Indirect Nitrous Oxide Emissions from Surface ... - ACS Publications

Jul 12, 2012 - Nitrogen removal rates in a frigid high-altitude river estimated by ... Baker , Ke Xiao , Peter A. Turner , Ming Chen , John Zobitz , K...
0 downloads 0 Views 4MB Size
Article pubs.acs.org/est

Indirect Nitrous Oxide Emissions from Surface Water Bodies in a Lowland Arable Catchment: A Significant Contribution to Agricultural Greenhouse Gas Budgets? Faye N. Outram* and Kevin M. Hiscock School of Environmental Sciences, University of East Anglia, Norwich Research Park, Norwich, NR4 7TJ, U.K. S Supporting Information *

ABSTRACT: In the UK agriculture is by far the largest source of nitrous oxide (N2O) emissions. Direct N2O emissions as a result of nitrogen (N) application to soils have been well documented in the UK, whereas indirect emissions produced in surface waters and groundwaters from leached N are much less understood with limited data to support IPCC emission factors. Indirect emissions were studied in surface waters in the Upper Thurne, a lowland drained arable catchment in eastern England. All surface waters were found to have dissolved N2O concentrations above that expected if in equilibrium with ambient concentrations, demonstrating all surface waters were acting as a source of N2O. The drainage channels represented 86% of the total indirect N2O flux, followed by wetland areas, 11%, and the river, 3%. The dense drainage network was found to have the highest dissolved N2O concentrations of all the water bodies studied with a combined N2O flux of 16 kg N2O−N per day in March 2007. Such indirect fluxes are comparable to direct fluxes per hectare and represent a significant proportion of the total N2O flux for this catchment. Separate emission factors were established for the three different surface water types within the same catchment, suggesting that the one emission factor used in the Intergovernmental Panel on Climate Change (IPCC) methodology for predicting all indirect N2O emissions is inappropriate.



INTRODUCTION The tropospheric concentration of nitrous oxide (N2O), a potent greenhouse gas, has increased by 9.4% since preindustrial times as a result of anthropogenic perturbation of the natural N cycle, increasing at the rate of 0.2% each year between 1981 and 1996.1 Not only does N2O have a global warming potential 310 times that of carbon dioxide, it also has a long residence time in the atmosphere of about 114 years, meaning that N2O emissions have a long-term effect.2 In the UK, agriculture is considered by far the biggest source of N2O,3 being formed from direct sources such as fertilized soils or animal manures, or indirect sources, including groundwaters, surface drainage, rivers, and estuaries receiving nitrogen runoff from agriculture. While direct sources of N2O from agriculture have been fairly well documented,4−6 indirect N2O emissions from groundwater, streams, and rivers receiving N-rich drainage water from agricultural soils are less well understood.7 The Intergovernmental Panel on Climate Change (IPCC) provides guidelines on calculating national inventories of N2O emissions associated with agriculture. Indirect N2O emission factors (EFs) are a method for expressing N2O emissions from a water body as a fraction of the original N flux into the system.8 Indirect EFs can be calculated in two different ways. The first method, referred to here as EF(A), is calculated by taking the total annual flux of N2O from the water body and dividing it by the total annual amount of N leached to the water body, which is the calculation used by the IPCC methodology to calculate national N2O inventories.8,9 However, as most studies are often lacking in such detailed mass balance © 2012 American Chemical Society

information, EFs are commonly calculated by using a N2O− N/NO3−N mass ratio derived using concentration data collected from the water body, referred to in the following as EF(B).8 The emission factor assigned to indirect emissions, collectively known as EF5 in the IPCC Guidelines for National Greenhouse Gas Inventories, was decreased after several studies called for its revision from 0.025 in the 1997 IPCC report to 0.0075 in the 2006 IPCC report.10,9 The specific emission factor for surface drainage and groundwaters, EF5 g, was decreased from 0.015 to 0.0025. However, the original IPCC EF5 g emission factor from 1997 was only based on six studies which had been carried out in Japan, Israel, and the United States, and the revision made to the emission factor in 2006 was only based on four new studies and a review paper which included the original six studies.10,9 Therefore, the emission factor EF5 g has been calculated with no distinction made among different topographies, climates, or land use, and considers N2O production mechanisms as a result of N-inputs from agriculture within surface drainage and groundwaters to be comparable. Further still, there is currently no IPCC emission factor to estimate the indirect emission of N2O from lakes. Neglecting lakes may account for serious uncertainties in regional N2O emission budgets, particularly in lake-rich landscapes.11 Received: Revised: Accepted: Published: 8156

April 16, 2012 July 11, 2012 July 12, 2012 July 12, 2012 dx.doi.org/10.1021/es3012244 | Environ. Sci. Technol. 2012, 46, 8156−8163

Environmental Science & Technology

Article

Figure 1. Location map of the Upper Thurne catchment within the Broads National Park, showing the broads, river, and main villages and sampling locations according to sample type: river, broad, pumped drainage, or drainage channel.

arable production, resulting in nutrient-rich drainage waters that feed directly into the wetland and river system. The purpose of this paper is to study N2O concentrations in different surface water bodies within this drained landscape, including the lakes, river, and drainage channels, to examine whether each surface water type functions differently in terms of N2O production and to calculate indirect fluxes. These data

The Broadland landscape in Norfolk and Suffolk in East Anglia in eastern England is an extremely valuable asset to the UK with land suitable for highly productive arable agriculture, as well as the biodiversity and amenity value of the largest wetland area in the country. The Upper Thurne catchment is typical of this landscape, where several large eutrophic wetlands exist side by side with under-drained land supporting intensive 8157

dx.doi.org/10.1021/es3012244 | Environ. Sci. Technol. 2012, 46, 8156−8163

Environmental Science & Technology

Article

Table 1. Analytical Results for Surface Water Samples in the Upper Thurne Catchment in March 2007, Grouped According to Surface Water Type (Values Given As Mean Concentrations with Sample Number in Brackets; nd = not detected) surface water type (no. of samples)

temp (°C)

pH

Elec. cond. (mS cm‑1)

Eh (mV)

DO (%)

NO3‑ (μM)

NH4+ (μM)

N2O (nM)

river (5) pumped drainage (12) broad (13) drainage channel (10)

9.2 9.2 10.3 9.4

6.6 6.6 7.3 6.8

7.1 5.5 5.5 8.2

199 165 211 90

143 115 99 93

115 175 56 38

22 19 4 26

62 129 22 116

Figure 2. (a) Box and whisker plot of dissolved N2O concentration in surface waters according to sample type and (b) a 3D-scatter plot showing the relationship between Eh, NO3−, and N2O. Samples collected in March 2007. Group 1 mainly consists of drainage channels, Group 2 consists of broads and the River Thurne, and Group 3 consists mainly of pumped drainage.

elevation and, therefore, require pumped drainage. The soils in the higher area of the Thurne tend to be noncalcareous, loamy soils and are ideal for agriculture. The main river is underlain by gleysols with side-valleys of naturally wet and fertile peat, with sandy soils bordering the dunes. Arable farming is the dominant form of land use, with large areas of grassland and fens surrounding the broads and rivers. These deposits are underlain by Pleistocene Crag which is formed of marine shelly sands and silty micaceous clays. Beneath the Crag are Tertiary clays of mudstone and siltstone with ash layers, all of which is underlain by Cretaceous Chalk, a fissured limestone. The Crag aquifer, which is present over much of eastern East Anglia, is of importance for groundwater supply and maintaining the wetland environment.12 Sample Collection. A catchment-wide sampling round was carried out in March 2007, in which samples were taken from the surface waters of the open wetland, the river, field drains draining agricultural land, and drainage channels leading to the land drainage pumps (hereafter referred to as pumped drainage) (Figure 1). On each sampling occasion, in situ measurements of pH, redox potential, temperature (Hanna HI 9025), electrical conductivity (Hanna HI 9635), and dissolved oxygen (Hanna HI 9143) were taken using hand-held meters. Water samples were taken for the analysis of nutrients by filtering on site using 0.45-μm cellulose-acetate filters into centrifuge tubes. Samples for dissolved N2O analysis were collected in 100-mL glass syringes with Luer Lock taps. All samples were transported on ice and returned to cold storage at 4 °C. Samples for N2O analysis were analyzed within 48 h of collection. Filtered samples for nutrient analysis were frozen

are then used to calculate and compare the EF5 g indirect emission factor with that from the IPCC methodology guidelines to test whether this general emission factor is appropriate for predicting N2O emissions from different surface water types in a temperate, lowland arable environment. A sensitivity analysis of the N2O flux calculation is used to test the significance of gas transfer coefficients when working out indirect fluxes. Finally, the indirect N2O budget for the catchment is estimated and compared to the direct N2O budget to ascertain whether the indirect contribution makes up a significant part of the catchment N2O budget.



EXPERIMENTAL SECTION Study Area. The Upper Thurne catchment, with an area of 110 km2, is part of the Broads National Park located in northeast Norfolk, with a catchment boundary reaching to the coast (Figure 1). There are several shallow lakes or “broads” that are interconnected via pumped dikes (ditches) to one another and to the River Thurne, which flows southwards, joining the River Bure, before eventually flowing out to the North Sea at Great Yarmouth. The tidal influence from Great Yarmouth extends up the River Thurne and can reach far as Hickling Broad. This tidal incursion prevents the wetland system from draining when tides are high, yet when tides are low, pump discharge dominates flow. The land has been drained by pumps for several decades, with gravitational drainage accounting for only a small proportion of the catchment flows. The Upper Thurne area is overlain by Holocene peat, clay, silts, and sand which are close to or just below sea level in 8158

dx.doi.org/10.1021/es3012244 | Environ. Sci. Technol. 2012, 46, 8156−8163

Environmental Science & Technology

Article

Figure 3. (a) Daily N2O flux for each surface water type in March 2007, presented as kg of N2O−N per day, and (b) kg N2O−N per hectare per day.

2). Dissolved N2O concentrations were highest in samples taken from the pumped drainage sites, with a mean concentration of 129 nM, followed by the drainage channels, which had a mean concentration of 115 nM. The mean concentration of the river water samples was roughly half that of the drainage channels at 62 nM. The broads had the lowest concentrations, with a mean of 22 nM. Although the pumped drainage sites had the highest mean concentration, the drainage channel sites had the highest range, with a minimum concentration of 33 nM and a maximum of 452 nM. The broads had the smallest range in dissolved N2O concentration, with a minimum of 13 nM and a maximum of 37 nM. The lowest concentration, observed in Hickling Broad, was just above the concentration of N2O expected when air is in equilibrium with water. A multiple regression analysis using SPSS (version 17.0.1) showed that NO3− concentration, NH4+ concentration, and redox potential (Eh) were all important factors predicting N2O concentrations; R2 = 0.67, F(5) = 12.35, NO3− β value of 0.52 (p = 0.001), NH4+ β value of 0.25 (p = 0.031), and Eh β value of −0.002 (p = 0.005). The 3D scatter plot in Figure 2b reveals three distinct groupings when NO3−, N2O, and Eh data are simultaneously plotted, with Group 1 consisting mainly of pumped drainage sites, Group 2 being mainly drainage channel sites, and Group 3 being a mixture of river and broads sites, showing that each surface water type functions differently in terms of surface water N2O production. A full description of the relationship between dissolved N2O and other parameters and possible N2O production mechanisms can be found in the Supporting Information. N2O Flux from Water Bodies. Daily N2O fluxes from surface water compartments were calculated with dissolved N2O concentrations collected during March 2007. The data are shown in Figure 3 as both the total flux in kg N2O−N day−1 and also flux in kg N2O−N ha−1 day−1 for each surface water type. The combined indirect N2O flux for all surface waters is calculated to be 18.4 kg N2O−N day−1. The drainage channels are the most significant contributor of N2O of all compartments, contributing a total of 16 kg N2O−N day−1, which represents 86% of the daily indirect N2O flux. The combined flux from all of the broads sites is about 2 kg N2O−N day−1, representing 11% of the daily indirect flux, whereas the River Thurne had the smallest flux of 0.5 kg N2O−N day−1, representing 3% of the daily flux. The drainage channels were found to have the highest mean N2O flux per hectare of 0.08 kg N2O−N ha−1 day−1 (Figure 3b). The River Thurne has the second highest flux per hectare and the broads have the lowest with fluxes of 0.02 and 0.009 kg N2O−N ha−1 day−1, respectively.

ready for batch analysis once the sampling round had been completed. Analytical Methods. Ammonium was determined by ion chromatography using a Dionex Dx 600 instrument, with a detection limit of 1.7 μM. Nitrate and nitrite concentrations were also determined by ion chromatography using a Dionex ICS2000 instrument, with a detection limit of 0.87 μM. A purge and trap preparation line connected to a Shimadzu GC-8A gas chromatograph fitted with an electron capture detector was used for dissolved N2O analysis. Full details of the analytical procedure are given in the Supporting Information. The N2O detection limit of the GC was found to be 0.07 nM. Repeat analyses of standards were found to have a precision of ±4.1% whereas water samples had a precision of ±2.5% (standard error expressed as a percentage of the mean). A water−air gas transfer model was used to calculate the flux of N2O from the surface waters to the atmosphere following previously published methods,13,14 whereby a gas-transfer coefficient of velocity across the water−air interface, appropriate to each surface water type, was used with the concentration of N2O in the water (corrected for salinity and temperature) and the concentration of dissolved N2O in surface water in equilibrium with the atmosphere. The water bodies of the Upper Thurne were divided into different compartments so that N2O fluxes could be calculated for different sections of the catchment. Each of the broads was considered to be an individual compartment, as too were larger ditches. The River Thurne was considered to be one compartment and the drainage ditches were combined together to form one compartment. The N2O concentration data from the pumped drainage and drainage channels were combined to calculate an average N2O flux from the entire drainage network. The flux calculated was then multiplied by the area of the compartment, obtained using a GIS, in order to calculate the total daily N2O flux for each compartment. Full details of the water−air flux calculation can be found in the Supporting Information.



RESULTS N2O Concentrations in Surface Waters. Table 1 shows the mean N2O and inorganic N species concentrations for each surface water type along with in situ parameters measured from the catchment-scale sampling in March 2007, while Figure 2a shows the variation in dissolved N2O concentrations for each surface water type. All of the sites were found to have dissolved N2O concentrations higher than would be expected when atmospheric N2O concentrations are in equilibrium with water, which is ∼12 nM, if the atmospheric concentration is 0.319 ppm (at 12 °C) (15; atmospheric N2O concentration data from 8159

dx.doi.org/10.1021/es3012244 | Environ. Sci. Technol. 2012, 46, 8156−8163

Environmental Science & Technology



Article

Indirect N2O Emission Factors. In this study, it is possible to estimate EFs using both methods EF(A) and EF(B), as presented in the Introduction. Daily N2O fluxes from each water body type (agricultural drainage channels, river, and open broads) calculated for the month of March have been scaled up to estimate potential annual fluxes. Annual N leaching rates were estimated using a calibrated export coefficient model whereby published regional export coefficients (leaching rates) were used alongside land use and N input data, such as fertilizer application rates, to predict the amount of N lost to surrounding water bodies (details in Supporting Information). The model was calibrated with N concentration and discharge data from six pump subcatchments before it was used to predict N leaching loads for the remaining pump subcatchment areas. These leaching loads were used along with scaled up N2O annual fluxes to estimate EFs for each surface water type using method EF(A). In comparison, the second method of using field data was applied to calculate EFs for each surface water type using method EF(B). The EFs produced by the two calculation methods are shown in Table 2.

DISCUSSION Indirect N2O Fluxes. The mean flux from the drainage channels was an order of magnitude greater than that found in a published study of headwater streams receiving under-drainage in North America7 and within the range found in agricultural drainage streams in Scotland.16 If the calculated daily N2O fluxes from the Upper Thurne catchment were representative of the whole of March the monthly flux from each compartment (agricultural drainage channels, open broads, and river) for the whole month would be 2.4, 0.6, and 0.2 kg N2O−N ha−1, respectively. These values could represent an over- or under-estimation for that month depending on the diurnal fluctuations and variability in inputs and processing within the system. The N2O concentration data collected here are limited in temporal resolution. In that the samples were collected in March only, it is impossible to calculate accurate annual fluxes from each surface water type. Previous studies have shown that N2O fluxes in rivers and streams can be highest in summer17,18 coinciding with maximum temperatures, or winter7,17,19 coinciding with maximum NO3− concentrations, or can even exhibit seasonal variation in one year and not the next,19 making it very difficult to predict how fluxes in this system would vary across different seasons. Nonetheless, these monthly fluxes are a good indication of possible fluxes and a good starting point for comparison with other systems. For the month of March 2007 alone, the drainage ditches exhibited an N2O flux that is comparable to annual direct N2O fluxes from cultivated soils in the UK, which commonly range from 0.3−3.6 kg N2O−N ha−1 a−1.4−6 The drainage ditches are, therefore, an important source of N2O in the Upper Thurne catchment, as the dense drainage network is a significant feature of this lowland landscape. The drainage channels in this catchment receive water from field under-drainage which, at times of low evapotranspiration and high precipitation, routes NO3−-rich groundwater directly to streams, resulting in high NO3− concentrations.19 Thus, the seasonality of precipitation and subsurface flow can control the transport of soil derived NO3− to streams with important implications for in-stream N2O production.19 The strong relationship between N2O and NO3− and NH4+ found in this study could imply in situ production of N2O in agricultural drainage ditches. Previous authors have confirmed that headwater streams are not only conduits for the emission of N2O derived from groundwater but also active sites of in situ N2O production, with in-stream denitrification accounting for at least 26% of emissions, and nitrification potentially contributing twice this amount.18 Alternatively, the drainage from the agricultural fields could contain high N 2 O concentrations at the same time as high NO3− and NH4+ concentrations. Either way, under-drainage provides a conduit for N-rich drainage waters to reach drainage networks with multiple pathways for N2O formation. Under-drainage is prevalent in the Upper Thurne catchment and the surrounding area of Broadland in the UK, but also elsewhere in much of Europe: the extent of drained land in major European countries calculated in a recent study showed that the total proportion of agricultural land that is under-drained exceeded 30% in seven countries, from 33% in the UK to 93% in Finland.20 Therefore, indirect N2O production in drainage waters receiving water from under-drainage is likely to be taking place across extensive agricultural areas in Europe.

Table 2. Emission Factors Calculated for the Drainage Channels, River, and Broads in the Upper Thurne Catchment via Two Different Methods, EF(A) and EF(B), and the Relevant 2006 IPCC EF for Each Surface Water Typea

EF(A)

EF(B)

IPCC 2006

area (ha) kg N2O−N ha−1 day−1 kg N2O−N ha−1 a−1 total kg N2O−N a−1 total N input (kg) EF N2O (mg N L−1) NO3− (mg N L−1) EF EF5 g

drainage channels

River Thurne

broads

196 0.081

26 0.019

219 0.009

30 5795 109,287 0.053 0.0032 0.532 0.0061 0.0025

7 180 19,135 0.009 0.0017 1.61 0.0011 0.0025

3 719 40,354 0.018 0.0006 0.784 0.0008 No EF

The daily flux of N2O used in EF(A) was calculated using N2O concentration data collected in March 2007 and the water−air flux calculation described in the Supporting Information; the total amount of N entering each surface water type in EF(A) was calculated from the mass balance presented in the Supporting Information; EF(B) calculations were carried out using average N2 O and NO 3 − concentrations for each surface water type collected in March 2007. a

The EFs calculated using EF(B) reveal that EFs are not uniform for the different surface water compartments. For the drainage channels, the EF of 0.006 is higher than the IPCC EF of 0.0025, whereas the EF for the river, 0.0011, is lower. There is no IPCC EF for lakes, so a comparison with the EF calculated for the broads cannot be made. This research shows that within one catchment the different surface water types are producing varying amounts of N2O, with a separate EF required for each compartment. Therefore, there is a need to distinguish among different hydrological environments when calculating emission factors, unlike the IPCC approach of using one EF for all surface water types. More investigation should be undertaken to understand indirect N2O fluxes from lowland, upland, and riparian streams in order to establish separate emission factors for more accurate calculations of the global anthropogenic N2O budget. Emission factors for lakes and 8160

dx.doi.org/10.1021/es3012244 | Environ. Sci. Technol. 2012, 46, 8156−8163

Environmental Science & Technology

Article

Table 3. Sensitivity Analysis for N2O Flux Calculationsa change made to N2O flux model

compartment

±10% March 2007 model values for temperature, windspeed and N2O concentration

drainage channels

change in gas transfer coefficient

drainage channels drainage channels drainage channels open broads

seasonal data from field scale study in 2008 used to estimate compartment flux

open broads

variable changed

% change in model output for drainage channel flux (kg N ha‑1 day‑1)

temperature windspeed N2O concentration power law relationship22

3.0 19.3 10.7 5.0

N2O samples collected in March 2008 N2O samples collected in June 2008 N2O sample collected in March 2008 N2O sample collected in June 2008

N2O flux (kg N ha‑1 day‑1)

0.007 0.039 0.050 0.005

Alterations to the water−air gas flux calculation are as follows: ±10% adjustment to the major inputs (temperature, wind speed, and N2O concentration) of the water−air gas transfer calculation using data from March 2007 for the drainage channel compartment; use of a different gas transfer coefficient in the water−air gas transfer calculation using data from March 2007 for the drainage channel compartment; and application of field data from two separate field-scale campaigns in March 2008 and June 2008 used to estimate potential seasonal variations in N2O fluxes for drainage channels and broads compartments, assuming the samples are representative for the whole compartment. a

compartment. In the original gas transfer calculation the “bilinear fit” was used, taken from a previous study.22 When the “power fit” coefficient from the same study was applied to the flux calculation it resulted in a 5% decrease in N2O flux for the drainage channel compartment. These first two approaches from the sensitivity analysis demonstrate the importance of choosing the correct gas-transfer coefficient when working out indirect N2O fluxes. The justification for applying transfer coefficients from small lakes to the drainage channel compartment is that the movement of water is largely dominated by the action of the pumps as opposed to gravitational flow, meaning that the water is standing when the pumps are not activated. In a low-lying coastal catchment such as the Upper Thurne, wind speeds are fairly high (March 2007 average wind speed was 6.2 m s−1 and 2007 annual average wind speed was 5.6 m s−1, using hourly measurements taken from the Gorleston weather station, obtained from the British Atmospheric Data Centre) and are therefore likely to be a very important factor when calculating water−air N2O fluxes where the dense drainage network is largely dissecting exposed arable fields. In a separate study of forested headwater streams, the slope was found to be the most important factor influencing gas transfers to the atmosphere, where the steeper the gradient the higher the transfer.23 This study further highlights the necessity for more detailed gas transfer studies covering a wide range of stream and drainage channel types, taking into consideration exposure (and therefore the importance of wind), slope, stream order, roughness, and hydrological regime with the aid of tracers, such as SF6 or propane. In catchments where wind speed is an important factor in controlling indirect N2O fluxes, highresolution wind speed data would need to be collected in unison with high-resolution N2O data. For the third approach to the sensitivity analysis, field data from two separate field-scale campaigns in March 2008 and June 2008 were used to estimate potential seasonal variations in N2O fluxes for the drainage channel and broads compartments, if the samples are assumed to be representative for the whole compartment (a summary of the field-scale data can be found in the Supporting Information). For the drainage channel sites, the fact that predicted fluxes were higher for June 2008 than March 2008 would mean that the annual N2O flux presented here from up-scaled data from March 2007 could potentially be

wetlands should also be included in the IPCC methodology as, in this instance, ignoring the indirect N2O flux from the broads would result in an underestimation of the catchment indirect flux by 11%. The EFs calculated using EF(A) are very different from those calculated using EF(B). In every case the EFs calculated are higher using EF(A) than EF(B). The EF calculated for the drainage channels using EF(A) is an order of magnitude higher than that using EF(B), with values of 0.053 and 0.0061, respectively, and is 2 orders of magnitude greater in the case of the broads with values of 0.018 and 0.0008, respectively. The EF calculated for the river using EF(A), while the same order of magnitude as that calculated using EF(B), is nine times as high, with values of 0.009 and 0.0011, respectively. EF(B) would be suitable for estimating EF5 g if groundwater acted solely as the domain of transport without any processing of NO3− or N2O8 or where exchange with the atmosphere is minimal, but in open stream channels this method may underestimate the fraction of the load converted to N2O because much of the N2O has already been lost to the atmosphere.7 It also neglects the role of other N species which can be involved in N2O production; this study has shown that it is likely NH4+ has a significant role in N2O production, particularly in the pumped drainage water. Greater temporal N2O flux data are required before an accurate EF(A) calculation can be made but the data here support the findings of another published study21 in proposing that investigations and evaluations of the IPCC methodology should be carried out in association with detailed watershed mass balance studies. Uncertainty in N2O Flux Calculations. Due to the temporal limitations of the data set, a sensitivity analysis for the N2O flux calculations was carried out using three different approaches (Table 3). In the first approach a ±10% adjustment was made to the three major inputs of the water−air gas transfer calculation using data from March 2007 to assess the effect on N2O fluxes in kg N ha−1 day−1 for the drainage channel compartment. Wind speed had the greatest effect on N2O flux when all other variables were held constant, resulting in a 19.3% change in drainage channel flux. In the second approach, a different gas transfer coefficient was used in the water−air gas transfer calculation to assess the effect on N2O fluxes in kg N ha−1 day−1 for the drainage channel 8161

dx.doi.org/10.1021/es3012244 | Environ. Sci. Technol. 2012, 46, 8156−8163

Environmental Science & Technology



an under-estimation for the drainage channel compartment. However, for the broads sites, the opposite pattern exists, meaning that the annual flux presented could be an overestimation. Therefore, gas transfer tracer studies should be carried out in unison with long-term temporal studies across a range of surface water types and hydrological environments before meaningful emission factors can be put forward for IPCC assessments. Comparison of Direct and Indirect N2O Emissions. The export coefficient model was also used to estimate the amount of N lost as a direct N2O flux in order to compare with the indirect flux calculated previously. The model predicts an input to the catchment of 1 × 106 kg N a−1 as a result of atmospheric deposition and fertilizer applications. If the IPCC EF for direct application of mineral fertilizer of 0.01 is applied, a total of 1 × 104 kg N a−1 is converted to N2O−N. The model predicts an input to the catchment of 1.1 × 105 kg N a−1 from livestock. If the IPCC EF for direct emissions from livestock inputs of 0.02 is applied, a total of 2.2 × 103 kg N a−1 is converted to N2O−N. This makes a total contribution from direct sources of 1.2 × 104 kg N2O−N a−1 out of a total of 1.89 × 104 kg N2O−N a−1 from all sources. The direct N2O emissions from arable and livestock sources account for 53% and 12% of the total N2O budget, respectively, while the indirect emissions from the drainage channels, river, and broads (as shown in Table 2), account for 31, 1, and 4%, respectively. Therefore, the contribution of indirect sources of N2O from surface waters to the overall N2O budget from the catchment is not insignificant, potentially contributing around a third of all N2O emissions, particularly considering that surface waters only comprise 4% of the total catchment area. In catchments such as the Upper Thurne where under-drained arable land dominates the landscape, the contribution from drainage channels to the overall N2O budget of the catchment could be more significant than from livestock, where low stocking densities are present.



REFERENCES

(1) Khalil, M. A. K.; Rasmussen, R. A.; Shearer, M. J. Atmospheric nitrous oxide: patterns of global change during recent decades and centuries. Chemosphere 2002, 47 (8), 807−821. (2) Forster, P., Changes in Atmospheric Constituents and in Radiative Forcing. In Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change; Solomon, S., Qin, D., Manning, M., Marquis, M., Averyt, K. B., Tignor, M. Miller, H. L., Chen, Z., Eds.; Cambridge University Press: Cambridge, United Kingdom and New York, 2007. (3) Kroeze, C.; Mosier, A.; Bouwman, L. Closing the N2O budget: a retrospective analysis 1500 - 1994. Global Biogeochem. Cycles 1999, 13 (1), 1−8. (4) Smith, K. A.; McTaggart, I. P.; Dobbie, K. E.; Conen, F. Emissions of N2O from Scottish agricultural soils, as a function of fertilizer N. Nutr. Cycling Agroecosyst. 1998, 52, 123−130. (5) Skiba, U. M.; McTaggart, I. P.; Smith, K. A.; Hargreaves, K. J.; Fowler, D. Estimates of nitrous oxide emissions from soil in the UK. Energy Convers. Manage. 1996, 37 (6−8), 1303−1308. (6) Dobbie, K. E.; Smith, K. A. Nitrous oxide emission factors for agricultural soils in Great Britain; the impact of soil water-filled pore space and other controlling variables. Global Change Biol. 2003, 9, 204−218. (7) Beaulieu, J. J.; Arango, C. P.; Hamilton, S. K.; Tank, J. L. The production and emission of nitrous oxide from headwater streams in the Midwestern United States. Global Change Biol. 2008, 14 (4), 878− 894. (8) Well, R.; Weyman, D.; Flessa, H. Recent research progress on the significance of aquatic systems for indirect agricultural N2O emissions. J. Integr. Environ. Sci. 2005, 2 (2−3), 143−151. (9) IPCC. N2O emissions from managed soils, and CO2 emissions from lime and urea application (Vol. 4, Chapter 11). In 2006 IPCC Guidelines for National Greenhouse Gas Inventories; Prepared by the National Greenhouse Gas Inventories Programme; Eggleston, H. S., Buendia, L., Miwa, K., Ngara, T., Tanabe, K., Eds.; IGES: Japan, 2006. (10) IPCC. Greenhouse Gas Inventory Reference Manual, (Vol. 3) Agriculture. In Revised 1996 IPCC Guidelines for National Greenhouse Gas Inventories; Bracknell, UK, 1997. (11) Huttunen, J. T.; Juutinen, S.; Alm, J.; Larmola, T.; Hammar, T.; Silvola, J.; Martikainen, P. J. Nitrous oxide flux to the atmosphere from the littoral zone of a boreal lake. J. Geophys. Res. 2003, 108, 4421. (12) Holman, I. P.; Hiscock, K. M.; Chroston, P. M. Crag aquifer characteristics and water balance for the Thurne catchment, north east Norfolk. Q. J. Eng. Geol. 1999, 32, 365−380. (13) Barnes, J.; Owens, N. J. P. Dentirification and nitrous oxide concentrations in the Humber estuary, UK, and adjacent coastal zones. Mar. Pollut. Bull. 1998, 37 (3−7), 247−260. (14) de Wilde, H. P. J.; de Bie, M. J. M. Nitrous oxide in the Schelde estuary: production by nitrification and emission to the atmosphere. Mar. Chem. 2000, 69, 203−216. (15) Weiss, R. F.; Price, B. A. Nitrous oxide solubility in water and seawater. Mar. Chem. 1980, 8 (4), 347−359. (16) Reay, D. S.; Smith, K. A.; Edwards, A. C. Nitrous oxide emission from agricultural drainage waters. Global Change Biol. 2003, 9, 195− 203. (17) Wilcock, R.; Sorrell, B. Emissions of Greenhouse Gases CH4 and N2O from Low-gradient Streams in Agriculturally Developed Catchments. Water, Air, Soil Pollut. 2008, 188 (1), 155−170. (18) Beaulieu, J. J.; Tank, J. L.; Hamilton, S. K.; Wollheim, W. M.; Hall, R. O.; Mulholland, P. J.; Peterson, B. J.; Ashkenas, L. R.; Cooper, L. W.; Dahm, C. N.; Dodds, W. K.; Grimm, N. B.; Johnson, S. L.; McDowell, W. H.; Poole, G. C.; Valett, H. M.; Arango, C. P.; Bernot, M. J.; Burgin, A. J.; Crenshaw, C. L.; Helton, A. M.; Johnson, L. T.; O’Brien, J. M.; Potter, J. D.; Sheibley, R. W.; Sobota, D. J.; Thomas, S. M. Nitrous oxide emission from denitrification in stream and river networks. Proc. Natl. Acad. Sci., U.S.A. 2011, 108 (1), 214−219.

ASSOCIATED CONTENT

S Supporting Information *

Additional information on the N2O analytical procedure, N2O water−air flux calculation, relationships between N2O and other parameters, production mechanisms for surface water N2O, field-scale N2O data, and the catchment nitrogen mass balance calculation. This material is available free of charge via the Internet at http://pubs.acs.org.



Article

AUTHOR INFORMATION

Corresponding Author

*Phone: (441)603-592922; fax: (441)603-591327; e-mail: f. [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research was funded with a PhD research award from the Natural Environment Research Council (NER/S/A/2005/ 13589) and was a CASE award with the Broads Authority. We thank the Broads Authority, the Environment Agency, the Water Management Alliance,and Norfolk Wildlife Trust for their support and cooperation. 8162

dx.doi.org/10.1021/es3012244 | Environ. Sci. Technol. 2012, 46, 8156−8163

Environmental Science & Technology

Article

(19) Beaulieu, J. J.; Arango, C. P.; Tank, J. L. The Effects of Season and Agriculture on Nitrous Oxide Production in Headwater Streams. J. Environ. Qual. 2009, 38 (2), 637−646. (20) De la Cueva, P. Identification of Agricultural Areas in Europe Subject to Different Types of Field Drainage. MSc Thesis. Cranfield University, Silsoe, UK, 2006. (21) Groffman, P. M.; Gold, A. J.; Addy, K. Nitrous oxide production in riparian zones and its importance to national emission inventories. Chemosphere: Global Change Sci. 2000, 2 (3−4), 291−299. (22) Crusius, J.; Wanninkhof, R. Gas transfer velocities measured at low wind speed over a lake. Limnol. Oceanogr. 2003, 48 (3), 1010− 1017. (23) Wallin, M. B.; Ö quist, M. G.; Buffam, I.; Billett, M. F.; Nisell, J.; Bishop, K. H. Spatiotemporal variability of the gas transfer coefficient (KCO2) in boreal streams: Implications for large scale estimates of CO2 evasion. Global Biogeochem. Cycles 2011, 25 (3), GB3025.

8163

dx.doi.org/10.1021/es3012244 | Environ. Sci. Technol. 2012, 46, 8156−8163