Content uploaded by Richard Beilfuss
Author content
All content in this area was uploaded by Richard Beilfuss on Feb 14, 2017
Content may be subject to copyright.
Adaptive management of
the invasive shrub Mimosa pigra
at Gorongosa National Park
Dr. Richard Beilfuss
Department of Scientific Services
December 2007
2
1. BACKGROUND ON MIMOSA
Mimosa pigra, one of the 100 "World's Worst" invasive species (Global Invasive Species
Database), is a highly aggressive shrub that poses a serious threat to biodiversity, water
resources, and tourism at Gorongosa National Park. Mimosa is native to tropical America,
but has invaded ecosystems worldwide including Africa, Asia, Pacific islands, and Australia,
favoring a wet-dry tropical climate with at least 750 mm annual rainfall. Mimosa is well-
established on the floodplains, pans, and waterways of Gorongosa, and may be poised to
expand rapidly in the future if it is not effectively managed.
1. 1 Biological characteristics of Mimosa pigra
Taxonomic name: Mimosa pigra L.
Common names: Mimosa or Giant Sensitive Plant (English); Columbi-da-lagoa or Malicia-
de-boi (Portuguese)
As a young plant mimosa grows as a single prickly stem, often as a ground creeper or
prostrate form (Photo 1). Mature mimosa tends to be a branched shrub with rose-like thorns
(Photo 2). The plant can reach a height of 3-6 m with a branching tap root extending 1−2 m
deep. Leaves are green, feathery and fern-like, with the central leaf stalk being prickly and
20−25 cm long. Each leaf contains up to 16 opposite segments, each segment 5 cm long and
divided into pairs of leaflets which fold up at nightfall or when touched or injured. The
flowers are round, fluffy, pink or mauve balls 1−2 cm across (Photo 3). Each flower head
produces a cluster of 1−30 seed pods which are 3−8 cm long and covered with dense hairs
(Photo 4). The pods turn brown when mature and break into segments, which fall away from
the pod leaving a skeletal outline. Each segment contains an oblong shaped seed 4−5 mm
long and 2 mm wide (Walden et al. 1999).
Photo 1. Typical prostrate form of mimosa on Urema floodplain (R. Beilfuss).
Mimosa typically germinates as flood waters recede. Plants mature quickly and can set seed
in their first year of growth (Walden et al. 1999). The first flowers tend to appear 6−8 months
after germination. Flowers are bee-pollinated or possibly wind-pollinated. Seeding occurs
approximately 3−6 weeks after the flower bud is formed.
3
Photo 2. Mature mimosa shrub (G. Howard).
Photo 3. Mimosa leaves, thorns, and flowers (R. Beilfuss).
Photo 4. Mimosa seed pods (R. Beilfuss).
4
The one seeded, hairy segment of mimosa pods are easily spread by humans, animals, and
water. Segments may become attached to people’s hair, shoes, and clothing. The segments
stick to the fur of animals and can pass unharmed through their digestive tract. Seeds pods are
buoyant, and are readily dispersed by flood waters and water currents.
The seeds have an extremely hard, often impermeable seed coat. Some are able to germinate
as soon as conditions permit, while others may remain dormant for 15 years or more
depending on the environment. Lonsdale (1983) counted more than 12,000 seeds per meter
squared from soil in a mimosa-infested area in northern Australia. Seed production and plant
life expectancy are highest on black cracking clays such as those found on the Lake Urema
floodplains.
Mimosa’s invasiveness is attributed to its aggressive growth. Once seedlings are established
growth is rapid; one-year-old plants with a stem diameter of 2.5 cm often attain a diameter of
7 cm in the second year. In experiments in the Northern Territory of Australia, regrowth from
young plants severed at ground level reached a height of 2.5 meters and covered an area of
6.3 square meters within 12 weeks (Land Protection 2006).
Mimosa often spreads with changes in land use and general "disturbance," including
alterations to the hydrological regime (Mumba and Thompson 2006). However, mimosa’s
buoyant seed pods can be spread long distances in flood waters and establish on a range of
soil types, enabling it to spread quickly from disturbed areas to natural floodplain systems
(Photo 5).
Photo 5. Spread of mimosa along the Malinde
Channel of the Kafue Flats, Zambia (G. Howard).
1.2 Ecological, social, and economic impacts of Mimosa pigra
Mimosa forms dense, monospecific, impenetrable thickets, 3−6 meters high, that establish on
waterways, floodplains, and wetlands similar to those found in Gorongosa National Park
(Photo 6).
5
Photo 6. Impenetrable mimosa thicket (Tran Triet).
Mimosa thickets significantly reduce available grazing areas for herbivores and eliminate
most other species of grasses and forbs; converting floodplains into unproductive scrubland
with reduced levels of biodiversity. Mimosa thickets block access to pans and waterways for
wildlife, livestock, irrigation, and recreation purposes. Mimosa colonizes water courses,
reducing water flow and increasing silt levels.
Mimosa will also invade agricultural areas, coastland, natural forests, planted forests,
range/grasslands, riparian zones, shrublands, and urban areas. When large infestations spread
to farmland, mimosa may reduce the area of cropland and grazing land and the overall
carrying capacity of the land. Furthermore, it may block access to water sources for livestock.
Mimosa may interference with the cultivation of other economically-important plants. For
example, mimosa is able to outcompete young palm trees in immature oil palm plantations
and cause a decrease in the production of palm oil (Praneetvatakul 2001).
There are numerous examples of the ecological, social, and economic costs of mimosa
invasions from around the world. I have personally observed significant mimosa invasions in
Thailand, Vietnam, northern Australia, Zambia, and Ethiopia over the past 20 years. In
Thailand, mimosa chokes irrigation systems that supply rice fields, reducing crop yield and
harming farming livelihoods. It also provides cover for of rats and crabs, which damage crops
(Praneetvatakul 2001). In Vietnam, mimosa began invading Tram Chim National Park in the
early 1990s, and now threatens the biodiversity of seasonally inundated grasslands in this
highly unique and vulnerable remnant of the original Mekong Delta ecosystem. According to
Triet et al. (2004), despite warnings from weed experts (including a study-visit of Vietnamese
resource managers to mimosa problem areas in Northern Australia that I hosted in 1994),
very little has been done in Tram Chim to control mimosa, and now the infestation has gone
beyond easy management (Photo 7). At the current rate of spread, and if there is no
improvement in weed control effort, mimosa is expected to invade all grassland areas of
Tram Chim in less than five years, seriously jeopardizing the life of native plants and animals
that depend on the native grassland habitat (Photo 8).
6
Photo 7. Mimosa thicket invading Tram Chim National Park in
the Mekong Delta of Vietnam (Tran Triet)
Photo 8. Mimosa displacing floodplain grassland of
Tram Chim National Park (Tran Triet).
In the Northern Territory of Australia, more than 80 000 ha of floodplains have been covered
by mimosa, and it threatens Kakadu National Park, a World Heritage Site (Photo 9). River
floodplains and swamp forests are threatened by dense thickets of mimosa, resulting in fewer
numbers of birds and lizards, less herbaceous plants, and fewer native tree seedlings. Mimosa
also prevents traditional food gathering by Aborigines on otherwise resource rich wetlands.
Mimosa has also invaded wetlands, floodplains, and waterways across much of sub-Saharan
Africa. It is reported (Global Invasive Species Database) as a problem species in Ghana,
Guinea, Kenya, South Africa, Swaziland, and Uganda. Photo 10 shows mimosa invasion to
the Boyo Wetlands in Ethiopia, a lake-floodplain system very similar to the Lake Urema
floodplain. A similar pattern was recently observed at Lake Chilwa in Malawi (M. Finlayson
pers. com.)
7
Photo 9. Mimosa invasion covering 100% of a tropical
floodplain in northern Australia (M. Finlayson).
Photo 10. Mimosa invasion in the Boyo Wetlands of Ethiopia. Mimosa
invaded on recently deposited sediments along drainage lines, and spread
laterally to the floodplain that surrounds Boyo Lake (R. Beilfuss).
The spread of mimosa across the Kafue Flats of Zambia has many important management
lessons for Gorongosa National Park. Mimosa has become a dominant species in the Chunga
Lagoon of Lochinvar National Park over the last 15 years (Photo 11). Before 1980 there was
only one known infestation of approximately 2 ha, at the head of Nampongwe steam which
flows from Chunga Lagoon (Mumba and Thompson 2005). By the mid-1980s it was
spreading and covered approximately 100 ha (Thompson 1986). Thompson (1986) suggested
that control measures were needed, but no steps were taken. A survey undertaken in 2003
indicated that this area had increased to around 2500 ha, with mimosa quickly replacing
previous dominant grassland vegetation in the flats such as Echinochloa stagina and Oryza
longistaminata (Photo 12) (Mumba and Thompson 2005). Most recent reports (G. Howard
pers. com.) indicate the mimosa infestation at Lochinvar now covers more than 3000 ha. G.
Howard also reports that mimosa is now spreading far from Chunga lagoon along the main
floodplain of the middle Nampongwe stream (an area that was “looking very healthy at the
8
end of the dry season in 2006’) and mimosa is expanding as well in Blue Lagoon National
Park on the northern side of the Kafue Flats.
Photo 11. Solid thickets of mimosa on the periphery of Chunga Lagoon
in 2003. Mimosa has reduced access to floodplain grazing lands for endemic
Kafue lechwe and Vulnerable Wattled Cranes (R. Beilfuss).
.
Photo 12. Dense mimosa thickets on the border of Chunga
Lagoon, which now cover more than 2500 ha (R. Beilfuss).
Although an action plan was drafted to control mimosa at Chunga Lagoon, aimed at
employing local laborers to hand cut and remove the shrubs, and donor funding was secured,
the Zambian Wildlife Authority (ZAWA) and the Star of Africa Tourism Camp chose not to
implement the plan and the invasion continues to expand (Photo 13). Only recently has a
serious effort been launched to combat the dense mimosa thickets, which are “a huge
problem to control” (G. Howard pers. com.)
9
Photo 13. View of Chunga Lagoon from the 5-star “Star of Africa” lodge in
Lochinvar National Park, blocked by mimosa (R. Beilfuss).
2. STATUS OF MIMOSA PIGRA AT GORONGOSA
Tinley (1977) makes only brief mention of mimosa in his comprehensive ecological study of
Gorongosa. Tinley recorded the presence of mimosa in the ephemeral mudflat annual
communities that occur on the edge of receding Lake Urema waters, and noted that “mimosa
is a favored browse food and only attains shrub growth form during inundations; the
remainder of the time it assumes a prostrate growth form in response to heavy utilization by
herbivores.”
Tinley monitored 30 m2 quadrats in each of 18 1-hectare sampling sites (total of 540 m2
quadrats) covering the microperennial Cynodon-Digitaria lawn grasslands south of Lake
Urema near the Sungue inflow channel. The prostrate form of mimosa was recorded in only
16 of 540 m2 quadrats, with a 1% relative frequency. However, the sampling approach used
by Tinley (transects along a catenal sequence of increasing moisture) may have under-
sampled mimosa in its preferred habitat at the edge of the inundation zone. Among the woody
saplings sampled along these transects, mimosa was clearly the dominant species (95%
relative density). Mimosa was not recorded in any of the other floodplain grassland
associations sampled, which included saline grassland; Setaria floodplain grassland;
Echinochloa stagnina marsh grassland; Vetiveria nigritana tall grassland.
Tinley’s accounts suggest that mimosa has certainly been present around Lake Urema for
decades, but was not a problem species due to intense herbivory (and likely corresponding
trampling) and perhaps other, unknown factors.
Preliminary field assessments during September to November 2007 suggest that mimosa now
occurs throughout the floodplain grasslands that form a ring around Lake Urema. It typically
occurs in prostrate form, 10-20 cm in height, with taller growth up to 1.5 m height along
drainage lines. It varies in density from a few shrubs to hundreds of shrubs per hectare, and is
the dominant non-graminoid ground cover around Lake Urema (Photo 14). Mimosa occurs at
highest density in the Cynodon-Digitaria short-grass associations, where it appears to be
expanding. It also occurs at lower density in the Echinochloa-Setaria medium-grass
10
associations on the floodplain. Mimosa is sparse or absent in floodplain areas where dense
stands of Acacia xanthophloea fever trees have established.
Photo 14. Mimosa prostate clusters along Picada 4, extending 100s of meters
onto the floodplain towards Lake Urema. These clusters are capable of
forming an impenetrable thicket over vast areas (R. Beilfuss).
Within the Chitengo trails system that can be easily monitored by road, mimosa is found
along almost the entire length of Picada 4 that dissects Cynodon-Digitaria short-grass
associations, on both sides of the road (Photo 15). Mimosa occurs in small isolated patches
on the eastern floodplains opposite Casa dos Leões, becoming increasingly dense near the
intersection of Picadas 4 and 6 and continuing to the interaction of Picadas 4 and 8. The only
observed gaps in distribution correspond to dense stands of Acacia xanthophloea and
Faidherbia albida, and some of the tall, unburned patches of Setaria floodplain grassland.
Mimosa also occurs on Echinochloa-dominated floodplains near Miradouro dos
Hipopótamos and on the edge of the larger pans, including Lago Mareza (Picada 12), Lago
Nhamutengo (Picada 4), and Lago do Paraiso (Picada 11)
Map 1 shows the generalized distribution of mimosa at Gorongosa National Park. Further
monitoring is needed to confirm whether mimosa is present in any of the other major pans
along the trail system, especially Lago Inhatite and Lago Nhansato, or the Rio Mussicadzi
and Rio Sungue channels. The species has not been observed in these areas to date.
A permanent plot of 1256m2 (20m swept radius around a permanent rebar) was sampled to
establish a baseline estimate of mimosa cover in an area where the species has recently
established. It is an easily observed area at the intersection of Picada 4 and 3 (GPS: S18º
12
54.172’; E034º 26.112’). For each mimosa patch observed, the distance from center rebar
and compass direction was recorded and the cover (rectangular area) was estimated. Total
cover of mimosa in the sample plot is 1.4% (Table 1).
Photo 15. Mimosa prostrate growth form along Picada 4 in
Cynodon-Digitaria grassland (R. Beilfuss).
Table 1. Patch size and location of Mimosa pigra in an experimental
monitoring plot (1256 m2) on the Lake Urema floodplain.
Orientation from
center (degrees)
Distance from
center (m)
Patch
dimensions (m)
Patch
area (m2)
80 10 1.7 x 1.4 2.4
84 8.5 2.1 x 1.7 3.6
85 13.5 1.6 x 1.3 2.1
92 16 2.1 x 1.9 4.0
256 15 5 x 1.1 5.5
Total area 17.5
Other plots in this same region range from near absence to >50% cover (Photo 16).
Additional sampling quadrats will be established for areas where the species has been
established for a longer time period, with repeat measurements to determine to changes in
species cover and density over the annual wet and dry cycle
13
Photo 16. Dense mimosa cover on floodplain at end of dry season (R. Beilfuss).
3. MIMOSA PIGRA MONITORING AND CONTROL PROGRAM
3.1 Is Mimosa pigra a problem invasive species at Gorongosa?
The evidence is inconclusive, but several factors suggest that Gorongosa National Park
Management should take urgent action to learn how to best control mimosa pigra:
1. The species is widespread around Lake Urema and very dense in the Cynodon-Digitaria
short-grass associations that form important grazing lands for my herbivores.
2. Although historical data are insufficient to quantify the change in density of mimosa over
the past 40 years, these data and other anecdotal evidence suggests that mimosa is
expanding significantly, especially in the Cynodon-Digitaria short-grass associations.
3. The density of herbivores on the Lake Urema floodplain is greatly reduced relative to
historical conditions. There is no evidence of significant herbivory on mimosa at present.
This reduced grazing pressure has enabled the species to photosynthesize throughout the
year, and its mostly prostrate growth form suggests that most of the mimosa energy has
been allocated to subsurface growth, establishing dense rooting networks.
4. There is evidence that a recent fire (probably during 2006) top killed many mimosa
patches, while stimulating basal regrowth (Photo 17). Although individual fires do not kill
the species, repeat annual fires may reduce plant vigor or eliminate some patches, and
will slowly deplete the seed bank. However, the floodplain area between Picada 4 and
Lake Urema remains fairly green and vigorous during the dry season due to high water
table conditions, and will not burn regularly without managed fires (see below).
5. As evidenced throughout the world, mimosa can remain in slow-growing prostrate form
for many years before expanding rapidly into dense impenetrable thicket (hence M.
Finlayson described mimosa as the “Rip Van Winkle” of the plant world).
14
Photo 17. Mimosa shrub that was top killed by fire during previous
growing season, as evidenced by large number of thick, dead stems.
Note vigorous regrowth at the base. (R. Beilfuss).
6. The rapid spread of mimosa after a long quasi-dormant period is often observed in
response to changes in hydrological conditions on floodplains. Many scientists suggest
that the Lake Urema system is already drying, which could trigger a mimosa response
(e.g., Kafue Flats in Zambia). Alternatively, an altered runoff and sediment regime
resulting from the land use change and deforestation on Gorongosa Mountain, the
Midlands, or Cheringoma escarpment could result in the rapid spread and infestation of
mimosa (e.g., Boyo Wetlands in Ethiopia).
7. There are severe negative consequences of a mimosa infestation around Lake Urema for
wildlife carrying capacity, access to water resources, and tourism development, and the
option of “doing nothing” entails great risk. These negative effects have been recorded
repeatedly throughout the world in floodplains similar to the Lake Urema system. Several
websites have been established to document these impacts
3.2 Monitoring and control plan
In response to these concerns, a mimosa monitoring and control program is urgently
recommended. The program should commence in the Cynodon-Digitaria short-grass
associations along Picada 4, with emphasis on floodplain areas on Picada 4 between Picada 6
and Picada 2 intersections (10.3 km), where the threat to tourism, wildlife, and the natural
structure and function of the floodplain is of greatest concern.
Permanent monitoring plots should be established to evaluate change in density and cover of
mimosa over a two-year evaluation period in response to the following management
treatments:
1. Fire
2. Herbicide
3. Manual removal
4. No action (monitoring only)
15
3.2.1 Fire management
Procedure and timetable:
1. Establish four 1-ha burn treatment plots along Picada 4. Mark the corners of each plot
with permanent rebar and GPS position. March 2008.
2. Measure baseline density and cover of mimosa in each plot. April 2008.
3. Conduct annual burn during the mid-dry season as soon as conditions are suitable to carry
hot fire. August 2008.
4. Measure density and cover of mimosa in each plot at the end of the dry season and
compare to other treatments. November 2008.
5. Repeat burn treatment during dry season in second year. August 2009.
6. Repeat measurement and evaluation procedure at end of second year. November 2009.
Justification:
As noted above, there is evidence that mimosa growing along Picada 4 was top-killed by fires
during the previous growing season, but in all cases resprouts occurred next to dry burned
stems. We will test if repeated annual burning of the same patches causes the mimosa to die-
out over time or show signs of significantly reduced vigor. Because there is limited grassy
understory in dense thickets of mimosa, it is difficult to destroy infestations with fire once
they become well established. Therefore, this prescribed burning program must be
implemented urgently while the species remains in its present, prostrate growth form within
heavy grass cover that can carry fire. Although mimosa seeds on the soil surface are also
destroyed by fire, germination of seeds from the seed bank, within 5 cm of the soil surface,
may be enhanced by this control program. It may stimulate seed germination due to the
removal of seed coats (Miller and Lonsdale 1992, in Walden et al. 1999). Therefore it is
important to continue the controlled fire experiment even if all the mimosa shrubs are
eliminated, to prevent new resprouts from establishing.
3.2.2 Herbicide application
Procedure and timetable:
1. Establish four 1-ha herbicide treatment plots for along Picada 4. Mark the corners of each
plot with permanent rebar and GPS position. March 2008.
2. Measure baseline density and cover of mimosa in each plot, and flag each patch for
identification by herbicide team. April 2008.
3. Herbicide all above ground foliage of mimosa. A locally available herbicide suitable to
woody species control with foliar spray or stump application will be used (the herbicide
“Access” or comparable). April 2008.
4. After two weeks, evaluate results of herbicide treatment and repeat herbicide of any live
mimosa foliage. May 2008.
5. Measure density and cover of mimosa in each plot (herbicide top-kill effectiveness and
resprouting) at the end of the dry season and compare to other treatments. November
2008.
6. In addition, herbicide all know isolated patches of mimosa, including especially the
eastern floodplains opposite Casa dos Leões where mimosa has just established at low
density, and small pans around Gorongosa NP (including Lago Mareza (Picada 12), Lago
Nhamutengo (Picada 4), and Lago do Paraiso (Picada 11). Record all treatment locations
with GPS. Check effectiveness of herbicide treatment after two weeks, and repeat
treatment as needed. April-May 2008.
7. Repeat herbicide treatment of plots (including 2-week recheck) and isolated patches of
regrowth during second year. April-May 2009.
8. Repeat measurement and evaluation procedure at end of second year. November 2009.
16
Justification:
Herbicides are widely used to control mimosa, especially in Australia. Herbicide should be
applied during the active period of growth of the mimosa and before any seed mature. In
Australia, this is during the wet season (Lonsdale 1988, Miller 1988, in Walden et al. 1999).
Ground-based herbicide application methods include direct injection, foliar/basal bark
spraying, and soil application (Walden et al. 1999).
3.2.3 Manual removal by cutting
Procedure:
1. Establish four 1-ha manual cutting plots along Picada 4. Mark the corners of each plot
with permanent rebar and GPS position. March 2008.
2. Measure baseline density and cover of mimosa in each plot, and flag each patch for
identification by manual cutting team. April 2008.
3. Manually remove all above ground foliage of mimosa as soon as access is possible
following the wet season. The crew should be provided with thick gloves to avoid thorns,
and should attempt to pull out the roots of each patch when soils are still damp. April
2008.
4. After two weeks, evaluate results of manual removal and remove any additional live
mimosa foliage that was missed. May 2008.
5. Measure density and cover of mimosa in each plot (resprouting) at the end of the dry
season and compare to other treatments. November 2008.
6. Repeat manual cutting of plots (including 2-week recheck) during second year. April-
May 2009.
7. Repeat measurement and evaluation procedure at end of second year. November 2009.
Justification:
Manual removal takes advantage of unskilled labor and readily available equipment.
However, it provides only temporary control. Hand weeding may be effective for controlling
seedlings in crops. Seeds should be collected and burnt before weeding commences. Roots
should be removed and destroyed. Long handled cutters, axes and machetes may be used but
any stumps left will resprout unless herbicide is applied immediately after removal
(Thamasara 1985, in Walden et al. 1999). Methods that cut plants off at ground level or
above (such as slashing or chaining) result in resprouting and will not control mimosa. Blade
ploughing is one method of physical control that cuts the plant off below ground level. It was
found to be very effective in preventing mimosa resprouting.
3.2.4 No action (monitoring control plots)
Procedure:
1. Establish four 1-ha no treatment (control) plots along Picada 4. Mark the corners of each
plot with permanent rebar and GPS position. March 2008.
2. Measure baseline density and cover of mimosa in each plot. April 2008.
3. Measure density and cover of mimosa in each plot at the end of the dry season and
compare to other treatments. November 2008.
4. Measure density and cover of mimosa in each plot at the end of the dry season and
compare to other treatments. November 2008.
17
Justification:
Provides clear baseline information about status and spread of mimosa without any control
action. May also provide insights about the impact of herbivory on mimosa, including the
possibility of increasing impact of browse over time as wildlife populations further recover in
the park.
3.2.5 Other recommended actions
At the end of the second year (November 2009), all the treatments will be evaluated to
determine the most effective and feasible course of action to be implemented on a large scale.
For all plots, we will record any observations of herbivory or other ecological activity in the
plots throughout the year and during the field measurements at the end of the dry season.
Ongoing reconnaissance throughout the park is important to ensure that any other isolated
infestations are identified and targeted before they expand to uncontrollable levels.
It is recognized that combinations of treatments may be part of the control program---for
example, a cutting and burning treatment, or an herbicide and burning treatment. The scope
of this first experimental program is to try to understand specifically the growth response
following each of these management treatments. The ideal scenario is that the infestation can
be controlled with fire until herbivore numbers have recovered adequately to maintain the
mimosa with browsing activity. Although fire is ineffective for managing mimosa shrubs
once they form dense thickets, it may be effective while they are in the more widespread,
prostrate growth form. Herbicide and manual cutting operations are much more labor
intensive and costly options, and combination treatments even more so.
This experimental program is based on the assumption that mimosa will continue to spread in
its prostrate form but will not undergo rapid transformation into thicket over the study period.
If sudden expansion or growth of mimosa is observed during 2008, more drastic measures
must be immediately implemented. These should be based in part on results of the
experimental treatments, if available, and in part on comparable experience from the region.
Mimosa control efforts recently implemented for the Kafue Flats illustrate a comprehensive
program involving a combination of treatments (G. Shanungu pers. com.). Forty field workers
from surrounding local communities are employed to physically remove the mimosa. Mimosa
is first cleared at the invasion front, aimed at eradicating isolated “founder” populations and
preventing further spread, followed by efforts focused on the larger well-established
infestations. The workers first remove any mimosa seeds and burn them to prevent entry into
the seed bank. For thicker stands of mimosa where hand picking of seeds is difficult, plants
(or thickets) are cut to ground level using machetes and the cut stems with foliage are
stacked, dried, and burned. Burning exposes the stumps and opens the areas for easier
access—the stumps are then uprooted by hand or with the use of mattocks, stacked to dry for
a period of 2-3 weeks, and burned. Burning is found to destroy any seeds near the soil
surface, and helps to break the dormancy of the seeds in the seed bank, causing seedlings to
sprout after a few weeks (on average two weeks after removal and burning). These seedlings
are then easily removed from cleared areas during follow-up action, further depleting the seed
bank. All areas are checked regularly to clear any regrowth of mimosa in target areas.
Chemical control is also under consideration for the flats, aimed at following-up the physical
removal process with the herbicide treatment (i.e., Glyphosate) on resprouts and seedlings
after the annual floods recede. The control team is concerned about the ecological sensitivity
of the area, however, and has decided not to apply herbicides on a large scale pending the
results of an Environmental Impact Assessment (EIA) are reviewed.
18
Results from the Kafue Flats control program thus far indicate that native vegetation is
growing back in cleared areas, most notably the herb Ambrosia maritima and grass species
such as Cynadon dactylon. Kafue lechwe have been observed utilizing areas now cleared of
mimosa.
Such an intensive program is not justified for Gorongosa at present, but conditions can
change and must be regularly monitored.
Educating local communities about mimosa impacts and spread could be an important tool
for identification and control of mimosa in the buffer zone.
Restricting the movement of vehicles, soil, and sand from infested areas is important to
prevent the spread of mimosa seeds.
3.2.6 Comments on biological control (possible control method in the future)
Although still under development, biological control is being considered as a long term
management option in some areas, including the Kafue Flats (G. Shanungu pers. com.).
PIER (2002) reports: “Six natural enemies have been released in Australia following
rigorous host-specificity testing, but have not yet had any discernible effect. Four of these
have also been released in Thailand. The six include two seed-feeding bruchid beetles, one
stem-feeding chrysomelid beetle, two stem-boring moths and, in January 1992, a flower-
feeding weevil.”
Cronk and Fuller (2001) report: “For long-term control, biological methods are probably the
most cost-effective considering the extent and ecology of this species. Palatability to higher
animals is low, but in its native range it is attacked by more than 200 species of insect
herbivores and fungal pathogens. The first insects introduced to Australia as controlling
agents were the seed-feeding beetles Acanthoscelides quadridentatus and A. puniceus
(Bruchidae) from Mexico. They were released in Australia in 19845 and 1985, respectively,
but have not attained high population densities and have had little impact on seed production.
Two stem-boring moths, Neurostrota gunniella (Gracillariidae) and Carmenta mimosa
(Sesiidae), were released in Australia in 1989; of these, N. gunniella established readily. The
young larvae mine leaf pinnules and the older larvae tunnel in the stems, causing them to die.
Carmenta mimosa complements the action of N. gunniella by tunneling stems of larger
diameter. Other important insects currently being tested for their host specificities in Mexico
and Australia are the seed- and flower-feeding weevils Apion sp., Chalcodermus serripes,
Sibinia fastigiata, S. ochreosa, S. pervana and S. seminicola.
“Two fungal pathogens, Phloeosporella sp. (Coelomycetes), and a rust, Diabole cubensis
(Uredenales), severely debilitate Mimosa pigra in Mexico. Phloeosporella sp. attack leaves,
branches, main stems and seed pods, causing leaf fall and cankers of the stems and leading to
ring barking and die-back. Diable cubensis causes chlorosis in stems and leaves resulting in
premature leaf fall. Both fungi are attacked by hyperparasitic fungi in their native range and
it seems likely that their effect on Mimosa pigra could be even more damaging in Australia if
they were to be introduced without their natural enemies. These fungi are under investigation
in Mexico and Britain.”
19
4. REFERENCES
Cronk, Q. C. B. and J.L. Fuller. 2001. Plant invaders. Earthscan Publications, Ltd., London.
241 pp.
Land Protection. 2006. Facts Pest Series: Mimosa pigra. Queensland Department of Natural
Resources and Waters, Australia. Link: http://www.nr.qld.gov.au/
Lonsdale, W. M., Miller, I. L. and I.W. Forno. 1995. Mimosa pigra L.. Pages 169-188 in
R.H. Groves, R. H., R.C. H. Shepherd, and R.G. Richardson (eds.). The Biology of
Australian Weeds, Volume 1. R. G. & F. J. Richardson, Melbourne.
Lonsdale, W.M., 1983. Rates of spread of an invading species—Mimosa pigra in northern
Australia. J. Ecol. 81, 513–521.
Mumba, M. and J.R. Thompson. 2005. Hydrological and ecological impacts of dams on the
Kafue Flats floodplain system, southern Zambia. Physics and Chemistry of the Earth.
Pacific Island Ecosystems at Risk Database (PIER). 2002. Compiled by Colin Wilson, Parks
& Wildlife Commission of the Northern Territory & Invasive Species Specialist Group
(ISSG) Palmerston, Australia. http://www.hear.org/pier/species/mimosa_pigra.htm
Paynter, Q. and G.J. Flanagan. 2004. Integrating herbicide and mechanical control treatments
with fire and biological control to manage an invasive wetland shrub, Mimosa pigra. Journal
of Applied Ecology 41(4).
Praneetvatakul, S. 2001. An Impact Assessment of ACIAR Research Projects on Biological
Control in Thailand. Kasetsart University (Department of Agricultural and Resource
Economics): Bangkok. In S. Isvilanonda, S. Praneetvatakul, C. Sangkapituk, A. Sattarasart,
C. Singhaprecha and P. Sirisupluxana. Impact Assessments of Forty-nine Thailand/Australia
Collaborative Projects Funded by ACIAR during 1983–1995 (Working Paper Series No. 38).
The Global Invasive Species Database (GISD). www.issg.org/database
Thompson, S.R., 1986. Report on new invasion by Mimosa pigra in an African National
Park. Aquaphyte, summer 1986.
Tinley, K.L. 1977. Framework of the Gorongosa ecosystem. Ph.D. dissertation. University of
Pretoria.
Tran Triet, Le Cong Kiet, Nguyen Thi Lan Thi and Pham Quoc Dan. 2004.. The invasion by
Mimosa pigra of wetlands of the Mekong Delta, Vietnam.
http://www.ento.csiro.au/weeds/pdf/mimosa_symposium/07Trietetal.pdf
Walden, D., C.M. Finlayson, R. van Dam, and M. Storrs. 1999. Information for a risk
assessment and management of Mimosa pigra in Tram Chim National Park, Viet Nam. Pages
160-170 in Proceedings of the EnviroTox’99 International Conference.
Weed management case study
1
How to attack
Mimosa pigra
on a grand scale
The successful
Mimosa pigra
control program on Melaleuca
Station in the Northern Territory
is the result of a practical,
reasoned and planned approach
to tackling a large-scale
infestation. The lessons learnt
over the past five to six years
provide valuable insights in the
way both large and smaller-scale
infestations of mimosa and other
damaging weeds can be tackled.
Tony Searle, Manager, Melaleuca Station,
Mary River District, NT.
Richard Fell, Tropical Savannas CRC,
Northern Territory University, Darwin.
Mimosa pigra growing wild. By 1993 the weed covered 10,000 ha of Melaleuca’s floodplain
Melaleuca Station lies in the Mary River catch
ment approximately 100 km east of Darwin
in the Northern Territory. It experiences a
monsoonal climate with an average annual rainfall be-
tween 1300 and 1600 mm. The station was subdivided
from Point Stuart Station in the early 1980s to form a
300 km2 property used for the purpose of domesticated
buffalo production. Three other such ‘buffalo blocks’
were formed at the same time.
History of the
Mimosa pigra
invasion
In the early 1980s mimosa was already established on
the property. It was a very small infestation and control
only took two men two days. However, by 1993, the
mimosa had spread to cover 10,000 ha of the flood plain.
This was a very large infestation and meant that almost
33 per cent of the property was not available for pro-
duction. At this time, the property changed hands and a
mimosa control program began.
The Sampan Project saw the Northern Territory
Government take initiatives under the Noxious Weeds
Act to control mimosa on navigable channels of the
lower Mary River. On Melaleuca, 2000 ha in the Red
Lily 1 paddock were cleared of mimosa.
The decision to act
In 1995 a new manager was faced with a restricted car-
rying capacity on the flood plain. Mimosa still covered
8000 ha, and the 2000 ha cleared by the government
program were in need of continuing maintenance to
keep them mimosa-free. It was decided to tackle the
problem in a systematic and planned way, building on
the good work already started. A proposed five-year
control plan including a detailed budget was put to the
directors of the company that owned Melaleuca in 1995.
The proposal was accepted and the program started in
1996. As of November 2000, the 2000 ha originally
cleared as part of the Sampan Project remains free of
mimosa and is back in production. Another 3000 ha
are under various stages of treatment, of which 2000
ha are used for dry-season cattle production at a capac-
ity of one beast to every two hectares (1:2 ha).
Mimosa control program
The aim of the mimosa control program is to:
•Maintain the 2000 ha area that has already been
cleared of mimosa in Red Lily 1 and return it to
production.
•Clear the country closer to the homestead and
work downstream along a line in 1000 ha blocks.
The line was selected early in the program and
stretched from Red Lily 1 to Rumby and Ackerie
Plains, and covered a further 3000 ha on top of
that already cleared.
•Clean up these areas and have all 5000 ha back
into production in five years.
The program is based on a year-by-year approach.
Year 1
A 1000 ha area of old-growth mimosa is selected and
in December a 100 metre perimeter around the area is
sprayed.
Year 2
The perimeter is chained, stick raked and burnt. In Oc-
tober the whole 1000 ha area is burnt inwards from the
perimeter. This has the effect of opening the country
Weed management case study
2
up and saves on one year of spraying. Floodwaters then
control the regrowth through the suppression of any
seedlings that germinate after burning.
Year 3
The whole area is chained in July, or when the
floodwaters recede and machinery can gain access. The
first bulk spray over the whole area to control regrowth
takes place in December. This is done meticulously us-
ing a run-by-run approach from fence line to fence line
so that no plants are missed.
Year 4
The area is now classed as ‘clean country’ and is stick
raked and bulk sprayed once again. Grasses are planted
in around May. Particular attention is paid to water-
courses so that seed can be spread through water move-
ment in the wet. Species planted depends on what is
available as seed or runners. Choice of species is a criti-
cal factor in ensuring that the area is revegetated, thus
becoming productive once more.
Year 5
Treatment is similar to Year 4. Another bulk spray may
be done if required and perhaps another stick rake, if
the season permits.
Year 6
The area is grazed lightly for a short period and atten-
tion is paid to any regrowth areas. The country is now
essentially clear of mimosa.
Progress: maintaining the pressure
Some unforeseen problems have been encountered
throughout the program. As water runs both ways on
the flood plain, mimosa seeds can be moved in two di-
rections. Control of regrowth therefore takes more time
than originally envisaged.
As more country is cleared, the effort required to main-
tain these areas also increases. Time, effort and resources
are then not available to clear more old-growth areas.
The fifth year of the program is largely on schedule.
The last 1000 ha on Rumby Plain were burnt in late
October 1999. The 2000 ha in Red Lily 1 are being
grazed and the other 2000 ha on Ackerie and Rumby
Plains are now in the maintenance program
The program aimed to have an additional 1000 ha
start the process in 1996, however, this did not end up
being feasible. 2000 ha went into Year 3 of the pro-
gram as outlined above, and additional areas were
cleared in 1997, 1998 and 1999 to bring up the total to
5000 ha under treatment at the present time. Thus the
program has slipped a year or so from the original tar-
gets, however, the full importance of maintaining the
pressure on the cleared areas has become well reinforced.
Chemical control
Spraying efficiency in the mimosa control program has
improved through experience. Application rates, spray-
ing conditions and delivery of chemicals have been
modified over time.
Application rates
The following chemicals and application rates are used
for the mimosa control program on Melaleuca:
Task Chemical Rates
Burning Napalm label rate
Bulk
spraying Brush-off 50% label
rate (see below)
Spot
spraying Star Rain label rate
Given that recommended chemical rates are for old-
growth mimosa, trials have been run in conjunction with
chemical companies to determine the rates needed for
mimosa regrowth. These indicated that regrowth can
be killed effectively at 50% of the label rate.
Spraying conditions
Air and ground conditions have to be right for an effec-
tive kill. Spraying must be done on the right day, at the
right time and in the right conditions. In particular:
•the temperature should not be over 35°C; and
•spraying should not be done in the late afternoon
when the leaves have curled up.
Delivery
The type and/or location of spraying determine which
helicopter is used. Three helicopters are used for spot
spraying; spraying within close range of the watering/
loading point (that is, within a ten minute reload cy-
cle); and spraying where a higher capacity and range is
required.
Revegetation
The type of pasture species used for revegetation after
clearing of mimosa depends on the seed and/or runners
that are available. Olive hymenachne (Hymenachne
amplexicaulis), para grass (Brachiaria mutica) and
aleman (german) grass (Echinochloa polystachya) are
the preferred species.
Some native species have been used and trials are
underway to further investigate their use.
Financial considerations
The mimosa control program is working to a strict
budget that was set over five years ago. Slowing of the
program occurred as the cost of controlling larger ar-
eas of regrowth overtook the cost of the initial kill.
Experience and increasing efficiency, however, enabled
the last 1000 ha of old-growth on Rumby Plain to be
tackled in 1999/2000.
Future of
Mimosa pigra
control
Over the next five years, the mimosa control program
will focus on maintaining the control achieved in the
Weed management case study
3
Principles of woody weed control
Principles of
Mimosa pigra
and woody weed control that have emerged
from the program at Melaleuca are:
•Don’t clear more than you can manage in the subsequent years
of the program
•Operate within your budget of finance, labour and resources
•Effective spraying means paying attention to the conditions
•Be set for the long haul and have ways to measure progress
•Use the natural environment to help. Floods can kill regrowth and
spread grass seed.
•Time the program to use the season to best effect. December,
October and July are critical times.
•Modify the options to suit your local environment.
Disclaimer
Information provided by the TS–CRC for the Prime Notes CD–ROM publication is provided as general advice only.
Professional advice should be sought if seeking to apply the information to specific circumstances. The TS–CRC has
tried to ensure this information is accurate at the time of publication. If in doubt, readers should enquire whether new
information is available on any particular subject matter.
For more information about
land-management issues
in northern Australia, go to
the Savanna Explorer sec-
tion of our website at:
http://savanna.ntu.edu.au/
For more information
about the Centre’s exten-
sive research program go
to our research section.
first five years and reducing the need to spend time and
money on regrowth. Another 5000 ha of old-growth
mimosa will then be tackled in the subsequent five years.
Step-by-step progress will be measured each year.
Conclusion
A successful mimosa control program at Melaleuca Sta-
tion is the result of a practical, planned and reasoned
approach, which began in 1996. A year-by-year pro-
gram of spraying, burning, revegetation and thorough
follow-up has been used. Over this time a number of
principles for large-scale mimosa control were devel-
oped and can be applied to the management of other
damaging infestations.
Acknowledgments
The NT Department of Primary Industry and Fish-
eries started mimosa control in the Mary River dis-
trict in 1993–94 as a result of government policy.
This proved to be the basis of a successful program
on Melaleuca Station. The Mimosa pigra subsidy
scheme provided the resources to keep the program
moving forward. Without this, the program would
have been much smaller. The directors of Melaleuca
have also shown a great deal of faith in the program
and allowing it to continue with little in the way of
economic return to date.
Also see the information sheet on this CD:
Native species for revegetation
