Water quality assessment for sustainable agriculture in the Wet Tropics—A community-assisted approach

Marine Pollution Bulletin 51 (2005) 99–112 www.elsevier.com/locate/marpolbul

Water quality assessment for sustainable agriculture in the Wet Tropics—A community-assisted approach John Faithful a

a,*

, Wendy Finlayson

b

Australian Centre for Tropical Freshwater Research, James Cook University, Townsville Campus, Qld. 4811, Australia b Cardwell Shire Catchment Management Association, P.O. Box 929, Tully, Qld. 4854, Australia

Abstract A number of studies in north Queensland over the past two decades have concluded that large amounts of nutrients and sediments are exported from agricultural watersheds, particularly during wet season rainfall events. With the co-operation of a number of growers, runoff from Queensland Wet Tropics banana and cane farm paddocks in two distinct tropical river catchments was examined to provide an estimate of nutrient and sediment concentrations and export, with comparison to water quality of flow through a small urban lakes system. Median total nitrogen concentrations in cane drainage runoff (3110 lg N/L) were higher than for banana paddock drainage (2580 lg N/L), although the maximum concentration was recorded from a banana paddock (20,900 lg N/L). Nitrogen losses during post-event drainage flow were supplemented by high proportions of NOX (nitrate + nitrite) sourced from groundwater inputs. Banana paddocks had the highest maximum and median total phosphorus and TSS concentrations (5120 and 286 lg P/L, and 7250 and 75 mg/L respectively) compared to the cane farms (1430 and 50 lg P/L, and 1840 and 14 mg/L respectively). The higher phosphorus and TSS concentrations in the banana runoff were attributed to higher paddock slopes and a greater proportion of exposed ground surface during the wet season. Highest nutrient and TSS concentrations corresponded with samples collected near the peak discharge periods; however, the rising stage of the drainage flows, where the highest nutrient and TSS concentrations are often reported, were difficult to target because of the manual sampling strategy used. This study shows that high concentrations of nutrients and TSS occur in the runoff from cane and banana paddocks. Median total nitrogen, total phosphorus and TSS concentrations in flow through the urban lakes were 369 lg N/L, 16 lg P/L and 11 mg/L, respectively. Flux estimates of 9.2 kg N, 0.8 kg P and 126 kg TSS/ha were determined for drainage runoff from a banana paddock during a single intensive storm event.
2004 Elsevier Ltd. All rights reserved. Keywords: Wet Tropics; Nutrients; Sediment; Agricultural runoff; Urban stormwater; Flux

1. Introduction A number of studies have been conducted in north Queensland measuring nutrient and sediment export from agricultural watersheds on a catchment scale (e.g. Hunter et al., 1996; Mitchell et al., 1996, 1997, 2001; Hunter and Walton, 1997; Mitchell and Furnas, 1997; Bramley and Muller, 1999; Cogle et al., 2000, 2003;

*

Corresponding author. Fax: +61 7 4781 5589. E-mail address: [email protected] (J. Faithful).

0025-326X/$ - see front matter
2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.marpolbul.2004.11.007

Johnson et al., 2001; Bramley and Roth, 2002; Rayment, 2002, 2003), a sub-catchment scale (e.g. Pearson et al., 2003) and farm scale (e.g. Prove, 1988; Prove and Hicks, 1991; McShane et al., 1993; Moody et al., 1996; Reghenzani et al., 1996; Hunter and Armour, 2001). Their conclusions have established that landuse activities in northern Queensland, such as grazing and intensive cropping, result in elevated nutrient and sediment discharges to receiving waters. These conclusions are similar to those examining the potential for runoff driven impacts upon north Queensland coastal marine environments (Crossland et al.,

100

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112

1996; Mitchell et al., 1996; Williams, 2001; Brodie, 2002; CRC Reef, 2003; Fabricius and DeÕath, 2004). However, it has been argued that because the marine environment is highly dynamic and extremely variable, especially with regards to increased sediment supply, the Great Barrier Reef has been able to exert substantial tolerance to environmental change having endured repeated physical stresses on a local, regional and global scale (Larcombe and Wolfe, 1999; Canegrowers, 2002). Paddock-scale investigations, such as reported here, enhance the understanding of the relative contribution of the different landuse activities within the Wet Tropics catchments to the overall nutrient and sediment export to the coast. This study was initiated with the co-operation of a number of banana and cane growers, catchment management groups (Barron River Integrated Catchment Management Association (BRICMA), Cardwell Shire Catchment Management Association and NRM Wet

Tropics Board). It provides a measure of the range of nutrients and suspended solids concentrations that exist in paddock scale drainage and urban lakes during rain events in the wet season within two distinct river catchments in north Queensland, i.e. the Tully–Murray Rivers catchment and the Granite Creek catchment (a component of the Barron River catchment southwest of Mareeba on the Atherton Tablelands).

2. Study area The two study catchments (the Tully–Murray Rivers and Granite Creek systems) are located in the Wet Tropics region on the northeast coast of Queensland, Australia (Fig. 1). Twelve farms (nine banana and three cane) were located in the Tully River catchment region and four farms (three cane and one banana) were lo-

Fig. 1. A map of the Wet Tropics bioregion on the northeast coast of Queensland, Australia, showing the locations of the major river catchments. The catchments areas in which the farm surveys were conducted are highlighted.

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112

cated within the lower Murray River catchment. Two farms (one cane and one banana) and two urban lakes sites were situated in the Granite Creek catchment. Farm data provided by the landholders listed the slope of all the cane paddocks, and seven of the 11 banana paddocks, including the Granite Creek farms, as less than 5%. The slope of the remaining four banana paddocks ranged between 10% and 30%. Other data relevant to each farm, such as soil types, fertiliser varieties and application data are provided in Faithful and Finlayson (2004). The Tully River is a component of the Tully–Murray Rivers Catchment Area (Fig. 1) and covers an area of approximately 1683 km2 (Brodie et al., 2001). It is AustraliaÕs least variable river with respect to annual discharge, making it relatively unique to Australian river systems renown for their extreme variability (Mitchell et al., 2001). The Tully River rises in the Cardwell Range, and 65% of the catchment area, predominantly within the mountain ranges, is in the Wet Tropics World Heritage Area. The Tully River floodplain has been extensively cleared and approximately 15% of the total catchment is under sugar cane agriculture, 2% under horticulture (of which the largest component is banana farming) and 19% is grazing land (Brodie et al., 2001). Sugar cane dominates agriculture land use activities on the floodplain, but in recent years banana agriculture has expanded throughout the floodplain. Mean annual rainfall within the catchment is 2855 mm, with mean annual runoff of 1954 mm/m2 and mean annual discharge of 3.3 km3 (3.3 · 106 ML). There has been significant alteration to the hydrological regime in the floodplain and pesticides and high nutrients loads (particularly nitrate) are evident in the lower reaches of the catchment (Brodie et al., 2001). The Murray River catchment has an area of 1107 km2 and also originates in the Cardwell Range. Almost half its area comprises grazing (47%) and the remainder is within the Wet Tropics World Heritage Area (47%) (Brodie et al., 2001). Other land uses included cane farming (5%) and horticulture (<1%), although increases in cane and horticulture areas within the catchment since 2000 are likely. Mean annual rainfall is 2098 mm, which is very similar to the Tully River catchment, although the mean annual runoff of 958 mm/ m2 and discharge of 1.1 km3 are approximately half and a third, respectively, of that of the Tully River. The Granite Creek catchment (Fig. 1) is a western tributary of the Barron River catchment. The area of the Barron catchment is 2902 km2, of which the Granite Creek catchment is 188 km2 (Cogle et al., 2000). Granite Creek is in the distinct Atherton Tablelands portion of the Barron River catchment, and has its confluence with the Barron River at the township of Mareeba. The Granite Creek catchment is moderately developed with respect to urbanisation (2% urban and 6% rural residen-

101

tial), agriculture (23%, comprised of irrigated mixed cropping such as sugarcane and horticulture) and grazing (6%). The amount of forested area encompasses 48% of the catchment and recently cleared land represents 12% of the catchment (Cogle et al., 2000). The mean annual rainfall for Mareeba is 910 mm, which is less than half that for the Tully–Murray Rivers catchment. Tinaroo Dam, on the Barron River, provides irrigation water for the farms within the Granite Creek catchment and therefore supplements flow through Granite Creek. The Bicentennial Lakes system in Mareeba was used to provide an indication of the water quality of urban runoff. These lakes, which comprise a series of three small interconnected lagoons, are sourced from Basalt Gully, a small rural catchment to the south of Mareeba. The lakes are supplemented by Tinaroo Dam irrigation supply to the region (approximately 1 ML/day, Jane Greer, BRICMA, personal communication, 2004) and wet season inflows from urban stormwater drains. The outflow from the lakes discharges into Granite Creek. The sampling points were located at the inflow point (Costin Street crossing) and at the outflow point (Keeble Street crossing).

3. Methods An important feature of this study was the involvement of the landholders and project catchment co-ordinators in collecting water samples and physico-chemical water quality measurements during drainage flow (no automated sampling was undertaken during this study). As part of the program, workshops were held to introduce the landholders to the sampling program, and train them to collect water samples, use a multi-parameter field meter and record relevant sampling information on field data sheets. There were three levels of farm investigation: primary sites with crump weirs installed on paddock drains to allow for continuous monitoring of drainage discharge and meteorological data that were sampled frequently during events; secondary sites that had gauging boards installed in the drainage channels and were sampled less regularly than the primary sites, and; random sites that were not gauged and were sampled opportunistically. Infrastructure works, such as weir construction, were completed by Queensland Department of Natural Resources, Mines and Energy (DNRM&E) hydrographers and landholders in November 2002. The weir structure for each of the primary sites was designed to accommodate twice the estimated annual flood; that is, approximately a 1 in five year flood event. Therefore, the type of weir selected for each primary site (e.g. v-notch or crump weir structures) was dependent on the paddock area. For example, the weir structure installed for the 2 ha paddock used for determining flux estimates during

102

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112

a storm event had a 0.73 cumec capacity, 1.4 m throat width, 1:10 crest slope crump weir. Each primary site was provided with Hydrological Services America TB3 1.0 mm (Ventura, USA) or TB5 0.5 mm tipping bucket rain gauges, with the core instrumentation consisting of a 3.5 m range Druck PTX1830 (Leicester, United Kingdom) pressure transducer coupled to an RRDLAN data logger through a locally fabricated vent box. The data logger was downloaded on a two-month service interval by DNRM&E hydrographers and the data emailed to the corresponding author. Water quality sampling began in November 2002. The project was initially established as a one-year pilot

project but was considered to be an initial step towards the development of a cost effective, regional water quality monitoring system based upon a significant level of landholder involvement that could be sustainable for a much longer timeframe. The 2002/2003 wet season was poor by Wet TropicÕs standards (Fig. 2) and as a consequence monitoring continued into the 2003/2004 wet season. 3.1. Field component Landholders and catchment co-ordinators collected water samples during periods when flow was evident in

800

Total Monthly Rainfall (mm)

700 600 500 400 300 200 100 0 Jul

Aug

Sep

Oct

Nov

Dec

Jan

Feb

Mar

Apr

May

Jun

Mar

Apr

May

Jun

July 2002 to June 2003 Tully RF

Total Monthly Rainfall (mm)

300

250

200

150

100

50

0 Jul

Aug

Sep

Oct

Nov

Dec

Jan

Feb

July 2002 to June 2003 Mareeba RF

Fig. 2. Monthly rainfall data for Tully (Tully Post Office, BoM Station No. 032070) and Mareeba (Mareeba Post Office, BoM Station No. 0031039) for July 2002 to June 2003. Included in the graph are average monthly rainfall totals (stalks) for that station (Tully data collected since 1983, and Mareeba data since 1952).

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112

the drainage channels. Drainage flow was very responsive to storm events, but for most paddocks only lasted for the duration of the storm event, dissipating very quickly after the cessation of the storm. Therefore, landholders had to respond quickly to these storm events so that they were able to monitor water quality and collect samples during drainage flow. Despite the spatial distance between the sites, the occurrence of major rainfall events was consistent between the primary sites, including the Granite Creek site for the events in February and March 2003. The primary paddock sites were subject to more comprehensive sampling on a regular and eventsbased regime, while the secondary and random sites were subject to less frequently sampling. Samples were collected from November 2002 to January 2004. Water samples were collected during drainage channel flow by the landholders at a point in the drain that represented the bottom of the paddock catchment (for the primary sites, this point was represented by the location of the weir) within the centre of the flow just below the water surface (approximately 10–15 cm depth). Samples for all TSS analyses were collected in 1 L polypropylene bottles and aliquots filtered through preweighed Whatman GF/C filters (nominal pore size 1.2 lm). Samples for total nitrogen (TN) and total phosphorus (TP) analyses were collected in 100 mL polyethylene bottles. Samples aliquots for total filterable nitrogen (TFN) and total filterable phosphorus (TFP), ammonia, NOX (nitrate + nitrite) and filterable reactive phosphorus (FRP) were filtered on-site through prerinsed Sartorius MiniSart filter modules (0.45 lm pore size) into 10 mL Sarstedt polypropylene vials. Total and filterable nutrient samples were frozen as soon as possible after collection. Samples for biochemical oxygen demand (BOD) were collected in 1 L clean polypropylene bottles and stored on ice. Samples collected in the Tully–Murray Rivers catchment area were returned to the Tully–Murray Rivers Catchment Co-ordinator who stored the frozen nutrient samples, filtered the samples for total suspended solids (TSS), and returned BOD samples to the Australian Centre for Tropical Freshwater Research (ACTFR) Water Quality Laboratory at James Cook University, Townsville, within 24 h of sample collection. TSS filter membranes and frozen nutrient samples were batched and transported to the ACTFR laboratory in regular submissions. The methodology for TSS filtration is described in APHA (1998, Method 2540 D). The involvement of landholders within this study varied between catchments and farms, made obvious by the much lower number of samples collected at the farms in the Granite Creek catchment, and where the full set of water samples was not collected during each sampling event (e.g. filtered nutrient samples were not always collected). The Granite Creek paddock drainage and urban lakes samples were collected by DNRM&E Mareeba

103

staff. Samples for TSS were analysed directly by the DNRM&E, and nutrient samples were immediately frozen, batched and transported to ACTFR in several batched submissions. Samples for BOD analyses were not collected from the Granite Creek sites due to the difficulty in returning samples to the ACTFR laboratory within the time constraints for this particular analysis (i.e. within a maximum of 24 h of sample collection). Quality assurance for field measurements and sample integrity was undertaken by ACTFR staff on two occasions during the study period, which coincided with intense rainfall events (February 29 and March 15, 2003). ACTFR field staff monitored the sample collection and physico-chemical measurement process undertaken by landholders and the Tully–Murray Rivers Catchment Co-ordinator to ensure that appropriate procedures were being adhered to for reliable water sample collection and field measurements of physico-chemical parameters. Duplicate samples were also collected on these occasions by the ACTFR field staff, and for all parameters the results correlated within acceptable levels (i.e. the variability of results for the replicates was less than 5%). 3.2. Laboratory component TSS filter membranes submitted by the Tully Catchment Co-ordinator were oven-dried at 103–105 C for 24 h and re-weighed to determine the dry TSS weight. Samples for TN and TP, and TFN and TFP, were digested using an alkaline persulfate technique (modified from Hosomi and Sudo, 1987) and the resulting solution also simultaneously analysed for NOX and FRP by segmented flow autoanalysis using an ALPKEM Flow Solution II (Alpkem Corporation, Wilsonville, Oregon, USA). The analyses for NOX, ammonia and FRP were also conducted using modified standard segmented flow auto-analysis techniques (APHA, 1998; Methods 4500NO3 F, 4500-NH3 G and 4500-P G). Particulate nutrient concentrations were estimated by the subtraction of the total filterable nutrient concentration from the total nutrient concentration. BOD analyses were conducted as per the standard method described in APHA (1998, Method 5210).

4. Results and discussion 4.1. Banana paddocks Table 1 provides a summary of results from the banana farms. Maximum nutrient concentrations are high, but these represented samples collected near peak flow periods in the drainage channel. The predominance of samples representing the falling stage of the flow

104

Table 1 Water quality data for the different landuse types utilised in the study for samples collected between November 2002 and January 2004 Catchment

Granite Creek

Sugarcane plots Tully/Murray Rivers

Granite Creek

Urban lakes Granite Creek

Particulate nitrogen (lg N/L)

n Min Max Median

115 630 20,900 2580

105 <50 9630 483

n Min Max Median

3 1620 4390 2030

n Min Max Median

Ammonia (lg N/L)

Nitrate + nitrite (lg N/L)

Total phosphorus (lg P/L)

Particulate phosphorus (lg P/L)

Filterable reactive phosphorus (lg P/L)

Total suspended solids (mg/L)

Biochemical oxygen demand (mg/L)

99 7 669 60

99 64 19,500 1440

113 12 5120 286

105 <5 4430 84

99 1 725 81

107 2 7250 75

27 <2 8.5 4.1

3 962 2370 1120

3 49 439 76

3 503 1320 705

3 782 1340 853

3 726 1270 799

3 36 72 41

3 2680 3900 2890

– – – –

81 270 15,500 3110

81 <50 1860 200

81 2 2250 20

81 40 13,500 2420

81 7 1430 50

81 <5 1420 32

81 <1 77 3

68 1 1840 14

27 <2 4.5 2.0

n Min Max Median

11 480 12,100 3080

11 40 1260 170

11 10 1200 137

11 75 8810 2480

11 43 140 65

11 9 59 34

11 4 20 11

11 7 107 63

– – – –

n Min Max Median

26 156 1570 369

26 <50 933 82

26 3 82 10

26 3 341 57

26 9 605 40

26 <5 419 16

26 2 171 14

26 1 323 11

– – – –

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112

Banana plots Tully/Murray Rivers

Total nitrogen (lg N/L)

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112

events is evidenced by the medians being considerably lower than the maximum concentration. TN concentrations ranged from 630 to 20,900 lg N/ L, and were comprised mostly of nitrate (the dominant component of Ônitrate + nitriteÕ) with negligible ammonia concentrations despite the use of both ammoniaand nitrate-based fertilisers on most farms (Table 1). This is due to the use of nitrogen based fertilisers (both nitrate, ammonia and urea forms) on soils that are capable of oxidizing the urea and ammonia-based fertilisers to nitrate. Fertiliser application on the banana paddocks occurred on a frequent basis (e.g. a number of times during the growing season), usually by fertigation, which is a process of fertilising and irrigating the crop at the same time (the fertiliser is applied in a dissolved form in the irrigation supply, not solid form). Nitrate is particularly soluble and moves through the soil profile and adsorbs to soil particles as an anion, accumulating in the groundwater supply. With each successive infiltration and consequent soil saturation after wet season rainfall, nitrate is remobilised and transported with soil water movement via runoff and horizontal movement to the paddock drain. This is a common feature in Queensland Wet Tropics floodplain agricultural regions (Rasiah and Armour, 2001; Rasiah et al., 2003). The TP was comprised mostly of particulate P (Table 1) because P applied to the crops as phosphate not immediately taken up by the crop rapidly adsorbs to fine soil particles. FRP ranged between 1 and 725 lg P/L in the runoff, but in much lower concentrations than TP (Table 1). The highest TP concentrations occurred during high flow rates in the paddock drains. The range of TSS concentrations was 2–7250 mg/L (Table 1). Like P, the highest TSS concentrations occurred during the height of the flow peaks. BOD concentrations for the Tully– Murray Rivers farms ranged between <2 and 8.5 mg/ L. This was attributed to the small drainage paddocks and the higher mass of organic material (trash) around the crops and possibly to sap from the plants that may be washed into the runoff via stem flow during rainfall events. The small number of samples (n = 3) collected from the banana paddock in the Granite Creek catchment makes any direct comparison with the Tully–Murray catchment data difficult. 4.2. Cane paddocks The cane farm data show some similar trends to the banana farms. Nitrogen concentrations in the drainage flow were high (Table 1), and as reported for the banana paddock drainage, were primarily comprised of nitrate. TN concentration ranged from 270 to 15,500 lg N/L with NOX concentrations between 40 and 13,500 lg N/ L. However, the concentrations of TP and TSS were much lower than that listed for the banana paddock

105

runoff (Table 1), varying between 7 and 1430 lg P/L, and 1 and 1840 mg/L, respectively. This reflects the obvious differences between crops and between farms, and land management practices. The lower TSS and phosphorus concentrations in the runoff is attributed primarily to a general lack of slope of the cane paddocks (<5%), the greater ÔcanopyÕ coverage minimizing soil exposure (particularly in the time of the year that the study was focused) and because fertiliser application occurs generally as a once-only event in the dry season prior to harvesting (June to November). The aerial coverage of the plant (the high plant density per unit area) significantly reduces the amount of direct rain contact with the soil on the paddock (dependent on the age of the plant), and wet seasonal rainfall influences (which will be more significant with higher rainfall intensity) are therefore generally limited to the period when the cane is in its early growing stage. BOD concentrations from the Tully–Murray Rivers farms ranged from <2 to 4.5 mg/L. The median and upper range BOD concentrations were approximately half that reported for banana, but the BOD samples were collected only during the early 2002/2003 wet season, i.e. the growing phase of cane, therefore negligible amounts of organic material and sugar accumulation in the soil would have been expected, as has been shown for harvested paddocks where BOD concentrations in drill drainage reached 300 mg/L (Bohl et al., 2002). Despite the lower number of samples collected from the Granite Creek cane paddock (11 samples cf. 81 from Tully/Murray farms), median nutrients and TSS concentrations from two catchment regions were generally similar, with NOX the most significant contributor to TN (Granite Creek, 81% compared to Tully–Murray, 78%). Particulate P was also the predominant contributor to TP (52% compared to 62%). Comparison of water quality data for small agricultural sub-catchments and paddock-scale watersheds in the Wet Tropics region of Queensland, as well as urban receiving lagoons in Townsville are provided in Table 2. Although the studies are limited, particularly those that report actual nutrient and sediment concentrations in runoff (as opposed to annual load estimates, e.g. McShane et al., 1993; Prove et al., 1995; Moody et al., 1996; Bloesch et al., 1997; Hunter and Walton, 1997), they generally refer to studies that have collected flow data during both ambient and flow event conditions (although Pearson et al., 2003, provides some intensive cane paddock event sampling data), from much larger watersheds than used in this study. Additionally, focus has been predominantly on cane farms. Nevertheless, the range of concentrations listed in Table 2 follow similar trends to that observed in this study, especially with respect to maximum nutrient and sediment concentrations. The median values are lower than reported for this study due to the larger watershed areas, and hence

106

Table 2 Water quality data for small agricultural sub-catchments, paddock-scale and urban watersheds in studies undertaken in the coastal tropics region of Queensland Site/landuse

Total nitrogen (lg N/L)

Particulate N (lg N/L)

Ammonia (lg N/L)

Nitrate (lg N/L)

Herbert River

Crisafulli Ck 72 ha 92% cane

n Min Max Median

20 662 6393 1697

34 33 825 197

37 2 1250 147

40 1 4584 34

Herbert River

Cane >70% cane

n Min Max Median

88 288 6393 1125

118 19 953 170

138 1 1250 49

Herbert River

Cane Storm Event

n Min Max Median

59 140 11,400 2140

– – – –

Wet Tropics

Cane

Min Max

1700 8500

Granite Creek (Tablelands)

Mixed

n Min Max Median

Ross River (Townsville)

Urban

n Min Max Median

Total phosphorus (lg P/L)

Total suspended solids (mg/L)

Reference

Particulate phosphorus (lg P/L)

Filterable reactive phosphorus (lg P/L)

28 32 518 199

35 2 364 47

39 3 205 37

39 4 268 78

Bramley and Muller (1999)

142 1 5600 33

98 17 518 119

120 1 364 37

140 1 237 25

93 1 268 33

Bramley and Muller (1999)

58 1 1200 469

43 5 9980 1715

58 11 958 230

– – – –

43 2 403 35

146 0.6 354 18

Pearson et al. (2003)

– –

– –

– –

210 1900

– –

– –

60 500

Brodie (2002)

112 12 12,700 282

– – – –

112 1 4850 14

112 1 2700 32

112 5 33,500 5

– – – –

112 1 463 11

112 <1 1800 6

81 53 1052 375

– – – –

23 4 454 8

23 6 106 7

81 8 135 43

– – – –

– – – –

58 2 44 14

Cogle et al. (2000)

Faithful (2000, 2003a,b)

This data includes ambient condition and storm event data, unless otherwise indicated. No count or median data was provided for the Wet Tropics cane water quality data in Brodie (2002).

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112

Catchment/source

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112

higher dilution potential, and the occurrence of ambient or baseflow conditions during sampling. The larger watersheds are comprised of a lower proportion of cropping area, and therefore nutrient and sediment exports from the cropping portion are ÔdilutedÕ with low nutrient and sediment runoff from the non-cropping component (e.g. forest, grazing, etc.). High nutrient and sediment concentrations are likely to be encountered in runoff from smaller watersheds.

5. Urban lakes The urban lakes water quality data are listed in Table 1. The nutrient and TSS concentration ranges of 156– 1570 lg N/L, 9–605 lg P/L and 1–323 mg/L, are much lower then the concentrations reported for the banana and cane farms, with higher nutrient and TSS concentrations occurring during storm rainfall periods and associated inputs to the lakes. In periods of no stormwater inputs to the lakes, the water quality is likely to represent that of the irrigation supply; that is, Tinaroo Dam supply. Nutrient and sediment concentrations from the lakes were similar to the range reported for coastal urban impoundments (Faithful, 2000, 2003a,b). The Townsville lagoonsÕ data (Table 2) have generally lower nutrient and sediment concentrations, with the exception of the median TN data, which may be a function of smaller watersheds with no agricultural landuse apart from garden or parkland fertilisation. The Bicentennial Lakes are a receiving aquatic environment, in contrast to the farm drainage channels, which can be considered as temporary conduits transporting runoff

107

to receiving water bodies, and are therefore subject to other water quality processes. 5.1. Flux estimate During the study, one banana farmer in the Tully– Murray Rivers Catchment intensively sampled over a major storm event in varying intervals between April 19 and May 6, 2003. Despite missing the rising stage of the drainage flow at the beginning of the storm event, the frequency of sampling through this event was extremely useful in that it provided a measure of the influence of a single pronounced rainfall event on the export of nutrients and TSS from a farm paddock. Fig. 3 shows the daily rainfall (mm) at the site and gauged flow data (cumec) between November 27, 2003 and May 1, 2004. The banana paddock had a watershed of 2 ha and during the study period, there was a total drainage discharge of 8.1 ML. Drainage flow was very responsive to rainfall events, especially those where the daily rainfall exceeded 50 mm. Fig. 4 illustrates the nitrogen data in accordance with the sampling during drainage flow over the study period. The insert in Fig. 4 presents the event data for the period April 18 to May 6, 2004 and illustrates the flashy flow response to rainfall. This is in contrast to the smooth flow transitions characteristic of flows in larger creeks or rivers. The ÔflashinessÕ is considered a function of the small watershed, the variability of the rainfall intensity and the paddockÕs hydrological characteristics. It is noted that the higher N concentrations occurred in the early stages of flow, which comprised a higher particulate nitrogen proportion. The proportion of nitrate with respect to

0.16

250

0.14

0.10

150

Fertiliser Application Periods

0.08

100

0.06 0.04

Daily Rainfall (mm)

Drainage Channel Flow (cumec)

200 0.12

50 0.02 0.00

0 01 Dec 02

01 Jan 03

01 Feb 03

01 Mar 03

01 Apr 03

Date - 2002/2003 Fig. 3. Daily rainfall data (mm) and flow gauging data (cumec) for the weir site at Farm No. 2148. The period of data logging (November 27, 2002 to May 6, 2003) coincides with the 2002/03 site investigation. Fertilizer application periods are also illustrated.

108

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112 0.20

5000

0.16

5000

4000 0.12 0.10

0.15

3000

4000

0.08 2000

0.06 0.04

1000

3000

0.02 0.00 18 Apr

0.10

21 Apr

24 Apr

27 Apr

30 Apr

03 May

0 06 May

2000

0.05

Nitrogen Fraction ( µg N/L)

Drainage Channel Flow Rate (cumec)

0.14

1000 Total N Oxidised N Ammonia

0.00

0 Dec 2002

Feb 2003

Apr 2003

Jun 2003

Aug 2003

Oct 2003

Dec 2003

Feb 2004

Survey Period Fig. 4. Nitrogen species concentrations (lg N/L) in samples collected during runoff events from a banana farm in the Tully–Murray Rivers catchment between November 2002 and February 2004. The insert shows the nitrogen speciation in water samples collected through a specific flow hydrograph in April 2003.

0.20

1200 600

0.16

500

0.12 400

0.10

0.15

1000

300

0.08 0.06

200

800

0.04 100

0.02

0.10

0.00 18 Apr

21 Apr

24 Apr

27 Apr

30 Apr

03 May

600

0 06 May

400 0.05

Phosphorus Fraction (µg N/L)

Drainage Channel Flow Rate (cumec)

0.14

200 Total P FRP

0.00

0 Dec 2002

Feb 2003

Apr 2003

Jun 2003

Aug 2003

Oct 2003

Dec 2003

Feb 2004

Survey Period Fig. 5. Phosphorus species concentrations (lg P/L) in samples collected during runoff events from a banana farm in the Tully–Murray Rivers catchment between November 2002 and February 2004. The insert shows the phosphorus speciation in water samples collected through a specific flow hydrograph in April 2003.

TN increased through the event probably as a result of increasing groundwater input through the falling stages of the flow. An estimate of N flux was determined using flow-weighted mean estimations from the data. For the period April 18 to April 30, the load of TN discharged over the weir was estimated as 9.2 kg N/ha, which for the fertiliser application on April 4 (refer to Fig. 4) represented 21% of that application. Fig. 5 shows the phos-

phorus data for the same period. Phosphorus response was directly related to flow rate, with higher P concentrations associated with high flow periods. During the storm event, the highest P concentrations were observed in the initial samples corresponding to the early stages of the drainage flow. In contrast to N, P concentrations were exhausted relatively quickly after the first stages of flow (Fig. 5—insert). Like N, it is important to

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112

109

0.20

800 0.16

200

0.12

150

0.10

0.15

0.08

600

100

0.06 0.04

50

0.02

0.10

0.00 18 Apr

21 Apr

24 Apr

27 Apr

30 Apr

03 May

400

0 06 May

0.05

200

0.00

Total Suspended Solids (mg/L)

Drainage Channel Flow Rate (cumec)

0.14

0 Dec 2002

Feb 2003

Apr 2003

Jun 2003

Aug 2003

Oct 2003

Dec 2003

Feb 2004

Survey Period Fig. 6. Total suspended solids concentrations (mg/L) in samples collected during runoff events from a banana farm in the Tully–Murray Rivers catchment between November 2002 and February 2004. The insert shows the TSS concentrations in water samples collected through a specific flow hydrograph in April 2003.

recognise that much higher concentrations of P are likely to have occurred in the rising stage of the event. The estimate of P loss through the event is 0.8 kg P/ha, which represented 10% of that applied in the application immediately before the event. Total suspended solids concentrations (Fig. 6) showed similar trends to that of TP concentrations, and for the specific event (Fig. 6—insert) an estimated loss of 126 kg/ha was determined. Sampling strategies in their current form will not be adequate to accurately determine the total flux of nitrogen, phosphorus or TSS for this event, primarily because the rising stage and the peak flows of minievents were not sampled. Therefore, it is anticipated that these estimates will grossly underestimate the

actual nutrient and sediment flux, which are required to determine the actual export of nutrients and sediment from paddocks during storm events. These data suggest that storm event runoff can transport significant amounts of nutrients and sediment in a single intense storm event, so wet season losses incorporating many storm events can be substantial. For example, banana farm export estimates from the Johnstone River catchment (also in the Wet Tropics) are 42 kg N, 6.8 kg P and 4000 kg TSS/ha per annum (Hunter et al., 2001), whereas annual average discharge estimates for the Tully Catchment (1987–2000) are 3.0 kg N and 0.8 kg P/ha (CRC Sugar, 2004), reflecting dilution effects from the whole catchment.

Table 3 Water quality data for Wet Tropics rivers River condition/season

River

Total N (lg N/L)

Filterable inorganic N (lg N/L)

Total P (lg P/L)

Filterable inorganic P (lg P/L)

TSS (mg/L)

Reference

Average flood

Herbert

1020

203

121

15

156

Mitchell et al. (1997)

Average wet

Tully South Johnstone Barron

400 434 594

230 200 112

32 50 57

7 8 8

– –

Mitchell et al. (1996) Mitchell et al. (1996) Mitchell et al. (1996)

Average dry

Tully

129

31 (NOX) + 15 (NH3)

10

2

4

Faithful and Brodie (1990) and Faithful (1990)

Annual average

Tully

–

–

5

–

Mitchell et al. (2001)

205

164 (NOX) + 29 (NH3) –

–

–

–

CRC Sugar (2004)

390

–

–

–

–

CRC Sugar (2004)

Upper Tully, Jarra Ck, Boulder Ck Lower Tully

110

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112

6. Conclusions The concentrations of nitrogen, phosphorus and TSS encountered in this study were often very high in comparison to those that are expected for Wet Tropics streams or rivers, especially during drainage events resulting from high rainfall periods, but may be typical of runoff from some agricultural landuse activities in the Wet Tropics. These concentrations are much higher than average nutrient and TSS concentrations reported for Wet Tropics rivers, particularly the Tully River during flood, wet season and annual flow conditions (Table 3). Median nitrogen concentrations in cane drainage runoff were higher (3110 lg N/L) than determined for banana paddock drainage (2580 lg N/L), although the maximum concentration resulted from a banana paddock (20,900 lg N/L). Banana paddocks had the highest maximum and median phosphorus and TSS concentrations (5120 and 286 lg P/L, and 7250 and 75 mg/L respectively) compared to the cane farms (1430 and 50 lg P/L, and 1840 and 14 mg/L respectively). These nutrient results generally correspond to the average fertiliser application of the banana and cane farms listed for the period October 2002 and July 2003 (cane: 143 kg N and 26 kg P/ha, n = 10; banana: 139 kg N and 40 kg P/ha, n = 11), where nitrogen application rates are equivalent, but more phosphorus was applied to the banana paddocks. The application rates were variable within the two crop types, but the characteristically higher slope and greater area of exposed soil on the banana paddocks exacerbated the greater export of phosphorus and sediment during storm events. The high median ratio of TN to TP in the sugar cane paddock drainage (Granite Creek, 47 and Tully–Murray, 64) is considered to be due to the high proportion of NOX in the TN, which persisted in the falling stage of the hydrograph flow, and lower TSS (and therefore TP) concentrations due to the flatter paddock slopes and lesser phosphorus fertiliser applications. The urban lakes nutrient and sediment concentrations were much lower than the agricultural drainage concentrations, because its catchment area is larger than the paddocks and any farm-sourced drainage inputs into Basalt Gully are diluted by the consistent irrigation supply from Tinaroo Dam prior to its flow into Bicentennial Lakes. The major outcomes from the project were the development of on-farm infrastructure and a high level of community involvement and awareness of water quality issues pertaining to nutrient and sediment runoff from agricultural paddocks. Community involvement in programs such as this is considered essential for conservation and landuse management (Williams, 2002). Information that a landholder can acquire regarding land management practices farming techniques from this type of study, such as reviewing paddock manage-

ment (e.g. reviewing fertiliser application rates and periodicity, weed control, the use of inter-row vegetation, vegetating headlands and drainage channels, etc.) are likely to result in more environmentally sustainable outcomes. The future aims of this study are to utilise duplicate paddocks to test the effectiveness of differing management practices between the paddocks, monitor for biocides in the paddock runoff, integrate a groundwater component within the paddocks to develop accurate water and nutrient balances for surface and sub-surface input to the drainage, and expand the investigation to include other land uses, such as horticulture (e.g. pineapples, mangoes, etc.), grazing and forestry.

Acknowledgments The landholders of the Tully–Murray Rivers and Granite Creek farms used in the study are thanked for their involvement in the project. The staff of the Natural Resources Management Board (Wet Tropics) Inc., Innisfail, Queensland, Cardwell Shire Catchment Management Association, the Barron River Integrated Catchment Management Centre, Queensland Department of Natural Resources, Mines and Energy (Mareeba and Innisfail Offices) for their assistance with the project, particularly Alan Hooper and Morgain Sinclair for installing and maintaining the weirs and data loggers, and providing the flow and rainfall data. Funding for the study was provided by the National Heritage Trust (NHT) under Project No. 2122015. Vivien McConnell, Jenny Cook, Sarah Thornton, and Joanne Knott (ACTFR Water Quality Laboratory) conducted analyses of the water samples. Jon Brodie, Damien Burrows, and two anonymous reviewers are thanked for critically reviewing the draft manuscript.

References APHA, 1998. Standard Methods for the Examination of Water and Wastewaters, 20th ed. American Public Health Association, American Water Works Association and Water Environment Federation, Washington, USA. Bloesch, P.M., Rayment, G.E., Pulsford, J.S., 1997. Regional total phosphorus budgets for sugar production in Queensland. Proceedings of the Australian Society of Sugar Cane Technologists 19, 213–220. Bohl, H.P., Bonnett, G.D., Fanning, D.J., Rayment, G.E., Davidson, A.B., 2002. Biological oxygen demand and sugars in irrigation water runoff from sugar cane fields. Proceedings of the Australian Society of Sugar Cane Technologists 21, 213–220. Bramley, R.G., Muller, D.E., 1999. Water quality in the lower Herbert River—The CSIRO dataset. CSIRO Land and Water Technical Report 16/99. CSIRO, Canberra. Bramley, R.G., Roth, C., 2002. Land-use effects on water quality in an intensively managed catchment in the Australian humid tropics. Marine and Freshwater Research 53, 931–940.

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112 Brodie, J., 2002. The effects of landuse on water quality in Australian northeastern catchments and coastal waterways. ACTFR Report No. 02/07, Australian Centre for Tropical Freshwater Research, James Cook University, Townsville. Brodie, J., Furnas, M., Ghonim, S., Haynes, D., Mitchell, A., Morris, S., Waterhouse, J., Audis, D., Lowe, D., Ryan, M., 2001. Great barrier reef catchment water quality action plan: a report to ministerial council on targets for pollution loads. Parts 1, 2 and 3. Great Barrier Reef Marine Park Authority, Townsville. Canegrowers, 2002. Industries in the great barrier reef catchment and measures to address declining water quality. Submission to the Productivity Commission, September 10, 2002, Queensland Cane Growers Organisation Ltd. Available from: . Cogle, A.L., Langford, P.A., Kistle, S.E., Ryan, T.J., McDougall, A.E., Russell, D.J., Best, E., 2000. Natural resources of the Barron River catchment 2: Water quality, land use and land management interactions. Information Series QI00033, Queensland Department of Primary Industries, Brisbane, Queensland. Cogle, A.L., Langford, P.A., Russell, D.J., Keating, M.A., 2003. Sediment and nutrient movement in a tropical catchment with multiple landuse. In: Dawson, N., Brodie, J., Rayment G., Porter, C. (Eds.), Proceedings of the 2nd National Conference on Aquatic Environments: Sustaining our Aquatic Environments—Implementing Solutions, Townsville, 20–23 November 2001, Queensland Department of Natural Resources and Mines, Brisbane, Australia. CRC Reef, 2003. Land use and the great barrier reef. Current state of knowledge (revised edition). CRC Reef Research Centre, Townsville. CRC Sugar, 2004. Final report 1995–2003. 8 Years of Achievement. Cooperative Research Centre for Sustainable Sugar Production, James Cook University, Townsville. Crossland, C.J., Done, T.J., Brunskill, G.J., 1996. Potential impacts of sugarcane production on the marine environment. In: Keating, B.A., Wilson, J.R. (Eds.), Intensive Sugarcane Production, Meeting the Challenges Beyond 2000, CAB International, pp. 432–436. Fabricius, K.E., DeÕath, G., 2004. A framework to identify ecological change and its causes: a case study of terrestrial run-off on coral reefs. Ecological Applications 14, 1448–1465. Faithful, J.W., 1990. Tully-Millstream hydroelectric scheme. Water quality sampling and testing program. Dry Season Monitoring Program. ACTFR Report No. 90/14, Australian Centre for Tropical Freshwater Research, James Cook University, Townsville. Faithful, J.W., 2000. Water quality monitoring program of the willows gardens ornamental ponds, Townsville. ACTFR Report No. 00/06, Australian Centre for Tropical Freshwater Research, James Cook University, Townsville. Faithful, J.W., 2003a. A brief limnological assessment of kingfisher lagoon: a small embayment in Black Weir, Ross River, Townsville. ACTFR Report No. 03/10, Australian Centre for Tropical Freshwater Research, James Cook University, Townsville. Faithful, J.W., 2003b. A follow-up limnological assessment of kingfisher lagoon: a small embayment in Black Weir, Ross River, Townsville. ACTFR Report No. 03/17, Australian Centre for Tropical Freshwater Research, James Cook University, Townsville. Faithful, J.W., Brodie, J.E., 1990. Tully-Millstream hydroelectric scheme. Water Quality Sampling and Testing Program. Post-Wet Season Monitoring Program. ACTFR Report No. 90/09, Australian Centre for Tropical Freshwater Research, James Cook University, Townsville. Faithful, J.W., Finlayson, W., 2004. Water quality assessment for sustainable agriculture: Tully–Murray Rivers catchment area and Granite Creek on the Atherton Tablelands. ACTFR Report 03/18, Australian Centre for Tropical Freshwater Research, James Cook University, Townsville.

111

Hosomi, M., Sudo, R., 1987. Simultaneous determination of total nitrogen and total phosphorus in freshwater samples using persulfate digestion. International Journal of Environmental Studies 27, 267–275. Hunter, H.M., Armour, J.D., 2001. Offsite movement of nutrients: contrasting issues at three Australian study sites. In: Proceedings of the Offsite Movement of Agrochemicals in Tropical Sugarcane Production Workshop, Bundaberg, May 8–9, 2001. ACIAR Project No. 9446. Hunter, H.M., Walton, R.S., 1997. From land to river to reef lagoon. Land use impacts on water quality in the Johnstone catchment. Queensland Department of Natural Resources, Brisbane, 10 pp. Hunter, H.M., Walton, R.S., Russell, D.J., 1996. Contemporary water quality in the Johnstone River catchment. In: Hunter, H.M., Eyles, A.G., Rayment, G.E. (Eds.), Downstream Effects of Land Use. Queensland Department of Natural Resources, Brisbane, pp. 339– 345. Hunter, H., Sologinkin, S., Choy, S., Hooper, A., Allen, W., Raymond, M., Peeters, J., 2001. Water Management of the Johnstone Basin. Queensland Department of Natural Resources and Mines, Brisbane. Johnson, A.K.L., Bramley, R.G., Roth, C., 2001. Land cover and water quality in river catchments of the Great Barrier Reef Marine Park. In: Wolanski, E. (Ed.), Oceanographic Processes of Coral Reefs Physical and Biological Links in the Great Barrier Reef. CRC Press, Boca Raton, pp. 19–37. Larcombe, P., Wolfe, K.J., 1999. Increased sediment supply to the Great Barrier Reef will not increase sediment accumulation at most coral reefs. Coral Reefs 18, 163–169. McShane, T.J., Reghenzani, J.R., Prove, B.G., Moody, P.W., 1993. Nutrient balances and transport from sugar cane land—A preliminary report. Proceedings of the Australian Society of Sugar Cane Technologists 15, 268–275. Mitchell, A.W., Furnas, M.J., 1997. Terrestrial inputs of nutrients and suspended sediments to the GBR lagoon. The Great Barrier Reef; Science, Use and Management, James Cook University, Townsville. CRC Reef Research Centre, James Cook University. Mitchell, A.W., Reghenzani, J.R., Hunter, H.M., Bramley, R.G.V., 1996. Water quality and nutrient fluxes from river systems draining to the Great Barrier Reef Marine Park. In: Hunter, H.M., Eyles, A.G., Rayment, G.E. (Eds.), Downstream Effects of Land Use. Queensland Department of Natural Resources, Brisbane, pp. 23– 33. Mitchell, A.W., Bramley, R.G.V., Johnson, A.K.L., 1997. Export of nutrients and suspended sediment during a cyclone-mediated flood event in the Herbert River catchment, Australia. Marine and Freshwater Research 48, 79–88. Mitchell, A.W., Reghenzani, J.R., Furnas, M.J., 2001. Nitrogen levels in the Tully River—A long-term view. Water Science and Technology 43, 99–105. Moody, P.W., Reghanzani, J.R., Armour, J.D., Prove, B.G., McShane, T.J., 1996. Nutrient balances and transport at farm scale— Johnstone River catchment. In: Hunter, H.M., Eyles, A.G., Rayment, G.E. (Eds.), Downstream Effects of Land Use. Queensland Department of Natural Resources, Brisbane, pp. 347– 351. Pearson, R.G., Crossland, M., Butler, B.M., Manwaring, S., 2003. Effects of Cane-ield drainage on the ecology of tropical waterways. ACTFR Report 03/04, Australian Centre for Tropical Freshwater Research, James Cook University, Townsville. Prove, B.G., 1988. Soil erosion research in cane fields on the Wet Tropical coast of north-east Queensland. In: Baldwin, C.L. (Ed.), Workshop on Nutrients in the Great Barrier Reef Region. Great Barrier Reef Marine Pak Authority, Townsville, pp. 30–32. Prove, B.G., Hicks, W.S., 1991. Soil and nutrient movements from rural lands of north Queensland. In: Yellowlees, D. (Ed.), Landuse Patterns and Nutrient Loading of the Great Barrier Reef Region.

112

J. Faithful, W. Finlayson / Marine Pollution Bulletin 51 (2005) 99–112

James Cook University of North Queensland, Townsville, pp. 67– 76. Prove, B.G., Doogan, V.J., Truong, P.N.V., 1995. Nature and magnitude of soil erosion in sugarcane land on the Wet Tropical coast of north-east Queensland. Australian Journal of Experimental Agriculture 35, 641–649. Rasiah, V., Armour, J.D., 2001. Nitrate accumulation under cropping in the ferrosols of far north Queensland Wet Tropics. Australian Journal of Soil Research 39, 329–341. Rasiah, V., Armour, J.D., Yamamoto, T., Mahendrarajah, S., Heiner, D.H., 2003. Nitrate dynamics in shallow groundwater and the potential for transport to off-site water bodies. Water, Air, and Soil Pollution 147, 183–202. Rayment, G.E., 2002. Water quality pressures and status in sugar catchments—exposure draft. CRC Sugar/Department of Natural Resources Mines and Energy, Indooroopilly.

Rayment, G.E., 2003. Water quality in sugar catchments of Queensland. Water Science and Technology 48, 35–47. Reghenzani, J.R., Armour, J.D., Prove, B.G., Moody, P.W., McShane, T.J., 1996. Nitrogen balances for sugarcane plant and first ratoon crops in the Wet Tropics. In: Wilson, J.R., Hogarth, D.M., Campbell, J.A., Garside, A.L. (Eds.), Sugarcane: Research Towards Efficient and Sustainable Production. CSIRO Division of Tropical Crops and Pastures, Brisbane, pp. 275– 277. Williams, D.Mc.B., 2001. Impacts of terrestrial runoff on the Great Barrier Reef world heritage area. Report to CRC Reef, CRC Reef Research Centre/Australian Institute of Marine Science, Townsville. Williams, W.D., 2002. Community participation in conserving and managing inland waters. Aquatic Conservation: Marine and Freshwater Ecosystems 12, 315–326.