Volumes 521–522, 15 July 2015, Pages 388–399
Environmental response of an Irish estuary to changing land management practices
Open Access
Highlights
- •
- Implications of improved environmental practices on nutrient transport were examined.
- •
- Reduced fertiliser usage and timing was linked to reduced estuarine nutrient loadings.
- •
- P and water column chlorophyll improved while N remained stable in the estuary.
- •
- Nutrient transport through an estuarine system depends on internal nutrient cycling.
Abstract
Anthropogenic
pressures have led to problems of nutrient over-enrichment and
eutrophication in estuarine and coastal systems on a global scale.
Recent improvements in farming practices, specifically a decrease in
fertiliser application rates, have reduced nutrient loadings in Ireland.
In line with national and European Directives, monitoring of Irish
estuarine systems has been conducted for the last 30 years, allowing a
comparison of the effectiveness of measures undertaken to improve water
quality and chemical and biological trends. The Blackwater Estuary,
which drains a large agricultural catchment on the south coast of
Ireland, has experienced a decrease in calculated nitrogen (N) (17%) and
phosphorus (P) (20%) loads in the last decade. Monitored long-term
river inputs reflect the reductions while estuarine P concentrations,
chlorophyll and dissolved oxygen saturation show concurrent improvement.
Consistently high N concentrations suggest a decoupling between N loads
and estuarine responses. This highlights the complex interaction
between N and P load reductions, and biochemical processes relating to
remineralisation and primary production which can alter the
effectiveness of the estuarine filter in reducing nutrient transport to
the coastal zone. Effective management and reduction of both diffuse and
point nutrient sources to surface waters require a consideration of the
processes which may alter the effectiveness of measures in estuarine
and coastal waters.
Graphical abstract
Keywords
- Nutrient source apportionment;
- Estuarine eutrophication;
- Agricultural policy measures;
- Mann–Kendall trend analysis;
- Nitrogen and phosphorus river loads
1. Introduction
Increased
nutrient enrichment derived from the rise of agricultural fertiliser
use, human population pressures and atmospheric deposition has resulted
in deleterious impacts on surface waters along the land–ocean continuum
over past decades (Kronvang et al., 1993 and Boesch, 2002). The European Union has specifically aimed at reducing nutrient inputs through the adoption of the Nitrates and Urban Waste Water Treatment Directives (1991) and the Water Framework Directive (WFD) 2000.
Decreases in the loadings of nutrients and organic matter that cause
eutrophication have been documented in a number of systems in Europe and
worldwide (Testa et al., 2008, Duarte et al., 2009 and Windolf et al., 2012). However, while measures have been shown to be effective in reducing the use of nitrogen (N) and phosphorus (P) fertiliser (Lalor et al., 2010 and Bouraoui and Grizzetti, 2011) and loss of nutrients from wastewater treatment plants and industrial discharges (Schindler, 2006 and Kronvang et al., 2008), future demands for food production will likely augment the intensity of agricultural practices in many countries.
Following
the implementation of mitigating measures, recovery of surface waters
from impairment is expected to vary depending on catchment
characteristics including forestry cover, agriculture and the degree of
urbanisation. Furthermore, natural factors such as geology, soils,
climate, and hydrology will largely determine background water quality
and legacy accumulation of anthropogenic nutrients in soils (Jordan et al., 2012, Taylor et al., 2012 and Vermaat et al., 2012).
The
specific response of estuarine and coastal systems to decreases in
diffuse and point source loads can differ greatly due to their inherent
chemical, biological and physical gradients and complex biogeochemical
cycles. Estuaries can act as a source of nutrients, especially P (Deborde et al., 2007 and Van Der Zee et al., 2007) and silica (Legovic et al., 1996 and Cabeçadas et al., 1999), due to organic material recycling, desorption and diffusion of P from sediment pore waters during early digenesis (Deborde et al., 2008 and Delgard et al., 2012).
Secondly, they may act as a sink or source of N through the balance
between nitrification–denitrification and ammonification-anammox (Abril et al., 2000, Garnier et al., 2006 and Seitzinger et al., 2006).
Finally, biological assimilation can also act to filter nutrients as
they pass through the estuarine system. However, the response of primary
producers to nutrient availability will depend largely on physical and
biological constraints such as light, residence time, grazing and ocean
exchange (Cloern, 2001, Carstensen et al., 2011 and O'Boyle et al., 2015).
Studies which trace N and P flows from the source to the coastal zone
allow the determination of not only the effectiveness of mitigation but
enhance understanding of response trajectories. This will assist in the
future targeting of actions to be applied specifically in light of
current and future programmes of measures to be undertaken under the
Nitrates Directive, WFD, and Marine Strategy Framework Directive (MSFD).
To
investigate the link between measures and improvements in water
quality, the Blackwater catchment and estuary in southern Ireland, which
has seen a substantial enhancement in water quality in the last decade,
was selected. The trophic status of the estuary, having previously been
classified as eutrophic by the EPA's Trophic Status Assessment Scheme
(TSAS), has shown a marked improvement and is now classed as unpolluted
with respect to eutrophication. Monitoring has been undertaken since
1990 to track river loads and since 1992 to evaluate the biochemical
status of the estuarine system. Coupled with this, an assessment of
nutrient source apportionment in the catchment has been carried out for
the years 1990, 2000 and 2010. This is done as part of a national source
apportionment exercise undertaken to meet the reporting requirements of
the Oslo–Paris Convention on the Protection of the North Eastern
Atlantic (OSPAR). The combination of these data is now a valuable tool
which can be used to determine the links between improvements in
practices and the response of an estuarine and coastal system.
The
objectives of this study were; 1) to determine whether decreases in
overall loads and changes in load apportionment to the estuarine
catchment have occurred in the last 20 years; 2) to examine potential
links between trends in calculated catchment nutrient loads, measured
river loads and downstream estuarine concentrations; 3) to determine the
impacts of any changes on physico-chemical and biological parameters
within the estuarine system and 4) to identify the measures that have
been most effective in reducing nutrient loss from the catchment to the
estuary.
2. Materials and methods
2.1. Study site
The Blackwater Estuary drains a large agricultural basin in southern Ireland with a catchment area of 3307.5 km2.
Livestock constitute the main farming activity in the area with over
50% of the agricultural land dedicated to pasture and 30% to the
production of silage. A number of small towns and villages also occupy
the catchment while the town of Youghal (treatment population 10,000),
which lacks a waste water treatment plant, lies at the estuary mouth.
The south of Ireland is a temperate region, with highest rainfall and
river flows occurring in the autumn/winter months. Median freshwater
discharge is 106.6 m3/s, with winter (October–March) flows
being twice those of the summer periods. The Blackwater Estuary is
shallow (average depth 4.2 m) and mesotidal with a tidal range of 3.6 m,
a surface area of 12.1 km2 and an intertidal area of 4.5 km2. The estuary is generally well-mixed although stratification occurs in the mid-estuarine region.
2.2. Catchment nutrient load estimations
The
quantification of nutrient sources to the Blackwater catchment was
based on historic reporting procedures which have been undertaken to
comply with requirements under OSPAR. In order to identify trends, load
calculations were undertaken for 1990, 2000 and 2010. These years were
chosen as the largest body of information was available at this time
step. A detailed account of load calculations for, diffuse (inorganic
and organic fertilisers, land use, unsewered population) and point
(waste water treatment plants, industry) sources of nutrients is
described below. In cases where actual data sets of direct discharges
and pathway processes are unavailable, coefficients have been applied
based on commonly agreed methods and previously measured rates (OSPAR, 2011 and O'Sullivan, 2002).
2.2.1. Inorganic and organic fertilisers
Agricultural
data for farm area usage (cereals, potatoes, silage, hay, pasture) and
livestock densities (cattle, sheep) were obtained from the Central
Statistics Office (CSO) of Ireland for the three years and are delimited
into area per electoral district (ED). National inorganic fertiliser
application rates of Nitrogen (N) and Phosphorus (P) per land use type
were sourced from a national farm study (Lalor et al., 2010).
As the survey only encompasses 1995–2008 the 1995 and 2008 application
rates were used for 1990 and 2010 respectively. Justification for the
use of 1995 and 2008 in lieu of 1990 and 2010 is based on the relatively
small change in recorded fertiliser sales between 1990 and 1995 (2.5%
increase for N and 0.7% decrease for P) and 2008 and 2010 (2008
fertiliser sales are within the standard deviation of the sales values
for 2009–2011 (years used in the comparison of actual river loadings
with calculated loads) for N (2008 = 309,000 tonnes;
2009–2011 = 327,670 ± 29,940) and P (2008 = 26,000;
2009–2011 = 26,000 ± 5200). As fertiliser sales are correlated with
fertiliser application rates (Lalor et al., 2010)
it can be assumed that the values used in the study are representative
of actual application rates. Annual excretion rates per livestock type
were obtained from the Good Agricultural Practice for Protection of
Waters S.I. No. 101 of 2009 (Government of Ireland, 2009). The calculation assumes that all excreted N and P are spread on the land during the year.
To
account for pathways of nutrients from field to surface water, loss
coefficients were applied for N and P which were adopted from
agricultural nutrient losses as per the NEUT 99 Screening Procedure for
Irish Coastal Waters with regard to Eutrophication Status. These
coefficients were also reported by the EPA to be within the ranges of
values quoted in PRAM 99/715-E Draft Guidance No. 6.
The agricultural loss estimates are also comparable with direct
measurements of export rates in agricultural areas in Ireland obtained
from catchment monitoring and management programmes (O'Sullivan, 2002).
It is recognised that this methodology, which estimated that 20% of
input agricultural N and 4% of input agricultural P reaches water bodies
in all areas, does not take account of variability in runoff risk
properties of soils or differences in transport processes relating to
bedrock or groundwater pathways (Tedd et al., 2014).
2.2.2. Leaching from landcover categories
Land
cover information was obtained from the European Corine (Coordination
of Information on the Environment) Land Cover Maps for 1990, 2000, and
2006 which are produced from satellite imagery. 2006 was used for the
2010 calculations as data from 2010 was unavailable. Nutrient leaching
from forestry, peatlands, inland marshes, scrub and different urban
areas were then estimated using standard coefficients which were
determined during a study undertaken in Ireland (O'Sullivan, 2002).
2.2.3. Unsewered rural populations
Standard
nutrient loading factors were utilised to estimate loadings to
watercourses from rural populations and septic tanks (PRAM 99/715 Draft
OSPAR Guidelines for harmonization of quantification and reporting
procedures (N = 9.0 g/person/day, P = 2.7 g/person/day)). In the absence
of data regarding the location of septic tanks, 50% were assumed as
standard loss rates and the remaining 50% assumed to be remote from
watercourses. The unsewered population was calculated by subtracting the
sewered population (CSO data source) from the total population.
2.2.4. Background losses
Losses
were estimated on the basis of total catchment area in accordance with
PRAM 99/7/5-E Draft Guideline No. 6 Annexes I and III. N and P losses
from background runoff were calculated as 0.75 kg N/ha/y and 0.05 kg
P/ha/y, respectively. These background rates are comparable with values
recorded for headwater sites obtained from catchment monitoring and
management programmes (O'Sullivan, 2002).
N and P inputs to waters originating from atmospheric deposition to the
catchment are considered to be accounted for in the background loss
estimates. Estimates of atmospheric deposition of total oxidized
nitrogen directly to surface water were determined from the European
Monitoring and Evaluation Programme Meteorological Synthesising
Centre-West (EMEP/MSC-W) model results for the three years of the study.
2.2.5. Wastewater treatment plants
Outflow
from Wastewater Treatment Plants was estimated according to operating
population equivalent (PE), assumed nutrient production loading
(N = 9.0 g/person/day, P = 2.7 g/person/day) and reduction factors
dependant on the level of treatment (PRAM 99/7/5-E Draft Guidance No. 7
Annex 1 Paragraph 6.3 NEUT Guidelines (Meeting of the Working Group on
Nutrients and Eutrophication, October 1999)). Reduction factors are
based on typical water quality data examples (OSPAR, 2011).
Under estimates of loads can result where treatment plants are not
operating efficiently while capacity overload at plants resulting in
overflow of partially treated or untreated effluents can also occur.
These data were used where actual population served was available
2.2.6. Unsewered industries
Historical
loading from licenced industries which discharge directly to
watercourses were estimated as 25% of maximum allowable discharge in
accordance with PRAM 99/7/5-E Draft Guideline No. 7 Annex 1 Paragraph
6.7. A compilation of monitoring data on actual emissions from 50 Irish
companies showed that nearly all companies discharged less than 25% of
the maximum licenced emission; hence this can be considered a reasonable
indication of actual emissions (EPA, 2000).
Loadings have been measured since 2007 in the case of larger industries
and can now be sourced from the Irish EPA's PRTR (Pollutant Release and
Transfer Register) database where available.
2.3. River load calculations
Monthly
nutrient loads to the Blackwater Estuary were calculated from
measurements undertaken under the OSPAR Riverine Inputs Programme (RID)
from 1990 to 2011. Instantaneous nutrient concentrations and flow were
measured monthly to give an instantaneous load. The load was
flow-weighted by monthly mass flow (measured daily) to give a monthly
load to the estuary, and all months were then summed to give annual
loads to the estuary in tonnes. Flow rates were sourced from hydrometric
data publically available from the Irish EPA HydroNet website
(hydronet.epa.ie).
2.4. Estuarine monitoring data
The
EPA has been monitoring the Blackwater Estuary on a seasonal winter
summer basis since 1997. The data set used incorporates 18 sampling
stations which are monitored once during winter and 3 times during the
productive period between May and September (Fig. 1).
Samples for the analysis of chlorophyll and nutrients were collected
using a 2-litre Hydrobios Ruttner bottle at the surface and 0.5 m above
the bottom. Dissolved Oxygen saturation (DO Sat) together with
temperature, salinity and depth were recorded using a Hydrolab datasonde
CTD. For practical purposes tidal sampling was scheduled to take place
in mid- to late morning (8.00–11.00 am) and again in mid- to late
afternoon (2.00–5.00 pm) to capture tidal variation.
- Fig. 1.The Blackwater Estuary on the south coast of Ireland. Monitored sampling stations are represented by black dots and labelled.
Samples
for the measurement of chlorophyll, a proxy for phytoplankton biomass,
were filtered using Whatman GF/C glass fibre filters and stored
overnight in the dark to prevent photo-degradation. Pigments were
extracted using hot methanol and absorbance (not corrected for the
presence of phaeopigments) was measured using a spectrophotometer (Standing Committee of Analysts', 1980).
Ammonia, total oxidized nitrogen (TON) and molybdate reactive
phosphorus (MRP) were measured according to Standard Methods for the
Examination of Water and Wastewater (2005). Water transparency at each
station was measured using a 25 cm diameter Secchi disc. Estuarine
parameters (DIN (dissolved inorganic nitrogen), MRP, chlorophyll, N:P
and DO Sat) were mapped with contouring software using a local
polynomial grid method (Surfer 11, Golden Software 2012).
