Water quality plays a major role in the health of organisms and ecosystems throughout the world. High quality water is essential for the support of life and the environment, whereas poor quality can cause the degradation of water bodies and consequently the flora and fauna that are reliant upon them. Hence, the quality of water, being the physical, chemical or biological parameters of the liquid, is a problematic issue and of concern globally (Sumok, 2001; Khan & Anasari, 2005; Smith, 2003; Suthar, et al. 2009; Rather, et al., 2016). Many studies have found that a major contributor to the decline in water quality across the globe is change in land use, particularly urbanisation (Sumok, 2001; Allan, 2004; Foley, et al., 2005; Suthar, et al., 2009; Miserendino, et al., 2010; Zhao, et al., 2014). Urbanisation of the watershed, otherwise known as the development of urban or agricultural land next to a body of water, elevates nutrients, bacteria loading, heavy metals, and suspended solids and reduces biodiversity in adjacent waterways. This contributes to the degradation of urban streams and rivers, otherwise known as the urban stream syndrome (Schoonover, Lockably & Pan, 2005; Walsh, et al., 2005; Elliot & Trowsdale, 2006; Dietz & Clausen, 2007; Haltstead, et al., 2014; Chen, et al., 2016). An increase in catchment imperviousness as a result of urbanisation is one source of the elevation of pollutant loading in waterways (Hatt, et al., 2004). Impervious surface area is increased through development with the construction of roads, paved areas and roofs, thus increasing stormwater overland flow and reducing baseflow (Hatt, et al., 2004; Schoonover, Lockably & Pan, 2005; Xiana, Craneb & Suc, 2007). Additionally, the increase in stormwater discharge is compounded by proficient drainage systems, the removal of vegetation and riparian zones, as well as the compaction of soils (Hatt, et al., 2004; Elliot & Trowsdale, 2006; Jacobson, 2011). As a result, the variation in both runoff and base flow can influence the volume of pollutants reaching rivers and streams (Hatt, et al., 2004; Rather, et al. 2016). Contaminants, both suspended and dissolved, gather on impervious surfaces, as a result surface runoff transports (either by overland flow or through piped stormwater systems) non-point source pollutants to water bodies instead of infiltrating into soil (Xiana, Craneb & Suc, 2007; Jacobson, 2011). Increased impervious area is not the only source of pollutants, urbanisation is the form of the cultivation of agricultural land through deforestation and the use of pesticides has been shown to increase nutrients in surface flow resulting in the eutrophication of waterways globally (Smith, 2003; Rather, et al., 2016). Pollution of waterways due to urbanisation may also come in the form of point source pollutants such as in areas where wastewater or sewage are inappropriately contained and managed (Schoonover, Lockably & Pan, 2005; Xiana, Craneb & Suc, 2007). As discussed above, the water quality in an urban setting is influenced by a multitude of the pollutant factors. A major cause of the increase in pollutants in urban rivers and streams can be attributed to land use (Foley, et al., 2005; Miserendino, et al., 2010). Urbanisation and development of watersheds can have a diverse impact on urban water bodies. If pollutants aren’t controlled they may cause a health risk to humans, loss of biotic environments and deprecation in recreational space (Schoonover, Lockably & Pan, 2005).
Globally there have been a large number of studies investigating the impacts of land use and urbanisation on water quality (Xiana, Craneb & Suc, 2007; Miserendino, et al., 2010; Halstead, et al. 2014; Rather, et al., 2016). The studies looked at the correlations between water quality and urban development as well as agricultural land use (Schoonover, Lockably & Pan, 2005; Halstead, et al. 2014). The multitude of these studies found that there is a link between land use and the reduction of river water quality, in particular physical, chemical and biological water quality indicators have all shown to be directly impacted by land use (Schoonoverb, Lockably & Pan, 2005; Rather, et al., 2016). For example, Schoonover, et al, (2005) observed that physical solids were found to have increased in concentration levels when waterways were within urban watersheds. Moreover, Almeida, et al, (2007) studied the impacts of urbanisation on rivers, finding significant degradation of chemical parameters downstream of urban areas after a five-year period. Another study by Miserendino, et al, (2010) found that the change in land use of a watershed has a direct impact on reducing biodiversity and increasing nutrient loading of waterways with urban landscapes being the most significant course of aquatic life degradation. Furthermore, reduction of water quality has been most prominent in cities with rapid expansion rates in both urban and industrial development, leading to increased impervious surface area (Wenhui, 2012). The increase of built up surface area has meant that additional non-point pollution has been known to dramatically decrease water quality by amplifying the rivers pollution (Kuang, 2012). Likewise, agriculture has in most cases been associated with an increase in non-point source pollutants, which can be attributed to the nutrients from farm use (Haidary, et al., 2012). However, recent studies have shown that this may be incorrect and agricultural use may reduce the amount of pollutants, when in the same catchment as other urban areas (Zhao, et al., 2014).
In recent years, integrated water resources management (IWRM) has been implemented in both developed and developing countries (Lubell & Edelenbos, 2013). These treatment methods have been proven to be effective globally, where stormwater is treated before discharge into the waterway. Zhao, et al., (2014) conducted a case study on a rectivular river network located in Shanghai, China and found that pollutants in the waterway have decreased since the implementation of IWRM. This was the case even when the amount of urban area had increased. Water sensitive urban design (WSUD) has been introduced in Australia to incorporate new IWRM systems, including bioretention systems and wetlands (Kazemig, et al., 2009). The introduction of these systems as micro-habitats, has led to an increase of species and diversity (Kazemi, et al., 2009). WSUD not only treats stormwater before discharge into waterways, but it also reduces the impact of storm events on waterways due to stormwater detention times in treatment systems (Roy, et al., 2007). A reduction in discharge of frequent stormwater events has meant ecosystems and the landscape surrounding the waterways is relatively unaffected (Roy, et al., 2007).
The majority of research focused on investigating the impact of land use on water quality is either, restricted in the number of parameters tested or doesn’t have great temporal variability. Furthermore, a large proportion of the research has been conducted outside of Australia (Singh et al., 2003; Parinet, Lhote & Legude, 2004; Schoonover, Lockably & Pan, 2005; Alberto, et al., 2006; Shrestha & Kazma, 2006; Bhat, et al., 2014; Zhao, et al., 2015; Rather, et al,. 2016). For example research by Rather, et al., (2016) into land use changes and their impact on the Jhelum River in the Kashmir Valley, India collected 15 water quality parameters from 12 sites, indicating that the study has good spatial variation and tested a wide range of contaminants. However the water quality samples were only collected over a 12-month period indicating poor temporal variation. Another study conducted by Schoonoverb, Lockably & Pan, (2005) looked at how a range of land uses (urban, developing, pasture, managed forest, unmanaged forest) of 16 watersheds impacted on chemical, biological and physical indicators of water quality in streams feeding the Chattahoochee River in Georgia, USA. Their study used 11 water quality indicators collected from rivers across the 16 watersheds. The data was collected in the winter and spring months between 2002-2004. This resulted in 2 years worth of data that was later split into two, one-year data sets for model calibration and validation. Again this study has good spatial variation with a reasonable amount of water quality indicators tested. However, it lacks temporal variation in the data set and as the data was only collected in the winter and spring months it most likely doesn’t represent fluctuations that may occur during summer and autumn months. Zhao, et al., (2015) also investigated the impact of land use on water quality. Their study focused on the Yangtze River Delta, China, with a total of 48 sites analysed for 7 water quality indicators over a 12-month period. This study again has great spatial variation, but only focuses on a small quantity of water quality indicators and again lacks in temporal variation with a data set only spanning 12 months. There are some studies that have reasonable temporal & spatial data sets, with a decent amount of water quality parameters tested. For example, Shrestha & Kazma, (2006) conducted a water quality study in the Fuji River Basin in Japan that tested 12 parameters at 13 sites over 8 years. Although, this study like many was conducted outside of Australia (Singh et al., 2003; Parinet, Lhote & Legude, 2004; Alberto, et al., 2006; Bhat, et al., 2014).
Many studies evaluating the impact of WSUD in minimising the impacts of urbanisation tend to lack spatial variability as they only look at a site-by-site basis. Few studies have evaluated the influence that multiple WSUD sites along an urban river have had on minimising the overall impact of new development. The research also tends to only report on limited water quality parameters over short time spans (Larm, 1999; Kovacic, et al., 2006; Dietz & Clausen, 2007). For example, Kovacic, et al., (2006) looked at the effectiveness of two experimental agricultural runoff wetlands at minimising non-point pollutants entering Lake Bloomington, USA. The wetlands were designed as an experiment to see if they could mitigate contaminants reaching the lake, which is used for drinking water. The study measured 3 nutrients, precipitation and inlet and outlet flow over a period spanning from April 1998 to December 1999 (21 months). They found that the wetlands were successful at reducing nutrient levels in the lake, particularly nitrogen levels. However, the study only focused on a small number of water quality parameters measured over a relatively short timespan. The two wetlands were also constructed within close proximity to each other draining small watersheds. Therefore it can be said that the study lacked spatial and temporal variability, it also only focused on agricultural land use. Another study conducted by Larm, (1999) evaluated the impact of a number of storm water treatment facilities constructed to minimise pollutants running of an urbanised watershed adjacent to a eutrophic lake in Sweeden called Lake Orlangen. The facilities consisted of an oil separation pond, pre-sedimentation pond, precipitation pond, constructed wetlands and a number of open ditches. The study had access to a wide range of variables including, inlet and outlet flow, water and air temperature, nutrients, bacteria, metals, pH and conductivity measured from October 1995 to August 1998. The air and water temperature, Total Phosphorous (TP), Total Nitrogen (TN), pH and conductivity were measured monthly and all other parameters four times a year. Sediment samples were also tested within the facilities yearly. However, despite this large data set the study selected to focus on a limited data set of TN, TP and three metal pollutants (Zn, Pb and Cu) as they were considered to have the largest impact on the lake. Again, the study was conducted over a relatively short period of time and only focuses on one treatment site that connects to a 9.2 km2 watershed, relatively small compared to the total watershed that drains into the lake which is 40 km2, thus the spatial and temporal variability of the data isn’t great.
Considering the current literature, the aim of this research is to understand the impact both sustained urbanisation and new developments are having on the water quality of an urban stream. Additionally the research aims to evaluate the success of WSUD techniques used in new developments at minimising the effect of urbanisation on river water quality. In order to achieve this an urban river data set spanning 20 years with approximately 19 water quality parameters will be analysed with the aim of answering the following questions:
1. How has land use distribution impacted the water quality of an urban river?
2. How has prolonged urbanisation over multiple decades impacted an urban river, what are the temporal trends?
3. Has the introduction of WSUD in new developments reduced the impact of urbanisation on water quality in an urban stream?
4. Have healthy river guidelines limits been breeched due to the impact of urbanisation?
As highlighted previously much of the current research is lacking in long-term data sets. Furthermore, more research is needed into the effectiveness of WSUD techniques on a whole river basis. For these reasons the research conducted in this paper will be unique. The results of this study will outline in more detail the impacts of prolonged urbanisation on a river body and may be useful in refining WSUD techniques in the better management of watersheds and rivers.
2. METHODS
2.1 DATA COLLECTION
The urban river selected for the analysis is called the Plenty River and is located in the north-eastern suburbs of Melbourne. A data set spanning for the last 20 years has been obtained from Melbourne Water. This data provides up to 19 physical, biological and chemical water quality indicators, including metals, nutrients and other conventional variables. This data has been collected at multiple sites along Plenty River, with all sites being located in urban areas before the confluence with Yarra River. All data that is received from Melbourne Water has been lab tested; therefore it is regarded as being of high quality.
A second set of data will also be obtained from Water Watch Victoria, who are a group of volunteers that visit sites across Victoria monthly and record a range of water quality parameters. Although the group conducts tests at multiple locations along Plenty River, it is limited to only testing for conventional variables such as EC, pH and turbidity, as well as phosphate, dissolved oxygen and macro-invertebrates. As the amount of different tests are limited and not lab tested, this data will be verified before it can be used.
NearMaps, a website that provides aerial imagery will be used to gather aerial photos of the Plenty River Catchment. NearMaps provides images dating back to October 2009, which can be used to map the change in land use around the Plenty River particularly the areas of new development.
2.2 DATA ANALYSIS
Four test sites will be selected along the Plenty River for analysis. Water quality indicators will be separated into three groups nutrients, metals and standard indicators. Multivariate statistical analysis will be used to analysis both spatial and temporal trends over the last 20 years. The exact statistical test to be performed on the data set is yet to be decided. The results from the analysis will demonstrate any significant trends and drivers of water quality over the last 20 years. This data will be used to determine how the spatial distribution of land between sites has impacted the river. The analysis will also show when changes in water quality have occurred and the impact of sustained urbanisation in the lower catchment area. The temporal analysis will be compared to satellite imagery of the catchment to evaluate the impact of new development on river quality in the upper catchment, in particular to see if WSUD is minimising the impact.
2.3 RESOURCES REQUIRED
The project is a desktop study, data has already been collected, for this reason we will not require water-testing equipment. Although a trip with WaterWatch to see how their data is collected is planned. Excel will be required to conduct the multivariate analysis of the data, this may require additional software.