Developing revised emission factors for nitrous oxide emissions from agricultural pasture treated with nitrification inhibitors
Results
Nitrogen application as excreta and fertiliser onto pastoral soils
For the Kyoto Protocol base year of 1990, we calculated there was 1,406,153 tonnes (hereafter, abbreviated as t) of N as excreta deposited onto soils by grazing animals (Table A.4 in the Appendix). We focus on sheep and dairy cattle because these animal types made the greatest contributions. Just over half of this total was calculated to have come from sheep. For sheep, 66 % of total excreta was in the form of urine and 34 % as dung. In contrast, we calculated that dairy cattle contributed 25 % of total excreta and 74 % of their excreta was urine with 26 % as dung. In 1990, the quantity of N fertiliser applied to pastoral soils was 41,429 t. This figure is 80 % of the (3 year running mean quantity of) N fertiliser sold in 1990 based on expert judgement described earlier.
By 2004, we calculated that total excreta N and fertiliser N applications had increased by 136,089 and 234,763 t, respectively (totalling 370,852 t with the increase averaging 26,489 t per year for 1990 - 2004). We calculated a similar increase for the combination of dairy cattle urine and fertiliser, 367, 217 t including increases of 161,799 t for urine and 205,418 t for fertiliser. The sheep contribution was calculated to have decreased by 124, 495 t including urine plus dung excreta decreasing by 136,233 t, while fertiliser increased by 11,738 t.
By 2010, with respect to 1990, total excreta and fertiliser were projected to have increased by 227,816 and 281,538 t, respectively (totalling 509,354 t). Between 2004 and 2010, the rate of increase of N application averaged 23,084 t per year or 87 % of that for the previous 14 years. For dairy cattle, compared to 1990, urine and fertiliser increased by 194,931 and 246,346 t, respectively, totalling 441,277 t. Compared to 1990, the sheep contribution decreased by 72,633 t including urine and dung excreta decreasing by 86,710 t, while fertiliser increased by 14,077 t. Between 2004 and 2010, the sheep contribution thus increased by 51,862 t including increases of 49,523 t for urine and dung and 2,339 t for fertiliser.
Nitrous oxide emissions inventories quantifying how the emissions are reduced below that which would have occurred in the absence of the use of the nitrification inhibitors
We developed three emissions inventory methods. Our first method, called method 1, is an aggregated N2O emissions inventory comparable to current calculations reported by government. The effects of nitrification inhibitors are calculated using ‘annualised’ revisions of emissions factors EF3(PRP) and EF1 and term FracLEACH. For the other two methods, separate calculations are done for October – April when nitrification inhibitors should not be used and May – September when nitrification inhibitors should be used because they will then be most effective. For method 2, the nitrogen applied to soils as excreta from grazing animals remains an aggregation of urine and dung, so calculations for October – April are the same as for method 1. For May – September, method 2 uses a second set of revised values for EF3(PRP) and EF1 and FracLEACH. Method 3 includes this disaggregation plus the excreta are disaggregated into urine and dung, so a third set of revised values for EF3(PRP), EF1, and FracLEACH are used for each of the two periods. To summarise:
Method 1: Completely aggregated, so comparable to the current method but DCD application uses ‘annualised’ revisions of emissions factors and leaching fraction
Method 2: Aggregated excreta but disaggregation into October – April calculations that are identical to method 1 but the May – September calculations, including DCD application, use a second set of revised emissions factors and leaching fraction
Method 3: Completely disaggregated with excreta separated into urine and dung and a third set of revised emissions factors and leaching fraction, one set for October – April and another for May – September including the effects of DCD application
The emissions calculations are tabulated in the Appendix (Tables A.4 – A.8) with a summary at the end (Table A.9). In 1990, by methods 1 and 2, we calculated that N2O emissions from pastoral soils totalled 31.2 Gg (Table A.9). For method 3, because lower values were used for EF3(PRP) of dung, this value was reduced to 18.5 Gg (Table A.9). For Kyoto Protocol accounting, a change of N2O emissions is calculated relative to the 1990 level. Consequently, the same method must be used to compute a change of emissions. For example, a change of emissions from 1990 to 2004 cannot combine the relatively large method 1 value for 1990 with the relatively small value for 2004 from method 3. Using methods 1 and 2, in the absence of nitrification inhibitors (DCD), there were emissions increases of 7.7 and 10.6 Gg by 2004 and 2010, respectively (Table A.9). During the year 2010, N2O emissions represent an average for the Kyoto Protocol’s five-year-long Commitment Period 1 (2008 – 2012). For method 3, the corresponding emissions increases were larger at 9.7 and 11.8 Gg (Table A.9).
As case studies, we now calculate changes in the emissions between 1990 and 2004 and 2010 if DCD was applied as recommended to all land grazed by dairy cattle. This allows us to utilise the excretion and fertiliser application rates discussed above. Method 1 does not include seasonal differences in N2O emissions (October – April versus May – September) or different emissions from urine and dung. Using method 1, there were emissions increases of 3.0 and 5.3 Gg by 2004 and 2010, respectively (Table A.9). Compared to the emissions increases in the absence of DCD, mitigation using DCD was 4.7 Gg in 2004 (7.7 – 3.0, Table A.9) and 5.3 Gg in 2010. For Commitment Period 1 according to method 1, DCD mitigation was 26.5 Gg (5.3 Gg per year multiplied by 5 years). For method 1, mitigation is proportional to the fraction of excreta and fertiliser exposed to DCD. For example, if DCD was applied to only half the dairy cattle excreta and fertiliser, the mitigation values would be halved (Commitment Period 1 mitigation becomes 13.2 Gg, half of 26.5 Gg). As a further illustration, when DCD was applied as recommended to all land grazed by all animals, emissions were actually reduced from the 1990 level by 3.2 and 1.2 Gg in 2004 and 2010, respectively (Table A.5).
Method 2 includes seasonal differences in N2O emissions (October – April versus May – September) but no difference made between emissions from urine and dung. If DCD was applied as recommended to all land grazed by dairy cattle, using method 2, there were emissions increases of 0.0 (that is, no increase above the 1990 level) and 1.8 Gg by 2004 and 2010, respectively (Table A.9). Compared to the emissions increases in the absence of DCD, mitigation was 7.7 Gg in 2004 and 8.8 Gg in 2010. For Commitment Period 1 according to method 1, mitigation was 44.0 Gg. For method 2, mitigation is also proportional to the fraction of excreta and fertiliser exposed to DCD.
Method 3 includes disaggregation of emissions into periods when DCD is not used (October – April) and when DCD is used (May – September). There is also disaggregation of excreta into urine and dung and different emissions factors applied to these components. If DCD was applied as recommended to all land grazed by dairy cattle, using method 3, there were emissions increases of 3.2 and 4.4 Gg by 2004 and 2010, respectively (Table A.9). Compared to the emissions increases in the absence of DCD, mitigation was 6.5 Gg (9.7 – 3.2, Table A.9) in 2004 and 7.4 Gg (11.8 – 4.4) in 2010. For Commitment Period 1 according to method 3, mitigation was 37.0 Gg. For method 3, mitigation is not simply proportional to the fraction of excreta and fertiliser exposed to DCD because of the disaggregation of excreta into urine and dung and their different EF3(PRP) values throughout the year.
Estimate future use of nitrification inhibitors and limitations to their use, and impact on emissions until the end of Commitment Period 1 to enable an estimate of future liabilities associated with N2O emissions
At present, there are many unknowns related to the future use of nitrification inhibitors. One example is the possibility of a public: private partnership whereby the involved cost is shared by farmers and the government on the basis that each gains from the investment. Farmers have the potential to gain through enhanced pasture production associated with improved nitrogen use efficiency, while the government can account for a financial liability associated with its ratification of the Kyoto Protocol. Nevertheless, we re-iterate that any predictions will probably have considerable, but at this stage, intractable uncertainty. Consequently, we do not believe this aspect of the objective can be constructively analysed here. However, we described some limitations of nitrification inhibitor use and the impact on emissions through Commitment Period 1 in our draft Part I report for this project and the previous section of this report, respectively.
Recommend how the revised factors should be monitored, including the long-term effectiveness of nitrification inhibitors
While Li and Kelliher (2005) proposed an underground method to monitor direct nitrous oxide emissions (EF3(PRP)) that allowed pasture grazing by farmed animals, this method has not yet been deployed operationally. The monitoring of nitrogen leaching (FracLEACH) includes spatial and temporal integration challenges that are beyond the scope of this report. We believe that monitoring should rely on the principles of observation and generalisation because measurements must be made at smaller space and time scales. After all, the emissions inventory relies on these principles. Consequently, we recommend that monitoring should be based on field trial, despite their limitations. We believe strongly that the site network approach used by de Klein et al. (2003) serves as a constructive model.
As mentioned earlier, there has been one New Zealand field trial that included dairy cattle grazing and quantified the effects of repeated use of DCD on pasture production and quality (Moir et al. 2007). For four years (2002 – 2006 at the Lincoln University Dairy Farm), DCD was applied to a Wakanui silt loam soil beneath grazed pasture at 10 kg ha-1 in early May in addition to dairy cattle urine (1000 kg N ha-1) with DCD applied again in early August. Comparisons were made with control plots that received no DCD. Each year, the DCD applications consistently corresponded with increased pasture herbage dry matter yield that averaged 21 % on an annual, whole paddock basis including urine patches and inter-urine areas. Pasture nitrogen, metabolisable energy and fibre contents were not affected by the DCD applications.
The production rate and number of animals (based on farm records) grazing on pastures treated and not treated with nitrification inhibitors and the soil’s drainage class (based on treated area soil inspections)
As stated in our Part I draft report, DCD use includes a requirement for accurate and verifiable records of the treated pasture/soils (land) area. Long-term data storage and availability for independent review are also required. A GPS system associated with application seems ideal. This system should be future proof and suitable for audit and accreditation.
Linked to the GPS record of land area covered by DCD application, farm records of additional information would be needed to determine the N loading rate onto soils. Firstly, nitrogen fertilizer rate on the land ‘treated’ with DCD needs to be recorded. Next, as described earlier in our Methods, the weight, production rate and number of grazing animals determines N excretion as urine and dung that is deposited onto soils during grazing. Hence farm records could be used to determine the number and type of animals (dairy cattle, beef cattle and sheep) ‘treated’ with DCD including grazing period on the farm for lambs and other animals slaughtered during the year. We acknowledge that access to farm records of production rate may be controversial and animal weights may not be available. Alternatively, the number and type of animals ‘treated’ with DCD could be compared to the national population, facilitating a proportional calculation of the treated animal’s N excretion rate.
An emissions factor, EF, that determines N2O emissions is the direct EF for excreta, called EF3(PRP). As described earlier, the EF3(PRP) data are cumulative values of direct N2O emissions over 5 to 10 months following an excreta application (fraction of applied nitrogen emitted to the atmosphere as nitrous oxide) based on field chamber measurements of the NzOnet field trials (Barton et al., 2000; de Klein et al., 2003, 2004; Sherlock et al., 2003a,b). The data for dairy cattle urine are given in Table A.1, including the soil’s drainage class. These data may be analysed to compute a statistic known as the geometric average, a robust measure of the central tendency. For the well-, imperfectly- and poorly-drained soils, the geometric average values of EF3(PRP) were 0.0061 (n = 7), 0.0063 (n = 3) and 0.0164 (n = 7), respectively. The well-drained and imperfectly-drained soils had virtually the same EF3(PRP) values, but the poorly-drained soils value was 2.7 times greater. We acknowledge the small samples sizes involved in these comparisons. To implement our comparisons, EF3(PRP) would need to be disaggregated by soil drainage class. Earlier, at the national scale, it was determined that 74 % of the grazed pasture land area has well-drained soils, while 17 and 9 % of the area has imperfectly- and poorly drained soils, respectively (Sherlock et al. 2001). For the land area covered by DCD application, determination of the soil’s drainage class could be done by three methods. Firstly, a high-level assessment may be possible using soil maps. However, the reliability of this approach may vary widely at the scale of individual paddocks. Alternatively, the land manager is probably best able to make the assessment based on experience from observation (for example, after heavy rainfall, poor drainage is indicated by an area prone to flooding). As a refinement, soils in the treated area could be examined following excavation but a sampling strategy and assessment protocol would need to be developed. As a rule of thumb, it is reckoned that about half of the spatial variance in the variables that determine drainage is likely within one metre of a measurement or inspection point. Consequently, though we can envisage reasonable, clear and unambiguous criteria for separation of inspected soils into well-, imperfectly- and pooly-drained classes, we are wary about including the need to dig holes and inspect soils throughout the area treated with DCD. This could involve a tremendous amount of work on farms, and hence expense, and the information acquired seems likely to be uncertain.
Contact for Enquiries
Sustainable Land Management and Climate Change
MAF
Pastoral House
25 The Terrace
PO Box 2526, Wellington
Tel: 0800 CLIMATE (254 628)
Contact this person
