Growth of aquatic alligator weed ( Alternanthera philoxeroides ) over 5 years in south-east Australia

The largest known infestation of alligator weed Alternanthera philoxeroides (Mart.) Griseb. in Victoria, Australia, was reported in January 2009 in an urban pond. To determine how long the infestation had been present, high quality digital aerial images (orthophotos) were gathered for the site. Since all infestations are subject to ongoing eradication programs in Victoria, historical orthophoto records provide a unique opportunity for retrospective analysis, to calculate and report on the uncontrolled growth of aquatic alligator weed. Using geographic information system software (GIS), orthophotos were visually assessed to delineate the area of infestation for each year, from which annual increases in area were calculated. The infestation increased in area from ca. 0.029 ha in December 2004 to ca. 0.73 ha in December 2009, to cover 33% of the 2.2 ha water body. The annual area expansion was 200% for the first year of record. This reduced each year, to 22% at the end of the five year period. The mean lateral rate of expansion for floating alligator weed over the five years was 4.3 m (SD 2.2) annually. The average biomass of alligator weed at the site in summer 2010 was 4.9 kg dry weight m. Using the area of infestation from the December 2009 orthophoto the total estimated biomass in the pond equated to 35.6 tonnes dry weight.

Alligator weed Alternanthera philoxeroides (Mart.)Griseb. is a serious weed that has invaded a wide range of habitats in Australia and other regions of the world including the USA, New Zealand, China and other parts of Asia (Sainty et al. 1998).It poses a significant threat to Australia's waterways, wetlands, floodplains and irrigation systems and has the potential to become far more widespread.Modelling has shown that most of eastern and southern areas of continental Australia are suitable for its growth (Julien et al. 1995).Alligator weed is a stoloniferous and rhizomatous perennial that grows rapidly in both terrestrial and aquatic habitats (Sainty et al. 1998).In aquatic situations floating mats can cover waterbodies, restricting human use, excluding desirable plant species, interfering with aquatic ecology and restricting water flow (Julien et al. 1992).Rooted in the soil near the water's edge or in the substrate beneath shallow water, mats of interwoven horizontal stems, from which upright stems arise, float and extend over the surface of deeper water (Julien 1995).Alligator weed does not reproduce sexually in its introduced range, its primary means of dispersal is through stem fragmentation where stem fragments break off or floating mats dislodge and drift, lodging elsewhere and creating new infestations (Julien et al. 1992 and Dugdale et al. in press).Julien et al. (1995) reported that alligator weed grows best and forms dense monospecific stands in sub-tropical to cool, but not cold, temperature climates.In cooler, higher latitude regions the shorter growing season and occurrence of frosts kills top growth and restricts biomass accumulation.Southern Victoria, Australia, 37°S, provides warm summer growth periods (December-March) and cool winters with infrequent frosts, where growth of alligator weed ceases.
Alligator weed was first recorded in Victoria, Australia, in 1995, being grown as a vegetable (Gunasekera and Bonilla 2001), and naturalised populations were subsequently found in 1997 (Gunasekera and Adair 1999).It is declared a high priority weed being targeted for eradication from the state.Currently, alligator weed infestations in Victoria are concentrated within the Melbourne metropolitan region, although there are a few outlier infestations.The largest known infestation of alligator weed in Victoria was reported to authorities in January 2009 in an urban pond in Springvale (37°57'18.65"S;145°10'10.06"E,Figure 1).The site consists of a 2.2 ha standing water body that is used for stormwater retarding and irrigation purposes.The pond receives high nutrient, urban stormwater runoff from Mile Creek (total nitrogen and phosphorus averaged 1.48 mg L -1 , and 0.073 mg L -1 , respectively, from monthly low flow sampling during the five year period (Melbourne Water 2010)).
There have been very few published studies reporting the growth rate of aquatic alligator weed, particularly over consecutive years (Julien et al. 1992), most likely due to control programs being enacted soon after detection.High resolution (15-35 cm) digital aerial images (orthophotos) dating back five consecutive years were available for the study site, providing an opportunity to document the uncontrolled growth and expansion of aquatic alligator weed.Aerial photography was the first remote sensing method to be employed for studying and mapping vegetation, with early studies dating back to the 1960s and 1970s (Silva et al. 2008).Aerial photography interpretation (API) has been used extensively around the world to detect changes in species composition and distribution and to evaluate estimates of habitat features when accompanied with complementary field investigations.API allows for comparison of aerial images taken over time, providing a cost effective, time efficient tool to detect and analyse change, while providing information from the past that has not previously been recorded (Fitzgerald et al. 2006).The limitations of API are varied; the process requires suitable aerial photography relevant to the question of interest, and an analyst to identify key elements of the image (including tone, colour, pattern, shape, shadow, texture, contrast) for the species or habitat of interest.Most analyses of aerial imagery rely on visual interpretation where plant species can be discriminated when using high spatial resolution images (Silva et al. 2008), although identifying boundaries between vegetation community types is a recognised problem (Adam et al. 2010).confirmed that all patches of alligator weed present in the pond were visible and comparable to the orthophotos.A thorough search of the inlet stream and associated water bodies upstream of the site was also conducted, which confirmed that this was the most upstream infestation of alligator weed in the catchment.Using ArcView® GIS software, orthophotos were visually assessed to delineate the area of infestation for each year.A polygon layer was created defining the outer boundary of the pond, using the February 2001 orthophoto.Using orthophotos from December 2004 to December 2009, polygons were then created outlining the extent of each infestation for each subsequent year (Figure 2).All polygon layers were created using a viewing scale between 1:50 to 1:125, dependent on image quality.API depends on the subjective judgement of the interpreter and the quality of photographs used (Finkbeiner et al. 2001and McGlone 2004cited in Zhu et al. 2007), and variability was minimised by using the same interpreter for all orthophoto analysis.Infestation polygons were clipped with the February 2001 outline polygon so that only the alligator weed present within the pond was included in the assessment.Using the outline of the pond as a boundary means any alligator weed growing on the margins was ignored and therefore the area of alligator weed present was underestimated, restricting reporting to the increase of infested area within the pond alone.The area of each patch of alligator weed was calculated, giving an estimated total area of the infestation for each consecutive year.The annual area expansion was determined by calculating the difference between the areas occupied in two consecutive years and dividing this value by the area occupied in the first of those years.These values were normalised to reflect "annual periods" associated with the differing number of days and months between available orthophotos.
No alligator weed was detected in the February 2001 orthophoto.Orthophotos were not available thereafter until December 2004, when two small infestations could be detected (Figure 2).It has been observed (Sainty et al. 1998) that a few stems at a site can grow into a patch 1-3 m     (1992) report that over-water biomass can double, or increase by 100%, in 41 days during the growing season.Over the next four years the infestation increased and by December 2009 was calculated to be ca.7,300 m 2 .Although the annual area expansion had decreased to 22%, there was still an increase in area of ca.1,200 m 2 during the last annual period of record (Figure 3).This represents an increase from 1.3% of the 2.2 ha pond covered by alligator weed to 32.7% in a five year period.As can be seen from the data, floating alligator weed mats are capable of extremely rapid growth in Victoria.
The annual rate of lateral expansion for alligator weed at the site was also determined.Infestations present were measured from a standard shoreline point along a perpendicular transect to the outermost edge of the weed bed for each year of orthophoto.The mean annual lateral expansion during the five year period was 4.3 m (SD 2.2) (Table 1).Mean values for the December 2004 to December 2005 annual period are based on the two infestations present in December 2004, and when new infestations appeared throughout the next five years data were added.By the January 2009 to December 2009 annual period, mean values were based on six individual infestations (Table 1).In December 2009 the interwoven mats of plant material extended laterally out to a maximum of ca. 31 m from the embankment.Julien et al. (1992) report mats growing to 70 m wide and Spencer and Coulson (1976) report that mats can extend ca. 100 m over the water surface.
Sampling was conducted in February 2010 to determine the biomass of alligator weed present.Six transects spread evenly around the pond were selected and transects traversed infestations from the shore to the outward edge of the floating mats.Twenty samples in total were taken at 5m intervals spread evenly across the six transects.Biomass samples were harvested by wading out on top of the floating alligator weed mats and cutting through the entire depth of the weed bed within a square-sectioned tube (0.0572 m 2 , 1.0 m in length).All plant material within the core was sorted into above-and below-water portions (emergent and submersed), placed into bags, cleaned later the same day, weighed wet and placed into drying ovens at 80°C until constant dry weight was achieved.At each sampling point, water depth, depth of the submersed alligator weed and height of the emergent alligator weed was recorded (Table 2).
Biomass averaged 4,867 g dry weight m -2 (SD 899) (Table 2) or 43.03 kg wet weight m -2 (SD 7.90).In December 2009 the infestation measured ca.7,300 m 2 , and using the average biomass from February 2010 the total estimated biomass in the pond was 35.6 tonnes dry weight (Table 2).Biomass values reported here are much higher than previously reported in the literature for aquatic alligator weed.In Sydney, Australia, peak biomass was attained in late summer and was 3,214 g m -2 for a freshwater site and 3,252 g m -2 for an estuarine site (Julien et al. 1992), while in South Carolina biomass peaked at 392 g m -2 in late summer for a stream infestation (Davis et al. 1983cited in Julien et al. 1992).Pesacreta (1999) reported mean biomass of ca.600 g m -2 in control plots in North Carolina between June and July 1998.On the margins of a lake in northern New Zealand, above ground biomass in patches of alligator weed ranged from ca. 700 to 1,700 g m -2 (Bassett et al. 2010).A study in Oksiopat Lake (Bishnupur), Manipur, India reported biomass between 19.94 g m -2 and 139.41 g m -2 over a two year study period (Devi and Sharma 2010).Other studies of biomass taken from lakes in India have reported biomass of less than 250 g m -2 (reviewed by Devi and Sharma 2010).From the results of this study we have shown that biomass accumulation of alligator weed can be high in Victoria if left uncontrolled.
The aquatic ecotype of alligator weed forms mats of entangled stems and has adventitious roots that may be attached to the substrate or the bank, or in deeper water may be free floating.With each consecutive season of growth new stems are produced from prostrate mats from the previous season's growth (Julien et al. 1992).Dense aquatic alligator weed infestations are reported to consist of mats of older stems up to 0.3 m thick, supporting erect stems up to 0.8 m tall (Julien, 1995), and may extend 1 m or more down into the water (Spencer and Coulson 1976).In our study the submersed alligator weed (including root material) accounted for an average of 60.4% of the biomass at each sampling site, while the overall vertical extent of the alligator weed mat averaged 1.64 m (Table 2).This very dense growth was robust enough to support two people wading over the weed bed without sinking more than ca.0.25 m into the water and is probably a result of good growing conditions provided by stable and slow moving water conditions, low interspecific competition and abundant nutrient input.The water depth at the sampling points ranged from 2-4 m and all alligator weed was free floating (Table 2).There were weak positive linear correlations between water depth and biomass (R² = 0.115), and distance from shore and biomass (R² = 0.1573), indicating there was no net accumulation of alligator weed biomass in the areas of the mat that have been intact the longest, i.e. those nearest the embankment.
This study demonstrates the potential growth rate of alligator weed if left unchecked in Victoria and provides an example of the insight that retrospective use of aerial photos can provide.A continued emphasis on the implementation and development of early detection programs and containment of new outbreaks when infestations are small should remain a priority in Victoria, minimising long term costs and problems caused by uncontrolled infestations.

Figure 1 .
Figure 1.The authors sampling an infestation of floating alligator weed at the study site.

Figure 3 .
Figure 3. Area of alligator weed present and annual increase in area.Bars represent the area of alligator weed present at the study site for each year of growth (bottom x axis).Line represents the percentage annual increase in area (top x axis) over the 5 year growth period.Values have been normalised to reflect annual periods.

Table 1 .
Annual lateral expansion (m) of alligator weed over a five year period.Values have been normalised to reflect annual periods.

Table 2 .
Means and ranges for water depth and alligator weed infestation characteristics at Springvale in February 2010.