Frequent extreme climate events globally threaten ecosystem services and functions, particularly in fragile regions. As one of the vital "ecosystem engineers" and "living skin", mosses have been demonstrated to impact soil biogeochemical cycles markedly, yet their role in shaping the soil multifunctionality of fragile regions under extreme climates remains unclear. A one-year in-situ field experiment was conducted to simulate the effect of moss crust on soil multifunctionality in intense karst rocky desertification regions under extreme drought and heavy precipitation. Compared with bare soil, the moss crust significantly enhanced soil multifunctionality, with increases of 44.0 %, 32.7 %, and 54.8 % in extreme drought, ambient precipitation, and heavy precipitation conditions, respectively. This was primarily attributed to moss crust promoting the related variable on soil carbon cycling, such as improving beta-1,4-glucosidase and microbial biomass carbon under extreme climate conditions. Furthermore, moss crust reduced the dependency of soil on precipitation by boosting above-ground biomass and modulating microbes to enhance soil multifunctionality. Collectively, this finding underscores the critical role of moss crusts in sustaining soil multifunctionality of vulnerable ecosystems under extreme climate contexts, and provides new insights into soil protection. However, the limited biomass input of moss crusts and the potential imbalance in nitrogen fraction conversion are critical issues that require careful consideration. Accordingly, combining moss crusts with diverse vegetation restoration approaches may be an effective strategy for ecological restoration and carbon sequestration enhancement in fragile ecosystems affected by climate change.

Rapid SOC enhancement in infertile soils requires substantial organic inputs, which can be efficiently supplied through the granulation of agricultural wastes. However, the mechanisms of granulated organic amendments (GOA) on SOC sequestration and saturation in infertile paddy soils remain unclear. This study aimed to investigate the dynamics of SOC functional pools and saturation under different organic amendment measures. A 2-year field experiment was established including four treatments: control (no amendment), composted manure (10 t ha−1 annually), 20 and 40 t ha−1GOA at the beginning of experiment. The accumulation and stabilization of SOC under different treatments were evaluated via the separated SOC functional pools and the calculated C saturation deficit. The application of GOA markedly increased SOC content, particularly by expanding the unprotected C sub-pool (by 20

Soil pH is a key attribute regulating biogeochemical processes, with its changes significantly influenced by land use patterns and acid deposition. While extensive research has focused on the causes of soil acidification, comprehensively understanding the mechanisms and extent of pH recovery in acidified soils is crucial for developing effective ecological management strategies. This study investigated the dynamic changes in soil pH recovery through 13 years of field sampling across different land use patterns, including woodland, paddy, and upland. Results indicate that acidified soils under various land use patterns exhibit varying degrees of pH recovery trends, with an overall upward trend in pH values. Notably, forest soils exhibited the most pronounced recovery, with average pH rising from 3.70 to 4.81. This was followed by paddy and upland. The calculated reduction in free acidity confirmed this recovery sequence: woodland (-0.46 mmol/kg) > upland (-0.32 mmol/kg) > paddy (-0.225 mmol/kg). The study identified key drivers of soil pH recovery, highlighting the decisive role of nitrogen and sulfur deposition alongside fertilizer application. Among these, reduced sulfur deposition emerged as the most influential factor, explaining the largest proportion of pH recovery across all land-use patterns. Furthermore, the study explored strategies to promote sustainable recovery of acidified soils in subtropical watersheds, emphasizing the importance of emission reductions. These findings provide valuable insights for regional applications and lay the groundwork for long-term ecosystem management, particularly in mitigating pollutant emissions and restoring soil health.

Barren paddy fields characterized by poor soil structure, shallow tillage layers and low organic carbon content are a common limitation to rice production in subtropical China. As a novel approach to soil improvement, granulated organic amendments offer significant potential. Previous studies have shown that granulated straw can improve soil physicochemical properties and rapidly increase the soil organic carbon (SOC) content. However, their effects on barren paddies remain underexplored. This study evaluated four soil amendment strategies: no organic amendments (CK), 10 t ha(-1) of composted manure (M10), 20 t ha(-1) of granulated organic amendment (G20), and 40 t ha(-1) of granulated organic amendment (G40). The objective was to assess the effects of these amendments on soil structure, the contents of aggregate-associated carbon (AAC), particulate organic carbon (POC) and mineral-associated organic carbon (MAOC), and the chemical stability of MAOC among various size aggregates in both topsoil (0-20 cm) and subsoil (20-40 cm). The results demonstrated that organic amendment inputs significantly increased the macroaggregate (>250 & micro;m) proportion and improved soil structural stability. These amendments also elevated the carbon concentration within aggregates of various sizes and facilitated the redistribution of organic carbon from microaggregates (53-250 & micro;m) and silt+clay fractions (<53 & micro;m) to macroaggregates. The proportion of POC to AAC declined with decreasing aggregate size, whereas the proportion of MAOC increased. In the topsoil, macroaggregate formation enhanced the protection of POC, supported the accumulation of non-hydrolyzable carbon within MAOC, and accelerated the formation of intra-microaggregates. In the subsoil, mineral-bound organic carbon remained the dominant form of carbon sequestration. In conclusion, the application of 40 t ha(-1) of granulated organic amendment proved to be a successful tactic for enhancing soil physicochemical structure, increasing SOC content, and improving carbon stability. This approach offers a promising and innovative solution for the sustainable management and restoration of barren paddy fields.

Industrial Haber-Bosch provides 32 Tg nitrogen (N) per year to global croplands. Such large amounts of N fertilization will reshape soil N cycles. One uncertainty is whether and how these long-term N inputs impact soil biological N2 fixation (BNF) associated with plant growth, a crucial microbial-driven process to decipher the balance between soil N and plant N in paddy fields. To fill this gap, we conducted a 10-year field experiment to systematically evaluate the impacts of three N fertilization rates (0, 75 and 150 kg N ha-1) on soil BNF activity, diazotrophic communities and environmental dynamics at two typical rice growth stages (tillering and maturation) in paddy fields. Moderate N fertilization at a 75 kg N ha-1 rate significantly stimulated BNF activity and nifH gene abundance at the tillering stage, with increases of 20.9% and 49.7%, respectively. This enhancement was primarily due to the supply of available carbon (C) sources from organic acids, which promoted the BNF process to meet the N requirement from rice at the early stage. In contrast, BNF activity and nifH gene abundance decreased with increasing N levels at the maturation stages, reflecting the negative impact of N enrichment on diazotrophs. The accumulation of inorganic N and amino acids reduced the dependence on the BNF process. Stage-specific key diazotrophs, such as Rhizobium at tillering and Azoarcus at maturation, played pivotal roles in the soil BNF. Soil-dissolved organic C and inorganic N were positively correlated with BNF at the tillering stage, but negatively at the maturation stage. Random Forest analysis further revealed that the BNF was primarily regulated by available nutrients at the tillering stage, while closely associated with the key diazotrophs at the maturation stage. This study highlights that the stage-specific BNF process depends on C availability and key diazotrophs with long-term N fertilization, providing a basis for optimized N management in paddy fields.Read the free for this article on the Journal blog.

Excessive nitrogen (N) fertilization in tea plantations often leads to substantial nitrous oxide (N2O) emissions, which exacerbate global warming, and to pronounced ammonia (NH3) volatilization, which is closely associated with air pollution and aquatic eutrophication. Although N transformation inhibitors and biochar have shown promise in mitigating these gaseous losses, their combined effects and the underlying mechanisms in tea fields remain poorly understood. A 2-year field experiment was conducted in a subtropical hilly tea plantation to evaluate the individual and combined effects of dual inhibitors (the urease inhibitor N-(n-butyl) thiophosphoric triamide, NBPT, and the nitrification inhibitor 3,4-dimethylpyrazole phosphate, DMPP) and biochar (28 t ha−1) on N2O and NH3 emissions. Four treatments were established: conventional N fertilization (CON), N fertilizer amended with dual inhibitors (NI), N fertilizer combined with both biochar and dual inhibitors (BNI), and a zero-N control (CK). The results showed that the CON treatment produced high cumulative gaseous emissions (N2O: 25.8 kg ha−1; NH3: 75.8 kg ha−1). The NI treatment reduced the N2O and NH3 emission factors by 54.5

Soil nutrients are essential for ensuring high and stable crop yields,which directly determine the fertility capacity and ecological function of agricultural soils.Dynamic monitoring of soil nutrients can help reveal the evolution trends of soil fertility under long-term fertilization management.Subtropical region of China is a key rice production base,where the improvement potential and the spatial heterogeneity of paddy soil nutrients are vital to the national food security.However,insufficient sharing of soil nutrient data at the regional scale leads to the uncertainty in assessing the effectiveness of different fertilization practices in improving soil fertility.This dataset is based on 12 long-term fertilization experiments located in Hunan Province initiated between 1986 and 2011.Fertilization treatments include no fertilizer(CK),conventional chemical fertilizer(CF),nitrogen-phosphorus fertilizer(NP),nitrogen-potassium fertilizer(NK),nitrogen-phosphorus-potassium fertilizer(NPK),high-rate nitrogen-phosphorus-potassium fertilizer(hNPK),organic fertilizer replacing 30%of chemical fertilizer N(30%OM),40%organic fertilizer replaces chemical fertilizer N(40%OM),50%organic fertilizer replaces chemical fertilizer N(50%OM),60%organic fertilizer replaces chemical fertilizer N(60%OM),100%organic fertilizer(OM),and rice straw combined with NPK fertilizer(NPKS).After the late rice harvest in 2024,soil samples from the 0-20 cm soil layer were collected,and soil nutrients were analyzed according to standard methods to obtain the dataset of soil nutrients,including soil organic carbon,total nitrogen,total phosphorus,total potassium,nitrate nitrogen,ammonium nitrogen,alkali-hydrolysable nitrogen,available phosphorus,available potassium,and soil pH.The dataset comprises five parts:basic information of experimental plots,initial soil physicochemical properties,fertilizer application amounts of plots,analytical methods of soil samples,and soil nutrient contents.This dataset could provide not only data support for studying soil nutrient cycling processes and succession trends in subtropical paddy soils,but also the basic information for the fertilizer application and optimized management of regional paddy fields.

Polyethylene (PE) mulching has been widely practiced in agriculture for decades, but its short-term impacts on heavy metal dynamics and crop safety under field conditions remain poorly understood. In this study, a one-season field trial was carried out in Cd-contaminated paddy to evaluate how PE mulching influences rhizosphere microbial communities, soil physicochemical properties, and Cd accumulation in rice. Results showed that PE mulching improved rice performance, increasing dry grain weight by 14.47% and thousand-grain weight by 1.10 folds, while reducing grain Cd concentration from 0.2307 to 0.1727 mg/kg, below the national safety threshold of 0.2 mg/kg. These effects were closely linked to elevated soil pH, decreased redox potential, and the enrichment of metal-reducing (Geobacteraceae, Desulfuromonadia) and sulfate-reducing (Desulfosporosinus, Methanospirillum) taxa, which promoted Cd immobilization into less bioavailable forms. A structural equation model (SEM) further confirmed that microbial abundance and Cd speciation were key factors associated with Cd uptake by rice. However, PE mulching also reduced microbial diversity and functional redundancy, disrupted co-occurrence networks, and potentially weakened rhizosphere ecosystem stability and resilience in the short term. This study provides field-based evidence that PE mulching reduces food safety risks and improves yield but destabilizes soil microbial communities, highlighting its short-term double-edged ecological effects and the need for balanced management to sustain productivity and soil health.

Accurate prediction of forest soil organic carbon (SOC) is essential for quantifying terrestrial carbon stocks and understanding climate-carbon cycle feedbacks. However, existing approaches often fail to fully capture spatial dependence and integrate multi-scale predictive information. Here, we developed an ensemble spatially enhanced TabPFN framework (ESE-TabPFN) for predicting SOC across Chinese forest ecosystems. A progressive VIF-Boruta-RFE framework identified 10 optimal predictors from 38 environmental variables representing topography, climate, vegetation, soil properties, and anthropogenic factors. Comparison of seven machine-learning algorithms identified TabPFN as the best-performing baseline model (R2 = 0.58). Incorporating a spatial enhancement strategy integrating multi-scale KNN, IDW-based spatial lag variables, and Gaussian kernel smoothing (SE-TabPFN) increased model performance to R2 = 0.67 while reducing RMSE by 11.4%. The final multi-variant ensemble framework (ESE-TabPFN) achieved an R2 of 0.70, representing a 20.7% improvement over the baseline model. The ESE-TabPFN framework was subsequently applied to generate a 1-km resolution map of forest SOC across China, with a predicted mean value of 5.05 kg C m–2. Predicted SOC exhibited a clear zonal pattern, generally increasing from southern to northern China and with elevation. Comparisons with publicly available datasets revealed strong spatial consistency, with a Pearson correlation coefficient of 0.77. Collectively, our results demonstrated that integrating spatial enhancement with ensemble learning substantially improves SOC prediction and provides a novel framework for digital soil carbon mapping.

Ammonia (NH3) volatilization is a major nitrogen loss pathway in rice paddy ecosystems, resulting in considerable economic costs and adverse environmental hazards. Yet, the specific characteristics of NH3 volatilization, its regional mitigation potential, and associated microbial responses under combined deep fertilizer placement and urease inhibitor application remain poorly understood. To address this knowledge gap, a two-year continuous high-frequency field monitoring experiment was conducted in a subtropical double rice-cropping system. Six treatments were setting: no N input (CK), conventional N input (CON), 30 % reduced N application (RN); and under the reduced N rate: deep placement (RD, 7 cm below the soil surface), urease inhibitor application (RU), and combined deep placement with urease inhibitor (RUD). The results demonstrated that the RUD treatment reduced cumulative NH3 volatilization by 68.0-73.8 % relative to the CON treatment. This mitigation effect was attributed to suppressed urea hydrolysis rates and soil urease activity, which significantly limited soil NH4+ -N accumulation. Furthermore, the RUD treatment slightly reduced soil microbial alpha-diversity indices, weakened the active of microbial taxa related to urease production and downregulated microbial carbohydrate metabolism pathways. Based on experimental data, we estimated that nationwide adoption of the RUD practice could reduce NH3 emissions from Chinese rice fields by 528.29 Gg N yr-1. Regionally, central south of China exhibited as largest NH3 reduction after RUD treatment, which accounted for 38 % of the total NH3 reduction in China. Our findings highlight that the combined deep placement and urease inhibitor strategy as a promising approach to support sustainable agricultural development in double-rice cropping systems, and underscore the importance of prioritizing implementation in high-emission regions to maximize environmental benefits.

As an innovative technology, incorporation of massive granulated organic materials has been proven effective in boosting soil organic carbon (SOC) in infertile croplands, but their effects on the soil quality index (SQI) and multifunctionality (SMF) remain largely unclear. Here, we conducted two micro-plot experiments with the application of granulated organic materials (incorporation amounts: 0, 30, 60, 90 t ha-1; material combinations: straw, straw+manure, straw+manure+biochar) in infertile upland and paddy soils. After one year, granulated organic materials incorporation rapidly improved SQI (28.7%–132.8% in upland and 40.8%–71.0% in paddy) and SMF (13.8%–168.7% in upland and 17.9%–272.4% in paddy) in both croplands, by enhancing soil physical, chemical and microbial properties. These improvements were dependent on the incorporation amounts and material combinations. With increasing granulated straw incorporation amount, SQI linearly increased in upland soil, but increased initially and then plateaued in paddy soil, likely owing to the inhibition of excessive straw decomposition under moistirrigated condition, thereby limiting the release of carbon and nutrients. SMF increased linearly in both croplands by improving C, N, and P-acquiring enzyme activities. Under equivalent organic carbon incorporation conditions, granulated straw+manure+biochar treatment generally had the highest values of SQI and SMF in both croplands. This benefited from the good porosity and surface charge characteristics of biochar, which enhanced C, N and P retention by reducing leaching and gaseous losses. Additionally, the SQI and SMF were positively correlated with crop yield in both croplands. Thus, this innovative practice of incorporating granulated organic materials into the plough layer is an efficient way to improve soil fertility and crop yield by increasing SQI and SMF in infertile croplands.

Straw amendments can improve soil fertility by loading organic carbon (C) into soils, but whether and how biological nitrogen fixation (BNF) occurs in long-term rice straw (RS)-incorporated paddy fields remain poorly understood. To fill this gap, we explored the effects of three rates of straw incorporation (0, 3 and 6 t ha-1; RS0, RS3 and RS6) on soil BNF activities inferred from acetylene reduction assay (ARA) and diazotrophic communities at three rice growth stages (tillering, elongation, and maturation) based on a 10-year field experiment. The ARA activities increased significantly in response to increasing straw incorporation rates across all three rice growth stages, while the effect decreased as rice growth progressed. Soil BNF was associated with key diazotrophs, such as Dechloromonas, Bradyrhizobium, and Azospirillum. Straw incorporation increased diazotrophic abundance, diversity and interactions, which consequently improved soil BNF activities and rice yields. Straw incorporation increased rice yield by 12.6% in RS3 and by 15.5% in RS6 compared with the control. Structural equation models (SEMs) suggested that microbial C turnover and nitrogenase gene expression were the key factors affecting soil stage-specific BNF associated with the decomposition of straw. These results revealed that C-rich straw incorporation reconfigured soil N dynamics, enabling simultaneous improvement of soil fertility and rice yields through demand-driven BNF patterns in paddy fields.

The accurate prediction of soil organic carbon (SOC) content constitutes a critical scientific challenge for optimizing soil health management strategies, assessing global carbon sequestration potential, and addressing climate change. This study proposes a Multi-scale Convolutional Wavelet-Attention Network that integrates multi-dimensional environmental variables (meteorological conditions, topography, vegetation, and soil properties), overcoming technical limitations of conventional methods in environmental variable coupling, data noise suppression, and multi-scale feature fusion. The framework comprises three key innovations: (1) The Attention Feature Extraction Module (AFEM) with dynamic weight allocation to enhance representation of critical environmental variables; (2) The Self-attention Enhanced Discrete Wavelet Transform Module (SEWTM) that synergizes discrete wavelet transform with attention mechanisms to eliminate cross-dimensional noise interference; and (3) The Multi-scale Attention Feature Fusion Module (MAFFM) enabling deep coupling of environmental features across scales. Experimental validation in forested regions demonstrated superior prediction accuracy (test set R²=0.60, RMSE = 5.16 g/kg), outperforming conventional random forest (RF) and convolutional neural network (CNN) approaches with R² improvements of 0.08–0.14. The proposed model establishes a novel deep learning framework for high-precision digital soil mapping while providing reliable technical support for sustainable agricultural development and global carbon governance.

Evaluating the current storage of soil organic C (SOC) and its future sequestration potential helps to develop goal-oriented strategies to enlarge the soil C pool. Here, we investigated the contents of particulate and mineral-associated organic C (POC and MAOC) and assessed the C sequestration potentials as stable mineral-associated fractions in 240 triplicated topsoils collected from adjacent woodlands, uplands, and paddies across four climate zones in eastern China. SOC content was generally lower in warmer climates (subtropics and tropics) than in cooler ones (mid- and warm temperate), with a generally descending order of paddies > woodlands > upland croplands in each climate zone. The SOC contents of all soils were dominated by MAOC, whereas with a lesser proportion of MAOC in warmer climates. Compared with woodlands, cropland use reshaped SOC compositions: paddies had larger POC and MAOC pools but with reduced SOC stability, whereas upland croplands had a lower total SOC content yet with higher stability. The average C/N ratios of particulate organic matter were comparable among the three ecosystems and four climates (15 +/- 3.4), while those of mineral-associated organic matter were lower in subtropics (10 +/- 2.0) than other climatic zones (13 +/- 1.7). These differences in the C/N ratios of mineral-associated fractions are explained by the fast organic matter turnover, intensive N fertilization and strong adsorption of N containing organic compounds by Fe/Al oxides in subtropics. Regardless of climate and land-use, MAOC accumulation in all ecosystems was not saturated as assessed by the 95(th) quantile regression between the mass proportion of the clay + silt fraction and the current MAOC content. Although the average MAOC saturation degrees exceeded 60%, its wide variation within and between individual soils suggested a further substantial C sequestration potential, particularly in the mid-temperate. These findings underscore that SOC management strategies must be tailored to local land-use and climate conditions to raise stable SOC pools.

Rapidly improving infertile croplands and enhancing their soil organic carbon (SOC) pool necessitate substantial organic materials incorporation. Converting loose crop straw into granulated form facilitates uniform incorporation within the plough soil layer. As an innovative soil amelioration approach, the efficiency and patterns of SOC accumulation remain unclear. Two field experiments were conducted in infertile subtropical upland and paddy soils with 0, 30, 60, and 90 Mg ha-1 granulated straw incorporation. After one year, SOC accumulation efficiency from straw input remained stable in upland (30.8-37.5%) with increasing amounts of straw incorporation, while declined from 60.0 to 38.3% in paddy. In both croplands, the contributions of lignin phenols to SOC increased with increasing straw incorporation, while the contributions from amino sugars remained constant at higher straw input levels. Subsequently, the ratios of lignin phenols to amino sugars increased with increasing straw incorporation, indicating faster plant residue accumulation compared to microbial necromass, as the granulation approach limited microbial involvement in straw transformation. Thus, single-time incorporation of substantial granulated straw presents an effective agricultural strategy for rapid amelioration of infertile croplands.

Microorganisms carryingcbbL, pmoA and coxL genes play crucial roles in regulating soil-atmosphere exchanges of carbon trace gases (CO2, CH4, and CO). However, the geographical distribution patterns of these functional genes in agricultural ecosystems and their environmental drivers remain poorly understood. Here, we surveyed agricultural soils across four climate zones (tropical, subtropical, warm temperate, and mid-temperate) in eastern China to quantify the abundances of CO2-assimilating bacteria (cbbLgene), methanotrophs (pmoAgene), and CO-oxidizing bacteria (coxLgene). We found significant ecosystem-specific patterns: the cbbL gene was more abundant in upland soils (averaging 9.46×109copies g-1) than in paddy soils (6.44×109copies g-1). In contrast, methanotrophs abundance was 1 to 3 orders of magnitude higher in paddy (averaging 1.17×108 copies g-1) than in upland (5.78×106 copies g-1)soils. The coxL gene maintained similar abundance levels across both soil types (averaging 6.12×108 vs. 5.91×108copies g-1). Structural equation models revealed that spatial factors primarily shaped cbbL and pmoA in uplands, whereas total bacterial abundance was the dominant predictor for all three genes in paddy soils. These results highlight distinct ecological controls on microbial functional groups and provide a predictive framework for how land use and climate change may regulate microbial mediation of carbon gas fluxes across a continental-scale transect in eastern China.