Fuller, R.A., Wilson, H.B., Kendall, B.E. & Possingham, H.P. 2009. Monitoring shorebirds using counts by the Queensland Wader Study Group. A report to the Queensland Wader Study Group and the Department of Environment and Resource Management. Brisbane, Australia.
Very little is known about how Australia’s biodiversity is changing. Here, we (i) summarize the distribution in space and time of counts conducted, and entered into a database, by the Queensland Wader Study Group, (ii) show how disturbance data collected as part of the surveys could be used to guide future management of Moreton Bay, (iii) develop two mathematical models that are powerful at detecting trends in the data, and (iv) assess the power to detect changes in shorebird population sizes, particularly in Moreton Bay, but also elsewhere in Queensland. We believe that the QWSG database is very valuable for informing the public and making management decisions in the future.
There is significant spatial and temporal variation in wader count coverage. However the number of counts reached a fairly consistent level around 1992–1995. This yields very useful time series data for waders in Moreton Bay from 1992 to 2008, although there is still significant variation in count effort within that period. We develop a method to account for this variation in effort when modelling shorebird population changes over time.
We investigate the utility of disturbance data collected in Moreton Bay as part of the QWSG counts. Overall, roost sites within national park zones showed a much lower frequency of disturbance than roosts in less stringent management categories. We map the causes of disturbance and show that these maps could be used to guide management and enforcement activity within Moreton Bay. Data on rates of disturbance could usefully be collected by QWSG observers during counts.
We show that a periodic model that accounts for seasonal fluctuations in bird numbers fits the QWSG dataset very well. We use this to help fit a between-year population model that accounts for different forms of stochastic variability in the data. We find that for the 22 most common migratory shorebirds, 7 show negative trends and 2 positive. The species declining show a loss of bird numbers of between 45% and 79%. None of the resident species shows significant change.
Power analyses are used to determine whether or not data collected is likely to be able to detect effects, e.g. changes in abundance, when they actually occur. Low statistical power means that real changes are likely to go unnoticed. A power analysis of the data reveals substantial power to detect past changes in shorebird numbers, but moderate to low power to predict whether these will continue into the future. More years of data are required to answer that question.
Based on our investigations, we make a number of recommendations for the way forward. Most importantly, we congratulate those involved in the assembly and maintenance of this important dataset, and urge a full analysis of the data to identify trends across species and sites, and potentially guide conservation management interventions.