A review of sediment and nutrient
concentration data from Australia for
use in catchment water quality models
Rebecca Bartley and William Speirs
1
eWater Technical Report
Review and summary of constituent concentration
data from Australia for use in catchment water
quality models
Rebecca Bartley1 and William Speirs
1
CSIRO, 120 Meiers Rd, Indooroopilly, Queensland 4068, Australia
October 2010
eWater Cooperative Research Centre Technical Report
i
Contact for more information:
Dr Rebecca Bartley
Catchment and Aquatic Systems Research Group
CSIRO Land and Water
120 Meiers Rd, Indooroopilly, QLD 4068
[email protected]
Please cite this report as:
Bartley, R. and Speirs, W. (2010) Review and summary of constituent concentration data from Australia for use in
catchment water quality models. eWater Cooperative Research Centre Technical Report
http://ewatercrc.com.au/reports/xyz.pdf
© eWater Cooperative Research Centre 2010
This report is copyright. It may be reproduced without permission for purposes of research, scientific
advancement, academic discussion, record-keeping, free distribution, educational use or other public benefit,
provided that any such reproduction acknowledges eWater CRC and the title and authors of the report. All
commercial rights are reserved.
Published online October 2010
ISBN: 978-1-921543-37-1
eWater CRC
Innovation Centre, University of Canberra
ACT 2601, Australia
Phone (02) 6201 5168
Fax
(02) 6201 5038
Email [email protected]
Web www.ewatercrc.com.au
eWater CRC is a cooperative joint venture whose work supports the ecologically and economically sustainable
use of Australia’s water and river systems. eWater CRC was established in 2005 as a successor to the CRCs for
Freshwater Ecology and Catchment Hydrology, under the Australian Government’s Cooperative Research
Centres
Program.
Executive Summary
Land use change is seen as the primary factor responsible for changes in sediment and
nutrient delivery to water bodies. Understanding how sediment and nutrient (or constituent)
concentrations vary with land use is critical to understanding the current and future impact of
land use change on aquatic ecosystems. Due to the large size of many catchments, and the
cost of water quality measurements, our understanding of the sources and rates of
constituent generation is largely dependent on the use of models. The subsequent
confidence in these models is reliant on access to quality input data.
This study collated runoff, concentration and constituent load data for Australian
catchments from published papers, reports and in some cases unpublished documents. We
collated water quality data from runoff events with a focus on areas that have a single or
majority of the contributing area under one land use. Based on a review of the literature
there are 5 broad methods used to calculate a mean event constituent concentration and a
discussion of the strengths and weaknesses of the various approaches is given.
Data were collated for total suspended sediments (TSS), total nitrogen (TN) and total
phosphorus (TP). Where possible, information on the dissolved forms of nutrients were also
collated. For each data point, information was included on the site location, land use type
and condition, contributing catchment area, runoff hydrology, lab analysis methods, number
of samples collected over the hydrograph and the constituent concentration calculation
method. The data were recorded in an excel data base and can also be viewed using
Google Earth.
A total of ~750 entries were recorded in the data base from 514 different geographical
sites covering 13 different land uses. Analysis of the data found that there was little
difference between data collected using “grab” sampling (over the hydrograph) versus autosampling (although the data set for this analysis was small). Analysis also showed that there
wasn’t a statistically significant difference (p<0.05) between data collected at plot and
catchment scales for the same land use. This is most likely due to differences in land
condition over-shadowing the effects of spatial scale. There was, however, a significant
difference in the concentration value for constituent samples collected from sites where
>90% of the catchment was represented by a single land use, compared to sites with <90%
of the upstream area represented by a single land use. This highlights the need for more
single land use constituent concentration and load data, preferably over a range of spatial
scales. Based on these analyses the final synthesis data is presented using the whole data
set, as well as for sites where a single land use is represented in >90% of the contributing
catchment area.
Overall, the land uses with the highest median TSS concentrations were Mining (~50,000
mg/l), Horticulture (~3000 mg/l), Cotton (~600 mg/l), Grazing on native pastures (~300 mg/l),
and Bananas (~200 mg/l). The highest concentrations of TN are from Horticulture (~32,000
ug/l), Cotton (~6,500 ug/l), Bananas (~2,700 ug/l), Grazing on modified pastures (~2,200
ug/l) and Sugar (~1,700 ug/l). For TP it is from Forestry (~5,800 ug/l), Horticulture (~1,500
ug/l), Bananas (~1,400 ug/l), Grazing on modified pastures (~400 ug/l) and Grazing on
native pastures (~300 ug/l). For the dissolved nutrient fractions, the Sugar land use has the
highest concentrations of dissolved inorganic nitrogen (DIN), dissolved organic nitrogen
(DON) and dissolved organic phosphorus (DOP), and the Urban land use had the highest
concentrations of dissolved inorganic phosphorus (DIP).
This report also demonstrates that all water quality data is extremely variable, even if
collected using the same methods, from the same land use in a similar region. There are a
range of factors that produce a constituent value (such as soil type, land condition, rainfall
pattern etc), and modellers need to be aware of these issues before applying the data. This
document discusses some of the options and issues associated with using water quality data
in catchment models, and modellers may want to use different data for different tasks. The
data base developed in this study may be used as a portal to access publications with water
quality data that may be suitable for a range of modelling applications.
3
1 Table of Contents
EXECUTIVE SUMMARY ................................................................................................................................. 3
TABLE OF FIGURES ....................................................................................................................................... 5
TABLE OF TABLES ......................................................................................................................................... 6
2
INTRODUCTION ................................................................................................................................... 7
3
DEFINING MEAN CONCENTRATION ...................................................................................................... 9
4
METHODS: DATA BASE FIELDS AND CLASSIFYING THE DATA ............................................................... 10
4.1
LOCATION INFORMATION ........................................................................................................................... 11
4.2
LAND USE TYPE AND CONDITION .................................................................................................................. 11
4.2.1 Comparison of the same land use in different condition ................................................................ 12
4.3
CONSTITUENTS MEASURED ......................................................................................................................... 13
4.4
SAMPLE COLLECTION METHODS ................................................................................................................... 13
4.5
LAB ANALYSIS PROCEDURE.......................................................................................................................... 14
4.6
CONSTITUENT CONCENTRATION CALCULATION METHODS ................................................................................. 14
4.7
NUMBER OF SAMPLES AND LOCATION ON THE HYDROGRAPH............................................................................. 15
4.8
DETERMINING IF THERE ARE DIFFERENCES IN CONSTITUENT CONCENTRATION FOR DIFFERENT PLOT SIZES AND
PROPORTIONS OF A CONTRIBUTING LAND USE ............................................................................................................. 15
4.9
LINKING THE DATA BASE TO GOOGLE EARTH .................................................................................................. 17
5
RESULTS ............................................................................................................................................ 18
5.1
SUMMARY OF DATA SETS USED IN DATA BASE ................................................................................................. 18
5.2
PRELIMINARY ANALYSIS TO DETERMINE DATA FOR INCLUSION IN FINAL RESULTS .................................................... 20
5.2.1 Comparison of constituent data collected using different methods ............................................... 20
5.2.2 Variation of data at different spatial scales ................................................................................... 22
5.2.3 Does the constituent concentration vary when the proportion of upstream land use changes? ... 24
5.3
SUMMARY OF WATER QUALITY DATA FOR DIFFERENT LAND USES ........................................................................ 27
5.4
COMPARISON OF EVENT AND BASE FLOW DATA FOR THE DOMINANT LAND USES.................................................... 33
6
DISCUSSION ....................................................................................................................................... 35
6.1
SUMMARY OF FINDINGS FROM THIS STUDY .................................................................................................... 35
6.2
ISSUES TO BE AWARE OF WHEN USING DATA IN CATCHMENT MODELS .................................................................. 36
6.2.1 Uncertainty in data collection, handling, processing and analysis ................................................. 36
6.2.2 Using ‘event mean concentration’ (EMC) versus ‘mean event concentration’ (MEC) .................... 37
6.2.3 Selecting appropriate mean concentration values for load calculations ........................................ 39
6.2.4 Differentiating between Event (EMC) and Base flow or dry weather (DWC) values....................... 40
7
CONCLUSIONS AND AREAS OF FURTHER RESEARCH............................................................................ 43
8
ACKNOWLEDGEMENTS ...................................................................................................................... 45
9
REFERENCES ...................................................................................................................................... 46
10
APPENDICES ...................................................................................................................................... 51
10.1
10.2
10.3
APPENDIX A: LIST OF GOVERNMENT WATER QUALITY DATA SITES ...................................................................... 51
APPENDIX B: REFERENCES USED IN THE WATER QUALITY DATA BASE.................................................................... 52
APPENDIX C: RECOMMENDATIONS FOR USING THE EXCEL DATA BASE CREATED FOR THIS STUDY ............................... 58
4
Table of Figures
Figure 1: Example of Grazing (on native pastures) for a site with low cover in poor condition (left) and
the same site 6 years later with high cover in good condition (right) (Source: Bartley et al., 2010a).12
Figure 2: (A) Comparison of annual average TSS EMC values for a 14 km2 grazed site in the Burdekin
catchment that has similar ground cover values of 36%, 37% and 38% in 2006, 2003 and 2005,
respectively (data from Bartley et al., 2010b) and (B) Comparison of annual average EMC values on
a 1.2 ha grazed hillslope for different cover levels (data sourced from Bartley et al., 2010a) ............ 13
Figure 3: Location of water quality data sites reported and entered into the data base are indicated by the
red markers ............................................................................................................................................... 19
Figure 4: An example of the data available for the the Bowen River catchment in Queensland for 2004 . 19
Figure 5: A description of the statistics represented by the box and whisker plots in subsequent sections 20
Figure 6. Discharge against sample concentrations of total nitrogen, total phosphorus (A) and total
suspended sediment (B) at Upper Mossman (Source: McJannet et al., 2005). Note nutrient data
units are mg/l in this specific graph. ........................................................................................................ 22
Figure 7: Comparison of data from the same land use (Grazing on native pastures) in the same catchment
(Burdekin River Basin) showing the variation in TSS values between plot scale studies (<1 km2) and
large catchment areas (~7,000 km2). Data sourced from Hawdon et al., (2008), Bartley et al.,
(2010a), Bainbridge et al., (2007), and Bartley et al., (2007a)................................................................ 23
Figure 8: The difference in TSS concentration values for sites at plot, medium-catchment and large
catchment scales for the (A) Forest and (B) Grazing (on native pastures) land uses. Note for both
sets of data the land use occupied >90% of the catchment. ................................................................... 23
Figure 9: The difference in (A) TN concentration and (B) TP values for sites at plot (<1 km2), subcatchment (1-100 km2) and large catchment scales (>100 km2) for the Sugar land use. Note that the
proportion of the catchment represented by sugar cane in each category varies. For the <1 km2 data
sugar is >95%, for 1-100 km2 it is between 25-100% and for the >100 km2 it is between 31-72%.
Note that the TSS plot for sugar has a very similar pattern as TP. ...................................................... 24
Figure 10: (A) TSS (B) TN and (C) TP for Grazing land use according to the % of that land use
represented above the water quality sampling point ............................................................................. 25
Figure 11: (A) TSS (B) TN and (C) TP for Forest land use according to the % of that land use
represented above the water quality sampling point ............................................................................. 25
Figure 12: (A) TSS (B) TN and (C) TP for the Sugar land use according to the % of that land use
represented above the water quality sampling point. Note that the % threshold of land that was
used for TSS was slightly different to that for TN and TP due to limited data, and the land use %
representations are different to the grazing and forest data sets. ......................................................... 26
Figure 13: Constituent concentrations for (A) Event TSS (mg/l) (B) Event TN (ug/l) and (C) Event TP
(ug/l) for land uses where n>3. Note the log (natural) y axis and Dairy has been presented with
Grazing on modified pastures .................................................................................................................. 28
Figure 14: Event (A) DIN (B) DON (C) DIP and (D) DOP (ug/l) concentrations for land uses (where n>3).
Note that Bananas and cotton have been combined with Horticulture. ............................................... 29
Figure 15: Comparison of event (EMC) and base flow (DWC) concentrations for (A) TSS (mg/l) for
Forest, Grazing (on native pastures) and Urban land uses, (B) TN (ug/l) for Forest, Forestry,
Grazing (on native pastures) and Urban land uses and (C) TP (ug/l) concentrations for Forest,
Forestry, Grazing (on native pastures) and Urban land uses................................................................ 34
Figure 16: Comparison of mean measured annual TSS concentrations and annual TSS EMC values for
Flume 1 and 2 ............................................................................................................................................ 37
Figure 17: (A) Event 1 on 16th and 17th of December 2001 and (B) Event 2 on 15th February 2002 showing
the high turbidity and sediment concentrations early in the event ....................................................... 40
Figure 18. Comparison of concentrations of TN, TP, turbidity and TSS for baseflow and event conditions
on the Mossman River, with event condition further divided into early (December 2003 to January
2004), mid (February to March 2004) and later (April to May 2004) stages of the wet season.
Baseflow concentrations are from EPA water quality data from 34 sampling campaigns for the
months May until November for the period 1994 to 1999. Baseflow data for TSS was only recorded
on three occasions, each being below the detection limit (< 2mg/L) and has been recorded on the box
plot as a single point at 2mg/L (Source: McJannet et al., 2005) ............................................................ 42
5
Table of Tables
Table 1: Summary of land uses and the number of sites included in the data base...................................... 11
Table 2: Summary of land uses and the number of sites included in the data base (including plot/hillslope
studies) ....................................................................................................................................................... 18
Table 3: Comparison of measured ‘median’ constituent concentration values for sites with the same land
use and similar climate, but using different sampling methods. Note: (i) these are measured
concentration data and not EMC values and (ii) the use of median values results in subtle
differences in nutrient balance calculations reported in the table ........................................................ 21
Table 4: Summary table showing when there was a significant difference (p<0.05) between sites that had
>90% of the catchment under a single land use compared with 70-90% of the area. ........................ 24
Table 5: Summary of data for TSS, TN and TP for the 10 major land uses identified in this study. Note
that n represents a data point from a single geographical site and in some cases this represents
individual event data, in other cases it represents annual or multi event averages. Data were only
presented when n ≥ 3................................................................................................................................. 30
Table 6: Summary of data for DIN, DON, DIP and DOP for the 10 major land uses identified in this
study for event (EMC) conditions only. There was insufficient data to present baseflow (DWC)
values. Note that n represents a data point from a single geographical site and in some cases this
represents individual event data, in other cases it represents annual or multi event averages. Data
were only presented when n ≥ 3 ............................................................................................................... 31
Table 7: Summary data for TSS, TN and TP for the 10 major land uses identified in this study where the
land use upstream of the sampling point is >90% under a single land use. Data is for event (EMC)
conditions only. There was insufficient data to present baseflow (DWC) values. Note that n
represents a data point and in some cases this represents individual event data, in other cases it
represents annual or multi event averages.............................................................................................. 32
Table 8: Annual median and mean measured TSS concentrations and the unit area sediment load and
annual EMC values for 6 years of monitoring data at two hillslope flume sites in the Burdekin
catchment (Source of data, Bartley et al., 2010a) ................................................................................... 38
Table 9: Sediment load (tonnes) calculated using the same hydrology with 4 different methods for
estimating the mean concentration for the 2001/02 data set ................................................................. 40
Table 10: Difference in constituent load values for the Upper Mossman River in north Queenland using
different methods for the period 1/12/03-30/06/04 (data sourced from McJannet and Fitch, 2005) . 41
6
2 Introduction
Understanding the generation and movement of excess sediments and nutrients through
catchments, and their subsequent impact on downstream aquatic ecosystems, is the ‘holy
grail’ for a large number of research agencies around the world. It is well documented that
phosphorus, when in excess, can limit the biological productivity of freshwater ecosystems
(e.g. Davis and Koop, 2006; Heathwaite, 2003) and that excess nitrogen, in particular nitrate,
appears to be detrimental for marine systems (e.g. Fabricius, 2005; Fabricius et al., 2005;
Smith et al., 2006). Excessive erosion and delivery of fine and coarse sediment to rivers and
coastal systems is problematic in many parts of the world, and the processes and impacts
are well documented (e.g. Meade, 1988; Prosser et al., 2001b; Rutherfurd, 2000; Walling,
2006). Constituent (e.g. sediment, nitrogen and phosphorus) generation, cycling, movement
and storage processes are complex and vary with watershed conditions such as land use,
rainfall and geology (Young et al., 1996). Land use change, however, is seen as the key
factor responsible for changes in sediment and nutrient delivery to downstream water bodies
(Harris, 2001). Therefore, understanding how constituent concentration varies with land use
is critical to understanding the current and future impact of land use change on downstream
water bodies.
Due to the large size of many catchments, and the cost of water quality measurements, our
understanding of the sources and rates of constituent generation at the catchment scale is
largely reliant on the use of models. Most models require data to help formulate the initial
conceptual structure, calibrate the model and test the model outputs (Bloschl and Sivapalan,
1995). In Australia, the most recent modelling platform for evaluating catchment scale
constituent losses is Source Catchments (formerly known as WaterCAST). This is a lumped,
semi-distributed, conceptual catchment modelling framework that operates on daily time
step. It allows for the construction of models by selecting and linking component models
from a range of options (Argent et al., 2008). Source Catchments conceptualises a range of
catchment processes using sub-catchments which are composed of Functional Units (FUs).
Each FU is characterized by similar pollutant generation processes, which are typically
determined using an event mean concentration (EMC), and/or dry weather concentration
(DWC) approach (e.g. Chiew and Scanlan, 2002). Concentration data are separated into
EMC and DWC values because concentrations can vary considerably between event and
baseflow conditions. Data representing the mean concentrations of sediment, nitrogen and
phosphorus are also required by other models used in Australia (e.g. MUSIC, see Fletcher et
al., 2004; and SedNet, see Wilkinson et al., 2004).
The new generation of catchment models (e.g. Source Catchments) incorporate their own
rainfall/runoff routines to simulate flows, yet to calculate constituent loads, estimates of
constituent concentrations are required. At present there is no single, peer reviewed
document that provides a summary of constituent concentration data for the whole of
Australia. As a result, models that are applied in areas that are relatively data poor end up
using data from elsewhere. This data can sometimes be well outside their geographic and
climatic region. Having knowledge of data collected from a range of sites and land uses
around Australia will provide model builders with a more robust data set from which to
parameterise their models. “Models are our attempt to encapsulate our understanding of the
real world. They are not the real world, and without data they are simply imagination and
computer games” (Silberstein, 2006). Hence, this report will provide a summary of suitable
water quality data for use in catchment models.
The reason that it is difficult to find suitable water quality data for use in catchment models is
that water quality data is also expensive to collect. Water quality monitoring is the systematic
collection and analysis of samples with the aim of providing information and knowledge
7
about a body of water and the factors affecting it. Monitoring of water quality (both surface
and sub-surface) is undertaken for a range of reasons including protecting public health and
aquatic ecosystems, environmental reporting, licence compliance and research. Water
quality monitoring is carried out by a wide range of organisations from local, state and
federal government as well as the private sector and community groups. It has been
estimated that Australia spends $142 - $168 m each year on water quality monitoring
(http://www.anra.gov.au/topics/water/pubs/national/water_quality.html).
Water
quality
monitoring is largely funded by public money, and therefore these data should be shared
where-ever possible. A lot of the water quality data collected is available on-line via the State
Government water quality web sites (see Appendix A). Some local councils also provide
data
and
reports
on
their
web
sites
(for
example,
see
http://www.hornsby.nsw.gov.au/environment/index.cfm?NavigationID=1040), and in some
states there are regional groups that monitor water quality, for example the South East
Queensland Ecosystem Health Monitoring Program (EHMP), Melbourne Water and the
Fitzroy Basin Association. At a federal level, the Australian Water Data Infrastructure Project
(AWDIP) was established under the national component of the Natural Heritage Trust to
facilitate
national
assessments
of
Australia’s
water
resources
(http://www.daff.gov.au/brs/water-sciences/ground-surface/awdi-project). The framework will
enable on-line access to data sets via a network of distributed jurisdictional databases. This
technology will be available to emerging water information systems as they become
operational (for example the Australian Water Resources Information System proposed by
the National Water Commission and being implemented by the Bureau of Meteorology). A
description of the different federal water data management projects can be found at
http://www.daff.gov.au/__data/assets/pdf_file/0005/382523/WRON_AWDIrelationship.pdf.
The data housed in government data centres provide important baseline information
regarding the health of our streams, rivers and in some cases groundwater. They also
provide a good spatial representation of water quality in many areas, and are often used to
check whether different regions are meeting the ANZECC water quality guidelines (e.g.
Moss et al., 2005). Unfortunately, however, the majority of these data are collected during
dry weather under base flow conditions without associated flow data. At present there is no
central accessible data base documenting data that has been collected in ‘event’ conditions.
The majority of water quality studies fall into two broad categories:
1. Base flow data collection which is used for routine monitoring and compliance checks
and sometimes for ecological research; or
2. Event based sampling which is used for estimating soil and nutrient erosion rates and
constituent load exports.
Most event data is collected over short duration and has not (until more recently) been
considered as part of routine monitoring program. Event monitoring has historically been
carried out by research groups within various agencies, and the associated data is written
into journal papers and reports.
This study focused on collating runoff, concentration and constituent load data for Australian
catchments. Data was taken primarily from published papers, but some data was taken from
reports and unpublished documents. This document builds on the work of Marston et al.,
(1995), Young et al., (1996), Fletcher et al., (2004) and Brodie and Mitchell, (2005). The aim
of this study was to collate water quality data that was collected under event conditions with
a focus on data collected from areas that have a single or majority of the contributing area
under one land use. Data from event conditions is considered important as the large events
contribute the majority of the sediment or nutrient load to downstream water bodies (Furnas,
2003). Dry weather or base flow concentration data were generally only reported when it
was measured at the same site as the event data. Data measured to describe the
processing of nutrients were not included in this study if they were only collected during base
flow over a short time period (i.e. a few days) (e.g. Hadwen et al., 2009). Similarly, data were
8
not included if they integrated a range of land uses or if they were largely measured in
estuarine rather than freshwater conditions (e.g. Eyre and Balls, 1999). Salinity, herbicides
and pesticides were not included in this review. Although a thorough review of the literature
was undertaken, it is acknowledged that there is likely to be additional data housed in
government and company reports that was not captured in this data base. Therefore this
data base should not be considered as the only resource for people looking for water quality
data.
The purpose of this document is not necessarily to provide a ‘final’ generic look up table of
values for modellers to apply to studies around Australia, although a summary table is
provided in Section 5.3. The main purpose of this report is to let the Australian hydrological
modelling community know that the data base exists, and should be used as a portal to
access publications with water quality data that may be suitable for a range of modelling
applications. There are a range of factors that produce a constituent value, and modellers
need to be aware of these issues before applying the data. It is also important to point out
that the spatial scale that constituent data is applied (i.e. the classification method for
defining a functional unit, FU) will vary depending on the specific model set up. Hence,
modellers may want to use different data for different tasks. The aim of this document is to:
• Summarise the water quality data collated for the dominant Australian land uses;
• Demonstrate that water quality data can be highly variable even if collected using the
same methods, from the same land use in similar region; and to
• Discuss the options and issues associated with using water quality data in catchment
models.
This report starts by outlining the different approaches used for measuring mean constituent
concentration (Section 3) and then outlines the methods used to classify and analyse the
collated data (Section 4). Section 5.1 provides an inventory of the data collated, and then
Section 5.2 provides a justification for the interpretation of water quality data according to the
different data collection methods, spatial scales and proportion of the catchment under a
single land use. This is then followed by a summary of the data in Sections 5.3 and 5.4.
Section 6 discusses some of the issues that need to be taken into consideration when using
water quality data in models. These include issues such as uncertainty, and the range of
constituent values possible due to the variability in land and flow condition. Section 7
provides a summary of the report and outlines the areas for further research.
3 Defining mean concentration
Defining what represents an event or quick flow verses base or slow flow can be conducted
using a range of methods including hydrological analysis (e.g. Grayson et al., 1996; Nathan
and McMahon, 1990) through to field observation. In this section we are not dealing directly
with how a hydrological event is defined, as this is dealt with within in the Source
Catchments model directly (see eWater-CRC, 2010). This section discusses how a mean
concentration is defined for an event flow.
A number of different uses of the term EMC have evolved over the last few decades and the
terms EMC and mean event concentration are often used interchangeably despite being
derived differently. The event mean concentration (EMC) is a flow-weighted average of
constituent concentration and is reported in units of mg/L (Charbeneau and Barrett, 1998).
Formally, EMC is defined as total constituent mass, M, discharged during an event divided
by total volume (V) of discharge during an event (Huber, 1993) (Equation 1).
Equation 1
9
The term EMC appeared to originate in urban and road stormwater runoff research,
however, it is now widely used for land use specific modelling of water quality for a range of
land use types (e.g. Chiew and Scanlan, 2002; Collins and Anthony, 2008; Kim et al., 2004).
Although EMC is often used in association with the Source Catchments model, there are 5
methods (found in the literature) used to calculate a mean event concentration. The methods
include:
(1) Dividing constituent load (t) by flow volume (ML) (according to Equation 1) where the
constituent load may be calculated using a range of methods (see Johnes, 2007;
Letcher et al., 1999; Marsh and Waters, 2009b for a discussion of different
approaches). All subsequent uses of the term EMC in this report refer to this
method.
(2) Using flow duration curves to differentiate between ‘event’ and ‘dry’ whether
conditions for which the concentration data samples were collected. For example,
Chiew and Scanlan (2002) defined ‘event’ samples as those collected during flow
events that are exceeded 5% of the time, and base flow or dry weather samples are
exceeded less than 20% of the time.
(3) Using runoff data to identify flow events, but it is not directly reporting the flow values.
In such cases, average measured concentrations are presented (e.g. Nelson et al.,
2006). This approach appears to be more common in studies with a focus on
eutrophication or ecological processes.
(4) Collecting concentration data during events without information on the specific flow
conditions at the time of sampling. This approach is often used in areas with poor
gauging and/or when gauges are not located at sites with a single land use (e.g.
Bainbridge et al., 2009).
(5) Converting unit area yield data (e.g. kg/ha/yr) to a concentration (mg/l) using
associated published flow data (e.g. Marston et al., 1995). Information on the sample
hydrograph is often not explicitly provided, and therefore it is difficult to ascertain if
the result is representative of ‘event’ or ‘base’ flow. Therefore the results are
considered as an integrated concentration.
To summarise, method (1) is considered as an event mean concentration (EMC) value,
whereas methods (2) to (5) are mean event concentrations (MEC) that directly use
measured concentration data. The data base developed, and the subsequent results
presented in this document, have utilised data collected and analysed using each of the
methods outlined above. Where ever possible, a description of the various techniques and
methods were provided in the data base. When concentration data is reported, some studies
report arithmetic ‘mean’ values and other studies report ‘median’ values. Median rather than
mean values are often used to reduce the influence of large outliers in the data. For the
purpose of this study the mean and median values are lumped to represent the ‘average’
water quality concentration coming from a given land use. The variations in water quality
observed at the National scale are considered to be larger than the variation resulting from
different statistical approaches.
4 Methods: data base fields and classifying the data
This section describes what data were collected and what considerations were made when
collating and cataloguing the data. In some cases the raw data was available, but in most
cases the interpreted data published in tables and figures within the relevant documents
were used to derive the water quality data for each site.
10
4.1
Location information
For each water quality entry, information on the following site characteristics were collected:
•
Basin name (e.g. Torrens River)
•
Basin ID using the Australian Water Resources Council basin numbers (e.g. 504)
as represented in the NLWRA (2000)
•
State or Territory
•
Gauge Number where applicable (e.g. A5040583)
•
Latitude and Longitude
4.2
Land use type and condition
Application of mean concentration values generally requires ‘single-land-use’ data due to the
high variability in constituent concentrations demonstrated for sites under different land uses
and conditions (Young et al., 1996). The breakdown and number of land use categories
used varies depending on the study, and ranges from 3 broad land use categories in the
study by Bramley and Roth (2002) to 11 categories in the study by Fletcher et al (2004). The
use of ~5 land use categories is typical for the majority of regional or catchment studies (e.g.
Bainbridge et al., 2009; Chiew and Scanlan, 2002; Cogle et al., 2000; Rohde et al., 2008)
and these increasingly match with the State or Federal mapping categories. This study
collated data for the whole of Australia and therefore the data were classified into 13
categories using the ALUM (Australian Land Use Mapping) classifications given in Table 1. It
is important to note that some land uses can fall into a number of categories. For example,
Dairy may be considered as 3.2.0 (Grazing modified pasture) or 4.2.0 (irrigated modified
pasture) or as 5.2.1 (Intensive animal production, dairy).
Table 1: Summary of land uses and the number of sites included in the data base
Land use
Bananas
Cotton
Dryland Cropping
Forest
Forestry
Grazing Modified Pastures (including dairy)
Grazing Native Pastures
Horticulture
Military
Mining
Mixed
Sugar Cane
Urban / Industry
ALUM classification
3.4.0
3.3.6
3.3.0
1.1.0
2.2.0
3.20, 4.2.0 or 5.2.1
2.1.0
4.4.0
1.3.1
5.8.3
3.3.5
5.4.2
For each data entry the following information were recorded when available:
• Primary land use (name and number) using the classification methods derived by the
Bureau of Rural Sciences (BRS, 2006). Note that no distinction was made between
irrigated and non-irrigated sugar cane;
• Australian
Land
Use
Mapping
(ALUM)
classification
(http://adl.brs.gov.au/mapserv/landuse/pdf_files/Web_LandUseataGlance.pdf) (Table
1);
• Secondary land use and % coverage;
• Vegetation type;
• Land condition;
• Soil type;
• Contributing catchment area.
11
Data from plot or hillslope scale studies were also included in the data base as these were
often the only data available for some land uses. Vegetation and soil types were given
generic descriptions based on the information provided in the references. Where precise
information was available that allowed classification according to accepted standards (e.g.
the Australian Soils Classification system), this information was recorded. No attempt was
made to go beyond the descriptions provided.
It is also important to note that some studies have identified sources or processes
responsible for constituent generation ‘within’ a given land use. For example, Fletcher et al.,
(2004) and Duncan (1999) have divided the urban landscape up into roads, roofs etc.
Similarly, Visser et al (2007) determined that headlands and drains were the dominant
source of sediment within sugar cane landscapes and Bartley et al., (2007b) identified
channels as the major source of sediment from grazed systems. In this study, however, no
attempt has been made to differentiate between generation processes occurring within a
landuse.
4.2.1 Comparison of the same land use in different condition
It is acknowledged that both land use, and land condition, can influence the quality of water
coming off a given site. The same land use type can yield very different constituent
concentrations when in poor condition, compared when it is good condition (Bartley et al.,
2010a; see Figure 1 and 2). This is due to differences in rainfall, runoff, vegetation, soil type,
slope and management practices. These factors collectively create different land condition.
To demonstrate how variable land condition can be in both space and time, data collected at
a range of scales from a 14 km2 grazed catchment in the Burdekin basin is presented in
Figure 2. These data demonstrate how the annual catchment total suspended sediment
(TSS) EMC value (derived from sediment load and runoff data) can vary by more than 100%
when the ground cover changes by an average of just 1% (see Figure 2A). This difference is
due to large differences in the amount and intensity of rainfall between years with similar
cover. On the other hand, it is also possible to obtain quite different TSS EMC values at a
site when the ground-cover changes significantly (by ~20%) through time on a given
hillslope (Figure 2B). Based on this analysis, knowledge of the soil and vegetation condition
were considered to have an important influence on the constituent water quality values
measured, and were captured in the data base where-ever available. Knowledge of the
rainfall pattern and intensity contributing to a given constituent concentration is also
important but was not explicitly recorded in this study.
Figure 1: Example of Grazing (on native pastures) for a site with low cover in poor condition (left) and the
same site 6 years later with high cover in good condition (right) (Source: Bartley et al., 2010a).
12
800
Rainfall
3500
700
3000
600
2500
500
2000
400
1500
300
1000
200
500
100
Total Supended sediment EMC
(mg/l)
EMC
4000
Rainfall (mm)
Total Supsended Sediment EMC (mg/l)
900
4500
36
37
800
700
600
500
400
300
200
100
0
35
0
0
900
38
55
75
average % cover on grazed hillslope
average % ground cover in catchment
(B)
(A)
Figure 2: (A) Comparison of annual average TSS EMC values for a 14 km2 grazed site in the Burdekin
catchment that has similar ground cover values of 36%, 37% and 38% in 2006, 2003 and 2005,
respectively (data from Bartley et al., 2010b) and (B) Comparison of annual average EMC values on a 1.2
ha grazed hillslope for different cover levels (data sourced from Bartley et al., 2010a)
4.3
Constituents measured
In this study, the three constituents of interest were suspended sediment, nitrogen and
phosphorus. Sediment concentration data were recorded as mg/l and nitrogen and
phosphorus ug/l. The total suspended sediment (TSS) data were assumed to represent
sediment <63um. Bedload was not explicitly recorded in the data base. Where data were
provided as kg/ha/yr, these data were converted to mg/l when flow and catchment area
information were provided.
As different forms of nutrients (e.g. dissolved versus particulate or inorganic versus organic)
are known to have different effects on food webs and freshwater and marine ecology
(Fabricius, 2005; Murray and Parslow, 1999), information on the form of nitrogen and
phosphorus was recorded where possible. The forms of N and P recorded were Dissolved
Inorganic Nitrogen (DIN), Dissolved Organic Nitrogen (DON), Dissolved Inorganic
Phosphorus (DIP) and Dissolved Organic Phosphorus (DOP). For some studies, only TN
and one form of dissolved nutrient data were provided, and in these cases the dissolved
concentrations were calculated by difference.
4.4
Sample collection methods
A monitoring program, whether for routine condition reporting, or for a research project,
needs to follow accepted standards where possible. In Australia, there are water quality
standards (AS/NZS 5667) and QA/QC guidelines that are outlined in the Australian and New
Zealand Guidelines for Fresh and Marine Water Quality (ANZECC, 2000), and the Australian
Guidelines for Water Quality Monitoring and Reporting (ANZECC and ARMCANZ, 2000b). In
this study, it is assumed that all field sample collection, handling, storage and processing
has met established protocols. Preservation and storage requirements were not explicitly
recorded, although quality assurance and quality control (QA/QC) issues and procedures
were assumed to have occurred.
It is acknowledged that the methods and approaches for collecting water quality samples
(e.g. pump sampler versus manual collection) can result in significantly different values for
constituent concentration (Facchi et al., 2007). Hence, when information on the data
collection method was provided, this information was recorded in the data base. This
included information on whether the sample was collected from:
• a pumping sampler, and the location of that sampler (e.g. edge or stream centre)
13
• grab samples and the location of the sample collection (e.g. bridge, stream edge etc)
• a boat
• or other device
In this study it was not possible to do a thorough investigation into the sources of variability
or error introduced at each stage of the water quality data collection process (as done in
Harmel et al., 2006b). There was, however, an opportunity to evaluate if the sample
collection method, and more specifically if knowledge of the flow or hydrograph at the time of
sample collection, influenced the event mean water quality. This was considered important
to evaluate as often areas that are suitable for obtaining single land use water quality data
are not located adjacent to hydrological gauging stations.
A simple comparison of three data sets from the same land use (forest) in similar geographic
and climatic region (wet tropics of North Queensland) is presented. In this comparison the
first study from the Tully catchment, data were collected manually (i.e. not using an
automatic sampler) when the rivers were in ‘event’ mode. No hydrograph data were
presented with the concentration data (Bainbridge et al., 2009). Data from the second
catchment were also collected manually, however, information about the hydrograph and
more importantly, where these samples were collected on the hydrograph (rising or falling
stages) were provided (Rohde et al., 2008). In the third catchment, 173 samples were
collected using an ISCO automatic sampler and stage recorder. Turbidity was also recorded
continuously and used to predict continuous TSS, TN and TP loads (McJannet et al., 2005).
4.5
Lab analysis procedure
It is acknowledged that the lab analysis method used can have profound implications on the
results of water quality analysis and thus impact on the analysis of historical water quality
trends. Therefore, where information regarding the lab methods used were available (e.g.
whether nitrogen is analysed using Kjeldahl digestion or persulfate autoclave digestion), this
information was recorded in the data base. It was also recorded, where possible, if the lab
was accredited with the National Association of Testing Authorities (NATA). There are
standard analytical methods for analysing samples from the water column, sediments and
biota and many of these have been cited in the NWQMS monitoring guidelines
(http://www.environment.gov.au/water/policy-programs/nwqms/). The primary reference for
methods of determining the concentrations of the various forms of nitrogen is the Standard
Methods for the Examination of Water and Wastewater 20th edition (APHA, 1998).
4.6
Constituent concentration calculation methods
There are five approaches used in literature to calculate and report mean event
concentration. These were described in more detail in Section 3, and are summarised below:
1. Dividing constituent load divided by the flow period;
2. Using flow duration curves to identify a consistent flow value to represent ‘event’ and
‘dry’ (or base) flow conditions;
3. Using flow equipment to define an event, but not reporting flow values;
4. Collecting concentration data without knowledge of the specific flow conditions at the
time of sample collection;
5. Converting unit area losses to concentrations by dividing by flow.
It is important note that mean event concentration (MEC), which is the average constituent
value measured during an event (represented by methods 2,3,4 and 5 above), is not the
same as event mean concentration (EMC), which is the concentration derived when dividing
total load by flow. Method 1 is the only method that formally calculates an EMC.
14
Each of the entries in the data base were given one of these classification values. This
approach meant that it was possible to include more data from around Australia, whilst
providing users with the opportunity to exclude data if they have reason to believe a given
approach distorts the final mean values.
4.7
Number of samples and location on the hydrograph
The number and location of samples collected during an event hydrograph can influence the
calculated load and thus the mean concentration calculated (e.g. Leecaster et al., 2002).
Load estimation using sample sizes less than 8 rarely predict a load within 50% of the true
mean (at 95% confidence level), and the best performing sampling methods are those with
samples on both the rise and fall of the event (Marsh and Waters, 2009b). Observational
errors associated with nutrient load estimates based on poor data can lead to a high degree
of uncertainty in modelling and nutrient budgeting studies (Johnes, 2007). It is also
acknowledged that the relationship between stream-flow and concentration can follow a
variety of non-linear forms which have important process implications (Nistor and Church,
2005; Sansalone and Cristina, 2004) and as a result, the large events may carry the highest
loads, but may not have the highest concentrations. Hence, it is important in any sampling
campaign that the whole hydrograph, including the first or initial events in the rainy season,
are collected to adequately represent average concentrations from a given area.
Therefore, where possible, data were classified according to the following:
1. More than 8 samples were collected evenly over the event hydrograph;
2. More than 8 samples were collected on rising or falling limbs of the event
hydrograph;
3. Less than 8 samples were collected, although they were measured evenly across the
event hydrograph, and/or at least the peak of the event was captured;
4. More than 8 samples were collected, however, this was dominated by low flow
conditions, and did not adequately capture the event hydrograph;
5. Less than 8 samples were collected and did not capture the whole hydrograph;
6. No hydrograph information was supplied with the concentration or load data;
7. Continuous turbidity was used as a surrogate for suspended sediment. Although the
TSS-turbidity relationship may change within and between events for a given site
(Sun et al., 2001), this is considered as the most accurate way of calculating the
‘true’ load (Walling and Webb, 1981).
It is also important to have information on the size of the event with respect to the long term
flow record (i.e. flow duration curves and/or exceedence probabilities). Therefore, where
possible, information was recorded on whether:
1. Flow data is provided and the location of the sampled event with respect to the long
term flow record is given;
2. Flow data provided, but no historical context given;
3. No flow data were provided.
4.8
Determining if there are differences in constituent concentration for
different plot sizes and proportions of a contributing land use
In the Source Catchments model, it is common to configure and characterise the Functional
Units (FU’s) within a model according to land use (e.g. Searle and Ellis, 2009). The model
set up is very flexible and therefore the FU size can vary from the size of plot (~< 1 km2) to
sub-catchment (1-100 km2) or larger, depending on the quality and size of the initial data
inputs (e.g. Digital Elevation Model). Due to the variable sizes of the FU’s it was important to
evaluate:
15
(1) If there is a difference between water quality data measured at plot and catchment
scales; and
(2) What the minimum % catchment area a particular land use should occupy before the
water quality data can be can considered to be dominated by the activities of that
land use.
It is well-established that sediment loss from small (i.e. 1 m2) plots will yield more sediment
than at the hillslope scale (e.g. 100 m2) due to sediment storage within the hillslope catena
(Ludwig et al., 2007; Parsons et al., 2006). Phosphorus concentrations also decrease with
increased plot scale (Sharpley and Kleinman, 2003) although there is less evidence for a
similar decline with nitrogen due to the pathways (i.e. surface and sub-surface) for dissolved
nitrogen loss on the hillslope. It is generally agreed that erosion rates from plot scale studies
should not be extrapolated beyond this scale (Parsons et al., 2006), however, for a number
of land uses in Australia (e.g. bananas), plot data is often all that is available. Therefore it is
important to find out if these data are suitable for application in catchment models, or
whether only catchment scale data should be used. To do this, data from the Forest and
Grazing (on native pastures) land uses were analysed to test if there was a difference in the
constituent values at different spatial scales. These were the only land use types where
>90% of the catchment was represented by a single land use at different scales. The scale
classes used were plot to small catchment (which is <1 km2 for grazing and < 10km2 for
forest due to limited data), medium catchment (1-100 km2) and large catchment scales (>
100 km2). Note that for the grazing land use the plot data were generally from sites less than
0.01 km2, but a 1 km2 threshold was used to accommodate the few studies that were larger
than 0.01 km2. There was no statistical difference between TSS concentrations for plots that
were <0.01 km2 and plots between 0.01 and 1 km2 (p=0.9, data not shown).
Knowledge of the minimum % contribution of a single land use has not been defined in the
literature. The land use percentages reported vary from as low as 5% for intensive high
fertiliser land uses (e.g. bananas) that rarely occupy whole sub-catchments (e.g. Bainbridge
et al., 2009), to 100% contribution for less intensive land uses such as forestry (e.g. Hunter
and Walton, 2008). As an example, Rohde et al., (2008) have classed some land uses as
‘sugar’ as the ‘major’ land use category, however, it only represents 25% of the catchment,
and no mention of the other land uses in this catchment are given. So although this
catchment has been classed as sugar, as this is the land use that dominates the water
quality signal (e.g. nitrate), the majority of the catchment is under forest.
This is important for applying these data in modelling situations. For example, if users of the
Source Catchments model are characterising the functional units (FU’s) in a model
according to land use, then it is important to be representing each specific land use
appropriately. This also implies having an understanding of the condition of the land use. In
the majority of studies, the water quality measured at a sub-catchment is a conglomerate of
the range of land uses upstream (e.g. 25% sugar, 50% forest, 25% grazing) and just
because one of the land uses may be more intensive (e.g. sugar) it does not necessarily
mean that the water quality downstream in representative of that land use on its own. In
other international studies, researchers have undertaken a land use specific compilation of
data for plot scale data alone (see Harmel et al., 2006a), however, the Source Catchments
model generally operates at a sub-catchment scale and therefore processes other than just
hillslope or sheet erosion need to be taken into consideration (e.g. gully erosion,
groundwater input etc). To test if there are differences in the constituent concentration
values for different proportions of a given land use, data from the Forest and Grazing (on unmodified pastures) land uses were analysed as these land uses had the most data. They
were grouped according to the proportion of area with a single land use (i.e. >90%, 70-90%,
50-70% and <50%). A Kruskal-Wallis One Way Analysis of Variance on Ranks was
performed to identify if there was a % threshold that gave a statistically different result for
TSS, TN and TP.
16
4.9
Linking the data base to Google Earth
The water quality database was developed using Microsoft Excel. An Excel macro was
created to convert the data into a Keyhole Markup Langauge (KML) file. This KML file can be
viewed using Google Earth TM, a spatial mapping program.
17
5 Results
5.1
Summary of data sets used in data base
A total of 757 entries were put into the data base from 514 different geographical sites
covering 13 different land uses (Table 2). A list of references used to create the data base
are given in Appendix B. A total of 44 entries (~6%) had data that could be used to calculate
both event concentrations (EMC) and dry weather or base-flow (DWC) concentrations at the
same site, and only 41% of sites had data for all three constituents: TSS, TN and TP. For the
total data base, ~70% of sites provided sufficient information to describe the sample
collection methods, ~85% provided information about the method used to calculate the
mean constituent concentration and ~26% described the number of samples collected over
the hydrograph. Only 17% of entries provided information on the laboratory analytical
methods used and <5% described the condition of the land use upstream of the sample
collection point in terms of vegetation cover or management practices.
Figure 3 presents a geographical summary of the location of the 514 sites in the data base.
It shows that there is a heavy bias to data from Queensland. This may reflect the amount of
research conducted in Queensland due to the link between catchment condition, water
quality runoff and the health of the Great Barrier Reef (GBR) or, to some extent, it may
reflect the author’s ability to access event data collected in this State. A copy of the KML file
used to generate the Google Earth Map of water quality sites is available from the eWater
CRC (http://www.ewater.com.au/).
Table 2: Summary of land uses and the number of sites included in the data base (including plot/hillslope
studies)
Land use
Bananas
Cotton
Dryland Cropping
Forest
Forestry
Grazing Modified Pastures
(includes dairy)
Grazing Native Pastures
(includes savannah /
woodland grazing)
Horticulture
Military
Mining
Mixed or unknown
Sugar Cane
Urban / Industry
Number of sites in data base
(primary land use)
4
4
27
121
27
Number of sites >90% land use
0
4
0
<17
<4
<17
70
<58
269
23
17
9
60
35
85
18
13
0
9
14
2
Figure 3: Location of water quality data sites reported and entered into the data base are indicated by the
red markers
Figure 4: An example of the data available for the the Bowen River catchment in
Queensland for 2004
19
5.2
Preliminary analysis to determine data for inclusion in final results
Initial analysis of the data was undertaken to determine if some samples should be excluded
from the final summary results due to:
(a) sub-optimal sampling protocols (e.g. collecting water quality samples at a site
without specific hydrographic data) (Section 5.2.1)
(b) variations in the spatial scale at which the samples were collected (Section 5.2.2); or
(c) the proportion of the land use area upstream of a water quality collection point not
being representative of a single land use (Section 5.2.3)
A number of box and whisker plots are presented in this report and Figure 5 describes the
statistical data represented in the figures.
Figure 5: A description of the statistics represented by the box and whisker plots in subsequent sections
5.2.1
Comparison of constituent data collected using different methods
A simple comparison of three data sets from the same land use (forest) in a similar
geographic and climatic region (wet tropics of North Queensland) is presented in Table 3.
Preliminary results suggest that data collected without specific hydrographic data are lower,
particularly for nitrogen, than for sites where samples were collected in association with
hydrographic data (Table 3). This result may be due the condition of the catchment,
magnitude of the event or other environmental factors as outlined in Section 4.2.1, although
it is also likely that the high concentrations that often occur at the beginning of events were
missed and, on average, this resulted in lower median concentrations of constituents. Figure
6 demonstrates the higher constituent concentrations measured at the beginning of an event
on the Mossman River, North Queensland. These early concentrations are much higher than
for subsequent events with similar discharge. These results suggest that concentrations may
be lower when samples are not collected simultaneously with flow data.
This section provided a simple demonstration of the possible ramifications of sampling
without appropriate hydrographic data, however, there is insufficient data to demonstrate a
statistically significant difference and thus little justification for excluding data collected
without associated hydrographic data. It is important to note that this is example is relevant
to event conditions. The reverse situation may also occur whereby insufficient base flow
sampling is undertaken to represent nutrient concentrations in groundwater dependent
systems or perennial streams.
20
Table 3: Comparison of measured ‘median’ constituent concentration values for sites with the same land use and similar climate, but using different sampling
methods. Note: (i) these are measured concentration data and not EMC values and (ii) the use of median values results in subtle differences in nutrient balance
calculations reported in the table
Case
Study
Catchment
Method of
sampling
1
Tully- Murray
catchment
(Davidson
Creek, Fishtail)
Pioneer (St
Helen’s Creek)
Mossman River
Manual
without
hydrograph
2
3
Manual over
hydrograph
ISCO sampler
with stage
recorder
Land use
type and
% land
use
Forest,
98%
Catchment
area (km2)
No. of
samples
Source of
data
TSS
(mg/)
TN
(ug/l)
PN
(ug/l)
TFN
(ug/l)
Median values
DON
DIN
TP
(ug/l)
(ug/l)
(ug/l)
PP
(ug/l)
TFP
(ug/l)
DOP
(ug/l)
FRP
(ug/l)
93
10
Bainbridge
et al (2009)
2
150
27
119
69
39
14
5
8
6
2.2
Forest,
98%
Forest,
99%
24
16
95
566
133
453
243
209
68
34
26
8
19
87
173
Rohde et
al., (2008)
McJannet et
al., (2005)
8
555
215
340
299
41
14
4
5
5
<1
21
150
3
100
2
50
1
0
0.15
0.10
0.05
Feb
Mar
Apr
May
Jun
200
B
180
Discharge
Total Suspended Sediment
200
160
140
120
150
100
80
100
60
40
50
20
0
0
Jan
Feb
Mar
Apr
May
Jun
0.00
Jul
Jul
Total Suspended Sediment (mg/l)
Jan
3
Discharge (m /s)
0.20
0
Dec
Dec
0.25
Total Phosphorus (mg/l)
4
Total Nitrogen (mg/l)
Discharge
Total Nitrogen
Total Phosphorus
200
250
0.30
5
A
3
Discharge (m /s)
250
Figure 6. Discharge against sample concentrations of total nitrogen, total phosphorus (A) and total
suspended sediment (B) at Upper Mossman (Source: McJannet et al., 2005). Note nutrient data units are
mg/l in this specific graph.
5.2.2
Variation of data at different spatial scales
Data from the Forest and Grazing (on native pastures) land uses were analysed to test if
there was a difference in the constituent values at different spatial scales. Statistical
transformation of the data was not possible, therefore a non-parametric Mann Whitney rank
sum test was used to determine if there was a statistical difference in the data at different
spatial scales.
For total suspended sediments (TSS) there was no statistical difference (at the p< 0.05
level) between plot (<1 km2) and medium catchment (1-100 km2) scale data for both the
forest (p = 0.427) and grazing land use (p=0.097). There was a statistical difference between
TSS values for Grazing between the moderate catchment (1-100 km2) and large catchment
(>100 km2) sizes (p<0.001) and this may reflect the incorporation of gullies as a sediment
source at the larger scale (> 100 km2), and hence higher mean TSS values for catchments
great than 100 km2 when compared to catchments between 1-100 km2 (Figure 8). There was
insufficient data to allow further analysis of the forest land use at larger scales. It could be
argued that using data from the whole of Australia contributed to the high variability within a
given data set, however, there can be as much variability in the constituent concentration
within the same land use of similar size as there is between different sizes (see Section
4.2.1 and Figure 7). Therefore the geographical spread of the data was not considered to
strongly influence this result. There were no statistical differences between plot and
catchment scale studies that justify analysing data collected at different scales separately.
22
Figure 7: Comparison of data from the same land use (Grazing on native pastures) in the same
catchment (Burdekin River Basin) showing the variation in TSS values between plot scale studies (<1
2
km2) and large catchment areas (~7,000 km ). Data sourced from Hawdon et al., (2008), Bartley et al.,
(2010a), Bainbridge et al., (2007), and Bartley et al., (2007a).
We had insufficient data to undertake the same analysis for sugar cane at different spatial
scales as fertiliser intensive land uses rarely occupy catchments greater than 100 km2.
Using TN values, however, there is a statistically significant decline (p <0.05) between TN
values measured at the plot or <1 km2 scale compared with catchment scales (> 1 km2)
(Figure 9). In this analysis the plot scale data were represented by > 95% sugar land use.
For the 1-100 km2 and >100 km2 groups, the proportion of sugar represented was between
25-100% and 31-72% which suggests that once data from other land uses are included, the
water quality concentrations or TN and TP in sugar cane are reduced. The reason why the
TP values increase again at the larger scale (> 100 km2) in Figure 9B is possibly due to the
increased contribution of channel erosion processes which would increase the amount of
sediment attached phosphorus.
(B)
(A)
Figure 8: The difference in TSS concentration values for sites at plot, medium-catchment and large
catchment scales for the (A) Forest and (B) Grazing (on native pastures) land uses. Note for both sets of
data the land use occupied >90% of the catchment.
23
(A)
(B)
Figure 9: The difference in (A) TN concentration and (B) TP values for sites at plot (<1 km2), subcatchment (1-100 km2) and large catchment scales (>100 km2) for the Sugar land use. Note that the
2
proportion of the catchment represented by sugar cane in each category varies. For the <1 km data
2
2
sugar is >95%, for 1-100 km it is between 25-100% and for the >100 km it is between 31-72%. Note that
the TSS plot for sugar has a very similar pattern as TP.
5.2.3
Does the constituent concentration vary when the proportion of upstream land
use changes?
The results of the Mann-Whitney Rank sum test suggest that, in the majority of cases, there
is a statistically significant difference between the constituent value measured from a site
where a single land use occupies >90% of the site, compared with sites that has < 90% of
the upstream catchment area dominated by one land use (see Table 4). This result appears
to be stronger for low intensity land uses such as forest (Figure 11) and high fertiliser land
uses such as sugar (Figure 12), than for moderate land uses such as grazing (Figure 10). A
detailed discussion of some of the reasons for this difference are given in Section 6.
Table 4: Summary table showing when there was a significant difference (p<0.05) between sites that had
>90% of the catchment under a single land use compared with 70-90% of the area.
TSS
TN
TP
Grazing
Yes
(p=0.025)
No
(p=0.187)
No
(p=0.589)
Forest
Yes
(p=0.05)
Yes
(p=0.03)
Yes
(p=0.01)
*Note that sugar had different % land use contribution thresholds than grazing and forest
24
Sugar*
Yes
(p = 0.003
Yes
(p=0.014)
Yes
(p=0.005)
(A)
(B)
(C)
Figure 10: (A) TSS (B) TN and (C) TP for Grazing land use according to the % of that land use represented above the water quality sampling point
(B)
(A)
(C)
Figure 11: (A) TSS (B) TN and (C) TP for Forest land use according to the % of that land use represented above the water quality sampling point
25
(A)
(B)
(C)
Figure 12: (A) TSS (B) TN and (C) TP for the Sugar land use according to the % of that land use represented above the water quality sampling point. Note that the %
threshold of land that was used for TSS was slightly different to that for TN and TP due to limited data, and the land use % representations are different to the
grazing and forest data sets.
26
5.3
Summary of water quality data for different land uses
Based on the results presented in Sections 5.2.1, Section 5.2.2 and 5.2.3, the final data is
presented in two sections:
1.
The main graphs (Figure 13, Figure 14 and Figure 15) and Table 5 and Table 6
present a summary of the plot and catchment data together for all land use data
based on the land use classification given in the original documents from which the
data were derived. These results are presented to show the range or spread of data
and allows the whole data base to be utilised in generating the summary statistics.
2.
Then Table 7 presents the constituent results for sites where the land use upstream
is dominated by (>90%) a single land use. This data could be considered more
robust when modellers are looking to represent land use specific concentrations,
however, it is important to be mindful that this smaller data set is biased towards
smaller plot sizes for the intensive land uses (e.g. sugar cane).
The land uses with the highest median TSS concentrations are Mining (~50,000 mg/l),
Horticulture (~3000 mg/l), Cotton (~600 mg/l), Grazing on native pastures (~300 mg/l), and
Bananas (~200 mg/l) (Figure 13A). The highest median TN concentrations are from
Horticulture (~32,000 ug/l), Cotton (~6,500 ug/l), Bananas (~2,700 ug/l), Grazing on modified
pastures (~2,200 ug/l) and Sugar (~1,700 ug/l) (Figure 13B). For TP it is Forestry (~5,800
ug/l), Horticulture (~1,500 ug/l), Bananas (~1,400 ug/l), Grazing on modified pastures (~400
ug/l) and Grazing on native pastures (~300 ug/l) (Figure 13C). A summary of the results is
presented in Table 5. For the dissolved nutrient fractions, Sugar has the highest
concentrations of DIN, DON and DOP, and the Urban land use has the highest
concentrations of DIP (Figure 14 and Table 6).
27
(B)
(A)
(C)
Figure 13: Constituent concentrations for (A) Event TSS (mg/l) (B) Event TN (ug/l) and (C) Event TP (ug/l) for land uses where n>3. Note the log (natural) y axis and
Dairy has been presented with Grazing on modified pastures
28
(B)
(A)
(C)
(D)
Figure 14: Event (A) DIN (B) DON (C) DIP and (D) DOP (ug/l) concentrations for land uses (where n>3). Note that Bananas and cotton have been combined with
Horticulture.
29
Table 5: Summary of data for TSS, TN and TP for the 10 major land uses identified in this study. Note that n represents a data point from a single geographical site
and in some cases this represents individual event data, in other cases it represents annual or multi event averages. Data were only presented when n ≥ 3
n
Bananas
Cotton
Forest
Forestry
Grazing Modified
Pastures (includes
Dairy)
Grazing Native
Pastures (includes
savannah/woodla
nd grazing)
Horticulture
Mining
Sugar Cane
Urban/Industry
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
th
10 percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
th
10 percentile
Median
Mean
th
90 percentile
10th percentile
Median
Mean
th
90 percentile
4
4
59
11
14
200
17
9
27
49
TSS (mg/)
Event
n
(EMC)
229
378
0
428
667
68
567
0
732
1529
5
26
54
77
228
5
33
0
50
56
23
188
n<3
256
650
24
330
30
837
2204
8
3104
0
5945
11719
84
51000
0
64474
136800
25
97
0
120
287
32
105
18
227
271
30
Baseflow
(DWC)
n
4
4
5
9
9
13
53
TN (ug/l)
Event
(EMC)
2204
2712
2745
3314
2122
6438
6436
10748
179
436
782
1533
n<3
54
6
14
19
38
137
17
1400
2235
6763
18880
558
1270
1593
2903
585
32181
31539
61414
10
35
44
73
51
1034
1710
2598
5742
180
530
648
1290
Baseflow
(DWC)
n
0
4
0
4
21
15
0
32
n
1
362
492
950
203
445
699
1162
6
48
n<3
41
68
250
880
1023
2050
137
0
17
0
0
0
32
15
1021
2280
2675
4062
36
TP (ug/l)
Event
(EMC)
1261
1360
1439
1679
72
230
239
414
16
70
222
585
76
5800
6175
12650
100
355
563
1295
66
310
482
1162
23
1451
3233
8315
n
Baseflow
(DWC)
0
0
40
15
10
30
44
112
6
14
35
100
n<3
43
9
30
66
184
0
0
62
223
292
731
80
161
1884
6165
0
14
87
160
348
402
Table 6: Summary of data for DIN, DON, DIP and DOP for the 10 major land uses identified in this study for event (EMC) conditions only. There was insufficient
data to present baseflow (DWC) values. Note that n represents a data point from a single geographical site and in some cases this represents individual event
data, in other cases it represents annual or multi event averages. Data were only presented when n ≥ 3
n
Bananas
Cotton
Forest
Forestry
Grazing Modified
Pastures (includes
Dairy)
Grazing Native
Pastures (includes
savannah/woodland
grazing)
Horticulture
Military
Mining
Sugar Case
Urban/Industry
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
th
10 percentile
Median
Mean
th
90 percentile
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
DIN
n
DON
n
DIP
n
0
0
0
0
0
0
0
0
27
96
210
531
49
0
35
69
119
131
222
0
73
188
2209
6530
43
190
214
434
112
369
780
1880
14
83
7
8
61
7
42
3
7
13
29
0
85
118
299
624
200
360
526
1290
185
212
986
2372
12
80
35
3
13
75
228
9
33
45
80
6
64
0
3
0
0
0
0
0
0
0
22
15
31
17
5
131
502
480
672
248
270
283
323
20
15
4
9
12
17
0
0
254
772
1534
4598
73
757
788
1587
DOP
3
30
121
526
13
43
63
145
17
7
14
17
16
18
3
11
14
27
1
2
3
4
18
27
33
58
7
14
15
25
Table 7: Summary data for TSS, TN and TP for the 10 major land uses identified in this study where the land use upstream of the sampling point is >90% under a
single land use. Data is for event (EMC) conditions only. There was insufficient data to present baseflow (DWC) values. Note that n represents a data point and in
some cases this represents individual event data, in other cases it represents annual or multi event averages.
n
Bananas
Cotton
Forest
Forestry
Grazing Modified
Pastures (includes
Dairy)
Grazing Native
Pastures (includes
savannah/woodland
grazing)
Horticulture
Mining
Sugar Cane
Urban/Industry
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
10th percentile
Median
Mean
th
90 percentile
th
10 percentile
Median
Mean
90th percentile
th
10 percentile
Median
Mean
th
90 percentile
10th percentile
Median
Mean
th
90 percentile
TSS (mg/l)
0
n
TN (ug/L)
0
68
567
732
1529
3
10
27
83
2
9
67
177
39
315
322
390
11
689
1571
4825
495
3774
7750
12440
84
51000
64474
136800
64
119
153
300
96
120
120
144
4
17
4
9
58
13
9
14
2
32
4
15
20
13
2122
6438
6436
10748
85
227
385
527
2
4
16
72
230
239
414
8
19
103
379
0
1652
2417
3044
4920
311
901
2175
3880
20272
41497
40652
65345
0
14
TP (ug/L)
0
0
9
n
17
19
13
174
429
726
2174
37
274
661
2000
643
2294
4195
8605
0
1492
2340
4050
8168
387
415
415
443
14
0
216
415
468
801
5.4
Comparison of event and base flow data for the dominant land uses
There were sufficient data to compare event (EMC) and base flow or dry weather (DWC)
TSS and TN concentrations for three land uses, and four land uses for TP (Figure 15). The
TSS values recorded for event flows were all statistically higher (p< 0.05) than for base flow
conditions for all land uses. For the TN data, there was no significant difference between
event and base flow concentrations for forests, but the grazing land use had significantly
lower DWC values than EMC values. Interestingly, the DWC TN values for the Urban land
use were significantly higher (p<0.05) than the EMC values. This appeared to be skewed by
data collected during baseflow conditions in the Melbourne city catchments (Mitchell et al.,
1998), and suggests that there may be a source of nitrogen in the urban environment (e.g.
garden fertilisers, dog manure and reticulated water leakage) that is reaching waterways
under baseflow conditions. For TP, Forest, Forestry and Grazing all had statistically higher
EMC concentrations than DWC (p<0.05), but there was no statistical difference between the
EMC and DWC values for the Urban land use.
33
(B)
(A)
(C)
Figure 15: Comparison of event (EMC) and base flow (DWC) concentrations for (A) TSS (mg/l) for Forest, Grazing (on native pastures) and Urban land uses, (B) TN
(ug/l) for Forest, Forestry, Grazing (on native pastures) and Urban land uses and (C) TP (ug/l) concentrations for Forest, Forestry, Grazing (on native pastures) and
Urban land uses
34
6 Discussion
6.1
Summary of findings from this study
The overwhelming outcome from this study is that there is a high variability in water quality
both within and between sites. The reasons for collecting water quality data vary
considerably, and so to do the field, lab and data analysis methods. Water quality data is
expensive to collect, and therefore very few studies appear to have measured all of the
constituents, and collated all of the required data (including information on land condition) at
a range of scales, under the most robust sampling and analysis conditions. Acknowledging
that there was considerable variability in the data, a number of analyses were undertaken to
determine if and when this variability affects the statistical result for a given constituent
representing a given land use.
An analysis of data collected from the same land use using different field collection and data
analysis procedures suggested that concentration data that is not collected simultaneously
with flow data, is likely to be lower in value than when data is collected using hydrographic
data (Table 3). Facchi et al., (2007) concluded that grab sampling is a relatively inefficient
methodology for capturing mean concentrations for rivers subjected to highly variable loads,
especially when it is restricted to office hours. Based on the analysis presented here it could
be argued that grab sampling is a suitable method for estimating concentration as long as
the water stage and location of the sample on the hydrograph can be determined. It is
recommended that flow data be captured (or at a minimum estimated) for each site where
constituent data is obtained. Sansalone et al., (2005) found that some constituents such as
suspended solids are driven by the hydrology of the event, as particulate matter is reliant on
the energy or stream power of an event to move the material. Dissolved constituents, on the
other hand, followed an exponential type decline during the rising limb of the hydrograph. It
is important therefore to collect hydrographic data at the same time as constituent data is
collected so that a robust understanding of constituent generation and movement pathways
can be calculated.
Based on the data sets available in this study, it was found that there wasn’t a statistical
difference between the median TSS concentration values at different spatial scales (Figure
8). It is important to note that sample sizes in this study were not large (n between 3 and 33),
nonetheless, this result was surprising given that other studies have advocated that it is not
suitable to extrapolate results from small scale plot studies to larger catchments scales
(Parsons et al., 2006). However, recent studies have shown that unit area sediment yields
decline with increasing plot length for undisturbed and moderately disturbed sites, but
actually increase for the highly disturbed sites (Moreno-de las Heras et al., 2010). Similarly,
Wilcox et al., (2003) concluded that site condition is likely to have as much or bigger
influence on the constituent concentration than the scale of study, and many studies have
shown that the type of vegetation on a plot can dramatically influence constituent
concentrations at a site (e.g. Michaelides et al., 2009). In addition, areas that have well
developed rill erosion, the sediment concentrations are likely to increase with increasing
area producing values that are similar to plot scale studies.
The results differed for the nutrient data which showed that plot scale concentrations were
higher than concentrations derived from catchment studies. These results were, however,
confounded by the proportion of land use represented in each of the size classes. Mitchell et
al., (2009) found a near perfect linear relationship between stream nitrate concentrations and
the proportion of land use upstream that used fertiliser, and similar results have been shown
in other studies (e.g. Feng et al., 2010). These results appeared to be independent of
catchment size and suggest that the proportion of land use upstream of the sampling point
that uses fertiliser is more important that the size of the contributing area upstream. It is
35
important to keep in mind that there may be considerable variability between plot and
catchment scale studies in certain situations. For example, a number of studies have shown
that concentrations of nutrients in sub-surface and deep drainage water can be as high or
higher than for surface runoff (e.g. Ridley et al., 2003) and in these cases nutrient
concentrations may increase at larger spatial scales when ground water and sub-surface
nutrient concentrations are incorporated.
The results presented in this study suggest that when you have <90% of a given land use,
you are potentially introducing contamination or influence from other land uses into the water
quality signal, and therefore data from sites with <90% are not necessarily representative of
that land use when used in a modelling context (e.g. Figure 10, Figure 11 and Figure 12).
Using data only from sites with >90% a single land use, although potentially more
appropriate for land use based modelling, does introduce a number of problems. The
primary issue is that it dramatically reduced the number of data points available for analysis
in this study from 757 to ~130. It also reduces the range of plot sizes incorporated into the
final analysis, particularly for the more intensive land uses (e.g. sugar cane) that rarely
occupy large parts of a catchment, and are rarely in steep headwater areas. Also processes
such as denitrification may have been taken into account in larger scale studies, but not plot
scale studies.
It is important to note that the number of data points in each land use category is different,
and this may have biased some of the results. For example, the TN values for Horticulture
are dominated by a single study (where they grew lettuce, spinach, capsicum) which was
based on 13 events with very high soil and nutrient loss rates (Hollinger et al., 2001). It is not
certain if this study is ‘typical’ of horticulture, and these results highlight where further data
collection is needed.
6.2
Issues to be aware of when using data in catchment models
This section describes some of the issues that modellers need to be aware of when using
and applying average water quality data within a catchment scale modelling framework.
6.2.1
Uncertainty in data collection, handling, processing and analysis
No attempt was made to estimate the uncertainty for data collected in this study, however, it
is acknowledged that understanding the level of uncertainty, and where it is introduced in the
data collection process, may be important for deciding what data is used in a modelling
project. The results presented in Harmel at el., (2006b) provide some estimate of the
cumulative uncertainty in water quality data for the purpose of providing defensible estimates
of data uncertainty to support water resource management. In summary, they determined
that:
‘Averaged across all constituents, the calculated cumulative probable uncertainty (±%)
contributed under typical scenarios ranged from 6% to 19% for streamflow measurement, from
4% to 48% for sample collection, from 2% to 16% for sample preservation/storage, and from
5% to 21% for laboratory analysis. Under typical conditions, errors in storm loads ranged from
8% to 104% for dissolved nutrients, from 8% to 110% for total N and P, and from 7% to 53% for
TSS.’
With respect to constituents, the study by Harmel at el., (2006b) showed that there is more
uncertainty related to measuring dissolved forms of N and P than for TN, TP or TSS, and the
study highlighted the importance of good QA/QC when collecting data. It is also important to
note that errors can be introduced during the data analysis process, particularly when
estimating constituent loads (see Kuhnert et al., 2009; Schwartz and Naiman, 1999).
36
6.2.2
Using ‘event mean concentration’ (EMC) versus ‘mean event concentration’
(MEC)
As described in Section 3 there are a number of different approaches for estimating average
event concentration for a site. There are advantages and disadvantages with each of these
methods, and it is important to highlight that use of each approach is dependent on samples
being adequately collected across the hydrograph (as discussed in Section 4.7 and 5.2.1).
The EMC approach (load divided by flow) is often considered one of the more accurate
techniques as it is a true flow weighted concentration, however, the amount of runoff can
have a large influence on the EMC value at a given site. Even when you have two sites in a
similar condition, a lower EMC value will be obtained in a year with high flow than in a year
with low flow due to the dilution of constituent loads. This can be demonstrated using the
data from two hillslope flume sites in the Burdekin catchment. Table 8 shows that in 2008
and 2009, the mean measured concentration from Flume 1 was similar at 88 and 98 mg/l,
respectively. However, due to the higher runoff in 2008, the total load in 2008 was much
higher than in 2009 and thus the TSS EMC value for 2008 (calculated as load divided by
flow) was ~1/3 of that calculated for 2009. It is also interesting to note that the calculated
EMC values are ~1.2 and 3.1 times higher than the average measured concentration values
for Flumes 1 and 2, respectively (Figure 16). This overestimation is largely due to the load
calculation (averaging) method used in this study. The load calculation method, and
associated uncertainty, can vary enormously between studies (Marsh and Waters, 2009a),
and this will influence the value of an EMC, but not the MEC. In some cases it may actually
be more appropriate to use MEC than EMC values if the load calculation method is
considered to bias the result.
Defining the mean concentration in an event by using flow duration curves (e.g. Chiew and
Scanlan, 2002) has many benefits as it provides a consistent method for selecting
concentrations at the same point on the hydrograph between sites and years. A limitation of
this approach is that it does not distinguish between different concentrations observed in
different parts of the event due to hysteresis (see Inamdar et al., 2006; Nistor and Church,
2005; Sansalone and Cristina, 2004), however, it is possible to adapt this approach to
calculate mean event concentrations at different points on the hydrograph that represent the
important constituent generating processes (see Section 6.2.4).
Flume 2
Flume 1
3000
y = 3.12x
R² = 0.92
TSS EMC (mg/l)
2500
2000
y = 1.23x
R² = 0.94
1500
1000
500
0
0
200
400
600
800
1000
Mean measured TSS concentration (mg/l)
Figure 16: Comparison of mean measured annual TSS concentrations and annual TSS EMC values for
Flume 1 and 2
37
Table 8: Annual median and mean measured TSS concentrations and the unit area sediment load and annual EMC values for 6 years of monitoring data at two
hillslope flume sites in the Burdekin catchment (Source of data, Bartley et al., 2010a)
Year
2002
2003
2004
2005
2006
2007
2008
2009
Annual
Rain
(mm)
Runoff (mm)
304
245
368
431
698
771
1308
630
16
32
33
30
82
147
309
58
Median of all
TSS samples
collected during
flow events in
that year (mg/l)
Flume 1
Mean
of all TSS
samples
collected during
flow events in
that year (mg/l)
845
439
288
240
88
62
67
81
868
460
299
258
118
60
98
88
*logger failure during wet season
38
EMC (mg/l) =
Total annual
load divided
by flow
1185
551
203
197
93
44
31
86
Sediment
loss (t/ha)
Runoff (mm)
Median of all
TSS samples
collected
during flow
events in that
year (mg/l)
0.27
0.25
0.09
0.08
0.11
0.09
0.14
0.07
11
21
24
68
115
>61*
66
1365
760
226
49
23
10
14
28
Flume 2
Mean
of all TSS
samples
collected
during flow
events in that
year (mg/l)
1365
759
226
621
45
19
19
28
EMC (mg/l)
2575
1160
1522
189
43
189
59
Sediment
loss (t/ha)
0.06
0.06
0.09
0.03
0.01
0.03
0.01
6.2.3
Selecting appropriate mean concentration values for load calculations
Given that there is considerable variability in the water quality data for any given land use,
there are a range of approaches available for selecting a constituent value. Young et al.,
(1996) summarised the three general options that modellers and catchment managers have
when selecting an approach for data use and application in catchment models, (i) use a ‘best
estimate’ value, (ii) use local experimental data that matches the situation or (iii) make a
subjective adjustment to the available figures to account for differences in management and
environmental attributes. Obviously (i) is a last resort and (ii) would be ideal, but in the
majority of cases, local data with the same attributes as the modelled catchment will not be
available, and therefore using a combination of (ii) and (iii) would be most appropriate.
To demonstrate the impact of using different concentration values on the end of catchment
loads, data collected at the outlet of a 14 km2 catchment in the Burdekin in 2001/02 wet
season was used (see Bartley et al., 2010b for description of sites and methods). The
2001/02 wet season had only two events that were well sampled for TSS and continuous
turbidity (see Figure 17). Walling and Webb (1981) viewed the load derived from the
continuous TSS-turbidity relationship is the most accurate representation of the ‘true’
sediment load against which to assess the accuracy of estimates produced by other indirect
load calculation procedures. The load calculated for the 2001/02 wet season was therefore
considered as the accurate load for the site (Table 9). The mean annual sediment load for
the site was then calculated using the same runoff and four different methods for estimating
the mean concentration for the site. The methods included:
• Option 1: using the average sediment concentration for grazing lands in the Burdekin
derived from the water quality data base developed for this project (~948 mg/l);
• Option 2: using the 90th percentile sediment concentration for grazing lands in the
Burdekin (derived from the water quality data base) (~2925 mg/l). This site was
considered to have contributed disproportionately high sediment yields based on
previous modelling applications (Prosser et al., 2001a), therefore an upper bound
estimate was considered appropriate;
• Applying the average concentration measured at the gauge for that year (~2395
mg/l);
• Applying the median concentration measured at the gauge for that year (1930 mg/l).
The results show that using the average TSS concentration for this particular land use, even
from the same region, resulted in sediment loads that were ~67% lower than the true load.
Using the measured mean and median concentration data from the same site resulted in
loads that were ~11% and 25% lower than the true load and using the 90th percentile EMC
from the region over-estimated the true load by ~1.7%. This example has highlighted that:
(i)
It is important to have an understanding of the condition and history of land use
for the site being modelled compared to other sites with the same land use; and
(ii)
It is important to obtain data from a site that has a similar condition to the site that
is being modelled, and this may not always be in the same geographic location.
39
Figure 17: (A) Event 1 on 16th and 17th of December 2001 and (B) Event 2 on 15th February 2002 showing
the high turbidity and sediment concentrations early in the event
Table 9: Sediment load (tonnes) calculated using the same hydrology with 4 different methods for
estimating the mean concentration for the 2001/02 data set
Sediment load (tonnes)
Option
1
2
3
95th Percentile
Using average
Mean concentration concentration for the
measured
for the Burdekin
Burdekin
concentration from
*
*
Method
(= 778 mg/l)
(=2925 mg/l)
the field site
4
True load
Using median
measured
concentration from Using TSS-turbidity
the field site
relationship
1
246
685
552
508
705
2
123
341
345
246
305
Total
369
*calculated from the data base
1026
897
754
1010
Event
6.2.4
Differentiating between Event (EMC) and Base flow or dry weather (DWC)
values
Models such as Source Catchments are increasingly being used to evaluate the
effectiveness of best management practice for different land uses on downstream
ecosystems such as the Great Barrier Reef (http://www.nrm.gov.au/funding/2008/reefrescue.html). They are also being used to link to estuarine and marine models to track the
movement of constituents from paddocks to the ocean (e.g. eReefs project). Such
application of these models are likely to require a greater temporal accuracy in estimating
constituent concentration than simply using average EMC or DWC values.
Figure 18 demonstrates how the concentration of sediments and nutrients can vary at
different times of the wet season on the Mossman River in North Queensland, and how the
event flow concentrations compare with base-flow (dry season) concentrations. It is
important to note that although field data show that the EMC varies from one storm to
another, use of an average EMC in multiple-event computer simulations may lead to
estimates of total load that are as accurate as those obtained by more complex models
(Charbeneau and Barrett, 1998). This is, however, conditional on obtaining accurate EMC or
mean concentration data. If the time scale of interest is daily or hourly data, in most cases it
will be necessary to compartmentalise the hydrograph into smaller units and apply
concentration data that reflects the processes occurring at that point in time. Table 10
provides an example of how the total constituent load can differ when the concentration
40
values presented in Figure 18 are applied to the equivalent time periods on the hydrograph,
versus only applying a whole of wet season mean concentration (Dec-May) and a mean
baseflow concentration (Jun-Nov). These load values are then compared to the constituent
load calculated using linear interpolation (Degens and Donohue, 2002; McJannet and Fitch,
2005). Using mean concentration data appears to dramatically over-estimate TP and TSS
concentrations compared to the using loads calculated by linear interpolation. Dividing the
runoff period into smaller time periods reduces this over estimation slightly, but in general
studies that have compared modelled outputs against measured data suggest that long-term
average results are better simulated than results for individual time periods (Jetten et al.,
1999). This example is from a forested stream and the base flow contribution of constituents
in this catchment was <3%. This suggests that in some cases, use of a single event mean
concentration without a base-flow component would be appropriate, and within the bounds
of load error estimates. In other cases, however, baseflow can contribute approximately twothirds of the mean annual nitrate export as Schilling (2004) found on the Raccoon River
watershed in west-central Iowa. Therefore it is important to have collated sufficient water
quality data to determine the dominant flow processes and constituent concentration values
prior to determining the level of hydrograph separation for calculating constituent loads.
Table 10: Difference in constituent load values for the Upper Mossman River in north Queenland using
different methods for the period 1/12/03-30/06/04 (data sourced from McJannet and Fitch, 2005)
Dividing the wet season into 4 time periods (early, mid and late wet
season and base-flow)
Use event average (mean of early, mid and late wet season) and
base-flow average values only
Event contribution
Base flow contribution
Load calculated using linear interpolation*
*see McJannet and Fitch (2005) or Marsh and Waters (2009) for a description of this method
41
TN
Load (tonnes)
TP
TSS
182
8
4858
217.58
8.49
5200.74
211.22
8.23
5200.36
6.36
0.26
0.38
151
4
2902
Total Phosphorus
0.20
0.00
0.10
Total Phosphorus (mg/L)
3
2
1
0
Total Nitrogen (mg/L)
4
Total Nitrogen
EPA Base
Flow
All Wet
Season
Early Wet Mid Wet
Season
Season
Late Wet
Season
EPA Base
Flow
All Wet
Season
Early Wet Mid Wet
Season
Season
Late Wet
Season
Late Wet
Season
150
100
50
0
80
60
40
20
0
Turbidity-Lab (NTU)
EPA Base
Flow
Early Wet Mid Wet
Season
Season
Total Suspended Solids
Total Suspended Solids (mg/L)
Laboratory Turbidity
All Wet
Season
EPA Base
Flow
All Wet
Season
Early Wet Mid Wet
Season
Season
Late Wet
Season
Figure 18. Comparison of concentrations of TN, TP, turbidity and TSS for baseflow and event conditions
on the Mossman River, with event condition further divided into early (December 2003 to January 2004),
mid (February to March 2004) and later (April to May 2004) stages of the wet season. Baseflow
concentrations are from EPA water quality data from 34 sampling campaigns for the months May until
November for the period 1994 to 1999. Baseflow data for TSS was only recorded on three occasions,
each being below the detection limit (< 2mg/L) and has been recorded on the box plot as a single point at
2mg/L (Source: McJannet et al., 2005)
42
7 Conclusions and areas of further research
This study has presented a summary of the known and readily accessible land use based
water quality data for Australian catchments and rivers. A total of 757 data points were
summarised into 10 land use types. The amount and quality of the data available has
increased considerably over the last 15 years since previous data collation exercises were
undertaken (Young et al., 1996). Despite the size of this data base, there are still large areas
of Australia for which no data exist (see Figure 3). There is also minimal or insufficient data
for land uses such as Horticulture, Cotton and other high intensity crops such as bananas,
particularly for large plot/catchment scales (> 1km2).
As well as presenting the summarised data collated in this study, this report provided some
examples of why and how constituent concentrations vary within and between different land
uses. This can be linked to processes such as the condition of the land use from where the
data were collected, the data collection procedure or load calculation method. It is hoped
that this study will provide modellers with an increased understanding of the processes
contributing to a given constituent value, and thus the power to choose appropriate data for
their modelling study. This document should be used in conjunction with other relevant
publications that provide important information regarding model choice and application (e.g.
Grayson et al., 1992; Jakeman et al., 2006; Jordan et al., 2010; Silberstein, 2006).
With respect to future research needs, studies have shown that information on the condition
of the land can greatly improve the model results (e.g. Jetten et al., 1999). Therefore, in
future studies it is critical that land use condition information is collected alongside water
quality data to help differentiate between the natural changes in water quality trends,
compared with changes resulting from land management (e.g. Bartley et al., 2010b; Cerdan
et al., 2010). This will be critical to evaluating the impact of land use and climate change on
water quantity and quality in the future.
There is also a need to improve our understanding of the amount, timing and form of
constituents leaving a given land use. The use of high frequency monitoring data is likely to
provide important insights into this process (Bende-Michl and Hairsine, 2010) although Beck
et al., (1990) raise the possibility that more intensive monitoring may simply uncover yet
greater heterogeneity and so require still more monitoring. The need, therefore, is for
catchment monitoring to be at a scale that is relevant to the question of interest. At present,
the high cost of water quality sample collection and analysis inhibits the broader temporal
and spatial collection of data. Therefore any methods that will reduce the costs associated
with collecting data, whilst simultaneously increasing our understanding of the various
pathways for constituents, are most welcome. As well as improved data collection methods
we will also need appropriate data analysis and statistical methods to interpret these data.
Use of EMC values is reliant on accurate load estimates, which require appropriate
calculation and uncertainty estimation (e.g. Kuhnert et al., 2007). If land use based modelling
remains as the most suitable approach for modelling constituent loss at the catchment scale,
then more land use specific data are required urgently. This finding needs to be
communicated to Government agencies that are initiating event based monitoring programs
around Australia.
43
Take home messages….
•
•
•
•
•
More data is needed from high fertiliser land uses such as cotton, horticulture and
bananas at both plot and catchment scales;
More data is needed from ‘single land use’ sites at a range of scales (plot, subcatchment and catchment);
Land condition and management data should be collected for the land surfaces
upstream of the sampling point where ever possible so that changes in water quality
due land use/management change can be separated from changes to do with climate
variability;
Collection of water quality concentration data is of limited value when not collected
with simultaneous flow data;
Modellers need to be aware of the quality of the water quality data they are using.
Factors such as rainfall intensity and amount, land condition, sample collection
methods and load calculation methods all influence the final concentration value.
44
8 Acknowledgements
The research presented in this publication was funded by eWATER CRC; Meat and
Livestock Australia and Department of Environment, Heritage and the Arts. Their support is
gratefully acknowledged. The first author also received a CSIRO Payne-Scott award which
contributed to this production of this publication. A number of people made their data
available for this study and alerted us to data sources, thanks go to Dave Waters, Mark
Silburn, Zoe Bainbridge. Finally, thankyou to Tim Ellis, Dave Waters and Tony Weber for
comments and review on early versions of this manuscript.
45
9 References
ANZECC, 2000. The Australian and New Zealand Guidelines for Fresh and Marine Water
Quality Australian and New Zealand Environment and Conservation Council and
Agriculture and Resource Management Council of Australia and New Zealand.
ANZECC and ARMCANZ, 2000b. Australian Guidelines for Water Quality Monitoring and
Reporting, Australian and New Zealand Environment and Conservation Council and
Agriculture and Resource Management Council of Australia and New Zealand.
APHA, 1998. Standard Methods for the Examination of Water and Wastewater, 20th edition.
American Public Health Association, Washington, DC.
Argent, R.M., Brown, A., Cetin, L., Davis, G., Farthing, B., Fowler, K., Freebairn, A.,
Grayson, R., Jordon, P.W., Moodie, K., Murray, N., Perraud, J.M., Podger, G.M.,
Rahman, J. and Waters, D., 2008. WaterCAST User Guide, eWater CRC, Canberra.
Bainbridge, Z., Lewis, S. and Brodie, J., 2007. Event-based community water quality
monitoring in the Burdekin Dry Tropics Region: Volume 1: 2006/07 Wet Season
Report, Burdekin Solutions Ltd, Townsville.
Bainbridge, Z.T., Brodie, J.E., Faithful, J.W., Sydes, D.A. and Lewis, S.E., 2009. Identifying
the land-based sources of suspended sediment, nutrients and pesticised discharged
to the Great Barrier Reef from the Tully-Murray Basin, Queensland, Australia. Marine
and Freshwater Research, 60: 1081-1090, 10.1071/MF08333.
Bartley, R., Corfield, J.P., Hawdon, A.A., Abbott, B.N., Wilkinson, S.N. and Nelson, B.,
2010a. Impacts of improved grazing land management on sediment yields, Part I:
hillslope processes. Journal of Hydrology, 389: 237-248,
doi:10.1016/j.jhydrol.2010.06.014
Bartley, R., Hawdon, A. and Keen, R., 2007a. Sediment and nutrient loads at the Myuna
Gauge in the Bowen Catchment (2006/07), CSIRO, Brisbane.
Bartley, R., Hawdon, A., Post, D.A. and Roth, C.H., 2007b. A sediment budget in a grazed
semi-arid catchment in the Burdekin basin, Australia. Geomorphology, 87: 302-321.
Bartley, R., Wilkinson, S.N., Hawdon, A.A., Abbott, B.N. and Post, D.A., 2010b. Impacts of
improved grazing land management on sediment yields, Part 2: catchment response.
Journal of Hydrology, 389: 249-259, 10.1016/j.jhydrol.2010.06.014.
Beck, M.B., Kleissen, F.M. and Wheater, H.S., 1990. Identifying flow paths in models of
surface-water acidifcation. Reviews of Geophysics, 28(2): 207-230.
Bende-Michl, U. and Hairsine, P.B., 2010. A systematic approach to choosing an automated
nutrient analyser for river monitoring. Journal of Environmental Monitoring, 12(1):
127-134, 10.1039/0110156j.
Bloschl, G. and Sivapalan, M., 1995. Scale issues in hydrological modeling - a review.
Hydrological Processes, 9(3-4): 251-290.
Bramley, R.G.V. and Roth, C.H., 2002. Land-use effects on water quality in an intensively
managed catchment in the Australian humid tropics. Marine and Freshwater
Research, 53(5): 931-940.
Brodie, J.E. and Mitchell, A.W., 2005. Nutrients in Australian tropical rivers: changes with
agricultural development and implications for receiving environments. Marine and
Freshwater Research, 56: 279-302.
BRS, 2006. Guidelines for land use mapping in Australia: principles, procedures and
definitions, Bureau of Rural Sciences, Canberra.
Cerdan, O., Govers, G., Le Bissonnais, Y., Van Oost, K., Poesen, J., Saby, N., Gobin, A.,
Vacca, A., Quinton, J., Auerswald, K., Klik, A., Kwaad, F., Raclot, D., Ionita, I.,
Rejman, J., Rousseva, S., Muxart, T., Roxo, M.J. and Dostal, T., 2010. Rates and
spatial variations of soil erosion in Europe: A study based on erosion plot data.
Geomorphology, 122(1-2): 167-177.
Charbeneau, R.J. and Barrett, M.E., 1998. Evaluation of methods for estimating stormwater
pollutant loads. Water Environment Research, 70(7): 1295.
46
Chiew, F.H.S. and Scanlan, P., 2002. Estimation of pollutant concentrations for EMSS
modelling of the South-East Queensland region, CRC Catchment Hydrology
http://www.catchment.crc.org.au/pdfs/technical200202.pdf, Melbourne.
Cogle, A.L., Langford, P.A., Kistle, S.E., Sadler, G., Ryan, T.J., McDougall, A.E., Russell,
D.J. and Best, E., 2000. Natural Resources of the Barron River Catchment 2 - Water
Quality, land use and land management interactions. Q100033, Department of
Primary Industries, Brisbane.
Collins, A.L. and Anthony, S.G., 2008. Predicting sediment inputs to aquatic ecosystems
across England and Wales under current environmental conditions. Applied
Geography, 28(4): 281-294.
Davis, J.R. and Koop, K., 2006. Eutrophication in Australian Rivers, Reservoirs and
Estuaries - A Southern Hemisphere Perspective on the Science and its Implications.
Hydrobiologia, 559(1): 23-76.
Degens, B.P. and Donohue, R.D., 2002. Sampling mass loads in rivers – a review of
approaches for identifying, evaluating and minimising estimation errors., Water and
Rivers Commission, Water Resources Technical Series no WRT 25.
Duncan, H.P., 1999. Urban Stormwater Quality: A Statistical Overview. CRC for Catchment
Hydrology, Clayton, Vic.
eWater-CRC, 2010. Source Catchments Scientific Reference Guide, eWater Cooperative
Research Centre, Canberra.
Eyre, B. and Balls, P., 1999. A comparative study of nutrient behavior along the salinity
gradient of tropical and temperate estuaries. Estuaries, 22(2A): 313-326.
Fabricius, K.E., 2005. Effects of terrestrial runoff on the ecology of corals and coral reefs:
review and synthesis. Marine Pollution Bulletin, 50: 125-146.
Fabricius, K.E., De'ath, G., McCook, L., Turak, E. and Williams, D.M., 2005. Changes in
algal, coral and fish assesmblages along water quality gradients on the inshore Great
Barrier Reef. Marine Pollution Bulletin, 51: 384-398.
Facchi, A., Gandolfi, C. and Whelan, M.J., 2007. A comparison of river water quality
sampling methodologies under highly variable load conditions. Chemosphere, 66(4):
746-756, 10.1016/j.chemosphere.2006.07.050.
Feng, M.L., Hu, R.G., Lin, S., Malghani, S. and Javed, I., 2010. Influence of land use on
nitrate concentrations in baseflow in a rural watershed of Three Gorges Reservoir
Area, China. Journal of Food Agriculture & Environment, 8(1): 332-337.
Fletcher, T., Duncan, H., Poelsma, P. and Lloyd, S., 2004. Stormwater flow and quality, and
the effectiveness of non-proprietary stormwater treatment measures - a review and
gap analysis, Cooperative Research Centre for Catchment Hydrology, Report 04/8,
Melbourne.
Furnas, M., 2003. Catchment and Corals: Terrestrial Runoff to the Great Barrier Reef.
Australian Institute of Marine Science, Townsville, 334 pp.
Grayson, R.B., Argent, R.M., Nathan, R.J., McMahon, T.A. and Mein, R.G., 1996.
Hydrological Recipes. Cooperative Research Centre for Catchment Hydrology,
Melbourne.
Grayson, R.B., Moore, I.D. and McMahon, T.A., 1992. Physically based hydrologic modeling.
2. Is the concept realistic. Water Resources Research, 28(10): 2659-2666.
Hadwen, W.L., Fellows, C.S., Westhorpe, D.P., Rees, G.N., Mitrovic, S.M., Taylor, B.,
Baldwin, D.S., Silvester, E. and Croome, R., 2009. Longitudinal trends in river
functioning: Patterns of nutrient and carbon processing in three Australian rivers.
River Research and Applications, DOI: 10.1002/rra.1321.
Harmel, D., Potter, S., Casebolt, P., Reckhow, K., Green, C. and Haney, R., 2006a.
Compilation of measured nutrient load data for agricultural land uses in the United
States. Journal of the American Water Resources Association, 42(5): 1163-1178.
Harmel, D., Qian, S., Reckhow, K. and Casebolt, P., 2008. The MANAGE Database:
Nutrient Load and Site Characteristic Updates and Runoff Concentration Data.
Journal of Environmental Quality, 37(6): 2403-2406, 10.2134/jeq2008.0079.
47
Harmel, R.D., Cooper, R.J., Slade, R.M., Haney, R.L. and Arnold, J.G., 2006b. Cumulative
uncertainty in measured streamflow and water quality data for small watersheds.
Transactions of the Asabe, 49(3): 689-701.
Harris, G.P., 2001. Biogeochemistry of nitrogen and phosphorus in Australian catchments,
rivers and estuaries: effects of land use and flow regulation and comparisons with
global patterns. Marine and Freshwater Research, 52(1): 139-149.
Hawdon, A., Keen, R.J., Post, D.A. and Wilkinson, S.N., 2008. Hydrological recovery of
rangeland following cattle exclusion, Sediment Dynamics in Changing Environments
IAHS Publ. 325, Christchurch New Zealand, pp. 532-539.
Heathwaite, A.L., 2003. Making process-based knowledge useable at the operational level: a
framework for modelling diffuse pollution from agricultural land. Environmental
Modelling and Software, 18(8-9): 753-760.
Hollinger, E., Cornish, P.S., Baginska, B., Mann, R. and Kuczera, G., 2001. Farm-scale
stormwater losses of sediment and nutrients from a market garden near Sydney,
Australia. Agricultural Water Management, 47: 227-241.
Huber, W.C., 1993. Contaminant transport in surface water. In: D.R. Maidment (Editor),
Handbook of Hydrology. McGraw-Hill, New York.
Hunter, H.M. and Walton, R.S., 2008. Land-use effects on fluxes of suspended sediment,
nitrogen and phosphorus from a river catchment of the Great Barrier Reef, Australia.
Journal of Hydrology, 356(1-2): 131-146.
Inamdar, S.P., O'Leary, N., Mitchell, M.J. and Riley, J.T., 2006. The impact of storm events
on solute exports from a glaciated forested watershed in western New York, USA.
Hydrological Processes, 20(16): 3423-3439, 10.1002/hyp.6141.
Jakeman, A.J., Letcher, R.A. and Norton, J.P., 2006. Ten iterative steps in development and
evaluation of environmental models. Environmental Modelling & Software(602-314).
Jetten, V., de Roo, A. and Favis-Mortlock, D., 1999. Evaluation of field-scale and catchmentscale soil erosion models. Catena, 37(3-4): 521-541.
Johnes, P.J., 2007. Uncertainties in annual riverine phosphorus load estimation: Impact of
load estimation methodology, sampling frequency, baseflow index and catchment
population density. Journal of Hydrology, 332(1-2): 241-258.
Jordan, P., Waters, D. and Cetin, L., 2010. Best Practice Guidelines for Modelling Using
eWater Source Catchments, eWater CRC, Canberra.
Kim, L.H., Kayhanian, M. and Stenstrom, M.K., 2004. Event mean concentration and loading
of litter from highways during storms. Science of the Total Environment, 330(1-3):
101-113.
Kuhnert, P., Bartley, R., Peterson, E., Browne, M., Harch, B., Steven, A., Gibbs, M., KinseyHenderson, A. and Brando, V., 2007. Conceptual and statistical framework for the
water quality component of the integrated report card for the Great Barrier Reef
Catchments. Final Milestone Report to MTSRF Project 3.7.7., CSIRO, Brisbane.
Kuhnert, P., Wang, Y.-G., Henderson, B., Stewart, L. and Wilkinson, S., 2009. Statistical
methods for the estimation of pollutant loads from monitoring data. Final Project
Report. Report to the Marine and Tropical Sciences Research Facility, Reef and
Rainforest Research Centre Limited, Cairns.
Leecaster, M.K., Schiff, K. and Tiefenthaler, L.L., 2002. Assessment of efficient sampling
designs for urban stormwater monitoring. Water Research, 36(6): 1556-1564.
Letcher, R.A., Jakeman, A.J., Merritt, W.S., McKee, L.J., Eyre, B.D. and Baginska, B., 1999.
Review of techniques to estimate catchment exports. NSW EPA Report 99/73,
Sydney.
Ludwig, J.A., Bartley, R., Hawdon, A., Abbott, B. and McJannet, D., 2007. Patch
configuration amplifies loss across scales in a grazed catchment in north-east
Australia. Ecosystems, 10: 839-845, 10.1007/s10021-007-9067-8.
Marsh, N. and Waters, D., 2009a. Comparison of load estimation methods and their
associated error, 18th World IMACS / MODSIM Congress, Cairns, Australia.
48
Marsh, N. and Waters, D., 2009b. Comparison of load estimation methods and their
associated error, 18th IMACS World Congress MODSIM, Cairns,
http://www.mssanz.org.au/modsim09/I4/marsh_I4.pdf.
Marston, F.M., Young, W.J. and Davis, R.J., 1995. Nutrient generation rates data book.
CSIRO Division of Water Resources, Canberra.
McJannet, D. and Fitch, P., 2005. Appendix G- Load calculations for Upper Mossman and
Upper Daintree Automatic monotoring stations for the 2003/2004 wet seaosn. In: D.
McJannet, P. Fitch, B. Henderson, B. Harch and R. Bartley (Editors), Douglas Shire
Water Quality Monitoring Strategy - Final Report. CSIRO Land and Water, Canberra,
pp. 208.
McJannet, D., Fitch, P., Harch, B., Thomas, S. and Bartley, R., 2005. Appendix EPresentation and interpretation of water quality data from automatic monitoring
stations for the 2003/04 wet season. In: D. McJannet, P. Fitch, B. Henderson, B.
Harch and R. Bartley (Editors), Douglas Shire Water Quality Monitoring Strategy Final Report. CSIRO Land and Water, Canberra, pp. 208.
Meade, R.H., 1988. Movement and storage of sediment in rivers systems. In: A. Lerman and
M. Meybeck (Editors), Physical and Chemical weathering in geochemical cycles.
Kluwer Academic Publishers, Netherlands, pp. 165-179.
Michaelides, K., Lister, D., Wainwright, J. and Parsons, A.J., 2009. Vegetation controls on
small-scale runoff and erosion dynamics in a degrading dryland environment.
Hydrological Processes, 23(11): 1617-1630.
Mitchell, A., Reghenzani, J., Faithful, J., Furnas, M. and Brodie, J., 2009. Relationships
between land use and nutrient concentrations in streams draining a 'wet-tropics'
catchment in northern Australia. Marine and Freshwater Research, 60(11): 10971108, 10.1071/mf08330.
Mitchell, V.G., Smolenska, K.A. and Argent, R.M., 1998. Nutrient and Sediment Loads and
generation Rates in the Port Phillip Catchment, Centre for Environmental Applied
Hydrology, University of Melbourne, Melbourne, Australia.
Moreno-de las Heras, M., Nicolau, J.M., Merino-Martín, L. and Wilcox, B.P., 2010. Plot-scale
effects on runoff and erosion along a slope degradation gradient. Water Resour.
Res., 46(4): W04503, 10.1029/2009wr007875.
Moss, A., Brodie, J. and Furnas, M., 2005. Water Quality guidelines for the Great Barrier
Reef World Heritage Area: a basis for development and preliminary values. Marine
Pollution Bulletin, 51: 174-185.
Murray, A.G. and Parslow, J.S., 1999. Modelling of nutrient impacts in Port Phillip Bay - a
semi-enclosed marine Australian ecosystem. Marine and Freshwater Research,
50(6): 597-611.
Nathan, R.J. and McMahon, T.A., 1990. Evaluation of automated techniques for base flow
and recession analysis. Water Resources Research, 26(7): 1465-1473.
Nelson, P.N., Cotsaris, E. and Oades, J.M., 2006. Nitrogen, Phorsphorus, and Organic
Carbon in Streams Draining Two Grazed Catchments. Journal of Environmental
Quality, 25: 1221-1229.
Nistor, C.J. and Church, M., 2005. Suspended sediment transport regime in a debris-flow
gully on Vancouver Island, British Columbia. Hydrological Processes, 19(4): 861-885,
10.1002/hyp.5549.
NLWRA, 2000. Ecosystem Health 2000, National Land and Water Resources Audit,
Canberra.
Parsons, A.J., Brazier, R.E., Wainwright, J. and Powell, D.M., 2006. Scale relationships in
hillslope runoff and erosion. Earth Surface Processes and Landforms, 31(11): 13841393, 10.1002/esp.1345.
Prosser, I., Moran, C., Lu, H., Scott, A., Rustomji, P., Stevenson, J., Priestly, G., Roth, C.H.
and Post, D., 2001a. Regional Patterns of Erosion and Sediment Transport in the
Burdekin River Catchment, Meat and Livestock Australia, Sydney.
49
Prosser, I., Rutherfurd, I., Olley, J.M., Young, W.J., Wallbrink, P. and Moran, C.J., 2001b.
Large-scale patterns of erosion and sediment transport in river networks, with
examples from Australia. Marine and Freshwater Research, 52: 81-99.
Ridley, A.M., Christy, B.P., White, R.E., McLean, T. and Green, R., 2003. North-east Victoria
SGS National Experiment site: water and nutrient losses from grazing systems on
contrasting soil types and levels of inputs. Australian Journal of Experimental
Agriculture, 43(7-8): 799-815, 10.1071/ea02090.
Rohde, K., Masters, B., Fries, N., Noble, R. and Carroll, C., 2008. Fresh and Marine Water
Quality in the Mackay Whitsunday Region 2004/05 to 2006/07, Department of
Natural Resources and Water, Mackay.
Rutherfurd, I., 2000. Some Human Impacts on Australian Stream Channel Morphology. In:
S. Brizga and B. Finlayson (Editors), River Management (The Australasian
Experience). John Wiley and Sons, England, pp. 11-50.
Sansalone, J.J. and Cristina, C.M., 2004. First flush concepts for suspended and dissolved
solids in small impervious watersheds. Journal of Environmental Engineering,
130(11): 1301-1314.
Sansalone, J.J., Hird, J.P., Cartledge, F.K. and Tittlebaum, M.E., 2005. Event-based
stormwater quality and quantity loadings from elevated urban infrastructure affected
by transportation. Water Environment Research, 77(4): 348-365,
10.2175/106143005x51932.
Schilling, K. and Zhang, Y.K., 2004. Baseflow contribution to nitrate-nitrogen export from a
large, agricultural watershed, USA. Journal of Hydrology, 295(1-4): 305-316,
10.1016/j.jhydrol.2004.03.010.
Schwartz, S.S. and Naiman, D.Q., 1999. Bias and variance of planning level estimates of
pollutant loads. Water Resources Research, 35(11): 3475-3487.
Searle, R. and Ellis, R., 2009. Incorporating variable cover in erosion algorithms for grazing
lands within catchment scale water quality models, 18th World IMACS/MODSIM
Congress, Cairns.
Sharpley, A. and Kleinman, P., 2003. Effect of rainfall simulator and plot scale on overland
flow and phosphorus transport. Journal of Environmental Quality, 32(6): 2172-2179.
Silberstein, R.P., 2006. Hydrological models are so good, do we still need data?
Environmental Modelling and Software, 21: 1340-1352.
Smith, V.H., Joye, S.B. and Howarth, R.W., 2006. Eutrophication of freshwater and marine
ecosystems. Limnology and Oceanography, 51(1): 351-355.
Sun, H., Cornish, P.S. and Daniell, T.M., 2001. Turbidity-based erosion estimation in a
catchment in South Australia. Journal of Hydrology, 253(1-4): 227-238.
Visser, F., Roth, C.H., Wasson, R. and Govers, G., 2007. A sediment budget for a cultivated
floodplain in tropical North Queensland, Australia. Earth Surface Processes and
Landforms, 32(10): 1475-1490.
Walling, D.E., 2006. Human impact on land-ocean sediment transfer by the world's rivers.
Geomorphology, 79: 192-216.
Walling, D.E. and Webb, B.W., 1981. The reliability of suspended sediment load data,
Proceedings of the Symposium on Erosion and Sediment Transport Measurement.
IAHS Publication No. 133, Florence, pp. 177-194.
Wilcox, B.P., Breshears, D.D. and Allen, C.D., 2003. Ecohydrology of a resource-conserving
semiarid woodland: Effects of scale and disturbance. Ecological Monographs, 73(2):
223-239.
Wilkinson, S.N., Henderson, A., Chen, Y. and Sherman, B.S., 2004. SedNet: User Guide
Version 2.0.0, CSIRO, Canberra.
Young, W.J., Marston, F.M. and Davis, R.J., 1996. Nutrient Exports and Land Use in
Australian Catchments. Journal of Environmental Management, 47(2): 165-183.
50
10 Appendices
10.1 Appendix A: List of Government water quality data sites
Victoria - http://www.vicwaterdata.net/vicwaterdata/home.aspx
SA - http://e-nrims.dwlbc.sa.gov.au/
TAS - http://water.dpiw.tas.gov.au/wist/ui
ACT - http://incp.environment.act.gov.au/water/index.aspx
NSW - http://waterinfo.nsw.gov.au/
NT
http://www.nt.gov.au/nreta/naturalresources/water/surfacewater/telemeteredsites/index.shtml
QLD http://www.derm.qld.gov.au/environmental_management/water/water_quality_monitoring/ass
essing_water_quality/water_quality_data.html
51
10.2 Appendix B: references used in the water quality data base
Australian Water Technologies, 1997. Pollutant loads to Berowra Creek from Pyes, Trunks
and Waitara Creeks 1995 1997 Prepared for Hornsby Shire Council, Australian
Water Technologies.
Bainbridge, Z. et al., 2006a. Event-based water quality monitoring in the Burdekin Dry
Tropics Region: 2004/05 wet season. ACTFR Report No. 06/01 for the Burdekin Dry
Tropics NRM., Australian Centre for Tropical Freshwater Research, James Cook
University, Townsville.
Bainbridge, Z., Lewis, S. and Brodie, J., 2007. Event-based community water quality
monitoring in the Burdekin Dry Tropics Region: Volume 1: 2006/07 Wet Season
Report, Burdekin Solutions Ltd, Townsville.
Bainbridge, Z. et al., 2006b. Monitoring of sediments and nutrients in the Burdekin Dry
Tropics Region: Interim Report for the 2005/06 wet season. ACTFR Report No. 06/13
for the Burdekin Dry Tropics NRM, Australian Centre for Tropical Freshwater
Research, James Cook University, Townsville.
Bainbridge, Z., Lewis, S., Davis, A. and Brodie, J., 2008. Event-based community water
quality monitoring in the Burdekin Dry Tropics NRM Region: 2007/08 wet season
update. ACTFR Report No. 08/19 for the north Queensland Dry Tropics., Australian
Centre for Tropical Freshwater Research, James Cook University, Townsville.
Bainbridge, Z.T., Brodie, J.E., Faithful, J.W., Sydes, D.A. and Lewis, S.E., 2009. Identifying
the land-based sources of suspended sediments, nutrients and pesticides discharged
to the Great Barrier Reef from the Tully–Murray Basin, Queensland, Australia. Marine
and Freshwater Research, 60(11): 1081-1090.
Bakri, D.A., Rahman, S. and Bowling, L., 2008. Sources and management of urban
stormwater pollution in rural catchments, Australia. Journal of Hydrology, 356: 299–
311.
Barlow, K., 2009. Nitrogen concentrations in soil solution and surface run-off on irrigated
vineyards in Australia. Australian Journal of Grape and Wine Research, 15(2): 131143.
Barlow, K., Nash, D., Hannah, M. and Robertson, F., 2006. The effect of fertiliser and
grazing on nitrogen export in surface runoff from rain-fed and irrigated pastures in
south-eastern Australia. Nutrient Cycling in Agroecosystems, 77: 69-82.
Bartley, R., Corfield, J.P., Hawdon, A.A., Abbott, B.N., Wilkinson, S.N. and Nelson, B., 2010.
Impacts of improved grazing land management on sediment yields, Part I: hillslope
processes. Journal of Hydrology, 389: 237-248, doi:10.1016/j.jhydrol.2010.06.014
Bartley, R., Wilkinson, S.N., Hawdon, A.A., Abbott, B.N. and Post, D.A., 2010. Impacts of
improved grazing land management on sediment yields, Part 2: catchment response.
Journal of Hydrology, 389: 249-259, 10.1016/j.jhydrol.2010.06.014.
Bartley, R., Hawdon, A. and Keen, R., 2007. Sediment and nutrient loads at the Myuna
Gauge in the Bowen Catchment (2006/07), CSIRO, Brisbane.
Bramley, R.G.V. and Muller, D.E., 1999. Water Quality in the Lower Herbert River - CSIRO
Dataset, CSIRO Land and Water.
Carroll, C., Merton, L. and Burger, P., 2000. Impact of vegetative cover and slope on runoff,
erosion, and water quality for field plots on a range of soil and spoil materials on
central Queensland coal mines. Australian Journal of Soil Research, 38: 313-27.
Chiew, F.H.S. and Scanlan, P., 2002. Estimation of pollutant concentrations for EMSS
modelling of the South-East Queensland region, CRC Catchment Hydrology
http://www.catchment.crc.org.au/pdfs/technical200202.pdf, Melbourne.
Chittleborough, D., 1983. The nutrient load in surface waters as influenced by land use
patterns. In: J.W. Holmes (Editor), The effects of change in land use upon water
resources. Water Research Foundation of Australia.
Cogle, A.L. et al., 2000. Natural Resources of the Barron River Catchment 2 - Water Quality,
land use and land management interactions. Q100033, Department of Primary
Industries, Brisbane.
52
Cosser, P.R., 1989. Nutrient Concentration-Flow Relationships and Loads in the South Pine
River, South-eastern Queensland. 1. Phosphorus Loads. Australian Journal of
Marine and Freshwater Research, 40: 613-30.
Costin, A.B., 1980. Runoff and soil nutrient losses from an improved pasture at Ginninderra,
Southern Tablelands NSW. Australian Journal of Agricultural Research, 31: 533-546.
Cullen, P., 1991. Regional catchment management and recieving water quality: The Monkey
creek project, Final report to LWRRDC.
Cullen, P., Farmer, N. and O'Loughlin, E., 1988. Estimating nonpoint sources of phosphous
to lakes. Verh. Internat. Verein. Limnol., 23: 588-593.
Department of Primary Industries and Water, 2010. Waterways monitoring report [electronic
resource]. In: Department of Primary Industries and Water (Editor). State of
Tasmania Hobart, Tasmania.
Drewry, J.J., Newham, L.T.H. and Croke, B.F.W., 2009. Suspended sediment, nitrogen and
phosphorus concentrations and exports during storm-events to the Tuross estuary,
Australia. Journal of Environmental Management, 90: 879-887.
Duncan, H.P., 1999. Urban Stormwater Quality: A Statistical Overview. CRC for Catchment
Hydrology, Clayton, Vic.
Esslemont, G., 2006a. Water Quality Event Monitoring - Baffle Creek (Mimdale) December
1988: Event Summary Load Calculation (WQEM 0634), Department of Natural
Resources and Water.
Esslemont, G., 2006b. Water Quality Event Monitoring - Baffle Creek (Mimdale) February
2003: Event Summary Load Calculation (WQEM 0636), Department of Natural
Resources and Water.
Esslemont, G., 2006c. Water Quality Event Monitoring - Baffle Creek (Mimdale) January
1991: Event Summary Load Calculation (WQEM 0635), Department of Natural
Resources and Water.
Esslemont, G., 2006d. Water Quality Event Monitoring - Baffle Creek (Mimdale) July-August
1973: Event Summary Load Calculation (WQEM 0633), Department of Natural
Resources and Water.
Esslemont, G., 2006e. Water Quality Event Monitoring - Barker-Barambah (Stonelends)
January 1996: Event Summary Load Calculation (WQEM 0637), Department of
Natural Resources and Water.
Esslemont, G., 2006f. Water Quality Event Monitoring - Lower Burnett (Walla Weir) March
1982: Event Summary Load Calculation (WQEM 0640), Department of Natural
Resources and Water.
Esslemont, G., 2006g. Water Quality Event Monitoring - Mary River (Bellbird) February
1995: Event Summary Load Calculation (WQEM 0641), Department of Natural
Resources and Water.
Esslemont, G., 2006h. Water Quality Event Monitoring - Mary River (Miva) June 2005: Event
Summary Load Calculation (WQEM 0652), Department of Natural Resources and
Water.
Esslemont, G., 2006i. Water Quality Event Monitoring -Barker-Barambah (Stonelends) May
1998: Event Summary Load Calculation (WQEM 0638), Department of Natural
Resources and Water.
Esslemont, G., 2006j. Water Quality Event Monitoring -Lower Burnett River (Mt Lawless)
January 1976: Event Summary Load Calculation (WQEM 0630), Department of
Natural Resources and Water.
Esslemont, G., 2006k. Water Quality Event Monitoring – Mary River (Bellbird) February
1999: Event Summary Load Calculation (WQEM 0643), Department of Natural
Resources and Water.
Esslemont, G., 2006l. Water Quality Event Monitoring – Mary River (Bellbird) May 1996:
Event Summary Load Calculation (WQEM 0642) Department of Natural Resources
and Water.
53
Esslemont, G., 2006m. Water Quality Event Monitoring – Mary River (Dagun Pocket)
February 1995: Event Summary Load Calculation (WQEM0644), Department of
Natural Resources and Water.
Esslemont, G., 2006n. Water Quality Event Monitoring – Mary River (Dagun Pocket)
February 1999: Event Summary Load Calculation (WQEM 0645), Department of
Natural Resources and Water.
Esslemont, G., 2006o. Water Quality Event Monitoring – Mary River (Gardners Falls)
February 1995: Event Summary Load Calculation (WQEM 0646), Department of
Natural Resources and Water.
Esslemont, G., 2006p. Water Quality Event Monitoring – Mary River (Gympie) 15-17th
November 2005: Event Summary Load Calculation (WQEM 0647), Department of
Natural Resources and Water.
Esslemont, G., 2006q. Water Quality Event Monitoring – Mary River (Home Park) April 1989:
Event Summary Load Calculation (WQEM 0651) Department of Natural Resources
and Water.
Esslemont, G., 2006r. Water Quality Event Monitoring – Mary River (Home Park) June 1983:
Event Summary Load Calculation (WQEM 0650), Department of Natural Resources
and Water.
Esslemont, G., 2006s. Water Quality Event Monitoring – Tinana Creek (Bauple East) April
1988: Event Summary Load Calculation (WQEM 0657), Department of Natural
Resources and Water.
Esslemont, G., 2006t. Water Quality Event Monitoring – Tinana Creek (Bauple East) June
2005: Event Summary Load Calculation (WQEM 0658), Department of Natural
Resources and Water.
Esslemont, G., 2006u. Water Quality Event Monitoring – Tinana Creek (Tagigan Road) May
1982: Event Summary Load Calculation (WQEM 0659), Department of Natural
Resources and Water.
Esslemont, G., 2006v. Water Quality Event Monitoring – Tinana Creek (Tagigan Road)
November 2005: Event Summary Load Calculation (WQEM 0660), Department of
Natural Resources and Water.
Esslemont, G., 2006w. Water Quality Event Monitoring : Lower Burnett River (Mt Lawless)
April -May1983: Event Summary Load Calculation (WQEM 0632), Department of
Natural Resources and Water.
Evans, K., Moliere, D., Saynor, M., Erskine, W. and Bellio, M., 2004. Baseline suspendedsediment, solute, EC and turbidity characteristics for the Ngarradj catchment,
Northern Territory, and the impact of mine construction, Darwin, NT.
Faithful, J.W., Brodie, J., Hooper, A., Leahy, P., Henry, G., Finlayson, W. and Green, D.,
2007. Plot-scale runoff of nutrients and sediments under varying management
regimes on banana and cane farms in the wet tropics, Queensland, Australia. A
Report to Task 1, Catchment to Reef Project. Australian Centre for Tropical
Freshwater Research (ACTFR) Report No 05/03, James Cook University, Townsville,
Qld.
Fitzroy Basin Association Inc., 2009. Priority Neighbourhood Catchments WaterQuality
Monitoring Program; 2005-09 Water Quality Monitoring Report, Fitzroy Basin
Association Inc.
Fleming, N., Cox, J., He, Y., Thomas, S. and Frizenschaf, J., 2009. Analysis of constituent
concentrations in the Mount Lofty Ranges for modelling purposes.
Forbes, G.G. and Birch, P.B., 1987. Nutrient loading into the Peel Harvey Estuary from the
Harvey, Murray and Serpentine Rivers (1977 - 1984), Environmental Protection
Authority, Perth, W.A.
Hawdon, A.A., Keen, R.J., Post, D.A. and Wilkinson, S.N., 2008. Hydrologic Recovery of
Rangeland following Cattle Exclusion, Sediment Dynamics in Changing
Environments. IAHS Publ., Christchurch, New Zealand.
54
Hollinger, E., Cornish, P.S., Baginska, B., Mann, R. and Kuczera, G., 2001. Farm-scale
stormwater losses of sediment and nutrients from a market garden near Sydney,
Australia. Agricultural Water Management, 47: 227-241.
Holz, G.K., 2010. Sources and processes of contaminant loss from an intensively grazed
catchment inferred from patterns in discharge and concentration of thirteen analytes
using high intensity sampling. Journal of Hydrology, 383: 194-208.
Hopmans, P. and Bren, L.J., 1999. Water Quality and Sediment Transport of Radiata Pine
Plantation and Eucalypt Forest Catchments in Victoria. In: J. Croke and P. Lane
(Editors), Proceedings of the Second Forest Erosion Workshop, Warburton, Victoria.
Hunter, H.M. and Walton, R.S., 2008. Land-use effects on fluxes of suspended sediment,
nitrogen and phosphorus from a river catchment of the Great Barrier Reef, Australia.
Journal of Hydrology, 356: 131-146.
Kernohan, A.K. and Townsend, S.A., 2000. Runoff water quality from an urban catchment
(Karama, Darwin), Resource Management Branch, Natural Resources Division,
Department of Lands, Planning and Environment. Report NR 2000/11.
Lane, P.N.J., Sheridan, G.J., Noske, P.J. and Sherwin, C.B., 2008. Phosphorus and nitrogen
exports from SE Australian forests following wildfire. Journal of Hydrology, 361: 186198.
Lewis, S., Bainbridge, Z., Brodie, J., Butler, B. and Maughan, M., 2008. Water quality
monitoring of the Black Ross Basins: 2007/08 Wet Season., ACTFR, Townsville.
Masters, B., Rohde, K., Gurner, N., Higham, W. and Drewry, J., 2008. Sediment, nutrient
and herbicide runoff from canefarming practices in the Mackay Whitsunday region: a
field-based rainfall simulation study of management practices, Queensland
Department of Natural Resources and Water for the Mackay Whitsunday Natural
Resource Management Group, Australia.
Mathers, N.J. and Nash, D.M., 2009. Effects of tillage practices on soil and water
phosphorus and nitrogen fractions in a Chromosol at Rutherglen in Victoria, Australia.
Australian Journal of Soil Research, 47: 46-59.
McCourt, M., 1983. Surface water drainage and land use patterns. In: J.W. Holmes (Editor),
The effects of changes in land use upon water resources: proceedings of a
symposium held at the Australian Mineral Foundation, Adelaide, on 5th May. The
Water Research Foundation of Australia, Adelaide, Australia.
McJannet, D., Fitch, P., Henderson, B., Harch, B. and Bartley, R., 2005. Douglas Shire
Water Quality Monitoring Strategy - Final Report.
http://www.clw.csiro.au/publications/consultancy/2005/douglas_shire_water_quality_r
eport.pdf, CSIRO Land and Water, Canberra.
McKergow, L.A., Prosser, I.P., Grayson, R.B. and Heiner, D., 2004. Performance of grass
and rainforest riparian buffers in the wet tropics, Far North Queensland. 2. Water
Quality. Australian Journal of Soil Research, 42(4): 485-498.
McKergow, L.A., Weaver, D.M., Prosser, I.P., Grayson, R.B. and Reed, A.E.G., 2003. Before
and after riparian management: sediment and nutrient exports from a small
agricultural catchment, Western Australia. Journal of Hydrology, 270: 253-272.
Mitchell, A., Reghenzani, J., Faithful, J., Furnas, M. and Brodie, J., 2009.
Relationships between land use and nutrient concentrations in streams
draining a 'wet-tropics' catchment in northern Australia. Marine and
Freshwater Research, 60(11): 1097-1108, 10.1071/mf08330.
Mitchell, C., Brodie, J. and White, I., 2005. Sediments, nutrients and pesticide residues in
event flow conditions in streams of the Mackay Whitsunday Region, Australia. Marine
Pollution Bulletin, 51: 23-36.
Mitchell, V.G., Smolenska, K.A. and Argent, R.M., 1998. Nutrient and Sediment Loads and
generation Rates in the Port Phillip Catchment, Centre for Environmental Applied
Hydrology, University of Melbourne, Melbourne, Australia.
Nash, D. and Murdoch, C., 1997. Phosphorus in runoff from a fertile dairy pasture. Australian
Journal of Soil Research, 35: 419-29.
55
Nelson, P.N., Cotsaris, E. and Oades, J.M., 2006. Nitrogen, Phorsphorus, and Organic
Carbon in Streams Draining Two Grazed Catchments. Journal of Environmental
Quality, 25: 1221-1229.
Noble, R.M., Duivenvoorden, L.J., Rummenie, S.K., Long, P.E. and Fabbro, L.D., 1997.
Downstream effects of land use in the Fitzroy Catchment, The State of Queensland,
Department of Natural Resources.
O'Reagain, P.J. et al., 2005. Nutrient loss and water quality under extensive grazing in the
upper Burdekin river catchment, North Queensland. Marine Pollution Bulletin, 51: 3750.
Packett, R., 2006a. Water Quality Event Monitoring - Fitzroy Catchment April 2006: Event
Summary Load Calculation (WQEM 0607), Department of Natural Resources and
Water.
Packett, R., 2006b. Water Quality Event Monitoring - Fitzroy Catchment January/February
2005: Event Summary Load Calculation (WQEM 0605), Department of Natural
Resources and Water.
Packett, R., Dougall, C., Rohde, K. and Noble, R., 2009. Agricultural lands are hot-spots for
annual runoff polluting the southern Great Barrier Reef Lagoon. Marine Pollution
Bulletin, 58: 976-986.
Padovan, A.V., 2001a. Catchment load monitoring during the 2000/01 wet season (Berry
Creek and Winnellie drain stations), Resource Management Branch, Natural
Resources Division, Department of Infrastructure, Planning and Environment.
Padovan, A.V., 2001b. The quality of run-off and contaminant loads to Darwin Harbour,
Resource Management Branch, Natural Resources Division, Department of Lands,
Planning and Environment.
Padovan, A.V., 2002. Catchment load monitoring during the 2001/02 wet season (Berry
Creek, Elizabeth River and Bees Creek Stations), Department of Infrastructure,
Planning and Environment, Darwin, NT.
Rattray, D.J., Freebairn, D.M., Gurner, N.C. and Huggins, J., 2005. Predicting Natural
Resource Management Impacts utilising Risk Assessment, Prioritisation and
Understanding Process (RAPUP), Department of Natural Resources & Mines,
Brisbane.
Ridley, A.M., Christy, B.P., White, R.E., McLean, T. and Green, R., 2003. North-east Victoria
SGS National Experiment site: water and nutrient losses from grazing systems on
contrasting soil types and levels of inputs. Australian Journal of Experimental
Agriculture, 43(7-8): 799-815.
Rohde, K., Masters, B., Fries, N., Noble, R. and Carroll, C., 2008. Fresh and Marine Water
Quality in the Mackay Whitsunday Region 2004/05 to 2006/07, Department of Natural
Resources and Water, Mackay.
RossRakesh, S., Gippel, C., Chiew, F. and Breen, P., 1999. Blackburn Lake discharge and
water quality monitoring program : data summary and interpretation, CRC for
Catchment Hydrology, Clayton, Vic.
Schofield, N.J. and Birch, P.B., 1986. Catchment management measures to reduce riverine
phosphorus inputs to the eutrophic Peel-Harvey Estuary, Western Australia,
Hydrology and Water Resources Symposium, pp. 113-118.
Silburn, D.M., 1994. Mt. Mort Field Visit - February 1994. In: Field Trip Notes - Eastern
Darling Downs and Lockyer Valley, Second International ‘Symposium on Sealing,
Crusting, Hardsetting Soils: Productivity and Conservation’, 7-11 February 1994,
University of Queensland, Brisbane.
Silburn, D.M., Carroll, C., Ciesiolka, C.A.A., deVoil, R.C. and Burger, P., In Review. Hillslope
runoff and erosion on duplex soils in grazing lands in semi-arid central Queensland I.
Influences of cover, slope and soil. Australian Journal of Soil Research.
Snowy Mountains Hydro-Electric Authority, 1971. Water quality records of the Snowy
Mountains Region, Australia.
Tan, H.T., 1991. Nutrients in Perth urban surface drainage catchments characterised by
applicable attributes, Water Authority of Western Australia, Perth.
56
Townsend, S.A., 1992. Nutrient suspended solid and metal inputs, from point and non-point
sources, into Darwin Harbour, November 1990 – October 1991, Water Resources
Division, Power and Water Authority.
Townsend, S.A. and Douglas, M.M., 2004. The effect of a wildfire on stream water
qualityand catchment water yield in a tropical savanna excluded from fire for 10 years
(Kakadu National Park, North Australia). Water Research, 38: 3051-3058.
Vink, S., Ford, P.W., Bormans, M., Kelly, C. and Turley, C., 2007. Contrasting nutrient
exports from a forested and an agricultural catchment in south-eastern Australia.
Biogeochemistry, 84: 247-264.
Waters, D., 2006a. Water Quality Event Monitoring - Balonne River (Barrackdale) November
1995: Event Summary Load Calculation (WQEM 0609), Department of Natural
Resources, Mines and Water.
Waters, D., 2006b. Water Quality Event Monitoring - Balonne River (Surat) December 2004:
Event Load Summary Load Calculation (WQEM 0601), Department of Natural
Resources, Mines and Water.
Waters, D., 2006c. Water Quality Event Monitoring - Balonne River (Surat) June 2004: Event
Load Summary Load Calculation (WQEM 0602), Department of Natural Resources,
Mines and Water.
Waters, D., 2006d. Water Quality Event Monitoring - Bungil Creek (Roma) Nov 2000: Event
Load Summary Load Calculation (WQEM 0608), Department of Natural Resources,
Mines and Water.
Waters, D., 2006e. Water Quality Event Monitoring - Condamine River (Brigalow) November
1995: Event Summary Load Calculation (WQEM 0611), Department of Natural
Resources, Mines and Water.
Waters, D., 2006f. Water Quality Event Monitoring - Condamine River (Chinchilla) December
2005: Event Summary Load Calculation (WQEM 0619), Department of Natural
Resources, Mines and Water.
Waters, D., 2006g. Water Quality Event Monitoring - Condamine River (Warwick) February
2001: Event Summary Load Calculation (WQEM 0606), Department of Natural
Resources, Mines and Water.
Waters, D., 2006h. Water Quality Event Monitoring - Condamine River (Warwick) June/July
2004: Event Summary Load Calculation (WQEM 0603), Department of Natural
Resources, Mines and Water.
Waters, D., 2006i. Water Quality Event Monitoring - Dogwood Creek (Miles) December
2005: Event Summary Load Calculation (WQEM 0621), Department of Natural
Resources, Mines and Water.
Waters, D., 2006j. Water Quality Event Monitoring - Event Summary Load Calculation
Charley’s Creek (Chinchilla) December 2005 (WQEM 0618), Department of Natural
Resources, Mines and Water.
Waters, D., 2006k. Water Quality Event Monitoring - Maranoa River (Cashmere) February
1997: Event Summary Load Calculation (WQEM 0610), Department of Natural
Resources, Mines and Water.
Waters, D., 2006l. Water Quality Event Monitoring - Maranoa River (Mitchell)
January/February 2001: Event Summary Load Calculation (WQEM 0607),
Department of Natural Resources, Mines and Water.
Waters, D., 2006m. Water Quality Event Monitoring - Warrego River (Cunnamulla) April
2006: Event Summary Load Calculation (WQEM 0620), Department of Natural
Resources, Mines and Water.
Waters, D., 2006n. Water Quality Event Monitoring - Weir River (Talwood) January 2006:
Event Summary Load Calculation (WQEM 0624), Department of Natural Resources,
Mines and Water.
Waters, D., 2006o. Water Quality Event Monitoring - Weir River (Talwood) July 2005: Event
Summary Load Calculation (WQEM 0623), Department of Natural Resources, Mines
and Water.
57
Waters, D., 2006p. Water Quality Event Monitoring - Yuleba Creek (Bruce Highway)
December 2005: Event Summary Load Calculation (WQEM 0625), Department of
Natural Resources, Mines and Water.
Waters, D., 2006q. Water Quality Event Monitoring -Surat Gauging Station (Surat)
December 2005: Event Summary Load Calculation (WQEM 0622), Department of
Natural Resources, Mines and Water.
Waters, D., 2008. Water Quality Event Monitoring – Warrego Catchment December 2007:
Event Load Calculation Summary (WQEM 0802), Department of Natural Resources,
Mines and Water.
Wilkinson, S. et al., 2008. Ecological Monitoring of the Townsville Field Training Area (TFTA)
2007/08, CSIRO Land and Water and ACTFR Client Report to Department of
Defence.
Wood, G., 1986. Mt Lofty Ranges Watershed: impact of land use on water quality and
implications for reservoir water quality management, Environment and Water Supply
Department South Australia, Adelaide.
10.3 Appendix C: Recommendations for using the Excel data base created for
this study
Harmel et al., (2006a) developed a similar data base for sites in the USA and in a
subsequent document (Harmel et al., 2008) he describes the MANAGE data base and its
application. In the first instance, we aim to make the Google Earth kml file available to
eWater partners. Further discussion is required about the arrangements under which the
data base would be made available to non eWater partners as well as how it is supported
and maintained into the future.
58