Tải bản đầy đủ - 0trang
III. Ecology of Upland Rice Weeds
WEEDS AND WEED MANAGEMENT
reported that continued application of butachlor caused a shift from monocots to dicots. Sankaran and De Datta (1984) reported that C. benghalensis,
an annual broadleaf weed, became dominant in upland rice when R. exaltata
was completely controlled by preemergence applications of pendimethalin. A
shift favoring slow-growing broadleaf weeds over fast-growing annual grasses
is desirable because it significantly reduces weeding time. Unfortunately, the
shift sometimes is from a moderately easy-to-control weed to one that is
difficult to control (Mercado, 1983).
c. ADAPTATIONAND GROWTHOF WEEDS IN UPLAND RICE
Iwata and Takayanagi (1974a) reported that D. ciliaris seeds were more
adaptable to low soil moisture content (40% of field capacity) than upland
rice and other weeds such as P . oleracea, Amaranthus retro$exus L., and E.
indica. E. indica was more adaptable to high soil moisture content, even
waterlogging, than other weeds.
Iwata and Takayanagi (1974b) studied the growth rate of D. ciliaris and
upland rice sown at the same time in monoculture and mixed culture. The
optimum growth period of Digitaria was shorter, but the growth increments
during the period were larger than those of rice. In mixed cultures, Digitaria
reduced the number of tillers, the plant height, and the dry weight of rice.
In West Africa, perennial weeds of cleared, drier forest and derived savanna
zones include Cyperus spp. and I. cylindrica. Their rhizomes make them more
difficult and expensive to control than annual weeds. Infested areas are
generally allowed to lay fallow for several years before rice is planted again
(Akobundu and Fagade, 1978).
IV. WEED COMPETITION
Weed competition largely governs the development of upland rice. The
usual low yield of upland rice has been attributed mainly to inadequate and
irregular moisture supply, heavy weed infestations, lack of suitable cultivars,
nutritional imbalance, and inadequate cultural practices, including inefficient
control of disease and insect pests (De Datta and Beachell, 1972). Among
these limiting factors (except poor water supply), inadequate weed control is,
perhaps, the most difficult constraint to increasing upland rice production
(De Datta, 1972; Madrid et al., 1972).
S. SANKARAN AND S. K. DE DATTA
FIG.1. Schematic diagram of an upland rice ecosystem showing the relationship between
weed management, weed pressure, and crop yield. (Adapted from OBrien, 1981.)
Figure 1 depicts the crop-weed balance in the upland rice ecosystem and
the various factors that influence it. Primary weed control methods affect rice
yield by reducing weed pressure. Secondary weed control methods, such as
seedbed preparation, moisture conservation, and crop rotation, directly
reduce weed pressure and increase rice yields. Other secondary methods, like
seeding method and density, increase the competitive ability of rice. Fertilizer
application increases weed and rice growth.
In the upland rice ecosystem, seed is sown when soil moisture is adequate
for germination. Rice and weed seedlings compete for moisture, nutrients,
and light during germination and growth (Akobundu and Fagade, 1978;
Borgohain and Upadhyay, 1980).
The rice crop and weeds use the same environmental factors for their
growth needs. Competition begins as soon as a factor cannot support the
normal growth needs of the crop and the weeds (Utomo, 1981).
Weeds in upland rice can withstand drought better than rice because they
have deeper roots and high root-length density to tap moisture from deeper
soil layers (Sankaran and De Datta, 1984).
High-cost inputs such as seeds and fertilizers are useless under upland
conditions without efficient weed control. In shifting cultivation, the land is
abandoned when weeds seriously reduce yields. After reviewing current
practices in shifting cultivation, Moody (1975) reported that existing weed
control methods are unsuitable for continuous or large-scale farming. He
emphasized the need for research to find economical, effective control
WEEDS AND WEED MANAGEMENT
The critical period of weed-crop competition is between early growth,
during which weeds can grow without affecting crop yield, and the point after
which weed growth does not affect yield (Zimdahl, 1980). Establishing the
critical period of competition is essential to develop effective, economical
weed control measures (Sharma et al., 1977). In upland rice, researchers
(Kawatei et al., 1966; Park and Kim, 1971; Mercado, 1979; Schiller and
Indhaphun, 1979; De Datta, 1980; Kolhe and Mittra, 1981) pinpointed the
optimum weed-free period (Table V).
The Critical Period of Weed-Crop Competition in Upland Rice
No. of days weeds
can be left in
Days after which
emerging weeds can
be left to compete
with the crop
Ghosh et al. (1977);
Sharma et al. (1977)
Upadhyay and Choudhary (1979)
Kolhe and Mittra (1981)
Sahai et al. (1983)
Kawatei et al. (1966)
Park and Kim (1971)
Wells and Cabradilla (1981)
Schiller and Indhaphun (1979)
Morales and Vargas (1976)
The first 15 days after seeding (DAS) rice seem to be the maximum period
during which weeds can be tolerated without affecting the final crop yield
(Morales and Vargas, 1976; Ghosh et al., 1977; Sharma et al., 1977; Schiller
and Indhaphun, 1979; Upadhyay and Choudhary, 1979; De Datta, 1980;
Wells and Cabradilla, 1981).The weed-free requirement for upland rice varies
from the first 30 to 60 DAS, depending on edaphic and climatic conditions
and weed flora.
In their studies on weed-crop competition, Wells and Cabradilla (1981)
found that competition began during the first 3 weeks (20 days) of the crop.
S. SANKARAN AND S. K. DE DATTA
Grain yield decreased as weed growth duration increased. Weed growth
increased exponentially during the first 60 days, reaching a maximum dry
weight of 6.6 ton/ha. Removing weeds at this stage gave little yield increase.
The critical period of weed competition was between 2 and 9 weeks after
Moody (1982) wrote that if weeds mature rapidly, shorter-duration rice
cultivars would compete for a proportionately longer time than longerduration cultivars. Because the late-maturing cultivar grows longer, it has
more ability to compensate for weed competition.
Yields from plots not weeded for 10 days after germination were equal to
those of the weed-free check (Ghosh et al., 1977). Weed presence for 20 days
or more reduced yield significantly. O n the other hand, keeping the plots
weed-free for the first 20 days was as good as keeping them weed-free for 50
days or throughout the growing period. No weed competition occurred and
yield was unaffected if weeds grew at 10 DAS. George et al. (1968), Morales
and Vargas (1976), Sharma et al. (1977), Schiller and Indhaphun (1979), and
Wells and Cabradilla (1981) had similar findings that weeds do not appreciably reduce yield if they begin to grow 15 DAS rice.
Merlier (1978) observed that weed dry weight and yield of the rice cultivar
Iguape cateto were negatively correlated (Fig. 2) and that increasing the
duration of grass weed competition progressively reduced panicle weight and
- - - Dry weight of weeds
FIG.2. The effect of dry weight of weeds after different periods of infestation and yield of
variety Iguupe cateto. (Adapted from Merlier, 1978.)
WEEDS AND WEED MANAGEMENT
In Ghana, Carson (1975) reported that weeding upland rice could be
delayed until 6 weeks after seeding without adversely affecting yield. Beyond
6 weeks, however, yield losses increased with a corresponding delay in weed
removal. Eight weeks of early weed-free conditions prevented substantial
yield reduction. The critical stage of weed competition was between 6 and 8
weeks after seeding.
The weed-free requirement of upland rice emphasizes the need for selecting
preemergence herbicides that are residually active and effective for the first 50
DAS to ensure optimum grain yield. Weed control after the critical period
could prove wasteful.
AND EFFECTOF COMPETITION
Blackman and Templeman (1938) reported that, in a normal rainfall year,
competition is principally for nitrogen. Chakraborty (1973) reported higher
nitrogen content in weed species at the vegetative, flowering, and postflowering stages, indicating severe competition for nitrogen throughout the upland
rice growing season. At IRRI,De Datta (1974a,b) and Okafor and De Datta
(1976b) studied the effect of nitrogen level on varying densities of C. rotundus
in upland rice. De Datta (1974b) reported a significant negative correlation
between C. rotundus dry weight and rice grain yield as influenced by nitrogen
application (Fig. 3).
120 kg N/ha
r = -09822**
r = -0
n- L., '
Dry weight of perennial nutsedge (g/rn21
FIG.3. Correlation and linear regression between weed weight and grain yield of upland rice
as influenced by nitrogen level. IRRI, 1972 wet season. *,Significant at 5% level; **, significant at
1% level. (From De Datta, 1974b.)
S. SANKARAN A N D S. K. DE DATTA
Severe weed infestation in upland rice depressed the height (Wells and
Cabradilla, 198 l), dry matter (Chakraborty, 1973), tiller and panicle numbers
(Okafor and De Datta, 1974; Sharma et a!., 1977; Kolhe and Mittra, 1981),
leaf area index, and light transmission ratio (Okafor and De Datta, 1974) of
Utomo (198 1) observed that rice plants and weeds compete in two ways.
The tops of the plants compete for light and the root zone competes for
nutrition (nitrogen and humus). In these competitions, rice plants develop
less chlorophyll and contain less nitrogen, which causes fewer panicles and
spikelets per plant.
Many factors influence competition. Weeds and rice modify the environment by influencing the growth of constituent plants. Plant species, plant
density, and distribution and duration of plants in the ecosystem directly
affect competition (Bleasdale, 1960, cited by Zimdahl, 1980). Climatic and
edaphic conditions modify rice-weed competition (Fig. 4).
Upland rice yield is influenced by weed density per unit area, especially
during early growth. Weed density and rice yield have a sigmoid relationship
(Utomo, 1981). Eussen (1981) reported that A . conyzoides densities from 32
to 1024 plants/m2 reduced grain yield of an upland rice cultivar in Bicol,
edaphic and climati
Upland r i c e 1
FIG.4. Schematic outline of the factors influencing the upland rice-weed complex. (Adapted
from Bleasdale, 1960.)
WEEDS AND WEED MANAGEMENT
Philippines, by 19 to 72%. Grain yields declined progressively as weed
density increased. Eussen and Hadi (1981) observed a similar effect of D.
ciliaris on grain yield. Takayanagi and Iwata (1978) reported that D. ciliaris
had a higher dry matter production at early growth stages.
C. rotundus, generally considered the world's worst weed (Holm et al.,
1977), reduced the rice yield 33 and 93% at 32 and 1024 plants/m2,
respectively. Grain yield reduction was mainly caused by reduced spikelet
and panicle number.
Eussen and Martoyo (1981) found that Porophyllum ruderale (Jacq.) Cass.
at 64 plants/m2 present for 60 DAS retarded tillering and reduced grain yield
by 50%. At 1024 plants/m2, the effect appeared 30 DAS and reduced yield by
91%. D. aegyptium populations reduced yield similarly (Utomo, 1981).
The level of competition varies with weed species. The extent of competition can be determined by calculating the relative space occupied by each
species following the De Witt (1960) method. Utomo (1981) reported that D.
ciliaris occupied more space than Digitaria fuscencens (Presl) Henr. and
Digitaria ternata (A. Rich.) Stapf.
In a greenhouse experiment, Soetrisno et al. (1981) showed that D. ciliaris
retarded rice growth and decreased yield. They found the relative yield value
of rice in mixed culture to be smaller than that of D. ciliaris. Weeds occupied
more space than rice.
Okafor and De Datta (1974) compared the effects of three groups of weeds
(annuals, C. rotundus, and annuals plus C. rotundus) on the growth and yield
of drilled and broadcast IR5 (Table VI). A combination of annual weeds and
Effect of Weed Composition on Leaf Area Index (LAI), Light Transmission
Ratio (LTR), and Grain Yield of IR5 Rice under Two Methods of Seeding"
Annuals plus perennial
IRRI, 1972 wet season. In a column, values followed by the same letter are not significantly
different at the 5 % level as determined by DMRT. From Okafor and De Datta (1974).
S. SANKARAN AND S. K. DE DATTA
C. rotundus caused the greatest reduction in leaf area index (LAI), light
transmission ratio (LTR), and grain yield.
Many soil and environmental factors modify weed competition. Soil
nutrient level influences the nature and duration of competition in upland
rice ecosystems (De Datta and Malabuyoc, 1976). Increasing nitrogen
fertilizer application increased nitrogen uptake of C. rotundus (Okafor, 1973).
Weeds grow better under adequate levels of nutrients, thus making them
Yield Reduction in Upland Rice Due to Uncontrolled Weed Growth in
Asia, Africa, and Latin America
Yield reduction (%)
Mani et al. (1968); Manna et al. (1971);
Mukhopadhyay et al. (1971); Spratt and
Chowdhury (1978); Singh and Sharma
(1981); Moorthy and Dubey (1981)
Syam and Effendi (1977)
IRRI (1967); Vega et al. (1967); De Datta
(1972); Okafor and De Datta (1974);
Mercado (1979); Sarkar and Moody (1981)*
Jayasekera and Velmurugu (1964)
Schiller and Indhaphun (1979); Teerawatsakul
Aryeetey (1970); Carson (1975)
Moody (1973); Fagade (1976)
Jones and Tucker (1978)
Burga and Tozani (1980)
‘Numbers in parentheses denote the range of grain yield reductions.
Mean of 10 wet seasons (1971-1980) yield data at IRRI and mean range of loss in yield at 9
locations in the Philippines.
WEEDS A N D WEED MANAGEMENT
more competitive. Eussen and Zulfaldi (1981) reported that at 100-300 kg N/
ha, weeds reduced grain yields by 25 %. Grain yield reduction was independent of nitrogen level. Therefore, weed control becomes more imperative as
fertilizer applications increase.
E. YIELD LOSSESDUETO WEEDS
Weeds cause two types of crop losses. The most important is the direct
yield loss resulting from competition, followed by indirect loss from reduced
crop quality (De Datta, 1980). All cereal crops yield less under severe weed
stress (Table VII).
Data from the upland rice experiments from 1971 to 1980 at the International Rice Research Institute (IRRI) show weeds reduce grain yield by 2 ton/
ha with a mean of 89% (Sarkar and Moody, 1981). Yield reduction fluctuated
from 0.9 to 3.3 ton/ha, with a mean loss of 2.3 tons/ha.
When weed pressure is extremely severe at early rice growth, the yield loss
can be 100% (IRRI, 1967, 1968; Vega et al., 1967; Manna et al., 1971; De
Datta, 1972; Madrid et al., 1972; Moody, 1973; Smith, 1973; Williams, 1975;
Fagade, 1976; WARDA, 1976). Dubey and Thomas (1977) reported 4 5 8 0 %
yield loss in upland rice, 42% of which occurred in the first 3 weeks of crop
V. LAND PREPARATION AND CROP ESTABLISHMENT
TECH NlQU ES
Land preparation reduces the weed problem in the following crop. The
rain-fed environment is so heterogenous that there is not one set of land
preparation methods for all situations (Pillai, 1981).
Most countries of Asia, Africa, and Latin America produce upland rice
during the rainy season, which normally lasts for 3 to 7 months. Generally, a
single rice crop is grown in uplands, after which the field remains fallow. This
practice facilitates weed growth and weeds complete their life cycle with
adequate soil moisture, which increases weed problems for the next crop.
Weekly harrowing in the dry season gives good soil tilth, which favors rice
seed emergence and weed growth. For example, Castin et al. (1983b) found
FIG.5. Effect of previous land usage on grain yield of upland rice. (Adapted from Castin
et al., 1983b.)
that planting maize (Zea mays L.) or mungbean in the dry season reduced
weed growth and weeding time and increased herbicide performance and rice
yields in the following season as compared with plots kept weed-free by using
paraquat or being maintained as weedy fallow (Fig. 5). In plots where maize
was planted during the preceding dry season, yields were 1.6 tons/ha
compared to 1.3 and 0.7 ton/ha, respectively, in the mungbean and weedy
OF SEEDBED PREPARATION IN DIFFERENT
In Asia, little mechanization is used to prepare seedbeds. Fields are plowed
and harrowed with animal-drawn implements when rain falls (De Datta and
Ross, 1975). In Africa, upland rice fields are tilled manually.
Deep plowing is also a weed control method in upland rice. Curfs (1975)
found fewer weeds on soils that were deep plowed and rototilled with a
powerful tractor than on those that were shallow plowed or zero tilled. Pande
and Bhan (1966a) also showed the effectiveness of deep (28 cm) tillage in
reducing weed biomass and increasing panicle number and grain yield as
compared to tillage at 7, 14, and 21 cm (Fig. 6). At IRRI, Lopez et al. (1980)
and De Datta and Llagas (1984) reported that increasing the number of
rototillings reduced weed population, weed biomass, and weeding time
(Lopez et al., 1980) and increased grain yields.
c . REDUCEDTILLAGE
AND WEED GROWTH
Crops can be grown with minimum soil disturbance and reduced energy
inputs for cultural operations if herbicides are used to control weeds.