The mobility of nitrification inhibitors under simulated ruminant urine deposition and rainfall: a comparison between DCD and DMPP

Urine patches within pasture soils are hotspots for nitrogen (N) cycling and losses, where nitrification inhibitors (NI) offer a means of reducing such losses. Within urine influenced soil, more research has been conducted for dicyandiamide (DCD) than 3,4-dimethylpyrazole phosphate (DMPP). Differences in the efficacy of these NI are often ascribed to a greater mobility of DCD, which may lead to spatial separation from NH4 + and nitrifying microorganisms. We tested the mobility of 14C-labelled DCD and DMPP relative to sheep urine-derived NH4 + in soil columns of contrasting texture and organic matter content, following simulated rainfall. We also assessed factors influencing the vertical mobility of these NI in soils, including solubility, sorption/desorption processes and microbial degradation and uptake. Following 40-mm rainfall, without the presence of sheep urine, the distribution of both NI were similar in the soil columns; however, there was a greater retention of DCD compared to DMPP in the top 1 cm. Both NI appeared to co-locate well with urine-derived NH4 +, and the presence of sheep urine altered the leaching profile of the NI (compared to rainfall application alone), but this effect was inhibitor and soil-type dependent. A greater sorption to the soil matrix was observed for DCD in comparison to DMPP in all three studied soils, and the presence of urine generally increased desorption processes. Of the NI applied to the soil columns, 18–66 % was taken up within 30 min by the microbial community. However, only small amounts (<1 %) were mineralized during this period. In conclusion, due to the greater adsorption of DCD as opposed to DMPP and similarity in the degree of co-location of both NI with urine NH4 +, the results of this study suggest that differences in microbial uptake and degradation may be more important parameters for explaining differences in the efficacy of reducing nitrification. Further work is required to determine the comparative efficacy of both NI in reducing nitrification rates under field conditions in a range of soil types and environmental conditions.


50
In pasture soils, high loadings of nitrogen (N) are deposited within ruminant urine patches and 51 these sites are particularly vulnerable to losses of N to the environment. Typically, 20% of 52 deposited urinary-N is leached as NO3 -, 13% is volatilised as NH3 and 2% is emitted as the 53 greenhouse gas N2O (Selbie et al. 2015). While N2O constitutes a small agronomic loss in terms 54 of magnitude of N, having nearly 300 times the global warming potential of CO2 (IPCC 2007), 55 it accounts for 46% of agricultural greenhouse gas emissions (Smith et al. 2007). The 56 agricultural sector will need to contribute to decreasing emissions (Misselbrook et al. 2014) in 57 order to achieve targets (80% reduction from 1990 baseline levels by 2050) set by the UK, and 58 other governments. Reducing N loss via NO3leaching and improving N use efficiency would 59 translate to a direct economic benefit for farmers and reducing N2O emissions from grasslands 60 could contribute to decreasing emissions from the livestock sector.

111
The objective of this study was to obtain information on how the combination of NI 112 characteristics and soil conditions can affect the mobility and co-location of NH4 + and NI in 113 soils, with ruminant urine as the source of NH4 + . We investigated physicochemical (solubility 114 and sorption/desorption) and biological (microbial uptake and degradation) factors 115 influencing the vertical mobility of DCD and DMPP, in soil columns of contrasting texture 116 and organic matter following a 40 mm rainfall event, with and without the presence of sheep 117 urine. We hypothesised that 1) DCD would move further down the soil profile than DMPP 118 following simulated rainfall, due to a greater sorption of DMPP, 2) a greater co-location 119 would be observed for DMPP with urine NH4 + , due to the lower mobility of DMPP, 3) the 120 presence of sheep urine would increase vertical movement and desorption of both NI, due to  had not been previously exposed to either DCD or DMPP. A summary of soil properties is 134 presented in Table 1. 135 Soil was sampled in triplicate (0-10 cm depth), sieved (< 2 mm) in order to reduce 136 sample heterogeneity and stored at 4°C until required. Soil moisture content was determined 137 by weight difference after oven drying (105°C), and organic matter was determined on dry soil  (Na + , K + and Ca 2+ ) were determined within 1:5 (w/v) soil-to-1 M NH4Cl extracts using a 150 Sherwood Model 410 Flame Photometer (Sherwood Scientific Ltd, Cambridge, UK).

151
Sheep urine was collected from Welsh mountain ewes fed a diet of 80% Lolium perenne 152 L. and 20% Trifolium repens L., where several urine samples from a single sheep were pooled.

153
The urine was frozen (unacidified) before use to avoid losses of N. The sheep urine had a pH 154 of 8.99 and an EC of 22 mS cm -1 ; the urine contained a total of 2.27 g N l -1 , 3.00 g organic C 155 l -1 , 1.71 g urea N l -1 , 44.9 mg NH4 + -N l -1 , 0.44 mg NO3 --N l -1 , 0.92 mg P l -1 , 7.16 g K l -1 , 1.11 156 g Na l -1 and 73.3 mg Ca l -1 . Properties were measured directly on the urine via the methods 157 described above and urea was measured using the method of Orsonneau et al. (1992). and Sapric Histosol soil columns, respectively (particle density was assumed to be 2.65 g cm -3 167 in the mineral soils and 1.4 g cm -3 in the organic soils; Rowell 1994). The bottom of the tubes 168 contained nylon mesh, to allow for drainage of leachate and to prevent any loss of soil. water movement down the soil profile. It should also be noted that these leaching rates also 175 approximate rates of water movement down preferential flow pathways in the soil profile under 176 lower rainfall events. Preliminary studies indicated that the wetting front generally reached, 177 but did not exceed the soil column length (15 cm).

Solubility of DCD and DMPP in water 197
To determine the water solubility of DCD and DMPP the OECD (1995) flask method was used.

198
Briefly, 5 g of NI was added to 10 ml of water (n = 3) and incubated at 30ºC on a rotary shaker 199 9 for 24 h. One replicate was then removed and incubated at 20ºC for 24 h with occasional 200 shaking, before centrifuging at 10 000 g. Samples were syringe filtered (0.2 µm) and analysed 201 for total dissolved C, as described above, and the amount of NI dissolved in the water 202 calculated. One of the remaining replicates was incubated for another 24 h at 30ºC and the final 203 replicate was incubated for an additional 48 h, before incubation at 20ºC for a further 24 h and 204 preparation of samples for analysis of dissolved C. This was conducted to ensure additional 205 time had no effect on the amount of NI dissolved.

266
Comparative vertical mobility of NI following simulated rainfall 267 The distribution of extractable DCD-14 C and DMPP-14 C following simulated rainfall was 268 generally similar within each soil type (Fig. 1). However, a greater retention of 14 C-DCD was 269 observed in comparison to 14 C-DMPP in the top 0-1 cm depth fraction of the sandy loam ( Fig.   270 1a) and sandy clay loam columns (Fig. 1b) respectively.

276
The presence of sheep urine reduced (p < 0.01) the quantity of extractable DCD-14 C 277 and DMPP-14 C in the top 1 cm of the sandy loam columns (Fig. 1a and d), increased (p < 0.01) 278 the amount of extractable DCD-14 C in the bottom 12-15 cm depth fraction and had no effect (p 279 > 0.05) on the extractable amount of DCD-14 C and DMPP-14 C in each remaining depth fraction.

280
The presence of sheep urine did not influence the extractable amount of DCD-14 C or DMPP-281 14 C in any studied depth fraction of the sandy clay loam columns (Fig. 1b and e). The presence 282 of urine had no effect (p > 0.05) on the amount of extractable DCD-14 C in each depth fraction 283 of the Sapric Histosol ( Fig. 1c and f). However, it decreased (p < 0.001) the extractable amount 3.54 and 47.9 ± 0.01% in the sandy loam, sandy clay loam and Sapric Histosol, respectively.

291
In conclusion, urine increased the total amount of DCD extracted from the soils, but had no 292 effect on DMPP.

294
Co-location of NI with urine-derived ammonium 295 In general, the distribution of both DCD-14 C and DMPP-14 C within the soil profile coincided 296 well with the urine-derived NH4 + (Fig. 2). A greater (p < 0.001) percentage of total column 297 extractable DCD-14 C in comparison to NH4 + was found in the top 1 cm in all three soil types 298 (Fig. 2a, b, and c), indicating a retention of DCD at the soil surface. A greater (p < 0.001) 299 13 percentage of total extractable NH4 + in comparison to DMPP-14 C was found in the 9-12 cm 300 depth fraction of the sandy loam columns (Fig. 2d). Greater (p < 0.001) amounts of total 301 extractable NH4 + in comparison to DCD-14 C were also found in the 9-12 cm depth fraction of 302 the sandy clay loam columns (Fig. 2e), indicating some dis-location of NI with NH4 + at depth.

303
No differences were observed at any depth fraction for DMPP-14 C and urine-NH4 + in the sandy 304 clay loam or the Sapric Histosol columns ( Fig. 2e and f,   Generally, the presence of urine increased total desorption (Fig. 5)  l -1 (p < 0.001; Fig. 5k) and 10 mg l -1 (p < 0.05; Fig. 5l). Interestingly, the same trend was not 332 observed for DCD-14 C in the same soil type ( Fig. 5i and j). Desorption of DCD-14 C in the CaCl2 333 matrix was greater (p < 0.05) in the sandy loam soil ( Fig. 5a and b) compared to the Sapric 334 Histosol ( Fig. 5i and j), but desorption was no greater (p > 0.05) in the sandy clay loam soil 335 (Fig 5e and f). In the urine matrix desorption of DCD-14 C was lower (p < 0.01) in the Sapric 336 Histosol ( Fig. 5i and j) compared to the sandy loam ( Fig. 5a and b) and sandy clay loam (   Histosol, but no such trend was observed for DCD-14 C. To consider the reasons behind these 447 results, a consideration of the soil, NI and urine properties are required.

19
The soils used in this study were all of a similar pH (Table 1), however the addition of 449 urine would have altered the soil pH and made conditions in the soil columns more alkaline.

450
As DCD is amphipathic, sorption has been shown to be pH dependent, where increases in pH 451 above pH 5 lead to increased sorption (Zhang et al. 2004). This may partially explain why the 452 vertical distribution of DCD-14 C was similar whether urine was present or not in the Sapric

453
Histosol. In the solubility assay, DMPP was found to be over 1.5 times more soluble in water The sandy loam and sandy clay loam soils had a similar CEC and organic matter content 462 (Table 1)

497
In this study even after 1 h, the microbial uptake of both inhibitors accounted for a large 498 proportion of that applied. Approximately 20% was taken up by soil microbes in the mineral 499 soils and > 50% of that applied was taken up by soil microbes in the Sapric Histosol, which 500 was likely to be a function of the greater microbial biomass in this soil. To be effective, the NI 501 would need to be acquired by the target microbial biomass (ammonium oxidizing bacteria and 502 archaea). Immobilisation into non-target microbial biomass could, therefore, equate to a fairly 503 large removal mechanism for these NI and this requires further investigation. No difference 504 was observed between DCD and DMPP in the proportion acquired by soil microbes in the 505 mineral soils. However, uptake was greater for DCD compared to DMPP in the Sapric Histosol 506 at the higher studied concentration. This suggests a slight preference of, or bioavailability of 507 DCD to the soil microbial community in the short-term.

508
The results of this study should be considered with care, as repacked soils were used