1 Introduction

China is a large country in coal resources and the largest coal consumer in the world. Underground mining, which has been the most common coal mining technique in the world, accounts for the most coal outputs in China (Hu et al., 2019). Large-scale underground coal mining in China has frequently resulted in soil collapse, damaging land resources irreversibly, and affecting agricultural production and the ecological integrity of the environment. At present, it is estimated that for each 1 × 104 t of coal that was mined, about 0.1 to 0.3 hm2 of land was sinking in China (Hu and Luo, 2006). Moreover, in mining areas where water tables are near the ground surface, surface subsidence has led to surface water accumulation, provoking shifts from terrestrial to aquatic ecosystems with dramatic changes in the original ecological equilibrium. The latter issue has also led to the destruction of large farmlands, threatening food security, and triggering conflicts between rural people who once worked and lived from the land. To circumvent these issues, many subsided areas are currently being reclaimed into agricultural land.
Agricultural land provides essential resources related to food security worldwide, in general, and in China, in particular. As a result, land reclamation has become an important approach to increase the surface of agricultural land after the subsidence of the land overlying coal mines in China (Bian et al., 2018). It is estimated that about 32.6% of agricultural land has been reclaimed from coal-grain composite areas in China (Qu et al., 2018). The coal-grain composite areas in the eastern regions of China are vital to support food security in this area, but land subsidence as a result of coal mining activities are jeopardizing the integrity of its agroecosystems (Hu, 2019).
Land reclamation in subsided mining areas involves refilling the receded area with topsoil followed by compaction operations (Qu et al., 2017; Hu, 2019). As a result, the natural structure of the soil is severely altered (Bai et al., 2018). Moreover, key soil attributes are also modified after land reclamation. Reclaimed soils may present less soil organic matter, altered soil texture, lower soil water retention capacity, less soil nutrients, and lower native soil microbial diversity than the original soil (Ahirwal and Maiti, 2016; Luo, et al., 2020). All these changes may limit crop yield and land development following land reclamation (Zhao et al., 2015). Natural succession of reclaimed soils is slow and ecological equilibrium is normally reached after a long time following reclamation (Bai et al., 2018). During this time, the reclaimed land may also be exposed to wind and water erosion, as well as to contamination (Hu, 2019). Therefore, monitoring the status of key soil attributes after reclamation and over time is essential for gaining insights into soil development and land restoration following reclamation (Ngugi et al., 2018; Shrestha and Lal 2011). However, only few studies have addressed this issue (Qu et al., 2018, Min et al., 2017) and even fewer have focused on the existing ecological networks between soil microbial communities (SMCs) after land reclamation in mining areas (Hu, 2019; Luo et al., 2020; Ma et al., 2020).
Soil microorganisms are an essential part of the soil ecosystem. SMCs play a key role in the processes of decomposing soil organic matter, nutrient cycling and utilization of nutrients by plants (Novianti et al., 2018; Yuan et al., 2018). After land reclamation, changes in the soil physical and chemical properties may provoke changes in key soil attributes (e.g. pH value) but also in the SMCs and their related functions in the soil ecosystem (Harris, 2009; Delgado-Baquerizo, et al., 2019). The reclamation of mining soil into farmland mostly focuses on restoring the integrity of the agroecosystem aboveground whilst the restoration of the integrity and function of the belowground compartments is largely overlooked (Liu et al., 2019). Detailed assessment of SMCs is essential to evaluate the success of soil reclamation and development (Dangi et al., 2012). In this regard, healthy soils will normally present diverse SMCs (Tu et al., 2020), indicating an adequate functioning of the soil ecosystem (Delgado-Baquerizo, et al., 2016) and an adequate provision of ecosystem services by the soil compartment (e.g. organic matter decomposition, nutrient cycling) (Griffiths and Philippot, 2013). However, little is known about the diversity of and the function provided by the SMCs found in soils subjected to reclamation after coal mining activities.
In recent years, high-throughput sequencing technology has become a common method to provide information related to SMCs. Nonetheless, the translation of the outputs from sequencing into new practical knowledge is challenging due to the size of the data generated throughout and the difficulty to correlate these with other environmental variables such as key soil attributes (Wang and Brose, 2018; Delgado-Baquerizo, et al., 2019). Molecular ecological networks (MENs), using the mathematical and bioinformatics methods to construct ecological association networks, and provide a conceptual framework to identify microbial interactions and key populations, and relate with key attributes and functions occurring in the soil compartment (Deng et al., 2016). Moreover, MENs can be useful to identify keystone taxa of soil microorganisms (Kitano, 2004). Keystone taxa play important roles in the microbiome, such as driving community composition and functionality, and their variation might cause changes in microbiome composition and functioning (Delgado-Baquerizo et al., 2018). While previous studies have reported keystone taxa in the soil under multiple environmental conditions, information on how keystone taxa may change over time following land reclamation in mining areas is severely lacking (Berry and Widder, 2014; Fierer, 2017; Ramirez et al., 2018; Banerjee et al., 2018 and 2019).
The aim of this study is to assess the restoration success of disturbed mine soil following land reclamation into arable land. To do so, we investigate changes in key soil attributes and in the composition of SMCs along a reclamation chronosequence of 17 years following land restoration. We also evaluate changes in the complexity of the ecological networks established between the SMCs and their interactions as an indicator of soil development after land reclamation. The results from our investigation will provide novel and practical insights into the ecological restoration of disturbed mine soils worldwide.