Water Management for Disease Vector Control - Biological and Environmental Control of Disease Vectors - page 167

China and India with the aim of achieving a
combination of reduced mosquito breeding, rice
yield increases and lower water use (van der
Hoek et al ., 2001; Keiser et al ., 2002). In the fi rst
half of the 20th century, researchers tested
dif erent options to reduce the time that water
was available on rice fi elds in order to control
vector breeding. The classic experiments in
Portugal resulted in lower water use and higher
rice yields and detected a more than 80%
reduction in Anopheles larvae compared with
continuous irrigation (Hill and Cambournac,
1941).
Most studies on the ef ects of AWDI have
been conducted in small experimental fi elds
assessing impacts on larval abundance.
However, there is a clear need for more studies
that focus on farmer-managed fi elds that
represent real life situations. Two studies
undertaken in Kenya (Mutero et al ., 2000) and
India (Krishnasamy et al ., 2003) highlighted the
limitations that can be encountered in large-
scale farmer-managed AWDI fi eld implement-
ations. In both studies, AWDI saved water at the
farm level, while maintaining rice yields on par
with other water management regimes tested,
but there were no signifi cant overall dif erences
in immature mosquito abundance between
AWDI and regular irrigation practices. In India,
in particular, the levelling of fi elds was
inadequate for proper drainage, and the interval
between wettings and dryings was too short
(farmers accepted only 1 day of drying) for
ef ective killing of stranded mosquito larvae
(Krishnasamy et al ., 2003). Rainfall was an
added complication that negated the impact of
periodic drying. Some of these factors can be
altered, such as ef ective levelling of fi elds,
improved irrigation infrastructure and irrigation
management capacity, while others cannot,
such as rainfall and soil conditions. Also, a well-
designed irrigation and drainage system that
allows for rapid fl ooding and drying and is
ei cient enough to allow for synchronous
irrigation and drainage of all fi elds within the
system is required before AWDI can be ef ective
(van der Hoek et al ., 2001). As a result, AWDI
might be more suitable for smaller irrigation
systems, or so called on-demand systems, where
more fl exibility in water allocation is available.
Larger irrigation systems are often run on fi xed
rotations, which in the case of large parts of

India and Pakistan mean that farmers will only
get water once every 2 or 3 weeks.
In addition, studies are needed to assess
how AWDI may impact breeding habitats
outside rice fi elds, such as seepage areas,
irrigation canals and drains. If AWDI does not
infl uence vector species of mosquito breeding in
canals, hydraulic structures, storage reservoirs
and seepage areas, then the control of water in
rice fi elds alone is likely to be of limited use.
Where agricultural drainage canals are a major
source of breeding itself, frequent draining of
fi elds could even exacerbate vector breeding.
Furthermore, even under AWDI, rice crops must
be kept under continuous fl ooded conditions
during the 10-15 days after transplantation of
seedlings, and vector production during this
period may be at a peak because the irrigated
area is greatest and the rice plants are short and
sparse, allowing easy access for female
mosquitoes to oviposit (Mogi et al ., 1984).
An example of the positive health impact of
AWDI is the dramatic reduction of malaria in
Sichuan Province, China over the period 1982-
1999, after a large scale implementation of
AWDI (Qunhua et al ., 2004). While several
control measures were used, the virtual
disappearance of malaria vector mosquitoes and
malaria in some areas was attributed to
elimination of breeding sites with the im-
plementation of AWDI. In India, AWDI was able
to control Cx . tritaeniorhynchus breeding in rice
fi elds, with a likely impact on JE transmission
(Kanojia, 2007). However, the knowledge of the
impact of AWDI on disease burden is hampered
by the fact that most studies so far have relied on
collection of mosquito larvae, thus only
determining ef ects on mosquito density. There is
a need for more epidemiological studies assessing
the impact on disease outcomes. Such studies
are needed as larval abundance might not be
linearly related to abundance of adult
mosquitoes, and adult mosquito abundance
might not translate into changes in human
disease. It is therefore essential that future
studies take human disease incidence and
prevalence as outcome measures of water
management interventions.
In most countries, AWDI has not become
an accepted practice among farmers and
irrigation managers, because of the more
complicated water management regime. Despite