Landcover and Water Quality in River Catchments of the Great Barrier Reef Marine Park Andrew K.L. Johnson, Robert G.V. Bramley, and Christian H. Roth CONTENTS Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Methods and Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 The Herbert River Catchment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Landcover. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Surface Water Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Contemporary Broadscale Landcover Change in GBRMP Catchments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Landcover Change in the Herbert River Catchment . . . . . . . . . . . . . . . . . . . . 26 Water Quality in the Herbert River Catchment . . . . . . . . . . . . . . . . . . . . . . . . 27 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 INTRODUCTION The Great Barrier Reef Marine Park (GBRMP) covers an area of approximately 350,000 km 2 and spans almost 2,000 km of the east coast of Queensland, Australia. The GBRMP is a marine ecosystem that is recognised internationally for its unique biological and physical features. Fifteen river catchments, covering an area of approximately 375,000 km 2 , drain directly into the GBRMP (Figure 1). Land use in these catchments is dominated in areal terms by grazing. Cropping, particularly sugar- cane production, is a major user of land resources in a number of catchments and is predominantly located on fertile coastal floodplains immediately adjacent to GBRMP waters (Table 1). 3 19 © 2001 by CRC Press LLC 20 Oceanographic Processes of Coral Reefs TABLE 1 Approximate Area of Major Land Uses in Catchments Adjoining the GBRMP 1996 Catchment Catchment Percentage of Catchment Area Name Area (km 2 ) Forest a Pristine b Grazing c Crops Urban NE Cape d 43,300 4.3 33.9 61.7 0.05 0.05 Daintree 2,130 37.7 31.7 26.7 1.9 2.0 Mossman 490 30.4 11.0 44.6 10.1 3.9 Barron 2,180 36.4 2.0 47.7 6.8 6.9 Mulgrave Russell 2,020 16.9 25.1 38.9 13.3 5.8 Johnstone 2,330 25.3 12.8 41.6 15.9 4.4 Tully 1,690 62.5 2.1 20.7 11.1 3.7 Murray 1,140 32.9 27.3 29.6 7.0 3.3 Herbert 10,130 9.5 9.7 71.1 7.0 2.7 Black 1,080 18.0 9.3 67.4 1.1 4.2 Haughton 3,650 0.8 10.8 74.0 10.9 3.5 Burdekin 129,860 1.0 1.3 94.8 1.0 2.0 Don 3,890 0.2 2.6 91.3 2.8 3.1 Proserpine 2,490 9.6 4.0 74.6 7.5 4.3 O’Connell 2,440 7.6 4.4 70.5 11.1 6.5 Pioneer 1,490 22.7 6.1 48.5 17.9 4.7 Shoalwater Bay-Sarina e 11,270 1.3 41.6 f 44.1 10.3 2.7 Fitzroy 152,640 6.7 2.3 87.5 3.3 0.2 Curtis Coast g 9,225 12.2 11.3 68.9 0.57 6.7 Total area 369,480 28,007 39,830 284,056 13,597 3,990 Total (%) 7.6 10.8 76.9 3.7 1.1 a Comprises state forests and timber reserves. b Comprises national parks and other reserves. c Comprises unimproved and improved pastures. d Comprises Jacky Jacky Creek, Olive-Pascoe, Lockhart, Stewart, Jeannie, Normanby, and Endeavour catchments. e Comprises Plane Creek, Styx, Shoalwater Creek, and Water Park Creek catchments. f Approximately 65% of this area occupied by the Shoalwater Bay Field Training Area of the Australian Defence Forces. g Comprises Calliope, Boyne, and Baffle Creek catchments. Source: QDPI, 1993; EPA, 1999; Johnson et al., 1999. Current environmental trends suggest a decline in coastal terrestrial and riverine systems, and on the adjacent GBRMP marine environment (Anonymous, 1993; Arthington et al., 1997). The vegetation of many of the river catchments adjoining the GBRMP has been extensively cleared (Russell & Hales, 1996) since the mid-19th century. Freshwater wetlands and riparian forests once covered large areas of coastal floodplains which are now used for agriculture (Tait, 1994; Johnson et al., 1999). © 2001 by CRC Press LLC Landcover and Water Quality in River Catchments 21 Prior to clearing, these wetlands would have provided extensive buffer strips and freshwater habitats adjacent to coastal river systems, estuaries, and shorelines. Clearing, notably for sugarcane cultivation, has left only remnants of these ecosys- tems (Russell et al., 1996). Present-day wetlands and riparian forests in many catch- ments are frequently narrow and sparsely vegetated and have been invaded by exotic weeds (Johns et al., 1997). It is likely those wetlands and riparian forests in such poor condition have suffered a corresponding degradation of their intrinsic ecological val- ues (Arthington et al., 1997). The status of freshwater wetlands and riparian forests in river catchments adja- cent to the GBRMP has been reviewed superficially by a number of authors (Arthington & Hegerl, 1988; Anonymous, 1993; Blackman et al., 1996). Accounts have increasingly confirmed their very high biological richness, diversity, geograph- ical extent, importance as habitat for a similarly rich and diverse biota, and funda- mental role in ensuring the health of key GBRMP ecosystems. Of the 19 Queensland wetlands identified as having national importance (Blackman et al., 1996), 8 are located in areas immediately adjacent to or within the GBRMP. While the present sta- tus of these ecosystems is known, there have been no detailed assessments of histor- ical changes in coastal wetlands and riparian forests in GBRMP catchments. Similarly, while the current extent of landcover in river catchments adjoining the GBRMP is generally known, the spatial and temporal distribution of landcover since European settlement is poorly understood. The aim of this chapter is to describe broad-scale changes in landcover in GBRMP catchments and to examine in detail changes that have occurred using a case study undertaken in the lower Herbert River catchment. We also describe the likely impact that these changes have had on the water quality of the Herbert River. While the focus of the chapter is not on the impacts of these changes per se, we discuss sig- nificant issues that are central to the maintenance and function of estuarine and marine ecosystems in the GBRMP. METHODS AND MATERIALS THE HERBERT RIVER CATCHMENT The Herbert River catchment drains an area of approximately 10,000 km 2 to the Coral Sea and is the largest of the river systems located in Australia’s sub-humid to humid tropical northeast (latitude 15 to 19°S, longitude 145 to 146°E) (Figure 2). Average annual rainfall is approximately 2500 mm. Mean annual runoff for the catchment is 4991 ϫ 10 9 m 3 or 493 mm, and the rainfall-to-runoff ratio approximately 37% (Hausler, 1991). Natural vegetation consists predominantly of open Eucalyptus and Melaleuca woodlands, with areas of open grassy plains and dense Melaleuca wetlands. Rainforest patterns occur on the creek and river levees and on some of the northern ranges. Large areas of the upper catchment remain under natural vegetation, although much of the lower catchment has been cleared for crop production or exotic pas- tures. Agricultural and pastoral activities are the largest users of land (in area) in the © 2001 by CRC Press LLC 22 Oceanographic Processes of Coral Reefs catchment. The catchment has a population of approximately 18,000 (1993 Census), of which 75% are located in the lower catchment. LANDCOVER A desktop study was conducted to collect data from a range of published and unpub- lished sources on landcover in catchments adjoining the GBRMP. This activity drew heavily on work undertaken by the Queensland Statewide Landcover and Tree Study (SLATS) (QDNR, 1999a and 1999b). The study utilised Landsat Thematic Mapper (TM) imagery (spatial accuracy Ϯ30 m) and ground surveys to map changes in woody vegetation cover (where woody vegetation was defined as approximately 12% foliage projective cover or greater) between 1988, 1991, 1995, and 1997. The study attempted to map vegetation change for all perennial woody plants of sizes that could be distinguished by Landsat TM imagery. Accuracy of areal interpretation for the whole state was reported as Ϯ8% at a 95% confidence interval. Error data associated with misclassification were not reported, although incidences of misclassification in areas of pasture and in highly fragmented landscapes (e.g., narrow riparian zones in coastal areas) were acknowledged. Anecdotal evidence from field-workers also suggests the existence of substantive misclassification in grazing lands (A. Ash, personal communication). QDNR (1999a and 1999b) describes the method used in more detail. Landcover in the Herbert River catchment was visually interpreted from scanned and rectified 1:25,000 aerial photography acquired in 1943, 1961, 1970, 1977, 1988, and 1992 (spatial accuracy Ϯ7 m) and 1:10,000 digital orthophotography acquired in 1993, 1994, and 1995 (spatial accuracy Ϯ1 m). An unsupervised classification of SPOT Panchromatic and MSS imagery was used to map landcover in 1996 (spatial accuracy Ϯ10 m). Landcover boundaries were mapped onto a geo-referenced digital base (spatial accuracy Ϯ10 m) in ARCINFO GIS. The classification methodology (Johnson et al., 1999 and 2000) drew heavily on previous vegetation (Tracey, 1982; Blackman et al., 1992; Perry, 1995) and soil (Wilson & Baker, 1990) surveys in the region. Validation of mapping units and mapped boundaries was conducted in 1996 by vehicle and foot traverses. Approximately 150 sites were visited. Classification of units and bound- aries not inspected in 1996 was undertaken by extrapolation from equivalent photo- graphic units. In addition to mapping observed landcover, an estimate of landcover prior to European settlement (circa 1860s) in the Herbert was developed from a simple rule base that related remaining stands of native vegetation and the known distribution of soils, topography, relief, hydrology, and rainfall. A time series was developed to elu- cidate spatial and temporal change in landcover (Johnson et al., 1999). SURFACE WATER QUALITY A number of sites were selected to reflect the major landcover classes, soil types (Wilson & Baker, 1990; Wood, 1984, 1985, and 1988), and sub-catchments in the lower Herbert floodplain (Figure 3) on the basis that water sampled at any given site reflected the biophysical characteristics of the land upstream of that site. © 2001 by CRC Press LLC Landcover and Water Quality in River Catchments 23 Beginning in October 1992, surface grab samples of river water were taken at each of these sites at monthly intervals and also in response to rainfall events of inten- sity greater than 50 mm d Ϫ1 . The samples were collected either by lowering a bucket from a bridge at a point above the centre of the flowing part of the channel, or more directly by wading into the stream. The samples were collected in acid-washed poly- ethylene bottles and were stored in a portable refrigerator for transfer back to the lab- oratory. On each sampling occasion, the distance between the surface of the water and a fixed arbitrary point such as a bridge rail was also measured for later estima- tion of actual water depth and then discharge. The laboratory procedures used in this study have been detailed by Muller et al. (1995). Total concentrations of nitrogen (N) and phosphorus (P) were determined according to USEPA (1984) methodology. Total suspended solids (TSS) were deter- mined by gravimetric measurement of the amount of particulate material retained on 0.45 ␮m cellulose acetate filter papers. For the analysis of land use impact on water quality, the landcover classification (Figure 3) was simplified into land under sugarcane, grazing (i.e., improved grazing or Eucalyptus-dominated patterns), and forestry (i.e., plantation forestry or natural rainforest). This was done to simplify the attribution of water quality differences, given that for the majority of sites, several land uses exist upstream of those sites (i.e., water quality measurements made at a particular site may integrate the effects of more than one landcover class). This simplification of landcover categories is also consistent with the results of Hunter and Walton (1997), who found that in the Johnstone catchment, whilst it was possible to discriminate between the effects of intensive and non-intensive land uses on water quality, it was not possible to dis- criminate within these broad groupings. For the purposes of the present study, time of sampling was treated as an inde- pendent variable because although several authors (e.g., Hunter et al., 1996 and ref- erences therein; Mitchell et al., 1996 and 1997) have demonstrated the strong seasonality of riverine discharge and water quality in north Queensland rivers and their links to the strongly seasonal climate, our purpose here was to examine the effects of landcover on water quality. RESULTS CONTEMPORARY BROADSCALE LANDCOVER CHANGE IN GBRMP CATCHMENTS Tables 2 and 3 show contemporary woody vegetation changes in GBRMP catchments for the period 1991 to 1997. They show: • Large areas of woody vegetation converted to pasture in the Fitzroy, Burdekin, Normanby, Don, Proserpine, and Baffle Creek catchments, implying a change from extensive grazing woodlands to more intensive forms of grazing on improved pastures • Large areas converted to crops in the Herbert, Murray, Haughton, Plane Creek, and Fitzroy catchments © 2001 by CRC Press LLC 24 Oceanographic Processes of Coral Reefs TABLE 2 Rates of Change from Woody Vegetation to Other Landcover Classes in GBRMP Catchments 1991–1995 Rate of Woody Vegetation Change (km 2 yr Ϫ1 ) Catchment Catchment New % Name Area (km 2 ) Regrowth a Pasture b Crops c Forest d Urban e Total Area Jardine 3,288 0.1 0.07 0 0 0 0.07 0.002 Jacky Jacky Creek 2,916 0 0 0 0 0 0 0 Olive-Pascoe 4,199 0 0 0 0 0 0 0 Lockhart 2,847 0 0 0 0 0 0 0 Stewart 2,694 0 0 0 0 0 0 0 Jeannie 3,886 0.02 0 0 0 0 0 0 Normanby 24,319 0.07 4.74 5.93 0 0.23 10.91 0.045 Endeavour 2,063 0.07 0 0.15 0 0.14 0.29 0.014 Daintree 2,130 0.01 0.13 0.12 0 0 0.26 0.012 Mossman 490 0.03 0 0.01 0 0.05 0.07 0.014 Barron 2,180 1.61 0.4 1.42 0.14 0.15 2.1 0.096 Mulgrave-Russell 2,020 0.04 0.05 0.73 0.03 0.12 0.94 0.047 Johnstone 2,330 0 0.4 1.21 0.03 0.02 1.65 0.071 Herbert 10,130 0.54 1.35 18.55 2.65 0.33 22.88 0.226 Tully 1,690 0 0.03 0.85 0 0.05 0.93 0.055 Murray 1,140 1.15 0 7.66 0.08 0.06 7.79 0.683 Burdekin 129,860 5.15 5.29 0.4 0 1.16 6.85 0.005 Black 1,080 0.53 0.22 0.94 0 0.22 1.38 0.128 Ross 1,346 0.75 0.43 0 0 0.61 1.04 0.077 Haughton 3,650 6.37 1.9 3.31 0 1.06 6.28 0.172 Don 3,890 2.68 3.42 0.05 0 0.12 3.59 0.092 Proserpine 2,490 3.03 2.85 0.35 0 0.09 3.29 0.132 O’Connell 2,440 0.55 2.31 0.22 0.11 0.06 2.69 0.110 Pioneer 1,490 0 0.21 0.03 0 0 0.24 0.016 Plane Creek 2,547 1.32 2.67 4.97 0.1 0.11 7.85 0.308 Styx 3,018 0.59 3.85 0 0 0.09 3.93 0.130 Shoalwater Creek 3,698 0.71 1.19 0 0 0.07 1.26 0.034 Water Park Creek 1,756 1.18 0.72 0.42 1.1 0.2 2.44 0.139 Fitzroy 152,640 25.07 17.34 0.5 0.01 0.45 18.28 0.012 Calliope 2,204 0.11 0.59 0 0 0.68 1.28 0.058 Boyne 2,473 0.02 0.72 0 0 0.35 1.07 0.043 Baffle Creek 4,106 1.14 1.55 0.93 0.01 1.33 3.82 0.093 Total 369,480 52.84 52.43 48.75 4.26 7.75 113.18 0.069 a New regrowth is defined as areas which have changed from non-woody to woody within the period. b Areas cleared to pasture. Includes clearing for grazing, rural residential, future urban land use, native forestry on private land, privately owned plantations cleared to pasture. c Cleared for growing crops. d State forest clearing including plantation and native forest. Includes cleared private plantations that are replanted. e Cleared for mining, infrastructure, and urban development. Source: QDNR, 1999a. © 2001 by CRC Press LLC Landcover and Water Quality in River Catchments 25 TABLE 3 Rate of Change from Woody Vegetation to Other Landcover Classes in GBRMP Catchments 1995–1997 Rate of Woody Vegetation Change (km 2 yr Ϫ1 ) Catchment Catchment New % Name Area (km 2 ) Regrowth a Pasture b Crops c Forest d Urban e Total Area Jardine 3,288 0 0.02 0 0 0.11 0.13 0.004 Jacky Jacky Creek 2,916 0 0 0 0 0.01 0.01 0.000 Olive-Pascoe 4,199 0 0 0 0 0 0 0.000 Lockhart 2,847 0 0 0 0 0.04 0.04 0.001 Stewart 2,694 0 0 0 0 0 0 0.000 Jeannie 3,886 0 0 1.6 0 0.34 1.93 0.050 Normanby 24,319 0 0.14 0 0 0.63 0.77 0.003 Endeavour 2,063 0.04 0.16 4.4 0.01 0.38 4.94 0.239 Daintree 2,130 0 0.09 0.17 0.07 0 0.34 0.016 Mossman 490 0 0.03 0.07 0.01 0.13 0.24 0.049 Barron 2,180 0.05 1.22 5.69 0.3 0.53 7.74 0.355 Mulgrave-Russell 2,020 0.05 0.1 0.11 0 0.23 0.44 0.022 Johnston 2,330 0.05 0.71 0.85 0.01 0.01 1.57 0.067 Herbert 10,130 7.54 2.92 7.59 0.23 0.92 11.66 0.115 Tully 1,690 0.09 0 1.41 0.01 0.12 1.54 0.091 Murray 1,140 1.63 0.45 3.4 0.22 0.08 4.15 0.364 Burdekin 129,860 5.57 14.72 1.24 0 2.92 18.88 0.015 Black 1,080 1.8 1.6 1.73 0.03 0.27 3.62 0.335 Ross 1,346 0.65 1.24 0.02 0 1.37 2.63 0.195 Haughton 3,650 2.93 0.72 10.81 0 0.51 12.04 0.330 Don 3,890 2.39 1.39 1.19 0 0.26 2.84 0.073 Proserpine 2,490 1.5 3.56 10.27 0 0.16 13.99 0.562 O’Connell 2,440 0.49 1.13 2.84 0 0.1 4.06 0.166 Pioneer 1,490 0 1.36 0.92 0.02 2.76 5.07 0.340 Plane Creek 2,547 0.53 0.58 19.58 0.01 0.18 20.34 0.799 Styx 3,018 1.14 5.39 0.91 0 0.12 6.43 0.213 Shoalwater Creek 3,698 0.79 1.47 0.04 0 0.24 1.73 0.047 Water Park Creek 1,756 0.26 1.06 0.4 1.06 0.82 3.34 0.190 Fitzroy 152,640 0.69 17.28 4.05 0.18 0.31 21.83 0.014 Calliope 2,204 0.03 0.93 0.01 0.05 0.33 1.32 0.060 Boyne 2,473 0.02 0.37 0 0 0.1 0.47 0.019 Baffle Creek 4,106 1.03 7.43 1.03 0.09 2.5 11.06 0.269 Total 369,480 29.27 66.07 80.33 2.3 16.48 165.15 0.102 a New regrow this defined as areas which have changed from non-woody to woody within the period. b Areas cleared to pasture. Includes clearing for grazing, rural residential, future urban land use, native forestry on private land, privately owned plantations cleared to pasture. c Cleared for growing crops. d State forest clearing including plantation and native forest. Includes cleared private plantations that are replanted. e Cleared for mining, infrastructure, and urban development. Source: QDNR, 1999b. © 2001 by CRC Press LLC 26 Oceanographic Processes of Coral Reefs • Large areas of woody regrowth occurring in the Fitzroy, Burdekin, and Herbert catchments • Small areas converted to urban and forest uses in all catchments These tables show that in terms of the total catchment area delivering to the GBRMP, the current rate of woody vegetation change is small. They indicate, how- ever, that the rate of change from woody vegetation to agriculture remains high in a number of catchments. In the smaller coastal catchments it is reasonable to expect that these changes are most likely to be occurring on the fertile coastal floodplains imme- diately adjacent to the GBRMP. In contrast to the bigger basins, in particular the Fitzroy and the Burdekin, substantial conversion of woody vegetation to more inten- sive agricultural use is taking place in the interior of these basins (e.g., Brigalow Belt). The SLATS data presented here should be interpreted with caution. First, it is important to point out that substantial areas of woody vegetation (e.g., the Brigalow Belt in the Fitzroy and Burdekin basins) have been cleared prior to the last decade, so that current rates do not adequately reflect the absolute change in land use. Moreover, in some instances the SLATS methodology has not always been able to correctly clas- sify vegetation or landcover classes. An example of this occurs in the Normanby catchment where large areas have been misclassified as regrowth and cleared for crops (see EPA, 1999; Johnson et al., 1999). Photointerpretation of chronosequences of aerial photos coupled with spatial analysis in GIS and underpinned by adequate ground truthing is a more reliable means of assessing landcover change, but is both an expensive and time-consuming methodology, restricting it to more detailed analysis in selected case studies. In the section that follows we describe in detail the changes that have occurred in the catch- ment of the Herbert River using such a case study approach. LANDCOVER CHANGE IN THE HERBERT RIVER CATCHMENT Landcover in the lower Herbert has changed substantially since European settlement in the 1860s. Johnson and Ebert (2000) describe changes in the catchment as a whole and show that since European settlement, approximately 7.5% of the total catchment area has been converted from native vegetation to other landcover types (95% con- verted to agriculture). It is likely that landcover in the middle catchment has remained unchanged due to its inaccessibility and more recent (post-1950) status as a national park. Landcover in the upper catchment has also remained virtually unchanged over the last 140 years, with only small areas (i.e.,Ͻ1%) being converted to mining, agri- culture, and urban uses. However, the increase in grazing pressure and change in fire regimes experienced since European settlement have caused a marked structural change in plant communities in the upper catchment and shifted the balance between shrub and herbaceous layers (Johnson et al., 2000). Johnson et al. (1999 and 2000) focussed on changes in the lower Herbert (i.e., the area immediately adjacent to the GBRMP) and showed that significant changes in landcover have occurred in this part of the catchment (Figures 4 and 5). It can be seen that prior to settlement, the area was dominated by open grassland, rainforest © 2001 by CRC Press LLC Landcover and Water Quality in River Catchments 27 patterns, mangrove patterns, Eucalypt woodlands, and Melaleuca communities. However, by the 1940s large losses of rainforest patterns and Melaleuca-dominated patterns had occurred and much of the native grassland had been converted either to grazing or sugarcane. Landcover remained relatively stable throughout the 1960s and early 1970s. However, the period between 1977 and 1996 saw a rapid expansion in the area under sugarcane. The consequences of this expansion have resulted in a decrease of approximately 65% in the area of Melaleuca-dominated patterns (comprising a 43% decrease between 1943 and 1996), a 60% decrease in the area of beachside vegetation, a 20% decrease in the area of Eucalypt woodland, and a 10% decrease in the areas of rain- forest patterns when compared to pre-European estimates. In contrast, the area of mangrove communities and open water has remained relatively stable since 1943, while the area of sugarcane has more than tripled between 1943 and 1996. As expected, the area of urban and industrial landcover has increased since European settlement, although the total area alienated is small within the context of total catch- ment area (i.e., Ͻ0.5%). WATER QUALITY IN THE HERBERT RIVER CATCHMENT Water quality, as expressed by median concentrations of TSS, significantly decreases in the lower Herbert catchment as the proportion of upstream land area under sugar- cane increases (Figure 6a). In contrast, as the proportion of upstream land area under grazing increases, median concentrations of TSS tend to be significantly lower; a similar but non-significant effect is observed for forestry. These results remain essen- tially unchanged irrespective of whether sampling sites which reflect the very large areas in the upper part of the Herbert catchment under grazing are included (Figure 6a) or excluded (Figure 6b) from the analysis, and also when other indices of water quality (e.g., fractions of N and P) are used, as concentrations of TSS and total N and total P are intercorrelated (Figure 7). The use of median concentrations as an indicator of water quality does not reflect the highly variable fluctuations of TSS and nutrients as the result of major rain events (Mitchell et al., 1996 and 1997), but rather can be assumed to better characterise the longer-term trends of low flow or base flow concentrations (i.e., a measure of “chronic” impact levels). Benthic faunas in tropical freshwater systems seem to be adapted to short-term “peaks” in key water quality parameters given the naturally high variations of flow, so assessing “chronic” changes in the levels of nutrient con- centrations may be a more meaningful method for assessing land use impacts on water quality in tropical systems (R. Pearson, personal communication). In the absence of robust discharge data for most of the sites sampled as part of the Herbert study, a more simplistic approach for assessing the likely contribution of the major forms of landcover in the Herbert toward sediment and nutrient discharge was used. All sites were grouped into three classes of landcover: sugarcane, grazing (Eucalyptus dominated patterns ϩ improved pastures), and forestry (plantation forestry ϩ natural rainforest), depending on which of these three land uses was pre- dominant upstream of any one sampling point. The threshold criterion to discriminate © 2001 by CRC Press LLC 28 Oceanographic Processes of Coral Reefs between relative dominance of any land use was Ͼ45%, with relative dominance in many cases Ͼ60%. Water quality parameters (TSS, total N and P) were then plotted as box plots. Numbers of samples analysed ranged from n ϭ 160 to 262, n ϭ 110 to 177, and n ϭ 104 to 185 for TSS, total N, and total P, respectively. As evidenced in Figure 8, sugarcane as a predominant land use clearly yields a significantly greater variation in concentrations of TSS, total N and P compared to grazing and forestry, with maximum values measured in high flow (i.e., 0.9 per- centile) — an order of magnitude or more than low flow values (i.e., 0.1 percentile). There is also a clear tendency for a greater variation and generally higher values of water quality parameters under grazing when compared to forestry. The data collected as part of this water quality study have been summarised by Bramley and Muller (1999). DISCUSSION The extent and nature of vegetation clearance can provide a useful indicator of envi- ronmental quality in GBRMP catchments, particularly given the significant link to water quality demonstrated above. As well as providing a direct indicator of the impact of agricultural and pastoral development on native vegetation, vegetation clearance can also act as an indicator of general ecosystem disturbance. Studies such as the ones reported in this chapter can assist decision-makers in assessing resource condition and addressing the broader requirements of natural resource policy devel- opment and planning. The evidence presented in this chapter clearly demonstrates a reduction in the area of native vegetation in GBRMP catchments. It also quantifies a substantial reduction in the area of native vegetation in the lower Herbert River catchment over the last 50 years. The trends observed on the Herbert River floodplain are not unique. For example, in the Johnstone River catchment, the area of coastal wetlands has decreased by approximately 60% since 1951 (Russell & Hales, 1996). The most sig- nificant losses have been of freshwater wetlands, particularly Melaleuca communi- ties. Melaleuca forests, notably those to the south of the Johnstone estuary, have been reduced by approximately 78%. There have also been significant reductions in other wetland categories, including a 64% reduction in palm- and pandanus-dominated wetlands and a 55% reduction in freshwater reed swamps. Freshwater wetlands to the north and west of the confluence of the North and South Johnstone Rivers have also almost entirely disappeared during this period. In contrast, the area of mangrove pat- terns has remained almost stable. Of the riparian forests assessed, 72% were in poor or very poor condition (Russell & Hales, 1996). Similar phenomena are manifest in the lower Burdekin, lower Pioneer, Fitzroy, Boyne, Mulgrave-Russell, Barron, Mossman, and Daintree River catchments (Congdon & Lukacs, 1995). In the Tully and Murray River catchments, less than 20% of coastal land systems suitable for agricultural production remains under native veg- etation (Tait, 1994). River catchments north of the Daintree River and in the Shoalwater and Styx catchments have, in comparison, remained largely undisturbed either as a result of their isolation or status as a national park (Johnson et al., 1997). © 2001 by CRC Press LLC [...]... sugarcane lands was damaging the world heritage status of the reef (Yellowlees, 1991 and references therein) More recently, the role of the grazing industry has received attention as a sediment source since grazing is the dominant land use upstream of the coastal floodplains of many of Queensland’s rivers Further, overgrazing in many areas has left soils bare and thus generated a large source of potential... justified and require further attention Given that agricultural industries in these areas are operating in the context of an ever-increasing © 2001 by CRC Press LLC Landcover and Water Quality in River Catchments 31 community expectation for the preservation of native vegetation (particularly riparian and wetland areas), conflict over the use of these resources is likely to grow in the future On the basis of. .. and other contaminants to the Great Barrier Reef Marine Park—a case study in the Herbert River catchment of North-East Queensland Marine Pollution Bulletin, in press Johnson, A.K.L., Ebert, S.P., & Murray, A.E 1997 Spatial and temporal distribution of wetland and riparian zones and opportunities for their management in catchments adjacent to the Great Barrier Reef Marine Park pp 82–101 in Haynes, D.,... Production: Meeting the Challenges Beyond 2000 C.A.B International, Wallingford, U.K., pp 2 43 –266 Yellowlees, D (ed) 1991 Land Use Patterns ad Nutrient Loading of the Great Barrier Reef Region James Cook University of North Queensland, Townsville © 2001 by CRC Press LLC 34 Oceanographic Processes of Coral Reefs FIGURE 1 Location of river catchments draining into the GBRMP FIGURE 2 Location of the Herbert... planning have been unsuccessful In response, policy, planning, and management reforms are required if remaining coastal ecosystems adjoining the GBRMP are to be protected or maintained However, their efficacy is likely to be substantively reduced in the absence of quantitative information on the ecological impact of landcover change In the meantime, it is of course incumbent on the individual landholder... as a land use is apparently the dominant source of sediment and nutrient in the lower Herbert, it is unclear whether this is coming from the cane paddocks themselves, or from the banks of the numerous man-made surface drains which dissect the cane-growing part of the catchment (Prove & Hicks, 1991) Recent data from Ripple Creek sub-catchment in the Lower Herbert suggest that plant cane paddocks and farm... Patterns of the Herbert River Floodplain Queensland Department of Primary Industries, Townsville Prove, B.G & Hicks, W.S 1991 Soil and nutrient movements from rural lands of North Queensland pp 67–76 in Yellowlees, D (ed) Land Use Patterns and Nutrient Loading of the Great Barrier Reef Region James Cook University of North Queensland, Townsville QDNR 1999a Landcover Change in Queensland 1991–1995 Queensland... the tasks facing policymakers and resource managers in catchments adjacent to the GBRMP, in terms of the way in which they manage riparian and wetland areas in the future, remains substantial The challenge facing government is to provide a stable environment in which locally relevant decision-making can occur and which is supported with appropriate and viable monitoring, cost-sharing, and regulatory... supported in part by funding from the Sugar Research and Development Corporation, the Land and Water Resources Research and Development Corporation, the CRC for Sustainable Sugar Production, and the CSIRO Divisions of Tropical Agriculture and Land and Water under the aegis of the CSIRO Coastal Zone Program The assistance of the many CSIRO technical staff members who contributed to the Herbert part of this... experience from the rest of the country, large areas of Southeast Asia, and most of Europe and the U.S.A., is that such a large proportion of these losses have occurred in the last 50 years and particularly in the last 20 years In Europe and the U.S.A., significant wetland and riparian zone restoration programs were operational 20 years ago, a time when large-scale losses were occurring in Australia Current . categories, including a 64% reduction in palm- and pandanus-dominated wetlands and a 55% reduction in freshwater reed swamps. Freshwater wetlands to the north and west of the confluence of the North and. 1. 73 0. 03 0.27 3. 62 0 .33 5 Ross 1 ,34 6 0.65 1.24 0.02 0 1 .37 2. 63 0.195 Haughton 3, 650 2. 93 0.72 10.81 0 0.51 12.04 0 .33 0 Don 3, 890 2 .39 1 .39 1.19 0 0.26 2.84 0.0 73 Proserpine 2,490 1.5 3. 56 10.27. to discriminate between the effects of intensive and non-intensive land uses on water quality, it was not possible to dis- criminate within these broad groupings. For the purposes of the present