Sessional Committee on Environment and Sustainable Development Written Submissions Received Volume 2 Issues associated with the progressive entry into the Northern Territory of Cane Toads October 2003
Tabled Paper 1123
Tabled Papers for 9th Assembly 2001 - 2005; Tabled papers for 9th Assembly 2001 - 2005; Tabled Papers; ParliamentNT
Tabled by Delia Lawrie
Made available by the Legislative Assembly of the Northern Territory under Standing Order 240. Where copyright subsists with a third party it remains with the original owner and permission may be required to reuse the material.
Written Submissions Professor Grigg Volume 2 Cane Toad Inquiry Report 95 Monitoring calling activity of frogs over long periods is not simple, especially as we wanted a method that recorded hours of data night after night in order to accommodate the variability in the wet-dry tropics. Most frog species are active only after heavy rainstorms, which occur patchily in both time and space. We therefore developed and deployed a novel automated censusing technique using machine learning technology to identify frogs by their calls. Identified calls are logged to memory automatically, along with environmental data. The information is then downloaded and analysed at the end of each wet season. The monitoring systems are (except for two) mounted in hollow steel poles approximately 5m high. A solar panel, microphone, rain gauge, microphone and thermometer are mounted on the top of the pole, which contains the recording device and rechargeable batteries. At two sites we have mounted the equipment on a metal tripod/quadripod, for easier installation. Throughout the wet season, each station turns itself on at dusk and logs calling intensity, rainfall and temperatures for 3-6 hrs, at intervals of approximately 11 minutes. The software is capable of identifying the calls of more than 20 species of frogs, as well as Cane Toads, on an intensity scale of 0-3. Monitoring at the Roper Valley sites started in the 1997-98 wet season. All of these sites are now within the toad's expanded range, indeed the invasion happened much more rapidly that we would have liked. We can now report a preliminary analysis of the results of the arrival of the toads. Monitoring within Kakadu National Park commenced in the 1998-1999 wet season, with four sites active. Before the 2000-2001 wet season we installed two more sites. During the current wet season, which has seen the arrival of the toads at Jabiru, we have six sites active. So for Kakadu, at this stage, we have four wet seasons of 'before toad' data at four sites and two seasons at the two newest sites. The results we got from the Roper Valley study area were provocative but not conclusive, so the Kakadu data will be very important. Results from the Roper Valley study area The experiment followed the general principles of a BACI ('before-after, controlimpact) design. We have five pairs of recording stations along the Roper Highway, about 25 km apart, from Mataranka to the end of the bitumen 110 kilometres east. We hoped that pairs of sites would be successively overrun by toads each year, so that we would have before toad-arrival and after toad-arrival data for most pairs, and data in most years from sites with both toads present and toads absent, so that we could allow for overall between-year variation in frog calling frequency. The much more rapid invasion of the toads than we expected, or wanted, meant that the western pairs were all engulfed within the space of the third wet season, so we had a shorter than desirable before-toad phase. Also, equipment failures in this developmental system led to some sites at which data were not recorded, leaving gaps in the data. However, a very large amount of data was collected, and some patterns emerge. The number of frog species calling per station declined markedly between the beginning of the study in 1997-98 and 2001-2002 (Figure 1). This pattern was consistent at each of the 10 stations. However, before this decline can be attributed unequivocally to the invasion of cane toads, a number of potentially confounding variables need to be considered. The shading on the graph