A 10-year-long field trial (between 2001 and 2010) was conducted to investigate the effect of paddy-upland rotation on rice yield, soil properties, and bacteria community diversity. Six types of paddy-upland crop rotations were evaluated: rice-fallow (control; CK), rice-rye grass (RR), rice-potato with rice straw mulches (RP), rice-rapeseed with straw incorporated into soil at flowering (ROF), rice-rapeseed incorporated in soil after harvest (ROM), and rice-Chinese milk vetch (RC). Analysis of terminal restriction fragment length polymorphism (T-RFLP) was used to determine microbial diversity among rotations. Rice yield increased for upland crops planted during the winter. RC had the highest average yield of 7.74 t/ha, followed by RR, RP, ROM, and ROF. Soil quality differences among rotations were found. RC and RP improved the soil mean weight diameter (MWD), which suggested that rice rotated with milk vetch and potato might improve the paddy soil structure. Improved total nitrogen (TN) and soil organic matter (SOM) were also found in RC and RP. The positive relationship between yield and TN/SOM might provide evidence for the effect of RC rotation on rice yield. A strong time dependency of soil bacterial community diversity was also found.
In China and other Asian countries, continuous rice planting has had a negative impact on soil properties, such as reduced soil nitrogen supply and organic carbon content [
In conventional paddy-upland rotation systems, farmers drain the fields after harvesting rice and then plant an upland crop, such as milk vetch, wheat, or oilseed [
Soil quality is a term used to describe the health of agricultural soils. It has been suggested as an indicator for evaluating sustainability of soil and crop management practices [
Soil physical properties are indicators of the impact of soil and crop management practices. Soil size distribution and water stability of soil aggregates would be influenced by crop types as well as soil management practices [
In China, paddy-upland crop rotation is a major cropping system utilized along the Yangtze River basin [
The study was conducted over a period of ten years at the experimental farm of the China Rice Research Institute (120.2 E, 30.3 N, and 11 m above sea level) located in Fuyang, Hangzhou, China. The long-term field experiment has been carried out in an irrigated rice paddy starting in 2001. The historical cropping background was monoculture rice cropping before this time. The area is characterized by a subtropical monsoon climate with an annual mean temperature of 13–20°C, ranging from 2°C in January to 35°C in July and mean precipitation of 1200–1600 mm per year, with about 80% falling between April and September. Soil is classified as Ferric-accumulic Stagnic Anthrosols [
The long-term field experiment included six different types of cropping rotation: continuous monoculture rice-fallow (CK), rice-ryegrass (
Experimental design of six paddy-upland crop rotation systems.
Label | Cropping system | Year of start of experiment | Rice season | Upland crop season | Herbicides | Annual N-P-K input Kg/ha | ||||
---|---|---|---|---|---|---|---|---|---|---|
Sowing | TP | Harvest | Sowing | Harvest | Rice | Upland crop season | ||||
CK | Rice-fallow phase | Continues | Early May | Early June | Late October | — | — | Yes | No | |
RR | Rice-ryegrass | 2001 | November | Harvest 2-3 times | Yes | b2001–2005: 180-50-180 | No | |||
RP | Rice-potato, with rice straw as mulch | 2001 | December | March | No | No | ||||
RC | Rice-Chinese milk vetch | 2001 | November | — | No | 2005–by now: 120-50-120 | No | |||
ROM | Rice-oilseed rape with burned straw return after harvest | 2001 | November | April | No | c20-20-20 | ||||
aROF | Rice-oilseed rape with fresh straw return as green manure at flowering | 2003 | November | March | No | No |
aOilseed rape was considered as manure crops but economic crops in ROF. This is a new cropping rotation under evaluating.
bFertilizer input was reduced due to the lodging problem of rice season in RP and RC treatment.
cFertilizer applied at bolting stage for oilseed rape.
The field experiment was planned with a large plot area (20 m long, 20 m wide) but without treatment replicates due to practical reasons. Although this design could lead to statistical problems, the cropping system has been used for decades with uniform fertilizer management. The small variability (CV < 5%) of the representative parameters of soil fertility in the treatment plots (Table
General soil properties before the experiment started in 2001 (0–20 cm soil depth).
Treatment | pH | SOM | TN | TP | aAvailable K | Bulk density |
---|---|---|---|---|---|---|
(g/kg) | (g/kg) | (g/kg) | (g/kg) | (g/cm3) | ||
CK | 6.51 | 26.90 | 2.53 | 0.62 | 0.22 | 1.17 |
RP | 6.69 | 24.50 | 2.49 | 0.65 | 0.24 | 1.21 |
RR | 6.82 | 25.60 | 2.50 | 0.63 | 0.23 | 1.12 |
RC | 6.62 | 29.10 | 2.46 | 0.66 | 0.23 | 1.24 |
ROM | 6.43 | 27.50 | 2.69 | 0.61 | 0.22 | 1.11 |
ROF | — | — | — | — | — | — |
Means | 6.61 | 26.72 | 2.53 | 0.64 | 0.23 | 1.17 |
bCV% | 2.30 | 4.81 | 3.32 | 3.08 | 3.61 | 4.85 |
aAvailable K was extracted with 1 mol NH4AC.
bC.V. is coefficient of variation.
Abbreviations: CK: rice-fallow phase; RR: rice-ryegrass; RP: rice-potato, with rice straw as mulch; RC: rice-Chinese milk vetch; ROM: rice-oilseed rape with burned straw return after harvest; ROF: rice-oilseed rape with fresh straw return as mulch at flowering.
In 2010, a survey of rice yield was carried out as follows: hills were harvested from the uniform part of each plot of 5 m2 with 4 subsamples. Unhulled (rough) rice was obtained after reaping, threshing, and wind selection. Rough rice from 80 hills was hulled and then put through a 1.8 mm sieve to remove any immature kernels. The weight of hulled rice (brown rice) was adjusted to a moisture content of 14%. Upland crop biomass was sampled from 1 m2 of four sub-sample plots. For RC, RP, ROM, and ROF, plants were sampled at harvest. For RR, sampling was carried on during the growth season. All samples were oven-dried at 70°C to a constant weight to determine the dry weight.
Soil samples from 0 to 20 cm depth at pretransplanting in early June 2010 were used for soil physical quality analysis. Bulk density was determined for undisturbed soil samples using a steel cylinder of 100 cm3 volume (5 cm in diameter, and 5.1 cm in height) [
Soil samples for chemical analysis were collected at pretransplanting in early June 2010 from two depths (0–10 and 10–20 cm). The field-moist soil samples were passed through an 8 mm sieve and air-dried. Cation exchange capacity (CEC) was measured according to the procedure used by Hendershot et al. [
Soil samples for T-RFLP analysis were collected from each plot using a soil auger (5 cm in diameter) at pretransplanting in early June 2010 (BR) and after harvest in late October 2010 (AR) from 0–20 cm soil depth. Samples were packed in sterile plastic bags and sent to the laboratory, then air-dried until the water content was about 75%. Later, the moist soil was passed through a 2 mm sieve and stored at 4°C for DNA extraction.
Genomic DNA of the soil samples was isolated using a SDS-hyperhaline buffer solution as used in Zhou et al. [
The eubacterial primers 8f (5′-AGAGTTTGATCCTGGCTCAG-3′) labeled at the 5’ end with 6-carboxyfluorescein (6-FAM) and 926r (5′-CCGTCAATTCCTTTRAGTTT-3′) were used to amplify approximately 920 bp of the 16S rRNA gene [
Data were analyzed by using Microsoft Excel 2003 and SAS 8.0 (2003). Means and standard deviations/standard errors are reported for each of the measurements. One-way analysis of variance (ANOVA) of Tukey’s test was used to compare the effects of rotations on soil properties determined for the two soil depths of 0–10 cm and 10–20 cm separately.
All T-RFLP community profiles were labeled for statistical analyses by rotation (CK, RR, RP, RC, ROM, or ROF), sampling time (BR or AR), restriction enzyme (HaeIII, HhaI, or HinfI), and field plot replicate (1, 2, or 3). T-RF peaks between 35 and 500 bp and peak heights of <50 fluorescence units were included in the analysis according to the range of the size marker. Generally, the error for determining fragment sizes with our automated DNA sequencer was less than 1 bp; however, in some cases, a higher variation was found. Therefore, T-RFs that differed by less than 1.5 bp were clustered unless individual peaks were detected in a reproducible manner. Three replicate samples of all rotations and particle sizes were analyzed individually, or a representative sample profile was determined in a way similar to that suggested by Dunbar et al. [
In order to determine similarities between T-RFLP profiles, a binary matrix that recorded the absence and presence of aligned fragments was generated. The distance matrix of fragments was generated according to the Jaccard index (1908) using NTSYS version 2.10e software for PC (Applied Biostatistics). Based on the distance matrix, cluster analysis was performed utilizing an unweighted pair group method with arithmetic average (UPGMA).
Rice yield increased in the plots with upland crops applied during the winter season (Figure
Effect of different paddy-upland crops rotations on the rice yield (a) and upland crops biomass production (b) during 2010-2011. Error bars represent standard deviation,
The soil bulk density, soil aggregation, and mean weight diameter (MWD) of different paddy-upland crop rotations are presented in Table
Soil physical properties in 0–20 cm depth.
Rotation | aSand | Slit | Clay | MWD | Bulk density |
---|---|---|---|---|---|
(g/kg) | (g/kg) | (g/kg) | (mm) | (g/cm3) | |
CK | 40.9b | 650.9 | 300.0b | 0.07b | 1.20b |
RC | 89.4a | 574.5 | 336.1b | 0.11a | 1.39a |
RR | 51.4b | 559.2 | 389.4a | 0.08b | 1.26ab |
RP | 76.2a | 586.2 | 337.5b | 0.10a | 1.21b |
ROM | 54.6b | 551.5 | 393.9a | 0.08b | 1.21b |
ROF | 42.6b | 531.4 | 426.0a | 0.06b | 1.35a |
Means on the same column and for the same sampling time followed by the same letter (or none) are not significantly different at
There was a strong depth-dependency of soil pH value, total soil nitrogen (TN), total soil phosphorus (TP), available potassium (K), and cation exchange capacity (CEC) in all rotations (see Table
Soil chemical properties in 0–10 cm depth.
Depth | Rotation | pH | Total N | Total P | Available K | CEC |
---|---|---|---|---|---|---|
(1: 2.5 |
(g/kg) | (g/kg) | (g/kg) | (Cmol/kg) | ||
0–10 cm | CK | 5.78a | 2.46b | 0.53b | 0.25b | 10.73b |
RC | 5.48b | 2.92a | 0.59b | 0.28ab | 12.23a | |
ROF | 5.21c | 2.82a | 0.70a | 0.30a | 10.97b | |
ROM | 5.64a | 2.56b | 0.59b | 0.33a | 11.92a | |
RP | 5.15c | 2.96a | 0.72a | 0.31a | 10.80b | |
RR | 5.40b | 2.67b | 0.67a | 0.24b | 11.23b | |
| ||||||
10–20 cm | CK | 5.86b | 2.36b | 0.47 | 0.22b | 10.26c |
RC | 5.82b | 2.61ab | 0.49 | 0.23b | 13.50a | |
ROF | 6.00a | 2.52b | 0.54 | 0.23b | 11.41b | |
ROM | 5.96a | 2.41b | 0.48 | 0.30a | 11.43b | |
RP | 5.84b | 2.70a | 0.53 | 0.25b | 11.01b | |
RR | 5.89b | 2.48b | 0.54 | 0.21b | 10.34c | |
| ||||||
Comparison of depth | 0–10 cm | 5.44b | 2.73a | 0.63a | 0.29 | 11.31 |
10–20 cm | 5.90a | 2.51b | 0.51b | 0.24 | 11.33 | |
| ||||||
ANOVA | Rotation | ** | ** | ** | ** | ** |
Depth | ** | ** | ** | ns | ns | |
|
* | * | * | ** | ** |
Means on the same column and for the same sampling time followed by the same letter (or none) are not significantly different at
For TN and TP, remarkable differences were found between depths, with 0–10 cm greater than 10–20 cm. The average values of TN/TP were 2.73/0.63 (g/kg) in 0–10 cm and 2.51/0.51 (g/kg) in 10–20 cm. However, the difference of available K between depths was not statistically significant. The rotation effect on soil TN, TP, and available K was significant. In comparison with CK, the values of TN were significantly increased in RC, ROF, and RP in 0–10 cm, with average increments of 18.7%, 14.6%, and 20.3%, respectively. However, similar results in 1–20 cm were only found in RC and RP, with average increments of 10.6% and 14.4%, respectively. There was a variation of rotation effects on TP in 0–10 cm, with ROF (0.70 g/kg), RP (0.72 g/kg), and RR (0.67 g/kg) significantly greater than CK (0.53 g/kg). However, little difference in TP was found in 10–20 cm depth among all rotations. The available K was greater in ROF, ROM, and RP compared with CK in 0–10 cm depth. However, little difference was found among rotations in 10–20 cm depth, except for ROM, which had the highest value among the six rotations. The rotation effect on soil CEC was variable with soil depth. For 0–10 cm depth, CEC in RC and ROM was significantly greater than that in CK, but only RC had significantly different results in 10–20 cm depth.
Soil organic matter (SOM), soil dissolved organic carbon content (DOC), and soil microbial biomass carbon content (MBC) of the six rotations in two soil depths were analyzed and are shown in Table
Soil organic carbon content, dissolved organic content, and soil microbial organic content in 0–10 and 10–20 cm depth
Rotation | SOM | DOC | MBC | SOM | DOC | MBC |
---|---|---|---|---|---|---|
(g/kg) | (g/kg) | (g/kg) | (g/kg) | (mg/kg) | (g/kg) | |
0–10 cm | 10–20 cm | |||||
| ||||||
CK | 21.6b | 0.08b | 0.76b | 22.2ab | 0.07b | 0.64b |
RC | 25.8a | 0.16a | 1.08a | 21.4b | 0.13a | 1.11a |
ROF | 20.3bc | 0.15a | 1.22a | 16.6c | 0.12a | 1.08a |
ROM | 18.9c | 0.09b | 0.51b | 17.1c | 0.15a | 0.44b |
RP | 23.7a | 0.20a | 0.57b | 23.4a | 0.08b | 0.60b |
RR | 21.3b | 0.19a | 1.09a | 18.6c | 0.14a | 1.04a |
| ||||||
Comparison of depth | ||||||
0–10 cm | 21.93a | 0.15 | 0.87 | |||
10–20 cm | 19.88b | 0.11 | 0.82 | |||
| ||||||
ANOVA | ||||||
Rotation | ** | ** | ** | |||
Depth | ** | ns | ns | |||
|
** | ** | ** |
Means on the same column and for the same sampling time followed by the same letter (or none) are not significantly different at
Terminal restriction fragment length polymorphism analysis of 16S rRNA gene fragments amplified from community DNA was applied to compare the bacterial community structure in the field sites described above. Consistent T-RFLP profiles were obtained from three sampling points of the same field site, as shown by respective replications of the six different rotations (Figure
Terminal restriction fragment length polymorphism profiles of soil bacterial communities derived from six different rotation fields before transplanting (BR): CK (rice-fallow phase), RR (rice-ryegrass), RP (rice-potato, with rice straw as mulch), RC (rice-Chinese milk vetch), ROM (rice-oilseed rape with burned straw return after harvest), and ROF (rice-oilseed rape with fresh straw return as mulch at flowering) are shown. Terminal fragments were generated by a HaeIII digestion of 16S rRNA gene fragments amplified from total community DNA. Selected terminal restriction fragments differing in their relative abundance between the studied sites are indicated.
As shown by cluster analysis, a total of eight rotations formed mainly two major separate branches: pretransplanting (BR) and postharvest (AR) (Figure
UPGMA dendrogram generated from all representative T-RFLP sample profiles. The scale indicated the coefficient between soil from paddy-upland crop rotation systems sampled at postharvest (AR) and pretransplanting (BR); Abbreviations: CK, rice-fallow phase; RR, rice-ryegrass; RP, rice-potato, with rice straw as mulch; RC, rice-Chinese milk vetch; ROM, rice-oilseed rape with burned straw return after harvest; ROF, rice-oilseed rape with fresh straw return as mulch at flowering.
Rice yield increased when upland crops were applied during the winter season (Figure
In paddy-upland crop rotations, soil puddling in advance of transplanting can foster high productivity [
Significant soil chemical quality diversity was found among rotations. Soil pH value was acid in the paddy field, which was consistent in our results. Application of ROM increased the pH value as well as available K compared with CK. This might be due to the use of rapeseed straw burned into ash and applied as fertilizer. Biomass ash is considered as a potassium fertilizer in China. Furthermore, improved TN and SOM were found in RC and RP. The positive relationship between yield and TN and SOM might provide the evidence for the positive effect of RC rotation on rice yield.
Soil quality is different for different crop species [
Soil bacterial communities were significantly affected by soil type and plant species as well as environmental factors. As shown in Figure
Significant differences in soil chemicals (i.e., soil pH, TN, CEC, SOM, and MBC), physical properties (soil bulk density), and soil bacterial communities were detected between cropping seasons within the year (rice and upland crops season), irrespective of different winter upland crop species. Rice-Chinese milk vetch and rice-rapeseed rotations improved the soil quality to some extent, which might result in the greatest yield performance in rice-Chinese milk vetch rotations among the tested rotations. Soil bacterial communities in CK and RP were remarkably different from those in the other rotations according to the T-RFLP of 16S rRNA genes.
This research was supported by grants from the “Five-twelfth” National Science and Technology Support Program (2011BAD16B14) and the National Natural Science Foundation of China (31171502).