Incorporation of fallow weed increases phosphorus availability in a farmer’s organic rice fields on allophanic Andosol in eastern Japan

ABSTRACT

We investigated the amount of soil phosphorus (P) in a farmer’s paddy fields under organic farming (OF) for various periods from 0 to 22 years as well as other farmers’ fields under conventional farming. All the fields are located in allophanic Andosol with long history of P fertilizer application, and some of them have been converted to OF across years. After conversion to OF, P was supplied only with winter fallow weeds mainly Foxtail (Alopecurus aequalis), rice residues (rice bran and straw), and guano. We determined total-P (Tot-P) and plant-available P (Av-P), which consists of Truog-P (Tru-P) and Bray-2-P under reducing condition with ascorbic acid (Asc-P), in soils of each field. For both Av-Ps, the ratio to Tot-P increased across years under OF following quadratic functions with both linear and quadratic terms being statistically significant. The ratios showed little changes for the initial 15 (Tru-P) or 10 (Asc-P) years and increased rapidly thereafter. These temporal changes in Av-P were consistent with the rapid increase of the amount of P accumulated in the winter fallow weeds and incorporated in the fields after beginning of OF. These results led us to the hypothesis that the incorporation of winter weeds has contributed to the increase of Av-P in the organic fields across years. We tested this hypothesis by investigating temporal changes of Av-P after suspending the weed incorporation for 2 consecutive years in plots of the organic fields. Both Av-Ps were significantly greater in plots with continued weed incorporation (CWI) than those in plots with its suspension. We further found that the increase of Asc-P in plots with CWI was 4.9-fold the input of total P in the incorporated weeds. This suggests that the incorporation of winter fallow weeds enhanced soil-P availability beyond the supply of P accumulated in the weeds.

KEY WORDS:

• Truog-P
• Bray-2-P
• phosphorus replenishment
• rhizosphere
• green manure

1. Introduction

Phosphorus (P) is one of the essential macro elements for plant growth and has been applied to agricultural lands as fertilizer. Particularly to farmlands located on Andosol in Japan, a large dose of P fertilizer has been commonly applied, since the applied P is bound to the soil and becomes hardly available for the plants. It has indeed been reported that, after substantial input of chemical P fertilizer for several years, suspension of fertilizer P input led to a decreasing trend of available P in soil (Shoji et al. 1971; Kogano 1984; Nira and Ito 2016). The heavy dose of P has led to its accumulation in soil of Andosol origin including the rice paddies (Shoji and Higuchi 1970; Imai et al. 1994). If the accumulated P in soil can be utilized for plant growth, the P input could be reduced for more efficient use of this globally limited resource (Egawa 1982). A recent study has actually shown such a possibility to utilize accumulated P by incorporating green manure in upland soils (Karasawa and Takahashi 2015).

Our previous study has shown that organic farming (OF) with little external inputs attained a rice yield comparable to that of the conventional farming, and that the fallow weed incorporation served as a major source of nitrogen for the rice production (Tanaka et al. 2012). In the literature also, attention have been given to beneficial effects of green manure on nutrient supply as well as accumulation of organic matter in soil (Cherr et al. 2006). Some studies have reported that green manure application enhanced organic P mineralization in upland soil (Cavigelli and Thien 2003; Randhawa et al. 2005). In paddy fields, however, no study has been done on the change in P bioavailability by green manure incorporation. In this article, we report on the effect of continued incorporation of fallow weeds on the soil P status of Andosol fields, where rice had been cultivated under organic practices for up to 22 years.

2. Materials and methods

2.1. The paddy fields and the farmer’s agronomic practices

The research was conducted at a farmer’s lowland paddy fields on allophanic Andosol in Nogi-town, Tochigi prefecture, Japan (latitude 36◦ 14ʹ N, longitude 139◦ 46ʹ E) for the years from 2013 to 2016. The farmer started OF in 1992 and had gradually expanded it to the other paddy fields. Under OF, rice straw was incorporated into soil in autumn, and so were rice bran and weeds that grew during off-season in the next spring. The off-season weeds were dominated by Foxtail (Alopecurus aequalis), a common fallow weed found in paddy field of Japan (Chisaka 1965; Arakawa and Abe 1984). In some fields, guano (citrate soluble P₂O₅ content of 18.3%, Surya Guano, Asunaro Corporation Ltd.) was also applied along with rice bran (P₂O₅ content of 5.66%). No synthetic fertilizers were applied, and no agrochemicals were used to control insects, diseases, or weeds. At the beginning of the cropping season in year 2013, more than 50 fields had been under OF for up to 22 years. More details of the cropping history and the application of rice bran and guano in individual fields are shown in Supplementary Table 1.

Rice fields under conventional farming (CF) were also studied for comparison with the organic fields. In the CF fields, P was applied according to the prefectural standard (Tochigi Prefectural Government 2017) as mixtures of nitrogen, phosphate, and potassium by synthetic fertilizer of various kind and some farmers used rice bran as a part of it.

2.2. Comparison of soil P contents among the fields under different periods of organic farming

We measured soil P contents in the paddy soil and related them to the length of period under OF to detect the effects, if any, of the organic farming on P availability for rice growth. The soil P was measured for 39 fields under OF for a period between 0 (the first year of OF) and 22 years as well as 8 fields under CF. In some fields, the measurement was conducted more than once during the study period from 2013 to 2016, and, hence, total number of the soil P measurements was 70 for OF fields and 14 for CF fields. More details of the soil sampling for the P measurements are given in Supplementary Table 2. The amount of P input was estimated at 61 kg P₂O₅ ha⁻¹ in OF fields with rice bran and guano application and 24 kg P₂O₅ ha⁻¹ in OF fields with only rice bran application. In CF fields, P was fertilized at a rate between 30 and 131 kg P₂O₅ ha⁻¹, with the mean being 87 kg P₂O₅ ha⁻¹ and SD 28 kg P₂O₅ ha⁻¹.

At the soil sampling, plow layer soils were collected by using hand auger from 5 points per location, and the composite samples were air-dried and ground to pass through 2 mm sieve to be used for chemical analysis. Available P was analyzed by the Truog method (Tru-P) (Truog 1930) and Bray-2 under reduced condition with ascorbic acid method (Asc-P) (Nanzyo et al. 1996). Total P content (Tot-P) was analyzed by HClO₄ acid decomposition method (Soil Standard Analysis Measurement Method Committee 2004). The outline of each method is as follows.

Tru-P: 200 mL of 0.01 mol L⁻¹ H₂SO₄ was added to 1.0 g of air-dried soil, shook for 30 min at 20°C, and then filtrate was collected. A coloring reagent, which was adjusted with ammonium molybdate liquid and ascorbic acid, was added to 10 mL of filtrate. Tru-P was determined by molybdenum blue method (Murphy and Riley 1962).

Asc-P: 1.0 g of air-dried soil was added to 10 mL of 10 g L⁻¹ ascorbic acid liquid, shook for 16 h at 20°C, and 10mL of extracting solution with 2.22 g of NH₄F in 1 L of 0.2 mol L⁻¹ HCl was added. Thereafter, shook vigorously by hand for 1 min and filtrate was collected. Asc-P was determined by molybdenum blue method.

Tot-P: 25 mL of mixed solution (H₂SO₄:HNO₃:HClO₄ = 1: 5: 20) was added to 5.0 g of air-dried soil, and the mixture was heated and pyrolyzed at 130°C until it became a syrup. Thirty milliliters of HCl and 50 mL of boiling distilled water were added after cooled down, then quickly heated to a temperature just before boiling, and the supernatant liquid was filtered. Tot-P was determined for the filtrate by molybdenum blue method.

2.3. Suspension of weed incorporation to detect changes in soil P availability

The comparison between the fields of different number of years under OF in soil P availability was augmented with an additional experiment, where soil P contents were compared between the plots with the weed incorporation being suspended for the years 2014 and 2015 and the other plots with continued weed incorporation (CWI). Details of this experiment are as follows.

On 15 May 2014, two plots, one with CWI and the other with suspended weed incorporation (SWI), were established in each of the six OF fields, of which three had been under OF for 10 years and the other three had been under OF for 18 years. The CWI and SWI plots were adjacent (0–3 m) to each other having an area of 9 m². Weed growth in SWI plots was about the same as that in CWI plots, when we set the treatment by removing all weeds with their subterranean parts being separated from the soil. The plots were maintained for the years 2014 and 2015 at the same place. In 2014, the weed biomass including the roots was determined for CWI plots by sampling from a subplot of 3 m², whereas weed biomass was not determined for SWI plots. In 2015, the weeds were sampled again on 6 May from a 1 m² subplot in each of CWI and SWI plots. The P content of weed was determined for samples taken on the same day as for the weed biomass.

Soil sampling and analysis of soil P contents were the same as those for the across-fields comparison experiment.

2.4. Comparison of P accumulation in fallow weeds among the fields under different periods of organic farming

Samples of weeds were taken from 14 fields under OF for a duration between 0 and 17 years on 11 and 12 May 2015. Three quadrats (0.7 m × 0.7 m) were set in each field, and the weeds were collected including the subterranean part to determine the dry biomass. In most fields, more than 90% of the weed biomass was that of Foxtail. Dried samples of Foxtail were subjected to the H₂SO₄-H₂O₂ decomposition (Mizuno and Minami 1980), and P content was determined by molybdenum blue method. P accumulation by the weeds was calculated by multiplying P content by the weed dry mass.

2.5. Statistical analyses

In the comparison of soil P contents between the fields under different periods of OF, a mixed linear model was fit to the measurements with the field as a random variable and the number of years of OF as a continuous variable and the application of guano as a fixed-effect variable. The effect of suspension of weed incorporation on soil P contents was analyzed with a mixed linear model of the field as a random effect, and the number of years under OF (10 y vs. 18 y) and the weed incorporation treatment (suspended vs. continued for 2 years) as fixed effects. Statistical analyses were conducted with JMP Pro ver. 13.0.0 (SAS Institute, Cary, USA).

3. Results and discussions

3.1. Comparison of soil P contents among the fields under different periods of organic farming

Total-P (Tot-P) concentrations in the fields under OF for various number of years (Fig. 1) were not significantly related to the number of years under OF (P = 0.270) or the application of guano (P= 0.404). To account for the field-to-field variability in Tot-P, the concentrations of Truog-P (Tru-P) and Bray-2-P with ascorbic acid (Asc-P) were divided by Tot-P for each field, and the ratios, Tru-P/Tot-P and Asc-P/Tot-P, were related to the number of years under OF and guano application.

Figure 1. Total P concentration in the fields under organic farming for various number of years with (filled circles) and without guano (unfilled circles) application.

For Tru-P/Tot-P, the effect of the number of years under OF was highly significant with its linear term (P = 0.0002) and the quadratic term (P = 0.0003). Since the field-to-field variability had no significant effect (P = 0.985), the relationship between Tru-P/Tot-P and the number of years under OF was described by a quadratic curve (R² = 0.355) (Fig. 2A).

Figure 2. Relationships between the number of years under organic farming and the ratio of Truog P (Tru-P) to total P (A) and that of Bray-2 P (Asc-P) to total P (B) with (filled circles) and without guano (unfilled circles) application. The dotted curves show the quadratic regression of the respective ratios on the number of years under organic farming.

The result was almost the same for Asc-P/Tot-P also. The effect of the number of years under OF was significant with its linear term (P < 0.0001) and the quadratic term (P = 0.0199). Field-to-field variability had no significant effect (P = 0.889), and the effect of the number of years under OF on Asc-P/Tot-P was described by a quadratic curve (R² = 0.417) (Fig. 2B).

Many investigations have been conducted on the transition of available P (Tru-P) after withholding P fertilizer application in Japan. One of such studies has shown that available P decreased exponentially at a rate of 6% per year for Andosol in the same prefecture (Tochigi) and soil type (Allophanic Andosol) as this study (Nira and Ito 2016).

The fields in this study also had received no synthetic P inputs but P from Guano, rice bran, and rice straw (Supplementary Table 1) in addition to P from the weeds which grew during the fallow period. The temporal increase of Tru-P/Tot-P and Asc-P/Tot-P (Fig. 2) could be attributed to either of the P inputs continued throughout the period under OF. Since the P accumulation in the weeds increased across the years after starting OF (Fig. 3), P supply from the incorporated weeds should have also increased gradually. If most P in the incorporated weeds joined the pool of plant-available P (Av-P), e.g., Tru-P and Asc-P, the Av-Ps would have increased at an increasing rate until the saturation of weed P amount (Fig. 3), which is consistent to the quadratic relationships between the ratios, Tru-P/Tot-P and Asc-P/Tot-P, with the number of years under OF (Fig. 2). We could therefore hypothesize that the incorporation of winter weeds has contributed to the gradual increase of the fraction of Av-P in Tot-P in OF fields. It must be noted, however, that contribution of the other P inputs such as guano to Av-Ps may not be disregarded. For example, a linear model of the accumulated P input from guano and that from rice bran (Supplementary Table 1) explains the variability of Tru-P/Tot-P between the fields equally well (R² = 0.353) as the model of the number of years under OF (R² = 0.355) as noted earlier.

Figure 3. Relationship between the number of years under organic farming and the amount of P in Foxtail, the dominant weed species. Circles are observations, and the curve is the model fitted to the observations.

It is indeed noteworthy that the relationships between Av-Ps and the number of years (Fig. 2) may not actually describe a temporal evolution across years of OF but the spatial variation among different fields under OF for different number of years. The above-noted hypothesis would therefore be supported only if the spatial comparison is interchangeable with the time evolution. Besides the uncertainty of the time-space interchangeability assumption, it remains uncertain if the amount of P input from the weed incorporation is large enough to account for the apparent increase of Av-P observed in the comparison between the fields.

The above-noted hypothesis along with the uncertainties was addressed by tracking soil P contents in the fields where the weed incorporation was suspended for 2 consective years.

3.2. Suspension of weed incorporation to detect changes in soil P availability

Results of the suspension of weed incorporation are shown in Table 1. On average, across the plots where the weed incorporation was suspended for 2 years, Tru-P concentration was 607 mg P₂O₅ kg⁻¹, which was less than 642 mg P₂O₅ kg⁻¹ for the plots with CWI. The effect of suspension of weed incorporation on Tru-P was significant (P = 0.0346), whereas the effect of the number of years under OF (P = 0.304) or its interaction with the weed incorporation treatment (P = 0.726) was not. Omitting the nonsignificant effects from the mixed linear model, the difference in Tru-P between the continued and SWI plots was estimated to be 35.0 ± 26.0 mg P₂O₅ kg⁻¹ with 95% confidence.

Table 1. Effect of suspension of weed incorporation in paddy fields on soil Tru-P and Asc-P. (mg P₂O₅ kg⁻¹).

			Tru-P^c		Asc-P
Years of organic farming	Field ID^d	Treatment^a		Difference between CWI and SWI		Difference between CWI and SWI	Change after treatment
18		Before^b	NA		4140
	8	CWI	626		4233		92
		SWI	605	21.0	4013	219.5	−127
		Before	NA		4530
	9	CWI	731		4735		205
		SWI	690	41.3	4379	356.6	−151
		Before	NA		4295
	10	CWI	719		4606		311
		SWI	689	30.1	4263	342.9	−32
10	51	CWI	449		2985
		SWI	367	81.7	2515	470.5	NA
	52	CWI	679		4403
		SWI	656	22.9	4316	86.5	NA
	53	CWI	649		4320
		SWI	636	12.9	4089	230.9	NA
	Mean of the fields	CWI	642		4214
		SWI	607	35.0	3929	284.5

^aTreatment was conducted in 2014 and 2015 rice seasons.

^bSoil sampling was done before treatment in October 2013, and soil sampling was done after treatment in October 2015. CWI: continuous weed incorporation, SWI: suspended weed incorporation.

^cTru-P denotes Truog-P. NA denotes not available. Asc-P denotes Bray-2-P after reduction by ascorbic acid.

^dHistory of the organic farming practices in the respective fields can be found in Supplementary Table 1.

The results were even clearer for Asc-P concentration, which was 3929 mg P₂O₅ kg⁻¹ for plots with SWI and 4214 mg P₂O₅ kg⁻¹ for those with CWI. The effect of suspension of weed incorporation on Asc-P was highly significant (P = 0.0090), whereas that of the number of years under OF (P = 0.320) or its interaction with the weed incorporation treatment (P = 0.735) was not. Omitting the nonsignificant effects from the mixed linear model, the difference in Asc-P between the continued and SWI plots was estimated to be 284.5 ± 140.2 mg P₂O₅ kg⁻¹ with 95% confidence.

It is noteworthy that Asc-P concentration was increased by the CWI and decreased by its suspension for the 2 years as compared to those before the treatments (Table 1). Such comparison is infeasible for Tru-P, however, because of the lack of observations in Tru-P before the treatment.

The above results support the hypothesis that the weed incorporation contributed to the increase in Av-P without the uncertainties mentioned earlier in the comparison of soil P contents among the fields under different periods of OF (3.1).

It must be noted that the timing of soil sampling for Asc-P was not consistent across years (Supplementary Table 2). Samples were taken after harvest in some years, whereas they were taken before the cultivation in other years. The difference in the sampling time could have affected the Asc-P measurements as reported by Saito et al. (2007), who showed that the aerobic soil condition after the harvest increased Bray-2-P as compared to that during the period of submergence. However, statistical testing with the mixed linear model with the random effect of field and the fixed effects of the soil sampling time, the number of years under OF, and the weed incorporation treatment on Asc-P showed a nonsignificant effect of the soil sampling timing (P = 0.623). This may have resulted from the prior treatment of soil with ascorbic acid which reduced some P compound once oxidized to its reduced form before Bray-2-P determination.

Some studies have reported that green manure application or incorporation of crop residues enhanced P bioavailability in upland soils (Cavigelli and Thien 2003; Hirata et al. 1999; Karasawa et al. 2015; Randhawa et al. 2005). Among the preceding studies, Hirata et al. (1999) should be the closest to this study as a long-term field experiment. They observed that incorporation of crop residues into upland soil of Andosol origin for over 9 years attenuated the differences in crop growth between the plots with and without continuous fertilizer P application. They also observed that Aluminum-bound P (Al-P) relative to Tot-P increased in quadratic manner, and that Al-P positively correlated with plant available P (Bray-2-P). From these results, Hirata et al. (1999) suggested that the increase in Al-P is driven by the changes in microbial biomass P under the continuous supply of organic matter from plant residues. Our results showed similar temporal changes for Tru-P/Tot-P and Asc-P/Tot-P in lowland paddy fields of Andosol origin with the fallow weed incorporated for many years. A common mechanism may be assumed in this and Hirata et al. (1999) studies for the increase in P availability.

3.3. Comparison of P accumulation in fallow weeds among the fields under different periods of organic farming

Amount of P in the biomass of Foxtail, the dominant weed species, plotted against the number of years under OF demonstrated a growth curve with rapid increase at the beginning and saturation after several years (Fig. 3). The apparent increase in the P amount at the beginning was due to the increase of the weed biomass rather than that of P concentration in plant with the latter showing no significant trend across years under OF (P = 0.114).

We further tested the hypothesis on the contribution of the weed incorporation to Av-P in soil by comparing the changes in Av-P with the amount of P input due to the weed incorporation. When the changes in Av-P concentrations in soil are converted to the amount of Av-P on land area basis, it is evident that the amount of P replenished in the soil was much greater (4.9-fold increase on average) than that contained in the incorporated weeds (Table 2). This points to an effect of the incorporated weeds (predominantly Foxtail) on soil P availability beyond its supply of P accumulated in the biomass.

Table 2. Increase of available P relative to P inputs due to the continued weed incorporation as compared with the plots with suspended weed incorporation for 2 years (2014–2015).

Field ID	Biomass of weeds		P content of weeds		P input from weeds for 2 years	Increase of available P^a		Ratio of available P increase to P input from weeds
	2014	2015	2014	2015		Tru-P^b	Asc-P^c
	g m^−2		%P2O5		mg P2O5 m^−2	mg P2O5 m^−2		Tru-P	Asc-P
8	275	518	0.65	0.63	5051	2103	21951	0.4	4.3
9	323	472	0.65	0.62	5011	4130	35655	0.8	7.1
10	383	434	0.66	0.75	5755	3013	34288	0.5	6.0
51	809	474	0.55	0.56	7154	8169	47045	1.1	6.6
52	502	446	0.61	0.66	6014	2294	8646	0.4	1.4
53	483	463	0.60	0.59	5621	1288	23091	0.2	4.1
MEAN	463	468	0.62	0.63	5768	3500	28446	0.6	4.9
SD	191	29	0.04	0.06	786	2479	13379	0.3	2.1

^aIncrease of available P on area basis was calculated from the difference in Av-P concentrations between the plots with continued weed incorporation and those in suspended weed incorporation plot on the assumptions of 15 cm plow layer soil depth and bulk density of 0.67.

^bTru-P denotes Truog-P.

^cAsc-P denotes Bray-2-P after reduction by ascorbic acid.

Randhawa et al. (2005) planted maize in soil amended with green manure from lupin (Lupinus angustifolius L.) at a rate equivalent to 45 mg P kg⁻¹ (15 g DW kg⁻¹). After repeating the cycle of soil amendment and maize planting for three times, they found a 5-fold increase in organic P mineralization as compared to that in unamended soil. Their finding is comparable to the 4.9-fold increase in available P (Asc-P) as compared to the amount of P input from Foxtail in this study.

Cavigelli and Thien (2003) measured P uptake of sorghum (Sorghum bicolor L.) plants grown in potted soils that were incorporated with either of the four cover crops: white lupin (Lupinus albus L.), pea (Pisum sativum L.), hairy vetch (Vicia villosa Roth.), or wheat (Triticum aestivum L.) at a rate from 0.7 to 1.9 mg P kg⁻¹. They also measured P uptake by the sorghum from soil and soil Bray-1 P before and after incorporating the cover crops. White lupin showed only a 1.4-fold increase in available P in soil compared to the P input by plant, whereas the other cover crops showed from 4 to 5.7-fold increase in available P. Although white lupin took up from 1.9 to 2.7 as much P from the soil than other cover crops, the increase of available P in soil was less than that in the other cover crops. From this discrepancy, they concluded that available P was not suitable to predict P supply potential to the following crops.

In comparison to the above studies, the amount of P incorporated by Foxtail in this study was 12 mg P kg⁻¹ on average as calculated from Table 2, and was about a quarter of that in Randhawa et al.’s (2005) experiment but 10 times that of Cavigelli and Thien’s (2003) experiment. In the latter experiment, the input and depletion of P may be too small to detect changes in soil available P. In comparison, the large amount of P incorporation in Randhawa et al.’s (2005) experiment suggests the enhancement of P mineralization of organic P in this study. It is possible that Foxtail has an ability to mobilize recalcitrant P in soil by specific reactions such as excretion of chelating substances in the rhizosphere. To explore such a possibility, studies are needed to compare the soil P fractions before and after planting Foxtail.

Other mechanisms could also account for the increase of bioavailability of P by weed incorporation. Tsutsuki and Ponnamperuma (1987) reported that a large amount of volatile fatty acids was formed under anaerobic incubation with green manure. The organic anions could compete with phosphate for sorption to soil particles as reported by Hernandez et al. (1986), and thereby increase bioavailability of P.

It must be noted, however, that earlier studies have been done with upland soils, and that subsequent chemical reactions of mineralized P with Ca, active Al, and active Fe in the paddy soil in this study could differ from those in the upland soils. Further studies are warranted to understand the mechanism(s) of the P replenishment by fallow weeds incorporation in organic rice farming.

Acknowledgments

The authors gratefully acknowledge the sincere cooperation of Mr. Hiroyuki Tateno of Nogi-town in Tochigi Prefecture for this study. We also thank Ms. Yoko Hoshino, Ms. Misao Akutsu, Mr. Kohei Hachisu, and Mr. Rintaro Yuki of Tochigi Prefectural Agricultural Experiment Station for their assistance in laboratory and field work. The authors’ appreciation is further extended to Dr. Satoshi Nakamura of Japan International Research Center for Agricultural Sciences for his valuable comments on a draft of this manuscript.

Supplementary material

Supplemental data for this article can be accessed here

Additional information

Funding

This work was supported by the Japan Society for the Promotion of Science Grant-in-Aid for Scientific Research Grant Number (B) 26310304.