4 Discussion

The results showed that land reclamation had a positive effect on the provision of soil ecosystem functions, which were measured in the light of the enzymatic activities (Figs. 2G-J). All the investigated plots subjected to land reclamation showed higher enzymatic activities than the control plots. The results suggested that URA, DHG and FDA increased rapidly after reclamation (Figs. 2G, 2I and 2J), indicating that long-time reclamation may change soil functions such as the conversion of soil nitrogen, or the soil respiratory metabolism (Delgado-Baquerizo, et al., 2016). Yet, an increase in the activity of PPO was only noticeable after 11 years from reclamation, increasing steadily subsequently (Fig. 2H), which might be related with the microbial activity of Chloroflexi . In the Fig. S5 (see supplemental material), the PPO showed the positive relation with phylumChloroflexi , whereas had negative relationships withProteobacteria and Bacteroidetes . In the Fig. 3A, the abundance of increased Chloroflexi and declinedProteobacteria and Bacteroidetes along the reclamation time. In addition, the activities of DHG and FDA presented some fluctuations over time, suggesting that continuous land management (such as crop planting, irrigation, or fertilization) is needed to maintain their activities levels high (Xiao et al., 2019; Hu et al., 2016). All the variations of soil enzyme activities in the study might suggest that the reclaimed soil became healthier, which might be related with the continuous agricultural production activities, as well as the natural successional processes (Dick, 1994; Sarathambal et al., 2016).
The effect of land reclamation on the studied key soil attributes was uncertain. Only the concentration of AK in the soil seemed to increase following reclamation (Fig. 2D), while the levels of soil N and P were lower in the reclaimed than in the control plots (Figs. 2E and 2F). The increasing AK in this study (Fig. 2D) might be related with coal gauges which contain potassium, and those coal gauges partially dissolving over time (Hu et al., 2015; Spargo and Doley, 2016). This observation could be also related to the fertilization schemes implemented in the studied plots by the local farmers as well as to the depletion of soil nutrients by the crops rotated in the studied plots (Yu et al., 2012; Wang et al., 2016). In addition, SOM showed a decreasing trend along the reclamation timeline and only 17 years after reclamation an increase in SOM was noticeable (Fig. 2A). The latter observation could be related to the accumulation of crop litter, leaves and roots over time, facilitating the formation of humus in the soil (Ezeokoli et al., 2019; Mukhopadhyay et al., 2016). It is also worth noting that the observed decreasing trend in SOM could be related to higher mineralization levels of SOM provided by a changing soil microbial community following reclamation, as we found herein (Zhong et al., 2015; Duval et al., 2016). The reclaimed soil composition has influenced the soil organic matter, which thus might affect the change of total N and Olsen-P contents, which indicated that the soil phosphorus content might be positively correlated with soil organic matter (Fig. S5, see supplemental material).
Contrariwise, the soil pH and EC levels increased over the reclamation timeline, indicating that soil alkalinity increased after reclamation. These observations contradicted earlier studies (Mukhopadhyay et al., 2014, Adeli et al., 2013) and were likely produced by the presence of a water table near the ground surface and with high concentration in salt (Qu et al., 2018) resulting from the release of cations (e.g. Ca2+, Mg2+) present in the coal gangue from past mining activities (Hu et al., 2015). Another possible source for the increase in soil pH and EC following land reclamation could be the application of inorganic fertilizers by the local farmers (Liu et al., 2018), or the soil materials used for reclaiming the land. The latter stresses that careful monitoring of the attributes and origin of the soil materials used in land reclamation should be considered to avoid worsening the attributes of the soil following restoration (Mukhopadhyay et al., 2016; Bai, et al., 2018).
In general, our results seemed to differ from previous studies indicating that land reclamation led to decrease the soil pH and to increase SOM and TN (Shrestha and Raj, 2010; Spargo and Doley, 2016; Li et al., 2014a; Sheng et al., 2015). However, as the plots under study have been subjected to intensive farming activities since they were reclaimed, it was hard to acknowledge which changes in the key soil attributes studied herein were influenced by land reclamation or by the existing farming practices. Future research will aim at disentangling the effects of farming and land reclamation on key soil attribute by identifying if key soil attributes played an important role in driving the multifunctionality in soil ecosystem (Delgado-Baquerizo, et al., 2016).
The results from analyzing the taxonomic composition of the SMCs (Fig. 3) indicated that the distributions at the phylum level differed among the reclamation plots (Fig. 3A), which evidenced the effect of land reclamation on the SMCs composition (Li, et al., 2014b). The relative abundance of phyla such as Bacteroidetes , Proteobacteriaand Verrucomicrobia declined along the reclamation timeline, whereas the relative abundance of the phyla Acidobacteria ,Actinobacteria , Chloroflexi , Planctomycetes andThaumarchaeota increased (Fig. 3A). These findings are somehow in disagreement with the results gathered in Li et al. (2014c), who suggested a decrease in the abundance of Actinobacteria andChloroflexi in old, restored sites, and who also reported a negative effect of land reclamation on the abundance ofActinobacteria . However, the increasing abundance ofAcidobacteria along the reclamation timeline was interpreted herein as a good indicator of soil health, provided the role that this phylum plays in relevant soil ecosystem processes (e.g. soil acidification; Tringe et al., 2005; Liu et al., 2016). In fact, we observed that Acidobacteria became the dominant phylum in all the reclaimed plots (Fig. 3A and Fig. 5), which is in agreement with previous studies evaluating the abundance of SMCs after land restoration (Li et al., 2014bc). Nonetheless, we detected a positive correlation between the abundance of Acidobacteria and soil pH (Fig. S5, see supplemental material), which differs from the findings reported before, that the phylum Acidobacteria decreased in relative abundance as soil pH increased (Chu et al., 2010; Griffiths et al., 2011). On other hand, the increasing levels of alkalinity found in the reclaimed plots (Fig. 2B) likely explained the increasing abundance ofActinobacteria (Fig. 3A), which has a great affinity for alkaline, arid soils. Yet this phylum was the seventh most abundant of the detected phyla, which is in opposition with previous studies reporting Actinobacteria as the most abundant phylum in reclaimed mine soil in semiarid climates (Bastida et al., 2013). Our finding may thus be related to the high soil moisture encountered during sampling (i.e. 35%, there was a five-day thunder and heavy rain prior to sampling;). In fact, the observed increase in abundance for the phylumChloroflexi was indicative of water tables present near the ground surface (Chen et al., 2017), and it correlates with the high soil moisture levels found at sampling time and with the relatively high abundance of the phylum Planctomycetes - i.e. aquatic bacteria commonly found in brackish water (Lindsay et al., 2001). Still, the steady increase in abundance of Chloroflexi along the reclamation timeline (Fig. 3A) may be related to the increase in the water retention capacity of the soil resulting from the observed increase in SOM in the plot assessed 17 years after reclamation - i.e. r17 (Fig. 2A).
Organic manure has been used as soil fertilizer in Eastern China since a long time ago. This popular farming practice could have distinctly affected the composition of the SMC found herein (Geng et al., 2008). The long-term application of organic fertilizer in the plots subjected to reclamation may have affected the soil physicochemical properties (Liang et al., 2014), which in turn may have affected the microbial microenvironments harboring different microbial communities (Jangid et al, 2008). For example, the increasing trend observed for the phylumThaumarchaea (Fig. 3A) could be related to the fertilization schemes implemented by the local farmers in the studied plots. It is worth noting that Thaumarchaea is an important ammonia oxidizer in aquatic and terrestrial environments, and that ammonia is a common form of nitrogen in organic manure (Brochier-Armanet et al., 2008). In fact, Thaumarchaea is the first archaea identified as being involved in the nitrification process (Brochier-Armanet et al., 2012). Hence, the observed abundance trend for Thaumarchaea may explain why the levels of soil TN decreased over the reclamation timeline, as discussed above (Fig. 2F) -i.e. more nitrification (Nitrification, the two-step oxidation of ammonia to nitrate via nitrite – is a critical component of the global nitrogen cycle. Biller et al., 2012) triggered by Thaumarchaeaderived into lower levels of soil N. Contrariwise, the decreasing trend observed for the phyla Proteobacteria can be related to the reduction in mine-based, soil pollution in the reclaimed plots over time (Banning et al., 2011; Lewis et al., 2012; Ma et al., 2017), stressing additional benefits of land reclamation on soil health. To clarify further the latter, we will investigate the levels of soil pollution (e.g. Polycyclic Aromatic Hydrocarbon (PAHs)), as well as the relationship between PAHs and phyla Proteobacteria in the study site in the future. As well known that, Protobacteria are the best bacteria to PAH degradation, containing PAH-degrading genes (e.g. naphthalene 1,2-dioxygenase, extradiol dioxygenase, Martin et al., 2012; Guarino et al. 2019). Guarino et al. (2019) have also reported thatProteobacteria was a phylum with higher proportions (82%), containing known species of PAH degraders (e.g. Pseudomonas ,Burkholderiales ) in the contaminated soil of Bagnoli brownfield site (Southern Italy). Other phyla, however, showed a relatively stable abundance trend along the reclamation timeline (Fig. S2, see supplemental material), as indicated in Li et al. (2014b). This can be due to having detected those SMCs at a stable stage in their succession at the time sampling was conducted (Novianti et al., 2018; Sun et al., 2019). Anyhow, the array of phyla for the SMCs detected in this study agrees with previous studies focusing on land reclamation (e.g. Li et al., 2014b; Chen et al., 2016; Luo et al., 2019), reflecting the eight major bacterial phyla commonly encountered in the soil (Fig. 3A, Fig. S2, see supplemental material; Chen et al., 2017). Moreover, the most abundant genera of SMCs found herein (i.e. RB41 and UTCFX1; Fig. 3B) were attributed to the high abundance of Acidobacteria andChloroflexi , given the demonstrated affinity between these phyla and the most abundant genera (Ngugi et al., 2018). Flavobacteriumand Pedobacter classified to Bacteroidetes , as well asSphingomonas and Luteimonas belonged toProteobacteria presented the similar phenomenon (Fig. 3B).
In this study, we identified strongly significant or significant correlations between the modules of the microbial ecological networks (MEN) and the studied key soil attributes, such as soil pH, OP, or DHG (Fig. 6), supporting that land reclamation had an effect on the SMC provided through changes in the key soil attributes. In this regard, Ngugi et al. (2018) have investigated the soil bacterial characteristics of 21 coal-mining sites, and found that soil bacterial structure was significantly correlated with soil OP value, which was similar to the conditions identified in the r8 network (Figs. 2 and 6). Phosphorus can be easily fixed in the soil, but its utilization rate is slow, making the concentration of available phosphorus (OP) in the soil small (Ngugi et al.,2018). Yet, some soil microbes, such as Acidobacteriamight produce acidic substances through metabolism, which then dissolve some insoluble phosphates and utilize them for their own metabolism (Chu et al., 2010), reducing the concentration of soil P. In the network identified for the plot sampled 11 years after reclamation (r11; Fig. 6B), we observed that the differences in terms of SMC were mostly affected by AK. Sun et al. (2019) pointed out similar findings, in which the distribution of bacteria was primarily affected by AK in coal-mining areas, although SOM and OP were also identified as key drivers of SMC change. We also identified a significant correlation between soil pH and the structure of SMC (Fig. 6), which is in agreement with earlier studies (e.g. Li et al., 2018; Hartman et al., 2008; Lauber et al., 2009), and which was mainly attributed to the role thatAcidobacteria may play on the soil pH (Table 2; Fig. 6). Although our result suggested that the key soil attributes present a complex relationship with soil microbial communities, which need further research to be disentangled, our results also confirmed that network analysis was an effective and feasible tool to analyze the relationship between environmental factors and microbial community structures.
The features of the microbial ecological networks (MEN) established between the detected SMCs varied across the five study sites (Tables 1-2; Fig. 4; Fig. S4, see supplemental material), suggesting that the time after reclamation substantially affected SMC stability and MEN complexity. MEN were more complex in the control plots (CK) than in the plots undergoing reclamation (Fig. 4). The latter may be related to the presence of more stable SMCs in the CK plots due to the absence of operations provoking soil disturbance as opposed to the plots that were reclaimed (Hunt and Ward, 2015; Helingerová et al., 2010). This is further supported by the higher MEN complexity found in r17 -i.e. higher SMC stability, and thus MEN complexity, is expected along the reclamation timeline (Luo et al., 2020; Ma et al., 2020). These findings may be directly related to the effect of time after reclamation on soil ecosystem health and stability, as more stable and complex SMCs are more resilient to environmental stress and ecosystem disturbance (Hunt and Ward, 2015; Dimitriu et al., 2010). Furthermore, the microbial species involved in each network was different, suggesting shifts in the dominant phyla along the reclamation timeline (Figs. 4 and 5). For example, the phyla Acidobacteria , Planctomycetes andProteobacteria were always present while the keystone phyla changed from plot to plot (Figs. 4 and 5, File S1-S5, see supplemental material). The variation here may cause changes in soil ecological functioning (Delgado-Baquerizo et al., 2018). Therefore, the observed changes in the composition of the SMCs and their interaction through MENs in relation to the time after reclamation is another indicator of the evolution of the soil ecosystem towards a more complex but stable system (Odum, 1969; Gonzalez-Ollauri and Mickovski, 2017) as time from reclamation (i.e. disturbance) progresses in spite of the soil disturbance triggered by past mining activities.