Fig. 6. Synergus punctatus Gillette, 1896 in Cymonomus curvirostris Sakai 1965

Changes of the Macrobenthic Faunal Community with Stand Age of a Non-native Mangrove Species in Futian Mangrove National Nature Reserve, Guangdong, China Ya-Fang Li1,2,3, Fei-Yan Du1,2,3,*, Yang-Guang Gu1,2,3, Jia-Jia Ning1,2,3, and Liang-Gen Wang1,2,3 1South China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Guangzhou 510300, China 2Guangdong Provincial Key Laboratory of Fishery Ecology and Environment, Guangzhou 510300, China 3Key Laboratory of South China Sea Fishery Resources Development and Utilization, Ministry of Agriculture, Guangzhou 510300, China (Received 18 September 2016; Accepted 20 June 2017; Published 19 July 2017; Communicated by Benny K.K. Chan) Ya-Fang Li, Fei-Yan Du, Yang-Guang Gu, Jia-Jia Ning, and Liang-Gen Wang (2017) Sonneratia apetala, a non-native superior rapidly growing mangrove species with wide environmental tolerance, has been introduced to Futian National Nature Reserve in Shenzhen, Guangdong, China, for mangrove restoration since 1993. However, the community structure of the associated macrobenthic fauna, a vital component of energy flow and nutrient recycling, remains obscure. The present study analyzed the macrobenthic faunal community, associated habitat characteristics and physico-chemical properties of sediment in rehabilitated S. apetala forests at stand ages of 8, 9, 14, 16 and 20 years from November 2014 to May 2015. Habitat complexity and stand structural heterogeneity varied with stand age. Sediment physico-chemical properties were similar for all stands analyzed, although soil organic matter (SOM) content was significantly higher in the 20-year-old stand than in others. Shannon-Weaver (H’) and Pielou’s evenness (J) indices of macrobenthic fauna were highest in 14and 16-yearold stands, respectively, and lowest in 8-year-old stands. In contrast, abundance and biomass peaked in 8-yearold stands and were lowest in 16-year-old stands. Multivariate analysis (cluster, ANOSIM and SIMPER) showed that the macrobenthic faunal community in the 20-year-old stand was different from other stand ages because of a greater abundance of small-sized mollusks and opportunistic species. Spearman correlation analysis showed that H’ was positively correlated with salinity. The distance-based linear model suggested that SOM was a significant predictor variable correlated with the macrobenthic faunal community. However, SOM was the only significant predictor variable explaining 12.7% of the total variation; this implies that the spatial variation of the macrobenthic faunal community here was mostly independent of the sediment properties measured. Therefore, we conclude that habitat characteristics such as vegetation characteristics can potentially explain the majority of the variation. Key words: Stand age, Habitat characteristics, Macrobenthic faunal community, Sediment physico-chemical properties, Sonneratia apetala. *Correspondence: E-mail: [email protected] BACKGROUND Mangroves serve as unique and vital ecosystems bridging the land and sea in tropical and subtropical coastal regions (Alongi 2008; Barbier et al. 2008; Morrisey et al. 2003), playing an important role worldwide. These ecosystems provide diverse ecological benefits including promoting biodiversity by offering habitats for coastal animals and birds, while protecting coastlines against disturbance, such as cyclones, tsunamis and water pollution (Luo et al. 2010). Despite their unique properties and ecological functions, mangrove ecosystems around the world risk being destroyed. During the last few decades, the population boom and rapid economic development in agriculture, aquaculture, industry and urban construction has greatly reduced Zoological Studies 56: 19 (2017) doi:10.6620/ZS.2017.56-19 1 the spatial extent of China’s mangrove forests (Chen et al. 2009a). In 1983, the United Nations Development Program and the Environment, Scientific, and Community Organization established a regional project concerned with the value of mangrove ecosystems in Asia and the Pacific; many countries have made considerable effort to restore mangroves (Bosire et al. 2008). The Chinese government has also made great efforts in mangrove reforestation since the early 1990s (Zheng et al. 2003). Because Sonneratia apetala from Bangladesh is a fast-growing species with a wide range of environmental tolerance, it has been planted in many locations along the coastline of Southern China, especially in Guangdong Province, even though the invasiveness of this non-native species has not been adequately analyzed (An et al. 2007; Tang et al. 2007; Chen et al. 2009a). Since 1993, as the pioneer species, S. apetala has been introduced to Futian National Nature Reserve in Shenzhen, which is the only mangrove forest located in the heart of a major urban area in China and is moderately contaminated with heavy metals. The wastewater discharge from the nearby electroplate factories and electronic instrument factories could be important pollution sources (Wang et al. 2013). As mangrove plants have relatively high heavy metal tolerance following the propagule stage, this species has rapidly become established and there are now forests of different ages up to 20 years. Lewis (2000) noted that ecosystem restoration should be used to replace persistent vegetative cover, a basic goal of mangrove rehabilitation programs. Macrobenthic fauna represent an important group in a mangrove ecosystem and strongly influence energy flow (Nordhaus et al. 2009). They form an important link between the primary detritus at the base of the food web and consumers of higher trophic levels (Macintosh 1984). In addition, some crabs graze on the propagules of mangrove plants, reducing the competition among seedlings and perhaps enhancing tree growth (Smith et al. 1991). Macrobenthic fauna are also sensitive to anthropogenic contamination; therefore, they can be useful indicator species for changes in ecosystem function (Lui et al. 2002). The recruitment of benthic fauna is one of the main criteria for judging the success of mangrove restoration programs (Field 1998). However, few studies have investigated the nature of the ecology of benthic fauna during rehabilitation of Futian mangroves, while most studies over the past two decades have focused mainly on productivity, seed dispersal and germination rates, litter dynamics, ecological assessment and leaf photosynthetic regimes in these planted forests (Zan et al. 2001, 2002; Chen et al. 2001, 2008, 2009b). Yang et al. (2014) conducted the only study in this region on the microbial community at different stand ages. However, the macrobenthic community structure and its changes with forest age is unknown. Information from these type of studies could provide essential information for improving the conservation and management of the Nature Reserve. It is; therefore, intriguing to determine how this non-native mangrove rehabilitation project affects the macrobenthic faunal community. The present study attempted to make this determination by analyzing the macrobenthic faunal community and environmental variables, including the physico-chemical properties of sediment and the characteristics of vegetation at different stand age. The findings will provide insight into how the plantation of mangroves influences the macrobenthic faunal community. MATERIALS AND METHODS Study area and sampling sites The study was conducted in the Futian National Reserve (22°30'-22°32'N and 113°56'- 114°03'E), located in northern Shenzhen Bay, which is the only mangrove forest located in the heart of a major urban area in Guangdong, China. This nature reserve features a subtropical monsoon climate, with a mean annual rainfall of about 1926 mm and a mean temperature of 22.2°C (Meteorological Bureau of Shenzhen Municipality: http://www.szmb.gov.cn/). This mangrove swamp, the sixth largest in China, covers 23.9 hm2 planted mangrove, including two non-native species, S. apetala and Sonneratia caseolaris, and 56.2 hm2 natural mangrove (Mao et al. 2012). In view of planting history, the sampling sites employed here were 8, 9, 14, 16, and 20 years old with plantation forests of S. apetala (abbreviated as SA08, SA09, SA14, SA16 and SA20) (Fig. 1), which were all within the same geographical location (according to the Guangdong Neilingding Futian National Nature Reserve Administration). page 2 of 14Zoological Studies 56: 19 (2017) Investigation of floral characteristics In November 2014, a 10 m × 10 m quadrat was randomly established in each of five sites. Parameters related to plant characteristics, including the density, height and diameter at breast height (DBH; ca. 1.5 m above ground) of mature individuals and seedlings of each mangrove species, were determined by the standard methods for plant analysis (HY/T, 081-2005) (Appendix 1). The vertical structure of the community was analyzed by Liu et al. (2016). The canopy Fig. 1. Locations of the sampling sites in the present study. (A) Vicinity map showing the locations of Shenzhen Bay, Hong Kong, and Shenzhen; (B) site locations in northeastern Shenzhen Bay; SA = S. apetala. The numbers in SA08, SA09, SA14, SA16, and SA20 represent different numbers of years since rehabilitation. N N (A) (B) page 3 of 14Zoological Studies 56: 19 (2017) coverage of vegetation type was estimated at noon on a sunny day by the decreased percentage of light intensity in the mangrove stands compared with that in the open sunshine. Collection and treatment of sediment samples for macrobenthic faunal community analysis In November 2014, and in February and May 2015, the macroinvertebrate fauna were collected from three 50 cm × 50 cm × 30 cm (length × width × depth) blocks of sediment at random locations in each plot. Initially, a steel frame was pressed into sediment to isolate the sample on the surface and prevent crabs from escaping during the excavation. Excavated samples were washed and sieved over a 0.5-mm mesh (Appendix 2). The collected fauna were temporarily preserved in 4% borax-buffered formalin before being identified and weighed. Above-ground gastropods were collected by hand from three replicates of 1 m2 quadrats randomly placed around the selected trees at each sampling plot. Some species, such as Neritina violacea, live within the prop roots of S. apetala; therefore, prop root gaps near the ground were checked carefully during collection. In the laboratory, all the fauna samples were sorted to species, and the biomass (fresh weights) and the number of each species was enumerated. Collection and treatment of sediment samples for sediment property analysis Five repeats in each sampling site were sampled using diagonal sampling from a single random point, and the soil in the upper 0-30 cm layer was excavated and mixed thoroughly; then, about 1 kg of each mixed soil sample was taken for analysis. The collected sediment samples were then homogenized, air-dried at room temperature, ground into powder and finally sieved through a 2-mm sieve for the subsequent analyses. Analyses of physico-chemical properties of sediment The particle size was determined by a particle size analyzer (Mastersizer 2000G Laser Diffraction Particle Analyzer, Malvern Instruments Ltd., UK) with a detection range of 0.02-2000 µm. The pH was measured by mixing 1 g dried sediment sample with 5 ml deionized water in a glass vial using a pH meter (EcoSense pH10A, Forestry Suppliers Inc., USA). Total nitrogen (TN) and carbon (C):N were determined by an elemental analyzer (Vario EL III, Elementar Analysensysteme GmbH, Germany). The salinity of sediment interstitial water was measured with a portable pH meter (Portable refractometer salinometer WYYII, Chengdu Hao Chuang Photoelectric Instrument Co., Ltd, Chengdu). Total phosphorus (TP) was extracted by the standards measurements and testing (SMT) method (González et al. 2005). Sediment organic matter (SOM) content and Cd, Pb, Cr, Ni, Cu and Zn concentrations were measured based on the methods described by Gu et al. (2014). Statistical analysis Two diversity indices, Shannon-Weaver (H’, base2 logarithm) and Pielou’s evenness (J’), were calculated for each macrobenthic faunal sample. One-way analysis of variance was used to test the differences in sediment physico-chemical properties between sites and seasons. Spearman’s correlation analysis was applied to examine the relationships between the univariate parameters of macrobenthic fauna and sediment properties. For multivariate analyses, the abundance data were first square-root transformed to down-weigh the very abundant species. The Bray-Curtis similarity coefficient for hierarchical cluster analysis was used to determine the similarity of macrobenthic fauna among sites. Similarity percentages (SIMPER) analysis was applied to determine the major species leading to the differences in community structure between sites. The best distance-based linear model (DistLM) using an Akaike information criterion (AIC) selection method was employed to select the predictor variables that could significantly explain the variation in the macrobenthic faunal community. The predictor variables used in the DistLM included Cd, Pb, Cr, Ni, Cu, Zn, TN, TP, C:N, SOM, pH and particle size in terms of sand, silt and clay fraction. Distancebased redundancy analysis (dbRDA) was used to provide a visual representation of the macrobenthic faunal community fitted to the significant predictor variables. All of the aforesaid analyses were performed using PRIMER 6 with the PERMANOVA + add-on, with the exception of correlation analysis that were performed using SPSS 19.0 for Windows. page 4 of 14Zoological Studies 56: 19 (2017) RESULTS Floral characteristics at different stand ages The floral characteristics analyzed here included vertical structure, individual density, average plant height, DBH, number of seedlings and branches per tree for five stand ages (Table 1). Five mangrove species, namely S. apetala, S. caseolaris, Kandelia candel, Aegiceras corniculatum and Acanthus ilicifolius, were found in this study area. Vertical structure of the community developed from simple to complex with stand age. SA08 comprised 4 mangrove species, including a tree layer (> 7 m) and a sapling layer (< 1.3 m). The tree layer consisted of only one species, S. apetala, while the sapling layer comprised A. ilicifolius, K. candel and Ae. corniculatum. SA09 included a tree layer, a sapling-shrub layer (1.32.5 m) and a sapling layer, comprising S. apetala, K. candel, A. ilicifolius, and Ae. corniculatum, respectively. SA14 was similar to SA09, except that the tree layer consisted of S. apetala and S. caseolaris. The tree layer in SA16 was the same as that in SA14, but A. ilicifolius replaced K. candel in the sapling-shrub layer and the sapling layer comprised K. candel and Ae. corniculatum. SA20 possessed the most complex vertical community structure, including two tree layers, an upper tree layer (> 7 m) and a secondary tree layer (2.5-7 m); the upper tree layer was S. apetala and the secondary tree layer was K. candel. The sapling-shrub layers were Ae. corniculatum and A. ilicifolius. Because of the density of A. ilicifolius (> 100 ind./m2), the sapling layer was not counted. The highest average plant height and DBH occurred in SA16 and the lowest were in SA14 and SA09, respectively. The highest number of branches were in SA20, and the lowest were in SA14. Sediment physico-chemical properties at different stand ages Because particle size of the sediment is persistent through time (Chapman and Tolhurst 2007), only the samples collected in November 2014 were analyzed for particle size. All sediments Table 1. Floral characteristics in the different stand ages of S. apetala in the Futian Mangrove National Nature Reserve in November 2014. DBH, diameter at breast height; DGH, diameter at ground height (used only for Acanthus ilicifolius because it is shorter than 1.5 m) site mangrove species Number of mature plants per 100 m2 Average height (m)a DBH or DGH (cm) Number of seedlings per 100 m2 Number of branches per tree SA08 Sonneratia apetala 67 9 30.3 39 1.3 Acanthus ilicifolius 67 1 1.3 0 Kandelia candel 26 Aegiceras corniculatum 2 SA09 Sonneratia apetala 27 10.6 33.9 11 1.4 Kandelia candel 13 1.3 7.6 47 Acanthus ilicifolius 68 1 1.3 Aegiceras corniculatum 15 SA14 Sonneratia apetala 33 7.7 32.9 6 1.2 Sonneratia caseolaris 3 12.3 50.2 0 1.3 Kandelia candel 2 1.4 1 Acanthus ilicifolius 109 1 1.3 0 Aegiceras corniculatum 2 SA16 Sonneratia apetala 13 10.9 52.7 0 1.5 Sonneratia caseolaris 712 51.5 0 1.1 Acanthus ilicifolius 400 1.5 1.3 0 Kandelia candel 16 Aegiceras corniculatum 3 SA20 Sonneratia apetala 24 9.5 40.1 9 1.7 Kandelia candel 2 2.5 15.5 3 1 Aegiceras corniculatum 1 2 7.3 22 1 Acanthus ilicifolius 10000 1.5 1.3 0 page 5 of 14Zoological Studies 56: 19 (2017) collected in the five different stand ages were clay-silt. The other physico-chemical properties of sediments are provided in figure 2. The sediment in the five sites were slightly acidic (pH < 7), except for SA16 in May 2015, which had the highest value (pH = 7.24); SA08 had the lowest pH. Generally, the spatial and temporal variations in pH were not obvious. The sediments were rich in SOM (6.95-13.49%) and this increased with stand age. Although the spatial variation was not remarkable, SA20 had significantly higher SOM content than the other sites (p < 0.05). Temporally, SOM content in sediment varied significantly (p < 0.05), with the lowest SOM content observed in May 2015. Similar to SOM, TN and C:N showed significant temporal differences, and the spatial variations in TN and C:N were not obvious except that high TN content and low C:N occurred in the older stand ages. Likewise, TP did not show any obvious spatial or temporal variations. Salinity varied temporally, with the lowest salinity concentrations observed in May. Heavy metal content did not vary significantly either spatially or temporally, except for Pb (Appendix 3). Macrobenthic faunal community at different stand ages At the five sites surveyed, a total of 35 taxa of macrobenthos were collected belonging to 5 phyla, 6 classes and 27 families (Appendix 4). The phylum Mollusca was the most diverse, composed of 16 species and accounting for 45.7% of the total number of species. The abundance, biomass and three diversity indices of macrobenthic fauna at different stand ages in the Futian mangrove swamp are shown in figure 3. Spatially, H’ and J indices ranged from 0.69 to 2.85 and 0.21 to 0.95, respectively, and were highest in SA14 and SA16, respectively, and lowest in SA08. In contrast, the abundance and biomass of macrobenthic fauna were highest in SA08 and lowest in SA16. Temporally, the H’ and J indices were highest in February 2015 and lowest in May; however, the abundance and biomass were highest in May 2015 and lowest in November 2014. Cluster analysis depicted the spatial and temporal variations of the macrobenthic faunal community (Fig. 4). Spatially, the macrobenthic faunal community in SA20 differed from the stands with other ages except SA08 in May 2015. No significant difference was observed in the macrobenthic faunal community in the other four ages within each sampling time. Results of the ANOSIM also indicated that significant differences existed in the macrobenthic faunal community among SA20 and other stand ages (R = 0.508, p = 0.002). Temporally, the samples in May 2015 of SA08 and SA20 differed from those at the other sampling times. Based on SIMPER analysis Fig. 2. Spatial and temporal variations of different physico-chemical properties of sediment. The numbers in SA08, SA09, SA14, SA16, and SA20 represent different numbers of years since rehabilitation. TN (%) = total sediment nitrogen content, TP (%) = total sediment phosphorus content, SOM (%) = sediment organic matter content, Salinity (‰) = the salinity content of sediment interstitial water, Cd (mg kg-1) = total sediment Cd content, Zn (mg kg-1) = total sediment Zn content, Cu (mg kg-1) = total sediment Cu content, Ni (mg kg-1) = total sediment Ni content, Cr (mg kg-1) = total sediment Cr content, Pb (mg kg-1) = total sediment Pb content, pH, C:N = the ratio of TC and TN. page 6 of 14 Zoological Studies 56: 19 (2017) Fig. 3. Spatial and temporal variations of Diversity, Evenness, Abundance and Biomass. SA08, SA09, SA14, SA16, and SA20 represent different site names in figure 1, with the numbers representing years since rehabilitation. Fig. 4. Dendrogram of hierarchical cluster analysis based on the square-root abundance data delineating the similarity of community structure among sampling sites at different sampling times. SA08, SA09, SA14, SA16, and SA20 represent different site names in figure 1, with the numbers representing years since rehabilitation. (Table 2), a higher abundance of Leamodonda punctigera, Capitella capitata, Assiminea sp., Stenothyra sp., Iravadia cochinchinensis and Assiminea brevicula in SA20 made the macrobenthic faunal community different from that in the other stand ages. In contrast, Ilyoplax dentimerosa was more dominant in the other four stand ages when compared with SA20. Comparing 0 0.5 1 1.5 2 2.5 3 3.5 0 0.2 0.4 0.6 0.8 1 1.2 0 200 400 600 800 1000 1200 1400 1600 SA20 SA16 SA14 SA09 SA08 0 20 40 60 80 100 120 140 160 SA20 SA16 SA14 SA09 SA08 Diversity Evenness Abundance(ind./m2) Feb Nov May Mangrove stand Biomass(g/m2) May May Nov Feb Feb May May Feb Feb Feb May Nov Nov Nov Nov Samples 100 80 60 40 20 Similarity Site SA20 SA16 SA14 SA09 SA08 page 7 of 14Zoological Studies 56: 19 (2017) the macrobenthic faunal community between the May samples from SA20, SA08 and the other samples, the polychaetes Neanthes glandicincta were more abundant in SA20 and SA08 in May 2015. Relationships between macrobenthic faunal community and sediment properties The Spearman’s correlation coefficients of the H’, J, abundance and biomass of the macrobenthic faunal community with the physico-chemical properties of the sediment are shown in table 3. H’ was positively correlated with salinity. J was negatively correlated with Ni. The dbRDA plot, which fitted the macrobenthic faunal community to the significant predictor variables using DistLM, is shown in figure 5. AIC selection of the predictor variables revealed that SOM, C:N, TP, Cu, TN, Pb, Ni, Zn, pH, Cd and Cr were the best combination of predictor variables, explaining 12.7%, 8.6%, 8.9%, 8.1%, 8.3%, 8.2%, 6.0%, 5.5%, 6.4%, 8.8% and 4.9% of total variation, respectively. From the plot, the macrobenthic faunal community in SA20 was different from that at other stand ages because of the high SOM and C:N values. Nevertheless, only one sediment property (SOM) was significant, and it explained 12.7% of the total variation, implying that the spatial variation of the macrobenthic faunal community in the reserve was independent of the sediment properties measured. Table 2. Similarity percentages analysis (SIMPER) indicating species leading to the difference in community structure between the 20-year-old stand (Site SA20) and stands with younger ages based on the squareroot abundance data. Cum. %: Cumulative percentage species Stand SA20 Younger stands Contribution % Cum.% Average Abundance Average Abundance Neanthes glandicincta 63.96 87.95 25.84 25.84 Leamodonda punctigera 39.10 0.00 19.04 44.88 Ilyoplax dentimerosa 14.21 27.06 9.72 54.61 Capitella capitata 15.99 6.88 8.05 62.66 Assiminea sp. 11.11 1.55 5.40 68.06 Stenothyra sp. 11.55 0.78 5.22 73.28 Iravadia cochinchinensis 11.11 0.56 5.17 78.45 Laonome albicingillum 3.55 8.22 3.03 81.48 Assiminea brevicula 5.33 2.89 3.59 85.07 Table 3. Spearman correlation coefficients of the Shannon-Weaver (H’), Pielou’s evenness (J), abundance and biomass of macrobenthic fauna with physicochemical properties of sediment (n = 15) Abundance biomass H’ J Cd -0.107 -0.189 0.029 0.061 Pb -0.500 -0.318 0.118 0.061 Cr 0.443 0.025 0.371 -0.032 Ni 0.057 0.139 -0.243 -0.227* Zn -0.500 -0.157 0.389 0.486 Cu 0.243 0.111 0.057 -0.171 pH 0.013 -0.175 -0.041 0.138 TC 0.307 -0.229 0.264 0.136 TN 0.381 0.281 0.272 0.209 TP -0.231 -0.240 0.254 0.252 C:N -0.375 -0.450 -0.057 -0.114 SOM -0.181 -0.273 0.359 0.061 Salinity -0.068 0.000 0.619* 0.097 *P < 0.05. page 8 of 14Zoological Studies 56: 19 (2017) DISCUSSION Effect of sediment properties and habitat structures on the macrobenthic faunal community In the present study, there was limited spatial variation of the sediment properties, except for SOM. The SOM increased with stand age; the SOM content in SA20 was significantly higher than in the other stand ages. The increase in SOM might be a result of the high productivity of these stands (Han et al. 2003; Yang et al. 2014; Tam et al. 1998). Additionally, the number and size of prop roots and pneumatophores increased with stand age, which could help the plants trap detritus and increase the amount of litter in the soil (Jonathan 2013). This was especially true for SA20, which contained abundant A. ilicifolius plants at densities that reached 100 individuals ind./m2. Regarding the temporal variation of the environmental factors, the exceptionally low concentrations of salinity and SOM in May were probably attributable to the different rainfall among months. Rainfall peaked in May (303.4 mm; Meteorological Bureau of Shenzhen Municipality: http://www.szmb.gov.cn/), concurrently with the lowest salinity and SOM (Dittmann et al. 2015). The responses of macrobenthic fauna to sediment properties were analyzed by Spearman correlation analysis and distance-based linear model (DistLM). The results showed that SOM was the best explanatory variable of the community structure of macrobenthic fauna; the abundance and biomass decreased with increasing SOM. Generally, high SOM represents an abundance of food resources, facilitating the growth and reproduction of macrobenthos. However, excessive organic matter might lead to reductions in species diversity and the proliferation of opportunistic species as a result of oxygen depletion and a buildup of toxic by-products (ammonia and sulfide) associated with the decomposition of these materials (Magni et al. 2015). The high SOM in SA20 explained the high abundance of opportunistic species such as C. capitata that occurred at this site based on SIMPER analysis. Ryu et al. (2011) found that organic enrichment selectively eliminated large-sized species resulting in the dominance of small-sized species in Fig. 5. dbRDA plot of macrobenthic faunal community fitted to significant predictor variables using BEST DistLM selection procedure and AIC selection criterion. SA08, SA09, SA14, SA16, and SA20 represent different site names in figure 1, with the numbers representing years since rehabilitation. -40 -20 0 20 40 60 80 dbRDA1 (28.1% of fitted, 24.3% of total variation) -60 -40 -20 0 20 40 dbRDA2 (20.9% of fitted, 18% of total variation) Site SA20 SA16 SA14 SA09 SA08 May May May May May Nov Nov Nov Nov Nov Feb Feb Feb Feb Feb SOM C:N TP Cu TN Pb Ni Zn PH Cd Cr page 9 of 14Zoological Studies 56: 19 (2017)