3.1 Effects of CS and G1 application on soil physical and chemical properties
Figure 1 showed that CS and G1 had significant effects on the soil bulk density, specific gravity, and porosity. The soil bulk density in B1 to B8 groups was significantly increased by 40.78–54.47%, compared with that in the control B0 (p = 0.000, P < 0.001). The soil specific gravity in B1 to B8 was 4.14–16.00% lower than that in the control group (B0), respectively, (P > 0.05), but the porosity in B1 to B8 was 17.17–24.39% higher than that in B0 group (p = 0.001, P < 0.05). Of all the treatment groups, B5 group exhibited the optimal bulk density, specific gravity, and porosity, and its bulk density reached up to 1.25 g·cm− 3, which was close to China national second soil census nutrient classification standards. The bulk density in the control B0 was only 0.87 g·cm− 3, which was far lower than the average level of local farmland soil (1.26 g·cm− 3) (LI et al. 2019). Bulk density in B1 group was lower than that in B0 group, which might be due to the fact that B1 group was added with G1 but with no CS, and humic acid, as one main component of G1 amendment, significantly increased the content of soil organic matter, thus decreasing soil bulk density (LI et al. 2019).
Generally, the aggregates with particle size of > 0.25 mm were defined as macroaggregates. The particle size of macroaggregates in all the groups except control group was within the range of 2-0.25 mm. Figure 2 showed that soil macroaggregates (> 0.25mm) in B5 group accounted for 34.2%, and those in the control B0 accounted for 32.6%. However, the percentage of the macroaggregates (with particle size of 0.25-5mm) in B5 group was 31.9%, which was the highest in all the treatment groups, while it was the lowest in B0 group (22.74%). One recent study has shown that soil nutrients are negatively correlated with the percentage of aggregates with particle size of > 5mm (Jiang et al. 2023). Our results showed that the addition of the CS and G1 effectively increased the percentage of the 0.25-2 mm aggregates in the soil, which was very important for soil drought resistance and soil moisture conservation (CHEN et al. 2019; Ju et al. 2023).
The soil chemical properties were shown in Table 2. Soil pH in all treatment groups was not significantly decreased, relative to that in B0 (p = 0.124, P > 0.01). Meanwhile, the soil available P in B1 to B8 groups was significantly increased by 13.95–55.69%, compared with that in B0, respectively (p = 0.000, P < 0.01), which showed G1 and CS had slow release on available P. And B5 exhibited the highest soil available P (124.39 mg·kg− 1). The available N in B1 to B8 groups was significantly decreased by 5.34–50.67%, compared with that in B0 (p = 0.048, P < 0.05). The reason might be that the available N in the soil was absorbed and fixed by the crops. The soil available K in B1 to B8 groups was significantly increased by 1.33–388.37% compared with that in B0, respectively (p = 0.000, P < 0.01). The soil organic matter in B2 to B8 groups was decreased by 35.71%, to 90.63%, compared with that in B0, while B1 was not decreased. It was suggesting that CS could decease soil organic matter content, respectively (p = 0.0002, P < 0.01). Only B1 exhibited an increase in soil organic matter (by 50.89%), indicating that G1 could increase soil organic matter concentration.
Table 2
Properties | B0 | B1 | B2 | B3 | B4 | B5 | B6 | B7 | B8 |
pH | 7.76 ± 0.0125a | 7.76 ± 0.028a | 7.67 ± 0.012ab | 7.58 ± 0.012ab | 7.59 ± 025ab | 7.57 ± 0.025b | 7.71 ± 0.01ab | 7.74 ± 0.03ab | 7.75 ± 0.005ab |
AN (mg kg− 1) | 72.60 ± 1.29a | 68.72 ± 1.69a | 60.71 ± 3.95b | 60.93 ± 1.81b | 38.64 ± 2.27d | 35.81 ± 1.52d | 46.47 ± 1.8c | 63.41 ± 2.8ab | 64.38 ± 1.7ab |
AP (mg kg− 1) | 28.07 ± 1.24e | 31.98 ± 1.24cd | 35.20 ± 1.27cd | 43.70 ± 2.93b | 41.00 ± 3.01b | 152.46 ± 2.95a | 28.19 ± 1.9e | 38.64 ± 1.0bc | 29.10 ± 3.3cd |
AK (mg kg− 1) | 125.46 ± 1.20f | 160.32 ± 4.96e | 148.64 ± 3.53ef | 328.26 ± 15.26b | 258.78 ± 17.45c | 610.08 ± 41.21a | 131.68 ± 25.7f | 208.15 ± 11.1d | 226.86 ± 18.8cd |
TN (g kg− 1) | 1.33 ± 0.043a | 1.36 ± 0.113a | 0.80 ± 0.011c | 0.70 ± 0.020 cd | 0.38 ± 0.009 f | 1.06 ± 0.031b | 0.53 ± 0.002de | 0.56 ± 0.003de | 0.60 ± 0.002 de |
TP (g kg− 1) | 0.92 ± 0.03a | 0.84 ± 0.002ab | 0.65 ± 0.001bc | 0.69 ± 0.001b | 0.61 ± 0.003bc | 0.82 ± 0.002a | 0.58 ± 0.001c | 0.67 ± 0.009bc | 0.63 ± 0.005bc |
TK (g kg− 1) | 1.30 ± 0.037c | 1.36 ± 0.004c | 1.77 ± 0.021b | 1.92 ± 0.003ab | 2.05 ± 0.001a | 1.95 ± 0.010ab | 1.92 ± 0.004a | 1.87 ± 0.003ab | 1.94 ± 0.01a |
OM(g− 1kg) | 10.67 ± 0.08b | 16.10 ± 0.97a | 6.86 ± 0.036c | 4.76 ± 0.045d | 1.00 ± 0.059f | 3.64 ± 0.041d | 3.09 ± 0.03e | 3.45 ± 0.01de | 4.33 ± 0.04de |
Notes: TN, total nitrogen; TP, total phosphorus; TK, total potassium; AN, available nitrogen; AK, available potassium. AP, available phosphorus. Different lower-case letters indicate significant differences (P < 0.05) among different samples based on a one-way ANOVA, followed by an LSD test. |
3.2 Bacterial α-diversity and taxonomy of bacterial communities
After chimeras was removed, the sequence data were analyzed at the OTU level. A total of 3,857 OTUs were generated from 9 samples sequenced on an Illumina MiSeq, and the number of OTUs ranged from 2,215 to 2,771 when sequence data were clustered at 97% similarity (Table S3). The coverage ranged from 97.01% (B3) to 97.47% (B7), indicating that more than 97% of the bacterial species were detected from each sample (Fig. 3a). These results suggested that the data collected were adequate to capture the diversity of the bacteria associated with each sample (Fig. 3b). The ACE and the Chao1 indexes were calculated to estimate the bacterial abundance. As shown in Fig. 3c and e, the overall bacterial abundance in the B3 estimated were higher than other treatments. Among all the groups treatments, it was showed that bacterial diversity was the highest in treatment B3 while it was the lowest diversity in treatment B4 (Fig. 3d and f). It is suggesting that the bacterial α-diversity was strongly influenced by the addition of CS.
Nine dominant phylogenetical phyla were identified including Proteobacteria (26.46–37.18%), Actinobacteria (16.36–23.47%), Chloroflexi (14.63–19.22%), Acidobacteria (7.93–17.89%), Gemmatimonadetes (3.25–4.97%), Bacteroidetes (1.51–4.28%), Nitrospirae (1.36–3.47%), Firmicutes (0.77–3.34%), and Planctomycetes (0.51–1.68%) (Fig. 4). The soil bacterial communities differed slightly between treatments and control. The abundance of Armatimonadetes (0.39–0.53%), Bacteroidetes (1.74–4.28%), and Cyanobacteria (0.36–4.27%) in soil samples of treatment groups was increased, compared with that in soil sample of control group (0.34%, 1.71%, and 0.33%, respectively). The abundance of Chlamydiae (0.01–0.05%), FCPU426 (0.00-0.01%), Ignavibacteriae (0.00-0.05%), Latescibacteria (0.04–0.51%), Nitrospirae (1.35–3.21%), Omnitrophica (0.01–0.09%), Parcubacteria (0.05–0.34%), Planctomycetes (0.51–1.59%), and RBG-1__Zixibacteria (0.00-0.09%) in treatment groups was decreased, compared with that in B0 (0.08%, 0.02%, 0.08%, 0.59%, 3.47, 0.19%, 0.38%,1.68%, and 0.26%, respectively).
A hierarchical cluster tree of the bacterial communities was constructed by the UPGMA at a 97% OUT similarity level. This tree showed that the bacterial communities were clustered into four distinct clades (Fig. 5a). Clade 1 contained bacterial communities from the sample of B4, Clade 2 included those from B5 and B7. Clade 3 included those from B3 and B8, and Clade 4 included those from B0, B1, B2 and B6. Afterward, a principal coordinate analysis (PCoA) of the major bacterial clades was performed with 46.77% of the observed variation explained. As Fig. 5b showed, B4 sample was separated from other 8 samples, and B0 was located on the left of the graph along PC1, and B4 was located on the right of the graph along PC1, whereas the other seven samples were distributed in the middle of the graph between B0 and B4. Bacterial sequences were assigned to a total of 585 classified and unclassified genera. The 50 genera with highest abundance were displayed in a heatmap (Fig. 5c), which revealed complex patterns in the genera abundances across samples. Some of the genera, such as Acidobacteria (77.05%), Sphingomonas (25.60%), Nitrospira (20.78%), Streptomyces (11.32%), and Gaiella (10.20%), were abundant in all 9 samples, with a total abundance of 10.20–77.05%.
3.3 Effects of soil amendments on maize growth in coal fly ash soil
After the maize samples were harvested, the root dry weight, the plant height, the stem dry weight, and the corncob dry weight were determined (Fig. 6a and b), respectively. The results showed that the root dry weight was significantly decreased by 7.98%-68.90% in B1 and B3-B8, respectively, while B2 was significantly increased by 52.59%, compared with that in B0 (P = 0.004, P < 0.01) (Fig. 6a). The root moisture content was significantly increased by 5.65%, 8.04%, 14.91%, 47.77%, 52.91%, 24.68%, and 53.15% in B1, B3, B4, B5, B6, B7, and B8, respectively, whereas that of B2 was significantly decreased by 10.38%, compared with that in B0 (P = 0.002, P < 0.01) (Fig. 6a). The result showed that the addition of G1 could promote the growth and water retention ability of plant roots. The plant height was significantly increased by 5.31%-12.24% in B1-B8, respectively, compared with that in B0 (P = 0.007, P < 0.01) (Fig. 6b). The corncob dry weight was significantly increased by 26.30%-76.37% in B1-B6 and B8, respectively, whereas B7 was significantly decreased by 3.81%, compared with that in B0 (P = 0.004, P < 0.01) (Fig. 6b). The stem dry weight was significantly decreased by 23.16%, 0.77%, 12.55%, 17.94%, 22.03%, 35.80%, 18.52%, and 25.24% in B1, B2, B3, B4, B5, B6, B7, and B8, respectively, compared with that in B0 (P = 0.000, P < 0.01) (Fig. 6b).
3.4 Effect of soil amendments on maize yield in coal fly ash soil
The corn kernel number per spike and hundred grain weight were shown in Fig. 7. The hundred grain weight was significantly increased by 9.60%, 22.31%, 23.01%, 22.80%, 32.17%, 10.04%, and 1.33% in B1, B2, B3, B4, B5, B6, and B8, respectively, while hundred grain weight in B7 was significantly decreased by 19.26%, compared with that in B0 (p = 0.027, P < 0.05). The kernel number per spike was significantly increased by 15.73%, 23.52%, 29.49%, 25.50%, 47.20%, 42.39%, 19.21%, and 23.52% in B1, B2, B3, B4, B5, B6, B7, and B8, respectively, compared with that in B0 (p = 0.018, P < 0.05). The values of hundred grain weight and kernel number per spike in B5 were higher than for the other groups, which were 32.17% and 47.2% higher than those in the control group, respectively.