Soil properties along the tea plantations
The temperatures at the sampling times, November 8, 2016, January 5, 2017, March 15, 2017, and May 10, 2017, were 26.5°C, 21.9°C, 15.1°C, and 29.7°C, respectively. The soil properties are presented in Fig. 1. The SA soil had the highest overall pH (5.41) followed by TA and CA soils, while the trend was reversed in EC of the soil with SA having the lowest EC (75.2 µS cm− 1). Total nitrogen and organic matter contents were the highest in TA soil followed by SA and CA soils. The extractable elements namely, P, K, Ca, Mg, Mn, Cu, and Zn in the SA soil were higher than that of the TA soil, followed by the CA soil (Fig. 2). Pearson correlation analysis indicated that soil pH was positively correlated with soil P and K contents and negatively correlated with soil EC (Fig. 3). Also, a significant correlation was observed between organic matter and total N.
Soil Enzymatic Activities Among The Tea Plantations
The activities of the enzymes, namely acid phosphatase, arylsulfatase, β-glucosidase, and urease, and nitrogen fixation involved in phosphorus, sulfur, carbon, and nitrogen cycles are presented in Fig. 4. Acid phosphatase in the TA soil was significantly higher than SA and CA soils (p < 0.001). Arylsulfatase and β-glucosidase activities were significantly higher in the SA soil (p < 0.001) than TA and CA soils. Though urease activity did not show significant difference between the management practices (p = 0.003), the activity of CA soil was 4.3-times higher than that of SA soil in January 2017 sample. Nitrogen fixation was not significantly different across the agricultural practices; however, it was comparatively higher it the SA soil.
Repeated measures-ANOVA for the enzymatic activities was performed to evaluate the interaction between temporal changes and agricultural management practices (Table 1). Acid phosphatase, arylsulfatase and urease activities showed significant difference between agricultural management practices and time. The β-glucosidase activity showed marginally significant differences across sampling times and agricultural management practice. However, in the nitrogen fixing activity, there were no significant changes in agricultural management systems along with the temporal scale.
Table 1
Statistical significance (p-value) for repeated measures ANOVA on acid phosphatase, arylsulfatase, β-glucosidase, urease, and nitrogen fixing activity
Model term
|
Enzymatic activity
|
Acid phosphatase
|
Arylsulfatase
|
β-Glucosidase
|
Urease
|
Nitrogen fixation
|
F
|
p
|
F
|
p
|
F
|
p
|
F
|
p
|
F
|
p
|
Test of within-subjects effects
|
|
|
|
|
|
|
|
|
|
|
Time
|
53.709
|
0.000**
|
7.899
|
0.001**
|
8.476
|
0.009**
|
80.172
|
0.000**
|
2.283
|
0.174
|
Time × Management
|
3.246
|
0.016*
|
3.510
|
0.011*
|
2.918
|
0.081
|
18.060
|
0.001**
|
0.703
|
0.552
|
Test of between-subjects effects
|
|
|
|
|
|
|
|
|
|
|
Intercept
|
6109.184
|
0.000**
|
4911.200
|
0.000**
|
1282.411
|
0.000**
|
243.964
|
0.000**
|
12.303
|
0.013*
|
Management
|
1297.194
|
0.000**
|
1116.072
|
0.000**
|
85.677
|
0.000**
|
7.665
|
0.022*
|
2.102
|
0.203
|
Significance is indicated by **p-value < 0.01, and *p-value < 0.05 |
Changes In The Soil Bacterial Community Structure
To estimate the diversity of bacterial distribution in soil, the total eubacterial community was examined using high-throughput sequencing. After raw reads were trimmed and chimeric reads were removed, the average clean reads and clean data of each sample were 108,342 reads and 31.125 Mb, respectively (Table S1). The OTUs obtained for the different agricultural management practices and sampling times are represented as a Venn diagram (Fig. S2). The number of unique OTUs in the SA soil was significantly higher than those of the CA and TA soils. Based on the sampling time, the order of presence of unique OTUs is May 2017 > November 2016 > March 2017 > January 2017. PCA revealed that the TA soil clustered closely with the CA soil but was distinct from the SA soil (57.26%) (Fig. 5). However, no grouping was observed for the sampling times.
The soil community structure at the phylum level showed that Proteobacteria, Acidobacteria, Actinobacteria, Chloroflexi, and Bacteroidetes were the predominant phyla (Fig. 6). The relative abundance of Bacteroidetes in the SA soil was 7.3 and 6.0-fold higher than that in the CA and TA soils, respectively and the differences are statistically significant (p < 0.01; p < 0.05). Gemmatimonadetes and Nitrospirae were also higher in SA soil compared to soils with other management practices. By contrast, the relative abundance of Actinobacteria in the SA soil was significantly lower than the TA and CA soils (p < 0.01; p < 0.001). The abundance of Chloroflexi was significantly lower in SA soil compared to CA soil (p < 0.05). The distribution of the major phyla among the management practices and sampling times are given as supplementary figure (Fig. S3). Further, at the class level Gammaproteobacteria was dominant having a relative abundance of 17.3% and 15.1% in the CA and SA soils respectively. In the TA soil, Alphaproteobacteria was dominant with a relative abundance of 13.2%. The heat map showing the clustering based on the abundance of each phylum is shown in Fig. 6. Three clades are clearly divided and of the SA soil is clustered in the same group. The CA and TA soils sampled in November 2016 and May 2017, and January 2017 and March 2017 showed the distribution of the bacterial community in two distinct clades.
Bacterial species abundance and diversity estimates such as observed specie richness, SOBS, Chao, ACE, Shannon, and Simpson diversity indices across the management and sampling times are shown in Fig. 7. The SA soil had a higher number of bacterial species and greater community richness than the soils managed under TA and CA practices. This finding is consistent with the species diversity displaying a high Shannon and low Simpson indices in the SA soil. The bacterial species and community richness did not differ considerably with the sampling time.
Correlation between bacterial community, soil enzymatic activity, and soil properties
Redundancy analysis was used to explore the relationship between bacterial community, soil enzymatic activities and soil properties. The RDA components explain 88.19%, 81.62%, and 99.46% of the variation in enzymatic activity, bacterial composition, and bacterial alpha diversity, respectively from the data obtained for the soil samples (Fig. 8). For the enzymatic activity, the first component explained 56.8% of the total variation and also, separated the CA, TA, and SA samples. Arylsulfatase activity was dominant in the SA soil and was related to the pH changes, while the acid phosphatase activity was predominant in the TA soil and is associated with the total nitrogen and organic matter of the soil. On the other hand, urease activity was dominant in the CA soil and was related to the EC value. Further, for the bacterial community, an association was seen between available phosphorus, soil pH and the abundance of Acidobacteria. Additionally, Chloroflexi and Actinobacteria were positively correlated with total nitrogen and organic matter; however, they were negatively associated with the soil EC. According to their close groupings, Acidobacteria and Proteobacteria occurred predominantly in the TA soil, while Chloroflexi and Actinobacteria were predominant in all the soils in November 2016. The observed species richness of bacterial community clustered along RDA axis-2, which is highly correlated with soil pH, K and P and negatively correlated to soil EC.
Pearson correlation between enzymatic activity, bacterial community, and soil properties is shown in Table 2. A significant difference was observed between different enzymatic activity and soil properties. Moreover, the observed species and abundance indices including SOBS, Chao, and ACE were significantly related to soil pH. Nevertheless, no significant relationship was perceived in the dominant bacteria and bacterial diversity such as Acidobacteria, Actinibacteria, Bacteroidetes, Proteobacteria, Shannon, and Simpson index. Chloroflexi showed a significant negative correlation (p-value < 0.05) with EC. Multivariate linear regression ANOVA with stepwise method was performed for modeling the enzymatic activity and bacterial community with soil chemical properties. To identify the major factors that influence the enzymatic activity, observed species, and bacterial abundance, the soil properties were used as independent variables for the stepwise multiple regressions. Based on this the regressions are constructed (Eq. 1–6) as follows:
Acid phosphatase = − 0.461 + 0.123 × organic matter………………….. (1)
(Model: r2 = 0.676, F = 20.888, p-value = 0.001)
Arylsulfatase = − 0.227 + 0.07 × pH……………………………………. (2)
(Model: r2 = 0.963, F = 260.678, p-value = 0.000)
β-Glucosidase = − 0.023 + 0.001 × K…………………………………...(3)
(Model: r2 = 0.646, F = 18.276, p-value = 0.002)
SOBS = 26.327 + 301 × pH…………………………………………....(4)
(Model: r2 = 0.725, F = 26.423, p-value = 0.000)
Chao = 117.262 + 330.738 × pH……………………………………….(5)
(Model: r2 = 0.761, F = 31.837, p-value = 0.000)
ACE = 181.788 + 317.184 × pH………………………………………(6)
(Model: r2 = 0.759, F = 31.486, p-value = 0.000)
Table 2
Pearson correlation between enzymatic activity, bacterial community, and soil properties.
Pearson correlation
|
pH
|
EC
|
Organic matter
|
Total nitrogen
|
P
|
K
|
Acid phosphatase
|
–0.157
|
0.004
|
0.822**
|
0.757**
|
0.619*
|
–0.037
|
Arylsulfatase
|
0.981**
|
–0.663**
|
–0.209
|
–0.124
|
0.655*
|
0.962**
|
β-Glucosidase
|
0.784**
|
–0.463
|
0.077
|
0.190
|
0.761**
|
0.804**
|
Urease
|
–0.207
|
0.447
|
–0.199
|
–0.156
|
–0.193
|
–0.222
|
Acidobacteria
|
0.293
|
0.019
|
–0.124
|
–0.044
|
0.376
|
0.290
|
Actinobacteria
|
–0.254
|
–0.148
|
0.159
|
0.130
|
–0.141
|
–0.348
|
Bacteroidetes
|
0.414
|
–0.089
|
–0.282
|
–0.298
|
0.202
|
0.508
|
Chloroflexi
|
0.020
|
–0.559*
|
0.190
|
0.238
|
–0.128
|
–0.079
|
Proteobacteria
|
–0.325
|
0.308
|
0.115
|
0.076
|
–0.323
|
–0.326
|
SOBS
|
0.852**
|
–0.591*
|
–0.279
|
–0.196
|
0.434
|
0.808**
|
Chao
|
0.872**
|
–0.577*
|
–0.235
|
–0.132
|
0.505*
|
0.832**
|
ACE
|
0.871**
|
–0.599*
|
–0.223
|
–0.127
|
0.501*
|
0.823**
|
Shannon
|
0.381
|
–0.396
|
–0.042
|
–0.019
|
0.295
|
0.438
|
Simpson
|
0.361
|
–0.094
|
–0.144
|
–0.106
|
0.141
|
0.246
|
Significance is indicated by **p-value < 0.01, and *p-value < 0.05 |