Local to national – the capacity for increasing the spatial scale of monitoring PDF Print E-mail


On this page you will find:




Project overview


The measurement and assessment of stream ecosystem health is usually undertaken at the local or site scale, while management actions are undertaken at a more regional or catchment scale. Local scale assessments can be reported at the catchment scale by aggregating site based data for a number of sites within a catchment. However, the manner in which this aggregation is undertaken differs for different programs and, depending on the arrangement and position of assessed sites, can provide widely differing overall assessments of catchment health. These sites-based assessments also preclude predicting ecosystem health along the stream network in areas where data are lacking. Recently, a number of studies have used landscape scale spatial data and site based indicator data to model temperature (Isaak et al. 2010), water quality and habitat (Perry & Bond 2009) along stream networks to assess the vulnerability, conservation risk and long-term persistence of a range of freshwater taxa at the regional scale. At an even larger scale, readily available large-scale global datasets have been used to undertake broad based assessments of water security and health – with predictions relevant at the national scale (see Vörösmarty et al. 2010). This project proposal is to convene a workshop with the aim exploring the feasibility of a spatial modelling approach to predict ecosystem health (comprising a number of separate indicators or metrics) along stream networks for entire catchments. This would provide a more robust mechanism for regional stream health reporting and allow reporting at the scale of management and restoration.


click to enlarge
Seining for fish (photo: Fran Sheldon)


click to enlarge
Flinders River (photo: Fran Sheldon)


click to enlarge
Murray River (photo: Doug Ward)



return to top


Products and outcomes


Final report


Local to National – the capacity for increasing the spatial scale of monitoring

FINAL REPORT available for download [0.7MB]

Download PDF



return to top



Workshop report (8-10 March 2012)


Monitoring is a key component of broad-scale environmental management, providing compliance, early-warning and diagnostic capabilities.  In the freshwater realm, monitoring is used for water supply, conservation planning, water quality, state of environment reporting, assessing management actions, and detecting disturbance. Monitoring is usually undertaken at individual sites, which are selected using a probability-based survey design. A “design-based” approach is typically used to provide a condition or trend assessment, with indicators spatially aggregated (e.g. averaged or summarised) to provide an unbiased assessment at broader spatial scales. Yet, managers often require detailed information about where poor conditions exist so that management interventions may be targeted appropriately. Statistical or process-based models can be used to address this need since they can be used to predict condition or trend, with estimates of uncertainty, at sites that were not sampled. However, a “model-based” approach is rarely used for assessment because the assessment is based on the assumption that the model is right. In other words, there is no guarantee that the assessment reflects reality if the model is wrong. Although there are advantages and disadvantages to each approach, a hybrid of the two approaches is rarely used. The goal of our workshop was to explore the possibility of using a hybrid-monitoring strategy that takes advantage of model-based methods to extract additional information from unbiased site-based assessments to aid managers in making more informed and cost-effective management decisions.  Experts from fields such as statistics, computing and ecology came together to consider these issues for our workshop.  The sixteen participants were from CSIRO, state government agencies, and universities, from the USA, Switzerland, Canada, NSW, TAS, QLD, ACT and VIC.


Although it was recognized that there are significant policy challenges, such as funding continuity, program inertia, and the perception of the social value of monitoring associated with implementing a hybrid monitoring program at a regional or national scale, the group primarily focussed on the technical (e.g. ecological, spatial, and statistical) challenges that must be considered before a hybrid monitoring program can be implemented.  Implications associated with survey design, indicator selection, indicator aggregation and reporting emerged. For example, the objective of many survey designs is to spread points out in space in order to reduce the amount of redundant information used in the condition assessment, whereas it may be useful to have some sites in close proximity to one another if the objective is prediction. In some cases, alternative methods already exist that would successfully meet multiple needs in a hybrid monitoring program, but in other cases new monitoring methodologies would need to be developed to realize the full advantages of a hybrid approach. However, if these challenges can be overcome, the modelling component could be used to provide more spatially-explicit assessments that account for uncertainty, to test alternative management scenarios, to feed-back into an adaptive survey design, and to identify process-knowledge gaps. While the focus of this group was on freshwater systems, the hybrid monitoring strategy we explored has applicability to other systems, such as terrestrial landscapes, estuaries and even marine systems.


click to enlarge
back row l-r: Don Stevens, Tim Page, Simone Langhans, Melissa Dobbie, Stuart Bunn, Lee Belbin, Peter Negus. middle row l-r: Jon Olley, Janet Stein, Wayne Robinson, Bruce Chessman, Nick Bond, Ben Stewart-Koster, front row l-r: Fran Sheldon, Erin Peterson and Trefor Reynoldson


return to top

Last Updated on Sunday, 08 February 2015 16:05