Confidential Report

Home , Cape Sorell

CONFIDENTIAL: VERSION 3 (FINAL)

CONFIDENTIAL REPORT

MACQUARIE HARBOUR DISSOLVED OXYGEN WORKING GROUP

6 OCTOBER 2014

CONFIDENTIAL: VERSION 3 (FINAL)

SUMMARY AND RECOMMENDATIONS 4

1. Changes in dissolved oxygen levels 4

2. Organic carbon loads and biological oxygen demand 4

3. Replenishment of dissolved oxygen by physical processes 4

4. Attribution 5

5. Recommendations 5

1. BACKGROUND 7

Biochemical processes 8

Physical processes 9

2. DATA DESCRIPTION AND PROCESSING 12

3. ORGANIC CARBON LOADS AND BIOLOGICAL OXYGEN DEMAND 14

Organic carbon loads 14

Biological oxygen demand 17

4. REPLENISHMENT OF DISSOLVED OXYGEN BY PHYSICAL PROCESSES 27

Long-term trends 27

High river discharge mixing events 27

Deep water recharge events – low frequency variability 35

Deep-water recharge events – higher frequency characteristics 36

4. DISCUSSION 40

REFERENCES 42

APPENDIX A: TERMS OF REFERENCE 44

Macquarie Harbour Dissolved Oxygen (MH DO) Working Group 44

APPENDIX B: CATCHMENTS AND RIVER DISCHARGES 47

Catchment data 47

Gordon River catchment 49

King River catchment 51

2 CONFIDENTIAL: VERSION 3 (FINAL)

Other catchments 53

APPENDIX C: SEALEVEL AND TIDES 54

APPENDIX D: WIND CHARACTERISTICS 55

APPENDIX E: WATER TEMPERATURE 58

In-situ temperature data 61

APPENDIX F: DISSOLVED OXYGEN QA/QC 63

3 CONFIDENTIAL: VERSION 3 (FINAL)

Summary and Recommendations

1. Changes in dissolved oxygen levels 1.1. There is a clear downward trend in the dissolved oxygen (DO) levels of the deep-waters (> 15m) of Macquarie Harbour over the period 2009- present. 1.2. DO levels less than 2 mg/l are now very common below 20 m and occasionally come to within 12 m of the surface. 1.3. There have been a number of significant changes over the period from 2009-present. River flow was historically low between 2009-12 and historically high in 2013. This period also coincides with a major expansion of salmon aquaculture.

2. Organic carbon loads and biological oxygen demand 2.1. Biological oxygen demand (BOD) in the harbour is controlled by the supply of labile (biologically available) organic carbon and potentially the conversion of refractory organic carbon to labile carbon. 2.2. It has been estimated that the total organic carbon load associated with river discharge into Macquarie Harbour is around 100 times that of salmon production in the harbour (ocean inputs are also expected to be small). However, the fraction of organic carbon in river discharges that is labile has not been measured in Macquarie Harbour and in other well forested catchments has been found to be small (Sun et al. 1997, Moran et al. 1999). The fractions of labile and refractory organic carbon exported to the ocean from the harbour are also unknown. 2.3. Based on the data available from recent in situ measurements aquaculture is estimated to be responsible for between 3 and 12% of the benthic BOD in Macquarie Harbour (for sediments deeper than 15 m). The remaining benthic BOD is presumably associated with particulates in river discharge and detritus derived from biological production within the harbour. 2.4. Previous studies have found that pelagic BOD is significantly higher than benthic BOD in many estuaries. However, pelagic BOD has not yet been measured in Macquarie Harbour.

3. Replenishment of dissolved oxygen by physical processes 3.1. Dissolved oxygen in the deep waters of Macquarie Harbour is mainly replenished through (i) mixing with higher DO surface waters; and (ii) higher DO ocean waters overflowing the sill at the mouth of the harbour and descending as a dense plume that recharges DO near the bottom. 3.2. Mixing with higher DO surface waters is most effective following flood events, where a large fraction of the upper water column has been displaced by relatively fresh high DO water. Weak stratification within this upper layer then allows energy from winds and currents to penetrate and continue to mix DO into the deep waters. This process has been observed to influence DO through the entire water column during

4 CONFIDENTIAL: VERSION 3 (FINAL)

flood events such as the series of events that occurred in July and August 2013. However, these conditions were rare through the low river discharge period of 2009-2012. 3.3. Recharge of bottom water is most common under low river discharge conditions when ocean inflow is not obstructed by the outflow of freshwater over the harbour sill. Wind, tide and pressure also play an important part in determining if salt water from outside of Hells Gate can enter the harbor. A significant number of recharge events were observed over the period 2009-present. However, with the exception of summer 2011-2012, the total input of oxygen during each event was quite small and enhanced DO near the bottom of the harbour tended to be rapidly diffused and/or consumed through respiration.

4. Attribution 4.1. There are a number of limitations in our knowledge about the biogeochemical processes in Macquarie Harbour that preclude any definitive attribution of the recent decline in deep-water DO. Primary among these are the absence of reliable estimates of labile organic carbon fluxes associated with river discharge and export to the ocean, and levels of pelagic BOD. 4.2. Relatively stable levels of average DO between 1993 and 2009 suggest there is tight coupling between demand and replenishment of DO in Macquarie Harbour, which could be influenced by additional labile organic carbon loads. However, limitations with the data available preclude an unambiguous assessment of the drivers of the deep water decline in DO since 2009. 4.3. Historically low and high river discharge and associated organic carbon loads are likely to have influenced both the physical resupply of oxygen and the BOD of bottom waters. 4.4. Historically, DO levels appear to have been maintained by a combination of regular (bottom-up) recharge and (surface-down) mixing events driven by highly variable river discharge. However, for 2009-present rainfall and river discharge conditions may have altered the frequency of these events. Detailed hydrodynamic modelling and continuous monitoring are required to quantify this relationship. 4.5. The rapid transition to high river discharge (2013) after a prolonged period of low discharge (2009-2012) is likely to have delivered an accompanying influx of accumulated organic carbon from the catchment. However, the influence of this influx on BOD in bottom waters is unclear. Measurements of labile and refractory loads and modelling to estimate retention verses export to the ocean are required to assess the role of river organic carbon loads on bottom water BOD.

5. Recommendations 5.1. Existing monitoring, including recently incorporated TOC, DOC and TN sampling, should be continued and supplemented by a strategic field campaign aimed at measuring the influence of organic carbon loads in river discharge. The latter should measure: (i) labile and refractory

5 CONFIDENTIAL: VERSION 3 (FINAL)

loads; (ii) rates of deposition and retention in the harbor; (iii) conversion of refractory to labile carbon; and (iv) the relative importance of pelagic and benthic BOD. 5.2. Similarly, the role of fish farm carbon and nitrogen inputs on BOD should be extended to provide greater resolution of the spatial extent of deposition and the influence on benthic and pelagic BOD. 5.3. Measurements in 5.1 and 5.2 should be used to support development of a more detailed quantitative carbon/nitrogen/oxygen budget for Macquarie Harbour. 5.4. Hydrodynamic and biogeochemical models should be used to help constrain the carbon/nitrogen/oxygen budgets and to explore past conditions and future scenarios relating to changes in river flows and/or salmon production. 5.5. Options for adaptive farm management approaches that allow timely responses to DO fluctuations should be explored. Such approaches could include implementing a Decision Support System (DSS) with a particular focus on oxygen.

6 CONFIDENTIAL: VERSION 3 (FINAL)

1. Background

In February 2014 the Tasmanian Salmon Growers Association (TSGA) established the Macquarie Harbour Dissolved Oxygen (MHDO) Working Group with the purpose of verifying the scope of dissolved oxygen (DO) reductions in the bottom waters of Macquarie Harbour and, to the extent allowed by available data, determine attribution. The Terms of Reference for this group are provided in Appendix A. This report documents the finding of the MHDO Working Group.

Macquarie Harbour is a large estuary supplied with freshwater from the Gordon and King Rivers (Figure 1.1). The harbour is approximately 33 km long and 9 km wide with a total surface area of 276 km2. While its maximum depth is around 50 m, a shallow sill at its mouth (< 5 m) restricts exchanges with the ocean. This isolation of deep water in the harbor has resulted in a naturally low DO environment (Cresswell et al. 1989) that may be vulnerable to further increases in oxygen demand. DO levels have been declining since 2009 (Figure 1.2) and further decreases may have a direct impact on both the ecology of the harbour and aquaculture production.

Figure 1.1: Bathymetry in Macquarie Harbour (Lucieer et al. 2009).

7 a) EPA Site 12 b) EPA Site 12

) 120 35

i t

a 30

r 100

u t

a 25

)

80 t

(

( 20

t e

60 i

i y

l 15

O S

d 10

v l

o 20 5

s i

D 0 0 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 CONFIDENTIAL: VERSION 3 (FINAL)Date Date

c) EPA Site 27 a) EPA Site 12 b) EPA Site 12 d) EPA Site 27

) 120 35

t ) 120a 30 35

r 100

n a

s 25

80 )

(

( 20

a 30 y

e 60

i g

r 100

y i

l 15

O t

40 S

d 10

a v

l 25 s

o 20

) s

t s

80i D 0 p

% 0 p

( 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5

9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1

1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ( 20

1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

n Date

Date y

t e

60 i

n i

y e) EPA Site 34 f) EPA Site 34

l 15

x a

) 120 35

S t

a 30

r 100

u t d 10

a 25

)

80 t

( p

(

l 20

y t

e 60 i

o 20

i y

l 15 5

O S

s i

d 10

l D

o 20 5 s

s 0 0 i D 0 0 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 3 3 44 5 6 57 8 69 0 17 2 3 84 5 96 7 80 9 011 2 23 4 35 4 5 36 4 5 76 7 88 9 09 1 2 03 4 15 6 72 8 930 1 42 3 54 5 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 9 9 9 9 9 99 9 99 0 09 0 090 0 90 0 0 0 00 0 0 00 0 00 0 0 90 9 9 09 9 09 9 0 0 010 0 10 0 10 0 010 0 10 0 10 0 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 9 1 91 1 19 1 191 2 92 2 92 2 292 2 02 2 02 2 20 2 2 02 0 0 01 1 01 1 101 1 02 2 02 2 20 2 2 02 2 02 2 20 2 2 02 2 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Date Date 0-5 m 5-15 m 15-25Date m 25-35 m 35-45 m Date

Figure 1.2: Long-term trend in dissolved oxygen within a number of depth ranges at EPA site 12. c) EPA Site 27 d) EPA Site 27 In order to resolve concerns regarding DO concentrations in the bottom waters

of Macquarie) 120 Harbour, a number of factors needed to be investigated and 35

n o

understood.i These include the processes that control DO in the harbour and t

a 30

whetherr the100 relatively low levels of deep-water DO observed recently are u

unusualt or part of the long-term natural variability of the system. Our a

s 25

understandin g of the contribution of aquaculture activities to the carbon and

80 )

t %

oxygen budgets of Macquarie Harbour needs to be assessed in the context of p

(

( 20 n

other sources of organic carbon that also drive biological oxygen demand (BOD), y

e 60 t

such as fluctuating river flows, oceanic inputs and biological production within i

y i

l 15

the harbour,x as well as the influence of physical exchanges that help replenish DO

a O

levels. 40 S

d 10

e v Biochemicall processes

o 20

s 5

DO decliness when BOD associated with microbial activity in the water column i

(pelagic)D and at the water/sediment interface (benthic) is greater than the 0 0 supply of DO by 3physical4 5 6 processes.7 8 9 0 It1 has2 3been4 5 demonstrated6 7 8 9 0 across1 2 3 other4 5 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 systems with 1similar1 1 geomorphology1 1 1 1 2 2 2 (e.g.2 2fjord2 2 systems2 2 2 ) that2 2 oxygen2 2 2 2 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 consumption per unit volume increases withDate decreasing basin depth and decreasing sill depth (Aure and Stigebrandt 1989). There are three primary Date sources of organic carbon driving BOD (Figure 1.3):

(a) River loads carry organic carbone) entering EPA Sitefrom 34the surrounding catchments. f) EPA Site 34 The BOD associated with this load is critically dependent on the mix of labile ) 120 35

materialn (readily biodegradable) and refractory material (low

o i biodegradability)t , as well as the more gradual conversion of refractory into

a 30

r 100

labile material. Some of the river load will remain in the surface waters and

u t

eventuallya discharge into the ocean, while the remainder will mix or sink into 25

)

80 t

( p

( 20

y t e 60

8 i

i y

l 15

O S

d 10

v l

o 20 5

s i

deeper waters as particulates. The proportion of the river load retained in the harbour is likely to increase during periods of low river discharge. (b) Local biological production in the harbor converts carbon dioxide in surface waters to organic carbon, which is eventually deposited into the deeper waters as detritus. While a significant proportion of detritus is highly labile, the underlying rates of photosynthesis are generally limited by the availability of light in Macquarie Harbour. (c) Aquaculture contributes highly labile organic carbon through fish excretion (and potentially settling of uneaten food pellets). The associated BOD can be estimated in terms of the feed inputs.

The balance of inputs and outputs are likely to be strongly dependent on the river conditions, both in terms of the flows and the loads of labile and refractory material (EIS 2011).

Figure 1.3: Schematic representation of the main organic carbon inputs and outputs in the harbor.

Physical processes DO declines when BOD in the water column (pelagic) and at the water/sediment interface (benthic) is greater than the supply of DO by physical processes. Three physical processes typically supply oxygen to deep-water in shallow-silled estuaries:

(i) Convective overturning occurs when the surface layer becomes denser than the deep water either through strong cooling or increased salinity due to strong evaporation. (ii) Vertical mixing gradually mixes surface and deeper waters so that bottom water salinity decreases and DO levels increase. The strength of vertical mixing is dependent on a number of processes. In the simplest sense, increasing stratification (caused by increasing river discharge) inhibits vertical mixing, while high winds or strong currents (caused by tides or high river discharge) enhance vertical mixing. Vertical mixing can be further enhanced over sloping bathymetry by internal wave breaking. (iii) Deep-water recharge occurs when dense, oxygen-rich, oceanic water penetrates over the shallow entrance sill and intrudes into the deep- water as a plume. Existing (low DO) water in the basin is forced upward sometimes resulting in a DO minimum at mid-depths. In most shallow- silled estuaries, recharge events occur episodically, with events typically

9 CONFIDENTIAL: VERSION 3 (FINAL)

lasting a few days. During the intervals between events, the deep-water tends to be isolated from the general circulation in the estuary and its properties are only modified through more gradual vertical mixing processes.

Of these three processes, convective overturning can probably be dismissed as a factor in Macquarie Harbour, since neither evaporation nor cooling are likely to be strong enough to overcome the strong salinity stratification. We have therefore focused on examining recharge and vertical mixing processes.

The surface layer gains its properties from mixing of river discharge with existing surface waters as it spreads across the harbour and eventually flows over the sill to mix with the ocean. The average residence time in the harbour of this water mass has been estimated at 50 days in spring and summer and less in other seasons (EIS 2011). However, the deeper layers will have much longer residence times dependent on the rate of recharge and vertical mixing.

Available measurements (described in more detail below) suggest that river discharge can influence the DO conditions in Macquarie Harbour through both vertical mixing and deep-water recharge events. Mixing with higher DO surface waters is most effective following flood events, where a large fraction of the upper water column has been displaced by relatively fresh high DO water (Figure 1.4a). The weak stratification of this upper layer remaining after the flood event then allows current and wind energy to penetrate and continue mixing DO into the deep waters. Under these lower river discharge conditions, the recharge process can also begin to replenish deep-water DO levels by displacing bottom waters (Figure 1.4c).

Other forcing such as tides and winds moderate these processes (Allen and Simpson 1998, Inall and Gillibrand 2010). For example, in some estuaries tidal mixing over the sill has been found to dilute the salinity of oceanic water entering the system, thereby restricting renewal of deep water to periods around the neap tide (Geyer and Cannon 1982). Alternatively, winds blowing brackish surface waters out of an estuary can increase the return flow of oceanic water and enhance recharge rates (e.g. Gillibrand et al. 1995), while other strong wind patterns will tend to trap freshwater in the system and favour vertical mixing and the inflow of oceanic water into the harbour.

10 CONFIDENTIAL: VERSION 3 (FINAL)

Figure 1.4: Schematic representation of physical processes operating in the harbor under a range of river discharge conditions. Typical forms of the associated DO profiles are shown along the right hand side. See Box 1 for flow definitions

BOX 1 Flow Definitions For the purpose of this study, low flow is defined as a monthly average discharge of less than 300 cumecs for total river discharge into Macquarie Harbour. High flow is defined as monthly average discharge exceeding 600 cumecs for total river discharge into Macquarie Harbour1. These definitions are based on the river flow conditions that broadly encompass the observed deep-water recharge and vertical mixing events. It is important to recognise that only a limited number of events were observed based on the available monthly data for dissolved oxygen. As such, these river flow definitions should not be regarded as definitive bounds on the conditions suitable for these processes. As we discuss, other factors such as wind, tide and pressure are also likely to be important in regulating these events. The recent deployment of continuous data loggers in the harbour will provide greater insight into the combination of drivers required to drive deep-water recharge and vertical mixing events.

It is also important to note that the relative importance of inflow variability and freshwater volume in the harbour for recharge and mixing processes cannot be unravelled based on the available monthly data for dissolved oxygen. The relationship with monthly average discharge indicates that pre-conditioning, and thus freshwater volume is important, however a more recent recharge event occurred during a period of significant flow variability. Again, the availability of continuous dissolved oxygen data from the harbour will provide greater insight into the important attributes of the drivers required for recharge and mixing processes.

1By way of contrast, the maximum discharge from the Gordon River Power station is approximately 270 cumecs.

11 CONFIDENTIAL: VERSION 3 (FINAL)

2. Data description and processing

The analyses presented in this report are based entirely on empirical data, which has been interpreted within the context of the conceptual models outlined above. Data sets relevant to DO levels in Macquarie Harbour can be divided into those associated with external forcing of the system (river discharge and loads, sea level, winds) and those associated with conditions in the harbour (temperature, salinity, DO, organic carbon, nutrients, currents). They have been collected across a wide range of spatial and temporal scales. The geographical distribution of the in-harbour sampling is summarized in Figure 2.1 and further details on periods and frequency of sampling for each data type are summarized in Table 2.1. Further detail on QA/QC assessment of the critical dissolved oxygen measurements can be found in Appendix F.

Figure 2.1: Distribution of sampling in Macquarie Harbour.

12 CONFIDENTIAL: VERSION 3 (FINAL)

Table 2.1: Summary of the main data types, locations and, and sampling period and frequency used in the study.

Data Type Location Sampling duration/frequency

Predicted Tidal Elevation Cape Sorell 1999 - 2014

Sealevel Strahan Wharf 09/05/2002 - 01/04/2014

Currents (ADCP) Liberty Point 20/07/2009 - 02/09/2009 Table Head 20/07/2009 - 24/08/2009 Eastern edge of Table Head 24/08/2009 - 02/09/2009 Table Head Central 26/10/2010 - 16/11/2010

River Discharge Collingwood River Below Alam River Hourly mean flow and temperature from 1990 to April 2014 Franklin River at Mt Fincham Track (Basslink Report presents plots Franklin Downstream dating back to 1958) Gordon at Power Station trailrace Gordon above Denison Gordon above Franklin Gordon Total at MH including Franklin

Wind (BOM) 3 hourly wind speed and direction Strahan Airport 01/09/1976 - 01/03/2014 Cape Sorell 01/01/1957 - 01/03/2014

Air Pressure Strahan Airport 16/06/1999 - 01/04/2014 Cape Sorell 01/01/2009 - 01/04/2014

Water quality data from DPIPWE Up to 35 stations across the harbour 1991 – 2014 (profiles) supplied by EPA: take 3-4 times per year. 1996 – 1998; 2013 (ammonia, - parameters incl. temperature, nitrate) salinity and DO profiles 2012/13 (TOC) - nutrient concentrations include ammonia, nitrate, DOC,TOC as 1997-2001 (DOC) surface, mid and bottom depths

Industry Monthly Monitoring data: 12 locations within Macquarie Harbour 15/10/2011 - 15/7/2014 and 1 station outside Hells Gate - temperature, salinity and DO profiles Additional stations (4) have been added since Nov-2013 - ammonium, nitrate, TN at 2m and 20m at 7 stations - chlorphyll-a at 2m and several stations at 12m - phytoplankton counts at 13 stations (2m)

Continuous data loggers (DO, Industry sites 23 (20, 30 and 40 m), 31 June 2014 – present Temperature and conductivity) (20, 30 and 35 m) and 46 ( 20 and 30 m)

EPA (next to WH2, see fig 2.2). Single logger string with sensors at 7 and 35m Nov 2013 – present FRDC Skate project . Liberty Point sites LPO5,LPO6,LPO7 in 7, 15 and 20m Nov 2013 - present respectively

13 CONFIDENTIAL: VERSION 3 (FINAL)

3. Organic carbon loads and biological oxygen demand

Organic carbon loads To understand the drivers of biological oxygen demand (BOD) in the harbor, we first need to estimate the relative contribution of the primary sources of organic carbon, namely river loads, local biological production and aquaculture inputs.

Catchment inputs The major sources of catchment organic carbon to Macquarie Harbour are the King River, Gordon River and Birch’s Inlet. Organic carbon loads from each of these sources were previously estimated (EIS 2011) using data obtained from the online database of the Australian Natural Resources Atlas (ANRA). The organic carbon load was further separated into dissolved and particulate fractions using prescribed ANRA values (Table 3.1).

There was very little local data to verify the ANRA values. In the case of organic carbon it was limited to dissolved organic carbon (DOC) data provided by the Tasmania EPA from sampling during 1997-2001. The flow weighted concentrations of DOC derived using the ANRA load data (4.69 mg/l) fell within the range measured by the Tasmanian EPA (4- 6 mg/l) and the DOC fraction (Table 3.1) was comparable to estimates from other blackwater rivers (Wetzel et al. 1995). However, more recently (2012-2013) the Tasmanian EPA has measured total organic carbon (TOC) concentrations throughout the harbour. The relationship of TOC concentrations with salinity indicates that river discharge into the harbour is delivering concentrations higher than previously assumed (Figure 3.1) with concentrations measured close to a salinity of 0 ppt indicating a source concentration of 10-12 mg/l.

To estimate total annual catchment loads of organic carbon corresponding to the period 1993-2013, a source concentration of 10 mg/l was assumed and multiplied by the total annual catchment flow (Figure 3.2). Because there was no data available for the recent sampling, we applied the same dissolved to particulate ratios provided by the ANRA.

Table 3.1: Summary of organic carbon loadings for King River, Gordon River and Birch’s inlet based on ANRA 2009 (EIS 2011). (* Catchment weighted concentrations.)

Gordon King River Birch’s Inlet Total River Mean annual flow (m3/s) 81.58 224.83 26.33 332.74 DOC (mg/l) 2.19 5.79 3.12 4.69* POC (mg/l) 0.55 1.45 0.78 1.17* Total organic carbon load (t/yr) 7050 51296.5 3238.9 61585.4 Despite significant inter annual variability in loads since 2007 there was a general decline in estimated loads until 2013. This was due to low rainfall and the implementation of the hydro rebuild strategy in June 2008 (Basslink 2013). In early 2013, river flows and thus estimated loads increased dramatically due to

14 CONFIDENTIAL: VERSION 3 (FINAL)

increased power generation at the Gordon Power Station in response to the fixed carbon price period (Basslink 2013). Further work is required to reduce uncertainty in river organic carbon loads, in particular, measurements of flow weighted concentrations and the reactivity of dissolved and particulate fractions.

)

(

n 10

a g

r 5

o t

0 0 5 10 15 20 25 30 35 Salinity (ppt)

Figure 3.1: Mixing plot of salinity (ppt) verses total organic carbon concentration (mg/l).

150000 dissolved organic carbon )

s particulate organic carbon

t (

100000

i n

a 50000

o t

0 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 year

Figure 3.2: Annual load estimates for total organic carbon entering the harbor from all major catchment sources.

Marine farming inputs Organic carbon inputs from fish farms are based on total feed input data provided by industry. Following the methods used in the EIS (2011), the volume of feed converted to faeces was estimated to be 21.2%, with approximately 20% of the faeces considered to be organic carbon. Consistent with industry

15 CONFIDENTIAL: VERSION 3 (FINAL)

expansion in the last 5-10 years, there has been a concomitant increase in organic carbon inputs (Figure 3.3).

1000

)

n n

o 800

(

e c

e 600

i 400

r a

c 200

o t 0 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 year Figure 3.3: Annual load estimates of total faecal organic carbon from fish farming in Macquarie Harbour.

Oceanic inputs There are few measurements available outside of Hell’s Gates to establish the exchange of organic carbon and other nutrients between the open ocean and harbour (EIS 2011). While there is ongoing monthly monitoring at one site outside of Hell’s gate, the tidal phase at the time of sampling is variable and the water sampled is most often a mix of both oceanic and freshwaters. Hence it is difficult to estimate the load of nutrients in either direction. Nonetheless, the relationship between salinity and total organic carbon concentration (Figure 3.1) indicates that the marine source concentration is significantly lower than the freshwater source.

Despite not having good estimates of the exchanges, it seems unlikely that inputs from the ocean are a significant source of organic carbon to the harbour relative to the other sources. It is more likely that a significant fraction of catchment organic carbon is exported to the ocean in the freshwater lens that often extends through Hell’s Gates. The only exception may be when a large phytoplankton bloom in the adjacent coastal waters, such as surface diatom blooms recorded off Ocean Beach (McLachlan et al. 1984), are swept into the harbour by the flood tide.

Internal production In the absence of primary productivity measurements in the harbour we must rely on Chlorophyll-a as an indicator of the biomass of photosynthetic organisms present in the harbour and as a proxy for organic carbon generated though primary production. For the purposes of this investigation determining whether there has been a change in primary production in the harbour is more critical than estimating absolute rates of organic carbon production per se. Data collected during industry monitoring indicates strong seasonality in water

16 CONFIDENTIAL: VERSION 3 (FINAL)

column production with peak biomasses through spring-summer (Figure 3.4). There is also some suggestion that production may have been higher in spring 2012, possibly related to the preceding low river flows and a concomitant increase in light penetration. However, there is no pattern that would indicate a stepwise increase in water column productivity over the course of the monitoring period.

8 a) KR1

) 6

(

l l

y 4

l h C 2

1 2 2 2 2 3 3 3 3 4

1 1 1 1 1 1 1

1 1 1

0 0 0 0

0 0 0

2 2 2 2

2 2 2

l l

r r

t t t

n n n

u u

c c c

p p

a a a

J J

A A

J J J

O O O 8 b) CHE

) 6

(

l l

y 4

l h C 2

1 2 2 2 2 3 3 3 3 4

1 1 1 1 1 1 1

1 1 1

0 0 0 0

0 0 0

2 2 2 2

2 2 2

l l

r r

t t t

n n n

u u

c c c

p p

a a a

J J

A A

J J J

O O O

8 c) WH2

) 6

(

l l

y 4

l h C 2

1 2 2 2 2 3 3 3 3 4

1 1 1 1 1 1 1

1 1 1

0 0 0 0

0 0 0

2 2 2 2

2 2 2

l l

r r

t t t

n n n

u u

c c c

p p

a a a

J J

A A

J J J

O O O

Figure 3.4: Chlorophyll a concentrations (mg/l) measured during monthly industry surveys at 2m depth at sites a) KR1; b) CHE; and d) WH2.

Biological oxygen demand The consumption of oxygen can occur via benthic and pelagic biochemical processes. Benthic rates of oxygen consumption were estimated for farm and non-farm sources using empirical data collected under FRDC project 2012-047. Pelagic rates of consumption have not been measured in Macquarie Harbour. We have therefore estimated the total oxygen consumption (pelagic and benthic) using the source loads of carbon and measured ratios of oxygen consumed to

17 CONFIDENTIAL: VERSION 3 (FINAL)

dissolved inorganic carbon produced. It is important to note that this is problematic for catchment inputs due to the absence of information on the proportions of refractory and labile catchment organic carbon and the volume of catchment organic carbon leaving though Hell’s Gates. These uncertainties are particularly significant given the magnitude of the source (~ 100 fold greater than fish farm organic carbon).

Benthic oxygen demand Benthic oxygen consumption was measured at 18 sites throughout the harbour in November 2012 and repeated at a subset of these sites in January, May and September 2013. Sites were distributed across both farm (local) and harbour (regional) scales. At the farm scale this included 2 sites at the edge of the cage edge (directly under the cage edge); one site 50m from the cage; and a control site a further 1 km away (Figure 3.5). This configuration was used for each of 2 deep-water farms and 1 shallow-water farm. At the harbour scale the remaining 6 sites covered the major boundaries, central harbour, and World Heritage Area (CHN, WH2, KR1, ECS, WH1, CH1).

Benthic BOD measured at the cage sites was significantly elevated compared to nearby control sites, while that at the 50 m sites was only slightly elevated (Figure 3.6). The oxygen consumption rates at the cage sites were comparable to those recorded elsewhere under salmon cage aquaculture (e.g. Hargrave et al. 1993; Pereira et al. 2004. The localised nature of the depositional footprint from the cages is consistent with what has been observed elsewhere under salmon aquaculture (See EIS 2011 and references therein).

There was no clear seasonal variation in BOD at control sites, but significant variation at farm sites (Figure 3.7). In the absence of more detailed spatial and temporal sampling, average fluxes (across space and time) were used in the calculations to estimate benthic BOD at both farm and control sites. As such, this should be considered a first order estimate and treated with a degree of caution given the size of the area and scaling required.

The assessment was restricted to benthic sediments in water deeper than 15 m, corresponding to the area of significant oxygen decline. The area was further split into zones of 15-25 m and > 25 m, differentiating between shallow water farms on the southwest shore and deeper water farms along the central harbour. The BOD measurements confirmed differences in both farmed and non-farmed areas between the depth zones. Within each of these zones, the fluxes were scaled as described in Figure 3.5.

18 CONFIDENTIAL: VERSION 3 (FINAL)

Area B Area C

Area A

100m ~1km

50m

Figure 3.5: Areas used to calculate benthic oxygen demand (not to scale) using benthic flux measurements (red crosses). For the area encompassing the cages at a lease (defined as a line joining the outer edges of all cages), the flux measurements taken at the cage edge sites were scaled up to this area (assuming that higher fluxes under the centre of the pen were offset by lower fluxes between cages. A buffer area was then defined as extending 100m out from the cage area and fluxes measured at the 50m from cage sites were assumed to represent the average flux in this area. The control or background area was defined as the total area of seabed deeper than 15m not including the buffer and cage areas.

)

- -1000



(

e -2000

d e

v -3000

i D

-4000

) )

)

2 1

1 2 1 2 1 2

m m m

m m

H H

H 0 0 0

E E E E E E

k k

5 5 5

W W

1 1

G G G G G G

1 2 3

A A A A A A

1 3

C C C C C C

M M M

1 1 2 2 3 3

R R R

A A A

M M M M M M

F F F

(

R R R R R R

(

A A A A A A

F F F F F F

E C P Figure 3.6: Rates of oxygen consumption at harbour and farm scales (2 cages at each farm) in November 2012. Note that the second of the Farm 2 sites was fallowed during the survey.

19 CONFIDENTIAL: VERSION 3 (FINAL)

a) Farm 1

Nov 12 Jan 13 May 13 Sep 13

)

r h

-1000

u (

-2000

l f

0m (cage 1)*

n e

g 0m (cage 2)*

y -3000 x 50m O 1000m *dashed line is fallowed period -4000 b) Farm 2 Nov 12 Jan 13 May 13 Sep 13

)

r h

-1000

u (

-2000

e g

y -3000

x O

-4000

Figure 3.7: Rates of oxygen consumption at a) farm 1 and b) farm 2 at the cage, 50m and 1km control sites through time. The dashed line indicates fallowing. For the purposes of the benthic budget we used the mean flux through time.

The total area in Macquarie Harbour below 15 m depth is 120 km2; of which 0.66 km2 (0.55%) was classified as cage area, 2.05 km2 (1.70%) as buffer, and the remaining 117.3 km2 (97.75%) as control (based on early 2012 cage distributions, although late 2013 distributions gave quite similar results). The BOD rates in the respective areas can be found in Table 3.2. Representing 2.3% of the benthic area (> 15 m), the cage and buffer area account for approximately 5% of the total benthic BOD (Table 3.2). This equates to an increase of almost 3% if we estimate BOD with and without farms. In the context of the total amount of DO in the water column below 15 m, total benthic BOD accounts for approximately 0.5 mg l-1 month-1 of which farm (cage and buffer) sediments account for approximately 0.025 mg l-1 month-1.

An important consideration when interpreting these budgets is that they assume that the sediment incubations capture the net oxygen consumption due to carbon mineralisation. In coastal sediments anaerobic respiration of carbon is often dominant, in which case alternative oxidants (i.e. nitrate, manganese and iron hydroxides, sulphate) are used when the demand of oxidants exceeds the supply of oxygen (Middleburg et al. 2004). In Macquarie Harbour sediments, the rate of dissolved inorganic carbon production (DIC) often exceeds the rate of

20 CONFIDENTIAL: VERSION 3 (FINAL)

oxygen consumption, particularly in enriched farm sediments (Table 3.2), indicating anaerobic respiration is common. The reduced compounds produced during anaerobic respiration (e.g. hydrogen sulphide produced during sulphate reduction of organic matter) can then be reoxidised, potentially consuming oxygen in the process. Carpenter et al. (1991) reported the presence of sulphate- reducing bacteria in Macquarie Harbour.

Here we assume that the incubation estimates of oxygen consumption encompass both oxidation and reoxidation. It is likely that there would be some reoxidation (utilizing oxygen) of reduced compounds in the water column that may not be captured in the incubations (e.g. reoxidation of ammonium and sulphide). However, reduced compounds are also likely to be removed from solution and bound in sediments (e.g. metal sulphides) and thus, may not lead to an increase in oxygen consumption (see Middleburg and Levin 2009). The degree to which these reduced compounds ultimately influence bottom water oxygen is difficult to estimate based on available information. A very conservative estimate would be to assume that all organic matter mineralisation will ultimately lead to an oxygen debt. To do this we have used measured DIC fluxes (i.e. rates of carbon minerilisation) and scaled those in the same way as described for the oxygen fluxes. Theoretical oxygen demand was then calculated based on a 1:1 molar ratio, in which case the cage and buffer area account for approximately 13% of the total benthic BOD, with aquaculture potentially accounting for approximately 12% of this (i.e. after removing the BOD that would occur naturally in this area in the absence of farming). The increase is because anaerobic respiration makes up a far bigger proportion of total respiration in cage and buffer sediments compared to control sediments.

Table 3.2: Benthic BOD rates in the cage, buffer and control zones defined in Figure 3.5, with corresponding benthic oxygen consumed and DIC produced for farm and non-farm areas.

Oxygen Depth Oxygen Area (km2) consumption DIC produced range consumed Zone rate tonnes tonnes m km2 % mol/m2/hr % % /day /day 15-25 0.17 900 0.341 Cage (farm) 0.49 1500 0.190 > 25 2.3 4.9 1.27 12.6 15-25 0.44 350 0.432 Buffer (farm) 1.61 750 0.245 > 25 Control (non- 15-25 53.91 240 9.937 97.7 95.1 8.84 87.4 farm) > 25 63.43 280 13.639

Totals 120.05 100.0 24.784 100.0 10.11 100.0

Nutrient generation As a result of fish farming, we might expect to see an increase in ammonia concentrations in surface waters where there are farms due to fish excretion, but also in bottom waters due to the mineralisation of farm waste (faeces/feed) and release of ammonia. Ammonia can also enter the harbour from the catchments

21 CONFIDENTIAL: VERSION 3 (FINAL)

either directly as dissolved ammonium or indirectly via organic matter and its subsequent mineralisation in the water column and sediments. As such we might also expect to see a change in ammonia concentrations in the harbour due to changes in catchment inputs, most notably between 2012 and 2013 when organic matter loads are likely to have increased dramatically.

Any change in bottom water nutrient concentrations also provides insight into the likely drivers of change in bottom water dissolved oxygen. If the decline in oxygen in bottom waters was due to biological demand, we might also expect an increase in the bottom water concentrations of the major products of organic matter mineralisation. The benthic flux study demonstrated that ammonia is the major nitrogen form produced, particularly at farm sites (Figure 3.8a). In contrast, nitrate is taken up by the sediments, most strongly at farm sites (Figure 3.8b). This is most likely because the process of nitrification that converts ammonium to nitrate is limited in Macquarie Harbour sediments due to the limited supply of oxygen. Denitrifying bacteria must therefore rely on nitrate sourced from the water column rather than that produced via sediment nitrification (Cornell et al. 1999).

There is no evidence that ammonium concentrations have increased during the period of industry monitoring (Figure 3.9) and these concentrations fall with the range recorded by the EPA from 1995-1998. However, this doesn’t preclude increased BOD and ammonia production because nitrification in the water column (as opposed to the sediments where it is limited due the lack of oxygen) may be converting the ammonia as soon as it is available.

Nitrate concentrations also do not appear to have changed over the industry monitoring (Figure 3.10), although the concentrations are elevated compared to those measured by the EPA from 1996-98. This stability in nitrate is despite an increase in farming and the likely spike in catchment inputs in 2013 and suggests a high degree of coupling between the process (e.g. ammonia production, ammonia conversion to nitrate {nitrification} and nitrate conversion to nitrogen gas {denitrification}). The fact that ammonia hasn’t built up in the bottom waters could indicate either no change in the rates of organic matter breakdown and ammonia production or an increase in water column nitrification that quickly converts the ammonium to nitrate. Similarly a lack of change in nitrate could indicate little change in mineralisation or an increase in denitrification that converts nitrate to nitrogen gas. Whilst we now have a reasonable understanding of nitrogen dynamics in Macquarie Harbour sediments, our knowledge of similar process in the water column is extremely limited. This understanding would help unravel the influence of organic matter loading on mineralisation rates and the ultimate fate of nitrogen.

22 CONFIDENTIAL: VERSION 3 (FINAL)

(a) 1300

) 900

- 500

l o

m 100

 (

100

i n

o 50

m m

A 0

) )

)

2 1

1 2 1 2 1 2

m m m

m m

H H

H 0 0 0

E E E E E

-50 E

k k

5 5 5

W W

1 1

G G G G G G

1 2 3

A A A A A A

1 3

C C C C C C

M M M

1 1 2 2 3 3

R R R

A A A

M M M M M M

F F F

(

R R R R R R

(

A A A A A A

F F F F F F

T H

(b) T

E C

50 P

)

- -50

 (

-100

t i

N -150

-200

) )

)

2 1

1 2 1 2 1 2

m m m

m m

H H

H 0 0 0

E E E E E E

k k

5 5 5

W W

1 1

G G G G G G

1 2 3

A A A A A A

1 3

C C C C C C

M M M

1 1 2 2 3 3

R R R

A A A

M M M M M M

F F F

(

R R R R R R

(

A A A A A A

F F F F F F

C P

Figure 3.8: Flux rates of ammonium and nitrate at harbour and farm scales in November 2012.

23 CONFIDENTIAL: VERSION 3 (FINAL)

0.015

a) KR1

)

l /

g 0.010

(

n o

m 0.005

m A

0.000

1 2 2 2 2 3 3 3 3 4

1 1 1 1 1 1 1

1 1 1

0 0 0 0

0 0 0

2 2 2 2

2 2 2

l l

r r

t t t

n n n

u u

c c c

p p

a a a

J J

A A

J J J

O O O 0.015

b) CHE

)

l /

g 0.010

(

n o

m 0.005

m A

0.000

1 2 2 2 2 3 3 3 3 4

1 1 1 1 1 1 1

1 1 1

0 0 0 0

0 0 0

2 2 2 2

2 2 2

l l

r r

t t t

n n n

u u

c c c

p p

a a a

J J

A A

J J J

O O O

0.015

c) WH2

)

l /

g 0.010

(

n o

m 0.005

m A

0.000

1 2 2 2 2 3 3 3 3 4

1 1 1 1 1 1 1

1 1 1

0 0 0 0

0 0 0

2 2 2 2

2 2 2

l l

r r

t t t

n n n

u u

c c c

p p

a a a

J J

A A

J J J

O O O

Figure 3.9: Ammonia concentrations (mg/l) measured during monthly industry surveys 2m above the bottom at sites a) KR1; b) CHE; and d) WH2.

0.15

a) KR1

) l

/ 0.10

(

t i

N 0.05

0.00

1 2 2 2 2 3 3 3 3 4

1 1 1 1 1 1 1

1 1 1

0 0 0 0

0 0 0

2 2 2 2

2 2 2

l l

r r

t t t

n n n

u u

c c c

p p

a a a

J J

A A

J J J

O O O 0.16 b) CHE

0.14

)

l /

g 0.12

(

a r

t 0.10

i N

0.08

0.06

1 2 2 2 2 3 3 3 3 4

1 1 1 1 1 1 1

1 1 1

0 0 0 0

0 0 0

2 2 2 2

2 2 2

l l

r r

t t t

n n n

u u

c c c

p p

a a a

J J

A A

J J J

O O O

0.15

c) WH2

) l

/ 0.10

(

t i

N 0.05

0.00

1 2 2 2 2 3 3 3 3 4

1 1 1 1 1 1 1

1 1 1

0 0 0 0

0 0 0

2 2 2 2

2 2 2

l l

r r

t t t

n n n

u u

c c c

p p

a a a

J J

A A

J J J

O O O

Figure 3.10: Nitrate concentrations (mg/l) measured during monthly industry surveys 2m above the bottom at sites a) KR1; b) CHE; and d) WH2.

24 CONFIDENTIAL: VERSION 3 (FINAL)

Total oxygen demand As noted previously, in many systems a significant fraction of the BOD can occur in the water column (pelagic BOD). In many estuaries, pelagic BOD is 2–4 times benthic BOD (Hopkinson and Smith (2004) and in a shallow silled estuary in Scotland, Gillibrand et al. (2006) found that pelagic BOD exceeded benthic BOD. However, pelagic rates of consumption have not been measured in Macquarie Harbour.

In the absence of this information, we can estimate total BOD (pelagic and benthic combined) based on organic carbon loads and measured values for likely oxidation ratios. Here, we have total BOD based on the ratio of carbon mineralised to oxygen consumed in benthic flux measurements, which equates to 0.75 gO2/gC and 2.15 gO2/gC for cage and buffer sediments respectively. Table 3.3 shows the change in estimated oxygen demand as the industry has grown in Macquarie Harbour. The BOD estimated for 2013 (3.55 tonnes O2/day) is higher than that predicted from benthic fluxes alone for the same year (1.21 tonnes O2/day).

The difference in the two BOD values may represent the contribution of pelagic BOD. It could equally be due to the implicit errors associated with scaling flux measurements to a large area (e.g. inadequate spatial sampling) or omission of significant processes (e.g. burial of feed or faeces before being broken down).

Table 3.3: Estimated farm related BOD (tonnes per day) in Macquarie Harbour for 2002-2013.

Year 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 BOD 0.83 0.81 0.95 0.54 0.71 0.89 1.09 1.62 2.42 3.05 3.29 3.55

Ultimately, though, a similar calculation for non-farm carbon inputs is required to put this in a proper context. However, for catchment organic carbon, there are too many unknowns (e.g. fraction exported, fraction labile) leading to significant uncertainty in estimating the associated BOD when considering the size of the input. Nonetheless, the mixing plot of total organic carbon and salinity from recent EPA measurements in the harbour does provide some useful insight and direction for future work (Figure 3.11). For samples collected above 15 m depth, the non-linear relationship between TOC and salinity is consistent with pelagic mineralisation (BOD) in surface waters. This suggests that at least some of the catchment organic carbon is broken down in the harbour before it is either exported or sinks to bottom waters. Unfortunately, the number of samples collected from below 15m precludes any meaningful conclusions about pelagic respiration in bottom waters. Given the lack of variation in bottom water salinity, direct measurements of pelagic oxygen demand will be more informative in determining the role of pelagic BOD.

25 CONFIDENTIAL: VERSION 3 (FINAL)

TOC >15m

)

l /

g TOC <15m

(

n 10

a g

r 5

o t

0 0 5 10 15 20 25 30 35 Salinity (ppt)

Figure 3.11: Mixing plot of salinity (ppt) and total organic carbon concentration (mg/l) for samples taken above and below 15m depth highlighted in blue and red respectively.

26 CONFIDENTIAL: VERSION 3 (FINAL)

4. Replenishment of dissolved oxygen by physical processes

Long-term trends Recent changes in DO levels in Macquarie Harbour need to be viewed in the context of historical patterns in environmental conditions. For example, there is evidence to suggest that average ocean temperatures off western Tasmania have increased by around 1.3°C over the past 100 years (Appendix E), corresponding to an average reduction in DO of approximately 0.36 mg/l.

Since 1978 the operation of the power station on the Gordon River has had a significant influence on river discharge into the harbour (Appendix B Figure B.2). Over most of this period pattern of extremes in flow (both high and low) have been modified, potentially resulting in fewer mixing and recharge events in the harbour. However, the flow regime changed again when Basslink came online in 2006 (to optimise returns from power generation). This has involved both extended periods of low discharges and very high discharges (Appendix B Figure B.3 & B.4)

Quarterly EPA sampling indicates an overall decline in DO below 15 m since 2009 over the length of the harbour (Figure 4.1). Surface waters (0 – 5 m) remain near saturation levels, while those in the halocline (5 – 15 m) show the highest variability. Salinity is highly variable in the upper 15 m reflecting patterns in river discharge (discussed below). However, at deeper levels there are no long-term changes in salinity to match the decline in DO (Figure 4.2).

The DO decline observed in the quarterly EPA data is also evident in monthly industry data taken at other deep-water sites around EPA sites 12 and 27 (right hand column of Figure 4.1). The data sets are remarkably consistent given differences in sampling times and locations. By mid 2014 all datasets represented in Figures 4.1 and 4.2 show minimum DO levels below 15% saturation and some much lower.

High river discharge mixing events Strong mixing conditions (Figure 1.4a) only appear to extend into the deep- waters of Macquarie Harbour following periods of high river discharge. The clearest example captured by the monthly sampling occurred in the August 2013 flood events when river discharges into the harbour were some of the highest on record (monthly average exceeding 900 m3s-1) and highly variable with several peaks well over 1000 cumecs (Figures 4.3 and 4.4a). Vertical mixing at this time may have been further enhanced by winds from the northwest (Figure 4.4c), which opposed inflow from the Gordon River and tended to maximize the shear in surface waters.

27 CONFIDENTIAL: VERSION 3 (FINAL)

a) 0-5 m b) 5-15 m 120 120 a) EPA Site 12 b) EPA Site 12 EPA CHE KR1 100 a) 0-5 m 100 b) 5-15 m - - – 120 120

80 ) 120 80 35 n 100 100

60 o 60

i t

a 30 4080 4080

r 100 u

2060 t 2060

a 25

Dissolved Oxygen (% saturation) (% Oxygen Dissolved saturation) (% Oxygen Dissolved

)

040 80 400 t p

19931994199519961997199819992000200120022003% 200420052006200720082009201020112012201320142015 19931994199519961997199819992000200120022003200420052006200720082009201020112012201320142015 p

20 ( 20

Date ( 20 Date

Dissolved Oxygen (% saturation) (% Oxygen Dissolved saturation) (% Oxygen Dissolved

0 0 y

t e

60 i g

19931994199519961997199819992000200120022003200420052006200720082009201020112012201320142015 1993n 1994199519961997199819992000200120022003200420052006200720082009201020112012201320142015

c) 15-25 m i d) 25-35 m

y 5-15 m l 15

x Date Date

120 120 a

O S

40 100 100 c) 15-25d m 10 d) 25-35 m 80 e 80

120 v 120 l

o 20 60 60 5

100 s 100 s

40 i 40

80 80 D 20 0 20 0 60 60

3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 Dissolved Oxygen (% saturation) (% Oxygen Dissolved Dissolved Oxygen (% saturation) (% Oxygen Dissolved 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 40 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 40 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 19931994199519961997199819992000200120022003200420052006200720082009201020112012201320142015 19931994199519961997199819992000200120022003200420052006200720082009201020112012252013-320145 m2015

20 Date Date 20 DateDate Dissolved Oxygen (% saturation) (% Oxygen Dissolved Dissolved Oxygen (% saturation) (% Oxygen Dissolved 0 0 a) 0-5 m CHE EPA 12 KR1 b) 5-15 m 19931994199519961997199819992000200120022003200420052006200720082009201020112012201320142015 19931994199519961997199819992000200120022003200420052006200720082009201020112012201320142015 120 120 Date c) EPA Site 27 Date 100 100 d) EPA Site 27 a) 0-5 m b) 5-15 m -EPA -CH1 -WH2 CHE EPA 12 KR1

12080 ) 120 12080 35 n

60 o 60

100 i 100 t

40 a 40 30

80 r 100 80 u 20 t 20

60 a 60

s 25

Dissolved Oxygen (% saturation) (% Oxygen Dissolved saturation) (% Oxygen Dissolved

0 80 0 )

40 40t %

19931994199519961997199819992000200120022003200420052006200720082009201020112012201320142015 1993p 1994199519961997199819992000200120022003200420052006200720082009201020112012201320142015

( p 20 20

Date ( 20 Date

e Dissolved Oxygen (% saturation) (% Oxygen Dissolved Dissolved Oxygen (% saturation) (% Oxygen Dissolved 60

0 0t i

n y

199319941995199619971998199920002001c)2002 15-25200320042005 m2006200720082009201020112012201320142015 1993i 19941995199619971998199920002001d)2002 25-35200320042005 m2006200720082009201020112012201320142015

l 15 15-25 m x

120 Date 120 a Date O

40 S

100 d 100 10

c)e 15-25 m d) 25-35 m v

80 l 80 120 120

o 20

60 s 60 5

100 s 100 i

40 D 40 80 0 80 0 20 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 20 60 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 60 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5

9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 Dissolved Oxygen (% saturation) (% Oxygen Dissolved Dissolved Oxygen (% saturation) (% Oxygen Dissolved 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 40 40 19931994199519961997199819992000200120022003200420052006200720082009201020112012Date201320142015 19931994199519961997199819992000200120022003200420052006200720082009201020112012252013-320145 m2015

20 Date 20 Date Date Dissolved Oxygen (% saturation) (% Oxygen Dissolved Dissolved Oxygen (% saturation) (% Oxygen Dissolved 0 0 CH1 EPA 27 WH2 19931994199519961997199819992000200120022003200420052006200720082009201020112012201320142015 19931994199519961997199819992000200120022003200420052006200720082009201020112012201320142015 Date e) EPA Site 34 f) EPA Site Date34

) 120 35

CH1 EPA 27 WH2 n

i t

a 30

r 100

u t

a 25

)

80 t p

( p

( 20

y t

e 60

i y

l 15

O S

d 10

v l

o 20 5

s i

Figure 4.1: Time series plots of median dissolved oxygen (% saturation) and salinity (ppt) in depth bins (0-5, 5-15, 15-25, 25-35, 35–45 m) at EPA sites 12 (a,b), 27(c,d) and 34(e,f) from 1993 to present. Comparisons of the EPA DO data from sites 12 and 27 with that from nearby industry monitoring sites are shown in the right hand column for two of the depth bins (15-25, 25-35 m) over the period 2011-2014. Site locations are shown in Figure 2.1.

28 CONFIDENTIAL: VERSION 3 (FINAL)

salinity 25-35 m dissolved oxygen 25-35 m salinity 35-45 m dissolved oxygen 35-45 m 100 35

30 Salinity (ppt) 60

40 25

20 Dissolved Oxygen (% saturation) (% Oxygen Dissolved

0 20

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Date Figure 4.2: Time series plot of median dissolved oxygen (% saturation) and salinity (ppt) in the deep-water depth bins (25–35, 35–45 m) at EPA site 12.

The DO response to the mixing event extends over most of the water column (Figures 4.5, 4.6 and 4.7) and therefore represents a significant change in the total DO content of the harbour. However, the change is only persistent in the upper 30 m (compare June and August in Figure 4.8). Below this the total oxygen content of the deeper layers remains relatively small, so that reversals evident in individual profiles (Figures 4.6 and 4.7) are short lived.

After the August 2013 mixing event there was a significant drop in DO to a depth of around 30 m (Figure 4.8) that cannot be explained by benthic BOD. This decline may be associated with deeper water being displaced upward by recharge events, as reflected by an increase in salinity at depth (Figures 4.5e and 4.9). Pelagic consumption (i.e. nitrification or pelagic mineralization) may also be playing a role in this decline, pointing to the need to examine these processes more carefully in Macquarie Harbour.

29 CONFIDENTIAL: VERSION 3 (FINAL)

(a) 1000

800

600

400

200

total river flow (30 day mean) running river flow total 0

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Date (b)

(c)

Figure 4.3: River discharge conditions: (a) 30-day running mean of total river discharge into Macquarie Harbour with blue dots identifying survey dates and the circled areas corresponding to the recharge events identified in figure 4.2; (b) annually averaged discharge from the Gordon River, King River and other catchments into Macquarie Harbour; and (c) comparison of the median monthly discharge from the Gordon Power Station during 1997-2013, 2006-2012 (post Basslink) and 2012-2013 (Source: Hydro Tasmania 2013b). All discharge rates are given in units of m3s-1.

30 CONFIDENTIAL: VERSION 3 (FINAL)

(a) Total river discharge (cumecs) 1000

1000

)

) c

n 800

)

m c

n 800

( g m

600

( g

n 600

e 400

l e

400

o 200

(

t 3

o 200 ( T 0 1 1 1 2 2 2 2 3 3 3 3 4 4 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 2 2 2 2 2 2 2 2 2 2 2 n g v b y g v b y g v b y u 1 u 1 o 1 e 2 a 2 u 2 o 2 e 3 a 3 u 3 o 3 e 4 a 4 J 1 A 1 N 1 F 1 M 1 A 1 N 1 F 1 M 1 A 1 N 1 F 1 M 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 2 8 2 6 2 6 2 4 2 2 2 0 2 1 2 9 2 7 2 5 2 6 2 n 3 g 2 v 2 b 2 y 2 g 2 v 2 b 2 y 1 g 1 v 1 b 1 y u o a o a o a J u e u e u e A N F M A N F M A N F M 1 0 8 6 6 4 2 0 1 9 7 5 6 3 2 2 2 2 2 Date2 2 1 1 1 1 (b) Low-0.24frequency sealevel at Strahan Wharf (m)

n Date o

i 0.24

)

o g

i 0.22

)

v g

0.22

t E

e 0.20

n 0.20

u W

r 0.18

0 W

a 0.16

(

a t

0 0.16

(

a r

t 0.14

S 1 1 1 2 2 2 2 3 3 3 3 4 4 1 1 1 1 1 1 1 1 1 1 1 1 1 0.14 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 2 2 2 2 2 2 2 2 2 2 2 n g v b y g v b y g v b y u 1 u 1 o 1 e 2 a 2 u 2 o 2 e 3 a 3 u 3 o 3 e 4 a 4 J 1 A 1 N 1 F 1 M 1 A 1 N 1 F 1 M 1 A 1 N 1 F 1 M 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 2 8 2 6 2 6 2 4 2 2 2 0 2 1 2 9 2 7 2 5 2 6 2 n 3 g 2 v 2 b 2 y 2 g 2 v 2 b 2 y 1 g 1 v 1 b 1 y u o a o a o a J u e u e u e A N F M A N F M A N F M 1 0 8 6 6 4 2 0 1 9 7 5 6 3 2 2 2 2 2 Date2 2 1 1 1 1 Date (c) Wind speed (m/s) and direction Wind Speed/Direction 100

Windspeed (m/s) -20

-40

-60

-80

-100 07/01/11 10/01/11 01/01/12 04/01/12 07/01/12 10/01/12 01/01/13 04/01/13 07/01/13 10/01/13 01/01/14 04/01/14

Figure 4.4: (a) Total river inflow (30 day running mean in ML); (b) Strahan wharf sealevel (30 day running mean) reflecting variations in atmospheric pressure (Appendix C); and (c) vector plots of wind speed and direction (7 day running mean). The circled areas correspond to the recharge events (red) and river discharge event (blue) marked in Figure 4.5.

31 CONFIDENTIAL: VERSION 3 (FINAL) 12 11 a) KR1 Dissolved Oxygen (mg/l) 10

0 9 (a) DO at KR1 8 7 -10 6 5 4 -20 3.512 b) (m) Depth CHE Dissolved Oxygen (mg/l) 311 2.510 -30 0 29 (b) DO at CHE 1.58 17 -10 0.56 0 5 -20 4

a) (m) Depth CH1 Dissolved Oxygen (mg/l) 3 12 2.5 -30 11 2 0 10 (c) DO at CH1 1.5 9 1 8 -10 0.5 7 0 6 -20 5 4

Depth (m) Depth 3 -30 12 2.5 11 c) WH2 Dissolved Oxygen (mg/l) 2 10 -40 1.5 0 9 1 (d) DO at WH2 8 0.5 7 -10 0 6 5 -20 4

b) (m) Depth CH1 Salinity (ppt) 3 36 2.534 -30 0 232 (e) Salinity at 1.530 28 CH1 1 26 -10 0.5 24 022 -20 20 18 16

Depth (m) Depth -30 14 12 10 -40 8 6 4 2 0

Figure 4.5: Contour plots of dissolved oxygen (contour interval = 0.5 mg/l) at sites a) KR1, b) CHE, c) CH1 (or near), and d) WH2, and (e) salinity at site CH1 (or near). The 3 mg/l contour is highlighted in green and the timing of recharge events (red) and a large river discharge event (blue) are indicated along the date axis in (e). Note CH1 finished in September 2013 and the data in the following months for the contour plot was from nearby sites CH5 (Oct and Nov 2013) and C8 (December 2013 to present)

32 CONFIDENTIAL: VERSION 3 (FINAL)

Figure 4.6: Dissolved oxygen (% saturation) at site WH2 from October 2011 to August 2014.

a) KR1 b) CHE c) CH1 d) WH2

Oxygen (mg/l) Oxygen (mg/l) Oxygen (mg/l) Oxygen (mg/l) 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 0 0 0 0

5 5 5 5

10 10 10 10

15 15 15 15

) ) ) )

m m m

( ( (

( 20 20 20 20

h h h

t t t t

p p p p

e e e

D D D D 25 25 25 25

30 30 30 30

35 35 35 35

40 40 40 40

45 45 45 45

Oct 2011 Nov 2011 Dec 2011 Jan 2012 Feb 2012 Mar 2012

Salinity (ppt) Salinity (ppt) Salinity (ppt) Salinity (ppt) 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 0 0 0

5 5 5 5

10 10 10 10

15 15 15 15

) ) ) )

m m m m

( ( (

( 20 20 20 20

h h h h

t t t t

p p p p

e e e e

D D D D 25 25 25 25

30 30 30 30

35 35 35 35

40 40 40 40

45 45 45 45 Figure 4.7: Depth profiles of dissolved oxygen (mg/l) and salinity (ppt) measured during monthly surveys before (Oct, Nov and Dec 2011) and during (Jan, Feb and Mar 2012) a recharge event in summer 2012 at sites a) KR1; b) CHE; c) CH1; and d) WH2.

33 CONFIDENTIAL: VERSION 3 (FINAL)

(a) Total DO in 5 m depth bins (b) Change in DO in 5 m bins (tonnes) (tonnes)

April 2013 April to June 2013

0 5,000 10,000 15,000 -3500 0 3500 0-5m 5-10m 10-15m

15-20m 20-25m 25-30m 30-35m Depth Depth (m) 35-40m 40-45m 45-50m

June 2013 June to August 2013 0 5,000 10,000 15,000 -3500 0 3500 0-5m 5-10m 10-15m

15-20m 20-25m 25-30m 30-35m Depth Depth (m) 35-40m 40-45m 45-50m

August 2013 August to September 2013 0 5,000 10,000 15,000 -3500 0 3500 0-5m 5-10m 10-15m

15-20m 20-25m 25-30m 30-35m Depth Depth (m) 35-40m 40-45m 45-50m

Figure 4.8: (a) Total oxygen content (tonnes) within 5 m depth bins during April, June and August 2013; and (b) change in total oxygen content over the corresponding two month intervals to August and one month interval August to September. The depth-integrated totals for April, June, August and September were 26.4 kT, 24.2 kT, 34.2 kT and 28.6 kT respectively.

34 CONFIDENTIAL: VERSION 3 (FINAL)

a) KR1 b) CHE c) CH1 d) WH2

Oxygen (mg/l) Oxygen (mg/l) Oxygen (mg/l) Oxygen (mg/l) 0 5 10 15 0 5 10 15 0 5 10 15 0 5 10 15 0 0 0 0

5 5 5 5

10 10 10 10

15 15 15 15

) ) ) )

m m m m

( ( (

( 20 20 20 20

h h h h

t t t t

p p p p

e e e e

D D D D 25 25 25 25

30 30 30 30

35 35 35 35

40 40 40 40

45 45 45 45

Apr 2013 May 2013 Jun 2013 Jul 2013 Aug 2013 Sep 2013

Salinity (ppt) Salinity (ppt) Salinity (ppt) Salinity (ppt) 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 0 0 0

5 5 5 5

10 10 10 10

15 15 15 15

) ) ) )

m m m m

( ( (

( 20 20 20 20

h h h h

t t t t

p p p p

e e e e

D D D D 25 25 25 25

30 30 30 30

35 35 35 35

40 40 40 40

45 45 45 45 Figure 4.9: Depth profiles of dissolved oxygen (mg/l) and salinity (ppt) measured during monthly surveys in mid-2013 (Apr-Sep 2013) at sites a) KR1; b) CHE; c) CH1; and d) WH2. They show deep- water recharge in May-June and top down vertical mixing in July-August.

Deep water recharge events – low frequency variability Superimposed on the long-term decline in DO were also frequent deep-water recharge events (Figure 1.4c). Examples of relatively recent events are indicated by red ellipses in Figure 4.2. Their signature below 25 m is an increase in both oxygen and salinity as new oceanic water arrives at the monitoring site. In each case, there river discharge is relatively low (indicated by red ellipses in Figure 4.3a).

Plots of DO profiles through time show four significant events since October 2011, by far the largest being in early 2012 (Figures 4.5 and 4.6). The average concentration of dissolved oxygen below 15 m depth increased from 2.6 to 3.8 mg/l in the early 2012 event. In the first two cases, the recharge events coincide with low river discharge (Figure 4.4a) and peaks in the (sub-tidal) sea-level signal at Strahan when average water depth over the sill is at its highest (Figure 4.4b, Appendix C), consistent with the general hypothesis in Figure 1.4c. In May- June 2013, the oceanic recharge follows a fall in river discharge but average flows remain elevated compared to the previous two events. In May-June 2013, the winds were strong and directed (NNW) into the harbour across the sill; this may be favoring the influx of oceanic water through direct wind forcing but also by pushing the freshwater layer back from the sill.

Individual profiles from the summer 2012 illustrate that this large recharge event was accompanied by an increase in the depth and salinity of the mixed layer, as well as an increase in deep-water salinity (Figure 4.9). This suggests

35 CONFIDENTIAL: VERSION 3 (FINAL)

that there was significant mixing over the sill before the plume descended and mixed with deeper waters.

Deep-water recharge events – higher frequency characteristics The analysis of recharge events has so far focused on their appearance in monthly sampling and the relationship of these events to low-frequency forcing (months to years). Recent deployments of continuous data loggers (Figure 4.10) have also allowed us to conduct preliminary analyses of their evolution over shorter timescales (hours to days). When combined with recent monthly profile data from May-June 2014, the continuous data provides additional insights into the deep-water recharge process.

Figure 4.10: Locations of DO, Temperature and conductivity loggers 23, 31 and 46 deployed in June 2014. Logger 23 had probes at depths of 20, 30 and 40 m; logger 31 at 20, 30 and 35 m; and logger 46 at 20 and 30 m.

36 CONFIDENTIAL: VERSION 3 (FINAL)

There were high flows in the Gordon River during the second half of May 2014 (up to 900 cumecs) and late June 2104 (700 cumecs) (Figure 4.13). However, between these events flows fell to around 300 cumecs and monthly profiles reveal recharge between May and June 2014 below 30 m at KR1, C10 and CHE, and below 35 m further south at C8 (Figure 4.11). This is further corroborated by profiles collected independently by the EPA (last plot in Figure 4.11) and the total DO content of these levels (Table 4.1). There is no evidence of recharge at the shallower most southern station WH2, which maintains very low DO levels below 20 m (Figure 4.11). This is consistent with recharge propagating from the sill end of the harbour. The DO profiles were also quite stable at mid-depths, providing further evidence that these changes are not related to mixing with overlying water.

Table 4.1: Estimated total DO in Macquarie Harbour for May and June 2014. The recharge event is highlighted in red.

Depth Tonnes of dissolved oxygen

June 2014 May 2014

0-5m 11765 12178 5-10m 5060 6691 10-15m 2465 2222

15-20m 942 866

20-25m 312 386 25-30m 160 258 30-35m 134 138 35-40m 198 81 40-45m 165 50 45-50m 21 TOTAL TONS 21227 22870

On the 19th June the continuous data loggers showed a significant increase in DO and decrease in temperature at 40m close to the sill (station 23), that were detected 2 hours later at 30 m, and then at 20 m after a further 8 hours (Figure 4.12). Two subsequent events were also detected at 40 m with approximately 48 hours between events. This timing suggests coupling with the tides and may be evidence that individual recharge events are initiated by favourable (flood) tide conditions. Alternatively, the signal may be associated with the return of water from the previous recharge with the tide. There is little evidence of multiple recharge events at shallower depths, although enhanced variability at 20 m (station 23) may indicate significant mixing and detrainment as the recharge plume began its descent along the sill.

Further from the sill at the deepest logger from station 31, DO and temperature experienced similar changes to those at station 23, but lagging by approximately 80 hours (cf. 4.12a,b or 4.12c,d). Assuming a direct pathway between the two logging stations 4.2 km apart gives a minimum propagation speed for the

37 CONFIDENTIAL: VERSION 3 (FINAL)

recharge plume of 1.5 cm/s. This is broadly consistent with ADCP measurements at 40 m (Cardinal Marker 5) of between 3 and 5 cm/s (including a tidal component that will tend to cancel out over the 80 hour propagation time). There was no evidence of the recharge event reaching the shallower depths of station 31 (20 and 30 m) or station 46 (20 and 30 m) a further 8 km southeast of station 31 (Figure 4.10).

Figure 4.11: Depth profiles of dissolved oxygen (mg/l) measured during May 2014 (yellow) and June 2014 (blue) at five stations. The last plot compares the June profiles from two of the stations with a nearby profile (MH12) collected 18 days later by the EPA.

38 CONFIDENTIAL: VERSION 3 (FINAL)

(a) DO at station 23 (mg/l)

(b) DO at station 31 (mg/l)

(d) Temperature at station 31 (°C)

Figure 4.12: Three hour running mean of DO at logger station 23 (depths 20, 30 and 40 m) and logger station 31 (depths 20, 30 and 35 m) and corresponding temperature time-series (all in hours since their deployment over the period 4-5 June 2014).

39 CONFIDENTIAL: VERSION 3 (FINAL)

Figure 4.13 Daily river flow (cumecs) at Gordon Power station (black line) and Gordon River above Franklin (red line) from January to August 2014.

4. Discussion

The process of identifying the underlying drivers for the observed decline in deep-water DO levels over the period 2009-present has been confounded by the coincidence of salmon aquaculture expansion with historical extremes in river discharge (driven by drought, the Hydro rebuild strategy, and then large releases aimed at capitalising on the fixed carbon price period).

There are significant gaps in our understanding of the oxygen dynamics in Macquarie Harbour, most notably the fraction of labile organic carbon entering via river discharge and the export of labile organic carbon into the ocean. In other estuarine and fjord systems a large proportion of the BOD has been found to be pelagic (Hopkinson and Smith 2004, Gillibrand et al. 2006), but this quantity has never been measured in Macquarie Harbour.

While our analysis suggests that aquaculture may be responsible for 3-12% of the benthic BOD (below 15 m), the implications for DO levels throughout the harbour are less clear. For example, if the benthic BOD associated with organic carbon inputs from the catchment is mainly via slow conversion of a layer of benthic refractory material to a more labile form (e.g. Middleburg and Levin 2009), then breakdown rates will be largely limited by the surface area of the bottom (i.e. highly buffered) and the natural levels of BOD may be relatively stable. Any labile organic material entering this system will still break down quickly and add to the background BOD. This is just one of a number of scenarios which need to be tested with more targeted field measurements.

River flow also plays an important role in replenishing deep-water oxygen. However, other factors such as wind, tidal height and atmospheric pressure also play significant roles in regulating oxygen replenishment. In regards to river

40 CONFIDENTIAL: VERSION 3 (FINAL)

flow the most effective flow regime appears to involve high variability that supports both frequent vertical mixing events (high river discharge) and frequent recharge events (low river discharge), which eventually coalesce in the water column through background vertical diffusion of DO. River discharge over the period of DO decline has been characterized by relatively low variability and less frequent switching between the two renewal processes. However, the net effect of these changes on DO and the relative importance of the other physical drivers of these replenishment process can only be quantified with more frequent collection of dissolved oxygen data (i.e. continuous data loggers) and within a more comprehensive modelling framework that integrates both the physical and biological processes.

Further data will clearly be required before the decline in DO can be definitively attributed. It is recommended that existing monitoring should be continued and supplemented by a strategic field campaign aimed at measuring (i) the labile and refractory organic carbon loads in river discharge and net exports to the ocean; and (ii) pelagic BOD in the harbour. Similarly, the role of fish farm organic carbon inputs on BOD should be extended to provide greater resolution of the spatial extent of deposition and the influence on benthic and pelagic BOD. These measurements will support development of a detailed quantitative carbon/oxygen budget for Macquarie Harbour. Further development and calibration of a biogeochemical model for Macquarie Harbour would not only assist with the understanding the spatial and temporal aspects of carbon/oxygen flows and transformations through the system, but would also provide a mechanism for exploring past conditions and future scenarios relating to changes in river flows and/or salmon production.

Environmental and industry management responses to the current situation will depend on how the Macquarie Harbour system responds to future conditions. While there has been some recent recovery in DO due to extreme weather conditions in early August 2014, we don’t know how long these conditions will persist. Factors such as river flows (without the influence of carbon pricing) and other system wide changes (particularly if deep-water conditions move from hypoxic to regular anoxic events) may be significant over the longer term. The future conditions cannot be accurately predicted at this stage and it is conceivable that the harbor will move to a new equilibrium based on events over the last 5 years.

41 CONFIDENTIAL: VERSION 3 (FINAL)

References

Allen, GL and Simpson, JH (1998) Deep water inflows to Upper Loch Linnhe. Estuarine Coastal and Shelf Science 47:487–498. Aure, J and Stigebrandt, A (1989) On the influence of topographic factors upon the oxygen consumption rate in sill basins of fjords. Estuarine Coastal and Shelf Science 28:59-69. Carpenter, PD, Butler, ECV, Higgins, HW, Mackay, DJ and Nichols, PD (1991) Chemistry of trace elements, humic substances and sedimentary organic matter in Macquarie Harbour, Tasmania. Australian Journal of Marine and Freshwater Research 42:625-54. Cornwell, J. C., W. M. Kemp, and T. M. Kana. 1999. Denitrification in coastal ecosystems: methods, environmental controls, and ecosystems level controls, a review. Aquatic Ecology 33:41-54. Cresswell, GR, Edwards, RJ and Barker, BA (1989) Maquarie Harbour, Tasmania – seasonal oceanographic surveys in 1985. Papers and Proceedings of the Royal Society of Tasmania 123, 63-66. EIS (2011) Environmental Impact Statement to Accompany the Draft Ammendment No. 1 to the Macquarie Harbour Marine Farming Development Plan October 2005. Gillibrand, PA, Turrell, WR and Elliott, AJ (1995) Deep-water renewal in the upper basin of Loch Sunart, a Scottish Fjord. Journal of Physical Oceanography 25:1488–1503. Hargrave, B. T., D. E. Dupliseas, E. Pfeiffer, and D. T. Wildish. 1993. Seasonal Changes in benthic fluxes of dissolved oxygen and ammonium associated with cultured Atlantic salmon. Marine Ecology Progress Series 96:249-257. Hopkinson, CS and EM Smith (2005) Estuarine respiration: an overview of benthic, pelagic, and whole system respiration. In: Respiration in Aquatic Systems. P. del Giorgio and P. Williams (ed) p122-146 (DOI:10.1093/acprof:oso/9780198527084.001.0001). Hydro Tasmania, (2013a) Basslink Review Report 2006-12: Gordon River Basslink Monitoring Program. Hydro Tasmania, Hobart. Hydro Tasmania, (2013b) Gordon River Monitoring Annual Report 2012–13. Hydro Tasmania, Hobart. Inall, ME and Gillibrand, PA (2010) The physics of mid-Latitude fjords: A review, Geological Society of London Special Publications 344:17-33. Liungman, O, Rydberg, L, Göransson, CG (2001) Modeling and observations of deep water renewal and entrainment in a Swedish sill fjord. Journal of Physical Oceanography 31:3401−3420. Lucieer, V, Lawler M , Pender A and Morffew M 2009, SEAMAP Tasmania – Mapping the Gaps. Tasmanian Aquaculture and Fisheries Institute, University of Tasmania 188pp. Middelburg JJ, Duarte CM, Gatusso JP (2004) Respiration in coastal benthic communities. In: del Giorgio PA, Williams PJIB, editors. Respiration in aquatic ecosystems. Oxford University Press, Oxford. Pp. 206–224 Middleberg, JJ and Levin, LA (2009) Coastal hypoxia and sediment biogeochemistry. Biogeosciences 6:1273-1293. Moran, MA, Sheldon WM Jr, and Sheldon, JE (1999) Biodegradation of riverine

42 CONFIDENTIAL: VERSION 3 (FINAL)

dissolved organic carbon in five estuaries of the southeastern United States. Estuaries 22:55-64. Palmer L., McConachy F. and Peterson J. (2001) Basslink Integrated Impact Assessment Statement, Potential Effects of Changes to Hydro Power Generation, Appendix 2, Gordon River Hydrology Assessment. Prepared for Hydro Tasmania June 2001. Pereira, P. M. F., K. D. Black, D. S. McLusky, and T. D. Nickell. 2004. Recovery of sediments after cessation of marine fish farm production. Aquaculture 235:315-330. Sun, L, Perdue, EM, Meyer, JL and Weis, J (1997) Use of elemental compostion to predict bioavailability of dissolved organic matter in a Georgia river. Limnology and Oceanography 42:714-721. Wetzel, RG, Hatcher, PG and Banchi, TS (1995) Natural photolysis by ultraviolet irradiance of recalcitrant dissolved organic matter to simple substrates for rapid bacterial metabolism. Limnology and Oceanography 40:1369-1380.

43 CONFIDENTIAL: VERSION 3 (FINAL)

Appendix A: Terms of Reference

Macquarie Harbour Dissolved Oxygen (MH DO) Working Group

Purpose / role of the group:

. The broad purpose of the group is to verify the scope of dissolved oxygen (DO) reductions in bottom waters of Macquarie Harbour and to determine attribution.

. The Group was established by the Tasmanian Salmon Growers Association in February 2014.

. The aims of the group are to work cooperatively in the study of DO issues in the Harbour and to guide studies undertaken under the project leadership of Scott Condie of CSIRO (replacing Philip Gillibrand).

. The Group may also provide approved (by the MH DO Working Group) information and advice to other stakeholders on matters relating to dissolved oxygen issues and related processes in Macquarie Harbour.

Membership:

. TSGA - Adam Main (Chair)

. Tassal - Matt Barrenger, Linda Sams

. Huon Aquaculture - David Whyte, Dom O’Brien

. Petuna - Lance Searle (Project Manager)

. CSIRO – Scott Condie (Project Leader)

. Aquadynamic Solutions (ADS) – Neil Hartstein, Alan Kerroux

. IMAS – Catriona MacLeod, Jeff Ross

. DPIPWE – Martin Read

Accountability:

. The Chairperson of the MH DO Working Group is Adam Main (TSGA) or other person as may be nominated by the Chair.

. The Project Leader (Scott Condie) will report to the MH DO Working Group at each meeting or as may be additionally required by the Chair.

44 CONFIDENTIAL: VERSION 3 (FINAL)

. The Project Manager (Lance Searle) will assist the Project Leader and members of the MH DO Working Group in matters of communications, data supply and day-to-day management. The Project Manager will report to the MH DO Working Group.

. Other members of the Working Group will participate and contribute as requested by the Chair, Working Group, Project Leader or the Project Manager.

Review:

. The MH DO Working Group will meet at least monthly (or more frequently as may be determined by the Chair) for project review, until its completion.

. The Working Group may review and amend the TOR as required to promote the best outcomes of the study.

Working methods / ways of working:

. The Project Leader will determine the best methods to be used for the study and the MH DO Working Group will approve this plan before commencement of work and on an ongoing basis.

. The TSGA will issue a contract for the work program to be undertaken by the Project Leader and may determine that contracts for both ADS and IMAS are required depending on the nature and scope of works requested of them by the MH DO Working Group.

. The Scope for the CSIRO Project Leader contract is as follows (but may be varied at the discretion of the MH DO Working Group):

. Outline the factors and processes contributing to DO levels in the harbour (particularly bottom water DO)

. Provide an analysis of factors and demonstrate the process whereby DO depression is occurring

. Provide reliable estimates of the relative contributions of sources that influence DO conditions in the harbour

. Draw valid conclusions on the attribution of the DO depression that will satisfy regulators, public/stakeholder scrutiny and peer review

. Describe the likely future scenarios for DO in the harbour based on projected inputs and influences from key sources

45 CONFIDENTIAL: VERSION 3 (FINAL)

. Provide ongoing expert advice and evidence as may be required in discussions and negotiations with regulators and stakeholders

. The MH DO Working Group may invite new members to join the group at its discretion.

. Secretariat for the group will be provided by the member nominated by the Chair at each meeting.

. Upon completion of the MH DO Project, the Project Leader will be responsible for the compilation of a report on the project detailing the scope of dissolved oxygen reductions in bottom waters of Macquarie Harbour and confirm attribution.

Sharing of information and resources (including confidential materials):

. All materials and information provided to the Group shall be regarded as confidential and must be kept in a secure place or on “secure” electronic systems.

. Information on salmon production data will be aggregated and confidential.

. Any company specific information that may need to be provided to any member as part of the study will remain confidential to that company unless confidentiality is specifically waived by the supplying company.

. Any intellectual property supplied and used in this study shall remain the property of the supplier.

. Any “new” intellectual property created by the study will be the property of The Tasmanian Salmon Growers Association (TSGA).

. The Chair shall be the only communicator of information from the study and any information to be shared outside the MH DO Working Group must be pre-approved for release by the MH DO Working Group.

Term of the MH DO Project:

. The Project has an expected completion date of 30th June 2014 but this can be amended as necessary by the MH DO Working Group.

46 CONFIDENTIAL: VERSION 3 (FINAL)

Appendix B: Catchments and River Discharges

Catchment data River discharge data was supplied to DPIPWE from Hydro Tasmania. A processed dataset was compiled and supplied by DPIPWE to TSGA for use in this study. The data set comprised estimated daily river discharge into Macquarie Harbour for the Gordon River, the King River and other miscellaneous unnamed catchments (Figure B.1). Gap in the raw flow data were filled using a combination of methods including conversion of rainfall to runoff and scaling of data from other stations. The analysis focused on the period from 1990-2013 and therefore did not consider conditions prior to construction of the Gordon River power station in 1978. In general terms, the operation of the power station has increased baseline flows and reduced peak flows (Figure B.2).

Figure B.1: Catchment delineations.

47 CONFIDENTIAL: VERSION 3 (FINAL)

Figure B.2 Gordon River discharge (m3/s) showing changes associated with operation of the power station from 1978 (Source: Palmer et al. 2001).

Following the processing of the Hydro Tasmania datasets, catchments relevant to Macquarie Harbour were delineated and the corresponding areas calculated (Table B.1). The subcatchments were aggregated in order to produce the catchment areas upstream of each station for which raw flow data was provided. This allowed extrapolation of the existing flow for a subset of the catchment area to the entire catchment in order to compare with the results obtained by DPIPWE (Table B.2).

Table B.1: Areas of catchment surrounding Macquarie Harbour (km2).

Upstream catchment area Catchment Area (km2) % of downstream River Station Section since catchment Catchment Total upstream last station unaccounted for Franklin Collingwood River below Alma River 272.41 - 85.06% Franklin Franklin River at Mt Fincham Track 884.20 611.79 51.50% Franklin Franklin downstream 1,823.20 939.00 0.00% Gordon Gordon at Power Station Tailrace 2,029.93 - 64.27% Gordon Gordon above Denison 2,160.34 130.41 61.98% Gordon Gordon above Franklin (without Franklin) - - 16.92% Gordon Gordon above Franklin (including Franklin) 3,205.77 1,045.43 43.58% Gordon Gordon Total at MH (without Franklin) 3,858.79 653.01 0.00% Gordon Gordon Total at MH (including Franklin) 5,681.98 2,476.21 0.00% King King below Queen 652.22 - 18.67% King King Total at MH 801.96 149.74 0.00% Other Ungauged catchments - no river 738.90 - -

Table B.2: Catchment and flow characteristics for the Gordon River.

% missing Average Discharge Corrected Discharge per Catchment downstream - period 2000-2013 Discharge - period catchment area Area (km2) catchment (m3/s) 2000-2013 (m3/s) (m3/s/km2) Franklin at Mt Fincham 884.20 51.50% 43.67 90.04 0.0494 Gordon above Franklin (without Franklin) 3,205.77 16.92% 175.78 211.58 0.0548 Franklin + Gordon (Raw) 4089.98 - 219.44 301.62 0.0537 % of DPIPWE (Raw) 71.98% - 75.38% 103.61% - MH inflow DPIPWE estimate 5,681.98 - 291.11 - -

48 CONFIDENTIAL: VERSION 3 (FINAL)

Over the past 24 years the Gordon River has been responsible for between 76 and 82% of the total discharge into Macquarie Harbour. The highest annual discharges on record occurred 2013 (Figure B.3) due to large releases at the Gordon Dam during the fixed-price carbon period (Figure B.4). The lowest discharges were in 2008, 2010 and 2012, all at least 20% less than average (Figure B.3, Table B.2). Indeed the entire period of 2008-2012 was characterized by low discharge rates (Figure B.3) due to a combination of low rainfall and a strategy to rebuild stored capacity at the Gordon Power Station (Figure B.4).

Figure B.3 Yearly average discharge (m3/s) distribution for the King, Gordon and Other Catchments into Macquarie Harbour 1990-2013.

Figure B.4 Yearly mean discharge at the Gordon Power Station tailrace 1990-2013.

Gordon River catchment Flows in the Gordon River are strongly influenced by the operation of the power station. In the first few years of Basslink (2007-09) there was increased hydro- peaking (Figure B.5), which may have increased sediment erosion. The initial years of Basslink also coincided with a drought and the resultant lack of water for run-of-river power stations meant greater reliance on the major storages such as Lake Gordon; by mid-2008 the Gordon was run to less than 20% full (Figure B.5). Although the water intake for the power station remained well off

49 CONFIDENTIAL: VERSION 3 (FINAL)

the bottom of the lake, the potential for collecting resuspended sediments and organic matter may have increased.

In June 2008 Hydro introduced a rebuild strategy and this will have led to the re- submergence of vegetation on the banks for the first time in close to a decade. Downstream of the dam, vegetated banks will have experienced increased periods of drying during the storage rebuild. In 2013 when discharge rapidly changed from historical low river flow to historical highs, a concomitant spike in organic matter loading is likely given both the rapid exposure of these banks to flow.

Figure B.5: Timeline of major factors affecting Gordon Power Station operation (including storage levels) relative to Basslink monitoring periods (upper). In 2007–2008 increased seepage erosion was related to the rapid draining of saturated banks when there was unrestricted hydro-peaking between two and three turbine operation. Seepage erosion appeared as net deposition as the bank slumped (lower; Source Hydro Tasmania 2013b).

In most years flows in the Gordon River are highest during winter and lowest during summer and occasionally spring (Figure B.6). During the low discharge period (2008-12) the summer flows were particularly low, while winter flows were more variable. The 2013 winter and spring flows were the highest on record.

50 CONFIDENTIAL: VERSION 3 (FINAL)

The influence of each sub-catchment over time is shown in Figure B.7. From 1990 to 2007 the catchment area above the Gordon Dam provided the largest portion of the water entering the harbour via the Gordon River. From 2008-2012 the influence of this catchment decreases markedly before rebounding in 2013.

Figure B.6 Seasonal averages for the Gordon catchment 1990-2013.

Figure B.7 Seasonal averages for the Gordon catchment 1990-2013.

King River catchment The discharge from the King River is only 15-20% of that from the Gordon River and away from its mouth is expected to have considerably less influence on the DO levels in Macquarie Harbour. Only one gauge station has been used by DPIPWE to reconstitute the full discharge time-series for the King River (starting in mid 1991).

51 CONFIDENTIAL: VERSION 3 (FINAL)

Data from 1992-2013 indicates that the annual King River discharge into Macquarie Harbour is relatively consistent, the largest change being a 50% increase from 2008 to 2009 (Figure B.8). The years 2008, 2010 and 2012 previously identified as the lowest annual flows in the Gordon are also relatively low in the King, although the total period 2008-12 does not stand out as such an anomaly.

Flow in the King River is highly seasonal. As in the Gordon River, winter is usually the season with the highest discharge into Macquarie Harbour, although there are notable spring peaks in 2002 and 2013 (Figure B.9). Summer is again the period with the least flow.

Figure B.8 Seasonal averages for the King River catchment 1992-2013.

Figure B.9 Seasonal averages for the King River catchment 1992-2013.

52 CONFIDENTIAL: VERSION 3 (FINAL)

Other catchments The remaining catchments around Macquarie Harbour are completely ungauged (i.e. no gauging stations). Their discharges were therefore estimated from rainfall and catchment area only. These estimates suggest that the total flow from these catchments is more than an order of magnitude lower than that observed from the Gordon River (Figure B.10).

Figure B.10 Yearly average discharge estimates for catchments other than the Gordon and King draining into Macquarie harbor 1990-2013.

53 CONFIDENTIAL: VERSION 3 (FINAL)

Appendix C: Sealevel and tides

Sealevel data from 2002 was obtained from Strahan Wharf where water level information has been collected every 15 mins. While Macquarie Harbour is characterised as “microtidal” (i.e. not more than 1m between highs and lows of the tidal component of sealevel), the full sealevel range at the Strahan Wharf is closer to 1.5m (Figure C.1). On average sealevel is also higher in winter than in summer by nearly 0.4 m (Figure 8b). All of these variations in sealevel beyond the tidal signal can largely be explained by variations in atmospheric pressure (EIS 2011).

Figure C.1 Sealevel at Strahan Wharf (m).

While the open ocean tides in this region are semidiurnal, lags associated with flow through the narrow mouth of the harbour result in a mixed diurnal- semidiurnal tide at Strahan. With significant river discharges into the harbour, the ebb tide typically runs twice as long as the flood tide (EIS 2011).

54 CONFIDENTIAL: VERSION 3 (FINAL)

Appendix D: Wind Characteristics

Wind data has been provided by BOM at two locations in and around Macquarie Harbour. These two stations are Cape Sorell out past the mouth of the harbour and Strahan Airport.

Cape Sorell has been in operation since 1901 and in recent times data has been saved every 3 hours for wind speed, direction, and other weather related parameters such as rainfall. Data has been collected consistently and there are no significant gaps in the data other than from mid 1992 to February 1993. However, on closer examination, wind speed data was not saved every three hour consistently until 2003, while other parameters were, which appears quite odd.

Wind direction and wind speed is averaged from 2003- March 2014 and presented in a wind rose (Figure D.1). Calm conditions, which are defined as wind speeds of less than 2 knots only occur 2.4% of the time. Wind is generally greater than 10 knots (5 m/s), and predominately from the north, north west and south west directions. There are very few periods, of easterly and south easterly winds. Wind speed and direction were also broken down into months and presented as monthly wind roses using data for the last 56 years (Note, pre 2003 wind strength data wasn’t always saved every 3 hours).

(a) (b)

Figure D.1 Wind speed (m/s) and direction at (a) Cape Sorell from Feb-1993-Mar 2014 and (b) Strahan Airport from Feb-1998-Mar 2014 (1 m/s = 2 knots).

Monthly wind data for the last 56 years also highlights key seasonal differences in wind direction and speed over time (Figure D.2). During summer unsurprisingly wind speed is lower, with the dominant wind direction from the South West. During winter and spring wind strength increases and wind generally is from the north, northwest and southwest. Starting from 2011 wind data is also collected at Cape Sorell every 30 minutes. This data set was not presented due to its short time period but should be pointed out its available and would be very useful for any future modelling work.

55 CONFIDENTIAL: VERSION 3 (FINAL)

Data collected at Strahan airport has been collected and analysed from Sept 1976 to March 2014. From 1976 to 1998 the data is inconsistent with a number of large gaps, or periods of time where regular time steps are missed. The data from 1998 to March 2014 is complete with no gaps and just the occasional missing bin. This data is presented in Figure D.3 in the form of a wind rose showing speed and direction. Wind speed is considerably weaker than that recorded at Cape Sorell with dominate wind directions from the north and north-northeast (NNE). Calm conditions of less than 2 knots were recorded only 4.5% of the time. Monthly wind data since 1976 also highlights key seasonal differences in wind direction and speed over time similar to those observed at Cape Sorell. During summer unsurprisingly wind speed is lower, with though the dominate wind direction is from the NNE rather than the SW like at Cape Sorrel. During winter and spring wind strength increases and wind generally is from the north and north west.

Figure D.2: Wind speed and direction at Cape Sorell broken down monthly from Jan 1957-Mar 2014 in m/s (1 m/s = 2 knots).

56 CONFIDENTIAL: VERSION 3 (FINAL)

Figure D.3: Wind speed and direction at Strahan Airport, broken down monthly from Sep 1976-Mar 2014 in m/s (1 m/s = 2 knots).

57 CONFIDENTIAL: VERSION 3 (FINAL)

Appendix E: Water temperature

Temperature is a parameter that influences Dissolved Oxygen saturation. Higher temperatures will result in a reduction in oxygen saturation, and this has potential influence on the volume of oxygen available during recharge events (especially summer time). Therefore, it is considered necessary to study its evolution over time.

Presently long term temperature data from within the Harbour is rare and there are only a few sources available. The EPA data set contains temperature data but sampling only occurs at 4 sampling periods per year and is not always consistent. While the industry monthly monitoring program, again isn’t particularly useful other than giving general seasonal changes as it is one off sampling every 30 days. While bottom water temperature within the harbor is relatively stable at 13-15 degrees (Figure E.1 & E.2), temperature in the upper part of the water column is high variable and depends on the solar radiation, river water temperature and in part offshore SST. Particular solar radiation forcing and its impact on upper body water temperatures within the harbor can change markedly in the space of days or even hours (Petuna, 2009); subsequently in-situ sampling is required to better understand temperature variability.

Figure E.1 An example of highly variable surface temperatures and lower variability in bottom temperatures with depth over several sampling periods in 1995-1996 (Marine Farming Planning Review of Macquarie Harbour 2011).

58 CONFIDENTIAL: VERSION 3 (FINAL)

Figure E.2 Examples of highly variable surface temperatures and lower variability with bottom temperatures with depth over several sampling periods in 2010& 2011 (from Marine farming planning review of Macquarie harbour 2011).

Two sets of in-situ data have been collected over a period spanning more than 5 years from permanently moored data loggers. The first set of data come from a data logger moored at the Tassal's Franklin lease at a depth of 5 meters. The second set of in-situ data comes from Petuna’s lease at Table Head where a data logger has been moored at 5m below mean sea level since 2006. Both sets of logger were programed to collect temperature data every 10-15 minutes. There are some small gaps in each data set but the majority of data is intact.

In addition to the physical data collected in the harbor Sea Surface Temperature (SST) data collected from satellites (and other remote sensing tools), boats, buoys and sea-surface modelling is available to look at very long term temperature trends. An example with coverage across Tasmania is the ERSST program developed by NOAA. Extended Reconstruction Sea Surface Temperature calculations have been made dating back to 1854 with the more recent data being that collected by remote sensing. Another data set of sea surface temperature can be found at the CSIRO website http://www.marine.csiro.au/las/servlets/dataset which can give Sea-surface temperature daily, weekly or in 2 weekly composites. This data is available from 1993 to 2005. Several other data sets are also available with different resolutions covering the last 15-20 years of so i.e. NASA http://thredds.jpl.nasa.gov/las/getUI.do.

59 CONFIDENTIAL: VERSION 3 (FINAL)

The analysis of temperature data can be made complicated by the overlapping cycles affecting it: inter-annual, seasonal, decadal to multi-decadal, etc. As such, it is more convenient to process the raw temperature records into anomalies, or values of departure from the average, sometimes referred to as “climatology”. This way, trends appear clearly when they exist. Anomalies can be computed for different parameters of the temperature (minimum, maximum, average, etc.) and for different frequencies (yearly, seasonal, monthly, weekly, etc.). For the purpose of the SST records, we presented the NOAA ERSST data, Version 3b as this is the longest record dating back to 1854 (Figure E.3 & E.4). This Version doesn’t use any satellite data rather it uses in-situ data collected from research and merchant vessels combined with modelling. As a result, the data is much coarser at 2 degrees of latitude and longitude, but the values are available on a monthly basis for a very long period (since 1854). The anomalies display a much clearer trend thanks to the greater temporal extent, showing an yearly average upwards increase in temperatures since approximately 1930 (Figure E.2 & E.3). Since approximately 1990 there appears to be half a degree increase in SST). This analysis has been achieved using extracted data for the point closest to Macquarie Harbour, which is located at the entrance of the harbor and extends along western Tasmania.

Figure E.3 Yearly Sea Surface Temperature over time.

60 CONFIDENTIAL: VERSION 3 (FINAL)

Figure E.4 Yearly and seasonal Sea Surface Temperature over time.

In-situ temperature data

As stated above only two sets of continuous water temperature data have been taken inside Macquarie Harbour that cover any significant period of time i.e. > 5 years.

The first of these from Petuna’s Table Head Farm site shows a summer increase during the month of February of approximately two degrees above that seen in the previous 4 years and during 2011 (Figure E.5). It should be pointed out that this data set was averaged (by month), and presented as average water temperature data only.

The high temperatures observed are 1-2 degrees above summer temperatures collected for the ten years prior to 2006 and presented in the 2011 EIS. Assuming that similar temperatures are observed across the harbour and most importantly near hells gate, we can expect a reduction in the capacity of the surface waters entering the harbour to hold oxygen. During summer time river levels are generally at their lowest (Figure E.5), and it has been observed that there have been significant periods of bottom water DO recharge over summer (i.e. Summer 2012), that coincides with this reduction of flow.

Higher average surface water temperatures over key summer recharge periods will result in potentially a reduction in DO being supplied down the sill into Macquarie Harbour bottom waters compared to summer periods with lower mean surface temperatures.

61 CONFIDENTIAL: VERSION 3 (FINAL)

Average Sea Surface temperatures at the mouth of Macquarie Harbour have increased by approximately 0.25 degrees over the last 10 years (Figure E.3 and E.4). This will result in a slight decrease in the availability of oxygen for bottom water recharge.

Water temperature data collected from the Tassal's Franklin lease and displaced as a time series indicates maximum temperatures of 19 degrees occurring every year (Figure 1.5). There is no sign of the very high average temperature recorded in February 2011 at Petuna's Table head site.

Figure E.5 Monthly average temperature, at 5m water depth from 2006 to 2012, Table Head. Unfortunately temperature data from 2013, onward is not available.

62 CONFIDENTIAL: VERSION 3 (FINAL)

Appendix F: Dissolved Oxygen QA/QC

An important first step in this study was to establish consistency in methods used through time and to confirm comparable trends in the independent industry monitoring data set. The EPA data period has seen changes in staff but also the instrument technology used to measure oxygen. Initial inspection of the data identified significant variation in the times taken to profile the water column, with profiles times often between 2-5minutes in the 1990’s compared to 10-20minutes in the 2000’s. Variation in profile times can also be seen in the industry data set, ranging from 8-40 minutes, but the majority is greater than 20 minutes. In April 2010 the EPA changed from a membrane based oxygen sensor to an optical sensor. These sensors are known to have different response times; optical sensors typically have a slower response time but they require less maintenance and are unaffected by gases such as hydrogen supplied.

In a highly stratified system, changes in sensor technology and profile times have the potential to confound the interpretation of long term trends. Subsequently the effect of profile time was tested in the harbour comparing 2, 6 and 35 min profiles. There was little discernable difference between the 6 minute profiles and 35 minutes profiles; the latter is the stand profile rate (i.e. ~ 70 sec/m) used in the industry monitoring (Figure F.2). The 2-minute profile was more variable but appears to represent changes in oxygen concentration with depth reasonably well. Thus, despite the short profile times used in the 1990’s the overall trend appears reliable.

We also investigated the potential consequences of the change from membrane to optical sensors and testing by YSI highlights the quicker response time of membrane probes (YSI Dissolved Oxygen Handbook). As such we wouldn’t expect to see any systematic bias in the sense that the membrane sensor wasn’t capturing the true low levels of bottom water oxygen in the 1990’s.

Figure F.2 Comparison of profile times at site WH2