove species differ in the chemistry of their root tissue (Ola et al. 2018), which may therefore contribute to variation in root decay rates among species.

This study aimed to assess the effect of variation in tissue quality, arising from the growth of mangroves in soils of differing BD, on the decomposition of root tissue from four mangrove species: Avicennia marina, Bruguiera gymnorrhiza, Ceriops australis and R. stylosa. It was hypothesized that the soil BD in which the roots were grown would affect litter quality and thus decomposition, which would also differ among species. Additionally, variation in root tissue chemistry was assessed to test the hypotheses that 1) plants growing in different BD soils may vary in the nutrient content of their tissues, C:N ratios and lignin content, which may influence decomposition; and 2) that species differences in nutrient levels and structural compounds would be associated with differences in the decomposition of the root tissue.

Materials and methods

Study sites

This study was conducted in the mangrove forest at Point Halloran Conservation Area (27.56°S, 153.28°E), Queensland, Australia (Fig. 1). The tidal regime at the site is semi-diurnal with a mean tidal range of 2 m (Australian Bureau of Meteorology, http://www.bom.gov.au/). The subtropical climate in the region is characterized by mean air temperatures ranging from 18 to 26 °C in the wet season (October–April) and 8 to 21 °C during the dry season (May–September) (Australian Bureau of Meteorology, http://www.bom.gov.au/). The bulk soil at the site has a relatively low density of 0.44 g cm−3, due to its high soil organic matter (SOM) content of 14.6% and a total porosity of 72.5% (Xiong et al. 2019). Porewater salinity at the site was 29 PSU (Lovelock et al. 2011). Further, N contents were 3.3 g kg−1 and P contents were 0.446 g kg−1 (Lovelock et al. 2014).

Fig. 1
figure 1

Location of the study site Point Halloran Conservation Area, Queensland, Australia

Full size image

Root material

The root material for the decomposition experiment was sourced from seedlings of four mangrove species (A. marina, B. gymnorrhiza, C. australis, and R. stylosa), which were grown for a period of approximately 200 days in pots (3 l) packed with soils with BD ranging from 0.2 to 1.2 g cm−3. Soil BD integrates a range of soil parameters. For example, increases in BD are inversely related to porosity, which limits oxygen supply and water availability (Kozlowski 2008), influencing both plant and microbial activity (Bengough et al. 2011; Cambi et al. 2015; Day and Bassuk 1994). In many studies of coastal wetland soils BD is reported and often used to calculate C densities. Thus, the experimental approach to growing roots in soils that varied in BD was to mix materials to simulate a range of soil BD that may be encountered by mangroves (Ouyang and Lee 2020). Variation in BD was achieved by gradually altering the proportions (Rivenshield and Bassuk 2007) of sand (BD: 1.5 g cm−3) and perlite (BD: 0.056 g cm−3; Chillagoe Perlite PTY Ltd., Mareeba, Australia) in the soil mix, assuming that the volumes of the individual components are additive (Ola et al. 2018). Characteristics of the soil mixtures can be found in Table 1. Plants were watered to field capacity 2–3 times per week using a salt solution (Ocean-Nature Sea Salt, Aquasonic, Wauchope, Australia) with a concentration of 20‰. Additionally, the plants were supplied with a tap water based nutrient solution containing 0.42 mM NH4NO3, 1.2 mM KNO3, 0.69 mM Ca(NO3)2, 0.1 mM NaH2PO4, 0.05 mM Fe-EDTA once a month (Hayes et al. 2017; López-Hoffman et al. 2007). Glasshouse temperatures ranged from 21.5 to 27 °C and relative humidity ranged from 57.9 to 64.9%. More details of the growth conditions and plant material can be found in Ola et al. (2020), as well as in the Supplementary materials (Table S1). The oven-dried (60 °C) root material was deployed in litterbags (see below). The chemical characteristics of the root tissue were assessed using five subsamples of root material for each BD x species combination, which were ground for nutrient analysis. The C and N contents were determined using dry combustion and infrared detection in a LECO analyser (CNS-2000, LECO Corporation, St. Joseph, USA), while the P content was determined using an acidified persulfate autoclave digestion followed by a colourmetric assay with ammonium molybdite and malachite green (Reef et al. 2010a). Lignin and cellulose content were determined using the acid detergent fibre method (Van Soest 1963a, b). To compare lignin and cellulose contents among A. marina, B. gymnorrhiza and R. stylosa, three subsamples from roots grown in soils with a BD of 0.8 g cm−3 were assessed due to limited root biomass availability in some treatments. Additionally, the influence of growth BD on cellulose and lignin content was assessed for root tissue of R. stylosa grown in soils with a BD of 0.4 and 1.2 g cm−3.

Table 1 Composition of the growth substrate (1000 ml) for all bulk densities (BD, g cm−3) tested and corresponding penetrometer resistance (kPa), porosity (%) and volumetric water content (%)
Full size table

Decomposition study

Root material was placed in litter bags (10 × 10 cm), which were made of nylon mesh (1 mm, Westernex, Perth, Australia) and heat-sealed after filling. The bags were filled with 4 g of oven-dried (60 °C) mangrove root material representing both coarse and fine roots. However, for litter bags of C. australis only 1 g of litter was used, due to limited root biomass. Four to five replicate litter bags were used for A. marina, B. gymnorrhiza, and R. stylosa. Due to limited root biomass production three replicate litter bags per growth soil BD were used for C. australis. The litterbags were inserted vertically into the soil, so that the tops of the bags were level with the soil surface (Middleton and McKee 2001). A field incubation period of three months (18/11/2017–18/02/2018) was chosen based on the observation by Hayes et al. (2017) for the same region that most labile fractions of mangrove root tissue are lost within less than six months of field incubation. After the collection of the litter bags, % mass loss of the litter was determined. Percent mass loss was calculated as percent initial dry mass of the litter minus final dry mass of the litter (Middleton and McKee 2001).

Statistical analysis

The influence of soil BD at which the roots were grown and species identity on decomposition (i.e. % mass loss) of root material was determined using a multiple linear regression analysis. A two-way Analysis of Variance (ANOVA) was performed to assess the effects of species and BD on % P of the root tissue. Robust two-way ANOVAs (using the R package rfit; Kloke and McKean 2012) were performed to analyse % N and C:N ratio, as the assumption of normality was not met by these variables. The robust ANOVA uses rank-based estimators and reduction of dispersion (RD) instead of traditional least squares (LS) estimators and LS reduction in sums of squares from the reduced to the full model (Kloke and McKean 2012). Interactions that were not significant were dropped from the models. Furthermore, post hoc analysis (Tukey) was applied to identify the differences among species. The differences in structural compounds (% lignin content, % cellulose, lignin:cellulose ratio) among species (A. marina, B.gymnorrhiza and R. stylosa), as well as the effect of BD on R. stylosa root tissue lignin content, were explored using one-way ANOVAs. All statistical tests were performed using R (R Core Team: www.R-project.org/).

Results

The decomposition of mangrove root material was influenced by soil BD conditions experienced by the plants during growth and species (F(5, 87) = 38.13, p < 0.001, with R2 = 0.69). There was a significant decrease in decomposition (coefficient = −5.576, t = −3.439, p < 0.001) with increasing BD of the soil in which the roots were grown (Fig. 2, Table 2). However, the effect of the soil BD in which plants were grown on mangrove root decomposition varies among species. For example, there is a limited effect of soil BD in which plants were grown on the decomposition of roots of R. stylosa compared to A. marina (Fig. 2). Additionally, decomposition differed among species. Decomposition was greatest in A. marina (coefficient = −2.501, t = −0.466) followed by C. australis roots, which did not differ significantly from the intercept. However, decomposition of root material was significantly smaller in R. stylosa (p = 0.002) and B. gymnorrhiza (p < 0.001) (Fig. 2, Table 2). Mass loss was 0.0046 g g−1 d−1 for A. marina, 0.0041 g g−1 d−1 for C. australis, followed by 0.0031 g g−1 d−1 for R. stylosa, and 0.0028 g g−1 d−1 for B. gymnorrhiza.

Fig. 2
figure 2

The relationship between bulk density of the growth media at which the tissue was grown, ranging from 0.2 to 1.2 g cm−3, and % mass loss of root material of mangrove species. Letters at the end of each regression line represent the species studied: Avicennia marina (A, filled circles), Bruguiera gymnorrhiza (B, open triangles), Ceriops australis (C, open circles) and Rhizophora stylosa (R, filled triangles)

Full size image
Table 2 Relationships from the multiple regression analyses for mangrove root decomposition in a common garden for roots grown over variation in bulk density (BD) of the growth media, which ranged from 0.2 to 1.2 g cm−3
Full size table

The BD of the soil in which the roots were grown had no effect on %P (p = 0.522), %N (p = 0.325) or the C:N ratio (p = 0.051) of the root material (Table 3). However, the root tissue of the species differed significantly in %P (p < 0.001), %N (p < 0.001) and C:N ratio (p < 0.001) (Table 3). Pairwise comparison of mean values showed that %P differed only between R. stylosa (0.027 ± 0.001%) and the other species within the family Rhizophoraceae (B. gymnorrhiza: 0.031 ± 0.001%, p = 0.016; C. australis: 0.033 ± 0.001%, p < 0.001) (Table 3, Fig. 3). Nitrogen content of the root material of R. stylosa (0.66 ± 0.02%) was significantly smaller than N levels in the root material of any of the other species (p < 0.001). This is also reflected in the C:N ratio, which was significantly greater in R. stylosa (49.4 ± 1.4) compared to the other species (p < 0.001). In addition, C:N ratios in the root material of A. marina (29.9 ± 0.9) were significantly smaller than that of B. gymnorrhiza (30.6 ± 0.8) or C. australis (36.4 ± 0.9, p < 0.001) (Table 3, Fig. 3).

Table 3 Result of the two-way ANOVAs for the effect of bulk density (BD) on root tissue chemistry (% phosphorus (P), % nitrogen (N), carbon to nitrogen ratio (C:N)) of the mangrove species Avicennia marina, Ceriops australis, Bruguiera gymnorrhiza and Rhizophora stylosa
Full size table
Fig. 3
figure 3

Differences between the mangrove species Avicennia marina, Ceriops australis, Bruguiera gymnorrhiza and Rhizophora stylosa in a) % nitrogen (N), b) % phosphorus (P), and c) carbon to nitrogen ratio (C:N) of root material. Values are means and standard errors (n = 4 to 5)

Full size image

Structural compounds i.e. lignin and cellulose, and associated lignin:cellulose ratio of root systems of plants grown in soil BD of 0.8 g cm−3 differed among species (for lignin the main effect of species F = 41.77, p < 0.001 and for lignin:cellulose F = 42.36, p < 0.001). Lignin content and the lignin:cellulose ratio were smallest in A. marina (lignin: 14.8 ± 1.2%), followed by B. gymnorrhiza (lignin: 26.7 ± 1.2%) and R. stylosa (lignin: 28.6 ± 1.1%; Fig. 4). Species did not differ significantly in % cellulose (F = 4.831, p = 0.056), although R. stylosa (26.2 ± 1.4%) tended to have smaller % cellulose than A.marina (33.1 ± 2.0%, Fig. 4b). Assessment of variation in lignin contents of R. stylosa found 28.7 ± 0.8% lignin at BD 0.4 g cm−3 and 26.1 ± 2.5% lignin at BD 1.2 g cm−3. However, mean lignin levels were not significantly different among BD treatments (F = 1.04, p = 0.342).

Fig. 4
figure 4

Root tissue content of a) lignin (%), b) cellulose (%), as well as c) lignin:cellulose ratio for mangrove seedlings of the species Avicennia marina, Bruguiera gymnorrhiza and Rhizophora stylosa grown in soils with a bulk density of 0.8 g cm−1 . Values are means with standard errors (n = 3)

Full size image

Discussion

Decomposition of mangrove root material was influenced by the soil BD conditions experienced by the plants during growth, with mass loss of root tissue declining with increasing soil BD of the growth substrate. There are no previous studies on the decomposability of root material of plants grown under different soil BD. The results suggest that BD affects C cycling in soils and soil C stocks by influencing the quantity of the SOM inputs (Ola et al. 2018, 2020) and tissue decomposability. The slope of the relationship between % mass loss and soil BD suggests that for each 1 unit increase of the BD at which roots grow, the rate of decomposition declines by 5.6%, which over decades to centuries could contribute to differences in C storage among mangrove forests growing on different substrates with different soil BD (e.g. organic vs mineral soil). For example, a decline in the decomposition constant, kd of 0.146 year−1 (Lovelock et al. 2017) by 5.6%, assuming 500 g m−2 year−1 root production (Castañeda-Moya et al. 2011) and root mass of 50% organic C, would result in a contribution to C stocks after 10 years of 72 gC m−2 compared to 67 gC m−2. In most field studies soil C declines with increasing BD, reflecting increasing mineral content of high BD soils. However, this relationship is often variable (e.g. Bhomia et al. 2016) suggesting that factors in addition to the soil mineral content, which could include characteristics of roots that influence the decomposability, may also contribute to observed levels of soil C in mangroves.

Plant roots undergo morphological and anatomical changes in response to variation in soil BD (e.g. Alameda and Villar 2012; Bengough et al. 2011). Previous work showed that BD had a significant effect on the morphology (number of roots, root length, root diameter, root surface area, and root volume) of the root material used here, although the effect was species specific (Ola et al. 2020). Further, some anatomical modifications such as changes in cell walls, providing additional strength for the penetration of dense soils, affect root structural compounds (Whalley et al. 1995), which has also implications for rates of decomposition. In mangroves, high BD soils triggered the development of fibrous structural features in R. stylosa stilt roots (Ola et al. 2019). Structural features are often associated with lignin, which is associated with slow decomposition rates (Bhatnagar et al. 2018; Silver and Miya 2001). However, lignin content did not vary significantly with the soil BD in which R. stylosa roots were grown, suggesting that the structural changes in roots grown at high BD may contribute to the protection of root material from microbial attack as was observed in terrestrial species (Talbot et al. 2012).

Mangrove species differed in mass loss during decomposition, which was greatest in A. marina, followed by C. australis, R. stylosa and B. gymnorrhiza. Previously reported root decomposition rates of Avicennia sp. (e.g. 0.0021 g g−1 d−1, Huxham et al. 2010) are smaller than those reported here (0.0046 g g−1 d−1). Decomposition rates of Ceriops sp. and Bruguiera sp. (0.0012 and 0.0013 g g−1 d−1, Huxham et al. 2010) are also slower than those measured here (0.0041 and 0.0028 g g−1 d−1, respectively). The more rapid rates of decomposition observed in Moreton Bay could be due to a range of factors that differ among studies, including differences in soil environmental conditions (e.g. oxygen, salinity, nutrient availability) during plant growth. Similar to observations of others, root tissue of Avicennia sp. decomposed at faster rates than Rhizophora sp. (Middleton and McKee 2001) or Bruguiera sp. (Huxham et al. 2010). Although decomposition of root tissue was generally slower than decomposition of leaf tissue, the rank order of decomposition of the genera in this study was similar for the above-ground litter decomposition with leaves from A. germinans decomposing faster than leaves of R. mangle (McKee and Faulkner 2000). In contrast to previous research (Huxham et al. 2010) decomposition rates of C. australis were very similar to that of A. marina, but the smaller mass of roots used for Ceriops sp. in this study, because of the limited availability of tissue, may have influenced rates of decomposition. The difference in decomposition rates of root material observed among species is consistent with a global analysis of variation in soil C stocks associated with those species (Atwood et al. 2017) and may indicate that differences in root morphology, anatomy and tissue chemistry among mangrove species are important contributors to global, regional and local variation in soil C stocks.

Nutrient levels of the root tissue differed among mangrove species, which may contribute to the differences in decomposition among the species observed. In a study of field-grown mature A. marina and R. stylosa trees, N and P contents were greater in A. marina than in R. stylosa (Alongi et al. 2003). The smaller nutrient content of R. stylosa root tissue reflects its high nutrient use efficiency compared to the other species (Reef et al. 2010b), which was also reflected in the high C:N ratio. Low tissue P and N concentrations accompanied by high C:N ratios, are often associated with slow decomposition rates (Albright 1976; Silver and Miya 2001), as observed in R. stylosa, whilst particularly low C:N ratios of A. marina likely contribute to the rapid decomposition of tissue observed for this species. The data suggest that the C:N ratio is a good predictor for the decomposition of A. marina and R. stylosa root material. However, N levels and C:N ratio do not explain the equally slow or rapid decomposition observed for B. gymnorrhiza and C. australis root material, respectively, suggesting that nutrient levels alone cannot explain decomposability of root tissue in all mangrove species.

Structural compounds such as lignin and cellulose influence decomposition rates in a wide range of terrestrial species (Silver and Miya 2001). In our study, root tissue lignin contents varied among mangrove species, with A. marina roots having the smallest lignin content and lignin:cellulose ratio. Lignin levels were similar in B. gymnorrhiza and R. stylosa, but B. gymnorrhiza had a smaller lignin:cellulose ratio, which corresponded with the greater decomposition for this species compared to R. stylosa. Indeed, lignin and cellulose levels are thought to contribute to species differences in decomposition rates of mangrove roots (Huxham et al. 2010; Middleton and McKee 2001). Differences in the quantity of lignin and its location within root tips have been observed for A. marina and B. gymnorrhiza (Lawton et al. 1981), which corresponds to the differences in mass loss between these species reported here. Recently, another study provided evidence for differences in the thickness of the root exodermis, as well as lignin and suberin content among mangrove species (Cheng et al. 2014). Greater levels of lignification and suberization were observed in apical, medium and basal roots of species in the family Rhizophoraceae (B. gymnorrhiza and R. stylosa) compared to Avicennia sp. (Cheng et al. 2014). However, lignin may also be present in other tissue such as endodermal and cortical cells (Gill and Tomlinson 1971; Lawton et al. 1981; Ola et al. 2019), affecting the quantity of the recalcitrant tissue within mangrove roots and the accessibility of C rich compounds to microorganisms (Talbot and Treseder 2012).

This experiment was conducted in a ‘common garden’ in order to assess how roots of different mangrove species grown at different BD decompose under the same conditions (i.e. in a ‘common garden’). However, a limitation of this experimental design is the inability to capture the potential effect of, for example, the soil microbial community, which can vary under different BD due to variation in water content and aeration, levels of nutrients and other factors. High soil BD are generally associated with dramatic reductions in the microbial biomass, soil respiration and enzyme activity (Cambi et al. 2015; Jordan et al. 2003; Li et al. 2004). For example, reductions in microbial biomass at BD of 1.7 g cm−3 impeded C mineralization compared to that in soils with a low BD (Beylich et al. 2010). Therefore, in order to understand the full effects of soil BD on root decomposition rates, incubations of root tissues in soils of different BD should also be considered, in addition to the effect of soil BD on the characteristics of root tissue.

In conclusion, the soil BD in which roots grow affects the decomposition of mangrove root material, which is likely due to variation in morphology and anatomy of roots growing under different soil BD conditions. Additionally decomposition differed among mangrove species, which may be attributed to differences in lignin content and N content of the root tissue. Hence, soil physical conditions such as soil BD play a crucial role in the C cycle of mangrove forests, not only by influencing the amount of the SOM inputs associated with root growth, but also by influencing the decomposability of the SOM inputs.