M. Becker et De. Johnson, Rice yield and productivity gaps in irrigated systems of the forest zone of Cote d'Ivoire, FIELD CR RE, 60(3), 1999, pp. 201-208
Much of the rapidly growing demand for rice in West Africa will be met from
increased production in irrigated lowlands, which cover about 12% of the r
egional rice-growing area. A large potential for expansion of irrigated are
as exists particularly in the inland valleys of the humid forest zone. Curr
ent production is characterized by large variability in productivity, manag
ement practices and production constraints. Quantifying the variability in
rice yield and identifying the determining factors are prerequisites to the
development of site-specific recommendations and to improved targeting of
technologies. Diagnostic on-farm trials were conducted on 64 irrigated lowl
and fields in the humid forest zone of southern Cote d'Ivoire, in 1995-1996
. This was a part of the regional gradient study of irrigated systems from
the desert margin to the humid forest zone. Cropping calendars, field opera
tions and input use were monitored. Weed biomass, rice N uptake, and grain
yield were determined in farmers' fields as well as in super-imposed, resea
rcher-managed subplots (clean weeding, no N control, and mineral fertilizer
N application). Rice yield potential was simulated by using the Oryza-S cr
op growth model. Yield losses were attributed to management factors based o
n performance of rice in researcher-managed subplots (management-related yi
eld gap) and by multiple regression with management options. Grain yields v
aried between 0.2 and 7.3 Mg ha(-1) with mean yields of 3.2 in partially an
d 4.2 Mg ha(-1) in fully irrigated systems, 44% and 57% of the potential yi
eld of 7.3 Mg ha(-1) respectively. Age of seedlings at transplanting, timel
iness of operations and application of P fertilizer were correlated to yiel
d and explained 60% of the observed variability. Grain yield was correlated
with N uptake (r(2) = 0.93***) but not with N application rate. Split appl
ication of mineral fertilizer N was associated with a 0.48 Mg ha(-1) yield
increase (p = 0.002), regardless of the quantity applied. Additional weedin
g increased yield only in systems with imperfect irrigation. Weed biomass w
as reduced with improved water control and it increased with age of seedlin
gs at transplanting, and was higher in direct-seeded than in transplanted r
ice. Echinochloa spp. were the most common weeds in fully irrigated systems
and Panicum laxum was more common in the imperfectly irrigated fields. Whi
le improved water management was associated with substantial rice yield inc
reases (1.16 Mg ha(-1)), the timeliness of transplanting, weeding and N fer
tilization appears to be the key to increased rice yields in the forest zon
e of West Africa. (C) 1999 Elsevier Science B.V. All rights reserved.