Comparing emission factors among systems and mitigation options for the aquatic pathway

To illustrate the impact of system properties such as residence time, N-retention, production and consumption of N2O and gas exchange on EFs, a simple process simulation is shown (Figure 8.2). During closed conditions, i.e. when there is no exchange of N species between a water parcel with adjacent domains, there is an ongoing progress of retention reactions leading eventually to complete consumption of reactive N, for example by reduction of NO3" to N2 via deni-trification (Figure 8.2a). During this process, N2O is produced and reduced, resulting in the typical pattern described above (Weymann et al, 2008). The magnitude of maximum N2O concentration depends on the balance between production and reduction and can thus be highly variable. The curves in Figure

8.2b represent EFs at a given stage of reaction progress, assuming that all N2O is emitted to the atmosphere at this stage, for example by discharge of ground-water via land drains or springs. By definition, CEF1 then follows the time course of N2O and thus depends on the reaction progress (RP) as described above. Low values of CEF1 can thus be found at early as well as late stages of the reaction. During the later phase, CEF3 must inevitably decrease with RP as the amount of produced N2 increases, while N2O decreases. This demonstrates that extended residence times under closed conditions in denitrifying systems lead to decreasing EFs. In semi-closed systems such as rivers and estuaries, there is emission of a certain fraction of produced N2O during ongoing reaction progress. This leads to final values of CEF1 and CEF3 >0 and both EFs may increase or decrease towards the end (Figure 8.2b).

The complete pathway of reactive N from the aquifer surface to the sea might be seen as a chain of connected systems where N2O dynamics follow varying patterns similar to the examples in Figure 8.2. Can the observed EFs in the various components of the aquatic pathway (Tables 8.1 and 8.2) be explained with these simple models of closed or semi-closed system? In riparian areas and constructed wetlands potential denitrification is often high due to richness in organic C. Residence time is very short at the interface between saturated and unsaturated layers. This might explain why the highest values of CEF1 and/or CEF3 are reported for riparian areas and constructed wetlands, i.e. semi-closed systems where reactive zones are in direct contact with or close to the atmosphere. The high EFs are thus in agreement with the EF pattern of semi-closed systems with high gas exchange (Figure 8.2b). The potential values of CEF1 and CEF3 of aquifers exhibit a large range, but EFs effective at the aquifer scale are probably much lower than maximum values of up to 0.5 (Table 8.1), which can also be seen from the effective CEF1 of 0.001 reported by Hiscock et al (2003). Due to almost closed conditions during aquifer transport, EFs can be expected to be at the low end.

Artificial land drainage decreases groundwater residence time. CEF3 may thus be larger compared to non-drained aquifers (Figure 8.2a). CEF1 of tile drainage might be lower or higher, depending on denitrification rates: if reaction progress is below the N2O maximum, artificial drainage decreases CEF1 and vice versa (Figure 8.2a).

For rivers and estuaries, CEF3 values derived from global budget calculations (0.03 and 0.06, respectively) are much higher than values based on measurements (Laursen and Seitzinger, 2004) or the default EFs that had been used in a river N2O model (Seitzinger and Kroeze, 1998). This discrepancy needs to be clarified. The river CEF3 of 0.003 derived from Laursen and Seitzinger (2004) is lower compared to CEF3 of riparian areas and constructed wetland and within the range of potential CEF3 of aquifers. From the process dynamics in rivers and estuaries one might expect that EFs are in between these groups, since the water column is more open compared to aquifers but there is no direct contact between active zones and the atmosphere such as at the saturated/unsaturated interface zone of wetlands and riparian areas.

What are the options for mitigating N2O fluxes from aquatic systems? Because anthropogenic discharge of reactive N is the driver for indirect N2O fluxes from aquatic systems, each measure to reduce this discharge certainly yields some mitigation effect. High mitigation efficiency can be expected from lowering discharge of reactive N to systems with high EFs, which might be riparian areas and constructed wetlands. Conversely, little effect will result, if systems with small EFs such as deep denitrifying aquifers receive lower N loads.

Principally, mitigation might also be achieved using end-of-the-pipe measures. Mathematically, N2O emission during the aquatic pathway is the product of CEF3 and N loss by denitrification within any time interval (Figure 8.2). Total N2O flux from denitrification during downstream movement of leached agricultural N through aquifers, wetlands, streams and estuaries is thus defined by the time course of denitrification and CEF3. Properties enhancing CEF3 thus increase total N2O fluxes. It might be possible to lower the high CEF3 values induced by near-surface processes in wetlands by assuring permanent flooding of these systems (Hernandez and Mitsch, 2006). Land drainage increases aquifer N2O fluxes since it enhances CEF3. Control of land drainage systems to achieve high water tables and thus longer residence times can be used to enhance denitrification within the drained groundwater zone. If CEF3 of this additional denitrification was lower than CEF3 of downstream systems, this measure would reduce overall N2O fluxes from the aquatic pathway. This needs to be investigated.

In rivers, measures to reduce flow velocity and increase depth, such as damming or restoration of natural meandering, might decrease the exchange of dissolved gases with the atmosphere. This might eventually reduce N2O fluxes, but care must be taken before generalizing because reduced gas exchange can lower oxygen concentrations and might thus favour N2O production by deni-trification.

N2O from groundwater abstraction wells might be trapped, for example by stripping the water and injecting the exhaust air into wastewater treatment plants. Because some drinking water plants use gas stripping to remove excess CO2 from the water, this measure might be realized with moderate effort. This might be an effective measure in cases of extremely high N2O levels as observed by Well et al (2005c).

Overall, each of these end-of-the-pipe measures can only mitigate a small fraction at various points within the rather long hydrological pathway, and each requires costly construction work. Moreover, there is a need to investigate the magnitude of potential mitigation effects. It is thus highly questionable whether these measures will be competitive compared to other more cost-effective greenhouse gas mitigation measures.

0 0

Post a comment