Tải bản đầy đủ - 0 (trang)
VII. Effective Management of Legume Nitrogen

VII. Effective Management of Legume Nitrogen

Tải bản đầy đủ - 0trang



2.3 t/ha S. rostrata green manure (76 kg N/ha) and transplanting on a

poorly buffered Aeric Palequult in northeast Thailand. Organic acids and

toxins, which accumulate during anaerobic decomposition of organic matter, can retard root elongation, shoot growth, and nutrient uptake by rice

(Cannell and Lynch, 1984). The adverse effects of organic acid on rice

seedlings increase with decreasing soil pH (Tanaka and Navasero, 1967;

Rao and Mikkelsen, 1977).

On many tropical soils, a 1- to 2-week period for anaerobic decomposition of incorporated green manure before transplanting can increase or

at least not be detrimental to rice yield. A 1-week anaerobic decomposition

of incorporated Sesbania green manure significantly increased rice yield

on a sodic soil in India with four levels of exchangeable sodium percentage

ranging from 16 to 48 (Swamp, 1988). A 5-day decomposition for incorporated S . aculeata green manure, but not for Sesbania sp. PL Se-17 green

manure, under flooded soil conditions significantly increased rice yield on

a sandy loam with electrical conductivity of I .7 dS/m (Ghai et al., 1988).A

15-day decomposition period following incorporation of S. aculeata green

manure made no significant difference on rice yield on a reclaimed saline

sodic clay loam in India (Tiwari et ul., 1980).

On flooded clay soils in the Philippines, initial growth of rice is occasionally retarded when transplanting immediately follows incorporation of

readily decomposable organic matter. Diekmann (1990) observed that S.

rostrata and A . afrasperu green manure incorporated one day before

transplanting significantly lowered initial rice growth rate as compared

with treatments receiving no N or only urea. Rice plants eventually recovered, and grain yield appeared unaffected by the early suppression of rice

growth. On the other hand, similar rates of green manure incorporated

immediately before broadcasting germinated seeds had no visible effect on

growth of the rice. Rice plants are known to compensate for retarded

growth following organic-matter addition, such that the reduction in grain

yield is negligible or much less than the reduction in early plant growth

(Cannell and Lynch, 1984). Nonetheless, it appears that more research is

merited on the factors responsible for initial retarded growth of transplanted but not broadcast-seeded rice and on the probable benefits of a

delay between incorporation of green manure and transplanting.

Literature suggests that the maximum contribution of legume N to rice

in the tropics occurs when soils are maintained anaerobic after incorporation of legume N. The relatively small contribution of incorporated legume

biomass to rice sown on aerobic soils (Chapman and Myers, 1987) might be

attributed to loss of legume N by denitrification or leaching. In temperate

regions, the maximum contribution of legume N to rice likely occurs when

the period between legume incorporation and permanent soil flooding is



insufficient for conversion of appreciable legume N to nitrate (Williams

and Finfrock, 1962).

Leguminous green manures in the tropics, particularly stem-nodulating

legumes, with an adequate supply of soil nutrients can accumulate more

than 120 kg N/ha (Table IV) when grown for about 50 or more days before

wet-season rice. Incorporation of such large quantities of green manure N

can exceed the initial N requirements for wet-season rice, which is frequently limited in yield potential by low solar radiation in the monsoon

season. Diekmann (1990) found that application of green manure N at high

rates (103 to 190 kg N/ha) resulted in no more yield of wet-season irrigated

rice in the Philippines than did application of 60 kg urea Nlha. Similarly,

increasing aboveground N accumulation of S . rostrata from 103 to 143 kg

N/ha or from 194 to 252 kg N/ha with fertilization of the S. rostrata had no

effect on yield of the following wet-season rice (Becker et al., 1990a).

Furoc and Moms (1989) concluded from field studies in the Philippines

that the additional N (approximately 60 kg N/ha) accumulated by S.

rostrata and S . cannabina between 48 and 60 days may offer little potential

to increase wet-season rice yield beyond that obtained with the 48-day-old

green manure.


Legume N remaining after the removal of grain or all aboveground plant

biomass will normally only partially meet the N requirements for a following high-yielding rice crop. Similarly, leguminous green manure N, except

when large quantities are applied before rice with limited potential for

responding to N (Becker et al., 1990a),will normally only partially substitute for industrial N fertilizer for the following rice. In a survey of current

green manure management practices in South, Southeast, and East Asia,

Garrity and Flinn (1988) found no case in which green manure substituted

entirely for industrial N fertilizer. Therefore, use of legumes as an N

source for lowland rice will typically require integrated use of industrial N

fertilizer for sustained high rice yields. The effectiveness of integrated

organic and industrial N fertilizer use depends on the magnitude of N

losses, the timing of industrial N fertilizer, and the sources of N applied.



Several pot and field experiments have shown that application of leguminous green manure with urea to flooded soil can reduce but not eliminate

loss of urea N. Application of vetch straw with urea reduced urea N losses,



as determined by "N balance, from 19 to 12% in a pot experiment (Huang

and Broadbent, 1989). Application of milk vetch reduced loss of urea I5N

from 59 to 46% in another pot experiment (Mo and Qian, 1983). In a field

study with transplanted rice in the Philippines (Diekmann et al., 1991),

basal incorporation of 30 kg S. rostrata N/ha with 60 kg urea N/ha reduced

loss of urea 'N from 54 to 46%. Basal incorporation of 30 kg S. rosfruta

N/ha with 30 kg urea Niha reduced loss of urea "N from 37 to 24%.

Biswas and De Datta (1988) reported that losses of urea "N (60kg N/ha

basal and 30 kg N/ha topdressed at 5 to 7 days before panicle initiation)

were reduced from 35 to 26% when one-half (30 kg N/ha) the basal urea

application was replaced with an equivalent quantity of cowpea green

manure N. However, the reduced loss of urea N reported by Biswas and

De Datta (1988) might also be due to decreased percentage loss of urea N

with decreasing urea application rate, as observed by Diekmann et al.


Results of Biswas (1988) and Diekmann (1990) indicate that ammonia

volatilization was the N loss mechanism reduced by application of green

manure. Ammonia volatilization is widely recognized as an important

mechanism for loss of industrial N fertilizer applied to tropical lowland rice

(Fillery and Vlek, 1986; De Datta e f al., 1989). High floodwater ammoniacal N concentrations following application of N fertilizer, high temperature common in the tropics, and elevated floodwater pH resulting from

photosynthetic activity create a favorable environment for ammonia loss.

Application of green manure with urea reduced floodwater pH (Fig. 7) and

Floodwater pH

















No Nab











Days after fertilizer application

FIG.7. Floodwater pH at 1300 to 1400 h as affected by Sesbania rostruta green manure

(Sr) and prilled urea (PU) incorporated immediately before transplanting rice in the Philippines. N60, 60 kg N/ha; N!N, 90 kg N/ha. (Adapted from Diekmann, 1990.)



pNH3 (Pa)








Days after fertilizer application

FIG.8. Partial pressure of ammonia (pNH3) at 1300 to I400 h as affected by Sesbaniu

rmtrata green manure (Sr) and prilled urea (PU) incorporated immediately before transplant-

ing rice in the Philippines. N30,30 kg Nfha; N60,60 kg N/ha; N90,90 kg N/ha. (Adapted from

Diekmann, 1990.)

consequently the partial pressure of ammonia (Fig. 8). The lower floodwater pH was attributed to production of carbon dioxide during decomposition of green manure (Biswas, 1988; Diekmann, 1990).

In other studies, when urea was applied after incorporation of green

manure, losses of urea N were not reduced. Goswami e? al. (1988) found

that the S. aculeata green manure incorporated 1 week before transplanting had no effect on loss of urea applied in equal doses at 10 and 30 days

after transplanting (DT). Nitrogen losses from 60 kg urea "N/ha were 44%

when rice followed a fallow and 42% when rice followed S. aculeata

incorporated as a green manure. Corresponding N losses for 120 kg urea

N/ha were 39% for each treatment. John et al. (1989b) similarly found no

effect of either cowpea green manure or residue, incorporated 15 days

before transplanting, on losses of urea applied to rice. Partial pressure of

ammonia following application of urea was also not affected by either

green manure or residue incorporated 29 days earlier.



Because the rate of carbon dioxide production decreases rapidly during

decomposition of legume green manure in tropical flooded soil (Beri et al.,

1989b), the period for effective reduction of ammonia loss is probably

brief. It is uncertain whether urea application can be briefly delayed after

green manure incorporation without reducing the beneficial effect of lowered ammonia loss. Nonetheless, the reported reductions in ammonia loss

with green manure application are small or moderate. These reductions in

N loss per se have not been demonstrated to significantly increase rice


In addition to loss by ammonia volatilization, inorganic N fertilizers can

also be lost by denitrification, which is controlled in flooded soils by supply

of available C and rate of nitrate formation. Application of leguminous

plant material could then conceivably increase loss of fertilizer N in soils

where available C limits denitrification. John et al. (1989a) reported that

incorporation of cowpea green manure (C/N = 15) to a flooded clay soil 15

days before transplanting had no effect on total N loss and evolution of N2

and N2O from urea N applied 30 days later at 15 DT. On the basis of high

levels of water-soluble organic C (101 mg C/kg in the top 15-cm soil layer)

in plots not receiving green manure, very low floodwater nitrate following

urea application, and very low evolution of urea N as NZand NzO, John et

al. (1989a) concluded that nitrate formation rather than available C limited

denitrification on this flooded clay.

Because nitrification in flooded soils is restricted to a small aerobic zone,

loss of added industrial fertilizer N by denitrification in most flooded rice

soils is probably limited by nitrate formation and nitrate diffusion to

anaerobic soil (Reddy and Patrick, 1986). The negligible soil nitrate levels

and the disappearance of added nitrate without application of a C source

observed in many flooded soils support this conclusion (Buresh and De

Datta, 1990). Carbon would be expected to gain importance as the factor

limiting denitrification when soils are low in available C and when flooded

soils undergo periodic drainage and drying.



Because the response of rice to N fertilizer depends on source and

timing of N application, calculations of industrial N fertilizer equivalence

for legume N depend upon the industrial N fertilizer treatment employed

by researchers. Thorough interpretation of N fertilizer equivalence values,

such as those shown in Tables VI and VII, requires knowledge of the

industrial N source and management used for determining the N response

models (Bouldin, 1988).



Mahapatra and Sharma (1989), for example, in a 2-year study on a clay

loam in India observed higher lowland rice yield (0.8 and 0.3 t/ha) for basal

application of 40 kg S. aculeata N/ha with 40 kg urea N/ha than for a split

application of 80 kg urea N/ha (50% basal, 25% at tillering, and 25% at

panicle initiation). Chakraborty et al. (1988), on the other hand, in a 3-year

study on a sandy loam in India found no greater rice yield for basal

application of 45 kg S. aculeata N/ha with 45 kg urea N/ha than for basal

application of 90kg urea N/ha. Comparison of results for these two studies

is confounded by the differences in N timing for the sole urea treatment.

Because rice yield correlates with soil ammonium N following rice transplanting (Mahapatra and Sharma, 1989), such conflicting reports of N

fertilizer equivalence by legume N could possibly result from differences

in timing and availability of urea N.

1 . N Fertilizer Timing

In addition to reducing the industrial N fertilizer requirements for rice,

application of legume N to rice might conceivably alter the optimum timing

and management of industrial N fertilizer for rice. Watanabe (1984) suggested that in temperate climates a basal application of 20 kg N/ha as

industrial N fertilizer together with green manure can increase rice yield by

increasing early tillering that can be retarded by toxins formed during

anaerobic decomposition and by slow N release from green manure. Some

researchers in the tropics (Meelu and Morris, 1988) have suggested delaying industrial N fertilizer application until about panicle initiation when

green manure is incorporated before transplanting.

Singh et al. (1987), in a field study with Sesbania green manure incorporated before transplanting on sandy loam soil, found greater rice yield

when 60 kg urea N/ha was applied in equal splits at 21 and 42 DT (8.4 t/ha)

rather than in equal splits at transplanting and 21 DT (7.7 t/ha) or as a

single application at 21 DT (7.8 t/ha). When 120 kg N/ha was applied,

grain yield was greater for three equal splits at transplanting, 21 DT, and

42 DT (9.0 t/ha) than for two equal splits at 21 and 42 DT (8.4 t/ha). The

same timings of urea were not examined at both N rates, and results were

for only 1 year. The transferability of these research findings, obtained on

coarse-textured, highly percolating soils, to fine-textured soils is unclear.

The effect of prerice treatment (fallow, cowpea incorporated at

flowering stage as a green manure, and cowpea grown to maturity with

grain and pods removed and remaining residues incorporated) on response

of lowland rice to two timings of urea was recently studied in the Philippines. Incorporation of green manure and residue increased dry-matter

accumulation and grain yield, but dry-matter accumulation (Table IX) and

yield (Table X) were consistently greater with the early rather than the late



Table IX

Effect of Urea Timing at 58 kg Nlha on Dry-Matter Accumulation of Lowland Rice

at Los Banos, Philippines"

Sampling timeh





Crop maturity ( 1 10)

Dry-matter accumulationc (t/ha)

Early N split

Delayed N split

















Unpublished IRRUIFDC collaborative research. All values are the mean of 3 prerice

treatments (fallow, cowpea green manure incorporated, and cowpea residue incorporated).

The prerice treatment by N timing interaction was not significant ( p = .05).

DT designates days after transplanting.

Early N split designates N applied two-thirds basal incorporation without standing water

and one-third broadcast at 43 DT (approximately panicle initiation). Delayed N split designates N broadcast one-half at 14 DT and one-half at 55 DT (approximately 10 days after

panicle initiation).

* Significant at .01 probability level. NS, not significant.

split of urea. The absence of a significant prerice treatment by urea timing

interaction suggests that the early N split, which is recommended in the

absence of organic N inputs, would also be the more effective urea timing

when integrated with incorporated legume N.

Table X

Effect of Prerice Treatment and N Source Applied at 58 kg Nlha on Grain Yield of

Lowland Rice at Los Baiios, Philippines"

Grain yield (t/ha)

Difference (t/ha)

Prerice treatment

Urea, DSb

Urea, ES'


DS vs. ES

DS vs. USG


Cowpea, green manure

Cowpea, residue

Cowpea, residue removed





















Adapted from John e t a / . (1989~).All values are the mean of 2 years.

one-half at 14 days after transplanting and one-half at 10 days after panicle


Applied two-thirds basal incorporation without standing water and one-third broadcast at

5 days before panicle initiation.

Urea supergranule all basal deep placed.

* Significant at .05 probability level.

** Significant at .01 probability level.

* Broadcast



Additional research is needed; experiments should include treatments

without and with legume N addition in factorial combination with urea

management practices at multiple urea N rates. The findings will aid in

determining whether the recommended N fertilizer management practice

for rice remains the optimal practice when legume N is added.

2. N Fertilizer Source

Placement of urea as large granules, referred to as urea supergranules

(USG), at about 10-cm soil depth between hills of transplanted rice is

frequently more effective than conventional broadcasting of prilled urea

on puddled rice soils (Savant and Stangel, 1990). Dhane et ul. (1991)

showed that the superiority of USG over conventional broadcast urea can

be enhanced with green leaf manuring. Basal application of 2 t/ha of

Gliricidiu sepium toppings increased rice yield by 0.2-0.7 t/ha for splitapplied urea and by 0.4-1 .O t/ha for USG. Rapid release of plant-available

N from the green manure presumably complemented the slower availability of N from deep-placed USG.

Whereas the results in Table X indicate no prerice treatment by N

interaction when comparing the two prilled urea timings, there was a

significant prerice treatment by N interaction when comparing prilled urea

and USG. The increase in yield from USG as compared with the yield from

the delayed split of urea was less for green manure than for the other three

prerice treatments. Nitrogen conceivably was no longer limiting yield in

the treatment with green manure and 58 kg N/ha as USG, and a comparable yield might have been obtainable with less USG N.

Experiments with multiple N rates for each N source are required to

clearly ascertain the possible interactive effect of legume N on effectiveness of urea placement.


Most research on N contributions from legumes in the tropics has

focused on short-duration legumes grown and subsequently incorporated

solely as green manures immediately before the monsoon rice crop. Food

legumes are frequently grown as postmonsoon rice crops and then followed by a fallow period during the dry season before remaining legume

residues may be incorporated with land preparation for the next monsoon

rice crop. Little is known about the transformations, losses, gains, and

recycling of N from postnce legumes.



More effective nodulation of legumes and selection of food legumes for

high residue production (App et al., 1980) merit consideration for increasing fixed N2 in rice-based cropping systems. Postrice forage legumes

cropped in association with food legumes or cereals may offer opportunities for providing fixed N2 to the subsequent rice crop in addition to food

and fodder production (Carangal et al., 1988; Miah, 1988). Little is known

about N transformations and contributions to rice in these forage-food


Nitrification during nonflooded periods between rice crops and then

subsequent denitrification when soil is flooded for rice may be an important avenue for N loss. Little is known about the influence of legumes on

these losses and the probable losses of legume N by nitrificationdenitrification. Research is needed to determine whether legumes influence losses and availability of native soil N by cycling soil nitrate N

through residues incorporated before rice (George et al., 1990).

The inclusion of legumes in rice-based cropping systems could conceivably result in enhanced formation of methane during anaerobic decomposition of leguminous plant material and in enhanced formation of nitrous

oxide during nitrification and denitrification of legume N. Methane and

nitrous oxide are greenhouse gases, and they have been linked to depletion

of ozone in the stratosphere. Research is needed to quantify methane and

nitrous oxide emissions in lowland rice-based cropping systems with and

without legumes. Cultural practices should be developed to minimize the

emissions of these gases from lowland rice fields.

Integrated use of industrial fertilizers will be essential for sustained high

yields in most rice-legume systems. Nitrogen from leguminous green

manure and residue normally only partially meets the N requirements for a

succeeding high-yielding rice crop. Moreover, P and micronutrients can

limit legume production, particularly in infertile, acid soils. Effective inoculation may also be necessary for high N accumulation by legumes, especially when soybean is newly introduced to an area. The limited estimates

of N2 fixation by legumes on ricelands suggest that in many situations with

food legumes, the N removed by the grain exceeds the N2 fixed. Moreover,

residues are often used for animal feed rather than as an N source for the

succeeding rice crop.

The N contribution of legumes in lowland rice-legume sequences, as in

upland crop-legume sequences, depend on the quantity of legume N

derived from N2 fixation, the NHI, the proportion of legume N mineralized, and the efficiency of use of this mineralized N by the succeeding crop

(Myers and Wood, 1987). However, the anaerobic-aerobic soil cycles

typical of lowland rice-legume sequences can lead to higher losses of N

than would normally occur in upland crop-legume sequences, in which the

soil is not flooded. Losses of nitrate N, formed from mineralization of soil

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

VII. Effective Management of Legume Nitrogen

Tải bản đầy đủ ngay(0 tr)