RRRD016: Developing integrated assessment metrics for reporting of water quality in the Great Barrier Reef lagoon

Vittorio E Brando1, Michelle Devlin2, Melissa Dobbie1, Aaron MacNeil3, Britta Schaffelke3 and Thomas Schroeder1

1 CSIRO
2 TropWATER James Cook University
3 Australian Institute of Marine Science

pdfDownload the RRRD016 Research Outcomes Report4.44 MB

Executive Summary

This project aimed to develop integrated assessment metrics for the reporting of water quality in the Great Barrier Reef (GBR) lagoon. The Reef Water Quality Protection Plan, including the Australian Government funded Reef Programme (formerly Reef Rescue), have the express purpose of improving the condition of water quality and ecosystem health in the Great Barrier Reef Marine Park. Long-term monitoring of spatial and temporal trends in broad-scale water quality within the GBR Marine Park is central to assessment of Reef Plan performance. This monitoring has been carried out since 2005 with funding from the Australian Government. The MMP is part of the Paddock to Reef (P2R) Integrated Monitoring Modelling and Reporting Program, which is a key action under Reef Plan to evaluate the efficiency of the implementation of Reef Plan and the progress towards its goal and targets.

The MMP water quality monitoring uses a combination of three complementary approaches to collect data at various spatial (site, location, region, and whole GBR lagoon) and temporal (snapshot, daily, 10-minute) scales, using traditional direct water sampling from research vessels, in situ data loggers at a small number of selected inshore reef locations and remote sensing techniques. Two indicator variables, suspended solids (an indicator of water clarity) and the plant pigment chlorophyll (an indicator of phytoplankton biomass and a proxy for nutrient availability), are measured by all three techniques. In addition, direct water sampling provides information on a wider suite of water quality variables, such as nutrients. While data loggers provide detailed information on the local variability in water quality parameters, remote sensing observations provide extensive spatial coverage at 1 km resolution, nominally on a daily basis (except e.g. overcast days).

Given the spatial and temporal variability of the data, an integrated framework that enables a comprehensive assessment of condition and trends of GBR water quality is required to provide environmental managers and policy makers with the information needed to make decisions for the management of land runoff. The integrated assessment method developed as part of this project combines a selection of key indicators to enable a reasonable evaluation of the overall status of coastal and marine waters. The assessment framework was developed to include the analyses of spatial and temporal variability of each data source, the development of metrics/indices that combine and scale up the three approaches used to collect water quality data (direct sampling, in situ data loggers and remote sensing), the estimation of uncertainty of combined indices, as well as the development of statistical methods to assess performance with regard to the existing GBR Water Quality Guidelines (GBRMPA, 2010).

The development of the water quality assessment and reporting framework was closely aligned with other nationally and international approaches for ecological monitoring, assessment and reporting. A review was carried out to assess the suitability for application in the GBR of national and international marine water quality assessment frameworks and their approaches and methods. This review found that the monitoring and reporting was often carried out at high temporal frequency and in small reporting areas where it should be reasonable to expect measurable change in response to variations of the environmental pressures that affect marine water quality.

In light of the findings of the review, we defined the criteria to determine the appropriate spatial and temporal scale for analysis and reporting for the assessment framework:

  • In the GBR, the direct influence of land run-off occurs at a smaller scale than at the scale of entire marine NRM regions, as currently used for P2R reporting. Hence we propose to define smaller regions that are directly influenced by land run-off and consistent with local oceanography (e.g., residence times, bathymetry, hydrodynamics). In these smaller reporting areas it is reasonable to expect measurable change over the next decade in response to improvement in land management practices on the GBR catchment.
  • To allow consideration of the variable influences during the dry and wet season, we propose to assess the condition of water relative to the seasonal guidelines (four seasons per year). This should provide an improved capacity to untangle the effects of land management practices from those directly due to the climate events, overcoming some of the limitations of the current P2R reporting of annual water quality summary data.

To aggregate multiple sources of data across various spatial and temporal scales in the reporting framework, a spatio-temporal statistical process model was developed. This allowed modelling of the average "response surfaces" for each indicator as well as the associated uncertainties. These seasonal "average response surfaces" for each indicator can then be used to assess compliance to the guidelines as well as provide a measure of the year-to-year variability and, ultimately, long-term trends.

For the proof of concept of the reporting framework we selected an area with the most data, in both space and time, from direct water sampling. A small reporting region around the Tully River mouth/Rockingham Bay was defined, using the criteria outlined above, which encompassed most of these data points. The reporting year 2010-11 was selected for this study, as in this year extreme floods occurred along the whole of the GBR coast. By selecting the most data-rich region and the most extreme year we aimed to maximise the capacity of the proposed reporting framework to show seasonal and spatial differences.

For the proof of concept of the reporting framework, two parameters were selected: (i) Total Suspended Solids measured by remote sensing, direct water sampling both during floods and ambient conditions, and water quality instruments on coral reefs, and (ii) Dissolved Inorganic Nitrogen and Particulate Nitrogen values measured only by direct water sampling both during floods and ambient conditions.

We demonstrated that the approach would be fully compatible with the current Paddock to Reef reporting and builds on the current model of tiered reporting across regional areas. The implementation of the proposed framework would need to be considered in the context of potential future changes to the Paddock to Reef monitoring and reporting structure.

For this case study, we have focused on a small reporting area (Tully marine area) as it the most data rich in the GBR. Implementation on a regular basis for whole Great Barrier Reef World Heritage Area (GBRWHA) is a complex computational exercise. To be able to perform the data integration across temporal and spatial scales at a GBR scale, the spatio-temporal statistical process model developed for the proof of concept would need to be implemented in parallel computing environments.

The proof of concept highlighted several research and development needs related to the implementation of the reporting framework for the inshore waters across the whole GBRWHA:

  • The proposed framework included the definition of new smaller, localised coastal reporting regions to provide a better reporting base for water quality assessments based on the influence of the river flow and water quality associated with that water flow. These smaller reporting regions need to be defined, based on the decision-making criteria proposed in the assessment framework.
  • In the proposed framework, all variables were normalised and no weighting was applied to the combination of the water quality variables. Future applications of this approach needs to consider the weighting of the different water quality variables based on understanding of relevant ecological thresholds for GBR ecosystems such as corals and seagrasses. Weighting and/or thresholds need to be considered in both seasonal and geographical context
  • Further development is required of the water quality guidelines for the GBR to define trigger values that are specific for regional characteristics and/or for coral and seagrass environments.
  • The future integration of other water quality indicators, such as pesticides, will require an investigation of the appropriateness of the proposed framework for spatial and temporal data aggregations.