catchments (Hunter and Walton, 2008), where baseflow has been observed to ...... 90,400. 31,900. 219,000. 9,700. 4,410. 21,800. 10,900. 1,970. 46,700. 1991.
Sediment and nutrient load estimates for major Great Barrier Reef catchments (1987–2009) for Source Catchment model validation
Prepared by Marianna Joo, Vivienne McNeil, Chris Carroll, David Waters, Satish Choy Water Planning Ecology Science Division Department of Science, Information Technology, Innovation and the Arts PO Box 5078 Brisbane QLD 4001 © The State of Queensland (Department of Science, Information Technology, Innovation and the Arts) 2014. The Queensland Government supports and encourages the dissemination and exchange of its information. The copyright in this publication is licensed under a Creative Commons Attribution 3.0 Australia (CC BY) licence
Under this licence you are free, without having to seek permission from DSITIA, to use this publication in accordance with the licence terms. You must keep intact the copyright notice and attribute the State of Queensland, Department of Science, Information Technology, Innovation and the Arts as the source of the publication. For more information on this licence visit http://creativecommons.org/licenses/by/3.0/au/deed.en ISBN: 978-1-925075-09-0
Disclaimer This document has been prepared with all due diligence and care, based on the best available information at the time of publication. The department holds no responsibility for any errors or omissions within this document. Any decisions made by other parties based on this document are solely the responsibility of those parties. Information contained in this document is from a number of sources and, as such, does not necessarily represent government or departmental policy. If you need to access this document in a language other than English, please call the Translating and Interpreting Service (TIS National) on 131 450 and ask them to telephone Library Services on +61 7 3170 5725
Citation Joo M., McNeil, V., Carroll, C., Waters, D., Choy, S. 2014. Sediment and nutrient load estimates for major Great Barrier Reef catchments (1987–2009) for Source Catchment model validation. Brisbane: Department of Science, Information Technology, Innovation, and Arts, Queensland Government.
Acknowledgements This report has been prepared by the Department of Science, Information Technology, Innovation and the Arts. Acknowledgement is made of number of internal and external reviewers for their valued input during the development of this report. Thanks also due to the Reef Plan’s Great Barrier Reef Catchment Loads Monitoring Program (formerly the GBRI5 program), Australian Institute of Marine Sciences (AIMS), James Cook University (JCU) and the Great Barrier Reef Marine Park Authority (GBRMPA) for providing the data on which this study is based. Front page image: Burdekin River flood plume. Courtesy of Robert Packett, Queensland Department of Natural Resources and Mines April 2014
Sediment and nutrient load estimates for major Great Barrier Reef catchments (1987–2009) for Source Catchment model validation
Contents Introduction ................................................................................................................................... 1 Study area ..................................................................................................................................... 2 Data ................................................................................................................................................ 5 Flow and concentration data
5
Flow representativeness of concentration data
5
Methods ......................................................................................................................................... 9 Development of Flow Range Concentration Estimator (FRCE) method
9
Likely upper and likely lower load range (LUR, LLR) calculation
10
Estimating concentrations by FRCE
11
Load estimation
12
Evaluation of daily time step
12
Results......................................................................................................................................... 13 Discussion................................................................................................................................... 17 Variability in concentration predictions
17
Comparison of FRCE annual loads to the best estimates of annual loads
19
Recommendations and future applications
21
Conclusion .................................................................................................................................. 23 References .................................................................................................................................. 25 APPENDICES .............................................................................................................................. 29 Appendix 1 – Concentration data assessment regarding flow coverage
30
Appendix 2 – Sensitivity analysis of flow stratification scenarios
35
Appendix 3 – FRCE sediment and nutrient concentrations
37
Appendix 4 – Annual sediment and nutrient loads.
68
i
Department of Science, Information Technology, Innovation and the Arts
ii
Sediment and nutrient load estimates for major Great Barrier Reef catchments (1987–2009) for Source Catchment model validation
Introduction The Reef Water Quality Protection Plan (Reef Plan) is a collaborative project between the Queensland and Australian governments to halt and reverse the decline in water quality from anthropogenic sediment, nutrient and pesticide loads entering the Great Barrier Reef (GBR) lagoon by 2013, and to ensure that by 2020 the quality of water entering the reef from adjacent catchments has no detrimental impact on the health and resilience of the Great Barrier Reef (Reef Plan 2003, 2009). This is accomplished through setting water quality targets against estimated anthropogenic baseline load (total load – pre-development load) and implementing best land management practices. To measure and report on progress towards the set water quality targets, a program has been established that combines monitoring and modelling at paddock through to catchment and reef scales; namely the Paddock to Reef Integrated Monitoring, Modelling and Reporting Program (Paddock to Reef Program) (Carroll et al. 2012). The eWater CRC Source Catchments modelling framework is used to report on progress towards reef water quality targets, and compensates for seasonal climatic variability by using a representative climatic period between July 1986–June 2009 (23 years). Six Source Catchment models have been produced for each of the Natural Resource Management regions adjacent to the GBR lagoon: Cape York, Wet Tropics, Burdekin Dry Tropics, Mackay–Whitsunday, Fitzroy and the Burnett–Mary. The Source Catchment models generate sediment, nutrient, and pesticide loads entering the GBR lagoon from the regions, and the changes in loads due to industry and government investments in improved land management practices (Carroll et al. 2012). Water quality monitoring is a critical point of truth for model validation in the Paddock to Reef program, and ensures continual refinement of the models whilst at the same time identifying data gaps and priorities for future monitoring. One of the Source Catchments’ important attributes is that modelling can be undertaken to compare loads from different locations and for various time periods, consequently quite disparate water quality datasets can be used for validation purposes. Under the Reef Plan, the GBR Catchment Loads Monitoring Program has collected water quality data since 2006 from high priority catchment and sub catchment sites; and this is a critical data set that is used to validated Source Catchment outputs over a relatively short time period. An additional validation for the 23-year modelled period is also required to provide confidence in the model outputs. However, pre-2006 concentration data was generally sporadically collected and often not sampled for critical parts of the hydrograph. Therefore, a method is required to estimate concentrations and loads for the entire 23-year modelled period. As most constituent loads are delivered under high flow events and most parameters are flow dependent, the accuracy of the load estimates based on water quality data are reliant on the frequency and flow representativeness of the constituent concentration data (Joo 2012; Joo et al. 2012). High frequency data collected since 2006 under the Reef Plan’s GBR Catchment Loads Monitoring Program will support the application of integrated tools, such as the Loads Tool, incorporated in the Water Quality Analyser software, (Tennakoon et al. 2007) to calculate current and future load estimates with greater certainty. However, an alternative approach is required for sporadically collected concentration data pre-2006 to support the earlier phase of the climatic cycle. The standard methods often used for this purpose were explored, including averaging, developing concentration to flow relationships (regression), the recently developed Load Regression Estimator (LRE), and the Beale Ratio (Joo 2012; Wang et al. 2011; Kuhnert et al. 2012; Marsh and Waters 2009; Richards 1999). Following testing, the linear regression method was rejected since the assumption requiring linearity on a log scale was not valid for most
Page 1 of 87
Department of Science, Information Technology, Innovation and the Arts
constituents at most of the monitoring sites. The more complex LRE was also rejected given it requires intensive mathematical analyses of the data for each constituent at each site, as well as extensive sensitivity analyses to optimise the parameter values. It was evident that this approach could not be applied for such as large project with given resources and time, even if sufficient unbiased data were available to support such a procedure. Of the approaches considered it was found that, with some modifications, the Beale ratio was the most appropriate approach to develop a Flow Range Concentration Estimator (FRCE) method to derive loads for the 23-year period. The aim of this project, therefore, was twofold: The first aim was to calculate mean annual sediment and nutrient loads based on all appropriate water quality data for the modelled period July 1986–June 2009 (23 years) for model validation. In line with this first aim, the Flow Range Concentration Estimator (FRCE) method was developed, which was also validated against best estimates based on reliable time series data sets. The second aim was to also to provide the likely upper and lower ranges (LUR, LLR) of loads for the key constituents.
Study area Mean annual sediment and nutrient loads for the 1987–2009 (i.e. July 1986–June 2009) period are required at the end-of-system sites of nine rivers: the Normanby in the Cape York region; the Barron, North Johnstone, South Johnstone, Tully and Herbert in the Wet Tropics; the Pioneer in the Mackay–Whitsundays, and the Burdekin and Fitzroy in the Dry Tropics (Figure 1, Table 1) for model validation purposes. These catchments cover a wide range of geo-climatic conditions including small to medium sized, wet coastal sugarcane catchments in the Wet Tropics and Mackay–Whitsundays, to the large, dry inland grazing catchments in the Dry Tropics.
Page 2 of 87
Sediment and nutrient load estimates for major Great Barrier Reef catchments (1987–2009) for Source Catchment model validation
Figure 1. Great Barrier Reef river catchments included in this study.
Page 3 of 87
Department of Science, Information Technology, Innovation and the Arts
Table 1 Description of the Great Barrier Reef monitored catchments
Gauging Region station
Adopted middle End-of-system location thread distance
River
(km) 105107A
Cape York
Normanby
110001D
Wet Tropics
Barron
Catchment Mean area above annual gauging discharge station 1987–2009 2
(km )
Kalpowar
71.0
12,950
(GL) 2,176
a
Myola
27.1
1950
649
b
Tung Oil
28.5
950
1,751
b
Upstream Central Mill
18.5
400
784
112004A
Wet Tropics
North Johnstone
112101B
Wet Tropics
South Johnstone
113006A
Wet Tropics
Tully
Euramo
17.5
1450
2,934
116001E
Wet Tropics
Herbert
Ingham
30.1
8600
3,459
16.7
1450
742
17.4
129,760
8,618
142.1
139,300
3,791
36.6
32,850
482
125013A
Whitsundays
Pioneer
Dumbleton Pump Station Head Water
120001A
Dry Tropics
Burdekin
Home Hill
1300000
Dry Tropics
Fitzroy
Rockhampton
136014A
Dry Tropics
Burnett
Ben Anderson Barrage Head Water
a
c
d
e
– modelled flow using Source Catchments as provided by MacCloskey (2012) – Johnstone is considered as one catchment by the Reef Plan (2003, 2009) – estimated from flow data at Mirani Weir GS 125007A ( = QMW * 1.226; TimeDP = TimeMW + 1.5 hours) d – based on flow data at Clare GS 120006B e – based on flow data at Walla GS 136001B up to 1999 and modelled flow from 1999 onwards as provided by Fentie (2012) using Source Catchments. b
c
Page 4 of 87
Sediment and nutrient load estimates for major Great Barrier Reef catchments (1987–2009) for Source Catchment model validation
Data Flow and concentration data The Source Catchment model has been set up to predicts daily sediment, nutrient and pesticide loads (Carroll et al. 2012). Sensitivity analysis, discussed below, showed that the daily time step is adequate (as opposed to sub-daily) for annual and mean annual load estimates for the selected rivers. Therefore, for consistency, loads have been calculated at a daily timestep in this study to provide the annual load estimates, which then are integrated and averaged to obtain mean annual loads for model validation. Accordingly, daily flow records were collated for the model run period of 1987–2009 from the Department of Natural Resources and Mines (DNRM) gauging station data. For a number of end-of-system gauges included in this study, flow records were not available for the full period of interest. For these catchments, missing flow records were estimated using nearby gauging stations or modelled data as indicated in Table 1 above. For the Normanby River, since flow monitoring only commenced in 2006, daily flows were provided by the Source Catchment model, and similarly for the Burnett (McCloskey, Fentie, pers. comm). Sediment and nutrient loads were estimated for nine parameters, including total suspended solids (TSS), particulate nitrogen (PN), total nitrogen (TN), dissolved inorganic nitrogen (DIN), dissolved organic nitrogen (DON), total phosphorous (TP) particulate phosphorous (PP), dissolved inorganic phosphorous (DIP), dissolved organic phosphorous (DOP). These were collected for all sites except for the Burnett River, where particulate and organic nutrients (PN, PP, DON, DOP) were not available. Concentration data for these parameters have been collected on an opportunistic base by DNRM (formerly the Department of Environment and Resource Management) under the Surface Water Ambient Monitoring Network (SWAN) program since 1973, and through high frequency sampling programs by the Australian Institute of Marine Sciences (AIMS), James Cook University (JCU), and the Great Barrier Reef Marine Park Authority (GBRMPA) since 1986 and by the Reef Plan’s Great Barrier Reef Catchment Loads Monitoring Program (formerly the GBRI5 program) since 2006.
Flow representativeness of concentration data High flow data are the most significant for load estimation because, in addition to flow, the concentration of sediment and nutrients tend to be elevated (Brodie and Mitchell 2006). Therefore, to assess our confidence in the estimates, the concentration records with regards to flow representativeness were evaluated based on two criteria (Joo and Yu, 2012). These criteria were the number of samples collected from the upper two percentiles of the flow range, and also the percentage of the maximum recorded flow rate that was sampled (Tables 2 and 3). The second criterion was necessary because the upper two percentile of flow range still includes most of the actual flow volume for most rivers (Appendix 1). According to these criteria, individual scores were assigned to each parameter at each site and integrated to obtain an overall score and final ratings (Table 4).
Page 5 of 87
Department of Science, Information Technology, Innovation and the Arts
Table 2. Criteria for assessing the number of samples in the upper two percentile of flow range. Number of samples in the upper two percentile flow range 0–9
Score 1
10–19
2
20–29
3
30–39
4
> 40
5
Table 3. Criteria for assessing the proportion of maximum recorded flow rate sampled. Percent of maximum recorded flow rate sampled
Score
0–9
0.5
Percent of maximum recorded flow rate sampled 50–59
10–19
1
60–69
3.5
20–29
1.5
70–79
4
30–39
2
80–89
4.5
40–49
2.5
90–100
5
Table 4. Flow representativeness rating based on the overall scores.
Page 6 of 87
Overall score
Rating
8–10
excellent
6–7
good
4–5
moderate
