Progress & News

PRELIMINARY RESEARCH FINDINGS (last updated: April 2006)


The major achievement of the first two years has been the close monitoring of plant and insect growth at 15 field sites around the country, now totalling 18 months of monthly samples for some of the sites. These data have revealed that all but one of the monitoring sites have total water phosphorous levels above the 0.1mg/l lower threshold which limits the growth of water hyacinth. This site is also the only one that can be considered to be under successful biological control. Degree-day modelling of immature insect development, using microsite temperature data from the field sites, shows that weevil and mirid immature development will stop for extended periods during the winter months at cold sites, and that warm sites will only support two to three generations of biocontrol insects per year. Trials with sub-lethal doses of herbicides have continued in the laboratory and have been extended to the field. A model to assist water managers decide on feasible levels of control has been developed, using both international data and that collected from the monitoring sites. Remote sensing of water hyacinth populations is being investigated as a monitoring and management tool.

Effects of temperature

?Continuous measures of temperature (every 30 minutes), facilitated by temperature buttons, are undertaken to collect microclimate data for each site. These measures include water temperature 5cm below the surface, air temperature within the water hyacinth canopy and ambient air temperature 1.2m above the water surface.

When these data are fitted to a degree-day model extracted from the literature, two patterns emerge (Figure 1). Firstly, that development of Neochetina eichhorniae immatures is almost halted for up to four months at the colder sites. Secondly, current estimates of development times of the weevil taken from the literature are grossly inaccurate, predicting up to 13 generations of N. eichhorniae in a year at the warmer sites. Comparison of these predictions (Figure 1) with actual population measures recorded at the field sites (Figure 2) indicates a huge discrepancy

Figure 1:
Accumulated degree day profiles for the development of Neochetina eichhorniae (Developmental threshold of 11.9 C and a Generation Time of 195 degree days) at three water hyacinth sites. * denotes accumulating generations of the weevil within a season, according to the degree day model.

In addition, studies of arthropod performance and plant productivity under conditions of chronic exposure to non-lethal low temperatures were run in environmental chambers, controlled for air temperature, humidity and photo phase. These evaluated the performance of N. eichhorniae, relative to low temperatures. Adult feeding activity and survival was significantly impaired in the ?cold? chamber (set up to approximate the winter conditions of a field site where the agent has had a negligible impact), indicating that even a relatively short exposure to chronic non-lethal temperatures is damaging to overwintering populations of adult weevils.

Effect of nutrients

Water nutrient sampling continues to be dogged by difficulties in receiving data from samples analysed by the Institute of Water Quality Affairs. These difficulties aside, our site selection procedure which characterized sites as high medium and low nutrient water bodies is validated by the data which has been collected. Water nitrogen and phosphorus appear to correlate with each other as do plant measures of these elements. However, we still don t have enough consecutive water samples analysed to indicate if one of the elements in the water correlates with its concentration in the plant.

From analysis of the available data two unsettling conclusions arise: Firstly, the regression of phosphorus levels against nitrogen concentrations reveals a ratio of N:P of 7:1 - which is ideal for water hyacinth growth (Wilson, 2002). Secondly, all but one of the monitored sites has an average phosphorus level greater than the 0.1mg/l threshold, above which water hyacinth growth is unrestrained by phosphorus availability (Wilson, 2002). That this site (New Years Dam) is the only one considered by us to be under biological control suggests that at high phosphorus levels water hyacinth cannot be suppressed by the available biocontrol agents. Circumstantial evidence to support this notion comes from Lake Victoria, where water hyacinth has been successfully controlled and phosphorus levels are below 0.1mg/l threshold. This hypothesis will be tested in the coming year in manipulative laboratory trials.

Effects of herbicide concentration

Laboratory trials in the summer of 2005, experimenting with 0.8% glyphosate as a retardant dose for water hyacinth, have supported results from Autumn 2004 and Spring 2005, in that application of 0.8 % glyphosate was found to inhibit growth of new leaves and daughter plants, without killing the plant or its resident beetles. Field trials using these concentrations have begun at Farm Dam, Mbozambo Swamp and Delta Park . However, early results from Farm Dam, where the plants were largely unaffected by glyphosate supplied and sprayed by Monsanto at 1.5%, did not correspond with laboratory trials, leading to the suspicion that spray volumes might need to be adjusted to plant mass. This discrepancy is under investigation with the assistance of Monsanto.

In the light of a recent flurry of scientific literature lambasting ?Roundup? for its effects on aquatic amphibian tadpoles (e.g. Relyea, 2005) a further review of the ecological impacts of ?Roundup? has been prepared, and several experiments have been planned to measure the effects of 0.8% glyphosate on aquatic vertebrates and invertebrates.

Plant and insect phenology

The monthly field site sampling has accumulated a wealth of data which awaits complete analysis. Some of the insect data has been shown above (Figure 2), which reveals weevil numbers peaking at successively later months in cooler sites. The decline in adults and larvae during the winter suggests that the weevils are overwintering as larvae, which drives the population into a bottleneck from which it has to recover in the spring. By contrast, plant phenology indicates that winter is a period of reproduction, with the production of new ramets occurring at all sites (Figure 3). Leaf elongation occurs during summer; root length is less seasonally variable, but may be correlated to nutrient levels at different sites. These data indicate that the plant is reproducing asexually at a time when weevil numbers are declining, which may allow new ramets to escape from colonization by the biocontrol agents. It also indicates that

herbicides should be applied in Autumn, when the plants are about to multiply - however herbicide uptake is probably hampered by low temperatures. Both of these hypotheses will be investigated further, as will the possibility that ramet addition is driven by plant density rather than photoperiod or temperature. If water hyacinth responds to a decrease in plant density by reproducing asexually then this will influence when herbicides should be applied, and reapplied on any water body that is under integrated management

Figure 3:
Phenology of water hyacinth at three field sites; reproduction and leaf addition.

Management model

A management model has been developed using data from water hyacinth infestations all over the world. Data collected from our monitoring sites has been added to this, and tempered with knowledge gathered during the course of sampling, to generate a series of scenarios under which water hyacinth populations may occur. The model then indicates which management options are available at different sites, and what outcome should be expected from different management strategies (Figure 4)

Figure 4:
Broad water hyacinth management recommendations under different abiotic conditions.

The scenario of ?Biocontrol only? indicates that the weevils are expected to eventually provide an adequate and permanent reduction in the plant population (i.e. a few fringing plants covering less than 5% of the water surface). However, an integrated management plan may be required before this stage is reached.

The conditions in the different panels of Figure 4 are defined as:

a) temp always good / good; deep / some shallows / or some marsh,

b) temp always good / good; marshy,

c) seasonal / short winter; no / minor frost damage; deep / some shallows,

d) seasonal / short winter; no / minor frost damage; some or all marshy,

e) seasonal / short winter; major / severe frost damage; deep / some shallows,

f) seasonal / short winter; major / severe frost damage; some or all marshy,

g) short / long winter; minor / most frost damage; deep / some shallows,

h) short / long winter; minor / most frost damage; some or all marshy.

The model is more of a decision-making tool than a predictive model (i.e. predicting exactly how all of the individual components in a water hyacinth system will interact). It is unlikely that we will ever get to a stage where we can closely predict how such a system will react, and this type of forecasting may be unnecessary if the decision-making model works in practice.


Remote sensing

On-the-ground measures of water hyacinth growth are valuable for an understanding of the way in which the plant grows in different circumstances, but the management goal of containing water hyacinth growth will need to be realised at a landscape scale. To this end we are investigating the feasibility of using remotely sensed satellite images of field sites as a management tool. Because waterside estimates are notoriously inaccurate we envisage satellite images being used as an accurate monitoring tool of mat growth, and possibly as predictors of mat growth, which will indicate when and where herbicide application should take place.

Matching on-the-ground measures of actual plant growth with remotely sensed images has rarely been tried because of insufficient ground data. Our extensive database from the field sites allows us to precisely ground-truth any growth patterns that are detected from satellite images.

This aspect will form an Honours project at Wits University, in which the student will be supported by the Working for Water Capacity Building Programme.

Long-term effects of herbicides

Data gathered to date indicate that sub-lethal herbicide applications do not affect the weevil populations in terms of feeding behaviour, adult survival or larval survival. However, these measures have only been taken over eight-week periods and do not address the long-term effects on the survival and reproduction of the plants. Consequently, we intend to investigate the effects of sub-lethal doses of glyphosate on aspects of the biology of both the plants and their biocontrol agents.

Herbicide resistance and tolerance

Although the use of sub-lethal doses of herbicide is predicted to reduce the selection pressure for herbicide resistance in water hyacinth populations, it is nevertheless a concern that has been raised in discussions with managers. Therefore we will undertake tests to examine the existence and inheritance of herbicide resistance in water hyacinth, using ramet production as a measure of herbicide resistance. We will use the same benchmark to determine if water hyacinth becomes herbicide tolerant after repeated applications of glyphosate.

Ecological effects of herbicides

Relyea published five papers in 2005 which present alarming results indicating that glyphosate, and ?Roundup? in particular, causes over 90% mortality of indigenous frog tadpoles in the USA . Though there are many debateable aspects of the way in which the trials were conducted, the impression remains that ?Roundup? is very damaging in aquatic ecosystems.

We will conduct laboratory trials with lethal and sub-lethal doses of glyphosate to determine what effects our recommended low dose of the herbicide have on indigenous amphibians.


Research team members have made the following presentations at conferences and workshops:

32nd Annual Weeds Workshop, 10-14 May 2004, Golden Gate, South Africa.
Byrne MJ: An integrated weed management programme for water hyacinth.
The 3rd Kruger National Park Science Networking Meeting, 4?8 April, 2005, Skukuza.
Brudvig R, King A, Robertson M, Hill M, Oberholzer I & Byrne MJ: Integrated control of water hyacinth in the Kruger National Park.
5th Congress of the Entomological Society of Southern Africa,10-15 July 2005, Rhodes University.
Byrne MJ, Hill MP & Robertson MP. Integrated management plan for control of water hyacinth.
33rd Annual Weeds Workshop, 14-16 July 2005, Rhodes University , Grahamstown South Africa.
Brudvig R & Byrne MJ: Integrated control of water hyacinth under different nutrient regimes.
Jadhav A & Byrne MJ: Impact of sub lethal doses of glyphosate on water hyacinth and its biocontrol agents.
King A & Byrne MJ: The effect of temperature on the biological control of water hyacinth.
The 4th Kruger National Park Science Networking Meeting: 12-17 March 2006, Skukuza.
Jadhav AJ, Brudvig R, King AM & Byrne MJ: Integrated control of water hyacinth in Kruger National Park .

Back to TOP

Photo credits

The majority of the photos displayed on this site are by members of the research team. For some images, however, the original source credit information could not be traced, and we apologise for any omissions of this nature. We welcome any information that could correct this.


Relyea RA, Schoeppner NM & Hoverman JT (2005). Pesticides and amphibians: the importance of community context. Ecological Applications 15 : 1125-1134.

Wilson JR (2002). Modeling the dynamics and control of Water hyacinth. PhD Thesis. Imperial College of Science, Technology and Medicine, Berkshire.