Journal of the Royal Society of Western Australia, 86(1), March 2003 
has no connection with the ocean. Here, inland 
mangroves, mostly Rhizophora stylosa, comprise 9.2 % of 
the island area. 
In the Indian Ocean, Van Steenis (1984) described a 
unique stand of mangroves (Bruguiera sp) on the eastern 
shore terrace of Christmas Island. The stand is 0.33 ha, 
120 m inland and occurs between the elevations of 24 to 
37 m above sea level (Woodroffe 1988). Corals in growth 
position on the terrace were dated to the last interglacial, 
suggesting that this stand of mangroves has persisted as 
a shoreline relict during the last 120 000 years of lower 
sea level. The stand occurs at a freshwater spring, where 
the mean annual rainfall is 2 000 mm. Large Bruguiera 
here indicate that the genus grows well in fresh water, 
with good regeneration. 
In northern Irian Jaya, Van Steenis (1963) 
described the mangrove Sonneratia caseolaris growing 
at a freshwater lake around 75 m above sea level. 
This is attributed to the geological uplifting of 
northern New Guinea. The same species has been 
observed growing inland at Timika in southern Irian 
Jaya, close to the Mile 21 replanting station of 
Freeport mine (J Ellison, University of Tasmania, 
personal observation). This location is 10 m above 
sea level, on a riverbank, and about 10 km inland 
from the estuarine mangroves. 
In Western Australia, Beard (1967) described a large 
mangrove community 25 km inland of the 80 Mile Beach, 
at Mandora. The location is a salt creek in limestone, 
which is lined by mangrove trees of up to 5 m (Avicennia 
marina). There is no tidal connection with the ocean at 
the site. 
These examples all indicate that mangroves occurring 
inland have been cut off from the open ocean by 
geological enclosure or sea-level change, most commonly 
on limestone islands. Thus they would have been isolated 
from the coastal mangrove genetic pool since the cut off 
time. The characteristics of inland mangroves at Lake 
MacLeod are investigated here. 
The objective of this study was to establish a baseline 
survey of the mangroves on Lake MacLeod, to 
characterise these unique mangroves, and to contribute a 
starting point for long-term monitoring. Monitoring sites 
were established at three representative areas of the 
mangrove margins, at the vents, on an island, and in the 
southern part of Ibis Pond. Measurements were made of 
stand structure, crown cover, seasonal production and 
phenology, annual production, growth rates, and plot 
biomass. 
Methods 
Lake MacLeod 
Lake MacLeod is situated in the Carnarvon Basin at 
23° 36-24° 38' S, 113° 30-113° 55' E (Fig 1), and is 
composed of calcareous marine deposits. It is a former 
sea embayment that was separated from the sea at the 
south end by development of sand dunes and mid- 
Holocene sea-level fail. There are few data on the past 
sea-level patterns from the region, although a mid- 
Holocene highstand has been modelled, reaching a 
maximum of 2 m above present sea level at Exmouth 
Gulf around 6 000 years BP (Lambeck & Nakada 1990). 
To the west of the lake is the Tertiary limestone Quobba 
Ridge, forming high sea cliffs in places, overlain with 
sand. 
Ocean water passes underground 18 km through the 
limestone barrier, driven by a hydraulic head of 3-4 m to 
rise in sinkholes ("vents") in the central west part of the 
lake bed (Handforth et al. 1984). The lake surface is at an 
elevation of 3-4 m below sea level, and consists of Cygnet 
(north) and Ibis (south) Ponds (Fig. 1). The sinkholes, 
outflow channels and ponds are permanent, although the 
southern lake varies in area depending on factors such as 
rainfall or prevailing winds. The vents are several meters 
in depth; the central pond is ca 1 m deep, and the 
southern pond cu 1.5 m deep. The salinity at the vents is 
close to seawater, with a salinity gradient rising towards 
the southern pond. 
Freshwater enters Lake MacLeod from several creeks 
and rivers. Surface inflow from Lyndon and Minilya 
Rivers is episodic (substantial discharge) or near-seasonal 
(only the nearby lake-bed inundated). Flooding from the 
Gascoyne River is infrequent, probably once in ten years, 
but it can cause the lake to be extensively inundated 
(such as 1989) or to fill (such as 1980 and 2000). However, 
the climate is arid. Median and mean annual rainfall at 
Gnaraloo (see Fig 1) are 203 mm and 230 mm 
respectively, mostly falling in May-July, and annual 
evaporation across the site is ca 2800-3000 mm (Lane et 
al. 1986). 
26 
