Few elements in surface waters are monitored as closely as phosphorus (P). Phytoplankton growth accelerates when water bodies such as lakes, reservoirs, and estuaries, become loaded with P, which element is often a limiting nutrient for growth of aquatic organisms (Drever, 1997). Eutrophication degrades water quality directly due to toxic algal blooms, increased turbidity, shifts in aquatic species, and increased costs in water treatment (Foy, 2005). Ultimately, eutrophication cycles into the decomposition of algae and the severe depletion of dissolved oxygen, causing a loss of biodiversity at all trophic levels (Carpenter, et al., 1998).
Many environmental protection plans include goals to reduce annual P loads exported from watersheds in efforts to limit the eutrophication of downstream water bodies. Metrics such as the Total Maximum Daily Load (TMDL) are calculated to set P load limits from all sources within a watershed and include an added margin of safety. The success of an environmental protection plan in limiting P exports from a watershed depends on locating and correctly characterizing critical source areas associated with instream P loads.
Sources of P fall into two general categories: point sources and nonpoint sources. A point source of P comes from a “clearly identifiable point of discharge”, such as dissolved P in the effluent piped out of a wastewater treatment plant (Pierzynski, et al., 2005). On the contrary, nonpoint sources of P are trickier to assess and control, mainly because of the diffuse nature and often widespread area of the source, such as eroded streambank materials (Carpenter, et al., 1998). Relatedly, a critical source area (CSA) is often identified where a nonpoint source overlaps with a hydrologically active area, resulting in disproportionately high amounts of mobilized P (Pionke, et al., 2000).
Locating a CSA allows for additional investigation into characteristics of the P source, including:
- Dominant forms of P
- Nature of the source (whether naturally occurring or anthropogenic)
- Possible pathways and processes through which the source mobilizes to nearby surface waters
- Quantitative estimates of P loading rates.
These insights help prioritize the goals of environmental protection plans by informing decisions made to limit P mobility from CSAs. However, locating and characterizing CSAs can be challenging in mixed-use watersheds with wide ranges of landscape characteristics. To locate and characterize CSAs in such a watershed requires one to understand the following:
- Spatial distribution of P sources in the watershed
- Temporal dynamics that affect source mobility
- Chemical forms of P from those sources and associated transfer pathways and processes.
1.1 Synoptic Sampling
Synoptic sampling is a proven strategy for capturing the spatial variation of water quality in surface waters across a study area or watershed that is not outfitted with numerous, automated monitoring stations (Banks and Palumbo-Roe, 2010, Grayson, et al., 1997). In its most basic form, synoptic sampling offers a summary of the flow rate and constituent concentration at a high density of sites across a study area at a single point in time and in a steady hydrologic state. Synoptic sampling has been most commonly used in baseflow conditions when streamflow is not influenced by recent flow events and constituent concentrations are potentially highest (Grayson, et al., 1997, Kimball, et al., 2002). The simple product of the flow rate and constituent concentration at a site provides an instantaneous load estimate.
After load estimates are calculated at all synoptic sampling sites, one gains a “spatially intensive snapshot” of constituent loading across the study area (Kimball, et al., 2002). Increases of instream constituent load readily reveal source areas. In this way, synoptic sampling is more effective in assessing constituent spatial loading patterns than either routine or event monitoring strategies (Eyre and Pepperell, 1999). Synoptic sampling has been applied in water quality studies, from sourcing heavy metal loads associated with acid mine drainage (Banks and Palumbo-Roe, 2010, Kimball, et al., 2002, Runkel, et al., 2013) to correlating land use changes with observed water quality parameters (Cox, et al., 2013, Eyre and Pepperell, 1999, Wayland, et al., 2003).
While useful for assessing water quality spatially, synoptic sampling is not commonly used as a strategy for capturing the temporal dynamics of instream P loads. The potential value of repeating synoptic sampling under different hydrologic conditions has been demonstrated. Wayland used temporally repeated synoptic sampling in baseflow conditions over two years in order to correlate certain water quality parameters with land use, but did not measure P (Wayland, et al., 2003). Banks used temporally repeated synoptic sampling to discriminate between point and nonpoint sources of zinc in a mining catchment. (Banks and Palumbo-Roe, 2010).
To forego repeated synoptic sampling may be acceptable if research is focused only on a single transfer pathway, as is the case with assessing subsurface transport of soluble metals associated with acid mine drainage, which can be assessed once in baseflow conditions. However, at the watershed scale and in a mixed landscape, a constituent such as P originates from numerous source areas, being mobilized through different pathways depending on annual hydrologic cycles and temporal changes in land use and cover (Harmel and Haggard, 2006, Jordan, et al., 2012, Pionke, et al., 1999). For example, in a predominately snow-fed watershed, the spring snowmelt period may last for many weeks, during which P is mobilized through snowmelt pathways, such as runoff over frozen soils. Later in the dry season, P transfer may shift to primarily subsurface pathways or to irrigation induced erosion. For reasons like these, temporally repeated synoptic sampling is essential for locating and characterizing CSAs of P in a mixed-use watershed.
1.2 Phosphorus Fractionation
When a CSA is located, classifying the forms of P helps further characterize the source. This is because certain forms of P are more likely to be mobilize to surface waters through certain processes and pathways. Knowing which forms of P dominate at different sites under different hydrologic conditions and land uses can in turn guide management strategies that limit those P exports.
In water quality monitoring, P is commonly reported as “Total P” (TP) because of the strong correlation in surface water bodies between mean TP concentration and chlorophyll a, an indicator of algal growth (Forsberg and Ryding, 1980, Prairie, et al., 1989, Smith, 1998). Total P represents P in all compounds and forms: both organic and inorganic P, both dissolved P and P associated with suspended particulates (Espinosa, et al., 1999, Haygarth, et al., 1997).
Total P in a water sample can be divided into general P fractions depending on a chosen combination of sample filtration and analytical method (Haygarth and Sharpley, 2000). These P fractions include: total dissolved P (TDP), dissolved reactive P (DRP), dissolved organic P (DOP), and particulate P (PP). Total dissolved P represents all P in solution, both inorganic and organic, since the samples are filtered prior to analysis. Particulate P (PP) can be derived by subtracting TDP from TP, and represents P associated only with particulates suspended in the water. Dissolved reactive P is a fraction of TDP, consisting of inorganic phosphates which are almost completely bioavailable for algal growth (Lee, et al., 1980, Reynolds and Davies, 2001). Dissolved organic P (DOP) is another fraction of TDP. It is derived by subtracting DRP from TDP, and consists of the non-reactive, organic P fraction in solution.
By analyzing these general forms of P at each synoptic sampling site, one can gain meaningful insight into the nature of upstream CSAs. For example, an increased PP load between two sites would indicate an increase in suspended sediment load, as the PP fraction does not include any dissolved P forms. Particulate P originates from sediment sources associated with either local stream channel erosion or surface runoff between the two sites. Since the increase of PP must have been driven by flowing water, the detachment and mobilization of those particles would have been related to soil saturation and hydrologic conditions (Pierzynski, et al., 2005). If, for example, the increase in PP load occurred during baseflow conditions, it follows that PP was most likely mobilized through overland return flows associated with irrigation or from eroded streambank material, and not from a rain event.
With this sort of information, one could identify the stream segment between the two sites as being associated with a source area of sediment-bound P. This assessment, however, would be limited by a lack of data on PP transport during non-baseflow conditions. Thus, it is essential to repeat synoptic sampling with P fractionation under different hydrologic conditions. If the stream segment showed similar increases in PP loads under different hydrologic conditions, the stream segment and associated areas could be classified as a CSA. By locating and characterizing this CSA, a more effective investigation could be conducted in the future to determine if the sediment was primarily originating from stream channels or from nearby fields, preparing the way for best management practices that limit PP transport and improve water quality.
The objective of our study was to combine repeated synoptic sampling with phosphorus fractionation into a unique strategy for locating and characterizing critical source areas of P. We evaluated this strategy in a montane, mixed-use watershed, based the analysis of five P fractions in surface waters collected during seasonally repeated and annually replicated synoptic sampling campaigns.
2.1 Study Area
The Wallsburg watershed (184 km2) is a subwatershed of the Provo River Basin, located within the Wasatch Mountains of north central Utah, USA. This watershed drains into the Deer Creek Reservoir, a major municipal water source that was classified as an impaired, eutrophic waterbody in the 1980’s. Despite substantial water quality improvements in the Deer Creek Reservoir, there is an ongoing effort to further limit P inputs (WCD, 2012; documents attached, right). To protect reservoir health, TMDLs were calculated by state officials, requiring that total P and total dissolved P concentrations of reservoir-feeding surface waters are limited to 0.05 mg/L and 0.025 mg/L, respectively.