The spatial characteristics, pollution sources, and risks of soil heavy metals were analyzed on Hainan Island. The results showed that the heavily polluted points accounted for 0.56%, and the number of mildly and above polluted points accounted for 15.27%, respectively, which were mainly distributed in the northern part of the study area. The principal component analysis–absolute principal component score–multiple linear regression (APCS-MLR) and the positive matrix factorization (PMF) revealed four sources of heavy metals: agricultural pollution sources for cadmium, (Cd), industrial and mining pollution sources for arsenic, (As), transportation pollution sources for zinc and lead (Zn and Pb), and natural pollution sources for chromium, nickel, and copper (Cr, Ni, and Cu). The human health risk assessment indicated that the average non-carcinogenic risk (HI) for both adults and children was within the safe threshold (<1), whereas Cr and Ni posed a carcinogenic risk (CR) to human health. In addition, the total non-carcinogenic risk (THI) indicated that heavy metals posed a potential non-carcinogenic risk to children, while the total carcinogenic risk (TCR) remained relatively high, mainly in the northern part of the study area. The results of the Monte Carlo simulation showed that the non-carcinogenic risk (HI) for all heavy metals was 1, indicating a potential health risk above the safe threshold. Meanwhile, nearly 100% and 99.94% of the TCR values exceeded 1 × 10⁻⁴ for children and adults, indicating that Cr and Ni are priority heavy metals for control. The research results provide the necessary scientific basis for the prevention and control of heavy metals in agricultural soils.

Conventional fertilization practices can lead to many ecological problems, such as nutrient imbalance, soil acidity, and reduced soil fertility, in natural rubber plantations. To address these challenges, a field investigation was strategically carried out to substitute inorganic fertilizer with organic fertilizer, consisting of six treatments: no fertilization (CK), inorganic fertilizer (NPK), 25% replacement of inorganic through organic (25% manure (M)), 50% replacement of inorganic through organic (50% manure (M)), 75% replacement of inorganic through organic (75% manure (M)), and 100% organic fertilizer (100% manure). The soil physicochemical properties (soil organic carbon (SOC), total nitrogen (TN), mineral nitrogen (N), ammonium nitrogen (NH₄⁺-N), and nitrate nitrogen (NO₃⁻-N)), C:N, pH, and the carbon- and nitrogen-converting enzymes β-1,4-glucosidase (BG), N-acetylglucosaminidase (NAG) and L-leucine aminopeptidase (LAP) were all determined. The partial substitution of inorganic fertilizer with organic fertilizer (i.e., 75% M at surface soil layer) showed higher SOC (14.52 g·kg⁻¹), TN (1.06 g·kg⁻¹), N (20.07 mg·kg⁻¹), C:N (14.63), NH₄⁺-N (10.63 mg·kg⁻¹), and NO₃⁻-N (11.06 mg·kg⁻¹) than NPK and CK. This increase in physicochemical properties after partial replacement of inorganic with organic fertilizer resulted from higher carbon and nitrogen enzyme activities (BG (143.17·nmol·g⁻¹·h⁻¹), NAG (153.96 nmol·g⁻¹·h⁻¹), and LAP (153.48 nmol·g⁻¹·h⁻¹)) compared to NPK and CK. Further, the Pearson correlation and redundancy analysis (RDA) analyses confirmed a significant positive correlation between SOC, N, and soil enzymes. This study presents a new strategy for assessing the impact of partially replacing inorganic fertilizer with organic fertilizer in rubber plantations in tropical regions, mainly by modifying the soil nutrient composition.

The structural characteristics of the rhizosphere soil’s microbial community is crucial to understanding the ecological function of mangroves. However, the mechanism influencing mangrove plants in soil microbial communities has yet to be determined. Here, the mangrove ecosystem of Xinying Mangrove National Wetland Park in Hainan Province was taken as the research object. The microbial communities, external regulatory factors, and the relationship between communities were analyzed using 16S rRNA high-throughput sequencing in the rhizosphere and non-rhizosphere sediments of mangrove forests under different spatiotemporal conditions. The results showed that there was no significant difference in the α-diversity of the bacterial community between the rhizosphere and non-rhizosphere sediments. However, β-diversity was significantly different. Redundancy analysis (RDA) showed that other environmental factors besides sulfide and Fe²⁺ affected the bacterial community structure in sediments. The co-occurrence pattern analysis of bacteria in the mangrove ecosystem indicates that the bacteria in rhizosphere sediments were more closely related than those in non-rhizosphere sediments. The results reveal significant differences between the rhizosphere and non-rhizosphere bacterial community diversity, structure, and their interaction in the mangrove ecosystem. Therefore, the ecological system of the mangrove wetland needs to be preserved and rehabilitated, which would have a tremendous impact on the sustainable development.

Agroforestry ecosystems are an efficient strategy for enhancing soil nutrient conditions and sustainable agricultural development. Soil extracellular enzymes (EEAs) are important drivers of biogeochemical processes. However, changes in EEAs and chemometrics in rubber-based agroforestry systems and their mechanisms of action are still not fully understood. Distribution of EEAs, enzymatic stoichiometry, and microbial nutrient limitation characteristics of rubber plantations under seven planting patterns (RM, rubber monoculture system; AOM, Hevea brasiliensis-Alpinia oxyphylla Miq; PAR, Hevea brasiliensis-Pandanus amaryllifolius Roxb; AKH, Hevea brasiliensis-Alpinia katsumadai Hayata; CAA, Hevea brasiliensis-Coffea Arabica; CCA, Hevea brasiliensis-Cinnamomum cassia (L.) D. Don, and TCA, Hevea brasiliensis-Theobroma Cacao) were analyzed to investigate the metabolic limitations of microorganisms and to identify the primary determinants that restrict nutrient limitation. Compared with rubber monoculture systems, agroforestry ecosystems show increased carbon (C)-acquiring enzyme (EEA_C), nitrogen (N)-acquiring enzyme (EEA_N), and phosphorus (P)-acquiring enzyme (EEA_P) activities. The ecoenzymatic stoichiometry model demonstrated that all seven plantation patterns experienced C and N limitation. Compared to the rubber monoculture system, all agroforestry systems exacerbated the microbial limitations of C and N by reducing the vector angle and increasing vector length. P limitation was not detected in any plantation pattern. In agroforestry systems, progression from herbs to shrubs to trees through intercropping results in a reduction in soil microbial nutrient constraints. This is primarily because of the accumulation of litter and root biomass in tree-based systems, which enhances the soil nutrient content (e.g., soil organic carbon, total nitrogen, total phosphorus, and ammonium nitrogen) and accessibility. Conversely, as soil depth increased, microbial nutrient limitations tended to become more pronounced. Partial least squares path modelling (PLS-PM) indicated that nutrient ratios and soil total nutrient content were the most important factors influencing microbial C limitation (−0.46 and 0.40) and N limitation (−0.30 and −0.42). This study presented novel evidence regarding the constraints and drivers of soil microbial metabolism in rubber agroforestry systems. Considering the constraints of soil nutrients and microbial metabolism, intercropping of rubber trees with arboreal species is recommended over that of herbaceous species to better suit the soil environment of rubber plantation areas on Hainan Island.

Soil carbon (C) distribution, which is deeply correlated with soil extracellular enzyme activity and its temperature sensitivity (Q₁₀), are important for predicting the soil organic carbon (SOC) pools under climate warming. However, the high spatial heterogeneity in tropical forest soils makes the predicted results vary significantly. Thus, a total of 87 soil samples of three forest types, eight slope orientations, and four vegetation types were selected from Wuzhishan in Hainan Island, China. SOC distribution, the activities of six soil enzymes, and Q₁₀ at 10, 20, 30, and 40 °C were investigated to supplement the tropical data. The results showed that: (1) SOC ranged from 24.82 to 87.72 g/kg. SOC in the primary forest was significantly higher than that of the secondary forest, and SOC of the cloud forests and hilltop scrub at higher elevations was significantly higher than that of the lowland rainforests and montane rainforests at lower elevations. However, the amount of recalcitrant carbon in the primary forest was the lowest. (2) Under lab temperature conditions, the acid phosphatase and β-1,4-glucosidase in the plantation forest were significantly higher than that of the secondary forest, and the polyphenol oxidase and catalase in the plantation forest were significantly higher than that of the primary forest. Enzyme stoichiometry analysis indicated that microbials were limited by nitrogen in the study area. The six soil enzyme activities were strongly correlated with SOC and total nitrogen. (3) The Q₁₀ of soil enzymes ranged from 0.61 to 1.92 under three temperature gradients. Most hydrolases enzymes (Q₁₀ > 1) showed a positive response with temperatures from 10 to 30 °C, and showed a negative response (Q₁₀ < 1) with temperatures from 30 to 40 °C. We concluded that the negative response of enzyme Q₁₀ with global warming would slow down the SOC decomposition. Primary tropical forests could still sequester SOC; however, their ability to do so may be vulnerable to climate change, as the amount of soil C quality index was low.