Modelling pollutant load changes due to improved management practices in the Great Barrier Reef catchments: updated methodology and results

Technical Report for Reef Report Card 2015

Prepared by Contact Gillian McCloskey and David Waters David Waters 07 4529 1395

Ownership of intellectual property rights

Unless otherwise noted, copyright (and any other intellectual property rights, if any) in this publication is owned by the State of Queensland.

Creative Commons

This material is licensed under a Creative Commons - Attribution 3.0 Australia licence. Creative Commons Attribution 3.0 Australia License is a standard form license agreement that allows you to copy, distribute, transmit and adapt this publication provided you attribute the work. A summary of the licence terms is available from www.creativecommons.org/licenses/by/3.0/au/deed.en.The full licence terms are available from www.creativecommons.org/licenses/by/3.0/au/legalcode.

To reference this volume McCloskey, G.L., Waters. D., Baheerathan, R., Darr, S., Dougall, C., Ellis, R., Fentie, B., Hateley, L. 2017. Modelling pollutant load changes due to improved management practices in the Great Barrier Reef catchments: updated methodology and results – Technical Report for Reef Report Cards 2015, Queensland Department of Natural Resources and Mines, Brisbane, Queensland.

Disclaimer

The information contained herein is subject to change without notice. The Queensland Government shall not be liable for technical or other errors or omissions contained herein. The reader/user accepts all risks and responsibility for losses, damages, costs and other consequences resulting directly or indirectly from using this information.

Acknowledgments: This project is part of the Queensland and Australian Government’s Paddock to Reef program. The project is funded by the Queensland Department of Natural Resources and Mines ‘Queensland Regional Natural Resource Management Investment Program 2013–2018’ with support from the Department of Science, Information Technology, Innovation and the Arts (DSITIA).

Technical support was provided by a number of organisations and individuals, including Environmental Monitoring and Assessment Science (DSITIA), Queensland Hydrology (DSITIA) and the Commonwealth Scientific and Industrial Research Organisation (CSIRO). We would also like to thank the reviewers for providing feedback on this report.

Source Catchments modelling – Updated methodology and results

Executive Summary

In this report the updated methodology and latest Great Barrier Reef (GBR) Report Card load reduction results are presented. Minor methodological changes have occurred since the release of the previous report card (Report Card 2014 (RC2014)). These include, recalibration of hydrology in a number of regions to better represent the baseflow portion of total flow, the implementation of instream deposition and remobilisation in three regions, and minor updates to adjustable parameters.

The recalibration of hydrology to align baseflow proportions with calculated values produced more typical ranges of baseflow particularly in the larger, drier, grazing catchments (Burdekin, Fitzroy and Cape York (CY)). In these regions many gauges were recalibrated and commonly the baseflow portion decreased from previous calibration proportions. The recalibration had negligible impact on average annual flow across the GBR with a small increase of 1% to 73,500,000 ML/yr.

Instream deposition and remobilisation functionality were implemented in the Wet Tropics (WT), Mackay Whitsunday (MW) and Burnett Mary (BM) regions for RC2015.

The average annual baseline loads for RC2015 are presented in Table 1. Generally the updated average annual baseline loads are higher than the RC2014 loads, except for the photosynthesis II toxic equivalent (PSII TE) pesticides which are approximately 10% lower than RC2014. The minor increase in loads from the previous Report Card are in part due to the recalibration of hydrology resulting in a minor increase in flow and therefore constituent loads. The decrease in PSII TE load from RC2014 is likely a product of a shift to alternative products.

Table 1 Baseline GBR constituent loads

TSS TN DIN DON PN TP DIP DOP PP PSII TE (kt/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (kg/yr)

9,925 55,083 12,028 18,303 24,753 13,418 2,143 1,074 10,201 6,758

The progress towards Reef Plan targets is presented in Figure 1. ‘Moderate’ progress has been made towards the fine sediment load reduction target of 20%. The cumulative reduction across the GBR is 12.3%, with the Burdekin region at 17.2%. ‘Very good’ progress is reported for Particulate Phosphorus (PP) reduction, which is well over half way towards the target at 14.7%. The WT is the only region to have reached the target having achieved a 20.7% reduction. For Particulate Nitrogen (PN), the cumulative reduction is 11.6%, which is classified as ‘moderate’ progress, with the Burdekin region showing the greatest progress with a reported 15.3% reduction thus far. Progress towards the DIN target of 50% is classified as ‘very poor’ as it stands at 18.1% for RC2015. Only the Burnett Mary region is making ‘good’ progress at 31.5%. Reductions in PSII TE load are rated as ‘moderate’ with a reduction of 33.7% across the GBR. Progress has been most improved in the MW region with a 44% reduction thus far.

3 Source Catchments modelling – Updated methodology and results

In summary, progress is being made for all constituents towards reaching the Reef Plan water quality targets. There is scope for greater reductions to be achieved managing DIN in the WT where progress is at 14.7% or ‘very poor’, and for sediment in the Fitzroy and BM regions, with both regions reporting ‘very poor’ progress at 5.5% and 3% respectively. The ratings of performance for load reductions are available at www.reefplan.qld.gov.au. It should be noted that the scoring system has changed since RC2014.

Figure 1 Progress towards Reef Plan targets till 2015 for the whole of GBR

4 Source Catchments modelling – Updated methodology and results

Contents

Executive Summary ...... 3

List of Tables ...... 11

List of Figures ...... 15

Acronyms ...... 18

Definitions of sources and sinks ...... 19

Modelling assumptions ...... 20

Introduction...... 22

Model Improvement Cycle ...... 23

Integrating modelling and monitoring ...... 24

Application of modelling outputs ...... 25

Quality assurance and peer review ...... 25

Methods...... 27

GBR methodology updates ...... 28

Hydrology ...... 28

Hillslope generation model ...... 29

Sugarcane constituent generation ...... 30

HowLeaky cropping constituent generation ...... 30

In-stream models ...... 30

Predevelopment Model ...... 31

Constituent load validation ...... 31

Performance ratings ...... 31

Scenario modelling – All A and All B management ...... 32

Baseline Model ...... 32

All B Scenario ...... 33

All A Scenario ...... 33

Cape York ...... 35

Hydrology ...... 35

Hillslope sediment, nutrient and herbicide generation ...... 35

Gully sediment and nutrient generation models ...... 35

Sugarcane constituent generation ...... 35

5 Source Catchments modelling – Updated methodology and results

HowLeaky cropping constituent generation ...... 35

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) ...... 35

In-stream models ...... 35

Predevelopment model ...... 35

Wet Tropics ...... 36

Hydrology ...... 36

Hillslope sediment, nutrient and herbicide generation ...... 36

Gully sediment and nutrient generation models ...... 36

Sugarcane constituent generation ...... 36

HowLeaky cropping constituent generation ...... 36

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) ...... 37

In-stream models ...... 37

Predevelopment model ...... 37

Burdekin...... 38

Hydrology ...... 38

Hillslope sediment, nutrient and herbicide generation ...... 38

Gully sediment and nutrient generation models ...... 38

Sugarcane constituent generation ...... 38

HowLeaky cropping constituent generation ...... 38

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) ...... 38

In-stream models ...... 38

Predevelopment model ...... 38

Mackay Whitsunday ...... 39

Hydrology ...... 39

Hillslope sediment, nutrient and herbicide generation ...... 39

Gully sediment and nutrient generation models ...... 39

Sugarcane constituent generation ...... 39

HowLeaky cropping constituent generation ...... 39

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) ...... 39

In-stream models ...... 39

Predevelopment model ...... 39

6 Source Catchments modelling – Updated methodology and results

Fitzroy ...... 40

Hydrology ...... 40

Hillslope sediment, nutrient and herbicide generation ...... 40

Gully sediment and nutrient generation models ...... 40

Sugarcane constituent generation ...... 40

HowLeaky cropping constituent generation ...... 40

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) ...... 41

In-stream models ...... 41

Predevelopment model ...... 42

Burnett Mary ...... 43

Hydrology ...... 43

Hillslope sediment, nutrient and herbicide generation ...... 43

Gully sediment and nutrient generation models ...... 43

Sugarcane constituent generation ...... 43

HowLeaky cropping constituent generation ...... 43

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) ...... 43

In-stream models ...... 43

Predevelopment model ...... 44

Results ...... 45

Hydrology and regional discharge comparison ...... 45

Modelled loads ...... 45

Contribution by land use ...... 49

Progress towards Reef Plan 2013 targets ...... 51

Scenario modelling – All A and All B management ...... 51

Cape York ...... 55

Hydrology calibration performance ...... 55

Modelled Loads ...... 55

Constituent load validation ...... 56

Contribution by land use ...... 58

Sources and sinks ...... 59

Progress towards Reef Plan 2013 targets ...... 60

7 Source Catchments modelling – Updated methodology and results

Wet Tropics ...... 61

Hydrology calibration performance ...... 61

Modelled Loads ...... 62

Constituent load validation ...... 64

Contribution by land use ...... 67

Sources and sinks ...... 69

Progress towards Reef Plan 2013 targets ...... 70

Burdekin...... 72

Hydrology calibration performance ...... 72

Modelled Loads ...... 72

Constituent load validation ...... 75

Contribution by land use ...... 75

Sources and sinks ...... 76

Progress towards Reef Plan 2013 targets ...... 77

Mackay Whitsunday ...... 81

Hydrology calibration performance ...... 81

Modelled Loads ...... 81

Constituent Load Validation ...... 84

Contribution by land use ...... 89

Sources and sinks ...... 90

Progress towards Reef Plan 2013 targets ...... 91

Fitzroy ...... 93

Hydrology calibration performance ...... 93

Modelled Loads ...... 94

Constituent load validation ...... 95

Contribution by land use ...... 99

Sources and sinks ...... 100

Progress towards Reef Plan 2013 targets ...... 101

Burnett Mary ...... 104

Hydrology calibration performance ...... 104

Modelled Loads ...... 104

8 Source Catchments modelling – Updated methodology and results

Constituent load validation ...... 106

Contribution by land use ...... 108

Sources and sinks ...... 109

Progress towards Reef Plan 2013 targets ...... 110

Discussion ...... 113

Particulate nutrients ...... 113

Progress towards targets ...... 114

Scenario modelling – All A and All B management ...... 115

Cape York ...... 116

Hydrology ...... 116

Fine sediment generation, sources and sinks ...... 116

Particulate nutrients ...... 116

Progress towards targets ...... 117

Wet Tropics ...... 118

Hydrology ...... 118

Fine sediment, particulates sources and sinks ...... 118

Nutrients ...... 119

Herbicides ...... 120

Progress towards targets ...... 120

Burdekin...... 121

Hydrology ...... 121

Fine sediment and particulates generation, sources and sinks ...... 121

Sugarcane constituent generation ...... 122

Herbicides ...... 122

Targeting land management for Reef WQ protection ...... 122

Progress towards targets ...... 123

Future model improvement ...... 123

Mackay Whitsunday ...... 124

Hydrology ...... 124

Constituent generation and validation ...... 125

Constituent sources and sinks ...... 126

9 Source Catchments modelling – Updated methodology and results

Progress towards targets ...... 126

Fitzroy ...... 128

Hydrology ...... 128

Fine sediment and particulates generation, sources and sinks ...... 128

Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) ...... 130

Particulate Nutrients ...... 131

Progress towards targets ...... 131

Burnett Mary ...... 132

Hydrology ...... 132

Fine sediment and particulate nutrient generation, sources and sinks ...... 132

Sugarcane dissolved nutrients and PSII loads ...... 133

Progress towards targets ...... 134

Conclusion...... 135

References ...... 136

Appendix 1 – Hydrology calibration results ...... 140

Cape York ...... 140

Wet Tropics ...... 141

Burdekin...... 142

Mackay Whitsunday ...... 144

Fitzroy ...... 148

Burnett Mary ...... 151

Appendix 2 – Dynamic SedNet global parameters and global requirements ...... 153

Grazing constituent generation ...... 153

In-stream models ...... 153

Sugarcane and cropping constituent generation ...... 154

Other land uses (EMC/DWC values) ...... 154

Appendix 3 – Report Card 2015 modelling results ...... 159

Appendix 4 – Model validation ...... 177

Wet Tropics model validation ...... 177

Burdekin model validation ...... 181

Mackay Whitsunday model validation ...... 183

10 Source Catchments modelling – Updated methodology and results

List of Tables

Table 1 Baseline GBR constituent loads ...... 3 Table 2 Summary of catchment modelling cycle ...... 24 Table 3 Model updates undertaken for Report Card 2015 ...... 27 Table 4 Number of gauges that underwent recalibration for baseflow ...... 29 Table 5 Performance ratings for recommended statistics for a monthly time-step (from (Moriasi et al. 2007), Moriasi et al. (2015)) ...... 32 Table 6 Summary of changes to sugarcane parameters in the WT ...... 37 Table 7 Summary of changes to sugarcane parameters in the BM ...... 43 Table 8 Total baseline loads for the GBR region ...... 46 Table 9 PSII pesticide load, PSII Toxic Equivalent (TE) load for the GBR region ...... 46 Table 10 Regional contribution of each constituent as a percentage of the GBR total baseline load .. 47 Table 11 Baseflow proportions (%) for RC2014, RC2015 and calculated for seven CY gauges ...... 55 Table 12 Contribution from CY basins to the total CY baseline load ...... 56 Table 13 Moriasi et al. (2007), Moriasi et al. (2015) annual statistical measures of performance for each constituent at the Kalpowar Gauge, Cape York ...... 58 Table 14 Sources and sinks of fine sediment in CY and the Normanby Basin (as percent of supply total) ...... 59 Table 15 CY grazing management practice proportions and change summary; values are percent of total area ...... 60 Table 16 WT baseflow proportions for the current and previous Report Card ...... 61 Table 17 Contribution from WT basins to the total baseline load ...... 62 Table 18 Moriasi et al. (2015) annual performance ratings at the North Johnstone River gauging station compared to GBRCLMP annual loads ...... 66 Table 19 Moriasi et al. (2015) monthly performance rating at the North Johnstone River gauging station ...... 66 Table 20 Moriasi et al. (2015) overall monthly performance rating for each constituent for the five gauge sites, Wet Tropics ...... 67 Table 21 TSS, DIN and PSII sugarcane areal rates by basin for the WT ...... 69 Table 22 Wet Tropics sugarcane management practice change summary; values are percent of total ...... 71 Table 23 Wet Tropics banana management practice change summary; values are percent of total... 71 Table 24 Burdekin baseflow proportions for six gauges the current and previous Report Card ...... 72 Table 25 Contribution from Burdekin basins to the total Burdekin baseline load ...... 73 Table 26 Moriasi et al. (2015) overall annual performance rating for each constituent at 120001A Burdekin River at Home Hill (120006B Burdekin River at Clare), n = 8 years for flow and n = 8 years for WQ) ...... 75

11 Source Catchments modelling – Updated methodology and results

Table 27 Burdekin grazing management change; values are percent of total ...... 78 Table 28 Burdekin sugarcane management practice change summary; values are percent of total ... 79 Table 29 Burdekin grains management change; values are percent of total ...... 80 Table 30 MW baseflow proportions for the current and previous Report Card ...... 81 Table 31 Contribution from MW basins to the total MW baseline load ...... 82 Table 32 Moriasi et al. (2007) annual statistics for GBCLMP versus modelled data for the Pioneer River at Dumbleton Pump Station (125013A) n = 8 year for flow data, n = 8 years for WQ data ...... 85 Table 33 Moriasi et al. (2007) annual statistics for GBCLMP versus modelled data for Sandy Creek at Homebush (126001A) n = 5 year for flow data, n = 5 years for WQ data ...... 86 Table 34 Modelled Fine Sediment export in the O’Connell River against LiDAR estimates for fine sediment export between 2010 and 2014 ...... 89 Table 35 MW grazing management practice change summary; all values are in percent ...... 91 Table 36 MW sugarcane management practice change summary; all values are in percent ...... 92 Table 37 Modelled flow differences for RC2014 and RC2015 compared to observed flows ...... 93 Table 38 Base flow proportions (%) for RC2014 and RC2015 compared to observed estimates ...... 94 Table 39 Contribution from Fitzroy basins to the total Fitzroy baseline load ...... 95 Table 40 RC2015 (1300000 Fitzroy River at Rockhampton, n = 7 years for flow and n = 7 years for WQ) ...... 98 Table 41 RC2015 (130504B Comet River at Comet Weir, n = 5 years for flow and n = 5 years for WQ) ...... 99 Table 42 Fitzroy Grazing Management Change ...... 102 Table 43 Fitzroy Grains Management Change (% change) ...... 103 Table 44 Burnett Mary baseflow proportions for six gauges the current and previous Report Card .. 104 Table 45 Contribution from Burnett Mary basins to the total Burnett Mary baseline load ...... 105 Table 46 Moriasi et al. (2015) overall annual performance rating for RC2015 (136014A Burnett River at Ben Anderson Barrage (HW), n = 8 years for flow and n = 5 years for WQ) ...... 108 Table 47 BM sugarcane management practice change summary; all values are in percent ...... 111 Table 48 BM grazing management Change; all values are percent of total ...... 112 Table 49 Cape York hydrology calibration results (1970–2014) ...... 140 Table 50 Wet Tropics hydrology calibration results (1970–2014) ...... 141 Table 51 Burdekin hydrology calibration results (1970–2014) ...... 142 Table 52 Mackay Whitsunday hydrology calibration results (1970–2014) ...... 144 Table 53 Fitzroy hydrology calibration results (1970–2014) ...... 148 Table 54 Baseflow proportions (%) for the current and previous Report Card for all gauges in the Fitzroy Basin ...... 149 Table 55 Burnett Mary hydrology calibration results (1970–2014) ...... 151 Table 56 Baseflow proportions (%) for the current and previous Report Card for all gauges in the Burnett Mary Basin ...... 152 Table 57 Hillslope erosion parameters ...... 153 Table 58 Gully erosion parameters ...... 153

12 Source Catchments modelling – Updated methodology and results

Table 59 Particulate nutrient generation parameter values ...... 153 Table 60 Parameter values for in-stream models in RC2015 baseline models ...... 153 Table 61 A summary of changes for RC2015 changes for cropping and sugarcane (Changes from 2014 Report Card) ...... 154 Table 62 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Cape York ...... 155 Table 63 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Wet Tropics ...... 155 Table 64 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Burdekin ...... 156 Table 65 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Mackay Whitsunday 156 Table 66 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Fitzroy ...... 157 Table 67 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Burnett Mary ...... 157 Table 68 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for predevelopment model ...... 158 Table 69 Constituent loads for predevelopment, baseline and 2015 Report Card change model runs for the CY NRM Region ...... 159 Table 70 Sources and sinks of TSS for each basin in CY (kt/yr) ...... 162 Table 71 Constituent loads for predevelopment, baseline and RC2015 change model runs for the WT NRM Region ...... 162 Table 72 Sources and sinks of each constituent in the Wet Tropics ...... 165 Table 73 Constituent loads for predevelopment, baseline and 2015 Report Card change model runs for the Burdekin NRM Region ...... 165 Table 74 Sources and sinks of each constituent in the Burdekin NRM Region ...... 168 Table 75 Constituent loads for predevelopment, baseline and 2015 Report Card change model runs for the Mackay Whitsunday NRM Region ...... 168 Table 76 Sources and sinks of TSS for each basin MW (kt/yr) ...... 170 Table 77 Constituent loads for predevelopment, baseline and RC2015 change model runs for the Fitzroy NRM Region ...... 171 Table 78 Sources and sinks of TSS for each basin in the Fitzroy NRM Region ...... 173 Table 79 Constituent loads for predevelopment, baseline and RC2015 change model runs for the Burnett Mary NRM Region ...... 174 Table 80 Sources and sinks of each constituent in the Burnett Mary NRM Region ...... 176 Table 81 Fine sediment and particulate nutrient comparison with GBRCLMP monitored data for five WT sites ...... 177 Table 82 Dissolved nitrogen species comparison with GBRCLMP monitored data for five WT sites 177 Table 83 Dissolved phosphorus comparison with GBRCLMP monitored data for five WT sites ...... 177 Table 84 Barron River model validation compared to GBRCLMP annual loads ...... 177 Table 85 Barron River model validation compared to Joo monthly loads ...... 178 Table 86 North Johnstone River model validation compared to GBRCLMP annual loads ...... 178 Table 87 North Johnstone River model validation compared to Joo monthly loads ...... 178 Table 88 South Johnstone River model validation compared to GBRCLMP annual loads ...... 179 Table 89 South Johnstone River model validation compared to Joo monthly loads ...... 179

13 Source Catchments modelling – Updated methodology and results

Table 90 Tully River model validation compared to GBRCLMP annual loads ...... 179 Table 91 Tully River model validation compared to Joo monthly loads ...... 180 Table 92 Herbert River model validation compared to GBRCLMP annual loads ...... 180 Table 93 Herbert River model validation compared to Joo monthly loads ...... 180 Table 94 RC2015 (120302B Cape River at Taemas, n = 7 years for flow and n = 7 years for WQ) .. 181 Table 95 RC2015 (119101A Barratta Creek at Northcote, n = 5 years for flow and n = 5 years for WQ) ...... 181 Table 96 RC2015 (120002C Burdekin River at Sellheim, n = 8 years for flow and n = 7 years for WQ) ...... 182 Table 97 RC2015 (120301B Belyando River at Gregory Development Road, n = 7 years for flow and n = 6 years for WQ) ...... 182 Table 98 RC2015 (120001A Burdekin River at Home Hill (120006B Burdekin River at Clare), n = 8 years for flow and n = 8 years for WQ) ...... 182 Table 99 RC2015 (120310A Suttor River at Bowen Development Road, n = 8 years for flow and n = 8 years for WQ) ...... 183

14 Source Catchments modelling – Updated methodology and results

List of Figures

Figure 1 Progress towards Reef Plan targets till 2015 for the whole of GBR ...... 4 Figure 2 Model build process ...... 26 Figure 3 Catchment area vs. stream width used to determine streambank erosion parameters ...... 41 Figure 4 Catchment area vs. bank height used to determine streambank erosion parameters ...... 42 Figure 5 Average annual modelled discharge for the GBR RC2015 (1986–2014) ...... 45 Figure 6 Predevelopment and anthropogenic baseline loads for all 35 basins in the GBR for TSS .... 48 Figure 7 Predevelopment and anthropogenic baseline loads for all 35 basins in the GBR for DIN ..... 49 Figure 8 Baseline GBR TSS load contribution by land use ...... 50 Figure 9 Baseline GBR DIN load contribution by land use ...... 50 Figure 10 Progress towards Reef Plan targets till 2015 for the whole of GBR ...... 51 Figure 11 Estimated TSS reduction under All A and All B management practice scenarios ...... 52 Figure 12 Estimated DIN reduction under All A and All B management practice scenarios ...... 53 Figure 13 Estimated PSII TE reduction under All A and All B management practice scenarios ...... 53 Figure 14 Estimated PN reduction under All A and All B management practice scenarios ...... 54 Figure 15 Estimated PP reduction under All A and All B management practice scenarios ...... 54 Figure 16 Average annual modelled versus measured load comparison for 2006-2014 period, Normanby River Gauge at the Kalpowar Crossing (105017A) ...... 57 Figure 17 Contribution to fine sediment, particulate nutrient and DIN export by land use for the CY region ...... 58 Figure 18 Fine sediment source sink budget diagram for the Normanby Basin ...... 59 Figure 19 TSS (kt/yr) loads for the WT basins, predevelopment and anthropogenic baseline contributions ...... 63 Figure 20 DIN (t/yr) loads for the WT basins, predevelopment and anthropogenic baseline contributions ...... 63 Figure 21 Fine sediment comparison between GBRCLMP measured loads and Source modelled loads for five sites in the WT ...... 64 Figure 22 DIN comparison between GBRCLMP measured loads and Source modelled loads for five sites in the WT ...... 65 Figure 23 PSII comparison between GBRCLMP measured loads and Source modelled loads for five sites in the WT ...... 65 Figure 24 Contribution to fine sediment, particulate nutrient and DIN export by land use for the WT region ...... 67 Figure 25 WT DIN load and areal rate per land use ...... 68 Figure 26 Fine sediment source sink budget diagram for WT ...... 70 Figure 27 TSS (kt/yr) loads for the BU basins, highlighting the predevelopment and anthropogenic baseline contributions ...... 74 Figure 28 DIN (kt/yr) loads for the BU basins, highlighting the predevelopment and anthropogenic baseline contributions ...... 74

15 Source Catchments modelling – Updated methodology and results

Figure 29 Contribution to fine sediment, particulate nutrient and DIN export by land use for the BU region ...... 76 Figure 30 DIN export load and areal rate for each land use in the BU ...... 76 Figure 31 Fine sediment source sink budget for the Burdekin ...... 77 Figure 32 TSS (kt/yr) loads for the MW basins, highlighting the predevelopment and anthropogenic baseline contributions ...... 83 Figure 33 DIN (kt/yr) loads for the MW basins, highlighting the predevelopment and anthropogenic baseline contributions ...... 83 Figure 34 Mean Annuals Modelled VS. GBRCLMP in the Pioneer River at Dumbleton Pump Station 84 Figure 35 Mean Annuals Modelled VS. GBRCLMP in Sandy Creek at Homebush ...... 84 Figure 36 Long-term annual TSS Loads Modelled VS. FRCE in the Pioneer River at Dumbleton Pump Station ...... 86 Figure 37 Long-term annual DIN Loads Modelled VS. FRCE in the Pioneer River at Dumbleton Pump Station ...... 87 Figure 38 Mean annual Source Catchments modelled and P2R at Multifarm during low and medium flow events ...... 88 Figure 39 Constituent contribution (%) by land use for four constituents ...... 89 Figure 40 DIN export load and areal rate for each land use in the MW ...... 90 Figure 41 Fine sediment source sink budget diagram for MW ...... 91 Figure 42 Fine sediment average annual modelled (Source Catchments) versus measured load comparison for 2006-2014* period, for the four monitoring sites in the Fitzroy region. *Data for Theresa Creek is only for two years, and for the Dawson River three years...... 96 Figure 43 Particulate nitrogen average annual modelled versus measured load comparison for 2006- 2014* period, for the four monitoring sites in the Fitzroy region. *Data for Theresa Creek is only for two years, and for the Dawson River three years...... 96 Figure 44 Particulate phosphorus average annual modelled versus measured load comparison for 2006-2014* period, for the four monitoring sites in the Fitzroy region. *Data for Theresa Creek is only for two years, and for the Dawson River three years...... 97 Figure 45 Constituent contribution (%) by land use for four constituents ...... 100 Figure 46 Fine sediment source sink budget for the Fitzroy ...... 101 Figure 47 TSS (kt/yr) loads for the BM basins, highlighting the predevelopment and anthropogenic baseline contributions ...... 106 Figure 48 DIN (t/yr) loads for the BM basins, highlighting the predevelopment and anthropogenic baseline contributions ...... 106 Figure 49 Comparison of monitored and modelled mean annual flow and constituent loads at Burnett EOS site (136014A) ...... 107 Figure 50 Contribution to fine sediment, particulate nutrient and DIN export by land use for the BM region ...... 108 Figure 51 DIN export load and areal rate for each land use in the BM ...... 109 Figure 52 Fine sediment source sink budget for the Burnett Basin ...... 110

16 Source Catchments modelling – Updated methodology and results

Figure 53 TSS, PN and PP trapping efficiencies (%) of storages...... 110 Figure 54 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in Lethe Brook Creek at Hadlow Road (122012A) ...... 144 Figure 55 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in the O’Connell River at Staffords Crossing (124001B) ...... 145 Figure 56 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in St.Helens Creek at Calen (124002A) ...... 145 Figure 57 Time Series, Flow Duration (log scale), and Scatter Plot Comparisons of Modelled VS. Measured Daily Flows in the Andromache River at Jochheims (124003A) ...... 145 Figure 58 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in the Pioneer River at Pleystowe Recorder (125001B) ...... 146 Figure 59 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in the Pioneer River at Sarichs (125002C) ...... 146 Figure 60 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in Cattle Creek at Gargett (125004B) ...... 146 Figure 61 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in Sandy Creek at Homebush (126001A)...... 147 Figure 62 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in Plane Creek at Sarina (126002A) ...... 147 Figure 63 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in Carmila Creek at Carmila (126003A) ...... 147 Figure 64 Long-term Annual PN Loads Modelled against FRCE and GBRCLMP Estimates in the Pioneer River at Dumbleton Pump Station ...... 183 Figure 65 Long-term Annual PP Loads Modelled against FRCE and GBRCLMP Estimates in the Pioneer River at Dumbleton Pump Station ...... 184 Figure 66 Long-term Annual DIP Loads Modelled against FRCE and GBRCLMP Estimates in the Pioneer River at Dumbleton Pump Station ...... 184 Figure 67 Long-term Annual DON Loads Modelled against FRCE and GBRCLMP Estimates in the Pioneer River at Dumbleton Pump Station ...... 185 Figure 68 Long-term Annual DOP Loads Modelled against FRCE and GBRCLMP Estimates in the Pioneer River at Dumbleton Pump Station ...... 185

17 Source Catchments modelling – Updated methodology and results

Acronyms

Acronym Description

BM Burnett Mary

BU Burdekin

CY Cape York

DIN Dissolved inorganic nitrogen

DIP Dissolved inorganic phosphorus

DNRM Department of Natural Resources and Mines DON Dissolved organic nitrogen DOP Dissolved organic phosphorus DR Delivery ratio e.g. HSDR hillslope delivery ratio Dynamic SedNet—a Source Catchments ‘plug-in’ developed by DNRM/DSITIA, which provides a suite of constituent generation and in-stream processing models that simulate the DS processes represented in the SedNet and ANNEX catchment scale water quality model at a finer temporal resolution than the original average annual SedNet model DSITI Department of Science, Information Technology, and Innovation Dry weather concentration—a fixed constituent concentration to base or slowflow generated DWC from a functional unit to calculate total constituent load Event mean concentration—a fixed constituent concentration to quickflow generated from a EMC functional unit to calculate total constituent load EOS End-of-system ERS Environment Resource Sciences Flow Range Concentration Estimator—a modified Beale ratio method used to calculate FRCE average annual loads from monitored data FU Functional Unit FZ Fitzroy GBR Great Barrier Reef GBRCLMP Great Barrier Reef Catchment Event Monitoring Program (supersedes GBRI5) HowLeaky Water balance and crop growth model based on PERFECT LH Lyne-Hollick baseflow MW Mackay Whitsunday NRM Natural Resource Management NSE Nash Sutcliffe Coefficient of Efficiency P2R Paddock to Reef Integrated Monitoring, Modelling and Reporting Program PSII herbicides Photosystem II herbicides—ametryn, atrazine, diuron, hexazinone and tebuthiuron PSII TE PSII herbicide Toxic Equivalents—the total load of PSII herbicides weighted for their toxicity RC2–RC5 Report Card 2–Report Card 5—published outputs of Reef Plan/Paddock to Reef An ongoing and key component of Caring for our Country. Reef Rescue represents a coordinated approach to environmental management in Australia and is the single largest Reef Rescue commitment ever made to address the threats of declining water quality and climate change to the Great Barrier Reef World Heritage Area RUSLE Revised Universal Soil Loss Equation

18 Source Catchments modelling – Updated methodology and results

Acronym Description Catchment model that constructs sediment and nutrient (phosphorus and nitrogen) budgets SedNet for regional scale river networks (3,000–1,000,000 km2) to identify patterns in the material fluxes TSS Total suspended sediment TN Total nitrogen TP Total phosphorus WT Wet Tropics

Definitions of sources and sinks

Supply Definitions

Particulate

Hillslope: Source of TSS, PP and PN supplied to the stream network by surface erosion processes (surface and rill erosion).

Gully: TSS, PP and PN supplied to the stream network by gully erosion processes).

Streambank: TSS, PP and PN supplied to the stream network by streambank erosion processes.

Channel remobilisation: TSS, PP and PN supplied by channel remobilisation processes.

Dissolved

Diffuse dissolved: Source of dissolved constituents supplied to the stream network from all land uses.

Seepage: Dissolved constituents supplied to the stream network from sugarcane via subsurface loss pathways.

Point source: Dissolved constituents supplied to the stream network from point sources; in this case Sewerage Treatment Plants (STP)

Loss Definitions

Storage deposition: TSS, PP and PN losses via the process of particle settling and resultant deposition in storages.

Floodplain deposition: TSS, PP and PN losses via the process of particle settling and resultant deposition in on floodplains.

Storage decay: Dissolved constituents lost via the process of decay processes.

Extraction: Dissolved and particulate constituents lost via extraction, (e.g. pumped losses for irrigation, stock or domestic water supplies) from the stream network.

Stream deposition: Particulate constituents lost to the stream network via instream and resultant deposition.

19 Source Catchments modelling – Updated methodology and results

Modelling assumptions

The key modelling assumptions to note are:  It is important not to lose sight of the main objective of the modelling, to report the water quality benefits as a result of management changes. Therefore the relativity between scenarios is important not the absolute load. However the goal of the modelling program is for continual improvement of modelled load predictions. This approach will continue.  Loads reported for each scenario are the modelled average annual load for the specified model run period (July 1986 – June 2014).  Land use areas in the model are static over the model run period and were based on the latest available QLUMP data in each NRM region.  The predevelopment land use scenario includes all water storages, weirs and water extractions are represented as in the baseline model. There is no change from the baseline scenario hydrology. This approach was recommended by the Independent Reef Science Panel (ISP), to isolate changes to water quality which are due solely to a change in land management practices. Therefore it is important to recognise that the term predevelopment in this report does not refer to a true representation of a pre-European scenario. The scenario is more reflective of a pre agricultural development scenario  Paddock model runs used in the catchment models represent ‘typical’ management practices in each management class and do not reflect the actual array of management practices being applied on a particular area of land.  This applies to application rates of herbicides and fertilisers used to in the paddock models which were derived through consultation with relevant local industry groups and technical experts in each region.  Distribution of baseline land management practices is not able to be captured (in its entirety) in a spatial format. Regional and sub-regional estimates of this distribution have been made, and these estimates are used as required in the preparation of the data inputs for the baseline scenarios used for each Report Card. The distribution will vary between reporting years, to accommodate the reported land management improvements that get supplied in spatial format.  All land management improvements to be represented in the relevant scenarios are supplied in a spatial format. This data is used to inform the data preparation for the necessary scenarios. In situations where the spatial data does not coincide with a representative modelling unit, the land management improvement is reallocated to the nearest suitable location for grazing and riparian investment. For cropping and cane management improvements, the area of improvement that is not coincident with a relevant modelling unit is applied evenly throughout the relevant region or sub-region for that industry.  Water quality improvements for the horticulture, dairy, and cotton industries are currently not modelled due to a lack of management practice adoption data and limited experimental data on which to base load reductions, and also the small area these industries occupy.

20 Source Catchments modelling – Updated methodology and results

 Modelling of the banana industry and the associated water quality improvements have been included in the Wet Tropics and Cape York models in Report Card 2014.  Dissolved inorganic nitrogen (DIN) load reductions are modelled for sugarcane systems with APSIM. DIN load reductions are not modelled in grains system as no DIN export are provided from these simulations and few if any investments are made.  The number of paddock model simulations available to represent the vast array of possible management practices within each region has increased markedly from the simple ‘ABCD’ approach used prior to RC2014. The increase in available simulations allows many more management improvements to be represented, however these are often representing changes to processes that yield a small benefit in terms of reducing pollutant generation. As a result the estimated loads reductions associated with this more detailed paddock modelling regime are significantly smaller than experienced previously. The smaller load reduction estimates are considered to be more accurate and reliable than estimates made using the old, limited suite of simulations.  For land uses that require spatially variable data inputs for pollutant generation models (USLE based estimates of hillslope erosion and SedNet-style gully erosion), data pre-processing captures the relevant spatially variable characteristics using the specific ‘footprint’ of each land use within each sub-catchment. These characteristics are then used to provide a single representation of aggregated pollutant generation per land use in each subcatchment.  Groundwater movement to streams is not explicitly modelled. Groundwater to streams is represented as a calibrated baseflow and ‘dry weather concentrations’ (DWC) of constituents. However, these loads are not subject to management effects.  Whilst DIN movement below the root zone is not explicitly modelled, a proportion of the DIN modelled by APSIM in sugarcane areas, can be delivered to the stream to represent lateral or seepage flow. The proportion of seepage load delivered to the stream is calibrated for the baseline and the loads will vary with the given management practice simulated.  It is important to note that this report summarises simulated, not actual, average annual load reductions of key constituents to the GBR lagoon. The modelled changes in water quality are a function of the area of improved land management adoption. Estimates of adoption areas are supplied by regional NRM groups. Results from this modelling project are therefore a useful indicator of the likely (theoretical) effects of investment in changed land management practices rather than a measured (empirical) reduction in loads.

21 Source Catchments modelling – Updated methodology and results

Introduction

The Great Barrier Reef (GBR) is the world’s largest living organism, and one of the seven natural wonders of the world. However, the health and resilience of the GBR is under threat from many pressures including declining water quality and climate change. Over the past 150 years, the GBR basins have been extensively modified for agricultural production and urban settlement, leading to a decline in water quality entering the GBR lagoon (Brodie et al. 2013). The largest contribution of poor water quality comes from agricultural land use activities within the GBR basins. Poor water quality, by way of high sediment, nutrient and pesticide loads, can affect the GBR ecosystem by reducing light availability, smothering corals, promoting algal growth and propagating crown of thorns (COTS) starfish outbreaks by increasing plankton blooms on which the COTS larvae feed (GBRMPA 2016).

The Reef Water Quality Protection Plan (Reef Plan) was established in 2003, updated in 2009 (Reef Plan 2009 (Cabinet 2009)) and again in 2013 (Reef Plan 2013 (Department of the Premier and Cabinet 2013)) in response to the decline in water quality entering the GBR lagoon. Reef Plan is a collaborative program that includes a set of water quality and land management practice targets for catchments discharging to the GBR. The long-term goal is to ensure the quality of water entering the GBR has no detrimental impact on the health and resilience of the Reef. The plan is a joint commitment of the Australian and Queensland Governments.

To strengthen the Australian and Queensland government’s commitment to the Reef, in 2014, Reef 2050 Plan (Australia 2015) was launched as the overarching framework for the long term (2015 to 2050) protection and management of the GBR. The Reef 2050 Plan is a key component of the Australian Government’s response to the recommendations of the UNESCO World Heritage Committee in 2015 (UNESCO 2015).

Reef Plan is the main mechanism for implementing the Reef 2050 Plan. The Paddock to Reef Integrated Monitoring, Modelling and Reporting (P2R) Program provides the monitoring and evaluation framework to measure and report on progress towards both Reef Plan and Reef 2050 Plan water quality targets. Monitoring and evaluation is undertaken for a range of targets including water quality targets in the annual Reef Report Card. These targets will be reviewed in 2017 and basin scale water quality targets will be set. The Report Cards show cumulative progress towards the Reef Plan 2013 water quality targets resulting from regional investments in improved land management practices each year. Six previous report cards (Report Cards 2009-2014) have been released (www.reefplan.qld.gov.au) tracking the good progress being made towards the Reef Plan targets. The Reef Plan 2013 water quality targets were set for the whole of GBR. Progress was reported for the whole of GBR plus the six contributing NRM regions: Cape York (CY), Wet Tropics (WT), Burdekin (BU), Mackay Whitsunday (MW), Fitzroy (FZ) and Burnett Mary (BM).

22 Source Catchments modelling – Updated methodology and results

The Reef Plan 2013 water quality targets state that by 2018 there will be:  a minimum 20% reduction in anthropogenic sediment and particulate nutrient loads at the end of catchment in priority areas

 a minimum 50% reduction in anthropogenic dissolved inorganic nitrogen (DIN) loads at the end of catchment in priority areas

 a minimum 60% reduction in pesticide loads at the end of catchment in priority areas

Catchment modelling is used as one of multiple lines of evidence to report on progress towards the Reef Plan water quality targets. In this technical report the Report Card 2015 modelled load reductions resulting from the adoption of improved land management practices for the 2015 year for the GBR and for the six NRM regions are summarised, and model updates are outlined. This report outlines:

 Changes to Source Catchments water quantity and quality model methodology since the previous report card  Updated predevelopment, baseline and anthropogenic loads for the 1986 – 2014 climate period  Progress towards meeting Reef Plan 2013 water quality targets over the past 12 months following the adoption of improved management practices.

Model Improvement Cycle The P2R Program strives for continual improvement in all areas. Therefore, the catchment and paddock modelling teams aim for continual improvement of the modelling approach to improve predictions of pollutant loads and the associated improvements in water quality as a result of improved management practice adoption.

In Table 2 the catchment modelling cycle is presented with the green highlighted cell indicating the current Report Card (RC2015). Phase 1 relates to the Reef Plan 2 period, which covers the models produced from 2010–2013. Report Card 2009 provided the first GBR-wide load estimates of predevelopment, total baseline and total anthropogenic loads using the best available monitoring and modelling data at the time (Kroon et al. 2010). Subsequent Report Cards used Source Catchment modelling framework to report on load reductions due to improved practice adoption. Major updates to the Source Catchment models occur on a five year cycle to align with the Reef Plan funding cycle. Minor changes occur between the individual Report Cards, where required. In RC2015 only minor model changes have been made. All model updates will be discussed at length in this Report. Each successive Report Card is considered the best representation of loads from each GBR NRM region, and the most recent Report Card should always be used when reporting or referencing Source Catchments GBR loads. A full summary of model changes for RC2015 are listed in Table 3.

23 Source Catchments modelling – Updated methodology and results

Table 2 Summary of catchment modelling cycle

Management Practice data Phase 1 Modelling 2009 - 2013 baseline year; climate period Report Card 2009 Kroon et al. (2010) model results Report Card 2010 Model build; hydrology calibration; 2009; Report Card 2011 Minor model changes 1986-2009 Report Card 2012 Minor model changes Report Card 2013 Minor model changes Management Practice data Phase 2 Modelling 2014 - 2018 baseline year; climate period Major model changes; update Source platform; hydrology recalibration; incorporate new management Report Card 2014 practice framework; climate period for reporting extended Minor model changes; hydrology calibration modified Report Card 2015 for better alignment of baseflow in some regions 2013; 1986-2014 Report Card 2016 Minor model changes (if required) Minor model changes; research major changes for Report Card 2017 post RC 2018 rebuild Minor model changes; research major changes for Report Card 2018 post RC 2018 rebuild

Integrating modelling and monitoring Modelling is a way to extrapolate monitoring data through space and time, and provides an opportunity to explore the climate and land management interactions and their associated impacts on water quality. Monitoring data is the most important point of truth for model validation and parameterisation. With the continuation of the Reef Plan projects, increased monitoring data becomes available, and as such, the models become more refined and a better representation of reality as they aim to integrate the sum of all information (measured and modelled data). Given the scale of the P2R program and its objectives, it would be impossible to measure the impact of land management practice change on water quality, hence models are applied that can predict changes in water quality when set against a baseline or benchmark (in this case the model run period of 1986–2014).

Combining the two programs (catchment modelling and monitoring) ensures continual improvement in the models while at the same time identifying data gaps and priorities for future monitoring. A key example of this is the recent endorsement by the Independent Science Panel of 16 new monitoring sites as part of the GBR Catchment Loads Monitoring Program. The selection of these sites followed a two-step process, including consultation with other Reef Plan programs, particularly the catchment modellers, as well as correspondence with each regional NRM body. The additional 16 sites means that 20 of the 35 GBR basins are now monitored, and measure approximately 95% and 89% of the TSS and DIN loads (respectively) discharged to the GBR lagoon.

24 Source Catchments modelling – Updated methodology and results

Application of modelling outputs While the sole objective of the Catchment Modelling program is to report on changes to water quality discharging to the GBR lagoon as a result of improved land management practices, the modelled load estimates have been used by a number of other Reef Plan funded programs. Most recently the RC2015 catchment modelling results will be used a major input dataset to the GBR Scientific Consensus Statement (Waterhouse in prep., specifically Chapter 2). In the Scientific Consensus Statement the scientific knowledge of water quality issues in the GBR were reviewed and synthesised to reach a consensus statement on the current understanding of the system (Waterhouse in prep.). Also, the latest modelling results are incorporated into a spatial prioritisation project to manage sediment, nutrient and pesticide runoff in the Burdekin NRM region (Waterhouse et al. 2016).

Scenario modelling – All A and All B management Modelling provides us with the ability to assess alternative management scenarios such as moving from current practices to some future improved state. One such example is modelling of the water quality response if all landholders were to adopt all A class management practices (All A scenario run) or an All B scenario. These scenarios were previously run using an earlier version of the GBR Source modelling framework (Waters et al. 2013).

Following the release of Report Card 2015, the All A and All B scenarios were rerun at the request of the Office of the Great Barrier Reef (OGBR) to assess whether the updated targets in Reef Plan 2013 could be achieved taking into account the updated modelling framework and the significant progress made to date.

Quality assurance and peer review At the program inception, the eWater CRC best practice modelling guidelines (Black et al. 2011) were used to develop the modelling approach and Quality assurance process described in Waters and Carroll (2012). A systematic model build process has been developed by the catchment modelling team, which is presented in Figure 2. Modelled loads generated for annual reef report cards are reviewed in the first instance internally by scientists from DNRM and DSITI. Once approved, loads and the associated reductions are reviewed externally, before final approval by the Queensland Government. A Coordination and Advisory Group (CAG) and an Independent Science Panel (ISP) oversee model development and proposed updates or changes (see Waters and Carroll (2012) for description of project governance arrangements). An external review of the catchment and paddock modelling is required every three years. The first was in 2012 (Jakeman et al. 2012) the second in 2015 (Jakeman et al. 2015). In addition the Queensland Audit Office undertook a review of the Paddock to Reef Program in 2015 (Queensland Audit Office 2015). Recommendations from the peer review process and the Audit form the basis for future model development and improvements.

25

Figure 2 Model build process

Methods

The Source Catchments water quantity and quality modelling framework is used to simulate how catchment and climate variables affect runoff and constituent transport, by integrating a range of component models, data and knowledge. Between Report Card releases there have been several upgrades to the Source Catchments platform, developed and maintained by eWater Ltd. In conjunction with the eWater upgrades, there have been some upgrades to the GBR component models maintained by DNRM. These upgrades, and in particular, the changes to input data sets are summarised in Table 3 and described in detail below. Those methods that have not changed since RC2014 modelling are not discussed, but rather can be reviewed in the series of Technical Reports completed post–2013 Report Card (see Waters et al. (2014) and the six supplementary regional reports), and following RC2014 (McCloskey et al. 2017).

Table 3 Model updates undertaken for Report Card 2015

Report Modification Rationale Impact page number  Baseflow proportion decreased in most instances of Baseflow was not well represented Recalibration of flow to recalibration. This across the GBR when compared to better represent baseflow increased surface 28 the Lyne-Hollick calculated baseflow proportion – all regions runoff volumes hence of monitored data saw a slight (<10%) increase in loads exported Previously, a daily implementation of ‘maximum annual erosion load’ was used to scale transported daily Use of maximum quickflow  Typically increase loads, however this approach does concentration for capping RUSLE TSS load in all 29 not allow for daily runoff differences hillslope erosion regions to be reflected in transported sediment masses, nullifying the use of a daily RUSLE model  Increase in TSS supply due to remobilised Measured data is now available in a sediment from Instream deposition and limited number of GBR catchments instream channel remobilisation were which can assist in constraining this  Losses in sediment 30 enabled in three additional component of the sediment budget. due to deposition. regions Can be a significant component of the sediment budget.  Generally aim for remobilisation/depositi on to be in equilibrium  Resulted in improved In a number of regions, under- estimates of particulate Replacement of subsurface prediction of particulate nutrients. nutrients. For example particulate nutrient layers Identified derivation of subsurface CY now within 30% of 35,36 with surface layers nutrient layers have high degree of observed PP/PN loads uncertainty. previously 60% in early RC2015 iterations. Default in previous models was 80:20 ratio DIP/DOP. DOP was  A decrease in DIP Adjustment of DIP:DOP ratio consistently under predicted against export loads, and an 36,43 in sugarcane observed loads, so this ratio was increase in DOP export adjusted regionally so results would loads better match monitored loads

Source Catchments modelling – Updated methodology and results

Report Modification Rationale Impact page number Changes to CY HowLeaky models  Additional pesticides HowLeaky cropping to better represent cropping now incorporated into 35 systems in CY CY model Current baseline streambank erosion was in some instances less  Predevelopment Application of A class than predevelopment scenario streambank management to because the riparian vegetation contribution is now 31 streambanks in the percentage was greater in the always lower than that predevelopment scenario baseline model thus reducing of the baseline model streambank erosion in a predevelopment scenario.

The methodology updates described below are split into two sections: updates applied across the whole of GBR, and regionally specific updates listed under NRM region headings. Only changes to the methodologies are described in this report.

GBR methodology updates

Hydrology Hydrology underwent a major recalibration prior to RC2014, where the rainfall-runoff model was changed from SimHyd to Sacramento to improve the calibration of high flows. Following the release of RC2014 it was noted that the proportion of baseflow runoff was higher than would be expected in many regions. A recalibration was undertaken to include a baseflow optimisation term to align with an estimates derived from observed gauged flow.

Sacramento rainfall-runoff model calibration for baseflow The hydrology model was recalibrated in all regions to better align baseflow to observed flow estimates. This was important because sediment generation and transport are associated with different flow components. To improve the modelled baseflow proportion, the existing Objective Function used in calibration was modified to include a term which enables the minimisation of the difference between the simulated baseflow and an ”observed” baseflow proportion derived from gauged flow data. The observed baseflow was estimated using the Lyne-Hollick (1979) filter method which applies a smoothing filter to daily flow which provides a means of splitting daily flow into baseflow and quickflow. The baseflow proportion was then calculated for each calibration gauge over the calibration period. If the difference between the modelled baseflow proportion in the previous calibration and the Lyne-Hollick baseflow estimate derived from gauged data was greater than 10% then that gauge was selected for re-calibration (Table 4). It is evident from Table 4 that the wetter regions such as WT and MW with high baseflow proportions required considerably less recalibration.

28 Source Catchments modelling – Updated methodology and results

Table 4 Number of gauges that underwent recalibration for baseflow

Number of gauges NRM Number of gauges recalibrated to Region calibrated originally improve baseflow CY 15 15 WT 30 9 Burdekin 46 34 MW 10 4 Fitzroy 44 32 BM 29 29

Hillslope generation model

Concentration cap The daily RUSLE model (as implemented in the Dynamic SedNet plugin) is susceptible, on occasion, to generating large daily eroded soil masses when there is not an equally significant runoff volume available to transport this mass (via the rainfall-runoff model pathway). This can occur due to the fact that daily eroded mass is calculated (from the RUSLE equation) as function of the daily rainfall and not the rainfall-runoff model generated daily runoff. This can be the result of erroneous topographic data inputs, or poor rainfall-runoff parameterisation, or both. To avoid transporting disproportionate eroded soil masses on days of smaller runoff hence concentration, the RUSLE model (as implemented in the Dynamic SedNet plugin) uses a “maximum daily quickflow concentration parameter” to keep the transported daily fine sediment load within acceptable tolerances of the runoff magnitude. Previously, a daily implementation of ‘maximum annual erosion load’ was used to scale transported daily loads, however this approach does not allow for daily runoff differences to be reflected in transported sediment masses, nullifying the use of a daily RUSLE model. The values implemented for the maximum daily quickflow (QF) concentration parameter for each NRM region are regionally specific based on monitoring data, and are presented in Appendix 2, Table 57.

Subsurface Nitrogen and Phosphorus layers change In the majority of regions there was a considerable under-prediction of particulate nutrient loads when compared to the measured load at monitoring locations in RC2014 and early iterations of RC2015. The existing particulate nutrient generation model utilises a clay percentage layer from the ASRIS database to estimate the associated particulate nutrient loads, on the assumption that particulates are exclusively bound to the clay fraction (Packett et al. 2014). Hence the model may have underestimated the particulate contribution from the non-clay sediments (e.g. fine silt) if these non-clay sediments have the ability to retain and carry particulate nutrients. This may be indicating further investigations are required to better estimate the export of particulate nutrient to the GBR lagoon. Wilkinson et al. (2010) describe the differences between the ASRIS spatial data sets and local point data. The surface and subsurface ASRIS nutrient datasets are extrapolated from direct measurements of soil nutrient content to provide a uniform spatial layer as required by the model. In remote areas such as CY there may be very few samples from which to extrapolate a continuous dataset, and hence there can be significant differences

29 Source Catchments modelling – Updated methodology and results between observed and modelled values. Often point source data of nutrient content is sourced from areas of intensive agriculture and as such the input datasets may be more reliable in these areas (Wilkinson et al. 2010). A review of the chemistry data for CY (CSIRO data in Biggs and Philip (1995)), suggests that there was no significant difference between surface and subsurface TN and TP chemistry. However, for the ASRIS derived layers surface concentrations of nutrients in CY were typically three times greater than that of sub-surface concentrations.

The sediment budget estimates suggest the main source of particulate nutrients was from subsurface derived particulate loads. Therefore, the subsurface nutrient concentration layers were replaced with surface layer in CY and WT regions. This is in line with the recommendation by Wilkinson et al. (2010) who suggested that users should choose the appropriate dataset for their region including from locally derived measurements or the ASRIS dataset.

In a number of the regions the hillslope PN and PP enrichment ratios were increased within acceptable ranges as reported in the literature (Sharpley 1985, Rose and Dalai 1988, Sharpley et al. 2002) to further increase the particulate nutrient load in line with measured data.

Sugarcane constituent generation Irrigation and drainage modelling were heavily revised in the BU paddock modelling (Shaw 2016), as were the management scenarios. This will impact the drainage, runoff and DIN components of the APSIM runs. Soils data was also reviewed in the Burdekin and WT. A summary of APSIM paddock model changes are presented in Table 61, Appendix 3. In addition to changes to the APSIM paddock model runs, there were a number of changes to specific parameters for sugarcane, especially in the WT and BM regions which are presented below.

HowLeaky cropping constituent generation For RC2014 a uniform slope of 1% was applied across all cropping areas modelled by HowLeaky and model runs were corrected for slope within Source Catchments. For RC2015 slope was accounted for in the HowLeaky paddock modelling as opposed to within the Source Catchments model.

In-stream models In-stream processes are an important component of regional sediment budgets. This functionality was only enabled in the CY model in RC2014. The in-stream model was applied in CY due to the availability of measured sediment data in the Normanby basin which provided a source of data to constrain the model parameters. All regions, with the exception of Burdekin and Fitzroy, have in-stream deposition and remobilisation enabled in RC2014. The in-stream models were enabled in RC2015 to incorporate this important component of the sediment budget, however the initial aim was to take a conservative approach and keep deposition and remobilisation in equilibrium until such time as more relevant local data becomes available. The parameter values used in each of the regional models are listed in Appendix 2, Table 60.

30 Source Catchments modelling – Updated methodology and results

Predevelopment Model In all regional models there instances where the load from the predevelopment scenario is higher than that for the baseline scenario, particularly for streambank and gully erosion. In order to manage this it is possible to scale down the predevelopment loads using a ‘management factor’ parameter for both gullies and streambanks. This is based on the assumption that the predevelopment load should always be less than the baseline, or current, load. In RC2015 and the streambank management factor has been set to ‘A’ class (0.6), while gullies remained at 1 (‘C’ class). This change was made on the assumption that streambanks were in a better condition in the predevelopment scenario, and that the changes already applied to gullies (e.g. reduction of cross-section by 90%) were sufficient to reflect a predevelopment scenario. That is, gullies were assumed to be of a lesser volume than current conditions. For the predevelopment model, the ‘management factor’ for both gullies and streambanks has been previously set to 1. A full description of how the management factor is applied to each gully and streambank management class is presented in McCloskey et al. (2014). In short, streambank and gully erosion rates are scaled relative to a ‘C’ class practice (adapted from Thorburn and Wilkinson (2012), where ‘C’ class management practice factor equals 1.

Constituent load validation Two main approaches were used to evaluate the GBR Source Catchments modelling. Firstly, a short- term comparison was made using load estimates from monitoring results that commenced in 2006 in ten high priority catchments (Turner et al. 2012, Joo et al. 2012). Secondly, a long-term comparison was made with catchment load estimates derived from all available measured data for the high priority catchments for the 23 year modelling period, using a Flow Regression Concentration Estimator (FRCE model) (Joo et al. 2014). Both of these sources are discussed in detail in the previous Report Card modelling technical reports (Waters et al. 2014).

As the number of years of load estimates from the short term monitoring program increases (currently eight years) there is less reliance on the long term load estimates as a source of validation given their high uncertainty. As a result the short term loads form the basis for validation in all regions with long term load estimates used in some regions where they are deemed to be of sufficient quality for load comparison.

Performance ratings Model evaluation performance ratings were established by Moriasi et al. (2007) and Moriasi et al. (2015), uing four quantitative statistics: Nash-Sutcliffe coefficient of efficiency (NSE), percent bias (PBIAS), the ratio of the root mean square error to the standard deviation of validation data (RSR), and the coefficient of determination (R2) (Table 5). In this report modelled and measured loads are assessed against these ratings where applicable.

The PBIAS performance measure can be applied for all constituents at an annual scale, as can flow for R2 and NSE. However, sediment and nutrient ratings for the R2, NSE, and RSR measures are

31 Source Catchments modelling – Updated methodology and results recommended to be used at a monthly scale (Moriasi et al. 2015). In the absence of another method for rating model performance against measured data, the monthly performance rating have been applied at an annual scale. It was not possible to calculate daily or monthly loads for all sites due to the highly variable sampling frequency not being appropriate for the load estimation technique applied.

Table 5 Performance ratings for recommended statistics for a monthly time-step (from (Moriasi et al. 2007), Moriasi et al. (2015))

Performance Performance evaluation criteria Constituent Measure Very good Good Satisfactory Unsatisfactory 2 2 2 2 R2 Flow R >0.85 0.75 < R ≤ 0.85 0.60 < R ≤ 0.75 R ≤ 0.60 (Moriasi, TSS/P R2 >0.80 0.65 < R2 ≤ 0.80 0.40 < R2 ≤ 0.65 R2 ≤ 0.40 2015) N R2 >0.7 0.60 < R2 ≤ 0.70 0.30 < R2 ≤ 0.60 R2 ≤ 0.30

NSE Flow NSE > 0.80 0.70 < NSE ≤ 0.80 0.50 < NSE ≤ 0.7 NSE ≤ 0.5 (Moriasi, TSS NSE > 0.80 0.70 < NSE ≤ 0.80 0.45 < NSE ≤ 0.70 NSE ≤ 0.45 2015) N/P NSE > 0.65 0.50 < NSE ≤ 0.65 0.35 < NSE ≤ 0.50 NSE ≤ 0.35

PBIAS (%) Flow PBIAS < ±5 ±5 ≤ PBIAS < ±10 ±10 ≤ PBIAS < ±15 PBIAS ≥ ±15 (Moriasi, TSS PBIAS < ±10 ±10 ≤ PBIAS < ±15 ±15 ≤ PBIAS < ±20 PBIAS ≥ ±20 2015) N/P PBIAS < ±15 ±15 ≤ PBIAS < ±20 ±20 ≤ PBIAS < ±30 PBIAS ≥ ±30

RSR Flow (Moriasi, TSS RSR ≤ 0.5 0.50 < RSR ≤ 0.60 0.60 < RSR ≤ 0.70 RSR > 0.7 2007) N/P

Scenario modelling – All A and All B management Modelling the water quality response as a result of all landholders were to adopt ‘All A’ class management practices or ‘All B’ class practices were run following the release of RC2014. These scenarios were previously run using an earlier version of the GBR Source modelling framework (Waters et al. 2013).

There were a number of limitations to the modelling approach reported in Waters et al (2013), including the inability to account for improved gully or streambank management and the inability to represent investments in recycle pits on cane farming systems. The GBR Source modelling framework was updated in 2014 to incorporate the new functionality and any updated input data sets were added. The following outlines the approach taken and any assumptions used in the running of these scenarios.

Baseline Model The RC2015 loads and land management proportions were used as the baseline data sets from which the All A and All B land management proportions and associated load reductions were derived.

A practice improvement that was added to the cane model simulations in RC2015 was the installation of recycle pits (Shaw and Silburn 2016). Projects funded recycle pit installation in the BU and BM regions.

32 Source Catchments modelling – Updated methodology and results

When applying hypothetical management improvements to the whole of the grazing industry within each regional catchment model, it is necessary to spatially define the baseline management distribution. The baseline distribution of the hillslope, gully and streambank management are not provided spatially. The method used previously to convert the non-spatial distributions into a useful spatial layer have relied upon the relationship between grazing (hillslope) management and ground cover (Scarth et al. 2006, Ellis and Searle 2014, Waters et al. 2014). This method only allows the hillslope baseline distribution to be represented spatially, yet the All A and All B scenarios require that gullies and streambanks are also addressed. For this reason, the spatial representation of baseline hillslope management was also used to create spatially variable parameter values for the ‘management factors’ for all relevant gully and streambank models. The gully management factors were calculated in this manner for grazing areas only, and the management factors for streambank models were calculated in a way that recognised the proportionality of grazing within a sub-catchment (as a surrogate for determining how much streambank actually aligned with grazing).

All B Scenario Cropping & Sugarcane:  Areas with allocated management ‘worse’ than B class in the baseline were modelled with B class management simulations (specific soils x climate simulation allocations were provided by paddock modellers)  In the Burnett Mary, under All B, recycle pits were implemented on all sugarcane properties. There was no specific data available to reflect differences in how these pits might be managed between A and B class management hence were modelled the same in All A and All B scenarios in this region  In the Burdekin area ‘recycle pits’ were implemented on all sugarcane properties, with their trapping efficiency reflecting B class management practices

Grazing:  In the hillslope generation model, the C factor (i.e. cover term) was re-calculated with the assumption that all areas classified as having C or D class management in the baseline scenario, were improved to B class reflected as improved cover  Gully and streambank management factors were recalculated under the assumption that areas in the baseline scenario at C or D class management were improved to B class  No specific adjustment was made to the percentage of riparian vegetation under an All B scenario in the stream bank erosion calculations

All A Scenario Cropping & Sugarcane:  Areas with management ‘worse’ than A class were modelled with A class management (specific simulation allocations were provided by paddock modellers)  In the Burnett Mary, under All A, recycle pits were implemented on all sugarcane properties. There was no specific data available to reflect differences in how these pits might be managed

33 Source Catchments modelling – Updated methodology and results

between A and B class management hence were modelled the same in All A and All B scenarios in this region  In the Burdekin area ‘recycle pits’ were implemented on all sugarcane properties, with their trapping efficiency reflecting A class management practices

Grazing:  Gully and bank management factors were recalculated under the assumption the and B, C and D class management was improved to A class  The percentage of riparian vegetation parameter, used in the stream bank generation calculations, was increased to 100% where possible  Hillslope generation model inputs were re-calculated with C Factor data adjusted assuming that B, C and D class management was improved to A class

34 Source Catchments modelling – Updated methodology and results

Cape York The following section describes methodology changes that are only applicable to the CY region.

Hydrology Recalibration was undertaken to improve baseflow representation, as described on page 28.

Hillslope sediment, nutrient and herbicide generation Maximum quickflow concentration adjusted, as described above on page 29. Particulate nutrient enrichment ratios were increased to account for under-predicted modelled loads. To further increase the particulate nutrient load contribution, the surface N and P concentration layers were used as inputs for both surface and subsurface data as described on page 29. These combined changes have increased particulate nutrient export loads to better align with measured loads, however both N and P loads remain under-predicted and this will be investigated in future modelling.

Gully sediment and nutrient generation models Nil changes occurred since RC2014.

Sugarcane constituent generation Not applicable to CY.

HowLeaky cropping constituent generation A CY specific farming system was applied for irrigated and dryland cropping in the HowLeaky paddock model runs. Previously the Fitzroy farming system had been previously applied due to the lack of better localised information. The CY system incorporated peanuts and maize into the crop cycle, and removed crops not grown in CY (e.g. wheat).

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) The EMC/DWC values were updated to better align with monitored data. The values used in this Report Card are in Appendix 2, Table 62.

In-stream models Nil changes occurred since RC2014.

Predevelopment model Streambank management factor set to A-class, as described above. Predevelopment EMC/DWC values are in Appendix 2, Table 68.

35 Source Catchments modelling – Updated methodology and results

Wet Tropics The following section describes methodology changes that are only applicable to the WT region.

Hydrology Recalibration to improve baseflow representation, as described on page 28.

Hillslope sediment, nutrient and herbicide generation Maximum quickflow concentration adjusted, as described above on page 29. Particulate nutrient enrichment ratios were increased to account for under-predicted modelled loads. To further increase the loads of particulate nutrients the surface N and P concentration layers were used to as inputs for both surface and subsurface data as described on page 29. This has the desired effect of increasing particulate loads, however both N and P remain under-predicted, and this will be investigated for future report cards.

Gully sediment and nutrient generation models Gully cross sectional area was increased to 10m2 (from 5m2) to increase the amount of subsurface erosion in line with empirical evidence. Gully cross section was doubled to increase the amount of subsurface erosion (gully and streambank erosion) being generated from grazing lands. This has resulted in a doubling of gully erosion supply. The increase in subsurface erosion aligns with sediment tracing work in the Herbert (Tims et al. 2010) and this is further outlined in the discussion.

Sugarcane constituent generation Nil changes made since RC2014.

HowLeaky cropping constituent generation The ratio of FRP:DOP in sugarcane is an adjustable parameter in the model and in Report Card 2014 was set at a default of 80:20. Following further investigation, this was changed to 50:50 in Report Card 2015 to correct the under prediction of DOP at model evaluation sites, and is supported by field data. Instream water quality monitoring data collected in a sugarcane subcatchment in the Tully Basin showed that the spilt was 50:50 (Bainbridge et al. 2009). Additionally, data from the banana paddock scale trial at South Johnstone showed that the average of FRP:DOP was 50:50 (n =147) (Masters et al. 2016, unpublished data). Additional parameter changes for sugarcane include a reduction to the DWC for fine sediment, and a reduction to the HSDR for TSS and particulate nutrients. This was to better align with measured load estimates at the five monitoring sites (Table 6). The PP enrichment ratio was also reduced as it was over predicted at some sites, especially at Tully Euramo for the 2014 Report Card. In the WT, there were soils included in the modelling in RC2014 that we excluded in RC2015 and reallocated to different APSIM simulations.

36 Source Catchments modelling – Updated methodology and results

Table 6 Summary of changes to sugarcane parameters in the WT

NRM Region Parameter RC2014 RC2015 FRP:DOP ratio 80:20 50:50 TSS DWC 72 50 WT TSS & particulate HSDR 20 15 PP enrichment ratio 5 3 PN enrichment ratio 2 2

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) For land uses where Event Mean Concentration/Dry Weather Concentration (EMC/DWC) were used to generate loads, a number of changes were made to EMC/DWC concentrations to better align with loads estimated from GBRCLMP monitoring data (Wallace et al. 2015). Table 63, in Appendix 2 shows the values for RC2015. Concentrations of DIN, TSS and particulates upstream of the Barron River at Myola gauge were reduced by 25% and increased by 25% respectively to align better with the monitored loads at this gauge.

In-stream models For the WT, the instream and floodplain settling velocities were applied based on the stream erosion and deposition graph in Earle (2015). In-stream deposition and remobilisation was implemented in this Report Card. The WT instream model parameters are in Appendix 2, Table 60.

Predevelopment model As described on page 31 there are often instances where the predevelopment streambank load is greater than that of the baseline, or current, load, and the streambank management factor can be used to address this. The streambank management factor set to A-class, as described above. Predevelopment EMC/DWC values are in Appendix 2, Table 68.

37 Source Catchments modelling – Updated methodology and results

Burdekin The following section describes methodology changes that are only applicable to the BU region.

Hydrology Recalibration to improve baseflow representation, as described on page 28.

Hillslope sediment, nutrient and herbicide generation Maximum quickflow concentration adjusted, as described above on page 29.

Gully sediment and nutrient generation models Nil changes occurred since RC2014.

Sugarcane constituent generation Nil changes occurred since RC2014.

HowLeaky cropping constituent generation Nil changes occurred since RC2014.

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) The EMC/DWC values were updated to better align with monitored data. The values used in this Report Card are in Appendix 2, Table 64.

In-stream models Nil changes occurred since RC2014.

Predevelopment model As described on page 31 there are often instances where the predevelopment streambank load is greater than that of the baseline, or current, load, and the streambank management factor can be used to address this. The streambank management factor set to A-class, as described above. Predevelopment EMC/DWC values are in Appendix 2, Table 68.

38 Source Catchments modelling – Updated methodology and results

Mackay Whitsunday The following section describes methodology changes that are only applicable to the MW region.

Hydrology Recalibration to improve baseflow representation, as described on page 28.

Hillslope sediment, nutrient and herbicide generation Maximum quickflow concentration adjusted, as described above on page 29.

Gully sediment and nutrient generation models Nil changes occurred since RC2014.

Sugarcane constituent generation The DWC for TSS was reduced from 70mg/L to 10mg/L to better match the raw TSS concentration measurements collated during baseflow conditions as part of various water quality monitoring programs.

HowLeaky cropping constituent generation Nil changes occurred since RC2014.

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) The DWC for TSS was reduced from 25 to 10 in conservation and from 35 to 10 in forestry land uses to better represent the dry weather concentration measurements recorded as part of water quality monitoring programs (Turner et al. 2012, Turner et al. 2013, Garzon-Garcia et al. 2015, Wallace et al. 2015). The EMC/DWC values used in this Report Card are in Appendix 2, Table 65.

In-stream models Instream deposition and remobilisation were implemented in RC2015. As was the case for all models in which the in-stream models were enabled, deposition and remobilisation were kept in a similar magnitude where there was a lack of better local information. The MW instream model parameters are in Appendix 2, Table 60.

Predevelopment model As described on page 31 there are often instances where the predevelopment streambank load is greater than that of the baseline, or current, load, and the streambank management factor can be used to address this. The streambank management factor set to A-class, as described above. Predevelopment EMC/DWC values are in Appendix 2, Table 68.

39 Source Catchments modelling – Updated methodology and results

Fitzroy The following section describes methodology changes that are only applicable to the FZ region.

Hydrology Recalibration to improve baseflow representation, as described on page 28.

Hillslope sediment, nutrient and herbicide generation The soil erodibility layer (k-factor) was altered by applying a rock factor that reduced soil erodibility in areas where surface rock fragments are present. The rock factor was produced using the approach described by McCloskey et al. (2014) and was applied across all land uses. Applying a k-factor that accounts for rock fragments on the soil surface was implemented to account for high hillslope erosion rates from steep, usually rocky, areas.

Gully sediment and nutrient generation models Two related changes were made to the Fitzroy gully model. The gully full maturity year was lowered from 2010 to 1980 to reflect the reduction in the growth rate of gullies. Published literature indicates that gully activity has been declining in the Fitzroy (Hughes et al. 2009) and neighbouring Burdekin (Wilkinson et al. 2013). The activity factor, which works in combination with the gully maturity year parameter, was raised to 0.7 and now better reflects gully activity in the Fitzroy region based on visual assessments of gully erosion in aerial photography.

Sugarcane constituent generation Not applicable in the Fitzroy.

HowLeaky cropping constituent generation Hillslope erosion in cropping areas is modelled within HowLeaky and daily sediment loads are imported into the regional Source models. For RC2014 a uniform slope of 1% was applied across all cropping areas modelled by HowLeaky. Once imported into Source Catchments the slope-factor was used to scale the loads to account for the variable slope across cropping land uses. For RC2015 accounting for the spatial variability of slope was dealt with within the HowLeaky paddock modelling as opposed to in the Source Catchments model previously (Owens 2016).

Hillslope erosion has not been modelled in irrigated cropping land uses for this report card. In previous Report Cards, dryland cropping simulations were used to simulate hillslope erosion in irrigated cropping. This was not an accurate estimate of loads from irrigated cropping due to tailwater capture generating minimal runoff to stream except in large events. This approach will be adopted until a better understanding of the proportion of pollutants transported off irrigated cropping areas is acquired. The Queensland Department of Agriculture and Fisheries (DAF) are currently investigating irrigation practices and runoff capture within the Fitzroy region. The results of this research will be used to improve the modelling of hillslope erosion in irrigated cropping areas within the Fitzroy region.

40 Source Catchments modelling – Updated methodology and results

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) Concentrations for DIN, DON and DIP were reduced by a factor of 0.5, 0.8 and 0.8 respectively for all land uses and tebuthiuron concentrations were decreased by a factor of 0.8 for open grazing to better align with observed load estimates (Joo et al. 2012, Turner et al. 2012). The EMC/DWC values used in RC2015 are in Appendix 2, Table 66.

In-stream models Large increases in hillslope erosion were recorded for this report card due to increased loads from dryland cropping. This resulted in high export loads being generated compared to monitored loads. As a result the bank full flow recurrence interval was lowered to 4.7 years to increase floodplain deposition which reduced exported loads and improved model estimates compared to GBRCLMP monitored loads.

To better reflect stream characteristics updated regression relationships between upstream catchment area and stream width (Figure 3) and bank height and upstream catchment (Figure 4) were derived from local gauging station data and from imagery interpretation of channel width.

Instream deposition and remobilisation were investigated prior to the release of this report card, however more work is required to fully understand the interactions between load reductions and deposition and remobilisation of sediment and particulate nutrients. Further research into these processes will be conducted with a view to incorporating these processes into the model for the next report card.

Figure 3 Catchment area vs. stream width used to determine streambank erosion parameters

41 Source Catchments modelling – Updated methodology and results

Figure 4 Catchment area vs. bank height used to determine streambank erosion parameters

Predevelopment model As described on page 31 there are often instances where the predevelopment streambank load is greater than that of the baseline, or current, load, and the streambank management factor can be used to address this. The streambank management factor set to A-class, as described above. Predevelopment EMC/DWC values are in Appendix 2, Table 68.

42 Source Catchments modelling – Updated methodology and results

Burnett Mary The following section describes methodology changes that are only applicable to the BM region.

Hydrology Recalibration to improve baseflow representation, as described on page 28.

Hillslope sediment, nutrient and herbicide generation Enrichment ratios for PN and PP have been increased from 4 to 9 and from 5 to 10 respectively, to improve model agreement with loads estimated from catchment water quality monitoring.

Gully sediment and nutrient generation models Application of a 100% delivery ratio in the previous report cards contributed to the overestimation of TSS by the model. Because of this and the expectation that not all fine sediment generated from gullies will be supplied to the streams, the gully delivery ratio has been reduced from 100% in RC2014 to 80% in RC2015.

Sugarcane constituent generation The conversion factor parameter for DOP and DIP in APSIM (which is simply a calibration parameter to be able to scale modelled DOP and DIP values up or down) has been increased from the default of 0.8 to 1 and 1.2, respectively. This is to account for the application of mill mud (Moody and Schroeder 2014), which is not otherwise captured in the model.

Table 7 Summary of changes to sugarcane parameters in the BM

NRM region Parameter RC2014 RC2015 FRP:DOP ratio 80:20 100:120* TSS & Particulate HSDR 20 5 BM PP enrichment ratio 5 10 PN enrichment ratio 4 9

HowLeaky cropping constituent generation Nil changes made since RC2014.

Other land uses: Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) The DWC for USLE FUs (grazing, conservation and forestry land uses) for fine sediment and particulate nutrients has been set to zero. This was implemented to offset the high hillslope erosion loads being generated in the model in RC2014. The EMC/DWC values used in this Report Card are in Appendix 2, Table 67.

In-stream models Instream deposition and remobilisation were implemented in RC2015. As was the case for all models in which the in-stream models were enabled, deposition and remobilisation were kept in a similar magnitude where there was a lack of better local information. However, the relative effect of this on

43 Source Catchments modelling – Updated methodology and results modelled sediment and particulate loads needs to be investigated further in line with published literature.

Predevelopment model As described on page 31 there are often instances where the predevelopment streambank load is greater than that of the baseline, or current, load, and the streambank management factor can be used to address this. In the BM the streambank management factor was set to a constant value (0.71) as this was the average of the values applied in an A-class scenario model run. For consistency with the other regions, the universal value of 0.6 will be considered and tested. Predevelopment EMC/DWC values are in Appendix 2, Table 68.

44 Source Catchments modelling – Updated methodology and results

Results

The overall GBR results, and the results section for each of the six NRM regions are separated into hydrology and modelled loads. In the hydrology component the updated flow and baseflow calibration results are presented. The modelled loads section includes the updated baseline load and the overall progress towards the Reef Plan 2013 targets. For each NRM region, the predevelopment loads, constituent load evaluation, loads by land use, identification of sources and sinks, and management change loads are also presented.

Hydrology and regional discharge comparison The modelled average annual flow (1986–2014) for the GBR was 74,000,000 ML/yr, with the Wet Tropics contributing 32% (23,500,000 ML/yr) of the total flow (Figure 5). The Burnett Mary was the lowest contributor (7%). Flow increased by 750,000ML/yr, or 1% as a result of the baseflow recalibration. The greatest changes to average annual flow were in the BM region with an increase of 16% and CY has a decrease of 3% average annual flow.

Figure 5 Average annual modelled discharge for the GBR RC2015 (1986–2014) Modelled loads For Report Card 2015, the modelled average annual load of fine sediment exported from the six GBR NRM regions is approximately 9,925 kt/yr. Of this, approximately 20% is the predevelopment load, and ~7,894 kt/yr anthropogenic load. In Table 8 the total constituent baseline load for all regions is presented, while in Table 10 this data is presented as a per cent contribution.

The Wet Tropics is the greatest contributor for five of 11 constituents (Table 10), including TN, DIN, and PN. Almost 50% of the DIN baseline load exported to the GBR comes from the Wet Tropics. The

45 Source Catchments modelling – Updated methodology and results

Burdekin contributes the greatest fine sediment load of all regions, while the Fitzroy contributes the greatest TP, DIP and PP loads. The MW region has the greatest load of PSII herbicides, primarily a function of land use, specifically large areas of sugarcane, at ~4,000 kg/yr. Typically, CY was the lowest contributor for most constituents which again is a function of land use and the limited area of intensive agriculture. However, CY was the highest contributor for DON which is consistent with monitoring data. The total photosystem II (PSII) and the diuron toxic equivalent (TE) pesticide loads are presented in Table 9. The third column is the ratio of the TE PSII load to the PSII load. The closer the number is to 1, the higher the proportion of diuron making up the total load. The WT and MW regions have a high ratio compared to all other regions indicating that diuron is the dominant chemical in use (according to the modelled simulations). It is important to note that on average whilst MW exports approximately 1,000 kg/yr more PSII load than WT, they export a similar toxic equivalent load and a similar proportion of PSII/PSII TE load, meaning that WT exports more diuron.

Table 8 Total baseline loads for the GBR region

TSS TN DIN DON PN TP DIP DOP PP Basin (kt/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) Cape York 526 6,854 423 4,536 1,896 682 76 152 454

Wet Tropics 1,516 16,577 5,496 4,385 6,696 2,301 192 376 1,733

Burdekin 3,781 8,793 2,574 2,562 3,657 2,823 440 144 2,239 Mackay 818 4,806 1,355 1,301 2,150 1,306 249 65 992 Whitsunday Fitzroy 1,824 10,907 1,140 3,406 6,361 4,665 1,047 260 3,357

Burnett Mary 1,459 7,146 1,040 2,113 3,993 1,642 139 77 1,426

Total 9,925 55,083 12,028 18,303 24,753 13,418 2,143 1,074 10,201

Table 9 PSII pesticide load, PSII Toxic Equivalent (TE) load for the GBR region

PSII Toxic Proportion of PSII Basin Equivalent (TE) PSII TE load to (kg/yr) load (kg/yr) PSII load Cape York 16 1 0.06 Wet Tropics 3,090 2,415 0.78 Burdekin 2,394 1,284 0.54 Mackay Whitsunday 4,044 2,925 0.72 Fitzroy 2,210 47 0.02 Burnett Mary 346 86 0.25 Total 12,100 6,758 0.56

46 Source Catchments modelling – Updated methodology and results

Table 10 Regional contribution of each constituent as a percentage of the GBR total baseline load

TSS TN DIN DON PN TP DIP DOP PP PSII PSII TE Basin (kt/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (kg/yr) (kg/yr) Cape York 5 12 4 25 8 5 4 14 4 0 0

Wet Tropics 15 30 46 24 27 17 9 35 17 26 36

Burdekin 38 16 22 14 15 21 21 13 22 20 19 Mackay 8 9 11 7 9 10 12 6 10 33 43 Whitsunday Fitzroy 18 20 9 19 26 35 49 24 33 18 1 Burnett 15 13 9 12 16 12 6 7 14 3 1 Mary Total 100 100 100 100 100 100 100 100 100 100 100

Across the GBR, the modelling suggests that current fine sediment loads are approximately five times the pre-development load (Figure 6). This varies for each region from twice the pre-development load in CY, to seven times the load in the Fitzroy and Burdekin NRM regions. Total phosphorus loads are approximately three times the pre-development load across the GBR, ranging from twice the pre- development load in CY to three times in the Fitzroy and Burdekin. Total nitrogen is more consistent across the regions with an approximate doubling of the load. The predevelopment load for DIN is also about half of the anthropogenic load (Figure 7) in those regions where sugarcane is prevalent. In the other regions (CY and FZ) there is minimal anthropogenic DIN load.

47 Source Catchments modelling – Updated methodology and results

Figure 6 Predevelopment and anthropogenic baseline loads for all 35 basins in the GBR for TSS

48 Source Catchments modelling – Updated methodology and results

Figure 7 Predevelopment and anthropogenic baseline loads for all 35 basins in the GBR for DIN Contribution by land use Grazing (43%) and streambank erosion (31%) are the greatest contributors to the baseline TSS load from the GBR (Figure 8). Nature conservation (13%) and sugarcane (7%) also make notable contributions to total loads. Sugarcane (43%) is by far the greatest contributor to the GBR DIN load (Figure 9), followed by grazing (23%) and nature conservation (20%). Unlike their contributions to TSS, the DIN load from bananas, horticulture and urban/other is no longer negligible despite their small area. Bananas are modelled only in the WT and CY, and have a greater load per unit area (12.3kg/ha/yr) of DIN than sugarcane (9.6kg/ha/yr) does across the GBR. Whilst the contribution from land uses such as bananas, dairy, and horticulture are negligible to DIN on a whole of GBR basis (Figure 8), within a

49 Source Catchments modelling – Updated methodology and results region individual industries may contribute a reasonable amount of the total load exported. For example in the WT, bananas contributes 5% of the total DIN load (Figure 24).

Figure 8 Baseline GBR TSS load contribution by land use

Figure 9 Baseline GBR DIN load contribution by land use

50 Source Catchments modelling – Updated methodology and results

Progress towards Reef Plan 2013 targets Progress towards the Reef Plan targets for the whole of GBR is presented in Figure 10. Fine sediment has been reduced by 0.5%, with the greatest reduction occurring in the Fitzroy region. Total nitrogen has been reduced by 0.5%, with the greatest reductions in the Burdekin (DIN reduction of 4%) and MW region (DIN reduction of 1%). A 3% reduction has been achieved for the toxic equivalent load of PSIIs, with the greatest reductions in the WT (3.5%), Burdekin (3.5%) and MW (3%) regions.

Figure 10 Progress towards Reef Plan targets till 2015 for the whole of GBR

Scenario modelling – All A and All B management Modelled load reductions were calculated from the All A and All B management practice adoption scenarios for TSS, both particulate and dissolved nutrients and for the five PSII herbicides of interest, aggregated to a PSII TE. The reductions reported represent the modelled load reductions from 2009 – 2015.

The modelling results suggest that under an All B scenario, the 20% TSS target is close to being met for the whole of GBR and can be achieved in the MW and Fitzroy regions (Figure 11). Under an All A scenario, all regions can achieve the target with the exception of CY, achieving just under the 20% target.

For DIN, the 50% target cannot be met for the whole of GBR under the All A or All B scenario. The Burdekin region is the only region that achieved the target under the All B scenario (Figure 12). For pesticides, the PSII TE target can be achieved under an All B scenario (Figure 13). Alternative,

51 Source Catchments modelling – Updated methodology and results innovative management practices should be explored that could assist in reaching the Reef Water Quality targets.

For particulate nitrogen, under an All B scenario the 20% target can be met for the whole of GBR and can be achieved in the MW and Fitzroy regions (Figure 14). Under an All A scenario all regions can achieve the target (Figure 14). For particulate phosphorus, under an All B scenario the 20% target can be met for the whole of GBR and can be achieved in the majority of regions namely WT, BU, MW, and FZ (Figure 15). Under an All A scenario all regions can achieve the target with the exception of CY, achieving just under the 20% target (Figure 15).

In summary, the modelling results suggest that the TSS and particulate nutrient targets are achievable under an All A scenario with the majority of the regions close to achieving the TSS target under an All B scenario. The DIN target may be more challenging to achieve even under the All A scenario. The PSII TE target is achievable even under an All B scenario.

Figure 11 Estimated TSS reduction under All A and All B management practice scenarios

52 Source Catchments modelling – Updated methodology and results

Figure 12 Estimated DIN reduction under All A and All B management practice scenarios

Figure 13 Estimated PSII TE reduction under All A and All B management practice scenarios

53 Source Catchments modelling – Updated methodology and results

Figure 14 Estimated PN reduction under All A and All B management practice scenarios

Figure 15 Estimated PP reduction under All A and All B management practice scenarios

54 Source Catchments modelling – Updated methodology and results

Cape York The results of the Cape York modelling are presented below, separated into hydrology and modelled loads. In the hydrology component, the results of the updated calibration process will be presented, as well as the impact on the total flow for the CY region. The modelled loads section includes the results of the total baseline, anthropogenic baseline and predevelopment loads. The validation of the CY Source Catchments modelled data is then presented using load estimates from measured data. Progress towards targets due to investment is reported against the anthropogenic baseline for RC2015. A summary of the total baseline load by land use, and land use by basin is also reported, as well as a mass balance summary of the sources and sinks by constituent. For a full list of the CY region loads for RC2015 refer Appendix 3, Table 69.

Hydrology calibration performance Recalibrating to improve baseflow did not change the number of gauges meeting the performance criteria. All gauges continued to meet the R2 and PBIAS criteria, and eight of 15 continued to meet the daily NSE criteria. The purpose of the baseflow recalibration was to improve the proportion of flow attributed to quick and baseflow. The Lyne-Hollick filter (Lyne and Hollick 1979) was used to separate baseflow in the observed data (Table 11); the modelled baseflow percent was also calculated. There is a very clear improvement in the RC2015 hydrology calibration in which the modelled baseflow proportion better matches the Lyne-Hollick calculated baseflow.

The recalibration of hydrology in CY to improve the baseflow component has seen average annual flow reduced by 3% of total annual flow (~500,000 ML/yr) to an average annual flow of 16,415,407 ML/yr.

Table 11 Baseflow proportions (%) for RC2014, RC2015 and calculated for seven CY gauges

RC2014 RC2015 LH calculated Gauge Basin baseflow baseflow baseflow (%) (%) 104001A Stewart 63 43 29 105001B Normanby 75 42 42 105105A Normanby 68 29 29 105107A Normanby 95 40 33 107001B Annan 65 38 38 106001A Jeannie 66 36 35 106002A Jeannie 78 31 29

Modelled Loads Cape York contributes 5% of the total baseline TSS load being exported to the GBR. CY was the lowest contributor of TSS load, and is one of the lowest contributors for all constituents other than DON. Within the CY region, the Normanby River Basin was the greatest contributor for all constituents (Table 12). All pesticide loads modelled in CY come from cropping and banana land uses in the Normanby, Endeavour or Jeannie River basins.

55 Source Catchments modelling – Updated methodology and results

Table 12 Contribution from CY basins to the total CY baseline load

TSS TN DIN DON PN TP DIP DOP PP PSII PSII TE Basin (kt/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (kg/yr) (kg/yr) Jacky 52 1,064 67 728 269 77 11 22 44 0 0 Jacky Olive 72 1,613 98 1,064 451 126 17 34 75 0 0 Pascoe

Lockhart 67 896 49 532 315 118 8 15 95 0 0

Stewart 49 516 30 333 153 60 5 10 45 0 0

Normanby 186 1,428 105 1,070 253 142 21 43 78 11.2 0.40

Jeannie 42 617 35 378 204 59 6 12 41 0.2 0.01

Endeavour 59 721 40 430 251 101 8 16 77 4.4 0.16

Total 526 6,854 423 4,536 1,896 682 76 152 454 16 1

Constituent load validation The extension of the modelling period through to 2014 for RC2015 provides eight years of water quality monitoring data from the Kalpowar gauging station (Joo et al. 2012, Turner et al. 2012, Turner et al. 2013, Garzon-Garcia et al. 2015, Wallace et al. 2015), collected as part of the GBR Catchment Loads Monitoring Program (GBRCLMP). Having eight years of water quality data collected by the GBRCLMP, the data calculated by Kroon et al. (2012) and Joo et al. (2014) used as previous sources of validation has not been utilised for RC2015. The GBRCLMP data sets are well sampled over the majority of the events, this is in comparison to previous estimates which use historic ambient data often collected sporadically and in low flow conditions.

A comparison was made between the mean GBRCLMP flow and loads (averaged over eight years, 2006–2014) and the modelled flow and load for the same location and the same time period (Figure 16). The modelled flow was within ±11% of the observed flow and constituent loads within ±35% at the Kalpowar Crossing gauging station for the eight year period. Across the monitoring period, the model is under-predicting eight of nine constituents, with only TSS over-predicted. Particulate nitrogen (PN) and DON are the two constituents that have the greatest difference compared to monitored data, at - 35% and -29% respectively. The under-prediction of particulate nutrients has been previously discussed (page 29), with questions around the reliability of the input data (number of samples used in extrapolation across the region). With respect to DON, the monitoring data will be reviewed further to improve the modelled load estimates. For example monitoring data indicates that nature conservation areas are a large contributor of DON to the end of system load.

56 Source Catchments modelling – Updated methodology and results

Figure 16 Average annual modelled versus measured load comparison for 2006-2014 period, Normanby River Gauge at the Kalpowar Crossing (105017A)

The performance of the RC2015 baseline model to predict the range of constituents modelled across the CY region was also assessed using a number of quantitative statistics (Moriasi et al. 2015). The statistics and the corresponding model performance ratings are presented in Table 13. An overall performance rating was established as recommended quantitative statistics revealed conflicting/unbalanced performance ratings for the majority of the constituents. For ease of presentation, ‘very good’ and ‘good’ rated sites are coloured green, while ‘satisfactory’ and ‘indicative’ are coloured orange, signalling that further work is required to improve the representation of these constituents.

The RSR, R2 and NSE yield similar results for all constituents, with a mix of ‘good’ or better rankings and some ‘indicative’ rankings. Eight of the 10 constituents achieved a ‘satisfactory’ or better rating. Flow was ranked as ‘very good’ across three measures; TP, DOP and DIP also ranked ‘satisfactory’ or better, while TSS was only ranked ‘good’ through the PBIAS measure. DIN was better than ‘good’ for all measures while TN, PN and DON achieved mixed results.

57 Source Catchments modelling – Updated methodology and results

Table 13 Moriasi et al. (2007), Moriasi et al. (2015) annual statistical measures of performance for each constituent at the Kalpowar Gauge, Cape York

RSR PBIAS NSE R2 Constituent Rating Rating Rating Rating Value (Moriasi Value (Moriasi Value (Moriasi Value (Moriasi 2007) 2015) 2015) 2015) Flow 0.3 Very good 11.4 Satisfactory 0.9 Very good 1.0 Very good TSS 1.3 Indicative -25.2 Indicative -0.8 Indicative 0.1 Indicative TN 0.4 Very good 21.2 Satisfactory 0.9 Very good 1.0 Very good PN 0.9 Indicative 36.8 Indicative 0.2 Indicative 0.0 Indicative DIN 0.5 Good 1.2 Very good 0.8 Very good 0.8 Very good DON 0.4 Very good 18.1 Good 0.9 Very good 0.8 Very good TP 0.6 Good 21.8 Satisfactory 0.7 Very good 0.9 Very good DIP 0.7 Satisfactory 20.0 Good 0.6 Good 0.6 Satisfactory PP 0.4 Very good 20.4 Satisfactory 0.8 Very good 0.7 Good DOP 0.6 Good 19.1 Good 0.7 Very good 0.5 Satisfactory

Contribution by land use Nature conservation and grazing (open and closed) account for 61% and 37% of the CY area respectively and generate 99% of the fine sediment load and over 95% of all other constituent loads exported (Figure 17). PSII toxic load is only derived from cropping land uses.

Figure 17 Contribution to fine sediment, particulate nutrient and DIN export by land use for the CY region

58 Source Catchments modelling – Updated methodology and results

Sources and sinks Of the total fine sediment generated (902 kt/yr) in CY, 51% is exported to the GBR lagoon. Of the material remaining in the catchment, 92% is deposited on the floodplain, and 8% instream. Over 90% of TN is exported, while 80% of the TP is exported. DON makes up over half of the TN load, while PP makes up 74% of the TP load. Thus, less TP is exported to the GBR than TN, because more TP is deposited on the floodplain or instream via PP. In Appendix 3, Table 70 the sources and sinks for each constituent are presented. The source sink breakdown for fine sediment is presented as a percent of total supply in Table 14, for both CY and the Normanby Basin, and diagrammatically in Figure 18 for the Normanby Basin. For the Normanby Basin, the improved gully density map (updated in RC2014) has resulted in a more realistic spatial representation of gullies based on field measurements gully erosion is now the major sediment source in that Basin.

Table 14 Sources and sinks of fine sediment in CY and the Normanby Basin (as percent of supply total)

CY (%) Normanby (%) SUPPLY Channel remobilisation 3 4 Gully 45 75 Hillslope surface soil 47 17 Streambank 5 4 LOSS Floodplain deposition 38 58 In-stream deposition 3 5

Figure 18 Fine sediment source sink budget diagram for the Normanby Basin

59 Source Catchments modelling – Updated methodology and results

Progress towards Reef Plan 2013 targets Minimal management changes were reported for CY for RC2015, with two gully improvement projects (total 93Ha) and riparian fencing (four projects, total 36,391 Ha) reported in the Normanby Basin (Table 15). The gully improvement projects resulted in a 0.01% shift out of C class management practices into B; while the riparian fencing projects are reflected as streambank management practise and resulted in a 0.25% shift into C class, and a 0.72% shift into B class. There were no other grazing, cropping or banana improvements reported. These small changes resulted in no impact on load reduction targets from CY.

Table 15 CY grazing management practice proportions and change summary; values are percent of total area

Hillslope Baseline RC2014 Change RC2015 Change A 0 0 0 0 0 B 20 20 0 20 0 C 47 47 0 47 0 D 33 33 0 33 0 Gully Baseline RC2014 Change RC2015 Change A 0 0 0 0 0 B 34.21 34.21 0 34.22 0.01 C 46.55 46.55 0 46.54 -0.01 D 19 19 0 19 0 Streambank Baseline RC2014 Change RC2015 Change A 0 0 0 0.00 0 B 27.8 27.8 0 28.47 0.72 C 42.0 42.0 0 42.27 0.25 D 30.2 30.2 0 29.26 -0.97

60 Source Catchments modelling – Updated methodology and results

Wet Tropics The results of the Wet Tropics modelling are presented below, separated into hydrology and modelled loads. In the hydrology component, the results of the updated calibration process will be presented per gauge, as well as the impact on the total flow for the WT region. The modelled loads section includes the results of the total baseline, anthropogenic baseline and predevelopment loads. The validation of the WT Source Catchments modelled data is then presented using load estimates from measured data. Progress towards targets due to investment is reported against the anthropogenic baseline for RC2015. A summary of the total baseline load by land use, and land use by basin is also reported, as well as a mass balance summary of the sources and sinks by constituent. For a full list of the WT region loads for RC2015 refer to Appendix 3, Table 71.

Hydrology calibration performance The modelled hydrology was recalibrated to improve the baseflow proportion contributing to total flow. There is an improvement in the RC2015 baseflow proportion when compared to the Lyne-Hollick calculated baseflow (Table 16). All gauges continued to meet the daily Nash-Sutcliffe, R2 and PBIAS criteria. There was little difference in the range of total flow bias between this Report Card and RC2014. Overall the annual flow for the WT increased by 1% to ~23,500,000 ML/yr with the biggest changes occurring the Tully (6% increase) and in the Barron (8% decrease). All other basins had a <1% change in total flow. The range of baseflow proportions of the 10 gauges before recalibration was 23%–97% and after calibration was 31%–55%. The biggest changes in baseflow proportions from RC2014 and RC2015 was from the Herbert gauge 116014A from 82% to 31%, closely matching the Lyne-Hollick filter value of 29%. The other biggest change was in the Barron gauge 110016A from 97% to 55%, again aligning closer to the Lyne-Hollick filter value of 47%.

Table 16 WT baseflow proportions for the current and previous Report Card

RC2014 RC2015 LH Gauge Basin baseflow baseflow baseflow (%) (%) filter (%) 108002A Daintree 31 47 47 110016A Barron 97 55 47 110001A-D Barron 23 40 42 110020A Barron 27 40 44 111101A-D Russell-Mulgrave 31 48 52 113006A Tully 46 51 61 114001A Murray 36 48 49 116001A-F Herbert 24 38 39 116014A Herbert 82 31 29 116010A Herbert 62 47 45

61 Source Catchments modelling – Updated methodology and results

Modelled Loads The WT contributes the third highest proportion of the total baseline TSS load being exported to the GBR at 15%, equal to the Burnett Mary region (Table 10). For particulate nutrients, the WT and the Fitzroy had the highest contributions of PN ~27% each and the third highest contribution of PP at 17%. For dissolved nutrients, the DIN contribution was 46% of the GBR export load, roughly double that of the second highest contributor, the Burdekin. The WT contributes the second highest contribution of PSIIs (26%) after the Mackay Whitsundays (33%). The contribution to total WT load for each basin is shown in Table 17. Within the WT region, the Herbert and/or Johnstone basins produced the largest loads for most constituents, with the Mulgrave-Russell Basin also an important contributor. The Johnstone, Mulgrave-Russell and Tully produced the largest PSII and PSII TE export loads. These three basins contributed 74% of the PSII TE export load from the WT. Of the total PSII load, 73% was attributed to Diuron. The proportion of PSII TE load to PSII load is 0.78. The total pesticide load exported from the WT was 8,953 kg/yr, which includes both PSIIs and non-PSIIs. Compared to the baseline load for RC2014, the biggest changes in export loads (± 20%) are for phosphorus species. All phosphorus loads decreased except for DOP load which had a large increase (89%). The TSS load decreased by 9% and the export loads for nitrogen species increased (4–15%). The PSII load decreased by 16% and the PSII TE load decreased by 14%.

Table 17 Contribution from WT basins to the total baseline load

TSS TN DIN DON PN TP DIP DOP PP PSII PSII TE Basin (kt/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (kg/yr) (kg/yr) Daintree 103 1,841 478 784 579 146 19 44 83 118 98 Mossman 17 323 160 62 101 28 3 8 17 75 65 Barron 55 568 152 215 201 72 12 15 45 120 37 Mulgrave- 253 2,769 934 609 1,227 392 40 76 276 746 625 Russell Johnstone 379 3,799 1,059 754 1,986 864 35 72 756 922 753 Tully 157 2,155 777 503 875 259 24 54 181 492 414 Murray 74 1,121 414 307 400 121 12 25 84 367 310 Herbert 478 4,001 1,522 1,152 1,327 419 45 82 291 248 113 Total 1,516 16,577 5,496 4,385 6,696 2,301 192 376 1,733 3,090 2,415

Anthropogenic baseline and predevelopment loads The anthropogenic baseline load was calculated by subtracting the predevelopment load from the total baseline load. The TSS anthropogenic baseline load was 936 kt/yr or 62% of the total baseline load with the remaining 38% attributed to the predevelopment load (Figure 19). For all basins except the Daintree and Mossman the anthropogenic load was over half of the total TSS load (Figure 20) and (Table 71 (Appendix 3). The anthropogenic baseline DIN load was 2,750 t/yr or 50% of the total baseline load. The Mossman Basin had the highest proportion of the anthropogenic DIN baseline load to the total load at 65%. The Herbert, Barron and Murray basins had similar anthropogenic proportions of ~57% each. Predevelopment DIN proportions to the total load were highest in the Daintree (72%), followed by the Mulgrave-Russell, Johnstone and Tully basins (just over 50% each). Comparing across

62 Source Catchments modelling – Updated methodology and results the constituents, the highest anthropogenic baseline loads were PP at 70% followed by FRP at 65% and TSS at 62%.

Figure 19 TSS (kt/yr) loads for the WT basins, predevelopment and anthropogenic baseline contributions

Figure 20 DIN (t/yr) loads for the WT basins, predevelopment and anthropogenic baseline contributions

63 Source Catchments modelling – Updated methodology and results

Constituent load validation The first source of validation data was the GBRCLMP monitored loads for six sites within the WT NRM region, for the period 2006–2014. For the modelled baseline loads, there is a general under prediction in modelled TSS and particulate loads compared to the GBRCLMP loads, except for Tully River at Euramo (Figure 21). The average percent bias across the five under predicted sites was -33% for TSS, -29% for PN and -40% for PP. For Tully River at Euramo, the average across TSS and particulates is within +50%. There was also a general under prediction for DIN modelled loads, except for Barron River at Myola (Figure 22). All modelled DIN loads are within ±43% of the measured estimates. DON, FRP and DOP modelled loads are all within ±38% of the measured estimates, mostly under predicted. Two of the three PSII loads were also under predicting compared to the load derived from measured estimates (Figure 23). See Table 81 to Table 83, Appendix 4 for the tabular results of the GBRCLMP validation.

Figure 21 Fine sediment comparison between GBRCLMP measured loads and Source modelled loads for five sites in the WT

64 Source Catchments modelling – Updated methodology and results

Figure 22 DIN comparison between GBRCLMP measured loads and Source modelled loads for five sites in the WT

Figure 23 PSII comparison between GBRCLMP measured loads and Source modelled loads for five sites in the WT

The performance of the RC2015 baseline model was also assessed using a number of quantitative statistics recommended in Moriasi et al. (2015). An overall performance rating was established as recommended quantitative statistics revealed conflicting/unbalanced performance ratings for most of the constituents. The overall performance was described conservatively using the least rating of the recommended statistics. For ease of presentation, ‘very good’ and ‘good’ rated sites are coloured green, while ‘satisfactory’ and ‘indicative’ are coloured orange, signalling that further work is required to improve the representation of these constituents.

65 Source Catchments modelling – Updated methodology and results

Of the 10 modelled constituents (flow, sediment and nutrients) across the five sites, 60% were ranked as ‘indicative’, 18% ‘satisfactory’ and 22% ‘good’/’very good’. The North Johnstone site (Table 18 and Table 19) ranked as the best overall, followed by South Johnstone and Barron. DON had the highest performance ratings followed by DIP and flow. The lowest performing basin was Tully, ranked as ‘indicative’ for all model parameters. TSS and PP at all sites ranked as ‘indicative’. The data at each of the five sites is presented in Appendix 4. Statistics could not be calculated at the daily or monthly time- step due to the load calculation method used in the GRBCLMP data. Load comparisons were not undertaken for the Tully Gorge at Tully river site nor for any of the PSII herbicides as there were less than the required five years of data available.

Table 18 Moriasi et al. (2015) annual performance ratings at the North Johnstone River gauging station compared to GBRCLMP annual loads

Time period 2006-2014 Site 112004A RC7 Annual (excludes 2008-2009) Modelled RSR rating PBIAS rating NSE rating R2 parameter Flow Good Satisfactory Good Very good TSS Indicative Indicative Indicative Satisfactory TN Good Good Very good Very good PN Indicative Satisfactory Satisfactory Very good DIN Very good Very good Very good Very good DON Very good Very good Very good Very good TP Satisfactory Satisfactory Good Good DIP Very good Good Very good Very good PP Good Indicative Satisfactory Satisfactory DOP Indicative Satisfactory Satisfactory Very good

Table 19 Moriasi et al. (2015) monthly performance rating at the North Johnstone River gauging station

Site 112004A RC7 monthly Time period 1986–2009 Modelled RSR rating PBIAS rating NSE rating R2 parameter Flow Very good Very good Very good Very good TSS Good Indicative Satisfactory Good TN Good Indicative Very good Very good PN Indicative Indicative Satisfactory Very good DIN Very good Very good Very good Very good DON Very good Very good Very good Very good TP Satisfactory Indicative Good Very good DIP Very good Very good Very good Very good PP Satisfactory Indicative Good Very good DOP Indicative Indicative Satisfactory Very good

The same quantitative statistics were also calculated on the long-term monitored monthly loads (Joo et al. 2014) (1986–2009) versus the modelled loads with the overall ranking for each constituent at each of the five gauging stations (see Table 20). Of the 10 modelled constituents (flow, sediments and

66 Source Catchments modelling – Updated methodology and results nutrients) across the five sites, 50% were ranked as ‘indicative’, 10% ‘satisfactory’ and 40% ‘good’/’very good’. The South Johnstone site ranked as the best overall, followed by the Herbert site. Flow performed as the best parameter followed by DON. The lowest performing basin was the Barron.

Table 20 Moriasi et al. (2015) overall monthly performance rating for each constituent for the five gauge sites, Wet Tropics

Overall rating RC2015 (1986–2009) Monthly Modelled North South Barron Tully Herbert parameter Johnstone Johnstone Flow Very good Very good Very good Very good Very good TSS Indicative Indicative Satisfactory Indicative Indicative TN Indicative Indicative Good Indicative Very good PN Indicative Indicative Satisfactory Indicative Satisfactory DIN Indicative Very good Very good Very good Indicative DON Good Very good Very good Good Satisfactory TP Indicative Indicative Good Indicative Good DIP Very good Very good Indicative Satisfactory Very good PP Indicative Indicative Good Indicative Indicative DOP Indicative Indicative Indicative Indicative Indicative

Contribution by land use Sugarcane and nature conservation account for 8% and 45% of the WT area respectively, and generated 77% of the DIN export load (Figure 24), with sugarcane generating the largest load 2,572 t/yr (Figure 25). By areal rate, bananas generate the highest areal rate 16 kg/ha/yr, followed by sugarcane 14 kg/ha/yr (Figure 25). The average DIN areal rate for all land uses in the WT is 3 kg/ha/yr.

Figure 24 Contribution to fine sediment, particulate nutrient and DIN export by land use for the WT region

67 Source Catchments modelling – Updated methodology and results

Figure 25 WT DIN load and areal rate per land use

68 Source Catchments modelling – Updated methodology and results

Sugarcane areal rates of TSS, DIN and PSIIs by basin The average annual sugarcane areal rate for TSS was 2.4 t/ha/yr, with the Johnstone Basin approximately 2.5 times the WT average (Table 21). The Daintree and Mossman basins had the highest DIN areal rates of ~ 21.5 kg/ha/yr each. For PSIIs, the average areal rate was 16 g/ha/yr and the highest rates were for Johnstone and Mulgrave-Russell basins at ~31 g/ha/yr each.

Table 21 TSS, DIN and PSII sugarcane areal rates by basin for the WT

Sugarcane Sugarcane Sugarcane Basin TSS DIN PSIIs (t/ha/yr) (kg/ha/yr) (g/ha/yr) Daintree 1.6 22 26 Mossman 1.2 21 16 Barron 0.3 9 7 Mulgrave-Russell 4.3 16 30 Johnstone 5.6 15 32 Tully 3.1 16 26 Murray 1.7 14 21 Herbert 0.8 13 3 WT 2.4 14 16

Sources and sinks The sources, sinks and resultant export of TSS is presented in Appendix 3, Table 72, and the source sink budget is presented diagrammatically in Figure 26. The greatest sources of TSS in the WT were from hillslopes 66% and streambanks 26%, with contribution from gullies 5%. Instream channel remobilisation and deposition were enabled in RC2015 and channel remobilisation accounts for the remaining 3% of supply. The gully supply load doubled from RC2014 due to a doubling of the gully cross sectional area. Of the TSS generated (1,766 kt/yr) 86% is exported to the GBR lagoon. Of the 250 kt/yr TSS loss, the majority of this was floodplain deposition (48%) and instream deposition (45%).

A large proportion of the dissolved nitrogen supply (DIN and DON) (99%) was from diffuse dissolved sources such as agriculture and nature conservation areas with the remaining 1% from point sources (sewage treatment plants). A large proportion (99%) of the dissolved nutrients were exported. The DIN export load for sugarcane is spilt into a surface runoff load and a seepage load. Of the total DIN supply load from sugarcane (2,582 t/yr), 90% of the DIN load to the stream was from seepage (2,325 t/yr) and 10% to surface runoff (257 t/yr). The modelled PSII pesticide supplied to the stream was 3,505 kg/yr, with 88% of supply (3,090 kg/yr) exported to the end of system. For herbicides, the main loss pathway was from instream decay.

69 Source Catchments modelling – Updated methodology and results

Figure 26 Fine sediment source sink budget diagram for WT

Progress towards Reef Plan 2013 targets In RC2015, investments in improved management practices were reported for sugarcane and bananas. No changes were reported for grazing, streambank or gully management. Soil and nutrient management changes for sugarcane were typically less than 1%. For example, the greatest change was a shift in area into A class soil management of 0.8%. Pesticide management for sugarcane reported larger shifts, with an increase of 2.8% from D and C class management into B class, and an increase of 1.1% into A class. The management changes for sugarcane are presented in Table 22. There was a much greater shift out of D and C class practices in RC2014 compared to RC2015.

For bananas there were shifts out of D and C class practices for soils, into B class (3.9%) with a similar trend occurring for nutrients. No shifts into ‘A’ class management was reported for soil or nutrients for bananas in the WT. The management practice summary for bananas is presented in Table 23. Similar to sugarcane, there was a much greater shift out of D and C class practices into B in RC2014 compared to RC2015.

The impact of RC2015 sugarcane management changes for pesticides is a 3.5% reduction in PSII toxic load, while reductions for other constituents were in the range of 0.1–0.2% combined for sugarcane and banana practice changes. The cumulative reductions since RC2010 are 13.6% for TSS, 14.7% for DIN and 31.9% for PSII TE.

70 Source Catchments modelling – Updated methodology and results

Table 22 Wet Tropics sugarcane management practice change summary; values are percent of total

Baseline Soil RC2014 Change RC2015 Change 2013 A 9.2 13.4 4.2 14.2 0.8 B 31.4 31.9 0.4 32.0 0.1 C 43.4 42.4 -1.0 41.6 -0.8 D 16.0 12.3 -3.7 12.2 -0.1 Nutrients A 2.0 7.9 5.9 7.9 0.0 B 10.4 11.1 0.7 11.6 0.5 C 84.3 77.7 -6.6 77.2 -0.5 D 3.2 3.2 0.0 3.2 0.0 Pesticides A 3.9 4.1 0.2 5.3 1.1 B 8.4 15.9 7.5 18.7 2.8 C 80.9 74.9 -6.0 71.2 -3.7 D 6.9 5.2 -1.7 4.9 -0.2

Table 23 Wet Tropics banana management practice change summary; values are percent of total

Baseline Soil RC2014 Change RC2015 Change 2013 A - - - - - B 45 53.5 8.75 57.4 3.9 C 38 29.4 -8.46 28.1 -1.3 D 17 17.2 -0.29 14.6 -2.6 Nutrients A - - - - - B 42 52.3 10.74 55.6 3.3 C 50 40.0 -10.07 37.3 -2.7 D 8 7.7 -0.67 7.1 -0.6

71 Source Catchments modelling – Updated methodology and results

Burdekin The results of the Burdekin modelling are presented below, separated into hydrology and modelled loads. In the hydrology component, the results of the updated calibration process will be presented per gauge, as well as the impact on the total flow for the Burdekin region. The modelled loads section includes the results of the total baseline, anthropogenic baseline and predevelopment loads. The validation of the Burdekin Source Catchments modelled data is then presented using load estimates from measured data. Progress towards targets due to investment is reported against the anthropogenic baseline for RC2015. A summary of the total baseline load by land use, and land use by basin is also reported, as well as a mass balance summary of the sources and sinks by constituent. For a full list of the Burdekin region loads for RC2015 refer Appendix 3.

Hydrology calibration performance Recalibrating to improve baseflow did not change the number of gauges meeting the performance criteria. All gauges continued to meet the R2 and PBIAS criteria, and 42 of 46 continued to meet the daily NSE criteria. The purpose of the baseflow recalibration was to improve the proportion of flow attributed to quick and baseflow. In Table 24 the Lyne-Hollick filter (Lyne and Hollick 1979) was used to separate baseflow in the observed data; and the modelled baseflow percent is also presented. There is a very clear improvement in the RC2015 hydrology calibration in which the modelled baseflow proportion better matches the Lyne-Hollick calculated baseflow.

In terms of impacts on discharge, the recalibration of hydrology in the Burdekin Region to improve the baseflow component has seen average annual flow increased by 4% (~450,000 ML/yr) from RC2014 to an average annual flow of 12,720,701 ML/yr.

Table 24 Burdekin baseflow proportions for six gauges the current and previous Report Card

RC2014 RC2015 LH baseflow Gauge Basin baseflow (%) baseflow (%) filter (%)

119101A Haughton 60 16 17 120302B Cape 100 14 16 120304A Belyando Suttor 24 17 11 120305A Belyando Suttor 48 17 9 120306A Belyando Suttor 53 6 6

Modelled Loads The Burdekin region contributes 38% of the total baseline TSS load being exported to the GBR. The Burdekin Region was the highest contributor of TSS load, and is one of the highest contributors for all constituents other than DON and DOP. The estimated baseline loads from RC2015 have not changed significantly from RC2014, with a 2% increase in TSS modelled. Within the Burdekin region, the Burdekin River Basin was the greatest contributor for all constituents, apart from PSII which is sourced

72 Source Catchments modelling – Updated methodology and results from the Haughton (Table 25). All pesticide loads modelled in the Burdekin Region come from sugarcane and cropping land uses, with the load dominated by sugarcane from the Burdekin Irrigation area.

Table 25 Contribution from Burdekin basins to the total Burdekin baseline load

TSS TN DIN DON PN TP DIP DOP PP PSII PSII TE Basin (kt/y) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (kg/y) (kg/yr)

Black 62 385 97 145 143 105 21 7 77 2 2

Ross 62 416 180 138 98 94 33 11 50 2 0

Haughton 183 1,474 1,016 235 223 236 73 21 141 1,708 937

Burdekin 3,260 5,849 1,104 1,857 2,887 2,177 285 94 1,797 503 255

Don 213 669 177 186 305 211 27 10 174 179 90

Total 3,781 8,793 2,574 2,562 3,657 2,823 440 144 2,239 2,394 1,284

Anthropogenic baseline and predevelopment loads The TSS increase factor for the whole of the BU region is ~5.6; the Haughton, Don and Burdekin basins have the greatest increase of ~6.3 and ~5.9 (both Don and Burdekin) times respectively, while the Ross Basin has a ~3.8 times increase and the Black Basin has the least increase of ~1.2 times to the predevelopment load (Figure 27).

In regard to DIN, the anthropogenic increase factor was 1.01 for the BU region. The Haughton Basin was the main contributor to anthropogenic DIN with an increase factor of nine. All the other basins had anthropogenic increase factors of less than 2.2 times, with the Burdekin (0.2 times), Black (0.3 times) and DON (0.6 times) having a negligible anthropogenic DIN load. (Figure 28).

73 Source Catchments modelling – Updated methodology and results

Figure 27 TSS (kt/yr) loads for the BU basins, highlighting the predevelopment and anthropogenic baseline contributions

Figure 28 DIN (kt/yr) loads for the BU basins, highlighting the predevelopment and anthropogenic baseline contributions

74 Source Catchments modelling – Updated methodology and results

Constituent load validation Of the 10 modelled constituents (flow, sediment and nutrients) across the six sites, 70% were ranked as ‘satisfactory’ or worse, with the remaining 30% being rated as ‘good’ or ‘very good’. The Burdekin River at Home Hill site (Table 26) ranked as the best overall, followed by the Suttor River and Burdekin River at Selheim sites (Appendix 4). Flow and DON were the best performing parameters. TN and PP were the worst performing parameters, being rated as ‘indicative’ at each of the six sites, apart from the Suttor where PN is rated as ‘good’. The data at each of the five sites is presented in Appendix 4. Statistics could not be calculated at the daily or monthly time-step due to the load calculation method used in the GRBCLMP data.

Table 26 Moriasi et al. (2015) overall annual performance rating for each constituent at 120001A Burdekin River at Home Hill (120006B Burdekin River at Clare), n = 8 years for flow and n = 8 years for WQ)

RSR PBIAS NSE R2 Rating Rating Rating Parameter Rating (Moriasi Value Value (Moriasi Value (Moriasi Value (Moriasi 2007) 2015) 2015) 2015) Flow 0.17 Very good 0.29 Very good 0.97 Very good 0.97 Very good TSS 0.45 Very good 2.46 Very good 0.80 Good 0.80 Very good TN 0.67 Satisfactory 33.23 Indicative 0.56 Good 0.87 Very good PN 0.88 Indicative 47.64 Indicative 0.22 Indicative 0.77 Very good DIN 0.69 Satisfactory -9.19 Very good 0.53 Good 0.85 Very good DON 0.38 Very good 18.90 Good 0.85 Very good 0.96 Very good TP 0.50 Very good 23.80 Satisfactory 0.75 Very good 0.93 Very good DIP 0.46 Very good -4.13 Very good 0.79 Very good 0.83 Very good PP 0.58 Good 28.76 Satisfactory 0.67 Very good 0.93 Very good DOP 0.89 Indicative 44.41 Indicative 0.21 Indicative 0.50 Indicative

Contribution by land use Grazing (open and closed) and Conservation account for ~90% and ~5% of the Burdekin Region respectively by area and therefore area generate 99% of the fine sediment load, and over 95% of all other constituents, apart from DIN and PSII. DIN and PSII toxic loads are predominantly derived from sugarcane with sugarcane accounting for ~45 % of the modelled DIN export.

75 Source Catchments modelling – Updated methodology and results

Figure 29 Contribution to fine sediment, particulate nutrient and DIN export by land use for the BU region

Sugarcane and grazing account for <1% and 90% of the BU area respectively, and between them generate 87% of the DIN export load (Figure 30). Sugarcane generated the largest load (1,175 t/yr). Per unit area, sugar also generates the highest average annual load (11 kg/ha/yr), while grazing has the lowest (<0.1 kg/ha) (Figure 30). The large grazing load is a function of its large area.

Figure 30 DIN export load and areal rate for each land use in the BU Sources and sinks Of the total modelled fine sediment load (6,527 kt/yr) in the Burdekin region (Appendix 3, Table 74), 58% is exported to the GBR lagoon with the remaining 42% largely lost to reservoir deposition (61%),

76 Source Catchments modelling – Updated methodology and results floodplain deposition (32%), and the remainder lost to extractions (7%). No instream deposition of fine sediment was modelled due to a lack of supporting data to quantify the deposition component at the time of model development, though instream deposition and remobilisation could be enabled in future models if data to support it become available. For TN, 78% of the generated load is exported, while 64% of the generated TP load is exported. The Burdekin fine sediment source sink budget is presented diagrammatically in Figure 31.

Figure 31 Fine sediment source sink budget for the Burdekin Progress towards Reef Plan 2013 targets Management practice changes that occurred in grazing lands are presented in Table 74. Small changes were made in streambank and gully management, with a 2% shift from C to B practices occurring in hillslope targeted practices. No changes were recorded for sugarcane soil erosion (Table 27), with minor changes occurring in cropping (grains) land management for soil erosion (Table 28) where a shift from C to A class was recorded. Collectively these changes result in no load reduction reported for fine sediment from the Burdekin region, as the changes result in a less than 0.5% reduction in load.

With respect to nutrients, for sugarcane (Table 27) there was a shift from D/C classes to C/B for nutrients, while no management changes were reported from cropping (grains) (Table 28). The land management practice improvements in sugarcane resulted in a 4% reduction in DIN from the Burdekin region. This contributed to the 1% overall GBR reduction.

Improvements were recorded in both sugarcane and cropping (grains) land uses for pesticide use. In sugarcane there was a big shift (10%) out of C class practices into A and B practices, while for cropping areas there was a 4.5% shift from B to A. This resulted in a 3.5% reduction in PSII toxic equivalent loads from the Burdekin region, contributing to the 3% overall GBR reduction.

77 Source Catchments modelling – Updated methodology and results

Table 27 Burdekin grazing management change; values are percent of total

Hillslope Baseline RC2014 Change RC2015 Change A 0.4 0.7 0.3 0.7 0.0 B 28.7 29.2 0.5 31.5 2.3 C 59.1 58.6 -0.5 56.3 -2.3 D 11.8 11.6 -0.2 11.6 0.0 Gully Baseline RC2014 Change RC2015 Change A 5.5 6.0 0.5 6.0 0.0 B 20.7 20.2 -0.5 20.5 0.3 C 55.9 55.9 0 55.5 -0.3 D 18.0 18.0 0 18.0 0.0 Streambank Baseline RC2014 Change RC2015 Change A 32.6 32.6 0 32.7 0.1 B 29.0 29.1 0.1 29.2 0.1 C 13.8 13.7 -0.1 13.6 -0.1 D 10.0 9.9 -0.1 9.9 -0.1 NA 0.4 0.7 0.3 14.7 0.0

78 Source Catchments modelling – Updated methodology and results

Table 28 Burdekin sugarcane management practice change summary; values are percent of total

Baseline Soil RC2014 Change RC2015 Change 2013 A 3 3 0 3 0 B-A 2 2 0 2 0 B 11 12 1 12 0 C-B 32 32 0 32 0 C 26 25 -1 25 0 D-C 12 12 0 12 0 D 14 14 0 14 0 Baseline Nutrients RC2014 Change RC2015 Change 2013 A 3 3 0 3 0.4 B 5 7 2 10 2.9 C-B 12 11 -1 13 1.3 C 54 53 -1 50 -3.3 D-C 16 15 -1 15 -0.1 D 11 10 -1 9 -1.1 Baseline Pesticides RC2014 Change RC2015 Change 2013 A 7 7 0 8 0.6 B 19 18 -1 29 10.2 C 47 47 0 37 -10.3 Df 28 28 0 27 -0.5 Baseline Irrigation RC2014 Change RC2015 Change 2013 A 9 9 0 9 0.0 B-C 8 8 0 10 1.2 D 83 83 0 81 -1.2 Baseline Recycle Pits RC2014 Change RC2015 Change 2013 A 19 20 1 21 1.8 B 16 19 3 26 6.6 C 27 25 -2 22 -2.7 D 38 36 -2 31 -5.7

79 Source Catchments modelling – Updated methodology and results

Table 29 Burdekin grains management change; values are percent of total

Soil Baseline 2013 RC2014 Change RC2015 Change A 4 8 4 9 1.2 B 82 83 1 83 0.0 C 14 9 -5 8 -1.2 D 0 0 0 0 0.0 Nutrient Baseline 2013 RC2014 Change RC2015 Change A 5 5 0 5 0.0 B 43 43 0 43 0.0 C 52 52 0 52 0.0 D 0 0 0 0 0.0 Pesticide Baseline 2013 RC2014 Change RC2015 Change A 5 6 1 10 4.5 B 26 26 0 21 -4.5 C 69 69 0 69 0.0 D 0 0 0 0 0.0

80 Source Catchments modelling – Updated methodology and results

Mackay Whitsunday The results of the Mackay Whitsunday modelling are presented below, separated into hydrology and modelled loads. In the hydrology component, the results of the updated calibration process will be presented per gauge, as well as the impact on the total flow for the MW region. The modelled loads section includes the results of the total baseline, anthropogenic baseline and predevelopment loads. The validation of the MW Source Catchments modelled data is then presented using load estimates from measured data. Progress towards targets due to investment is reported against the anthropogenic baseline for RC2015. A summary of the total baseline load by land use, and land use by basin is also reported, as well as a mass balance summary of the sources and sinks by constituent. For a full list of the MW region loads for RC2015 refer to Appendix 3, Table 75.

Hydrology calibration performance The Proserpine River Basin produces the highest modelled mean annual flows (~35%) in the region, followed by the O’Connell River, Plane Creek, and Pioneer River Basins contributing approximately 29%, 20% and 16% for the regional flows respectively. This is consistent with RC2014. Only four gauges were recalibrated for hydrology (which are highlighted green in Table 30), and in all instances the baseflow contribution increased. However, the recalibration of hydrology to better represent the regional baseflow has reduced the average annual flow only by approximately 0.5% compared to RC2014.

Table 30 MW baseflow proportions for the current and previous Report Card

RC2014 RC2015 LH baseflow Gauge Basin baseflow (%) baseflow (%) filter (%)

122012A Proserpine <1 9 9 124001B O’Connell 23 23 20 124002A O’Connell 15 28 28 124003A O’Connell 28 28 29 125001B Pioneer 24 24 30 125002C Pioneer 21 21 22 125004B Pioneer 33 33 38 126001A Plane 6 14 14 126002A Plane 18 18 18 126003A Plane 10 11 18

Modelled Loads In Table 31 the contribution to total MW load of each constituent by basin is presented. The estimated baseline TSS load exported from MW has increased by 34% from RC2014 (611kt/yr). The O’Connell River Basin has the highest modelled mean annual TSS exports of 314 kt/yr (~38% of MW load); this is consistent with RC2014. Plane Creek Basin has the lowest modelled mean annual TSS export of 146 kt/yr, which is ~18% of the regional TSS export; in RC2014 the Proserpine Basin was the lowest exporter though only 3kt/yr was the difference between Plane Creek and Proserpine Basins. Dominance

81 Source Catchments modelling – Updated methodology and results of the modelled sediment erosion from grazing land use and stream banks in the O’Connell River Basin are the key reasons for its highest contribution to the regional TSS load.

The Plane Creek Basin has the highest proportion of sugarcane land use (~42%) when compared to other MW Basins. Consequently, the Plane Creek Basin has the highest modelled mean annual exports of PSIIs 1,758 kg/yr and DIN 464 t/yr for the MW region. This equates to 43% of the PSII load and 34% of the DIN load for the MW region. In contrast, the Proserpine Basin, with the least sugarcane land use (~15% of the MW sugarcane area) had the lowest modelled mean annual loads for total toxic load 331 kg/yr (~8% of MW load). Similarly, the modelled mean annual load for DIN is lowest with 256 kg/yr (~19% of the MW load) in the Pioneer River Basin, which has the third lowest sugarcane land use (~22% of the MW sugarcane area), next to the O’Connell River Basin’s contribution of 21% to the MW sugarcane area. Dominance of the grazing land use in the Proserpine River Basin (~29% of the MW grazing area), compared to that of the Pioneer River Basin (~14% of the MW grazing area), may have increased DIN contributions in the Proserpine River Basin, given that the RC2015 model indicated the grazing land use contributes approximately 22% of the regional DIN load.

Table 31 Contribution from MW basins to the total MW baseline load

PSII TSS TN DIN DON PN TP DIP DOP PP PSII Basin TE (kt/y) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (kg/y) (kg/yr) Proserpine 131 1,066 310 370 387 255 73 19 164 331 248

O'Connell 314 1,528 325 358 845 495 62 16 416 971 701

Pioneer 227 933 256 227 450 237 42 11 184 985 707

Plane 146 1,279 464 347 468 319 73 19 228 1,758 1269 Creek Total 818 4,806 1,355 1,301 2,150 1,306 249 65 992 4,044 2,925

Anthropogenic baseline and pre-development loads The TSS increase factor for the whole of the MW region is ~3.6; the Pioneer and O’Connell basins both have the greatest increase of ~4.3 times, while the Plane Creek Basin has a ~3.1 times increase and the Proserpine Basin has the least increase of ~2.3 times to the predevelopment load (Figure 32).

In regard to DIN, the anthropogenic increase factor was 3 for the MW region. All four basins had a significant anthropogenic contribution to DIN, primarily due to sugarcane production in the area. The Proserpine had the lowest increase factor of 2, while the Plane Creek Basin was exporting nearly 4.7 times the predevelopment load (Figure 33).

82 Source Catchments modelling – Updated methodology and results

Figure 32 TSS (kt/yr) loads for the MW basins, highlighting the predevelopment and anthropogenic baseline contributions

Figure 33 DIN (kt/yr) loads for the MW basins, highlighting the predevelopment and anthropogenic baseline contributions

83 Source Catchments modelling – Updated methodology and results

Constituent Load Validation

GBRCLMP Comparisons The modelled mean annual flows and pollutant loads are compared against the GBRCLMP load estimates in the Pioneer River at Dumbleton Pump Station (Figure 32) and Sandy Creek at Homebush (Figure 33).

Figure 34 Mean Annuals Modelled VS. GBRCLMP in the Pioneer River at Dumbleton Pump Station

Figure 35 Mean Annuals Modelled VS. GBRCLMP in Sandy Creek at Homebush

Overall the RC2015 baseline model estimated mean annual flows and loads of most pollutants fairly well at the Pioneer River and Sandy Creek End of System (EoS) sites. However, the model overestimated the mean annual DIN load at the Sandy Creek EoS site. The model also overestimated total toxic loads at both EoS sites. It should be noted that sugarcane covers ~50% and ~20% of the upstream area of the Sandy Creek and Pioneer River EoS sites respectively. It should be noted that

84 Source Catchments modelling – Updated methodology and results the modelled DIN and pesticide (PSII) loads will not match observed loads because typical rates and timing of fertiliser and pesticide applications are used in the model not actual values. Hence, any inconsistency between the amount and timing of fertilisers and pesticides applied on field and the values used in the APSIM model simulations will be reflected in outputs. Therefore the comparison between modelled and observed loads should only be seen as indicative until local data becomes available.

The performance of the RC2015 baseline model to predict the inter-annual loads across the MW region was also assessed using a number of quantitative statistics recommended in Moriasi et al. (2015). The estimated quantitative statistics and the corresponding model performance ratings are presented in Table 32 and Table 33 for the Pioneer River and Sandy Creek EoS sites. An overall performance rating was established as recommended quantitative statistics revealed conflicting/unbalanced performance ratings for most of the constituents. The overall performance was described conservatively using the least rating of the recommended statistics. For ease of presentation, ‘Very Good’ and ‘Good’ rated sites are coloured green, while ‘Satisfactory’ and ‘Indicative’ are coloured orange, indicating that further work is required to improve the representation of these constituents.

Based on the overall performance rating the water quality model at the Pioneer River EoS site achieved a ‘Very Good’ rating for annual TSS loads. The model achieved a rating of ‘Good’ for annual particulate phosphorus loads while it was ‘indicative’ only to estimate annual particulate nitrogen loads. The model performance was ‘very good’ at the Sandy Creek EoS site to estimate annual particulate phosphorus while it was ‘indicative’ to estimate annual TSS and particulate nitrogen loads. The RC2015 baseline model was ‘Indicative’ to estimate DIN loads at both EoS sites.

Table 32 Moriasi et al. (2007) annual statistics for GBCLMP versus modelled data for the Pioneer River at Dumbleton Pump Station (125013A) n = 8 year for flow data, n = 8 years for WQ data

RSR PBIAS NSE R2 Rating Constituent Rating Rating Rating Value Value Value Value (Moriasi (Moriasi 2007) (Moriasi 2015) (Moriasi 2015) 2015) Flow 0.2 Very good 1.4 Very good 1.0 Very good 1.0 Very good TSS 0.4 Very good -1.2 Very good 0.9 Very good 1.0 Very good TN 0.6 Satisfactory 26.1 Satisfactory 0.6 Good 0.9 Very good PN 0.8 Indicative 36.8 Indicative 0.4 Satisfactory 1.0 Very good DIN 0.9 Indicative -3.4 Very good 0.3 Indicative 0.3 Satisfactory DON 0.5 Good 21.0 Satisfactory 0.7 Very good 0.9 Very good TP 0.5 Good 16.5 Good 0.7 Very good 1.0 Very good DIP 0.5 Very good 3.3 Very good 0.8 Very good 0.8 Very good PP 0.5 Good 13.9 Very good 0.7 Very good 1.0 Very good DOP 0.7 Indicative 44.3 Indicative 0.3 Indicative 0.9 Very good

85 Source Catchments modelling – Updated methodology and results

Table 33 Moriasi et al. (2007) annual statistics for GBCLMP versus modelled data for Sandy Creek at Homebush (126001A) n = 5 year for flow data, n = 5 years for WQ data

RSR PBIAS NSE R2 Rating Rating Constituent Rating Rating Value (Moriasi Value Value Value (Moriasi (Moriasi 2015) (Moriasi 2015) 2007) 2015) Flow 0.1 Very good 3.3 Very good 1.0 Very good 1.0 Very good TSS 0.7 Indicative -56.1 Indicative 0.5 Satisfactory 0.8 Very good TN 0.3 Very good -16.7 Good 0.9 Very good 0.9 Very good PN 0.2 Very good 2.3 Very good 1.0 Very good 0.9 Very good DIN 1.6 Indicative -102.9 Indicative -1.5 Indicative 0.2 Indicative DON 0.2 Very good 2.5 Very good 1.0 Very good 0.9 Very good TP 0.3 Very good -16.4 Good 0.9 Very good 1.0 Very good DIP 0.2 Very good 15.8 Good 1.0 Very good 1.0 Very good PP 0.5 Very good -34.9 Indicative 0.8 Very good 0.9 Very good DOP 0.3 Very good -3.0 Very good 0.9 Very good 0.9 Very good

FRCE Comparisons The following plots compare the long-term annual TSS (Figure 36) and DIN (Figure 37) loads estimated by the model against that estimated by Joo et al. (2014). There is good agreement between the FRCE and modelled loads for TSS across the model period. Long-term comparisons for DIN also showed good agreement though a larger range of under/over prediction compared to TSS. For both TSS and DIN it is evident from the graphs that for the period 2009-2014 the model has improved its predictive capability and modelled loads are more in line with measured loads. This is given the availability of more data against which to calibrate, and validate, the models. Plots comparing the annual loads of the rest of the pollutants estimated by the model and Joo et al. (2014) are presented in Appendix 4.

Figure 36 Long-term annual TSS Loads Modelled VS. FRCE in the Pioneer River at Dumbleton Pump Station

86 Source Catchments modelling – Updated methodology and results

Figure 37 Long-term annual DIN Loads Modelled VS. FRCE in the Pioneer River at Dumbleton Pump Station

P2R Comparisons The performance of the RC2015 baseline model to predict sub-catchment scale water quality was assessed by comparing the modelled flows and pollutant loads leaving the Multifarm subcatchment against those estimated as part of the P2R monitoring program (Joo et al. 2012). The multi-farm sub- catchment is located in the Plane Creek basin, approximately 9 km north-west of the Sandy Creek at Homebush gauging station. The sub-catchment covers approximately 3000 hectares and contains 48 farms (some whole and others part farms). The Multifarm subcatchment is dominated by sugarcane (~78% sugarcane, 14% grazing, 6% other and 2% nature conservation and forestry) and thus comparison of modelled and measured flows and pollutant loads is critical to assess the model performance to represent sugarcane related water quantity and quality processes. The modelled annual flows and pollutant loads at the Multifarm subcatchment were modified to compare against those measured by the P2R program. This was done to rectify the irregularities observed between the modelled and measured sub-catchment area and discharge characteristics of the subcatchment. Detailed descriptions of the modifications to the modelled and measured flows and loads are summarised in Appendix 1.

In Figure 38 the modelled mean annual flows and pollutant loads during the low and medium flow events are compared against those of the P2R estimates at the Multifarm site. Overall the RC2015 baseline model estimated mean annual flows and loads of most pollutants well at the Multifarm site during low and medium flow events.

87 Source Catchments modelling – Updated methodology and results

Figure 38 Mean annual Source Catchments modelled and P2R at Multifarm during low and medium flow events

Streambank Erosion Model Validation A recent stream stability assessment (Alluvium, 2014) estimated sediment export from major erosion sites due to streambank erosion in the following five reaches of the O’Connell River system:  Reach 1- Extends from Dingo Creek confluence to the sea.  Reach 2 - Extends from a point two hundred metres downstream of the Boundary Creek confluence to the Dingo Creek confluence.  Reach 3 –Extends from Horse Creek confluence to two hundred metres downstream of the Boundary Creek confluence.  Reach 4 –Extends from Cathu-O’Connell River Road to Horse Creek confluence.  Reach 5 –Extends from the headwaters to Cath-O’Connell River Road.

The volume of sediment eroded between 2010 and 2014 was estimated using digital elevation models (DEM) derived from LiDAR data sets acquired in those years. It should be noted here that the sediment export estimated using temporal analysis of LiDAR derived DEMs between 2010 and 2014 is likely to include erosion of both fine and coarse sediment fractions. Therefore, sediment export estimated as part of the stream stability assessment were halved to assess the performance of the RC2015 baseline model to estimate only the fine sediment export from streambank erosion along the O’Connell River between 2010 and 2014 (Table 34).

The comparison is encouraging given the modelled streambank erosion being within 11, 20, and 33 percent of the Alluvium estimates for reaches 1, 3, and 4-5 respectively. Overall the RC2015 model has underestimated streambank erosion only by approximately 16% in the O’Connell River. The model under-predicts streambank erosion at three of four reaches but a maximum of 33% (Table 34).

88 Source Catchments modelling – Updated methodology and results

Table 34 Modelled Fine Sediment export in the O’Connell River against LiDAR estimates for fine sediment export between 2010 and 2014

Stream Stability Corresponding Assessment Modelled Fine River Reaches Catchments in the Estimates for Fine Sediment Export (m³) Model Sediment Export (m³) Reach 1 SC #42,SC #43,SC #45 85,000 68,146 Reach 2 SC #148 2,500 8,112 Reach 3 SC #46 25,000 16,534 Reach 4 & Reach 5 SC #47 28,500 25,452

Contribution by land use Sugarcane and grazing (closed and open) account for approximately 19% and 44% of the MW area respectively, and between them generate approximately 86% of the DIN export load (Figure 40). Grazing and sugarcane contribute approximately 75% of the regional TSS and particulate nitrogen and phosphorus loads (Figure 39). Sugarcane land use produces most of the PSIIs (and thus total toxic) loads and the largest DIN load (~67%) in the region.

Figure 39 Constituent contribution (%) by land use for four constituents

Sugarcane generated the largest average annual DIN load (~ 885 t/yr). Per unit area, sugarcane generate the highest average annual load (~ 5.3 kg/ha/yr), followed by cropping (~ 2.5 kg/ha/yr) (Figure 40). The average DIN per unit area rate for all land uses in the MW is approximately 1.8 kg/ha/yr.

89 Source Catchments modelling – Updated methodology and results

Figure 40 DIN export load and areal rate for each land use in the MW

Sources and sinks The fine sediment source (hillslope/gully/streambank) and sink (extractions/deposition) budget for MW is presented in Figure 41. Hillslope erosion is the dominant erosion supply (~54% of total) generated from grazing, sugarcane, and cropping land uses. Streambank erosion is the other dominant supply (~31%). Erosion from all other land use types (e.g. nature conservation, forestry, urban, etc.) constitutes ~14% of the regional TSS supply. Gully erosion contribution for the regional TSS supply is low (~1%). RC2015 baseline model indicates that approximately 96% of the regional TSS supply is exported, while only ~4% is lost to reservoir deposition, extraction and floodplain deposition. Reservoirs in the region trap approximately 50% of the total TSS load that is lost, while extraction and flood plain deposition account for 32% and 19% of the total TSS loss respectively.

90 Source Catchments modelling – Updated methodology and results

Figure 41 Fine sediment source sink budget diagram for MW Progress towards Reef Plan 2013 targets Overall there has been only marginal progress made towards the reef targets for the regional TSS, DIN and total toxic loads following the RC2015 management practice investments and associated improvements (Table 36 and Table 35). RC2015 models estimated < 0.1% reduction in TSS and particulate nutrient loads due to 252ha of hillslope, 252ha of gully, 252ha of riparian, and approximately 1.7 km of streambank improvement projects around grazing management. Approximate reductions of 1.1% and 2.9% were estimated for DIN and total toxic loads.

Table 35 MW grazing management practice change summary; all values are in percent

Hillslope Baseline RC2014 Change RC2015 Change A 8.00 8.01 0.01 8.01 0.00 B 30.00 30.75 0.75 31.02 0.27 C 25.00 24.61 -0.39 24.44 -0.17 D 37.00 36.63 -0.37 36.53 -0.10 Gully Baseline RC2014 Change RC2015 Change A 0.02 0.02 0.00 0.03 0.01 B 37.00 37.00 0.00 37.08 0.07 C 38.00 38.00 0.00 37.97 -0.02 D 24.00 24.00 0.00 23.94 -0.06 Streambank Baseline RC2014 Change RC2015 Change A 0.96 0.96 0.00 1.02 0.05 B 17.00 17.10 0.10 17.23 0.13 C 46.00 45.94 -0.06 45.78 -0.15 D 35.00 34.96 -0.04 34.93 -0.03

91 Source Catchments modelling – Updated methodology and results

Table 36 MW sugarcane management practice change summary; all values are in percent

Baseline Soil RC2014 Change RC2015 Change 2013 A 1.00 1.00 0.00 1.00 0.00 B-A 3.21 3.24 0.02 3.24 0.02 B 36.79 36.76 -0.02 36.88 0.10 C-B 14.00 14.00 0.00 13.75 -0.25 C 17.00 17.00 0.00 18.97 1.97 D-C 8.96 9.01 0.05 7.18 -1.78 D 19.04 18.99 -0.05 18.99 -0.05 Baseline Nutrients RC2014 Change RC2015 Change 2013 A 0.00 0.00 0.00 0.00 0.00 B 19.45 19.89 0.44 20.88 1.43 C-B 28.50 28.34 -0.17 30.65 2.14 C 30.29 30.12 -0.17 29.66 -0.63 D-C 19.68 19.57 -0.11 19.68 0.01 D 2.09 2.09 0.00 -0.86 -2.95 Baseline Pesticides RC2014 Change RC2015 Change 2013 A 4.13 4.13 0.00 4.13 0.00 B 32.48 32.48 0.00 36.50 4.02 C 30.81 30.81 0.00 27.55 -3.26 Df 32.58 32.58 0.00 31.82 -0.76 Baseline Irrigation RC2014 Change RC2015 Change 2013 A 4.60 4.60 0.00 4.60 0.00 B-C 14.00 14.07 0.07 14.07 0.07 D 81.40 81.33 -0.07 81.33 -0.07 Recycle Baseline RC2014 Change RC2015 Change Pits 2013 A 3 3 0 3 0 B 19 19 0 19 0 C 46 46 0 46 0 D 31 31 0 31 0

92 Source Catchments modelling – Updated methodology and results

Fitzroy The results of the Fitzroy modelling are presented below, separated into hydrology and modelled loads. In the hydrology component, the results of the updated calibration process will be presented per gauge, as well as the impact on the total flow for the Fitzroy region. The modelled loads section includes the results of the total baseline, anthropogenic baseline and predevelopment loads. The validation of the Fitzroy Source Catchments modelled data is then presented using load estimates from measured data. Progress towards targets due to investment is reported against the anthropogenic baseline for RC2015. A summary of the total baseline load by land use, and land use by basin is also reported, as well as a mass balance summary of the sources and sinks by constituent. For a full list of the Fitzroy region loads for RC2015 refer Appendix 3, Table 77.

Hydrology calibration performance The hydrology in the Fitzroy region was recalibrated to improve the baseflow proportion. Of the 44 gauges used for the previous calibration, 39 were recalibrated. Gauges in the Calliope and Boyne basins were not recalibrated as the base flow proportions in these basins were considered satisfactory.

Recalibrating the hydrology in the Fitzroy region to improve the baseflow proportion has resulted in the total annual flow for the Fitzroy region decreasing by 1% from RC2014, a reduction of approximately 50,000 ML/yr. This resulted in a total average annual flow for this report card of ~9,200,000 ML/yr. There was a decrease of approximately 90,000 ML/yr for the Fitzroy Basin and a ~20,000 ML/yr increase in the Styx and Shoalwater basins. There was little change (<1%) for flow in the remainder of basins. The difference between the estimated model flow and the observed flow at gauges for RC2014 and this RC2015 are shown in Table 37 for a number of key gauges that were recalibrated.

Table 37 Modelled flow differences for RC2014 and RC2015 compared to observed flows

RC2014 RC2015 Flow Flow Gauge Gauge Name Difference Difference (%) (%) 129001A Waterpark Creek at Byfield 0 0 130005A Fitzroy River at The Gap 1 -7 Theresa Creek at Gregory 130206A 9 0 Highway 130322A Dawson River at Beckers 0 0 130401A Isaac River at Yatton 0 2 130506A Comet River at The Lake 5 -4

An assessment of the models predicted baseflow compared to the Lyne-Hollick baseflow filter applied to observed data shows that the average difference between observed and modelled baseflow for all calibrated gauges improved from a 29% difference for RC2014 to 2% for RC2015. In the previous report card 15 gauges had a difference between the modelled base flow and the Lyne-Hollick calculated base flow of greater than 50%, for this report card no gauge had a difference greater than 10%. In Table 38

93 Source Catchments modelling – Updated methodology and results a sample of the base flow percentages for the recalibrated gauges and the improvement in estimated base flow volumes that occurred is shown.

Table 38 Base flow proportions (%) for RC2014 and RC2015 compared to observed estimates

RC2014 RC2015 LH Gauge Gauge Name baseflow baseflow baseflow (%) (%) filter (%) 129001A Waterpark Creek at Byfield 47 29 27 130005A Fitzroy River at The Gap 67 31 26 Theresa Creek at Gregory 130206A 34 10 8 Highway 130322A Dawson River at Beckers 89 24 19 130401A Isaac River at Yatton 99 23 18 130506A Comet River at The Lake 98 23 20

Modelled Loads The Fitzroy region is the second largest contributor of fine sediment (18%) to the GBR behind the BU and is the highest contributor of DIP (49%) and PP (33%). It is also second largest contributor for PN (26%). In terms of total PSII load, the Fitzroy is the fourth highest contributor (18%), but when this is converted to diuron toxic equivalent load the contribution is less than 1% due to the low amount of diuron used in the region. There has been a 2% increase in TSS loads from RC2014

The Fitzroy Basin occupies the largest area in the Fitzroy region at 142 000 km2 or 91% of the area and as a result, it produces the largest load for all constituents. It contributes 83% of the TSS load (1,507 kt/yr), 58% of the TN load (6,276 t/yr), and 59% (2,750 t/yr) of the TP load (Table 39). Both the Styx and Shoalwater contribute significant nutrient loads for their size, in part due to their proximity to the coast. The Styx at 1.9% of the region area contributes 11% (1,210 t/yr) and 12% (541 t/yr) of the TN and TP loads respectively, while the slightly larger Shoalwater Creek Basin (2.3% of the total area) contributes 9% for both TN and TP. There has been a significant increase in PN (55%), with decreases in DIN and DON. PP has decreased by 37%, as has DIP (11%) while DOP increased. PSII TE has remained about the same.

94 Source Catchments modelling – Updated methodology and results

Table 39 Contribution from Fitzroy basins to the total Fitzroy baseline load

TSS TN DIN DON PN TP DIP DOP PP PSII PSII TE Basin (kt/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (kg/yr) (kg/yr) Styx 104 1,210 91 293 829 541 91 22 428 203 4 Shoalwater 67 984 100 259 646 408 80 20 308 107 2 Creek Water Park 65 1,478 65 161 1,268 579 50 12 517 15 0 Creek Fitzroy 1,507 6,276 799 2,425 3,066 2,750 743 185 1,822 1,749 38

Calliope 57 638 47 152 439 281 47 12 221 97 2

Boyne 24 264 37 116 113 105 36 9 60 39 1

Total 1,824 10,851 1,140 3,406 6,361 4,665 1,047 260 3,357 2,210 47

Constituent load validation Three data sources were used to assess model performance. The key validation dataset used to validate the model were the GBRCLMP 2006–14 monitoring program established by the Qld State Government (Turner et al. 2012). Further validation was conducted against the long-term (1986 – 2009) loads report (Joo et al. 2014). The third source was event sampling data collected throughout the Fitzroy Basin via a number of short term monitoring programs.

A comparison was made between the mean GBRCLMP flow and loads (averaged over eight years, 2006–2014) and the modelled flow and load for the same location and the same time period. Due to limited availability of measured data the comparison for Theresa Creek only covers two years and for the Dawson River only three years. The modelled fine sediment load matches well at the Fitzroy River at Rockhampton site, with a slight over-prediction (~3%), and also matches well at the Comet River (10%) (Figure 42). Fine sediment is over-predicted at the Dawson River and Theresa Creek sites. For PN, the model is under-predicting at three of four sites (Figure 43), only over-predicting at Theresa Creek. PP is under-predicted at the Fitzroy River at Rockhampton and Comet River at Comet Weir sites, and over-predicted at the other sites (Figure 46).

95 Source Catchments modelling – Updated methodology and results

Figure 42 Fine sediment average annual modelled (Source Catchments) versus measured load comparison for 2006-2014* period, for the four monitoring sites in the Fitzroy region. *Data for Theresa Creek is only for two years, and for the Dawson River three years.

Figure 43 Particulate nitrogen average annual modelled versus measured load comparison for 2006-2014* period, for the four monitoring sites in the Fitzroy region. *Data for Theresa Creek is only for two years, and for the Dawson River three years.

96 Source Catchments modelling – Updated methodology and results

Figure 44 Particulate phosphorus average annual modelled versus measured load comparison for 2006-2014* period, for the four monitoring sites in the Fitzroy region. *Data for Theresa Creek is only for two years, and for the Dawson River three years.

A number of quantitative statistics, recommended in Moriasi et al. (2015) were calculated using the measured (GBRCLMP data (2006-2014)) and modelled (RC2015 baseline) loads across the Fitzroy region. The estimated quantitative statistics and the corresponding model performance ratings are presented in Table 40 and Table 41 for the Fitzroy and Comet Rivers, respectively. An overall performance rating was established as recommended quantitative statistics revealed conflicting/unbalanced performance ratings for most of the constituents. The overall performance was described conservatively using the least rating of the recommended statistics. For ease of presentation, ‘very good’ and ‘good’ rated sites are coloured green, while ‘satisfactory’ and ‘indicative’ are coloured orange, signalling that further work is required to improve the representation of these constituents.

At the Fitzroy River site, all four measures yield similar results for all constituents, typically with ‘very good’ rankings. This represent an extremely good result overall and indicates that the baseline model is performing well predicting loads for all constituents at the end of system site. Only PN had two ‘satisfactory’ ratings for the RSR and R2 measures. The PBIAS measure shows the model results were very good results with a bias within 8% for all constituents but PN. Statistics could not be calculated at the daily or monthly time-step due to the load calculation method used in the GRBCLMP data. At the Comet River site, the ratings for RSR, NSE and R2 were generally consistent, with largely very good ratings for all constituents except for TSS which was rated as indicative. With respect to the PBIAS measures at the Comet River site, TSS, TN, DIN, DON, PP and DOP are all rated good or better, while flow, PN, TP and DIP can be considered indicative at best.

97 Source Catchments modelling – Updated methodology and results

Table 40 RC2015 (1300000 Fitzroy River at Rockhampton, n = 7 years for flow and n = 7 years for WQ)

RSR PBIAS NSE R2 Parameter Rating Rating Rating Rating Value (Moriasi Value (Moriasi Value (Moriasi Value (Moriasi 2007) 2015) 2015) 2015) Flow 0.2 Very good -0.3 Very good 1.0 Very good 1.0 Very good TSS 0.3 Very good -2.7 Very good 0.9 Very good 0.9 Very good TN 0.3 Very good 6.1 Very good 0.9 Very good 1.0 Very good PN 0.7 Satisfactory 13.1 Very good 0.5 Good 0.6 Satisfactory DIN 0.3 Very good -3.9 Very good 0.9 Very good 0.9 Very good DON 0.2 Very good 0.9 Very good 1.0 Very good 1.0 Very good TP 0.4 Very good 4.0 Very good 0.8 Very good 0.9 Very good DIP 0.3 Very good 0.3 Very good 0.9 Very good 1.0 Very good PP 0.6 Good 8.0 Very good 0.7 Very good 0.7 Very good DOP 0.5 Very good 1.2 Very good 0.8 Very good 0.9 Very good

98 Source Catchments modelling – Updated methodology and results

Table 41 RC2015 (130504B Comet River at Comet Weir, n = 5 years for flow and n = 5 years for WQ)

RSR PBIAS NSE R2 Parameter Rating Rating Rating Rating Value (Moriasi Value (Moriasi Value (Moriasi Value (Moriasi 2007) 2015) 2015) 2015) Flow 0.2 Very good 14.4 Satisfactory 1.0 Very good 1.0 Very good TSS 0.8 Indicative -10.3 Good 0.4 Indicative 0.4 Indicative TN 0.2 Very good 11.9 Very good 1.0 Very good 0.9 Very good PN 0.3 Very good 21.2 Satisfactory 0.9 Very good 0.9 Very good DIN 0.2 Very good 14.4 Very good 1.0 Very good 0.9 Very good DON 0.2 Very good 14.1 Very good 1.0 Very good 1.0 Very good TP 0.4 Very good 25.1 Satisfactory 0.9 Very good 0.9 Very good DIP 0.5 Good 49.5 Indicative 0.7 Very good 1.0 Very good PP 0.4 Very good 15.6 Good 0.9 Very good 0.8 Very good DOP 0.6 Good 0.0 Very good 0.7 Very good 0.7 Good

Contribution by land use The major contributors to the TSS load are open and forested grazing (68%) (Figure 45). Dryland cropping’s contribution to the export load increased from 3% in RC2014 to 15% in RC2015 due to the cropping model simulation now accounting for slope, this equates to an increase in the generated areal rate of 0.2 t/ha/yr for RC2014 to 2.0 t/ha/yr for this report card which is a better estimate of typical erosion rates in cropping areas of Central Queensland (Thornton et al. 2007, Waters and Packett 2007).

The particulate nutrient contribution is dominated by open and forested grazing, and nature conservation. The contribution from nature conservation has increased significantly for this report card and is the second highest contributor at ~28% behind forested grazing (~35%). Together the three land uses contribute approximately 85% of the export load for both particulate nitrogen and phosphorus.

All dissolved nutrient loads are dominated by contributions from open and forested grazing, with ~75% of the total export load coming from these two land uses. Conservation and forestry essentially supply the remaining load for DIN, DON, DIP and DOP however dryland cropping does make a small contribution to the DIN and DON export loads.

99 Source Catchments modelling – Updated methodology and results

Figure 45 Constituent contribution (%) by land use for four constituents

Sources and sinks The total fine sediment load generated for the Fitzroy region is 8,650 kt/y. Of this load 52% is generated by gully erosion, 29% by hillslope erosion and 19% by streambank erosion. As a result of losses throughout the system only 21% of the sediment generated throughout the region is delivered to the GBR lagoon. This results in an average annual load of 1,824 kt/y of fine sediment being exported to the GBR lagoon. Sediment is lost throughout the system to floodplain deposition (61%), reservoir deposition (19%) and the remainder is lost to extractions. The Fitzroy source sink budget for fine sediment is presented in Figure 46; while the source sink table for all constituents can be found in Appendix 3, Table 78.

A higher proportion of generated nutrient load is exported to the GBR (than for sediment) with 44% of the generated nitrogen load and 37% of the generated phosphorus load being transported to the GBR lagoon. Hillslope erosion is the largest contributor to both these loads through the generation of particulate nutrients. The total nitrogen load exported to the GBR is 10,957 t/y. Particulate nitrogen contributes almost three-quarters of the total generated nitrogen load. A significant portion of the PN is lost throughout the system through deposition and as a result only contributes 59% to the exported load. DON makes up 31% of the exported load and DIN 10%. Particulate phosphorus contributes 87% to the total generated phosphorus load. A significant proportion of PP is lost throughout the system to floodplain and reservoir deposition, however it still contributes the largest proportion to the total exported phosphorus load at 72%. Minor amounts of DIP and DOP are lost throughout the system and they contribute 23% and 6% respectively to the exported phosphorous load. The total phosphorous load exported to the GBR is 1,689 t/yr.

100 Source Catchments modelling – Updated methodology and results

Figure 46 Fine sediment source sink budget for the Fitzroy Progress towards Reef Plan 2013 targets Management changes reported for the Fitzroy region for RC2015 included riparian fencing, gully erosion and cropping management practices. No changes were reported for hillslope management in grazing land uses. The modelling results indicate that the management changes resulted in a 15 kt/y reduction in the exported sediment load or a 1% reduction of the anthropogenic load. A reduction of 0.2% for the total phosphorous load occurred which was solely the result of a 5 t/y reduction in the particulate phosphorous load. Total nitrogen also had a 0.2% reduction of the anthropogenic load which was a result of an 11 t/y reduction of the particulate nitrogen load. The toxic load reduced by 0.3% predominantly due to a 3.75 kg/y reduction in the atrazine load.

The majority (78%) of the streambank protection investment that was reported for the GBR occurred in the Fitzroy Basin. A total of 399 km of streambank fencing was constructed in the Fitzroy region. This represents an improvement in streambank management for 1% of the 40,970 km of streambank modelled for the Fitzroy region. The streambank management changes resulted in the greatest reduction of fine sediment exported to the reef from the Fitzroy region and reduced the load by 13 kt/yr. The streambank fencing resulted primarily in a shift from C and D class management to B and some A class management. There was also a small shift from B class to A class management.

Sixty-six kilometres of streambank fencing undertaken within the Styx and Shoalwater coastal catchments was not modelled this year. Due to the flat terrain in these coastal catchments, little or no streambanks were defined from the DEM. Therefore investment in streambank protection in these areas cannot be modelled. This issue will be investigated further to ensure that investment that occurs along coastal waterways can be included next year.

101 Source Catchments modelling – Updated methodology and results

Limited gully management change occurred in the Fitzroy region for this Report Card. A total of 533 ha changed from C management to B class management. This change in management represented only 1% of the total gully management change that occurred for the entire GBR. Within the Fitzroy region, the modelled gully erosion within the project area that the management change occurred in, represents only 0.01% of the total Fitzroy modelled gullied area. The management change produced a 19t/yr reduction in the generated load and negligible change to the export load.

Table 42 Fitzroy Grazing Management Change

Hillslope Baseline RC2014 Change RC2015 Change A 7.9 7.9 0.01 7.9 0.0 B 13.3 13.9 0.66 13.9 0.0 C 49.2 48.6 -0.59 48.6 0.0 D 29.6 29.6 -0.08 29.6 0.0 Gully Baseline RC2014 Change RC2015 Change A 3.9 4.0 0.06 4.0 0.0 B 16.2 16.3 0.05 16.3 0.0 C 54.6 54.7 0.02 54.6 0.0 D 25.2 25.1 -0.14 25.1 0.0 Streambank Baseline RC2014 Change RC2015 Change A 18.0 18.1 0.14 18.6 0.5 B 17.3 17.3 0.00 17.6 0.3 C 16.7 16.9 0.26 16.5 -0.4 D 35.6 35.2 -0.39 34.8 -0.4 NA 12.5 12.5 0.00 12.5 0.0

Within the Fitzroy region, improvements in cropping management practices occurred for nutrient and soil management. No changes were reported for pesticide specific management change, however as a result of other management changes and the combination of scenarios modelled in HowLeaky a reduction in the pesticide load occurred. All cropping nutrient management change for the GBR occurred in the Fitzroy region with 0.6% (5,540 ha) of dryland cropping transferred from C class to A class management. Three-quarters of all soil management change occurred in the Fitzroy, with 17,536 ha of cropping that was in D class management transferred to A, B and C class management with slightly more shifting to B class than the other two classes.

The cropping management changes produced a reduction of 23,723 t/yr in the fine sediment generation rate which translated to a 2,160 t/yr reduction in the fine sediment export load. The Atrazine load was reduced by 3.75 kg/yr.

102 Source Catchments modelling – Updated methodology and results

Table 43 Fitzroy Grains Management Change (% change)

Soil Baseline RC2014 Change RC2015 Change A 12.74 12.74 0.00 13.30 0.56 B 29.13 29.57 0.44 29.98 0.41 C 56.14 55.79 -0.35 55.45 -0.35 D 1.99 1.90 -0.09 1.27 -0.63 Nutrient Baseline RC2014 Change RC2015 Change A 1.28 1.28 0.00 1.28 0.00 B 52.26 52.26 0.00 52.86 0.61 C 40.11 40.11 0.00 39.50 -0.61 D 6.35 6.35 0.00 6.35 0.00 Pesticide Baseline RC2014 Change RC2015 Change A 2.90 2.90 0.0 2.90 0.0 B 67.13 67.36 0.2 67.36 0.0 C 28.52 28.32 -0.2 28.32 0.0 D 1.45 1.42 0.0 1.42 0.0

103 Source Catchments modelling – Updated methodology and results

Burnett Mary The results of the Burnett Mary modelling are presented below, separated into hydrology and modelled loads. In the hydrology component, the results of the updated calibration process will be presented per gauge, as well as the impact on the total flow for the BM region. The modelled loads section includes the results of the total baseline, anthropogenic baseline and predevelopment loads. The validation of the BM Source Catchments modelled data is then presented using load estimates from measured data. Progress towards targets due to investment is reported against the anthropogenic baseline for RC2015. A summary of the total baseline load by land use, and land use by basin is also reported, as well as a mass balance summary of the sources and sinks by constituent. For a full list of the BM region loads for RC2015 refer Appendix 3, Table 79.

Hydrology calibration performance All gauges in the BM region were recalibrated to improve the baseflow proportion, with six examples presented in Table 44. Baseflow was reduced in 76% (22 out of 29 sites) of the gauges as a result of the recalibration process. In terms of mean annual flow, the impact of recalibration caused small decreases of less than 5% in average annual flow in the Baffle, Kolan and Burnett basins. There was a 1% increase in average annual flow in the Burnett Basin and a 35% increase in average annual flow in the Mary River Basin as a result of the recalibration. Overall, there was a regional increase in average annual flow of 13%.

Table 44 Burnett Mary baseflow proportions for six gauges the current and previous Report Card

RC2014 RC2015 LH baseflow Gauge Basin baseflow (%) baseflow (%) filter (%) 136002D Burnett River at Mount Lawless 78 23 21 136094A Burnett River at Jones Weir Tailwater 86 21 18 136209A Barker Creek at Glenmore 96 23 18 136306A Cadarga Creek at Brovinia Station 56 12 8 137003A Elliott River at Dr Mays Crossing 81 24 29 137101A Gregory River at Isis Highway 88 15 12

The hydrology recalibration produced similar agreement with gauged flow data compared to RC2014. Out of the 29 gauges used in the recalibration, (52%) met the daily Nash-Sutcliffe criterion (NSE >0.75), compared to 60% in RC2014; 97% met the total volume error criterion (PBIAS of ± 20%) compared to 100% in RC2014. The improved estimate of baseflow estimates due to the recalibration outweighs these minor compromises in the RC2015 calibration results compared to RC2014. Overall this can be regarded as an extremely good calibration result given that a global review of hydrology calibrations rated Nash Sutcliffe E values > 0.75 as “very good” performance (Moriasi et al. 2007).

Modelled Loads Despite being the third largest GBR NRM region (13% of total GBR area), the BM region exports the second lowest TSS, TP, TN and PSII inhibiting herbicides of the six regions draining to the GBR lagoon. The relatively low contributions in TP, PP, TN and PN export are attributed to the BM having the lowest

104 Source Catchments modelling – Updated methodology and results annual runoff and a high degree of constituent trapping in the 11 storages included in the model. The model estimated 21% of TSS and particulate nutrients generated, were trapped in storages. The contribution from each of the basins in the BM, to the total load exported from the BM region are presented in Table 45.

Within the BM region, the Mary Basin is the greatest contributor to export of TSS and particulate nutrients mainly due to the high modelled streambank erosion in this catchment. The Mary also exports the largest proportions of the other constituents with the exception of PSII inhibiting herbicides for which the Burnett Basin exports the largest load.

Table 45 Contribution from Burnett Mary basins to the total Burnett Mary baseline load

PSII TSS TN DIN DON PN TP DIP DOP PP PSII Basin TE (kt/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (t/yr) (kg/y) (kg/yr) Baffle 75 546 58 220 268 148 14 9 125 8 3 Creek

Kolan 40 312 78 133 101 48 7 4 37 70 31

Burnett 548 1,407 246 494 667 317 32 17 268 119 22

Burrum 26 420 199 163 58 44 13 6 25 81 16

Mary 770 4,460 459 1,102 2,899 1,084 72 41 972 68 14

Total 1,459 7,146 1,040 2,113 3,993 1,642 139 77 1,426 346 86

Anthropogenic baseline and predevelopment loads There has been a fivefold increase in TSS and DIN loads and a threefold increase in particulate nutrient loads from predevelopment scenario (Figure 47). The greatest anthropogenic load is from the Mary Basin (six-times the predevelopment load), followed by the Burnett Basin (three-times the predevelopment load). The anthropogenic DIN loads are presented in Figure 48. While the overall BM anthropogenic contribution is five-times the predevelopment load, in the Burrum Basin it is 14-times the predevelopment load. The predevelopment and anthropogenic baseline loads for all constituents in each of the five BM basins are presented in Appendix 3, Table 79. The DIN and PSII anthropogenic baseline exports are associated with the prevalence of sugarcane land use in each catchment in the region.

105 Source Catchments modelling – Updated methodology and results

Figure 47 TSS (kt/yr) loads for the BM basins, highlighting the predevelopment and anthropogenic baseline contributions

Figure 48 DIN (t/yr) loads for the BM basins, highlighting the predevelopment and anthropogenic baseline contributions Constituent load validation The constituent loads for the BM region were validated against load estimates from catchment monitoring at four sites in the Burnett Basin and two sites in the Mary Basin. Generally, loads estimated

106 Source Catchments modelling – Updated methodology and results by the RC2015 model more closely match loads estimated from catchment loads monitoring than loads modelled in the previous report card. Modelled and monitored loads at the Burnett Basin end of system (EOS) site (Burnett River at Ben Anderson Barrage) are compared in Figure 49. Five of the 10 modelled constituent loads were within 30% of the monitored load (Figure 28). In contrast, only one constituent was within 30% of the monitored load in RC2014. Overestimation of flow by the model at this site is attributed to the fact that monitored flow at this site has been estimated by summing up flows from upstream gauging stations that the sum of catchment areas of is less than the catchment area of the Burnett EOS site.

Figure 49 Comparison of monitored and modelled mean annual flow and constituent loads at Burnett EOS site (136014A)

The performance of the RC2015 baseline model was also assessed using a number of quantitative statistics recommended in Moriasi et al. (2015). The estimated quantitative statistics and the corresponding model performance ratings are presented in Table 46 for the Burnett River. An overall performance rating was established as recommended quantitative statistics revealed conflicting/unbalanced performance ratings for most of the constituents. The overall performance was described conservatively using the least rating of the recommended statistics. For ease of presentation, ‘very good’ and ‘good’ rated sites are coloured green, while ‘satisfactory’ and ‘indicative’ are coloured orange, signalling that further work is required to improve the representation of these constituents.

Over 90% of the rating achieved a ‘good’ or ‘very good’ ranking for RSR, R2 and NSE for all constituents. For PBIAS, about half of the constituents achieved a ‘good’ or ‘very good’ ranking, while flow, TSS, PN, DIN, DIP and DOP were indicative or ‘satisfactory’. Statistics could not be calculated at the daily or monthly time-step due to the fact that Beale Ratio which was one of the methods used to calculate loads from the GRBCLMP data needed at least three data points, which were not available in most cases.

107 Source Catchments modelling – Updated methodology and results

Table 46 Moriasi et al. (2015) overall annual performance rating for RC2015 (136014A Burnett River at Ben Anderson Barrage (HW), n = 8 years for flow and n = 5 years for WQ)

RSR PBIAS NSE R2 Rating Rating Rating Rating Value (Moriasi Value (Moriasi Value Value (Moriasi 2015) (Moriasi 2015) Parameter 2007) 2015) Flow 0.1 Very good -15.4 Indicative 1.0 Very good 1.0 Very good TSS 0.9 Indicative -89.1 Indicative 0.1 Indicative 1.0 Very good TN 0.3 Very good 15.5 Good 0.9 Very good 0.9 Very good PN 0.5 Very good 26.7 Satisfactory 0.8 Very good 0.8 Very good DIN 0.5 Very good 23.8 Satisfactory 0.8 Very good 0.8 Very good DON 0.2 Very good -12.9 Very good 1.0 Very good 1.0 Very good TP 0.3 Very good 12.1 Very good 0.9 Very good 0.9 Very good DIP 0.4 Very good 30.9 Indicative 0.8 Very good 1.0 Very good PP 0.3 Very good 11.7 Very good 0.9 Very good 0.9 Very good DOP 0.6 Good 25.1 Satisfactory 0.7 Very good 0.7 Very good

Contribution by land use Grazing contributed 21% of the TSS export from the region. With a contribution of 65% (the Burnett and Mary contributing 28% and 34%, respectively, and the remaining 3% contributed by the other three basins), streambank erosion is the dominant source of TSS export from the region. Grazing contributes 38% and 39% of the PN and PP, respectively, and is the dominant source of export of particulate nutrients, most of which is generated from the Mary Basin. Sugarcane contributes 56% of the DIN export, followed by grazing accounting for 24%. All of the PSII inhibiting herbicides export is from the sugarcane land use.

Figure 50 Contribution to fine sediment, particulate nutrient and DIN export by land use for the BM region

108 Source Catchments modelling – Updated methodology and results

Despite accounting for less than 2% of the BM region by area, sugarcane contributes 58% of the DIN export from the region. Grazing is the dominant land use accounting for 69% of the region and contributes 25% of the DIN export from the region (Figure 50). In terms of areal rates, sugarcane at 6.7 kg/ha/y, urban at 0.54 kg/ha/y and horticulture at 0. 48 kg/ha/y are the dominant sources of DIN in the BM region. The average DIN per unit area rate for all land uses in the BM is 0.19 kg/ha/yr.

Figure 51 DIN export load and areal rate for each land use in the BM

Sources and sinks In RC2015, streambank erosion contributes the greatest load (52%) to total TSS export from the Burnett Mary region, followed by hillslope erosion (23%) and gullies (15%). The remaining 10% was contributed by instream fine sediment remobilisation which was not included in previous report cards. The constituent budget also shows that 21% of the TSS supplied to the stream network is trapped in storages and 3% is taken out of the stream network through water extractions for different uses with floodplain deposition, losses and stream deposition each accounting for 4% of the total TSS loss. Given the significance of the percentage trapped in storages in the constituent budget, it is important that this component is verified by future monitoring data which has been lacking to date in the region and the GBR as a whole. The fine sediment source sink diagram for the Burnett Basin is presented in Figure 52. The sources and sinks for each constituent are presented in Appendix 3, Table 80.

109 Source Catchments modelling – Updated methodology and results

Figure 52 Fine sediment source sink budget for the Burnett Basin

Storage trapping is a significant loss pathway for fine sediment and particulate nutrients in the BM region, accounting for 21% of TSS being lost. In Figure 53 the trapping efficiencies of storages in the BM model are shown. It should be noted that the relatively low TSS trapping efficiency of the Burnett River Dam (aka Paradise Dam - 33%) is due to the two extreme floods in 2010/11 and 2012/13 the region experienced, which accounted for over 90% of sediment export from the region in the 28 years modelling period (1986–2014).

Figure 53 TSS, PN and PP trapping efficiencies (%) of storages Progress towards Reef Plan 2013 targets There has been very little progress towards meeting the reef targets in RC2015, as was the case in RC2014. This is due to the very low level of investment in management changes required to achieve water quality targets for these two report cards. Percentage load reductions of all constituents from the

110 Source Catchments modelling – Updated methodology and results

BM region are less than the average GBR-wide percentage reductions, which is attributed to the low level of investment in management interventions in this region during this report card. The Burnett and Burrum basins were the only basins where there were some DIN load reductions as management interventions were limited to these areas for this report card. Management changes for sugarcane nutrients was limited to a 0.5% partial improvement from a Cf (C-full) practice to Cp (C-partial). Sugarcane and grazing management practice change summaries are presented in Table 47 and Table 48. For ease of presentation, management classes are aggregated for sugarcane investments (partial steps are lumped to whole).

Table 47 BM sugarcane management practice change summary; all values are in percent

Baseline Soil RC2014 Change RC2015 Change 2013 A 7.75 7.79 0.04 7.79 0 B 30.5 30.83 0.33 30.83 0 C 47.75 47.38 -0.37 47.38 0 D 14 14 0 14 0 Baseline Nutrients RC2014 Change RC2015 Change 2013 A 1 1 0 1 0 B 12 12.14 0.14 12.14 0 C 59.6 59.52 -0.08 59.52 0 D 27.4 27.35 -0.05 27.35 0 Baseline Pesticides RC2014 Change RC2015 Change 2013 A 18 18.53 0.53 18.53 0 B 29 28.95 -0.05 29.13 0.17 C 48 47.51 -0.49 47.34 -0.17 Df 5 5.00 0 5.00 0 Baseline Irrigation RC2014 Change RC2015 Change 2013 A 6 6 0.00 6 0 B-C 28.5 28.54 0.04 28.54 0 D 65.5 65.46 -0.04 65.46 0 Recycle Baseline RC2014 Change RC2015 Change Pits 2013 A 14 14.04 0.04 14.13 0.09 B 11 11.05 0.05 11.08 0.04 C 44 43.92 -0.08 43.79 -0.13 D 29 29 0 29 0

111 Source Catchments modelling – Updated methodology and results

Table 48 BM grazing management Change; all values are percent of total

Hillslope Baseline RC2014 Change RC2015 Change A 1.66 1.68 0.02 1.68 0 B 44.04 44.14 0.10 44.16 0.03 C 54.27 54.16 -0.11 54.14 -0.02 D 0.02 0.02 0.00 0.02 0 Gully Baseline RC2014 Change RC2015 Change A 8.76 8.76 0.00 8.77 0.01 B 20.98 20.99 0.01 21.04 0.04 C 54.12 54.12 0.01 54.10 -0.02 D 16.14 16.12 -0.02 16.09 -0.03 Streambank Baseline RC2014 Change RC2015 Change A 28.98 28.98 0.01 29.01 0.03 B 14.61 14.65 0.04 14.69 0.04 C 18.40 18.39 -0.02 18.34 -0.04 D 27.54 27.51 -0.02 27.49 -0.02 NA 10.47 10.47 0.00 10.47 0

There have been very little reductions in export loads in RC2015 across all constituents of concern, reflecting the level of investment in management changes during these two investment years. This is generally similar to other GBR regions with less than 0.1% TSS export reduction from the BM compared to 0.3% from the GBR as a whole. There has also been a very low DIN export reduction of about 0.03% from the BM compared to 1.1% from the GBR as a whole. The story is similar for PSII herbicides with 0.2% reduction from the BM region compared to 2.5% from the GBR as a whole. These small reductions from the BM region indicate the relatively low level of investment in the BM region for RC2015.

112 Source Catchments modelling – Updated methodology and results

Discussion

Report Card 2015 is the sixth iteration of the Source Catchments modelling of load reductions for reporting on progress towards Reef Plan Water Quality targets. The second external review of the modelling was undertaken in November 2015. Recommendations from the external review and the Queensland Audit Office, have formed the basis of model improvements. In contrast to RC2014, RC2015 saw a number of minor updates to the modelling approach, with only three more significant improvements to note.

Hydrology calibration was updated to ensure modelled estimates of baseflow better aligned with baseflow estimates at gauging stations. The recalibration resulted in small changes in average annual runoff across the six regions with an average of +/- 5% from RC2014 with improved baseflow alignment observed at all gauges (refer Results on page 45).

Instream deposition and channel remobilisation functionalities were enabled in four of the six regions to represent this important component of the sediment budget. In general, the deposition and remobilisation were maintained at a similar magnitude. Further work will be undertaken to refine this component of the sediment budget.

It is important not to lose sight of the main objective of the modelling, to report the water quality benefits as a result of management changes. Therefore the relativity between scenarios is important not the absolute load. However the goal of the modelling program is for continual improvement of modelled load predictions. This approach will continue.

Particulate nutrients There was an under-prediction of particulate nutrient loads in CY, WT and BM. Particulate nutrient under-prediction was also an issue in the remaining regions however it was not as pronounced as in CY and WT.

The gully and streambank models use the subsurface Total Nitrogen (TN) and Total Phosphorus (TP) layers from ASRIS (Henderson et al. 2001) to estimate the particulate nutrient contribution. Wilkinson et al. (2010) describe the differences between the ASRIS spatial data sets and local point data. The surface and subsurface ASRIS nutrient datasets are extrapolated from direct measurements of soil nutrient content to provide a uniform spatial layer as required by the model. In remote areas such as CY there may be very few samples from which to extrapolate a continuous dataset, and hence there can be significant differences between observed and modelled values. Often point source data of nutrient content is sourced from areas of intensive agriculture and as such the input datasets may be more reliable in these areas (Wilkinson et al. 2010). A review of the chemistry data for CY (CSIRO data in Biggs and Philip (1995)), suggests that there was no significant difference between surface and subsurface TN and TP chemistry. However, for the ASRIS derived layers surface concentrations of nutrients in CY were typically three times greater than that of sub-surface concentrations.

113 Source Catchments modelling – Updated methodology and results

Therefore, the subsurface nutrient concentration layers were replaced with surface layer in CY and WT regions. This is in line with the recommendation by Wilkinson et al. (2010) who suggested that users should choose the appropriate dataset for their region including from locally derived measurements or the ASRIS dataset. The updated layers resulted in PN and PP estimates improving from under- predicting loads by 60% to now being within 30% of the observed loads for the Kalpowar gauge in CY.

Progress towards targets Overall the percentage load reductions were low (<1%) ranging from 0.1 – 4.3% across the GBR. These reductions are of a similar order to RC2014 which was the first year under the new water quality management framework.

The cumulative total GBR load reduction for TSS for RC2015 is 12.3% - an increase of 0.3% from RC2014. Four regions contributed to the reduction: WT (0.1%), BU (0.2%), MW (0.1%) and FZ (1%). The Fitzroy reduction was due to significant riparian fencing and investment in improved hillslope management in grains and grazing. The reductions in other regions were due to a mix of improved practices in grazing, sugarcane and banana industries

The cumulative total GBR load reduction for DIN for RC2015 is 18.1% - an increase of 1.1% on RC2014. The majority of the investment occurred in the WT, BU and MW sugarcane areas, with a small amount of investment in improved practices banana growing areas in the WT also contributing.

The cumulative total GBR toxic equivalent load reduction for PSII TE for RC2015 is 33.7% - an increase of 3.2% on RC2014. Small reductions (< 0.3%) were modelled in the Fitzroy and Burnett Mary Regions whilst the Wet Tropics, Burdekin and Mackay Whitsundays all had modelled reductions over 3% (3.4%, 3.6% and 3% respectively). The majority of the reductions were attributed to sugarcane industry investments.

The trend of significant progress in the early GBR Report (RC2010 – RC2103) has slowed in the past two Report Cards (RC2014 and RC2015). This is mainly due to the updated management practice water quality framework resulting in more conservative reductions but likely a more realistic representation of investment and water quality response. A significant feature of the P2R program is that the data produced from both monitoring and modelling can be used to target funding to maximise water quality investment. For example, an outcome of the Reef Water Quality Taskforce investigation was that investment in innovative management practices and monitoring should be undertaken in two key regions (WT and BU) to achieve an improved water quality response for DIN and TSS. The Major Integrated Projects (MIPs) in these regions are well into the planning phase, and the results of the projects will be used to further inform model development, and well as, ideally, have a measurable benefit to GBR water quality.

114 Source Catchments modelling – Updated methodology and results

Scenario modelling – All A and All B management The modelling results suggest that under an All B scenario the 20% TSS Reef Plan water quality target could be met in the MW and Fitzroy regions. Under an All A scenario, the majority of regions can achieve the target, the exception being CY, which was just under the 20% target. For DIN, the 50% Reef Plan water quality target cannot be met under the All B scenario. The Burdekin region is the only region that achieved the target under the All B scenario. For pesticides, the PSII TE target can be achieved under an All B scenario. In summary, the modelling results suggest that the TSS and particulate nutrient targets are achievable under an All A scenario with the majority of the regions close to achieving the TSS target under an All B scenario. The DIN target may be more challenging to achieve even under the All A scenario and new, innovative management practices need to be explored. The PSII TE target is achievable even under an All B scenario. It is important to note that variations in load reductions between regions are related to differences in the dominant sources of pollutants within a region. Therefore reductions in constituent loads from hillslope/gully and streambank sources may not occur proportionally.

The catchment models were set up to assess relative reductions in exported loads due to investment within a region (that might be implemented over an annual funding cycle). The hypothetical representation of the All A and All B scenarios reflect large changes in constituent loads (that would not be expected to be implemented in a short timeframe).

The results of the modelling scenarios should be seen as an approximation of reality. The results do however offer an estimate of the relative magnitude of change that may be required to move closer to achieving the Reef Plan targets. However the many other sources of information available to guide investment should be utilised. Modelling should not be seen as the only driver for decision making.

The following sections provide an overview discussion of results for all six NRM regions.

115 Source Catchments modelling – Updated methodology and results

Cape York The results from RC2015 are considered to be the most refined set of model runs for the Cape York region thus far. The availability of additional datasets against which to compare the model results, in particular, has enabled specific components of the model to be more comprehensively reviewed and refined. This is especially applicable to the representation of the major sediment sources within the region. Furthermore, minor inconsistencies in model parameters have been addressed and updated for this Report Card.

Hydrology The recalibration of hydrology in CY to improve the baseflow component has seen average annual flow reduced by ~3% compared to RC2014, with almost half the reduction in the Normanby Basin. Recalibration has seen a significant improvement in the Normanby Basin baseflow, previously comprising 95% of the total flow, which now contributes around 35% of total flow.

Fine sediment generation, sources and sinks The hillslope fine sediment load is higher in RC2015 than for RC2014 (increased by 42%), resulting in a higher end of system TSS load for CY (RC2014 baseline TSS was 371 kt/yr, compared to 526 kt/yr in RC2015). The modelled load is now within 25% of the Kalpowar gauge monitored load compared to 45% for RC2014. The capping of daily instream sediment concentrations and the reduction in hillslope delivery ratio (HSDR) from 10% to 5% has been the major reason for the improved estimates. Applying the 5% HSDR was in agreement with Rustomji et al. (2010) whilst Brooks et al. (2013) reported much lower generation rates from hillslope plots and higher HSDR.

The updated model has resulted in changes to the fine sediment source sink budget. In the current Report Card, the source sink breakdown in the Normanby Basin remains approximately the same as the previous Report Card, however for the whole of CY there has been an increase of 11% to the hillslope supply load. The gully and hillslope load for CY are now similar (45% and 47% respectively), whereas in RC2014 gullies were the major contributor at 57%. The streambank load remained constant between the two Report Cards. Further investigation of the impact of setting a maximum concentration parameter will be carried out, and application of a spatially variable value will be considered. Due to the lack of monitoring data in CY catchments, data from other regions (e.g. rainforest areas in the WT for application in CY steep, wet catchments) will be considered when setting this value variably.

Particulate nutrients The modelling of particulate nutrients were another component of significant change in this Report Card. Initial iterations of the model showed particulate nutrient loads were well under predicted (up to 60%) for RC2015 compared to loads estimated from measured data. The under-prediction of particulate nutrients was not observed for CY in RC2014 due to minor oversights. In RC2014, a dry weather concentration (DWC) was not applied to the hillslope TSS component in an attempt to better match monitored instream data from the Kalpowar gauge. However, this was not carried over to particulate nutrients supplied from the hillslope. That is, given that particulate nutrients are attached to the fine

116 Source Catchments modelling – Updated methodology and results sediment load, parameters applied to fine sediment should similarly be applied to particulate nutrients. Also in RC2014, the TSS hillslope delivery ratio (HSDR) for particulate nutrients was higher than for TSS. This has been corrected in RC2015 with a consistent HSDR of 5% now applied for particulate nutrients and TSS. To compensate, the resultant decrease in particulate nutrients (by significantly reducing the HSDR from 20% to 5%), combined with reduced baseflow runoff volume (and hence the loss of particulate nutrient generation by implementing a zero DWC) has meant that the enrichment ratio for PN and PP has been increased. This parameter is only applicable to the hillslope erosion derived particulate nutrients, not the gully or streambank derived load. Considering that in some areas up to 80% of the load in CY is derived from gullies, the parameters used to derive gully and streambank particulate nutrients were reviewed. As described, the surface nutrient ASRIS data layers were applied to represent both the surface and sub-surface nutrient concentrations to increase the particulate nutrient load from gully and streambank erosion.

Progress towards targets There were only minor management changes reported in CY for RC2015, with some gully improvements (two projects, total 93Ha) and riparian fencing (four projects, total 36,391 Ha) reported in the Normanby Basin. There were no other grazing, cropping or banana management practices reported. These small changes resulted in no impact on load reduction targets from CY, for any constituents. At the completion of RC2015, the cumulative reduction of fine sediment from CY is 8%, with no contribution from RC2014 or RC2015.

117 Source Catchments modelling – Updated methodology and results

Wet Tropics The major model enhancements for this Report Card include: recalibration of hydrology for selected gauges, in-stream deposition and remobilisation functionality being enabled and adjustments of dissolved phosphorus model in APSIM for sugarcane. Following is a discussion of each of these factors and how they have influenced the modelling results, producing in an improved estimate of constituent loads from WT.

Hydrology The recalibration of hydrology in WT to improve the baseflow component has seen average annual flow increase by 1% compared to RC2014. However it did change total flow in the Barron and Tully basins. Of the 10 gauges that were recalibrated, the baseflow proportion increased for eight of them.

Fine sediment, particulates sources and sinks The total TSS supply increased by 4%, with increases in subsurface erosion (stream and gully) and instream remobilisation compared to the RC2014 baseline model. The small increase in discharge due to the recalibration accounts for the increase in streambank erosion with no other changes made to the streambank model since the last Report Card. Gully cross section was doubled to increase the amount of subsurface erosion (gully and streambank erosion) being generated from grazing lands. This has resulted in a doubling of gully erosion supply. The increase in subsurface erosion aligns with sediment tracing work in the Herbert (Tims et al. 2010). Tims et al. (2010) found that the subsurface to surface ratio was approximately 91% to 9% from two tributaries draining mostly grazing land. Using a comparable subcatchment in the Herbert (Rudd Creek) to that of the Tims et al. (2010) work, with available monitored data to compare to, showed that increasing the gully cross sectional area has resulted in a modelled spilt of 59% subsurface to 41% surface erosion in a predominantly grazed catchment. The modelled value instream at the Rudd Creek gauge was 198 mg/L and the modelled average annual load was 14,265 t/yr. While the monitored instream concentration was 90 mg/L and the monitored load was 8,000 t/yr. It is likely that the hillslope contribution from this area is too high, while the subsurface contribution is still under-predicted despite changes to gully parameters increasing the load from this source. The gully mapping initiative that is currently being undertaken by DNRM across many reef catchments will provide an improved gully density map for the Herbert Basin. It is anticipated that this may result in an increased load of sediment from gully sources.

The implementation of instream deposition and remobilisation in RC2015 resulted in a large proportion of TSS lost to instream deposition and a small proportion of TSS supplied from instream remobilisation. Significant changes in the amount of TSS ‘lost’ occurred compared to RC2014. The large increases in losses of instream and floodplain deposition were due to increasing the sediment settling velocities required for deposition. The sediment settling velocities were updated based on the stream erosion and deposition graph in Earle (2015). This change increased floodplain deposition x20. It is acknowledged that the floodplain deposition may be too high therefore further investigation of the bank full recurrence interval term will be conducted and implemented spatially prior to the next report card. Reducing the amount of floodplain deposition will help to improve the under prediction of TSS at five sites.

118 Source Catchments modelling – Updated methodology and results

Particulate nutrients were mostly under-predicted (-57% to -11%) compared to loads from measured estimates across the five model evaluation sites, except at Tully at Euramo where both PN and PP are over-predicted. To increase the particulate nutrient loads, the sub-surface nutrient concentration layers were replaced with the surface concentration layers used in the calculation of subsurface erosion contribution (refer to overall discussion). Further, due to the general under prediction of fine sediment and particulate nutrient, storage deposition was disabled. Initial iterations of the model for the Tully at Euramo site, showed both PN and PP to be over-predicted. To counter this, the PP enrichment ratio was reduced from five to three in this Report Card to reduce the PP load to better match the monitored loads as it was significantly over-predicted in RC2014 (129%). The Tully at Euramo site is a nested gauge, where the Tully Gorge site (50 km upstream) drains predominantly Nature Conservation, and all of TSS, PN and PP are under-predicted. However, the modelled loads of TSS, PN and PP are over- predicted at the Euramo gauge indicating a significant contribution of these constituents from the major land uses between these gauging stations: sugarcane, grazing and bananas, is occurring. Given the over-prediction at Tully at Euramo, and the under-prediction at all remaining sites, a spatial approach to parameter application should be considered for future Report Cards.

For model evaluation, there was no change in the overall TSS ranking at each site for the short term annual modelled loads or the long term monthly modelled loads compared to RC2014. The modelled loads of TSS and particulates had overall better performance ratings at the long-term monthly scale compared to the shorter annual scale. The South Johnstone River performed overall the best compared to the other sites for TSS and particulates at the monthly scale.

Nutrients For dissolved nutrients, there were large changes in the export load of DOP and DIP, and small increases of DIN and DON export loads compared to RC2014. The largest change occurred for DOP where the export load increased by 89% compared to RC2014. This was due to changing the ratio of DOP to DIP from 0.2 to 0.5 of total dissolved P for sugarcane and increasing the DOP EMC and DWC values where applicable. The average percent difference between the modelled loads and the GBRCLMP observed loads at the six sites was -61% in RC2014 and decreased to -27% in this report card. The DIN loads at all gauging stations were under predicting compared to the load derived from measured data (-33% to -9%) except for Barron River at Myola (43%), but were all within ±50%. Of the five sites that were under predicting DIN, the average was -25% for RC2014 and decreased to -22% in this Report Card. The DIN seepage delivery ratios remained the same as RC2014 and were 100% for all basins except for upstream Barron at Myola (15% DR) and upstream Herbert at Ingham (50% DR). The DWC was set to zero. For the next Report Card, an appropriate DWC will be used and delivery ratios will be further adjusted to align with monitored loads. Similar to sediments and particulates, dissolved nutrients performed better at the long-term monthly scale compared to the shorter annual time scale. The North Johnstone site overall had the best performance rankings for dissolved nutrients at the monthly scale. The incorporation of DIN in deep drainage from banana land use is planned for future Report Cards.

119 Source Catchments modelling – Updated methodology and results

Herbicides The PSII export load for the baseline scenario decreased by 16%, compared to RC2014, predominantly in the Herbert Basin. The differences are due to the approaches used in the allocation of management practice proportions across the APSIM simulations. RC2014 used a database approach to allocate management across available simulations while in RC2015 the assignment was done using expert opinion.

Progress towards targets The WT progress towards the targets for RC2015 were small and less than the changes in RC2014. A number of factors have contributed to the smaller reductions compared to previous Report Cards, which have been discussed in (McCloskey et al. 2017).

For sugarcane, 4% of the sugarcane area had a soil management change, 1% of the area for nutrient management (only partial step changes) and 5% of the area for pesticide management (all full step changes). However, the modelling currently represents four levels of management for pesticide applications (Af, Bf, Cf and Df) and there is no improvement modelled unless a full step change has been achieved. Due to the full step changes that occurred in this Report Card, all of the sugarcane area that underwent a herbicide management change contributed to a reduction in the herbicide load.

For soil management 4% of the banana area underwent a change and for nutrients 3% of the area. Overall, sugarcane and banana represent 8% and 0.7% of the WT area. The overall percent reductions in the export loads are sensible given that the area of management change is very small.

120 Source Catchments modelling – Updated methodology and results

Burdekin

Hydrology The RC2015 model utilised a new hydrology model parameter set, which was produced from a recalibration focused on improving the representation of quickflow (runoff) and slowflow (baseflow) components. The recalibration of hydrology in the Burdekin Region to improve the baseflow component has seen average annual flow increased by 4% (~450,000 ML/yr)from RC2014 to an average annual flow of 12,720,701 ML/yr. In terms of hydrology changes, volume and performance statistics are similar to that of RC2014. However, the prediction of quickflow and slow flow has improved significantly due to the more robust and detailed calibration objective functions. For example, a gauge in the upper Burdekin for RC2014 had a base flow proportion of 67% of the total flow. After recalibration with a baseflow optimiser, the baseflow proportion was reduced to 34%. This aligned with a baseflow estimate of 30% for observed flow estimated using the Lyne-Hollick baseflow separation method (Lyne and Hollick 1979). The improved calibration has led to better process simulation, therefore improving the model comparison dataset for all constituents.

Fine sediment and particulates generation, sources and sinks The RC2015 baseline or current load scenario has generally seen a small increase in sediment (~2%), particulate nutrient (~3%) and dissolved nutrient loads compared to RC2014 (~3-5%). These changes are well within the uncertainty boundaries of load estimates. Individually the model modifications have produced varied results, the changes and the effect they had on model loads are discussed below. The RC2015 baseline sediment sources are dominated by gully and streambank erosion, which aligns with sediment tracing data (~80% subsurface, mainly gully scald, and 20% hillslope) (Hancock et al. 2013, Wilkinson et al. 2013), and also the RC2014 results.

Hillslope In RC2014 spatial patterns in sediment supply at a sub-catchment and sub-basin scale were improved, through the use of improved gully mapping, and seasonal cover combined with more realistic erosion based cover values and the inclusion of a generation ceiling for steep hillslopes. For RC2015, the methodology for calculating hillslope erosion was changed, with the introduction of a new daily sediment ceiling parameter. The RC2014 model used a maximum allowable daily load (t/ha/day), while the RC2015 model parameter uses a maximum allowable daily concentration (mg/L/day). The driver behind this was observation of unrealistically high sediment concentrations being generated on days of low flows in the MW region. The change has improved baseflow loads (TSS, PN, PP) in other GBR catchments and importantly ameliorated technical issues with reporting RUSLE management change.

Particulate nutrients, especially PN, are under-predicted against measured estimates in RC2015. This was also an issue in the RC2013 and RC2014 models despite increases in particulate delivery ratios. It is still unknown which components of the particulate nutrient simulation are the cause of the low estimates (that is, the streambank, hillslope or gully PN or PP loads). The ASRIS nutrient soil

121 Source Catchments modelling – Updated methodology and results concentration layers used in the calculation of subsurface stream and gully erosion are another potential source of error, due to a lack of soil core data or poor spatial extrapolation of boundaries. Further investigation of the ASRIS layers will be undertaken by the modelling team.

Cropping The change in methodology that altered the way slope was applied to dryland cropping resulted in the generated hillslope load for dryland cropping increasing from 28 kt/yr to 168 kt/yr. This increase may be a better estimate of the loads being generated from dryland cropping, however there is no water quality data at an appropriate scale in the Suttor region to test model performance. Average hillslope generation rates in dryland cropping are now 1.2 t/ha/yr whereas the average was ~0.2 t/ha/yr for previous report cards. The increase in the dryland cropping load resulted in minimal changes in the contribution of cropping to end of valley loads due to the high trapping efficiency of the Burdekin Falls Dam and the relatively small loads.

Sugarcane constituent generation The Burdekin region was the second highest contributor of DIN (22%) to the total GBR DIN export load, which is consistent with previous Report Cards. At the regional scale the Burdekin Basin is the largest contributor of DIN, followed by the Haughton and this ranking is the same as previous report cards.

Conceptual improvements were made to the sugarcane paddocking modelling scenario set by the paddock modelling team. These included changes to how soil water balance is modelled across soil type. Previously the minor soil types had sub optimal representation of the runoff and drainage split. This facilitated better modelling of irrigation practices in these minor soils. An extra irrigation management level has been added and there has been some slight modifications to the annual irrigation amounts applied in the Burdekin River and Delta irrigation areas (BRIA). Overall these changes have had little impact on the modelled loads. For example The Burdekin region DIN export load decreased by 5% compared to RC2013, which is well within the uncertainty boundaries of the monitoring data, sugarcane contributing 45% of this load and this is comparable to RC6 (44%) . In addition the majority of the FU scale sugarcane DIN loss is still from seepage (~90%) and the remaining (~10%) from surface runoff. Seepage DRs were set at 100% to better match the limited monitoring’s data at Barratta Creek.

Herbicides The BU contributes approximately a fifth of the PSII herbicide load to the GBR. There was a minor decrease in the baseline RC2015 PSII load compared to the RC2014 PSII load of ~100 kg/yr.

Targeting land management for Reef WQ protection There has been relatively small changes in RC2015 loads compared to RC2014. Therefore readers are referred to the work of Waterhouse et al. (2016). Here the RC2014 catchment modelling results along with other lines of evidence are synthesised in detail providing information on appropriate prioritisation.

122 Source Catchments modelling – Updated methodology and results

Progress towards targets Collectively, management practice changes for grazing and cropping for sediment resulted in zero reported load reduction for fine sediment from the Burdekin region, as the load reduction was less than 0.5%. It is important to acknowledge the considerable difficulties in reporting small changes in grazing land management across the large Burdekin region. In general the scale of investment in terms of area are logical in terms of expectations from investments, with considerable variation in area to reduction ratios to be expected. However as the modelling further improves confidence at these smaller scales should increase. At the completion of RC2015, the modelled cumulative fine sediment reduction for the Burdekin is 17%. The land management practice improvements in sugarcane for nutrients resulted in a 4% reduction in DIN from the Burdekin region, with the cumulative reduction for DIN from the Burdekin being 20%. Improvements were recorded in both sugarcane and cropping (grains) land uses for pesticide use. In sugarcane there was a big shift (10%) out of C class practices into A and B practices, while for cropping areas there was a 4.5% shift from B to A. This resulted in a 3.5% reduction in PSII toxic equivalent loads from the Burdekin region and this is as a result of investment in sugarcane pesticide management.

Future model improvement There is much potential for improvement in the modelling of all constituents. In terms of data inputs large model improvements may be possible through the inclusion, and construction, of LIDAR derived data sets for coastal catchments (it is envisaged that a broader complete of high resolution DEM for the GBR will exist within the next few years). It would be pertinent to start looking at trialling approaches at smaller scales. This would allow for an improved parameterisation of streambanks, gullies and slope factors for the USLE. In addition, approaches could be trialled for a variable hillslope delivery ratio, plus better delineation of scaled “low cover” areas, using higher resolution remote sensing. The sourcing of improved rainfall data, or calibration of the actual rainfall data itself, could further improve hydrology prediction in the Burdekin. Also perhaps some improvement may arise from the inclusion of a rainfall intensity factor, but this idea is more speculative in nature.

123 Source Catchments modelling – Updated methodology and results

Mackay Whitsunday The results from RC2015 are considered to be the best representation of landscape processes for the Mackay Whitsunday region thus far due to the model improvements undertook as part of this report card. The stream stability assessment of the O’Connell River (Alluvium 2014) has provided a new source of data for validating the streambank erosion component of the model.

Hydrology The performance of the RC 2015 baseline model to predict the hydrology across the MW region was assessed using a number of recommended quantitative statistics (Moriasi et al. 2007), namely Nash- Sutcliffe efficiency (NSE), percent bias (PBIAS), and coefficient of determination (R2). Estimated NSE and PBIAS values for the ten gauges (see Table 52, Appendix 1 – Hydrology calibration results) used in the calibration of the regional hydrologic model were compared against a general performance ratings developed in Moriasi et al. (2007). The quantitative performance of the hydrologic model for the O’Connell and Pioneer River Basins were rated as ‘Very Good’. The quantitative performance for Plane Creek Basin was rated as ‘Very Good’ at Sandy Creek at Homebush (126001A) and ‘Good’ at Plane Creek at Sarina (126002A) and Carmila Creek at Carmila (126003A). The Proserpine River Basin was rated ‘Satisfactory’ due to the low NSE and high PBIAS statistics at Lethe Brook at Hadlow Road (122012A). Gauge 122012A may be not suitable for calibration because it is situated at the downstream of a highly regulated catchment for water extraction for irrigation and inflow from Peter Faust Dam (Packett et al, 2014). This may be indicating further investigations are required to assess the suitability of this gauge (or any other gauge) for better calibrating the hydrologic properties of the Proserpine River Basin.

Both quantitative and qualitative assessments of catchment model hydrologic calibrations are critical to ensure that the hydrologic performance was suitable for not only predicting long term data such as mean annual flows and loads, but also to ensure that the characteristics of the key flow events were also simulated correctly. Hence, the qualitative model performance was assessed using time-series plot, flow duration curve (log scale) and one to one scatter plot (Appendix A).

The RC2015 baseline hydrologic model predicts the baseflow contributions well at six of 10 gauges across the MW region. The recent hydrology calibration has improved estimates of baseflow contributions at three gauges in the Proserpine River, O’Connell River and Plane Creek basins. The model still underestimates the baseflow contributions at three gauges in the Pioneer River and Plane Creek 1-7% lower than the Lyne-Hollick filter estimates.

The RC2015 baseline model estimated mean annual discharges very well against that of GBRCLMP estimates in the Pioneer River at Dumbleton and in Sandy Creek at Homebush sites and P2R estimates at Multifarm site. The RC2015 model also well matched the long-term annual flows estimated by Joo et al. (2014) in the Pioneer River at Dumbleton.

124 Source Catchments modelling – Updated methodology and results

Constituent generation and validation Flows and pollutant loads estimated as part of the Great Barrier Reef Catchments Load Monitoring Program (GBRCLMP) and P2R programs were compared against the RC2015 baseline model estimates to assess the model performance at catchment and sub-catchment scales respectively. Flows and pollutant loads estimated by Joo et al. (2014) in the Pioneer River at Dumbleton Pump Station between 1987 and 2009 were also utilised to assess the RC2015 baseline model performance to capture the hydrologic and water quality responses to the long-term (23 years) seasonal climatic variability.

The RC2015 baseline model showed good agreement with the mean annual flows and pollutant loads estimated at two GBRCLMP EoS sites. The model also showed good agreement with the mean annual flows and pollutant loads estimated at the P2R monitoring site during low and medium flow events. The model showed good correlation to the long-term estimates derived by Joo et al. (2014). It should also be noted that there is a degree of uncertainty around the annual loads estimated by the reef water quality monitoring programs and Joo et al. (2014). Estimation of uncertainty bounds around the observed loads will be beneficial to better assess the model performance more fully.

Several improvements were made to enhance the performance of the RC2015 baseline model. One of the main improvements to the RC2015 baseline model was the recalibration of hydrology to better represent the baseflow indices1 estimated from long-term gauged flow measurements using the Lyne and Hollick filter (Lyne & Hollick, 1979), a widely used approach for baseflow separation. The recalibration of hydrology for the MW region has seen average annual flow reduced by approximately 0.5% compared to RC2014.

The GBR Source catchments model platform utilises the Revised Universal Soil Loss Equation (RUSLE) to estimate daily hillslope soil loss from the grazing land use in the Mackay-Whitsunday region. In the RC2015 models, a ‘maximum concentration’ cap was applied as opposed to previous ‘maximum load’ cap. The modification has improved modelled predictions against raw concentration measurements available for grazing dominant catchments. Moreover, the RC2014 baseline model underestimated fine sediment (<20 µm) contribution from the grazing land use. Given that the grazing land use dominates the Mackay-Whitsunday region, the relative contribution of fine sediment from the grazing land use was increased in the RC2015 baseline model.

The RC2014 baseline model under-predicted fine sediment and associated particulate nutrient loads estimated at the Pioneer River at Dumbleton Pump Station. It was noted that the RC2014 model was set to trap significant amount of fine sediment and associated particulate nutrient loads (~50% of the exported loads) at weirs located along the Pioneer River system (i.e. Mirani, Marian, and Dumbleton). Trapping at these weirs was turned off for RC2015 model as minimal trapping of fine sediment is

1 Baseflow index is the ratio of long-term baseflow to total stream flow and it represents the contribution of seepage of water from the ground to river flow.

125 Source Catchments modelling – Updated methodology and results expected at those weirs. The trapping model can be calibrated/validated should water quality data be made available for both up and down stream of the weirs.

Constituent sources and sinks Hillslope erosion from grazing, sugarcane, and cropping land uses is the dominant (~54%) source of TSS supply in the MW region (see Table 3) followed by streambank erosion (~31%). Erosion from the remaining land uses constitutes ~14% of the regional TSS supply. Gully erosion contribution for the regional TSS supply is low (~1%). RC 2015 baseline model indicates that approximately 96% of the regional TSS supply is exported, while only ~ 4% is lost due to reservoir deposition, extraction and floodplain deposition. Reservoirs in the region trap approximately 50% of the total TSS load that is lost, while extraction and flood plain deposition claims approximately 32% and 19% of the total TSS loss.

Hillslope erosion dominates the particulate nitrogen and phosphorus supply by respectively generating approximately ~63% and ~73% of the regional loads. Although streambank erosion is the second highest source for the regional TSS supply, it only contributes approximately 8% and 6% to the regional particulate supply. This indicates that the regional particulate nutrient levels are higher in the surface soil than sub-surface soil.

Approximately 46% (~635 t/yr) of the regional dissolved inorganic nitrogen supply is through deep- drainage in sugarcane, while approximately 20% (~273 t/yr) of the regional DIN supply is from sugarcane surface runoff. The RC2015 baseline model indicated that only 20% of the total DIN leached into the deep-drainages reaches the surface water, while a significant portion of DIN (~2,500 t/yr) is permanently lost to the groundwater system. This needs further investigation as DIN contribution by fallow legumes in the region may be overestimated in regional paddock models. The remaining land uses contribute ~33% of the regional DIN supply. Point source (e.g. Sewage treatment plants) DIN contribution is low (~2%). Extractions dominate the loss of DIN in the region, with approximately 3% of the total DIN supply is lost via extractions. PSIIs, and thus the toxic load, supply is dominated by sugarcane, while only approximately 2.6% of the PSII supply is lost through extractions.

The comparison between the modelled fine sediment export due to streambank erosion along the O’Connell River against that estimated as part the stream stability assessment conducted by Alluvium Consultancy is encouraging given the modelled streambank erosion is underestimated only by approximately 16% for the O’Connell River.

Progress towards targets RC2015 models estimated marginal reductions in fine sediment and particulate nutrient loads (< 0.1 %), while estimating 1.1% and 2.9% reductions in dissolved inorganic nitrogen (DIN) and total toxic loads respectively. All-A and All-B scenario modelling outcomes indicated that Reef Plan water quality targets can be achieved for all pollutants, except DIN, should all land use practices improve to meet ‘B’ management standards. The modelling indicated that the WQ target for DIN cannot be achieved even if all land use practices are improved to meet ‘A’ management standards. This may suggest further

126 Source Catchments modelling – Updated methodology and results investigations are required to assess the efficacy of the proposed best practice to manage DIN reductions for water quality improvement as well as the accuracy of methods used to represent DIN management practices in paddock and (or) catchment models.

127 Source Catchments modelling – Updated methodology and results

Fitzroy The updates to the model for RC2015 have resulted in modelled loads being better aligned to observed loads. The inclusion of the rock cover layer provides an improved representation of hillslope erosion rates in nature conservation and forestry areas; while improvements to paddock scale modelling of dryland cropping have significantly improved the representation of these FUs in the model. Furthermore, minor inconsistencies in model parameters have been addressed and updated for this Report Card.

Hydrology The modelled baseflow and quick flow ratios have been improved by the recalibration of the hydrology. There has been a significant improvement in the modelled baseflow when compared to the Lyne-Hollick baseflow filter estimate for the observed baseflows. The average difference between observed and modelled baseflow at gauges improved from 29% for the previous report card to 2% for this report card. Compared to the previous report card, the changes have significantly increased the proportion of surface runoff (quick flow) for many catchments and as a result have increased erosion in these catchments. The effect of this change on the various erosion processes is discussed in the relevant sections below. The improvement in the models baseflow and quick flow ratios has resulted in small changes to the total predicted flow. Previously the model over predicted the combined flow at all gauges used for calibration by 1%; however for this report card, the model under predicts the flow by 2%. The reduction in flow was observed at most gauges and on average each gauge had a reduction in flow of 1%.

Fine sediment and particulates generation, sources and sinks The changes to the RC2015 input data, parameters and methodology has produced a small increase (2%) in the total fine sediment load exported to the GBR lagoon from the Fitzroy region. Individually the modifications have produced varied results, the changes and the effect they had on model loads are discussed below.

The model results indicate that the Fitzroy Basin contributes 83% of the TSS load for the region and greater than 70% of the exported load for all other constituents except PP (~48%) and PN (~54%). Of the minor contributors, Water Park Creek, the smallest basin in area (~1%), produces the second highest contribution to the export load for PN at ~20% and PP at ~15%. This is due to the high proportion of the export load that is derived from hillslope erosion as a result of the high rainfall and steep slopes within this basin and its proximity to the GBR lagoon. The Boyne has the lowest contribution to export loads for all constituents.

Hillslope The rock factor layer was introduced into the model to reduce soil erodibility factor in conservation and forestry areas with steep slopes that were producing very high erosion rates which in reality have low erosion rates due to the surface rock cover. The rock factor layer was applied to the entire Fitzroy region and produced reductions in soil erodibility in approximately half of the Fitzroy region. In ten per cent of

128 Source Catchments modelling – Updated methodology and results the region, soil erodibility was reduced by over half of the original values and the overall average reduction was a 27% decrease in the soil erodibility factor.

The largest reductions in soil erodibility occurred in the Isaac, Styx and Fitzroy catchments along the Connors, Broadsound and Boomer ranges as well as in an area between Marlborough and Yeppoon. These reductions occurred in forestry, grazing and conservation areas. Similar reductions occurred in the Boyne and Calliope catchments. For the majority of the Fitzroy Basin there were little or no reductions in soil erodibility; however significant reductions occurred in Callide Creek area in an area northwest of Biloela and some smaller reductions occurred in an area south of Emerald. These reductions occurred in open and forested grazing, forestry and conservation areas.

The reduction in soil erodibility enabled the hillslope delivery ratio to be altered in conservation and forestry areas where a very low delivery ratio of 2.5% had previously been used to reduce very high erosion rates. The delivery ratio was made consistent with the surrounding areas and was changed to 5% for the Fitzroy Basin and 10% in the wetter and steeper coastal catchments.

The combined effect of the reduction in soil erodibility factor and increased delivery ratio resulted in a 45% reduction in generated hillslope loads in grazing and forestry areas and a 50% increase in the conservation generated hillslope load compared to the previous report card.

Cropping The incorporation of spatially variable slope as opposed to an average slope, into the HowLeaky cropping model resulted in the generated hillslope load for dryland cropping increasing more than tenfold from 147 kt/yr to 1,591 kt/yr. This large increase is considered to be a better estimate of the loads being generated from dryland cropping. Average hillslope generation rates in dryland cropping are now 2 t/ha/yr whereas the average was ~0.2 t/ha/yr for previous report cards. The increase in the dryland cropping load resulted in the total generated hillslope load increasing by 60% and the exported hillslope load increasing by 20%.

In previous Report Cards, the dryland cropping HowLeaky simulations were used to approximate the irrigated cropping land use in the Fitzroy NRM region. A decision was made not to use this approach for RC2015, because there is no evidence to support the assumption that irrigated cropping produces similar erosion rates as dryland cropping. Typically, within irrigated cropping areas of the Fitzroy region all irrigation tail water is captured and stored on farm. The decision was made not to model hillslope erosion within these areas until a better understanding of the proportion of sediment that is transported off irrigated cropping areas is acquired. The Queensland Department of Agriculture and Fisheries (DAF) is currently investigating irrigation practices and runoff capture within the Fitzroy region. The results of this research will be used to improve the modelling of hillslope erosion in irrigated cropping areas within the Fitzroy region. By choosing not to use dryland cropping HowLeaky runs, there is now a zero fine sediment contribution for hillslope erosion for irrigated cropping areas. This change had a minor impact,

129 Source Catchments modelling – Updated methodology and results as irrigation is <1% of the catchment, thus only reducing the exported sediment load by 10kt/yr compared to the previous Report Card.

Gully Changes to the gully full maturity year and gully activity parameters implemented to reduce gully erosion contribution, resulted in a slight increase in the total fine sediment load generated by gully erosion. The change to the gully full maturity year resulted in a reduced rate of gully erosion activity being applied to the entire model period rather than just the last few years as used in the previous report card.

Streambank The combined effect of the changes described above prompted a reassessment of end of system loads, sources of sediments and losses. An assessment of bank full flow recurrence interval and floodplain deposition indicated that there was potential to increase the deposition by reducing the recurrence interval from 6 years to 4.7. This change increased floodplain deposition by 8% and improved the model’s end of system estimation. Losses to floodplain deposition now accounts for 48% of the sediment generated by the model which fits within the findings of Trimble (1983), Phillips (1991) and Hughes et al. (2009) who reported that half of river sediment loads can be deposited on floodplains. The application of a variable bank full flow recurrence interval across the Fitzroy region will be explored for future report cards to improve the spatial distribution of streambank erosion and floodplain deposition.

The reduction in the recurrence interval produces a small reduction in the sediment load generated by streambank erosion and the changes to the regression relationship for channel height also reduced streambank erosion. The combined result of these changes produced a 5% reduction in the generated streambank load.

As a result of all the changes the average annual modelled fine sediment load at Rockhampton is 1,424 kt/yr which is 7% below the combined GBRCLMP monitored load (2006–2014) and the FRCE estimate (1986–2006). The modelled average annual sediment concentration for fine sediment is 247 mg/L which provides a good estimate of the combined GBRCLMP monitored load (2006–2014) and the FRCE estimate (1986–2006) average concentration of 274 mg/L.

Event Mean Concentration (EMC)/Dry Weather Concentration (DWC) The adjustment of Event Mean Concentration/Dry Weather Concentration (EMC/DWC) values improved prediction of nutrient loads, as expected. Constituents modelled solely with the EMC/DWC approach (DIN, DON, DIP and DOP) are within ±2% of monitored loads (GBRCLMP) with most within ±1%. Concentrations similarly are within ±1% of monitored concentrations. DIN and DON are generally over predicted during high flow events whereas DIP and DOP are under predicted during these events.

130 Source Catchments modelling – Updated methodology and results

Particulate Nutrients The model under predicts particulate nutrient loads by 9 % for nitrogen and 5% for phosphorus. Concentrations of particulate nutrients were also under predicted by similar proportions. With both particulate nitrogen and phosphorous being under predicted by the model future investigation will be focused on the relationship between reservoir deposition of sediment and particulate nutrients within the model.

Progress towards targets Management changes within the Fitzroy region for RC2015 were again relatively small. Whilst they were larger than the previous report card the changes were generally smaller than previously reported changes. The changes in management practices resulted in load reductions for fine sediment, particulate nutrient loads and toxic load. The largest reduction in the fine sediment load across the GBR occurred within the Fitzroy region.

Improvements occurred in gully, streambank and cropping management practices. A 1% reduction in the anthropogenic fine sediment load was modelled for the Fitzroy region for RC2015 and resulted in the cumulative reduction within the Fitzroy region increasing to 5.5%. Nitrogen and phosphorus loads were both reduced by 0.2% producing a cumulative load reduction of 3.2% and 6.7% respectively. The toxic load reduced by 0.3% which improved the cumulative reduction to 4.3%.

Significant management changes are required to be implemented in the coming years to achieve the Reef Plan targets within the Fitzroy region. The size of the required changes exceeds all changes reported thus far.

131 Source Catchments modelling – Updated methodology and results

Burnett Mary The results from RC2015 are an improvement on previous model runs for the Burnett Mary region thus far. The main reason for the improved representation of constituent loads in this Report Card is the recalibration of hydrology for baseflow, and the resultant increase in average annual flow from the BM region to better align with gauged flows. Furthermore, minor inconsistencies in model parameters have been addressed and updated for this Report Card.

Hydrology The modelled hydrology has been improved by recalibrating the model to better predict the proportion of total flow represented as baseflow. Overall this increased the average annual flow of the BM region by 16% compared to the baseline model for RC2014. Baseflow was reduced in 76% of the calibration gauges. In three basins (Baffle, Kolan and Burrum) the average annual flow was reduced by less than 5%. In the Burnett Basin average annual flow increased by 1%, while in the Mary, the wettest basin, it increased by almost 35%. In the Mary Basin there were six gauges used for hydrology calibration. In RC2014 the average baseflow for these gauges was 33% (ranging from 1% to 60%), while the recalibration in RC2015 resulted in an average baseflow for the same six gauges of 24% (ranging from 18% to 29%). This decrease in baseflow leads to an increase in quickflow; which is the major cause of the 35% increase in average annual modelled flow for the Mary Basin. In summary, the reason for the discrepancy between modelled flows in the two report cards is the adjustment of baseflow proportions as the result of the hydrology re-calibration in RC2015. Since modelled baseflow proportions in RC2015 better align with baseflow proportions from monitored data, the hydrology model results in RC2015 are an improvement over those in RC2014. It should also be noted that increase in flow in the Mary will lead to resultant increases in constituent generation in this basin.

Fine sediment and particulate nutrient generation, sources and sinks In RC2015, the BM Source Catchments modelling estimates the TSS contribution exported from the region to the coast is 23% from hillslope erosion, 15% gully, with the largest contribution from streambank erosion (52%). The remaining 10% is contributed by instream fine sediment remobilisation which was not included in previous report cards. This is in contrast to results from RC2014 where contributions to TSS export were 38%, 20% and 42% from hillslope, gully and streambank erosion, respectively. Since streambank erosion is a function of flow, the fact that flow modelling in RC2015 is an improvement from that in RC2014 implies that the percentage contribution from streambank erosion in RC2015 is also an improvement on that in RC2014.

The contribution from hillslope erosion to the fine sediment load has decreased in the BM from RC2014 by 15%, and there are a number of parameter changes which have caused this. In the first instance, the improved quickflow/baseflow representation has meant an overall increase in average annual flow, which will result in increased constituent generation for the quickflow component. To offset this, the fine sediment (and particulate nutrient) hillslope delivery ratio has been decreased from 20% to 5%, in order to better match monitored loads. Further, the DWC for USLE parameterisation has been reduced from 100 mg/L to 0 mg/L. In RC2014 baseflow was much greater than for RC2015, hence more TSS load

132 Source Catchments modelling – Updated methodology and results will have been generated by the DWC component of the model. By setting this value to zero in RC2015 there is now no load generated during baseflow (or slow flow) conditions.

Streambank Streambank erosion is the dominant source of fine sediment in the Mary Basin, accounting for 65% of the sediment export from the basin itself and 34% of the regional TSS export. In both RC2015 and RC2014 the Mary catchment had the greatest contribution of streambank erosion from the BM region.

The dominance of streambank erosion as a source of TSS export is a consequence of relatively high channel capacity and stream power, as a result of large historical channel incision and widening coupled with localised clearing of riparian vegetation. In both RC2014 and RC2015 the Mary Basin had the greatest contribution of streambank erosion from the region, and this was the case from earlier modelling undertaken by DeRose et al. (2002) and the National Land and Water Audit (NLWRA,2001). Whilst there is a general lack of measured data to evaluate the extent and severity of streambank erosion in the GBR as whole (Bartley et al. 2015), there is evidence that streambank erosion is a very important issue in the region (Simon, 2014 and Department of Natural Resources and Mines, 2015).

Storage trapping The constituent budget for the Burnett Mary region shows that 21% (498 kt/y) of the TSS supplied to the stream network is trapped in storages and a further 7% (174 kt/y) is taken out of the stream network through water extractions and other losses (e.g. sediment still in route through the system), floodplain, and stream deposition, respectively, account for 4% (91 kt/y) and 5% (101 kt/y) of the fine sediment budget. The Burnett catchment generated almost 48% (1,119 kt/y) of the regional TSS load to the stream network, with 38% (548 kt/y) exported from the basin. On the other hand, the Mary catchment supplied 44% (1,017 kt/y) of the total TSS to the stream network and exported 53% (770 kt/y). The relatively small TSS per cent contribution to export from the Burnett catchment in RC2015 compared to the Mary is a result of the high sediment trapping efficiency of six of the 11 sediment water storages in the lower reaches of the catchment and losses associated with water extractions during the modelling period. In terms of the overall constituent budget, it is worth noting that sediment deposition in storages is the major sink of TSS within the region accounting for about 21% of the TSS supply. The presence of fewer and smaller water storages in the Mary catchment (three out of 11 can partly explain the large contribution of TSS export from this basin). Given the significance of storage trapping on the constituent budget, it is important that this component is verified by future monitoring data which has been lacking to date in the region and in the GBR as whole.

Sugarcane dissolved nutrients and PSII loads The load conversion factor in the sugarcane parameteriser is an adjustable parameter in the model and in RC2014 (and all previous Report Cards) was set at a default of 0.8 for DIP and 0.2 for DOP. As a result of the under-prediction of DIP and DOP loads at model evaluation sites, these were changed to 1.0 for DIP and 1.2 for DOP in RC2015. This change can be justified by the fact that mill mud which is commonly applied to supplement nutrients in sugarcane has not been accounted for by the model and

133 Source Catchments modelling – Updated methodology and results the existence of evidence there bioavailable phosphorous increases with mill mud application (Moody and Schroeder 2014).

DIN and PSII herbicide exports are predominantly associated with the sugarcane and cropping land uses in each catchment in the region. In the Burnett Mary region as a whole, DIN export increased by 20% in this report card compared to the previous report card due to the increase in flow in the Mary and Burrum basins. PSII loads reduced by 38% compared to the previous report card. These differences are attributed baseline management distributions were improved between RC2014 and RC2015.

Most of the DIN export is associated with fertiliser losses in sugarcane growing areas. Adoption of the 6 easy steps program for nutrient management is expected to result in major contribution to reduction of DIN export. All of the PSII inhibiting herbicides exported are due to the application of these chemicals to control weeds in sugarcane farms. The sugarcane industry in the Burnett, Burrum, Mary and Kolan basins contributes 82% of the PSII export.

Progress towards targets There has been very little progress towards meeting the reef targets in RC2015, as was the case in RC2014. This is due to the very low level of investment in management changes required to achieve water quality targets for these two report cards. Moreover, the impact of management changes on loads in both report cards has been moderate compared to previous report cards. Percentage load reductions of all constituents from the BM region are less than the average GBR-wide percentage reductions, which is attributed to the low level investment in management interventions in this region during this report card. The Burnett and Burrum basins were the only basins where there were some DIN load reductions as management interventions were limited to these areas for this report card.

134 Source Catchments modelling – Updated methodology and results

Conclusion

The catchment scale water quality modelling as described in this report is one of multiple lines of evidence used to report on progress towards Reef Plan 2013 targets. Investments in improved land management practices between 2008 and 2015 have resulted in a fine sediment load reduction of 12.3%. Particulate nitrogen and phosphorus have declined by 11.6% and 14.7% respectively. Progress towards meeting the 50% DIN target by 2018 has been ‘very poor’, with an 18.1% reduction reported to date. PSII TE load, however, is over halfway to its 60% reduction target, with a 33.7% reduction reported.

The baseline loads presented in this report are typically higher than for previous Source Catchments sediment and nutrient loads. The reasons for the higher loads are methodological changes such as hydrology model recalibration to better represent baseflow proportion in line with modelled data. Generally this had the effect of reducing the baseflow portion and increasing quickflow which in turn increases constituent generation. Also, instream deposition and remobilisation were enabled in four regions, and while the relative contribution of both is similar, this has also increased end of system loads slightly.

Overall, the catchment scale water quality modelling has been successful, and the aim of reporting towards Reef Plan 2013 targets had been achieved. The results show that while progress is being made towards meeting the sediment, nutrient and pesticide targets, further work still needs to occur to ensure we reach the targets by 2018.

135 Source Catchments modelling – Updated methodology and results

References

Alluvium (2014). O'Connell River Stability Assessment.

Australia, C. o. (2015). Reef 2050 Long-Term Sustainability Plan.

Biggs, A. J. W. and S. R. Philip (1995). Soils of Cape York Peninsula. Brisbane, Queensland Dept. of Primary Industries.

Black, D., P. Wallbrink, P. Jordan, D. Waters, C. Carroll and J. Blackmore (2011). Guidelines for water management modelling: Towards best-practice model application. Canberra, Australia, eWater Cooperative Research Centre.

Brodie, J., J. Waterhouse, B. Schaffelke, F. Kroon, P. Thorburn, J. Rolfe, J. Johnson, K. Fabricius, S. Lewis, M. Devlin, M. Warne and L. McKenzie (2013). 2013 Scientific consensus statement, Land use impacts on Great Barrier Reef water quality and ecosystem condition. Brisbane, Reef Water Quality Protection Plan Secretariat.

Brooks, A., J. Spencer, J. Olley, T. Pietsch, D. Borombovits, G. Curwen, J. Shellberg, C. Howley, A. Gleeson, A. Simon, N. Bankhead, D. Klimetz, L. Eslami-Endargoli and A. Bourgeault (2013). An emprically-based sediment budget for the Normanby Basin: sediment sources, sinks, and drivers on the Cape York Savannah, Australian Rivers Institute, Griffith University: 506pp.

Cabinet, D. o. t. P. a. (2009). Reef Water Quality Protection Plan, 2009, For the Great Barrier Reef World Heritage Area and adjacent catchments, The State of Queensland and the Commonwealth of Australia. www.reefplan.qld.gov.au.

Department of the Premier and Cabinet (2013). Reef Water Quality Protection Plan, 2013, Securing the health and resilience of the Great Barrier Reef World Heritage Area and adjacent catchments, Reef Water Quality Protection Plan Secretariat. www.reefplan.qld.gov.au.

Earle, S. (2015). Physical Geology. Physical Geology. British Columbia, British Columbia.

Ellis, R. J. and R. D. Searle (2014). Dynamic SedNet Component Model Reference Guide. Concepts and algorithms used in Source Catchments customisation plugin for Great Barrier Reef catchment modelling. Bundaberg, Queensland, Queensland Department of Science, Information Technology, Innovation and the Arts.

Garzon-Garcia, A., R. Wallace, R. Huggins, R. D. R. Turner, R. A. Smith, D. Orr, B. Ferguson, R. Gardiner, B. Thomson and M. S. J. Warner (2015). Total suspended solids, nutrient and pesticides loads (2013-2014) for rivers that discharge to the Great Barrier Reef - Great Barrier Reef Catchment Loads Monitoring Program. Brisbane.

GBRMPA (2016). "Great Barrier Reef Marine Park Authority." from www.gbrmpa.gov.au.

Hancock, G. J., S. N. Wilkinson, A. A. Hawdon and R. J. Keen (2013). "Use of fallout tracers 7Be, 210Pb and 137Cs to distinguish the form of sub-surface soil erosion delivering sediment to rivers in large catchments." Hydrological Processes.

Henderson, B., E. Bui, C. Moran, D. Simon and P. Carlile (2001). ASRIS: Continental-scale soil property predictions from point data. Canberra, CSIRO Land and Water. 28/01.

Hughes, A. O., J. M. Olley, J. C. Croke and L. A. McKergow (2009). "Sediment source changes over the last 250 years in a dry-tropical catchment, central Queensland, Australia." Geomorphology 104(3- 4): 262-275.

Jakeman, A. J., R. Argent, J. Brodie and M. Littleboy (2012). External Review: Catchment and Paddock Modelling, Paddock to Reef Integrated Monitoring, Modelling and Reporting Program. Brisbane, Department of the Premier and Cabinet.

136 Source Catchments modelling – Updated methodology and results

Jakeman, A. J., I. Prosser, R. Grayson and M. Littleboy (2015). An Independent Review of the Paddock Modelling and Catchment Modelling Program. Brsibane, Department of Natural Resources and Mines: 13.

Joo, M., V. McNeil, C. Carroll, D. Waters and S. Choy (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 (unpublished report). Report.

Joo, M., M. A. A. Raymond, V. H. McNeil, R. Huggins, R. D. R. Turner and S. Choy (2012). "Estimates of sediment and nutrient loads in 10 major catchments draining to the Great Barrier Reef during 2006- 2009." Marine pollution bulletin 65(4-9): 150-166.

Kroon, F., K. Kunhert, A. Henderson, R. Turner, R. Huggins, S. Wilkinson, B. Abbott, J. Brodie and M. Joo (2010). Baseline pollutant loads to the Great Barrier Reef, CSIRO: Water for a Healthy Country Flagship Report series ISSN: 1835-095X. 41 pp.

Kroon, F. J., P. M. Kuhnert, B. L. Henderson, S. N. Wilkinson, A. Kinsey-Henderson, B. Abbott, J. E. Brodie and R. D. R. Turner (2012). "River loads of suspended solids, nitrogen, phosphorus and herbicides delivered to the Great Barrier Reef lagoon." Marine pollution bulletin 65(4-9): 167-181.

Lyne, V. and M. Hollick (1979). Stochastic time variable rainfall-runoff modelling, Perth, Institution of Engineers National Conference Publication.

McCloskey, G. L., R. Ellis, D. K. Waters and C. Carroll (2014). Modelling reductions of pollutant loads due to improved management practices in the Great Barrier Reef Catchments: Cape York NRM region, Technical Report, Volume 2. Cairns, Queensland Department of Natural Resources and Mines.

McCloskey, G. L., D. Waters, R. Baheerathan, S. Darr, C. Dougall, R. Ellis, B. Fentie and L. R. Hateley (2017). Modelling reductions of pollutant loads due to improved management practices in the Great Barrier Reef catchments: updated methodology and results. Technical Report ffor Reef Report Card 2014, Queensland Department of Natural Resources and Mines.

Moody, P. W. and B. Schroeder (2014). Mill mud and mill mud products: efficacy as soil amendments and assessment of environmental risk. Brisbane, Department of Science, Information Technology, Innovation and the Arts, Queensland Government.

Moriasi, D. N., J. G. Arnold, M. W. Van Liew, R. L. Bingner, R. D. Harmel and T. L. Veith (2007). "Model evaluation guidelines for systematic quantification of accuracy in watershed simulations." Transactions of the ASABE 50(3): 885-900.

Moriasi, D. N., M. W. Gitau, N. Pai and P. Daggupati (2015). "Hydrologic and water quality models: Performance measures and evaluation criteria." Transactions of the ASABE 58(6): 1763-1785.

Office, Q. A. (2015). Managing water quality in Great Barrier Reef catchments. Report 20: 2014-15. Brisbane, Queesland Audit Office.

Owens, J. S. S., M. (2016). Modelling Reductions of Pollutant Loads Due to Improved Management Practice in the Great Barrier Reef Catchments Paddock Modelling for Grains, Technical Report. Report Card 2014 and Report Card 2015. Brisbane, Queensland Department of Natural Resources and Mines.

Packett, R., R. Ellis, D. K. Waters and C. Carroll (2014). Modelling reductions of pollutant loads due to improved management practices in the Great Barrier Reef Catchments – Mackay Whitsunday NRM region, Technical Report, Volume 5. Cairns, Queensland Department of Natural Resources and Mines.

Phillips, J. D. (1991). "Fluvial sediment budgets in the North Carolina Piedmont." Geomorphology 4(3– 4): 231-241.

137 Source Catchments modelling – Updated methodology and results

Rose, C. W. and R. C. Dalai (1988). "Erosion and runoff of nitrogen." Advances in Nitrogen Cycling in Agricultural Ecosystems: 212-235.

Rustomji, P., J. Shellberg, A. Brooks, J. Spencer and G. Caitcheon (2010). A catchment sediment and nutrient budget for the Mitchell River, Queensland. A report to the Tropical Rivers and Coastal Knowledge (TRaCK) Research Program, CSIRO Water for a Healthy Country National Research Flagship.

Scarth, P., M. Byrne, T. Danaher, B. Henry, R. Hassett, J. Carter and P. Timmers (2006). State of the paddock: monitoring condition and trend in groundcover across Queensland. In: Proc. of the 13th Australasian Remote Sensing Conference, Earth observation: from science to solutions, Canberra, Spatial Sciences Institute (Australia).

Sharpley, A. N. (1985). "The selective erosion of plant nutrients in runoff." Soil Science Society of America Journal 49(6): 1527-1534.

Sharpley, A. N., P. J. A. Kleinman, R. W. McDowell, M. Gitau and R. B. Bryant (2002). "Modeling phosphorus transport in agricultural watersheds: Processes and possibilities." Journal of Soil and Water Conservation 57(6): 425-439.

Shaw, M. and D. M. Silburn (2016). Modelling reductions of pollutant loads due to improved management practices in the Great Barrier Reef catchments – Paddock Scale Modelling, Technical Report - Report Card 2010 to 2013. Brisbane, Queensland Department of Natural Resources and Mines.

Shaw, M. F., G.; Silburn, M. (2016). Modelling Reductions of Pollutant Loads Due to Improved Management Practice in the Great Barrier Reef Catchments Paddock Modelling for Sugarcane, Technical Report. Report Card 2014 and Report Card 2015. Brisbane, Queensland Department of Natural Resources and Mines.

Thorburn, P. J. and S. N. Wilkinson (2012). "Conceptual frameworks for estimating the water quality benefits of improved agricultural management practices in large catchments." Agriculture, Ecosystems and Environment.

Thornton, C. M., B. A. Cowie, D. M. Freebairn and C. L. Playford (2007). "The Brigalow Catchment Study: II. Clearing brigalow (Acacia harpophylla) for cropping or pasture increases runoff." Australian Journal of Soil Research 45(7): 496-511.

Tims, S. G., S. E. Everett, L. K. Fifield, G. J. Hancock and R. Bartley (2010). "Plutonium as a tracer of soil and sediment movement in the Herbert River, Australia." Nuclear Instruments and Methods in Physics Research, Section B: Beam Interactions with Materials and Atoms 268(7-8): 1150-1154.

Trimble, S. W. (1983). "A sediment budget for Coon Creek basin in the Driftless Area, Wisconsin, 1853- 1977." American Journal of Science 283(5): 454-474.

Turner, R., R. Huggins, R. Smith and M. S. J. Warne (2013). Total suspended solids, nutrient and pesticide loads (2010-2011) for rivers that discharge to the Great Barrier Reef. Great Barrier Reef Catchment Loads Monitoring 2010 -2011. Brisbane, Department of Science, Information Technology, Innovation and the Arts.

Turner, R., R. Huggins, R. Wallace, R. Smith, S. Vardy and M. S. J. Warne (2012). Sediment, Nutrient and Pesticide Loads: Great Barrier Reef Catchment Loads Monitoring 2009-10. Brisbane, Department of Science, Information Technology, Innovation and the Arts.

UNESCO (2015). Decisions adopted by the World Heritage Committee at its 39th session (Bonn, 2015). Bonn, Germany, UNESCO, World Heritage Committee.

Wallace, R., R. Huggins, R. A. Smith, R. D. R. Turner, A. Garzon-Garcia and M. S. J. Warne (2015). Total suspended solidas, nutrient and pesticide loads (2012-2013) for rivers that discharge to the Great Barrier Reef - Great Barrier Reef Catchment Loads Monitoring Program (2012-2013. Brisbane.

138 Source Catchments modelling – Updated methodology and results

Waterhouse, J., P. Duncanson, S. Attard, R. Bartley, K. Bristow, J. Brodie, C. Coppo, A. Davis, C. Dougall, S. Lewis, M. Smith, M. Star, A. Roberts and S. Wilkinson (2016). Spatial prioritisation and extension of information to manage sediment, nutrient and pesticide runoff in the Burdekin NRM region.

Waters, D. and C. Carroll (2012). Great Barrier Reef Paddock and Catchment Modelling Approach and Quality Assurance Framework. Queensland, Department of Natural Resources and Mines.

Waters, D., C. Carroll, R. Ellis, G. L. McCloskey, L. Hateley, R. Packett, C. Dougall and B. Fentie (2013). Will we meet our reef water quality targets? Catchment modelling results. MODSIM 2013, 20th International Conference on Modelling and Simulation, Adelaide, South Australia, Modelling and Simulation Society of Australia and New Zealand.

Waters, D. and R. Packett (2007). Sediment and nutrient generation rates for Queensland rural catchments - an event monitoring program to improve water quality modelling. Proceedings of the 5th Australian Stream Management Conference. Australian Rivers: Making a Difference, Thurgoona, NSW, Charles Sturt University.

Waters, D. K., C. Carroll, R. Ellis, L. R. Hateley, G. L. McCloskey, R. Packett, C. Dougall and B. Fentie (2014). Modelling reductions of pollutant loads due to improved management practices in the Great Barrier Reef Catchments: Whole of GBR, Technical Report, Volume 1. Brisbane, Queensland Department of Natural Resources and Mines.

Wilkinson, S., F. Cook, R. Bartley, A. Henderson, R. Searle, T. Ellis, P. Jordan and A. Miller (2010). Specification for Sediment, Nutrient and Pesticide Generation and Transport Modules in Source Catchments. Canberra, eWater Cooperative Research Centre.

Wilkinson, S. N., G. J. Hancock, R. Bartley, A. A. Hawdon and R. J. Keen (2013). "Using sediment tracing to assess processes and spatial patterns of erosion in grazed rangelands, Burdekin River basin, Australia." Agriculture, Ecosystems and Environment 180: 90-102.

139 Source Catchments modelling – Updated methodology and results

Appendix 1 – Hydrology calibration results

Below is the objective function used to recalibrate the hydrology to improve the baseflow component:

푛 0.5 0.5 2 푛 0.5 0.5 2 ∣(∑Oi – ∑Si∣ ∣(Obf – Sbf)∣ SDEB= {[α∑ (Oi – Si ) + (1- α) ∑ (Ek – Sk ) ] × [1 + ]}×{ 1 + } 푖=1 푘=1 ∑Oi Obf

Where n is the number of calibration days, Oi is the observed flow on day i (Ml/d), Si is the simulated

th flow on day i (Ml/d), Ek and Sk are the k values of the observed and simulated flow series of the exceedance curve, α is a weighting factor. The term ‘1 + ∣(Obf – Sbf∣)’ calculates the simulation bias of Obf baseflow component. The designated PEST’s parameters optimisers search a broad range of parameter combinations to minimise the difference between the simulated daily flow, total flow volume and exceedance time series and also the difference between the simulated baseflow rate and the value calculated by Lyne-Hollick filter.

Cape York

Table 49 Cape York hydrology calibration results (1970–2014)

Total flow High flow Catchment Years of Gauge Gauge name R2 NSE volume volume area (km2) record^ bias bias Pascoe R. at 102102A 1,313 44 0.806 0.634 -5.0 -6.9 Garraway Creek Stewart R. at 104001A 470 44 0.752 0.504 1.2 0.5 Telegraph Road Hann R. at Sandy 105001B 984 44 0.839 0.678 0.3 0.4 Creek Normanby R. at 105101A 2,302 44 0.818 0.63 0.6 1.0 Battle Camp Laura R. at 105102A 1,316 44 0.723 0.468 1.3 0.4 Coalseam Creek Kennedy R. at 105103A 1,083 18 0.731 0.489 -0.2 -1.2 Fairlight Deighton R. at 105104A 590 18 0.794 0.596 -1.7 -1.8 Deighton East Normanby R. 105105A at Mulligan 297 44 0.774 0.537 2.2 1.3 Highway West Normanby R. 105106A 839 19 0.746 0.463 0.9 0.9 at Sellheim Normanby R. at 105107A 12,934 9 0.876 0.734 0.1 0.3 Kalpowar Crossing McIvor R. at 106001A 175 18 0.644 0.299 3.0 3.0 Elderslie Jeannie R. at 106002A 323 18 0.687 0.369 3.7 4.3 Wakooka Road Starcke R. at 106003A 192 18 0.69 0.383 4.3 3.5 Causeway Endeavour R. at 107001B 337 44 0.705 0.378 1.1 1.2 Flaggy Annan R. at Mt 107002A 373 21 0.787 0.561 0.5 1.1 Simon NSE (Nash Sutcliffe coefficient of Efficiency) ^ Years of record = number of years of flow data that was within the hydrology calibration period (1972–2014). (EOS) end-of-system. It refers to the furthest downstream gauge on a stream or river

140 Source Catchments modelling – Updated methodology and results

Wet Tropics

Table 50 Wet Tropics hydrology calibration results (1970–2014)

Total flow High flow Catchment Years of Gauge Gauge name R2 NSE volume volume area (km2) record^ bias bias Daintree River at 108002A 911 42 0.85 0.71 -0.1 -1.2 Bairds (EOS) Bloomfield River at 108003A# 264 42 0.86 0.73 0 -0.2 China camp Mossman River at 109001A# 106 41 0.85 0.62 -2 7.4 Mossman (EOS) South Mossman 109002B-C# River at Loco 54 20 0.9 0.8 -7.3 -4.6 Bridge Barron River at 110001A-D 1,945 42 0.95 0.91 0 -0.7 Myola (EOS) Barron River at 110003A 228 42 0.92 0.85 -0.1 -0.1 Picnic Crossing Flaggy Creek at 110011B 150 42 0.88 0.77 -3 -3.9 Recorder Barron River @ 110016A 1,434 13 0.94 0.86 1.8 7.4 Koah Barron River at 110020A 1,258 22 0.88 0.77 0 -0.4 Bilwon Mulgrave River at 111007A 520 42 0.91 0.83 0 -0.8 the Fisheries (EOS) Russell River at 111101A-D# 315 34 0.95 0.9 -1.5 -3.6 Buckland’s (EOS) Babinda Creek at 111102A-B# 82 16 0.9 0.81 -6.2 -4.5 Babinda North Johnstone 112001A* River at Tung Oil 936 42 0.75 0.52 0.5 1.1 (EOS) North Johnstone 112003A 165 42 0.96 0.92 -1 -1.7 River at Glen Allyn South Johnstone 112101A-B River at Upstream 401 42 0.94 0.87 -0.1 -0.9 Central Mill (EOS) Liverpool Creek at 112102A 78 42 0.9 0.81 0.7 1.2 Upper Japoonvale Cochable Creek at 113004A# 95 42 0.88 0.76 0 0.4 Powerline Tully River at 113006A 1,450 42 0.93 0.86 1 -0.6 Euramo (EOS) Murray River at 114001A Upper Murray 156 42 0.87 0.75 -3.7 -5.6 (EOS) Herbert River at 116001A-D 8,581 42 0.92 0.84 -4.6 -5.7 Ingham (EOS) Herbert River at 116004A-C 5,236 42 0.92 0.85 3.1 -0.8 Glen Eagle Herbert River at 116006A 7,440 42 0.93 0.87 -0.3 -2.5 Abergowrie College Gowrie Creek at 116008B# 124 42 0.78 0.61 -4.8 -6.9 Abergowrie Blencoe Creek at 116010A 226 42 0.88 0.77 -2.1 -3.3 Blencoe Falls Cameron Creek at 116012A 360 42 0.77 0.59 -7.7 -10.3 8.7 km

141 Source Catchments modelling – Updated methodology and results

Total flow High flow Catchment Years of Gauge Gauge name R2 NSE volume volume area (km2) record^ bias bias Milstream at Archer 116013A 308 42 0.92 0.84 0 0 Creek Wild River at Silver 116014A 591 42 0.92 0.83 0.4 0.1 Valley Blunder Creek at 116015A 127 42 0.85 0.7 -0.3 -0.1 Wooroora Rudd Creek at 116016A 1,450 42 0.88 0.75 -0.1 0.3 Gunnawarra Stone Creek at 116017A 157 42 0.89 0.79 0.1 0.4 Running Creek NSE (Nash Sutcliffe coefficient of Efficiency) * The flow from 112004A was added onto flow from 112001A. ^ Years of record = number of years of flow data that was within the hydrology calibration period (1972–2014). (EOS) end-of-system. It refers to the furthest downstream gauge on a stream or river #rainfall data for the sub-catchment draining to the gauge were scaled

Burdekin

Table 51 Burdekin hydrology calibration results (1970–2014)

High Total flow Catchment Years of flow Gauge Gauge name R2 NSE volume area (km2) record volume bias bias Black River at Bruce 117002A 256 41 0.85 0.7 0.0 -0.3 Highway Bluewater Creek at 117003A 86 41 0.85 0.7 -3.3 -3.9 Bluewater Bohle River at Mount 118001B 183 29 0.87 0.7 31.7 28.1 Bohle Bohle River at Hervey 118003A 143 29 0.86 0.7 -4.0 -4.3 Range Road Little Bohle River at 118004A Middle Bohle River 54 24 0.87 0.7 10.8 10.5 junction. Haughton River at 119003A 1,773 42 0.87 0.7 8.9 12.0 Powerline Haughton River at 119005A 1,133 42 0.90 0.8 0.0 -0.1 Mount Piccaninny Major Creek at 119006A 468 36 0.79 0.6 -2.8 -3.3 Damsite Barratta Creek at 119101A 753 40 0.83 0.7 0.0 0.0 Northcote Burdekin River at 120002C 36,260 42 0.96 0.9 0.0 0.0 Sellheim Burdekin River at 120004A 114,654 11 0.96 0.9 -3.7 -4.9 Burdekin Falls D/S Bogie River at 120005B 1,031 23 0.88 0.8 -0.1 -0.6 Strathbogie Burdekin River at 120006B 129,876 24 0.97 0.9 -1.1 -2.7 Clare Keelbottom Creek at 120102A 193 42 0.78 0.6 -4.9 -5.2 Keelbottom Basalt River at Bluff 120106B 1,301 42 0.88 0.8 0.0 0.3 Downs Burdekin River at Blue 120107B 10,528 42 0.89 0.8 0.1 -1.5 Range Fletcher Creek at 120108C 1,034 17 0.92 0.8 0.0 1.7 Fletchervale Burdekin River at 120110A 17,299 42 0.95 0.9 0.0 -0.6 Mount Fullstop

142 Source Catchments modelling – Updated methodology and results

High Total flow Catchment Years of flow Gauge Gauge name R2 NSE volume area (km2) record volume bias bias Burdekin River at 120111A 6,277 17 0.79 0.6 0.1 -0.7 Lucky Downs 120112A Star River at Laroona 1,212 42 0.82 0.7 -2.8 -3.2 Clarke River at 120113A 1,802 22 0.84 0.7 0.4 -0.7 Wandovale Douglas Creek at 120114A 663 17 0.91 0.8 0.4 1.1 Kangaroo Hills Burdekin River at 120118A 1,087 17 0.82 0.6 -0.4 -0.1 Lake Lucy Fanning River at 120119A 498 19 0.59 0.3 -3.6 -4.1 Fanning River Running River at Mt. 120120A 490 39 0.85 0.7 -0.6 -0.8 Bradley Burdekin River at 120121A 2,216 38 0.86 0.7 -3.0 -4.8 Lake Lucy Dam Site Burdekin River at 120122A 26,316 10 0.94 0.9 0.9 -1.9 Gainsford Bowen River at 120205A 7,104 42 0.90 0.8 0.0 -0.9 Myuna Pelican Creek at Mt 120206A 545 34 0.81 0.6 -0.1 -0.4 Jimmy Broken River at 120207A 1,103 42 0.83 0.6 5.4 7.2 Urannah Bowen River at Jacks 120209A 4,305 42 0.88 0.7 3.2 5.7 Creek Emu Creek at The 120212A 431 17 0.69 0.4 -0.1 0.6 Saddle Grant Creek at Grass 120213A 325 17 0.72 0.5 0.0 0.8 Humpy Broken River at Mt. 120214A 2,269 22 0.89 0.7 1.0 0.9 Sugarloaf Bowen River at Red 120219A 8,280 14 0.90 0.8 -0.2 -3.2 Hill Creek Belyando River at 120301B Gregory Development 35,411 38 0.87 0.7 -1.1 -0.2 Rd. Cape River at 120302B 16,074 42 0.90 0.8 0.0 0.4 Taemas Suttor River at St 120303A 50,291 42 0.91 0.8 -0.2 -0.7 Anns Suttor River at 120304A 1,915 42 0.70 0.4 0.0 -0.1 Eaglefield Native Companion 120305A 4,065 42 0.89 0.8 -0.5 -0.5 Creek at Violet Grove Mistake Creek at 120306A 2,583 22 0.78 0.6 0.0 -0.5 Charlton Suttor River at Bowen 120310A 10,758 8 0.84 0.7 0.0 0.0 Developmental Road Don River at Ida 121001A 604 42 0.80 0.6 0.0 0.4 Creek Elliot River at 121002A 273 41 0.86 0.7 -4.8 -4.8 Guthalungra 121003A Don River at Reeves 1,016 30 0.87 0.7 0.1 0.6 Euri Creek at 121004A 429 16 0.82 0.7 -5.7 -6.3 Koonandah NSE (Nash Sutcliffe coefficient of Efficiency) ^ Years of record = number of years of flow data that was within the hydrology calibration period (1970–2014). (EOS) end-of-system. It refers to the furthest downstream gauge on a stream or river

143 Source Catchments modelling – Updated methodology and results

Mackay Whitsunday

Table 52 Mackay Whitsunday hydrology calibration results (1970–2014)

Total flow High flow Catchment Years of R Gauge Gauge name NSE volume volume area (km2) record^ Squared bias bias Lethe Brook at 122012A C 89 13 0.8 0.63 -25.1 -26.5 Hadlow Road O’Connell River at 124001B 342 44 0.9 0.8 -1.9 -2.1 Staffords Crossing St.Helens Creek at 124002A 118 41 0.92 0.85 -2.2 -3.2 Calen Andromache River 124003A 230 38 0.91 0.81 0 0 at Jochheims Pioneer River at 125001B C Pleystowe 1434 12 0.94 0.88 -0.1 -0.4 Recorder Pioneer River at 125002C 757 44 0.91 0.82 -1.5 -1.9 Sarichs Cattle Creek at 125004B 326 44 0.94 0.89 -0.7 -0.9 Gargett Sandy Creek at 126001A 326 44 0.88 0.77 0 0.3 Homebush (EOS) Plane Creek at 126002A C 92 16 0.85 0.68 1 1.2 Sarina Carmila Creek at 126003A 84 41 0.79 0.62 -4.4 -4.7 Carmila NSE (Nash Sutcliffe coefficient of Efficiency). C closed/historic gauging station. ^ Years of record = approximate number of years of flow data that was within the hydrology calibration period (1970–2014). (EOS) end-of-system. It refers to the furthest downstream gauge on a stream or river.

Figure 54 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in Lethe Brook Creek at Hadlow Road (122012A)

144 Source Catchments modelling – Updated methodology and results

Figure 55 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in the O’Connell River at Staffords Crossing (124001B)

Figure 56 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in St.Helens Creek at Calen (124002A)

Figure 57 Time Series, Flow Duration (log scale), and Scatter Plot Comparisons of Modelled VS. Measured Daily Flows in the Andromache River at Jochheims (124003A)

145 Source Catchments modelling – Updated methodology and results

Figure 58 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in the Pioneer River at Pleystowe Recorder (125001B)

Figure 59 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in the Pioneer River at Sarichs (125002C)

Figure 60 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in Cattle Creek at Gargett (125004B)

146 Source Catchments modelling – Updated methodology and results

Figure 61 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in Sandy Creek at Homebush (126001A)

Figure 62 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in Plane Creek at Sarina (126002A)

Figure 63 Time Series, Flow Duration (log scale), and Scatter plot Comparisons of Modelled VS. Measured Daily Flows in Carmila Creek at Carmila (126003A)

147 Source Catchments modelling – Updated methodology and results

Fitzroy

Table 53 Fitzroy hydrology calibration results (1970–2014)

Total flow High flow Catchment Years of R Gauge Gauge name NSE volume volume area (km2) record^ Squared bias bias Waterpark Creek at 129001A 213 44 0.88 0.78 0.0 -0.1 Byfield Fitzroy River at 130003B 131,413 44 0.94 0.89 0.3 -0.7 Riverslea Raglan Creek at 130004A 389 44 0.80 0.58 0.0 0.3 Old Station Fitzroy River at The 130005A 135,788 44 0.93 0.85 0.9 -2.4 Gap Neerkol Creek at 130008A 505 25 0.91 0.83 -1.3 -1.7 Neerkol Marlborough Creek 130009A 679 16 0.95 0.90 -2.0 -2.9 at Slopeaway Mackenzie River at 130103A 45,367 29 0.90 0.72 9.7 12.2 Carnangarra Mackenzie River at 130105A 76,659 43 0.94 0.88 -4.3 -5.5 Coolmaringa Mackenzie River at 130106A 50,851 43 0.88 0.77 0.0 -0.1 Bingegang Blackwater Creek 130108B 776 37 0.82 0.64 -0.1 0.3 at Curragh Mackenzie River at 130109A Tartrus Weir 75,182 26 0.94 0.82 11.9 12.1 Headwater Theresa Creek at 130206A 8,485 44 0.90 0.81 9.2 9.2 Gregory Highway Theresa Creek at 130208A 758 35 0.83 0.66 0.1 0.3 Ellendale Nogoa River at 130209A 13,871 42 0.92 0.84 0.0 -2.4 Craigmore Theresa Creek at 130210A 4,421 43 0.81 0.60 0.0 -0.8 Valeria Nogoa River at 130219A 27,146 21 0.92 0.84 2.2 1.9 Duck Ponds Dawson River at 130302A 15,847 44 0.84 0.70 -0.1 -0.8 Taroom Dawson River at 130305A 27,332 32 0.89 0.74 0.0 2.7 Theodore Don River at 130306B 6,806 44 0.93 0.87 -1.1 -1.1 Rannes Recorder Palm Tree Creek at 130313A 2,660 44 0.77 0.55 4.3 4.8 La Palma Callide Creek at 130315C 547 32 0.82 0.67 -0.1 0.5 Stepanoffs Mimosa Creek at 130316A 2,487 44 0.76 0.52 -2.8 -0.8 Redcliffe Dawson River at 130317B 28,503 44 0.93 0.86 6.3 6.9 Woodleigh Dawson River at 130322A 40,501 44 0.91 0.80 0.2 0.4 Beckers Dawson River at 130324A 6,039 44 0.88 0.77 -0.2 -0.6 Utopia Downs Callide Creek at 130327A 4,468 43 0.90 0.74 1.6 1.6 Goovigen Juandah Creek at 130344A 1,679 40 0.82 0.65 0.0 0.1 Windamere Don River at 130349A 594 38 0.85 0.69 0.0 1.2 Kingsborough

148 Source Catchments modelling – Updated methodology and results

Total flow High flow Catchment Years of R Gauge Gauge name NSE volume volume area (km2) record^ Squared bias bias Isaac River at 130401A 19,736 44 0.92 0.84 0.0 -0.2 Yatton Connors River at 130403A 1,292 44 0.86 0.74 -6.6 -7.3 Mount Bridget Connors River at 130404A 8,741 44 0.92 0.85 0.1 -1.6 Pink Lagoon Funnel Creek at 130406A 1,065 44 0.95 0.91 -0.1 -0.3 Main Road Nebo Creek at 130407A 257 44 0.89 0.79 -0.5 0.3 Nebo Isaac River at 130410A 4,091 44 0.82 0.65 -0.7 -1.0 Deverill Denison Creek at 130413A 757 43 0.88 0.77 -0.3 -0.3 Braeside Isaac River at 130414A 1,213 31 0.72 0.45 0.0 -0.3 Goonyella Comet River at 130504B 16,423 43 0.87 0.72 0.6 0.6 Comet Weir Comet River at The 130506A 10,188 42 0.89 0.79 4.6 5.5 Lake Carnarvon Creek at 130509A 351 29 0.82 0.64 0.0 0.6 Rewan Calliope River at 132001A 1,288 44 0.91 0.83 0.0 0.1 Castlehope Calliope River at 132002A 165 18 0.90 0.79 2.1 2.7 Mount Alma Boyne River at 133001B 2,258 18 0.77 0.59 -1.7 -3.9 Riverbend Diglum Creek at 133003A 203 42 0.88 0.76 -0.1 0.3 Marlua Boyne River at 133004A 1,170 21 0.93 0.86 -0.8 -1.2 Milton NSE (Nash Sutcliffe coefficient of Efficiency) ^ Years of record = number of years of flow data that was within the hydrology calibration period (1970–2014). (EOS) end-of-system. It refers to the furthest downstream gauge on a stream or river

Table 54 Baseflow proportions (%) for the current and previous Report Card for all gauges in the Fitzroy Basin

RC 2015 RC 2014 LH Gauge Basin baseflow baseflow baseflow (%) (%) filter (%) 129001A Waterpark Creek 29 47 27 130003B Fitzroy 29 82 25 130004A Fitzroy 17 10 19 130005A Fitzroy 39 67 26 130008A Fitzroy 10 2 12 130009A Fitzroy 10 10 8 130103A Fitzroy 18 6 21 130105A Fitzroy 28 80 24 130106A Fitzroy 18 97 19 130108B Fitzroy 15 15 5 130109A Fitzroy 23 16 24 130206A Fitzroy 10 34 8 130208A Fitzroy 13 13 7 130209A Fitzroy 13 90 11 130210A Fitzroy 10 99 8

149 Source Catchments modelling – Updated methodology and results

RC 2015 RC 2014 LH Gauge Basin baseflow baseflow baseflow (%) (%) filter (%) 130219A Fitzroy 23 41 22 130302A Fitzroy 15 25 13 130305A Fitzroy 17 71 14 130306B Fitzroy 13 23 11 130313A Fitzroy 17 92 12 130315C Fitzroy 15 27 14 130316A Fitzroy 21 74 19 130317B Fitzroy 20 74 18 130322A Fitzroy 24 89 19 130324A Fitzroy 20 88 15 130327A Fitzroy 12 69 9 130344A Fitzroy 4 4 4 130349A Fitzroy 12 12 10 130401A Fitzroy 23 99 18 130403A Fitzroy 18 26 14 130404A Fitzroy 21 98 17 130406A Fitzroy 15 8 20 130407A Fitzroy 22 49 20 130410A Fitzroy 11 13 9 130413A Fitzroy 15 15 17 130414A Fitzroy 17 41 12 130504B Fitzroy 20 41 20 130506A Fitzroy 23 98 20 130509A Fitzroy 19 19 18 132001A Calliope 13 13 14 132002A Fitzroy 30 30 27 133001B Boyne 25 25 20 133003A Boyne 29 29 23 133004A Boyne 22 22 24

150 Source Catchments modelling – Updated methodology and results

Burnett Mary

Table 55 Burnett Mary hydrology calibration results (1970–2014)

Total High Years Catchment R flow flow Gauge Gauge name of NSE area (km2) Squared volume volume record^ bias bias

134001B Baffle Creek at Mimdale 1,401 46 0.854 0.725 -0.4 -0.2 135002A Kolan River at Springfield 548 49 0.915 0.81 9.1 8.9 135004A Gin Gin Creek at Brushy Creek 537 49 0.836 0.695 -0.2 -0.6 136002D Burnett River at Mount Lawless 2,130 39 0.945 0.892 2.5 3.6 136007A Burnett River at Figtree Creek 1,318 18 0.824 0.675 1.9 1.9 136011A Degilbo Creek at Coringa 686 28 0.929 0.838 0.2 0.3 136019A Perry River at Mt Rawdon 155 19 0.931 0.833 -2.4 -2.0 136094A Burnett River at Jones Weir Tailwater 2,817 33 0.907 0.822 0.9 2.1 136101C Three Moon Creek at Abercorn 1,537 50 0.896 0.787 -0.3 -1.8 136103B Burnett River at Ceratodus 2,349 54 0.964 0.925 -0.1 -0.1 136106A Burnett River at Eidsvold 3,208 54 0.948 0.858 -20.3 -15.1 136207A Barambah Creek at Ban Ban 4,154 48 0.809 0.652 -0.3 -0.8 136209A Barker Creek at Glenmore 1,366 27 0.914 0.827 -0.1 0.1 136304A Stuart River at Proston Rifle Range 1,555 49 0.911 0.811 -0.3 0.7 136305A Auburn River at Dykehead 5,287 49 0.878 0.748 0.1 0.2 136306A Cadarga Creek at Brovinia Station 1,288 49 0.639 0.29 1.0 1.2 136315A Boyne River at Carters 1,619 35 0.92 0.841 1.8 1.6 136319A Boyne River at Cooranga 2,103 20 0.846 0.705 -5.3 -7.1 137001B Elliott River at Elliott 232 53 0.829 0.668 -0.1 -0.2 137003A Elliott River at Dr Mays Crossing 25 41 0.752 0.408 4.2 15.8 137101A Gregory River at Isis Highway 441 49 0.801 0.619 -1.2 -1.1 137103A Gregory River at Leesons 161 8 0.807 0.618 3.5 4.5 137201A Isis River at Bruce Highway 457 49 0.819 0.647 0.0 -0.1 138001A Mary River at Miva 1,713 105 0.833 0.674 1.9 0.1 138004B Munna Creek at Marodian 1,190 40 0.941 0.883 -0.4 -0.6 138007A Mary River at Fishermans Pocket 2,246 46 0.872 0.76 -0.2 1.7 138014A Mary River at Home Park 897 32 0.903 0.811 -9.6 -8.3 138111A Mary River at Moy Pocket 813 51 0.854 0.712 0.6 1.3 138903A Tinana Creek at Bauple East 756 33 0.941 0.885 -1.3 -1.6 NSE (Nash Sutcliffe coefficient of Efficiency) ^ Years of record = number of years of flow data that was within the hydrology calibration period (1972–2014).

151 Source Catchments modelling – Updated methodology and results

Table 56 Baseflow proportions (%) for the current and previous Report Card for all gauges in the Burnett Mary Basin

Lyne- RC 2014 RC 2015 Hollick Gauge Gauge name baseflow baseflow baseflow (%) (%) filter (%)

134001B Baffle Creek at Mimdale 24 21 18 135002A Kolan River at Springfield 23 18 15 135004A Gin Gin Creek at Brushy Creek 22 19 17 136002D Burnett River at Mount Lawless 78 23 21 136007A Burnett River at Figtree Creek 23 19 18 136011A Degilbo Creek at Coringa 52 17 14 136019A Perry River at Mt Rawdon 23 17 13 136094A Burnett River at Jones Weir Tailwater 86 21 18 136101C Three Moon Creek at Abercorn 21 16 15 136103B Burnett River at Ceratodus 9 12 13 136106A Burnett River at Eidsvold 9 19 21 136207A Barambah Creek at Ban Ban 59 21 19 136209A Barker Creek at Glenmore 96 23 18 136304A Stuart River at Proston Rifle Range 12 15 13 136305A Auburn River at Dykehead 48 15 10 136306A Cadarga Creek at Brovinia Station 56 12 8 136315A Boyne River at Carters 61 15 11 136319A Boyne River at Cooranga 0 14 16 137001B Elliott River at Elliott 51 20 18 137003A Elliott River at Dr Mays Crossing 81 24 29 137101A Gregory River at Isis Highway 88 15 12 137103A Gregory River at Leesons 33 21 18 137201A Isis River at Bruce Highway 1 12 10 138001A Mary River at Miva 60 27 25 138004B Munna Creek at Marodian 31 18 16 138007A Mary River at Fishermans Pocket 50 29 27 138014A Mary River at Home Park 1 22 23 138111A Mary River at Moy Pocket 18 27 29 138903A Tinana Creek at Bauple East 38 23 22

152 Source Catchments modelling – Updated methodology and results

Appendix 2 – Dynamic SedNet global parameters and global requirements

Grazing constituent generation

Table 57 Hillslope erosion parameters

Parameter CY WT Burdekin MW Fitzroy BM TSS delivery ratio 5 15/20/30 5/10 15/100 5/10 5 Max. QF 2,000 1,000 10,000/5,000/2,500 1,000 10,000 10,000 concentration DWC (mg/L) 5 20 0 10 0 0

Table 58 Gully erosion parameters

Parameter CY WT Burdekin MW Fitzroy BM Daily runoff power factor 1.4 1.4 1.4 1.4 1.4 1.4 Gully model type DERM DERM DERM DERM DERM DERM TSS DR % 100 100 100/25 100 100 80 Gully cross-sectional area (m2) 7.5 10 16/8 5 5 10 Average gully activity factor 1 1 1 1 0.7 1 Management practice factor Variable Variable Variable Variable Variable Variable Default gully start year 1870 1870 1870 1861 1870 1900 Gully full maturity year 2010 2014 2014 2010 1980 2007 Density raster year 2001 2001 2003 2001 2013 2003

Table 59 Particulate nutrient generation parameter values

Parameter CY WT Burdekin MW Fitzroy BM PP/PN enrichment ratio 8/8 3/2 2/1.2 4/2.4 10/6 10/9 Hillslope DR % 5 20 10/20 20 20 5 Gully DR % 100 100 100/200 100 100 80

In-stream models

Table 60 Parameter values for in-stream models in RC2015 baseline models

Parameter CY WT Burdekin MW Fitzroy BM Bank height proportion for fine 0.05 0.05 0 0.2 0 0.05 sediment deposition (%) Initial proportion of fine bed store 0.01 0.01 0 0 and 0.25 0 0.01 (%) Floodplain sediment settling 0.002 0.0003 0.000001 0.000001 0.0001 0.000001 velocity Sediment settling velocity for 0.5 0.1 0.1 0.1 0.1 0.1 remobilisation Sediment settling velocity for 0.002 0.0003 0.000001 0.000001 0.0001 0.00005 instream deposition

153

Sugarcane and cropping constituent generation Table 61 A summary of changes for RC2015 changes for cropping and sugarcane (Changes from 2014 Report Card)

Change CY WT Burd MW Fitz BM Possible effect Updated rainfall To 2015 To 2014 To 2015 To 2014 To 2015 To 2014 Dryland grains cropping - HowLeaky Area 20 km2 (<1%) 142 (0.7%) 1336 (0.9%) 3 km2 (<1) 7934 km2 (5%) 820 km2 (1.6%) Cape York specific No change No change No change No change No change Additional pesticides in Cape Management farming system York due to new farming scenarios modelled including systems “maize and peanuts” Irrigated grains cropping - HowLeaky Area 31 km2 (<1%) 8 km2 (<0.05%) 70 km2 (<0.0%) <1 km2 (<1%) 1213 km2 (1%) 466 km2 (0.9%)

No change No change No change No change “set to zero” plug in for No change No discharge from irrigated land Management the Fitzroy (due to uses for the Fitzroy scenario water recycling). Bananas - HowLeaky Row/Inter-row and Row/Inter-row and ------More consistent inter-annual Model set up Fallow modelled Fallow modelled erosion estimates – removes separately separately “fallow” peak. Sugarcane - APSIM --- Run of ‘Df’ soils Heavily revised No change --- Run of ‘Bp’ nutrient Management management to irrigation management to scenarios include additional management change fertiliser tillage operations scenarios applied --- Revised allocation of Reduced number No change --- Soils simulations to mapped of soils modelled. soils. Irrigation modelling --- No change Heavily revised No change --- No change Change in runoff, drainage, DIN Drainage modelling --- No change Heavily revised No change --- No change Change in runoff, drainage, DIN

Other land uses (EMC/DWC values) Although not all FUs have an EMC/DWC model applied, there are some constituents that use this approach, as well as those FUs that have an EMC/DWC model applied for all constituents (commonly, urban, other, and horticulture with some exceptions). During the development of the appropriate EMC/DWC

Source Catchments modelling – Updated methodology and results values all land uses were considered and a concentration value determined. This was important for the predevelopment model as nature conservation values were applied to all FUs, excluding those where the USLE approach was taken in the baseline model, and maintained in the predevelopment model (commonly, open and closed grazing, forestry, and nature conservation areas with some exceptions).

Table 62 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Cape York

TSS PN DIN DON PP DOP DIP FU TSS EMC PN EMC DIN EMC DON EMC PP EMC DOP EMC DIP EMC DWC DWC DWC DWC DWC DWC DWC Bananas HowLeaky 30 HowLeaky 0.1125 HowLeaky 0.03 HowLeaky 0.225 HowLeaky 0.02625 HowLeaky 0.015 HowLeaky 0.0075 Dryland cropping HowLeaky 30 HowLeaky 0.1125 HowLeaky 0.03 0.45 0.225 HowLeaky 0.02625 HowLeaky 0.015 HowLeaky 0.0075 Forestry USLE 20 USLE 0.075 USLE 0.02 0.3 0.15 USLE 0.0175 0.02 0.01 0.01 0.005 Horticulture 60 30 0.225 0.1125 0.06 0.03 0.45 0.225 0.0525 0.02625 0.03 0.015 0.015 0.0075 Irrigated cropping HowLeaky 30 HowLeaky 0.1125 HowLeaky 0.03 0.45 0.225 HowLeaky 0.02625 HowLeaky 0.015 HowLeaky 0.0075 Nature USLE 10 USLE 0.04 USLE 0.016 0.2 0.1 USLE 0.005 0.01 0.005 0.005 0.0025 conservation Other 40 20 0.15 0.075 0.04 0.02 0.3 0.15 0.0525 0.0175 0.02 0.01 0.01 0.005 Urban 40 20 0.15 0.075 0.04 0.02 0.3 0.15 0.0525 0.0175 0.02 0.01 0.01 0.005 Water 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Grazing closed USLE 0 USLE 0.06 USLE 0.016 0.24 0.15 USLE 0.014 0.016 0.008 0.008 0.004 Grazing open USLE 0 USLE 0.075 USLE 0.025 0.3 0.15 USLE 0.0175 0.02 0.01 0.01 0.005

Table 63 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Wet Tropics

TSS PN DIN DON PP PP DOP DIP FU TSS EMC PN EMC DIN EMC DON EMC DOP EMC DIP EMC DWC DWC DWC DWC EMC DWC DWC DWC Nature conservation 30 15 0.25 0.13 0.03 0.02 0.16 0.08 0.12 0.06 0.02 0.01 0.00 0.00 Forestry 40 20 0.25 0.13 0.03 0.02 0.20 0.10 0.72 0.36 0.02 0.01 0.01 0.01 Closed grazing USLE 20 USLE 0.48 USLE 0.02 0.16 0.08 0.30 0.15 0.03 0.01 0.02 0.01 Open grazing USLE 20 USLE 0.48 USLE 0.02 0.16 0.08 0.30 0.15 0.03 0.01 0.02 0.01 Dairy 150 75 0.48 0.24 0.04 0.02 0.19 0.09 0.12 0.06 0.03 0.01 0.02 0.01 Other 105 35 0.48 0.24 0.04 0.02 0.76 0.38 0.27 0.14 0.03 0.01 0.04 0.02 Urban 105 35 0.48 0.24 0.04 0.02 0.76 0.38 0.27 0.14 0.03 0.01 0.04 0.02

155 Source Catchments modelling – Updated methodology and results

TSS PN DIN DON PP PP DOP DIP FU TSS EMC PN EMC DIN EMC DON EMC DOP EMC DIP EMC DWC DWC DWC DWC EMC DWC DWC DWC Horticulture 120 60 0.50 0.25 0.10 0.05 0.41 0.21 0.21 0.11 0.03 0.02 0.05 0.02 Dryland cropping HowLeaky 60 HowLeaky 0.30 HowLeaky 0.10 0.75 0.38 0.50 0.25 HowLeaky 0.03 HowLeaky 0.01 Irrigated cropping HowLeaky 60 HowLeaky 0.30 HowLeaky 0.10 0.75 0.38 0.50 0.25 HowLeaky 0.03 HowLeaky 0.01 Banana HowLeaky 60 HowLeaky 0.55 HowLeaky 0.06 HowLeaky 1.50 0.21 0.11 HowLeaky 0.03 HowLeaky 0.02 Sugarcane APSIM 50 APSIM 0.30 APSIM 0.02 APSIM 0.00 0.50 0.30 APSIM N/A APSIM N/A Water 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Table 64 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Burdekin

TSS PN DIN DON PP PP DOP DIP FU TSS EMC PN EMC DIN EMC DON EMC DOP EMC DIP EMC DWC DWC DWC DWC EMC DWC DWC DWC Conservation USLE 0 USLE 0 USLE 0 0.1 0.1 0.2 0.2 0.01 0.01 0.03 0.03 Forestry USLE 0 USLE 0 USLE 0 0.1 0.1 0.2 0.2 0.01 0.01 0.03 0.03 Grazing Forested USLE 0 USLE 0 USLE 0 0.1 0.1 0.2 0.2 0.01 0.01 0.03 0.03 Grazing Open USLE 0 USLE 0 USLE 0 0.1 0.1 0.2 0.2 0.01 0.01 0.03 0.03 Horticulture 80 0 USLE 0 USLE 0 0.74 0.74 0.25 0.25 0.02 0.02 0.01 0.01 Dryland Cropping HowLeaky 0 HowLeaky 0 HowLeaky 0 0.5 0.5 0.37 0.37 HowLeaky 0 HowLeaky 0 Irrigated Cropping HowLeaky 0 HowLeaky 0 HowLeaky 0 0.5 0.5 0.37 0.37 HowLeaky 0 HowLeaky 0 Other 80 0 USLE 0 USLE 0 0.1 0.1 0.2 0.2 0.01 0.01 0.03 0.03 Sugarcane APSIM 0 APSIM 0 APSIM 0 APSIM 0 0.25 0.25 APSIM N/A APSIM N/A Urban 80 0 USLE 0 USLE 0 0.16 0.16 0.26 0.26 0.02 0.02 0.01 0.01 Water 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Table 65 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Mackay Whitsunday

TSS PN DIN DON PP PP DOP DIP FU TSS EMC PN EMC DIN EMC DON EMC DOP EMC DIP EMC DWC DWC DWC DWC EMC DWC DWC DWC Nature conservation 50 10 0.15 0.075 0.07 0.035 0.058 0.029 0.092 0.046 0.003 0.002 0.011 0.005 Forestry 70 10 0.4 0.2 0.1 0.05 0.077 0.039 0.122 0.0061 0.004 0.002 0.015 0.007 Closed grazing USLE 10 USLE 0 USLE 0 0.097 0.048 0.153 0.076 0.005 0.003 0.018 0.009 Open grazing USLE 10 USLE 0 USLE 0 0.193 0.097 0.306 0.153 0.01 0.005 0.036 0.018

156 Source Catchments modelling – Updated methodology and results

TSS PN DIN DON PP PP DOP DIP FU TSS EMC PN EMC DIN EMC DON EMC DOP EMC DIP EMC DWC DWC DWC DWC EMC DWC DWC DWC Other 109 55 0.448 0.224 0.127 0.064 0.193 0.097 0.306 0.153 0.01 0.005 0.036 0.018 Urban 218 109 0.895 0.448 0.255 0.127 0.386 0.193 0.611 0.306 0.021 0.01 0.073 0.036 Horticulture 109 55 0.448 0.224 0.127 0.064 0.193 0.097 0.306 0.153 0.01 0.005 0.036 0.018 Dryland cropping HowLeaky 0 HowLeaky 0 HowLeaky 0 0.58 0.29 0.917 0.458 Howleaky 0 Howleaky 0 Irrigated cropping HowLeaky 0 HowLeaky 0 HowLeaky 0 0.483 0.242 0.764 0.382 Howleaky 0 Howleaky 0 Sugarcane APSIM 10 APSIM 0.56 APSIM 0.159 APSIM 0 0.65 0.33 APSIM N/A APSIM N/A Water 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Table 66 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Fitzroy

TSS PN DIN DON PP PP DOP DIP FU TSS EMC PN EMC DIN EMC DON EMC DOP EMC DIP EMC DWC DWC DWC DWC EMC DWC DWC DWC Nature conservation USLE 0 USLE 0 USLE 0 0.12 0.06 0.246 0.123 0.019 0.01 0.077 0.038 Forestry USLE 0 USLE 0 USLE 0 0.15 0.075 0.492 0.246 0.038 0.019 0.154 0.077 Closed grazing USLE 0 USLE 0 USLE 0 0.12 0.06 0.394 0.197 0.03 0.015 0.123 0.061 Open grazing USLE 0 USLE 0 USLE 0 0.15 0.075 0.492 0.246 0.038 0.019 0.154 0.077 Other 80 40 USLE 0 USLE 0 0.15 0.075 0.492 0.246 0.038 0.019 0.154 0.077 Urban 80 40 USLE 0 USLE 0 0.15 0.075 0.492 0.246 0.038 0.019 0.154 0.077 Horticulture 120 60 USLE 0 USLE 0 0.225 0.112 0.738 0.369 0.057 0.029 0.23 0.115 Dryland cropping HowLeaky 0 HowLeaky 0 HowLeaky 0 0.225 0.112 0.738 0.369 HowLeaky N/A HowLeaky N/A Irrigated cropping 0 0 0 0 0 0 0 0 0 0 0 N/A 0 N/A Sugarcane 140 70 USLE 0 USLE 0 0.225 0.112 0.738 0.369 0.057 0.029 0.23 0.115 Water 0 0 0 0 0 0 0.01 0.005 0.025 0.012 0.001 0.001 0.003 0.002

Table 67 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for Burnett Mary

TSS PN DIN DON PP PP DOP DIP FU TSS EMC PN EMC DIN EMC DON EMC DOP EMC DIP EMC DWC DWC DWC DWC EMC DWC DWC DWC Conservation USLE 0 USLE 0 USLE 0 0.038 0.019 0.166 0.083 0.007 0.003 0.011 0.005 Forestry USLE 0 USLE 0 USLE 0 0.051 0.025 0.222 0.111 0.009 0.004 0.014 0.007 Grazing Forested USLE 0 USLE 0 USLE 0 0.064 0.073 0.270 0.227 0.011 0.005 0.018 0.004

157 Source Catchments modelling – Updated methodology and results

TSS PN DIN DON PP PP DOP DIP FU TSS EMC PN EMC DIN EMC DON EMC DOP EMC DIP EMC DWC DWC DWC DWC EMC DWC DWC DWC Grazing Open USLE 0 USLE 0 USLE 0 0.127 0.064 0.554 0.277 0.022 0.011 0.036 0.018 Other 110 55 0.521 0.260 0.156 0.078 0.127 0.064 0.554 0.277 0.022 0.011 0.036 0.018 Urban 220 110 1.041 0.521 0.312 0.156 0.254 0.127 1.108 0.554 0.044 0.022 0.072 0.036 Horticulture 275 137.5 1.302 0.651 0.390 0.195 0.318 0.159 1.385 0.693 0.055 0.028 0.090 0.045 Dryland Cropping HowLeaky 165 HowLeaky 0.781 HowLeaky 0.234 0.381 0.191 1.662 0.831 HowLeaky 0.033 HowLeaky 0.054 Irrigated Cropping HowLeaky 137.5 HowLeaky 0.651 HowLeaky 0.195 0.318 0.159 1.385 0.693 HowLeaky 0.028 HowLeaky 0.045 Sugarcane APSIM 0 APSIM 0 APSIM 0 APSIM 0 2 0 APSIM N/A APSIM N/A Water 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Table 68 Event Mean Concentrations/Dry Weather Concentrations (mg/L) for predevelopment model

TSS PN Basin TSS DWC PN EMC PP EMC PP DWC DIN EMC DIN DWC DON EMC DON DWC DOP EMC DOP DWC DIP EMC DIP DWC EMC DWC Cape York USLE/20 USLE/10 USLE/0.08 0.04 USLE/0.032 0.016 0.2 0.1 0.01 0.005 0.01 0.005 0.005 0.0025 Wet 30 15 0.25 0.13 0.03 0.015 0.160 0.080 0.119 0.060 0.020 0.010 0.003 0.003 Tropics Burdekin USLE 0 USLE 0 USLE 0 0.100 0.100 0.200 0.200 0.010 0.010 0.030 0.030 Mackay 50 10 0.15 0.075 0.07 0.035 0.058 0.029 0.092 0.046 0.003 0.002 0.011 0.005 Whitsunday Fitzroy USLE/40 20 USLE 0 USLE 0 0.120 0.060 0.246 0.123 0.190 0.010 0.077 0.038 Burnett Mary USLE 35 USLE 0 USLE 0 0.038 0.019 0.166 0.083 0.007 0.003 0.011 0.005

158

Appendix 3 – Report Card 2015 modelling results

Table 69 Constituent loads for predevelopment, baseline and 2015 Report Card change model runs for the CY NRM Region

Total fine sediment (kt/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Jacky Jacky 9 52 43 4.78 52 0.00 Olive 18 72 54 3.00 72 0.00 Pascoe Lockhart 13 67 54 4.15 67 0.00 Stewart 8 49 41 5.13 49 0.00 Normanby 35 186 151 4.31 186 0.01 Jeannie 11 42 31 2.82 42 0.00 Endeavour 31 59 27 0.87 59 0.00 TOTAL 126 526 400 3.17 526 0.00 Total phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Jacky Jacky 38 77 38 1.00 77 0.00 Olive 61 126 65 1.07 126 0.00 Pascoe Lockhart 38 118 80 2.11 118 0.00 Stewart 21 60 39 1.86 60 0.00 Normanby 69 142 73 1.06 142 0.00 Jeannie 27 59 31 1.15 59 0.00 Endeavour 65 101 36 0.55 101 0.00 TOTAL 320 682 361 1.13 682 0.00 Particulate phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Jacky Jacky 7 44 37 5.29 44 0.00 Olive 16 75 59 3.69 75 0.00 Pascoe Lockhart 15 95 80 5.33 95 0.00 Stewart 7 45 38 5.43 45 0.00 Normanby 24 78 53 2.21 78 0.00 Jeannie 11 41 30 2.73 41 0.00 Endeavour 47 77 30 0.64 77 0.00 TOTAL 127 454 327 2.57 454 0.00 Dissolved inorganic phosphorus (DIP) (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Jacky Jacky 10 11 1 0.10 11 0.00

Source Catchments modelling – Updated methodology and results

Olive 15 17 2 0.13 17 0.00 Pascoe Lockhart 8 8 0 0.00 8 0.00 Stewart 5 5 0 0.00 5 0.00 Normanby 15 21 7 0.47 21 0.00 Jeannie 5 6 1 0.20 6 0.00 Endeavour 6 8 2 0.33 8 0.00 TOTAL 64 76 11 0.17 76 0.00 Dissolved organic phosphorus (DOP) (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Jacky Jacky 21 22 1 0.05 22 0.00 Olive 30 34 4 0.13 34 0.00 Pascoe Lockhart 15 15 0 0.00 15 0.00 Stewart 10 10 0 0.00 10 0.00 Normanby 30 43 13 0.43 43 0.00 Jeannie 11 12 1 0.09 12 0.00 Endeavour 12 16 3 0.25 16 0.00 TOTAL 129 152 23 0.18 152 0.00 Total nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Jacky Jacky 847 1,064 217 0.26 1,064 0.00 Olive 1,267 1,613 346 0.27 1,613 0.00 Pascoe Lockhart 638 896 258 0.40 896 0.00 Stewart 396 516 120 0.30 516 0.00 Normanby 1,240 1,428 188 0.15 1,428 0.00 Jeannie 469 617 148 0.32 617 0.00 Endeavour 604 721 117 0.19 721 0.00 TOTAL 5,462 6,854 1,393 0.26 6,854 0.00 Particulate nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Jacky Jacky 52 269 217 4.17 269 0.00 Olive 108 451 343 3.18 451 0.00 Pascoe Lockhart 57 315 258 4.53 315 0.00 Stewart 33 153 120 3.64 153 0.00 Normanby 100 253 153 1.53 253 0.01 Jeannie 58 204 147 2.53 204 0.00 Endeavour 139 251 111 0.80 251 0.00 TOTAL 547 1,896 1,349 2.47 1,896 0.00 Dissolved inorganic nitrogen (DIN) (t/yr)

160 Source Catchments modelling – Updated methodology and results

Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Jacky Jacky 67 67 0 0.00 67 0.00 Olive 97 98 1 0.01 98 0.00 Pascoe Lockhart 49 49 0 0.00 49 0.00 Stewart 30 30 0 0.00 30 0.00 Normanby 96 105 9 0.09 105 0.00 Jeannie 34 35 0 0.00 35 0.00 Endeavour 39 40 1 0.03 40 0.00 TOTAL 412 423 11 0.03 423 0.00 Dissolved organic nitrogen (DON) (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Jacky Jacky 728 728 0.2 0.00 728 0.00 Olive 1,062 1,064 1.7 0.00 1,064 0.00 Pascoe Lockhart 532 532 0.1 0.00 532 0.00 Stewart 333 333 0.2 0.00 333 0.00 Normanby 1,045 1,070 25.3 0.02 1,070 0.00 Jeannie 377 378 0.8 0.00 378 0.00 Endeavour 426 430 4.1 0.01 430 0.00 TOTAL 4,503 4,536 32 0.01 4,536 0.00 PSII Load (kg/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Jacky Jacky 0 0 0.0 NA 0.0 0.00 Olive 0 0 0.0 NA 0.0 0.00 Pascoe Lockhart 0 0 0.0 NA 0.0 0.00 Stewart 0 0 0.0 NA 0.0 0.00 Normanby 0 11 11.2 NA 11.2 0.00 Jeannie 0 0 0.2 NA 0.2 0.00 Endeavour 0 4 4.4 NA 4.4 0.00 TOTAL 0 16 16 NA 16 0.00 PSII toxic equivalent load (kg/yr) Predevelopment Baseline Anthropogenic Increase Total Total (12/13) Baseline Factor (14/15) Change (%) Jacky Jacky 0 0 0.00 NA 0 0.00 Olive 0 0 0.00 NA 0 0.00 Pascoe Lockhart 0 0 0.00 NA 0 0.00 Stewart 0 0 0.00 NA 0 0.00 Normanby 0 0 0.40 NA 0 0.00 Jeannie 0 0 0.01 NA 0 0.00 Endeavour 0 0 0.16 NA 0 0.00

161 Source Catchments modelling – Updated methodology and results

TOTAL 0 1 1 NA 1 0.00

Table 70 Sources and sinks of TSS for each basin in CY (kt/yr)

Jacky Olive- Lockhart Stewart Normanby Jeannie Endeavour Jacky Pascoe Supply 54 81 77 59 512 47 71 Channel 1 2 1 1 22 1 1 remobilisation Gully 3 6 0 5 386 3 1 Hillslope 50 66 73 47 86 42 62 Streambank 1 7 3 6 18 1 6 Loss 2 10 10 10 326 6 12 Floodplain 2 9 8 10 299 6 11 deposition Stream deposition 0 0 1 0 26 0 2 Export 52 72 67 49 186 42 59

Table 71 Constituent loads for predevelopment, baseline and RC2015 change model runs for the WT NRM Region

Total fine sediment (kt/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Daintree 75 103 28 0.37 103 0.00 Mossman 10 17 6 0.60 17 0.00 Barron 23 55 32 1.39 55 0.01 Mulgrave- 97 253 156 1.61 253 0.18 Russell Johnstone 119 379 260 2.18 379 0.14 Tully 74 157 83 1.12 157 0.07 Murray 34 74 39 1.15 74 0.24 Herbert 147 478 331 2.25 478 0.02 TOTAL 581 1,516 936 1.61 1,516 0.09 Total phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Daintree 118 146 28 0.24 146 0.00 Mossman 18 28 10 0.56 28 0.00 Barron 35 72 38 1.09 72 0.01 Mulgrave- 165 392 227 1.38 392 0.15 Russell Johnstone 224 864 640 2.86 862 0.28 Tully 124 259 136 1.10 259 0.04 Murray 56 121 65 1.16 121 0.32 Herbert 194 419 224 1.15 419 0.03 TOTAL 934 2,301 1,367 1.46 2,298 0.18 Particulate phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%)

162 Source Catchments modelling – Updated methodology and results

Daintree 66 83 17 0.26 83 0.01 Mossman 10 17 7 0.70 17 0.00 Barron 19 45 27 1.42 45 0.01 Mulgrave- 89 276 187 2.10 276 0.18 Russell Johnstone 141 756 616 4.37 755 0.29 Tully 64 181 117 1.83 181 0.04 Murray 28 84 55 1.96 83 0.37 Herbert 101 291 191 1.89 291 0.03 TOTAL 518 1,733 1,215 2.35 1,731 0.20 Dissolved inorganic phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Daintree 8 19 11 1.38 19 0.00 Mossman 1 3 2 2.00 3 0.00 Barron 3 12 9 3.00 12 0.00 Mulgrave- 12 40 28 2.33 40 0.00 Russell Johnstone 14 35 21 1.50 35 0.00 Tully 10 24 14 1.40 24 0.00 Murray 4 12 8 2.00 12 0.00 Herbert 14 45 31 2.21 45 0.00 TOTAL 68 192 124 1.82 192 0.00 Dissolved organic phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Daintree 43 44 1 0.02 44 0.00 Mossman 7 8 1 0.14 8 0.00 Barron 14 15 2 0.14 15 0.00 Mulgrave- 64 76 12 0.19 76 0.00 Russell Johnstone 70 72 2 0.03 72 0.00 Tully 49 54 5 0.10 54 0.00 Murray 23 25 2 0.09 25 0.00 Herbert 79 82 2 0.03 82 0.00 TOTAL 349 376 27 0.08 376 0.00 Total nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Daintree 1,120 1,841 722 0.64 1,841 0.01 Mossman 176 323 147 0.84 323 0.03 Barron 258 568 309 1.20 567 0.02 Mulgrave- 1,594 2,769 1,176 0.74 2,768 0.07 Russell Johnstone 1,746 3,799 2,053 1.18 3,796 0.13

163 Source Catchments modelling – Updated methodology and results

Tully 1,222 2,155 934 0.76 2,155 0.01 Murray 558 1,121 563 1.01 1,120 0.29 Herbert 1,739 4,001 2,262 1.30 3,998 0.14 TOTAL 8,413 16,577 8,165 0.97 16,569 0.10 Particulate nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Daintree 520 579 59 0.11 579 0.00 Mossman 79 101 22 0.28 101 0.00 Barron 112 201 88 0.79 201 0.00 Mulgrave- 702 1,227 525 0.75 1,226 0.17 Russell Johnstone 769 1,986 1,217 1.58 1,984 0.16 Tully 535 875 340 0.64 875 -0.01 Murray 240 400 160 0.67 400 0.19 Herbert 629 1,327 697 1.11 1,327 0.03 TOTAL 3,587 6,696 3,109 0.87 6,693 0.10 Dissolved inorganic nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Daintree 344 478 135 0.39 478 0.03 Mossman 55 160 104 1.89 159 0.05 Barron 65 152 87 1.34 152 0.05 Mulgrave- 511 934 423 0.83 934 0.00 Russell Johnstone 560 1,059 499 0.89 1,058 0.15 Tully 393 777 384 0.98 777 0.04 Murray 182 414 232 1.27 413 0.58 Herbert 636 1,522 886 1.39 1,519 0.33 TOTAL 2,746 5,496 2,750 1.00 5,491 0.19 Dissolved organic nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Daintree 256 784 528 2.06 784 0.00 Mossman 41 62 20 0.49 62 0.00 Barron 81 215 134 1.65 215 0.00 Mulgrave- 381 609 228 0.60 609 0.00 Russell Johnstone 417 754 337 0.81 754 0.00 Tully 293 503 210 0.72 503 0.00 Murray 136 307 171 1.26 307 0.00 Herbert 474 1,152 679 1.43 1,152 0.00 TOTAL 2,080 4,385 2,305 1.11 4,385 0.00 PSII Load (kg/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%)

164 Source Catchments modelling – Updated methodology and results

Daintree 0 118 118 NA 118 0.00 Mossman 0 75 75 NA 75 0.00 Barron 0 120 120 NA 119 1.13 Mulgrave- 0 746 746 NA 738 1.09 Russell Johnstone 0 922 922 NA 866 6.11 Tully 0 492 492 NA 485 1.51 Murray 0 367 367 NA 347 5.44 Herbert 0 248 248 NA 248 0.03 TOTAL 0 3,090 3,090 NA 2,997 3.02 PSII toxic equivalent load (kg/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Daintree 0 98 98 NA 98 0.00 Mossman 0 65 65 NA 65 0.00 Barron 0 37 37 NA 36 3.22 Mulgrave- 0 625 625 NA 618 1.14 Russell Johnstone 0 753 753 NA 703 6.57 Tully 0 414 414 NA 407 1.61 Murray 0 310 310 NA 293 5.66 Herbert 0 113 113 NA 113 0.01 TOTAL 0 2,415 2415 NA 2,333 3.40

Table 72 Sources and sinks of each constituent in the Wet Tropics

Mulgrave- Daintree Mossman Barron Johnstone Tully Murray Herbert Russell Supply 109 18 108 267 410 163 76 612 Channel remobilisation 3 0 3 2 3 2 1 27 Gully 0 0 8 1 5 1 0 60 Hillslope 91 18 79 224 335 142 62 236 Streambank 15 1 18 41 67 19 12 289 Loss 6 1 53 14 31 7 2 134 Extractions 0 10 1 2 1 0 1 Floodplain deposition 5 1 14 12 11 3 1 73 Stream deposition 1 29 2 18 3 1 59 Residual 0 0 0 0 0 0 0 0 Export 103 17 55 253 379 157 74 478

Table 73 Constituent loads for predevelopment, baseline and 2015 Report Card change model runs for the Burdekin NRM Region

Total fine sediment (kt/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Black 28 62 34 1.21 62 0.00

165 Source Catchments modelling – Updated methodology and results

Ross 13 62 49 3.77 62 0.00 Haughton 25 183 157 6.28 183 0.10 Burdekin 475 3,260 2,786 5.87 3,256 0.18 Don 31 213 183 5.90 212 0.73 TOTAL 571 3,781 3,209 5.62 3,774 0.20 Total phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Black 82 105 23 0.28 105 0.00 Ross 27 94 67 2.48 94 0.01 Haughton 59 236 177 3.00 236 0.16 Burdekin 685 2,177 1,492 2.18 2,173 0.25 Don 67 211 144 2.15 209 1.13 TOTAL 919 2,823 1,903 2.07 2,817 0.30 Particulate phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Black 53 77 23 0.43 77 0.00 Ross 7 50 42 6.00 50 0.01 Haughton 23 141 118 5.13 141 0.12 Burdekin 319 1,797 1,479 4.64 1,794 0.24 Don 32 174 142 4.44 172 1.14 TOTAL 434 2,239 1,805 4.16 2,234 0.30 Dissolved inorganic phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Black 21 21 0 0.00 21 0.14 Ross 14 33 19 1.36 33 0.00 Haughton 27 73 47 1.74 73 0.25 Burdekin 275 285 11 0.04 285 1.33 Don 26 27 1 0.04 27 -0.49 TOTAL 362 440 78 0.22 440 0.32 Dissolved organic phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Black 7 7 0 0.00 7 0.14 Ross 6 11 5 0.83 11 0.00 Haughton 9 21 12 1.33 21 0.24 Burdekin 92 94 3 0.03 94 1.33 Don 9 10 0 0.00 10 -0.49 TOTAL 123 144 21 0.17 143 0.31 Total nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%)

166 Source Catchments modelling – Updated methodology and results

Black 309 385 76 0.25 385 0.00 Ross 177 416 239 1.35 416 0.00 Haughton 322 1,474 1,152 3.58 1,440 3.01 Burdekin 3,256 5,849 2,593 0.80 5,835 0.53 Don 347 669 322 0.93 657 3.80 TOTAL 4,411 8,793 4,382 0.99 8,732 1.39 Particulate nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Black 89 143 54 0.61 143 0.00 Ross 15 98 83 5.53 98 0.00 Haughton 36 223 187 5.19 223 0.12 Burdekin 480 2,887 2,407 5.01 2,882 0.21 Don 55 305 250 4.55 302 1.14 TOTAL 675 3,657 2,982 4.42 3,649 0.27 Dissolved inorganic nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Black 76 97 21 0.28 97 0.00 Ross 57 180 123 2.16 180 0.00 Haughton 102 1,016 914 8.96 982 3.75 Burdekin 933 1,104 171 0.18 1,095 5.06 Don 109 177 68 0.62 168 13.81 TOTAL 1,278 2,574 1,297 1.01 2,522 4.03 Dissolved organic nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Black 143 145 2 0.01 145 0.00 Ross 105 138 33 0.31 138 0.00 Haughton 184 235 52 0.28 235 0.39 Burdekin 1,844 1,857 14 0.01 1,857 1.07 Don 183 186 4 0.02 186 -0.47 TOTAL 2,458 2,562 104 0.04 2,561 0.32 PSII load (kg/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Black 0 2 2 NA 2 -0.01 Ross 0 2 2 NA 2 0.04 Haughton 0 1,708 1,708 NA 1,645 3.66 Burdekin 0 503 503 NA 481 4.37 Don 0 179 179 NA 179 0.27 TOTAL 0 2,394 2,394 NA 2,309 3.55 PSII toxic equivalent load (kg/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%)

167 Source Catchments modelling – Updated methodology and results

Black 0 2 2 NA 2 0.00 Ross 0 0 0 NA 0 0.04 Haughton 0 937 937 NA 902 3.68 Burdekin 0 255 255 NA 244 4.34 Don 0 90 90 NA 89 0.32 TOTAL 0 1,284 1,284 NA 1,238 3.57

Table 74 Sources and sinks of each constituent in the Burdekin NRM Region

Black Burdekin Don Haughton Ross Supply 64 5,943 221 194 104 Channel

remobilisation Gully 19 3,919 109 92 43 Hillslope 42 1,205 87 81 57 Streambank 4 819 25 20 4 Loss 2 2,683 8 11 42 Floodplain deposition 2 861 8 11 6 Stream deposition Reservoir Deposition 1,649 36 Extraction 167 Residual 0 6 0 0 0 Export 62 3,260 213 183 62

Table 75 Constituent loads for predevelopment, baseline and 2015 Report Card change model runs for the Mackay Whitsunday NRM Region

Total fine sediment (kt/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Proserpine 56 131 75 1.34 131 0.35 O'Connell 73 314 241 3.30 314 0.01 Pioneer 54 227 173 3.20 227 0.03 Plane 47 146 99 2.11 146 0.07 TOTAL 230 818 589 2.56 818 0.07 Total phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Proserpine 112 255 143 1.28 255 0.26 O'Connell 145 495 350 2.41 495 0.00 Pioneer 83 237 154 1.86 237 0.03 Plane 107 319 212 1.98 319 0.05 TOTAL 447 1,306 859 1.92 1,306 0.06 Particulate phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%)

168 Source Catchments modelling – Updated methodology and results

Proserpine 76 164 88 1.16 164 0.43 O'Connell 112 416 304 2.71 416 0.00 Pioneer 68 184 116 1.71 184 0.04 Plane 83 228 144 1.73 228 0.08 TOTAL 339 992 653 1.93 991 0.08 Dissolved inorganic phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Proserpine 29 73 44 1.52 73 0.00 O'Connell 26 62 37 1.42 62 0.00 Pioneer 12 42 30 2.50 42 0.00 Plane 18 73 54 3.00 73 0.00 TOTAL 85 249 165 1.94 249 0.00 Dissolved organic phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Proserpine 8 19 11 1.38 19 0.00 O'Connell 7 16 9 1.29 16 0.00 Pioneer 3 11 7 2.33 11 0.00 Plane 5 19 14 2.80 19 0.00 TOTAL 24 65 41 1.71 65 0.00 Total nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Proserpine 556 1,066 510 0.92 1,067 -0.20 O'Connell 579 1,528 950 1.64 1,528 0.01 Pioneer 308 933 625 2.03 930 0.50 Plane 417 1,279 862 2.07 1,270 1.06 TOTAL 1,860 4,806 2,946 1.58 4,795 0.38 Particulate nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Proserpine 161 387 226 1.40 386 0.36 O'Connell 221 845 624 2.82 845 0.00 Pioneer 145 450 305 2.10 450 0.03 Plane 163 468 305 1.87 468 0.07 TOTAL 690 2,150 1,460 2.12 2,149 0.08 Dissolved inorganic nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Proserpine 153 310 157 1.03 312 -1.16 O'Connell 139 325 186 1.34 325 0.03 Pioneer 63 256 193 3.06 253 1.56 Plane 98 464 366 3.73 455 2.43

169 Source Catchments modelling – Updated methodology and results

TOTAL 453 1,355 902 1.99 1,344 1.12 Dissolved organic nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Proserpine 242 370 127 0.52 370 0.00 O'Connell 219 358 139 0.63 358 0.00 Pioneer 100 227 127 1.27 227 0.00 Plane 156 347 191 1.22 347 0.00 TOTAL 717 1,301 584 0.81 1,301 0.00 PSII load (kg/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Proserpine 0 331 331 NA 329 0.74 O'Connell 0 971 971 NA 960 1.12 Pioneer 0 985 985 NA 983 0.23 Plane 0 1758 1758 NA 1651 6.06 TOTAL 0 4,044 4044 NA 3,922 3.02 PSII toxic equivalent load (kg/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Proserpine 0 248 248 NA 246 0.71 O'Connell 0 701 701 NA 693 1.10 Pioneer 0 707 707 NA 705 0.24 Plane 0 1,269 1,269 NA 1,194 5.93 TOTAL 0 2,925 2,925 NA 2,838 2.96

Table 76 Sources and sinks of TSS for each basin MW (kt/yr)

Proserpine O’Connell Pioneer Plane Supply 147 316 242 148 Channel remobilisation 0 0 0 0 Gully 5 4 2 2 Hillslope 117 241 100 119 Streambank 25 71 140 27 Loss 16 2 11 2 Floodplain deposition 1 1 2 2 Stream deposition 0 0 0 0 Reservoir deposition 15 0 2 0 Extraction 0 0 6 0 Export 131 314 231 146

170 Source Catchments modelling – Updated methodology and results

Table 77 Constituent loads for predevelopment, baseline and RC2015 change model runs for the Fitzroy NRM Region

Total fine sediment (kt/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Styx 9 104 96 9.40 104 0.59 Shoalwater 7 67 60 7.38 67 0.09 Water Park 8 65 57 7.13 65 0.00 Fitzroy 181 1,507 1,326 6.01 1,493 1.07 Calliope 7 57 50 7.14 57 0.04 Boyne 8 24 16 2.00 24 1.09 TOTAL 219 1,824 1,605 6.13 1,809 0.94 Total phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Styx 128 541 413 3.22 540 0.07 Shoalwater 122 408 287 2.34 408 0.01 Water Park 109 579 471 4.32 579 0.00 Fitzroy 1,054 2,750 1,696 1.57 2,745 0.32 Calliope 74 281 206 2.75 281 0.00 Boyne 76 105 29 0.38 105 0.18 TOTAL 1,563 4,665 3,102 1.96 4,659 0.19 Particulate phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Styx 64 428 364 5.69 427 0.08 Shoalwater 46 308 262 5.57 308 0.01 Water Park 60 517 457 7.62 517 - Fitzroy 558 1,822 1,265 2.19 1,817 0.43 Calliope 41 221 180 4.39 221 0.00 Boyne 48 60 12 0.25 60 0.44 TOTAL 817 3,357 2,540 3.03 3,351 0.23 Dissolved inorganic phosphorus (DIP) (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Styx 52 91 39 0.75 91 0.00 Shoalwater 61 80 20 0.33 80 0.00 Water Park 39 50 11 0.28 50 0.00 Fitzroy 397 743 345 0.87 743 0.00 Calliope 27 47 21 0.78 47 0.00 Boyne 22 36 14 0.64 36 0.00 TOTAL 597 1,047 450 0.75 1,047 0.00 Dissolved organic phosphorus (DOP) (t/yr)

171 Source Catchments modelling – Updated methodology and results

Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Styx 13 22 10 0.77 22 0.00 Shoalwater 15 20 5 0.33 20 0.00 Water Park 10 12 3 0.30 12 0.00 Fitzroy 98 185 86 0.88 185 0.00 Calliope 7 12 5 0.71 12 0.00 Boyne 5 9 3 0.60 9 0.00 TOTAL 148 260 112 0.76 260 0.00 Total nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Styx 372 1,212 840 2.25 1,212 0.04 Shoalwater 388 1,005 617 1.58 1,005 0.01 Water Park 333 1,494 1,161 3.49 1,494 0.00 Fitzroy 2,875 6,290 3,416 1.17 6,280 0.30 Calliope 208 639 430 2.06 639 0.01 Boyne 195 266 71 0.36 266 0.10 TOTAL 4,371 10,907 6,535 1.48 10,896 0.16 Particulate nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Styx 124 829 705 5.68 828 0.05 Shoalwater 98 646 548 5.59 646 0.01 Water Park 146 1,268 1,122 7.68 1,268 - Fitzroy 918 3,066 2,148 2.28 3,056 0.48 Calliope 81 439 358 4.42 439 0.01 Boyne 90 113 23 0.26 113 0.31 TOTAL 1,456 6,361 4,904 3.31 6,350 0.22 Dissolved inorganic nitrogen (DIN) (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Styx 81 91 10 0.12 91 0.00 Shoalwater 95 100 5 0.05 100 0.00 Water Park 61 65 4 0.07 65 0.00 Fitzroy 641 799 159 0.25 799 0.00 Calliope 42 47 6 0.14 47 0.00 Boyne 35 37 3 0.09 37 0.00 TOTAL 954 1,140 186 0.19 1,140 0.00 Dissolved organic nitrogen (DON) (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Styx 167 293 126 0.75 293 0.00

172 Source Catchments modelling – Updated methodology and results

Shoalwater 196 259 64 0.33 259 0.00 Water Park 126 161 35 0.28 161 0.00 Fitzroy 1,316 2,425 1,109 0.84 2,425 0.00 Calliope 86 152 67 0.78 152 0.00 Boyne 71 116 45 0.63 116 0.00 TOTAL 1,961 3,406 1,445 0.74 3,406 0.00 PSII Load (kg/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Styx 0 203 203 NA 203 0.00 Shoalwater 0 107 107 NA 107 0.00 Water Park 0 15 15 NA 15 0.00 Fitzroy 0 1,749 1,749 NA 1,745 0.21 Calliope 0 97 97 NA 97 0.00 Boyne 0 39 39 NA 39 0.00 TOTAL 0 2,210 2,210 NA 2,206 0.17 PSII toxic equivalent load (kg/yr) Predevelopment Baseline Anthropogenic Increase Total Total (12/13) Baseline Factor (14/15) Change (%) Styx 0 4 4 NA 4 0.00 Shoalwater 0 2 2 NA 2 0.00 Water Park 0 0 0 NA 0 0.00 Fitzroy 0 38 38 NA 38 0.35 Calliope 0 2 2 NA 2 0.00 Boyne 0 1 1 NA 1 0.00 TOTAL 0 47 47 NA 47 0.29

Table 78 Sources and sinks of TSS for each basin in the Fitzroy NRM Region

Styx Shoalwater Water Park Fitzroy Calliope Boyne River Creek Creek River River River Supply 120 75 67 8,153 80 146 Channel remobilisation 0 0 0 0 0 0 Gully 13 11 2 4,435 13 6 Hillslope 80 49 64 2,139 53 105 Streambank 27 15 1 1,579 14 35 Loss 16 8 2 6,646 23 122 Floodplain deposition 16 8 2 4,061 23 45 Stream deposition 0 0 0 0 0 0 Reservoir deposition 0 0 0 1,207 0 70 Extraction 0 0 0 408 0 5 Node loss 0 0 0 952 0 0 Residual 0 0 0 18 0 2 Export 104 67 65 1,507 57 24

173 Source Catchments modelling – Updated methodology and results

Table 79 Constituent loads for predevelopment, baseline and RC2015 change model runs for the Burnett Mary NRM Region

Total fine sediment (kt/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Baffle 23 75 53 2.30 75 0.11 Kolan 10 40 30 3.00 40 0.02 Burnett 122 548 426 3.49 548 0.07 Burrum 9 26 17 1.89 26 0.01 Mary 104 770 666 6.40 770 0.02 TOTAL 267 1,459 1,192 4.46 1,459 0.04 Total phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Baffle 60 148 88 1.47 148 0.15 Kolan 17 48 31 1.82 48 0.00 Burnett 140 317 177 1.26 317 0.03 Burrum 14 44 30 2.14 44 0.01 Mary 228 1084 856 3.75 1084 0.01 TOTAL 459 1642 1183 2.58 1641 0.02 Particulate phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Baffle 48 125 77 1.60 125 0.17 Kolan 11 37 25 2.27 37 0.00 Burnett 122 268 146 1.20 268 0.04 Burrum 8 25 17 2.13 25 0.01 Mary 182 972 789 4.34 972 0.01 TOTAL 372 1,426 1,054 2.83 1,426 0.03 Dissolved inorganic phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Baffle 7 14 7 1.00 14 0.00 Kolan 3 7 4 1.33 7 0.00 Burnett 11 32 21 1.91 32 0.00 Burrum 4 13 10 2.50 13 0.00 Mary 28 72 44 1.57 72 0.00 TOTAL 53 139 85 1.60 139 0.00 Dissolved organic phosphorus (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Baffle 5 9 4 0.80 9 0.00 Kolan 2 4 2 1.00 4 0.00 Burnett 7 17 10 1.43 17 0.00

174 Source Catchments modelling – Updated methodology and results

Burrum 2 6 3 1.50 6 0.00 Mary 18 41 23 1.28 41 0.00 TOTAL 34 77 43 1.26 77 0.00 Total nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Baffle 244 546 302 1.24 546 0.10 Kolan 91 312 221 2.43 312 0.00 Burnett 534 1,407 873 1.63 1,407 0.05 Burrum 90 420 330 3.67 420 0.02 Mary 1,075 4,460 3,385 3.15 4,460 0.01 TOTAL 2,034 7,146 5,112 2.51 7145 0.02 Particulate nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Baffle 105 268 164 1.56 268 0.19 Kolan 33 101 68 2.06 101 0.00 Burnett 326 667 341 1.05 667 0.06 Burrum 20 58 38 1.90 58 0.00 Mary 552 2,899 2,347 4.25 2,899 0.01 TOTAL 1,036 3,993 2,957 2.85 3,993 0.02 Dissolved inorganic nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Baffle 26 58 32 1.23 58 0.00 Kolan 11 78 68 6.18 78 0.00 Burnett 39 246 207 5.31 246 0.12 Burrum 13 199 186 14.31 199 0.03 Mary 97 459 361 3.72 459 0.00 TOTAL 186 1,040 854 4.59 1,039 0.03 Dissolved organic nitrogen (t/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Baffle 114 220 107 0.94 220 0.00 Kolan 47 133 86 1.83 133 0.00 Burnett 169 494 325 1.92 494 0.00 Burrum 57 163 106 1.86 163 0.00 Mary 426 1,102 676 1.59 1,102 0.00 TOTAL 812 2,113 1,301 1.60 2,113 0.00 PSII load (kg/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Baffle 0 8 8 NA 8 0.00 Kolan 0 70 70 NA 70 0.00 Burnett 0 119 119 NA 119 0.15

175 Source Catchments modelling – Updated methodology and results

Burrum 0 81 81 NA 81 0.00 Mary 0 68 68 NA 68 0.92 TOTAL 0 346 346 NA 345 0.24 PSII toxic equivalent load (kg/yr) Baseline Anthropogenic Increase Total Total Predevelopment (12/13) Baseline Factor (14/15) Change (%) Baffle 0 3 3 NA 3 0.00 Kolan 0 31 31 NA 31 0.00 Burnett 0 22 22 NA 22 0.06 Burrum 0 16 16 NA 16 0.00 Mary 0 14 14 NA 14 0.35 TOTAL 0 86 86 NA 86 0.07

Table 80 Sources and sinks of each constituent in the Burnett Mary NRM Region

Basin Baffle Kolan Burnett Burrum Mary Supply 79 77 1,119 31 1,017 Channel Remobilisation 9 11 174 7 22 Gully 13 15 288 3 32 Hillslope surface soil 48 24 104 16 339 Streambank 8 27 553 5 623 Loss 3 37 571 5 247 Extraction 1 3 17 1 58 Flood Plain Deposition 3 2 78 1 7 Node Loss 1 32 56 Reservoir Deposition 26 346 3 123 Residual Node Storage 2 5 Stream Deposition 4 93 3 Export 75 40 548 26 770

176 Source Catchments modelling – Updated methodology and results

Appendix 4 – Model validation

Wet Tropics model validation

Table 81 Fine sediment and particulate nutrient comparison with GBRCLMP monitored data for five WT sites

TSS (t/yr) PN (t/yr) PP (t/yr Site % % % Measured Modelled Measured Modelled Measured Modelled diff diff diff Barron River at Myola 208,770 86,025 -59 656 284 -57 143 74 -48 North Johnstone River 166,495 123,650 -26 924 697 -25 294 210 -28 at Tung Oil South Johnstone River at Upstream Central 77,850 45,556 -41 367 287 -22 140 96 -31 mill Tully River at Tully 25,522 19,990 -22 179 158 -11 39 19 -51 Gorge Tully River at Euramo 110,901 147,170 33 475 805 69 129 185 44 Herbert River at 566,999 459,453 -19 1,764 1,203 -32 438 245 -44 Ingham

Table 82 Dissolved nitrogen species comparison with GBRCLMP monitored data for five WT sites

DIN (t/yr) DON (t/yr) Site Measured Modelled % diff Measured Modelled % diff Barron River at Myola 57 81 43 294 250 -15 North Johnstone River at Tung Oil 311 282 -9 229 251 10 South Johnstone River at 160 122 -24 93 95 2 Upstream Central mill Tully River at Tully Gorge 148 99 -33 119 87 -27 Tully River at Euramo 778 657 -16 459 419 -9 Herbert River at Ingham 1,055 748 -29 1,190 926 -22

Table 83 Dissolved phosphorus comparison with GBRCLMP monitored data for five WT sites

FRP (t/yr) DOP (t/yr) Site Measured Modelled % diff Measured Modelled % diff Barron River at Myola 10 10 -4 17 16 -3 North Johnstone River at Tung Oil 15 13 -13 49 31 -37 South Johnstone River at 10 7 -29 18 13 -28 Upstream Central mill Tully River at Tully Gorge 3 3 0 20 12 -38 Tully River at Euramo 22 22 1 63 49 -22 Herbert River at Ingham 51 38 -25 111 77 -31

Table 84 Barron River model validation compared to GBRCLMP annual loads

Site 110001D RC7 Annual Time period 2006-2014 (excludes 2008-2009) Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter Flow Very good Satisfactory Very good Very good Satisfactory TSS Indicative Indicative Indicative Very good Indicative TN Indicative Indicative Satisfactory Very good Indicative PN Indicative Indicative Indicative Very good Indicative

177 Source Catchments modelling – Updated methodology and results

Site 110001D RC7 Annual Time period 2006-2014 (excludes 2008-2009) Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter DIN Indicative Indicative Indicative Very good Indicative DON Very good Good Very good Very good Good TP Indicative Indicative Satisfactory Very good Indicative DIP Very good Very good Very good Very good Very good PP Indicative Indicative Indicative Very good Indicative DOP Satisfactory Very good Good Good Satisfactory

Table 85 Barron River model validation compared to Joo monthly loads

Site 110001D RC7 monthly Time period 1986–2009 Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter Flow Very good Very good Very good Very good Very good TSS Indicative Indicative Satisfactory Very good Indicative TN Very good Indicative Very good Very good Very good PN Indicative Indicative Satisfactory Very good Indicative DIN Indicative Satisfactory Good Very good Indicative DON Very good Good Very good Very good Very good TP Good Indicative Very good Very good Good DIP Very good Very good Very good Very good Very good PP Satisfactory Indicative Good Very good Satisfactory DOP Good Indicative Very good Very good Good

Table 86 North Johnstone River model validation compared to GBRCLMP annual loads

Site 112004A RC7 Annual Time period 2006-2014 (excludes 2008-2009) Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter Flow Good Satisfactory Good Very good Satisfactory TSS Indicative Indicative Indicative Satisfactory Indicative TN Good Good Very good Very good Good PN Indicative Satisfactory Satisfactory Very good Indicative DIN Very good Very good Very good Very good Very good DON Very good Very good Very good Very good Very good TP Satisfactory Satisfactory Good Good Satisfactory DIP Very good Good Very good Very good Good PP Good Indicative Satisfactory Satisfactory Indicative DOP Indicative Satisfactory Satisfactory Very good Indicative

Table 87 North Johnstone River model validation compared to Joo monthly loads

Site 112004A RC7 monthly Time period 1986–2009 Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter Flow Very good Very good Very good Very good Very good TSS Good Indicative Satisfactory Good Good TN Good Indicative Very good Very good Good PN Indicative Indicative Satisfactory Very good Indicative

178 Source Catchments modelling – Updated methodology and results

Site 112004A RC7 monthly Time period 1986–2009 Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter DIN Very good Very good Very good Very good Very good DON Very good Very good Very good Very good Very good TP Satisfactory Indicative Good Very good Satisfactory DIP Very good Very good Very good Very good Very good PP Satisfactory Indicative Good Very good Satisfactory DOP Indicative Indicative Satisfactory Very good Indicative

Table 88 South Johnstone River model validation compared to GBRCLMP annual loads

Site 112101B RC7 Annual Time period 2006-2014 (excludes 2008-2009) Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter Flow Very good Good Very good Very good Good TSS Indicative Indicative Indicative Very good Indicative TN Good Good Very good Very good Good PN Satisfactory Satisfactory Good Very good Satisfactory DIN Indicative Satisfactory Indicative Satisfactory Indicative DON Very good Very good Very good Very good Very good TP Indicative Satisfactory Good Very good Indicative DIP Indicative Indicative Indicative Very good Indicative PP Indicative Indicative Satisfactory Very good Indicative DOP Indicative Satisfactory Indicative Very good Indicative

Table 89 South Johnstone River model validation compared to Joo monthly loads

Site 112101B RC7 monthly Time period 1986–2009 Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter Flow Very good Very good Very good Very good Very good TSS Satisfactory Very good Satisfactory Good Satisfactory TN Very good Good Very good Very good Very good PN Satisfactory Satisfactory Good Good Satisfactory DIN Very good Very good Very good Very good Very good DON Very good Very good Very good Very good Very good TP Good Very good Very good Very good Good DIP Indicative Indicative Indicative Very good Indicative PP Good Very good Good Good Good DOP Indicative Indicative Indicative Very good Indicative

Table 90 Tully River model validation compared to GBRCLMP annual loads

Site 113006A RC7 Annual Time period 2006-2014 (excludes 2008-2009) Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter Flow Indicative Indicative Indicative Very good Indicative TSS Indicative Indicative Indicative Very good Indicative TN Indicative Very good Satisfactory Very good Indicative PN Indicative Indicative Indicative Very good Indicative

179 Source Catchments modelling – Updated methodology and results

Site 113006A RC7 Annual Time period 2006-2014 (excludes 2008-2009) Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter DIN Indicative Good Indicative Satisfactory Indicative DON Indicative Very good Indicative Satisfactory Indicative TP Indicative Indicative Indicative Very good Indicative DIP Indicative Very good Indicative Indicative Indicative PP Indicative Indicative Indicative Very good Indicative DOP Indicative Satisfactory Indicative Satisfactory Indicative

Table 91 Tully River model validation compared to Joo monthly loads

Site 113006A RC7 monthly Time period 1986–2009 Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter Flow Very good Very good Very good Very good Very good TSS Very good Indicative Good Very good Very good TN Satisfactory Indicative Good Very good Satisfactory PN Indicative Indicative Indicative Very good Indicative DIN Very good Very good Very good Very good Very good DON Very good Good Very good Very good Very good TP Indicative Indicative Indicative Very good Indicative DIP Very good Satisfactory Very good Very good Very good PP Indicative Indicative Indicative Very good Indicative DOP Indicative Indicative Indicative Very good Indicative

Table 92 Herbert River model validation compared to GBRCLMP annual loads

Site 116001D RC7 Annual Time period 2006-2014 (excludes 2008-2009) Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter Flow Very good Satisfactory Very good Very good Satisfactory TSS Very good Indicative Very good Very good Indicative TN Very good Satisfactory Very good Very good Satisfactory PN Good Satisfactory Very good Very good Satisfactory DIN Indicative Good Satisfactory Indicative Indicative DON Very good Good Very good Very good Good TP Good Indicative Very good Very good Indicative DIP Very good Satisfactory Very good Very good Satisfactory PP Satisfactory Indicative Good Very good Indicative DOP Good Satisfactory Very good Very good Satisfactory

Table 93 Herbert River model validation compared to Joo monthly loads

Site 116001F RC7 monthly Time period 1986–2009 Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter Flow Very good Very good Very good Very good Very good TSS Good Indicative Satisfactory Very good Good TN Very good Very good Very good Very good Very good

180 Source Catchments modelling – Updated methodology and results

Site 116001F RC7 monthly Time period 1986–2009 Modelled RSR rating PBIAS rating NSE rating R2 Overall rating parameter PN Satisfactory Satisfactory Good Very good Satisfactory DIN Satisfactory Indicative Good Very good Satisfactory DON Very good Satisfactory Very good Very good Very good TP Good Very good Very good Very good Good DIP Very good Very good Very good Very good Very good PP Satisfactory Indicative Good Very good Satisfactory DOP Good Indicative Very good Very good Good

Burdekin model validation

Table 94 RC2015 (120302B Cape River at Taemas, n = 7 years for flow and n = 7 years for WQ)

RSR PBIAS NSE R2 Parameter Rating Rating Rating Rating Value (Moriasi Value (Moriasi Value (Moriasi Value (Moriasi 2007) 2015) 2015) 2015) Flow 0.22 Very good -5.06 Good 0.95 Very good 0.96 Very good TSS 0.88 Indicative 51.20 Indicative 0.23 Indicative 0.78 Good TN 0.88 Indicative 51.70 Indicative 0.22 Indicative 0.83 Very good PN 1.28 Indicative 80.72 Indicative -0.63 Indicative 0.56 Satisfactory DIN 2.86 Indicative -207.49 Indicative -7.18 Indicative 0.02 Indicative DON 0.64 Satisfactory 37.50 Indicative 0.59 Good 0.93 Very good TP 0.68 Satisfactory 15.12 Good 0.54 Good 0.45 Satisfactory DIP 5.17 Indicative -299.15 Indicative -25.70 Indicative 0.82 Very good PP 0.88 Indicative 35.77 Indicative 0.22 Indicative 0.25 Indicative DOP 0.88 Indicative 52.43 Indicative 0.23 Indicative 0.79 Very good

Table 95 RC2015 (119101A Barratta Creek at Northcote, n = 5 years for flow and n = 5 years for WQ)

RSR PBIAS NSE R2 Rating Rating Rating Parameter Rating Value Value (Moriasi Value (Moriasi Value (Moriasi (Moriasi 2007) 2015) 2015) 2015) Flow 0.40 Very good 32.58 Indicative 0.84 Very good 0.94 Very good TSS 0.76 Indicative 39.16 Indicative 0.42 Indicative 0.74 Good TN 0.68 Satisfactory 51.15 Indicative 0.54 Good 0.86 Very good PN 0.94 Indicative 76.52 Indicative 0.11 Indicative 0.64 Good DIN 0.92 Indicative -39.29 Indicative 0.16 Indicative 0.17 Indicative DON 0.86 Indicative 72.09 Indicative 0.26 Indicative 0.88 Very good TP 0.71 Indicative 52.41 Indicative 0.50 Satisfactory 0.82 Very good DIP 0.43 Very good 28.86 Satisfactory 0.81 Very good 0.69 Good PP 0.72 Indicative 51.60 Indicative 0.48 Satisfactory 0.82 Very good DOP 0.81 Indicative 64.47 Indicative 0.35 Indicative 0.87 Very good

181 Source Catchments modelling – Updated methodology and results

Table 96 RC2015 (120002C Burdekin River at Sellheim, n = 8 years for flow and n = 7 years for WQ)

RSR PBIAS NSE R2 Parameter Rating Rating Rating Rating Value (Moriasi Value (Moriasi Value (Moriasi Value (Moriasi 2007) 2015) 2015) 2015) Flow 0.20 Very good -3.48 Very good 0.96 Very good 0.96 Very good TSS 0.44 Very good 26.47 Indicative 0.81 Very good 0.90 Very good TN 0.74 Indicative 49.17 Indicative 0.45 Satisfactory 0.94 Very good PN 0.98 Indicative 68.27 Indicative 0.04 Indicative 0.90 Very good DIN 4.08 Indicative -131.73 Indicative -15.67 Indicative 0.64 Good DON 0.28 Very good 16.82 Good 0.92 Very good 0.99 Very good TP 0.63 Satisfactory 37.17 Indicative 0.60 Good 0.86 Very good DIP 0.28 Very good -15.25 Good 0.92 Very good 0.98 Very good PP 0.69 Satisfactory 40.47 Indicative 0.53 Good 0.83 Very good DOP 0.74 Indicative 44.05 Indicative 0.45 Satisfactory 0.72 Very good

Table 97 RC2015 (120301B Belyando River at Gregory Development Road, n = 7 years for flow and n = 6 years for WQ)

RSR PBIAS NSE R2 Parameter Rating Rating Rating Rating Value (Moriasi Value (Moriasi Value (Moriasi Value (Moriasi 2007) 2015) 2015) 2015) Flow 0.28 Very good -4.89 Very good 0.92 Very good 0.93 Very good TSS 1.84 Indicative -72.95 Indicative -2.40 Indicative 0.13 Indicative TN 0.49 Very good 34.02 Indicative 0.76 Very good 0.88 Very good PN 0.56 Good 36.82 Indicative 0.69 Very good 0.78 Very good DIN 6.16 Indicative -401.03 Indicative -36.99 Indicative 0.88 Very good DON 0.69 Satisfactory 52.82 Indicative 0.53 Good 0.94 Very good TP 0.60 Good 8.73 Very good 0.64 Good 0.54 Satisfactory DIP 0.69 Satisfactory 41.74 Indicative 0.53 Good 0.57 Satisfactory PP 0.65 Satisfactory -6.70 Very good 0.58 Good 0.54 Satisfactory DOP 0.73 Indicative 50.73 Indicative 0.47 Satisfactory 0.69 Good

Table 98 RC2015 (120001A Burdekin River at Home Hill (120006B Burdekin River at Clare), n = 8 years for flow and n = 8 years for WQ)

RSR PBIAS NSE R2 Rating Rating Rating Parameter Rating (Moriasi Value Value (Moriasi Value (Moriasi Value (Moriasi 2007) 2015) 2015) 2015) Flow 0.17 Very good 0.29 Very good 0.97 Very good 0.97 Very good TSS 0.45 Very good 2.46 Very good 0.80 Good 0.80 Very good TN 0.67 Satisfactory 33.23 Indicative 0.56 Good 0.87 Very good PN 0.88 Indicative 47.64 Indicative 0.22 Indicative 0.77 Very good DIN 0.69 Satisfactory -9.19 Very good 0.53 Good 0.85 Very good DON 0.38 Very good 18.90 Good 0.85 Very good 0.96 Very good TP 0.50 Very good 23.80 Satisfactory 0.75 Very good 0.93 Very good

182 Source Catchments modelling – Updated methodology and results

RSR PBIAS NSE R2 Rating Rating Rating Parameter Rating (Moriasi Value Value (Moriasi Value (Moriasi Value (Moriasi 2007) 2015) 2015) 2015) DIP 0.46 Very good -4.13 Very good 0.79 Very good 0.83 Very good PP 0.58 Good 28.76 Satisfactory 0.67 Very good 0.93 Very good DOP 0.89 Indicative 44.41 Indicative 0.21 Indicative 0.50 Indicative

Table 99 RC2015 (120310A Suttor River at Bowen Development Road, n = 8 years for flow and n = 8 years for WQ)

RSR PBIAS NSE R2 Parameter Rating Rating Rating Rating Value (Moriasi Value (Moriasi Value (Moriasi Value (Moriasi 2007) 2015) 2015) 2015) Flow 0.43 Very good -3.99 Very good 0.82 Very good 0.77 Good TSS 1.28 Indicative -99.47 Indicative -0.63 Indicative 0.64 Satisfactory TN 0.42 Very good 31.25 Indicative 0.82 Very good 0.80 Very good PN 0.37 Very good 19.97 Good 0.86 Very good 0.77 Very good DIN 5.37 Indicative -391.86 Indicative -27.79 Indicative 0.83 Very good DON 0.69 Satisfactory 58.37 Indicative 0.53 Good 0.76 Very good TP 0.39 Very good -10.74 Very good 0.85 Very good 0.72 Very good DIP 0.55 Good 34.64 Indicative 0.70 Very good 0.57 Satisfactory PP 0.73 Indicative -48.95 Indicative 0.46 Satisfactory 0.74 Very good DOP 0.92 Indicative 76.05 Indicative 0.15 Indicative 0.48 Satisfactory

Mackay Whitsunday model validation

Figure 64 Long-term Annual PN Loads Modelled against FRCE and GBRCLMP Estimates in the Pioneer River at Dumbleton Pump Station

183 Source Catchments modelling – Updated methodology and results

Figure 65 Long-term Annual PP Loads Modelled against FRCE and GBRCLMP Estimates in the Pioneer River at Dumbleton Pump Station

Figure 66 Long-term Annual DIP Loads Modelled against FRCE and GBRCLMP Estimates in the Pioneer River at Dumbleton Pump Station

184 Source Catchments modelling – Updated methodology and results

Figure 67 Long-term Annual DON Loads Modelled against FRCE and GBRCLMP Estimates in the Pioneer River at Dumbleton Pump Station

Figure 68 Long-term Annual DOP Loads Modelled against FRCE and GBRCLMP Estimates in the Pioneer River at Dumbleton Pump Station

185