Bp. Caton et al., A plant growth model for integrated weed management in direct-seeded rice.III. Interspecific competition for light, FIELD CR RE, 63(1), 1999, pp. 47-61
The model DSRICE1 was developed for analyzing integrated weed management st
rategies for direct-seeded rice. We have shown that DSRICE1 predicts monocu
lture rice growth well and accounts for water-depth effects on growth. Here
, the model is used to simulate competition for light between rice and two
weeds, Echinochloa oryzoides (early watergrass) and Ammannia spp. (redstem)
. Except for minor differences in phenology, weed growth was simulated as d
escribed for rice. Direct competition for Light depended on the species' ve
rtical distributions of leaf and stem areas (live and dead) and their extin
ction coefficients. Water also attenuates light, so species' early height g
rowth rates were important because they determined when plants emerged into
full light. Structural sensitivity analyses of rice in competition with th
e two weeds revealed that water-depth effects and leaf area distributions s
trongly affected competition, and shading by dead leaf and stem dry mass re
duced total production. Validation was based on independent data sets for r
edstem and watergrass competition using several statistical tests and indic
es. For rice-redstem competition, DSRICE1 simulated rice growth well becaus
e redstem competitive effects were small, but predictions of redstem growth
were good only when observed heights were matched in simulations. Redstem
competitiveness depended on height growth rate, perhaps due to its small se
ed size. For rice-watergrass competition, the growth of both species was pr
edicted well, except that watergrass growth in plots with early-season drai
nage was underpredicted. Watergrass parameters were similar to those for ri
ce except for faster height growth and higher photosynthesis rates. In a mo
del application, simulations in which rice seeding was delayed for a time a
fter flooding led to greater yield losses from redstem than from watergrass
because delays reduced the advantage of rice over redstem. The usefulness
of DSRICE1 for drained fields will be improved by better simulation of plan
t growth responses to drainage, but rice competition with redstem and water
grass in continuously-flooded fields was simulated well. (C) 1999 Elsevier
Science B.V All rights reserved.