Rivers and estuaries

According to model estimates (Dumont et al, 2005), the total global load of dissolved inorganic N transported by rivers to the oceans is 25Tg yr_1, of which 16Tg N yr"1 comes from anthropogenic sources, including sewage point sources (0.4Tg), diffuse agricultural emissions from mineral fertilizer (5.3Tg), biological N2 fixation (4.5Tg) and manure (3.8Tg). In addition, 12Tg organic N is transported by rivers to coastal waters (Harrison et al, 2005), plus 39Tg of particulate N (Beusen et al, 2005). The latter may not be all biologically available. N2O discharged by rivers to the ocean or emitted from river surfaces originates from various sources. Both nitrification and denitrification can occur in the water column, in the sediments and in the interior of suspended particles (Bange et al, 2006b). In addition to this in situ production, N2O can also be discharged to streams from groundwater or sewage plants (Seitzinger and Kroeze, 1998). The contribution from these sources is variable and difficult to estimate. However, Toyoda et al (2009) were able to show by using stable isotope signatures of N2O that peak N2O concentrations in an urban river were always dominated by sewage sources.

Model estimates of global N2O emissions from rivers and estuaries have been based on river loads of dissolved inorganic nitrogen (DIN) and the assumption that 50 per cent of the DIN is lost through denitrification and that all DIN is nitrified once (Seitzinger and Kroeze, 1998). It was also assumed that 0.3 per cent of the nitrified and denitrified N is lost as N2O, or 3 per cent when the average watershed N loading is relatively high (>10kg N ha-1 yr-1). This is similar to the basic concept of CEF3, but it includes nitrification and N-fluxes from natural sources. A direct comparison with CEF3 from other systems is thus not possible (Table 8.1). Further calculations with this model resulted in estimates for N2O emissions from rivers and estuaries of 1.26 and 0.25Tg N yr-1, respectively, totalling 1.5Tg N yr-1 (Kroeze et al, 2005). More recently, they estimated emissions from rivers and estuaries in 2000 to be 0.3-2.1Tg N2O-N, using updates of the global nutrient models (Kroeze et al, 2009).

Principally, the N-budget of rivers and estuaries might be used to derive EFs using the concepts of Equations 8.3 and 8.4. Assuming an average N retention in rivers of 50 per cent (Seitzinger and Kroeze, 1998) or 30 per cent (van Drecht et al, 2003) and using the total dissolved N export and N2O fluxes of about 40Tg N yr-1 and 1Tg N yr-1, respectively, yields FNin of 80Tg N yr-1 and 56Tg N yr-1, respectively, and CEF1 of 0.013 and 0.019, respectively. The lower value is close to the sum of IPCC-EF5r and IPCC-EF5e in the 1996 IPCC Guidelines (0.01). Both values are clearly higher than the mean values of the current IPCC Guidelines, where EF5r + EF5e = 0.005 (IPCC, 2006), but close to the upper limit of the respective uncertainty range (0.0004-0.017). An approximate CEF3 can be obtained from the ratio between N2O flux and N retention (16-40Tg yr-1) since the major fraction of N retention in rivers can probably be attributed to denitrification (Seitzinger and Kroeze, 1998). This would give a CEF3 in the range 0.025-0.063, but the actual value must be higher since some of the retention is not denitrification but burial of N in sediments. This estimate of river CEF3 is thus relatively large compared to the CEF3 of other systems (Table 8.2).

Reported N2O concentrations levels in rivers and estuaries range from 9 to 201nM, with medians of 25 and 83nM, respectively (Table 8.2). These levels are clearly lower than most of the reported groundwater concentrations (up to 100,000nM, with site medians always >100nM). There are several factors explaining the higher groundwater levels. Because degassing from aquifers is small (see above), N2O concentrations result mainly from the balance between production and reduction to N2. In the rivers, there is degassing of the groundwater-derived load and of N2O produced in situ during river transport. Another factor might be the fact that most groundwater data come from aquifers under agricultural land, whereas river data also include catchments without agricultural land use.

Table 8.2 Indirect N2O fluxes from constructed wetlands, rivers and estuaries and associated conceptual emission factors

Location and description





Constructed wetlands

Constructed river bank marshes, USA 1 to 3.5

Constructed wetlands for wastewater 1 to 590

treatment, Estonia

Rivers and estuaries

Model estimates of global rivers and estuaries N-budget model of global rivers and estuaries

Review of nine rivers 12.6 to 301 (83)3

Whole reaches of three small rivers in north-west Indiana and north-east 3 to 403 8.9 to 18.1

Illinois, USA.

Three lowland agricultural streams,

New Zealand 4 to 151 39 to 167

LII river, New Zealand 3 to 120 14 to 68

Seine river and estuary, France 12 to 22 Up to approx. 140

Review of 17 estuaries 8.6 to 220 (24.7)3

0.005 to 0.034 Hernandez and Mitsch, 2006

approx. 0.001 to 0.015 Mander et al, 2003

0.013 to 0.0186

0.00003 to 0.00017

0.003 or 0.031 0.033 to 0.0636


0.00006 0.02 to 0.042

Seitzinger and Kroeze, 1998

Toyoda et al, 2009

Laursen and Seitzinger, 2004

Wilcock and Sorrell, 2008 Clough etal, 2006,

2007 Garnier et al, 2006 Toyoda et al, 2009

Note: CEF,, CEF2 and CEF3 were estimated using Equations 8.1, 8.3 and 8.4. Individual data sets are only listed if EFs were given or can be calculated from presented data in the reference.1 0.03 if watershed N loading is relatively high (>10kg N ha-1 yr-1), otherwise 0.003.2 Estimated from the ratio between river-derived fluxes and estimated N-input within the river catchment.3 Total range (median in brackets).4 Denitrification estimated using C2H2 static core incubation, range of means.5 Denitrification by core incubation under He/O2. An approximate CEF3 range was estimated from the ranges of N2O and N2 fluxes since the range of (N2O/N2+N2O) ratios was not given.6 Estimate for rivers plus estuaries derived from global budget calculations (see text).

N2O dynamics in rivers are extremely variable since process-related factors such as concentrations of O2 and suspended or dissolved organic carbon depend on factors like flow velocity, depth, path length, sediment properties and nutrient load. For example, extremely low EF5r values were reported for a short, shallow, fast and well-oxygenated river in New Zealand (Clough et al, 2006, 2007), whereas high values were obtained for the deep, slow, partially sub-oxic Seine river system in France (Garnier et al, 2006) (Table 8.2).

Some studies report on the contribution of the N2O influx from groundwater and on in situ production. Xiong et al (2006) observed high N2O concentrations in river water similar to those in the groundwater wells in the Taihu river catchment and concluded that river N2O mostly originated from groundwater discharge in riparian areas. Laursen and Seitzinger (2004) measured N2O emission and denitrification in three rivers at the whole-reach scale by measuring dissolved N2, Ar and N2O, as well as concentrations of injected gas tracers. Den-itrification and N2O emission ranged from 0.4 to 60pmol N m"2 hour"1 (median 6.9pmol N m"2 hour"1) and from 0.31 to 15.91mmol N m"2 hour"1 (median 2.62mmol N m"2 hour"1), respectively. Using Equation 8.4, a CEF3 of 0.0026 (ratio of median N2O flux to median denitrification) can be calculated, which is similar to the median CEF3 of the reduced aquifers (Table 8.1). To fully characterize the fate of inorganic N during passage through aquifers, rivers and estuaries to the ocean it would be necessary to combine such studies with investigation of N2O loads and associated process dynamics in the aquifers of the catchment, for example using excess N2.

Principally, N2O dynamics in estuaries tend to be different from those in rivers, due to smaller flow velocities. This implies lower gas exchange and thus lower oxygen concentrations and higher residence times for N2O. Theoretically, there should be more N2O reduction to N2 and CEF3 should thus be lower than in rivers. Reported median N2O concentrations of reviewed estuaries (Table 8.2) are indeed lower than those of rivers (Table 8.2). Since gas exchange rates are also lower in the former locations, resulting N2O fluxes must be clearly smaller.

0 0

Post a comment