3 , 4-Dimethylpyrazole Phosphate ( DMPP ) Reduces N 2 O Emissions from a Tilled Grassland in the Bogot á Savanna

Grasslands are subject to a wide range of land management practices that influence the exchange of the three main agricultural greenhouse gases (GHGs) that are related to agriculture: carbon dioxide (CO2), nitrous oxide (N2O), and methane (CH4). Improving nitrogen fertilization management practices through the use of nitrification inhibitors (NIs) can reduce GHGs emissions. We conducted a field experiment at the Colombian Agricultural Research Corporation with four fertilization treatments: urea (typical fertilizer used in this region), ammonium sulfate nitrate (ASN), ASN plus the NI 3,4-dimethylpyrazole phosphate (ASN+DMPP), and an unfertilized control. The highest grassland yields (1956 and 2057 kg DM ha−1, respectively) and apparent fertilizer nitrogen recoveries (34% and 33%, respectively) were generated by the conventional urea fertilizer and ASN+DMPP. Furthermore, the use of ASN+DMPP reduced the N2O emissions that were related to N fertilization to the level of the unfertilized treatment (ca. 1.5 g N2O-N ha−1), with a significant reduction of N-yield-scaled N2O emissions (ca. 20 g N2O-N kg N uptake−1). These results support the application of DMPP as an alternative strategy to increase grassland yield while simultaneously reducing the environmental impact of N fertilization.


Introduction
Agriculture is responsible for just under a quarter (10)(11)(12) Gt CO 2 -eq year −1 ) of global anthropogenic greenhouse gas (GHG) emissions [1].Among the GHGs that were emitted through human activity, carbon dioxide (CO 2 ), methane (CH 4 ), and nitrous oxide (N 2 O) have the greatest links with intensive agriculture, mainly because of deforestation, livestock emissions, and soil and nutrient management [1,2].
The atmospheric concentration of N 2 O has increased by 20% over the last two and a half centuries, from 271 ppb in 1750 to 324.2 ppb in 2011 [3].Very few studies have been conducted in Colombia dealing with GHGs emissions from grasslands; however, GHGs emissions from agriculture are estimated to be 158.6 Gt CO 2 -eq year −1 , of which, 10% corresponds to N 2 O [4].The main source of N 2 O in agricultural soils comes from microbial activity in processes such as nitrification, denitrification, and nitrifier denitrification [5,6].Therefore, the use of nitrification inhibitors (NIs) has become an increasingly important strategy in recent decades to help reduce N 2 O emissions [1].3,4-dimethylpyrazole phosphate (DMPP) is one of the most widely used NIs.It is incorporated pre-blended in N fertilizers and delays the bacterial oxidation of ammonium (NH 4 + ) by inhibiting the activity of the ammonia monooxygenase enzyme [7].CO 2 is a major GHG.Its atmospheric concentration has increased by 40% since 1750 and from 278 ppm to 390.5 ppm in 2011 [3].It is generated by natural processes, such as respiration and decomposition of organic matter and by anthropogenic processes.Agricultural CO 2 emission is considered to be almost neutral because it is also associated with annual cycles of carbon fixation through photosynthesis.
The atmospheric concentration of CH 4 has increased by 150% since 1750 and from 722 ppb to 1803 ppb in 2011 [3].The main sources of agricultural emissions of CH 4 are the intestinal metabolism of domesticated ruminants, biomass burning, and production systems in flooded soils [8], because methanogenic bacteria generate CH 4 in anoxic environments.Contrastingly, grasslands and arable lands act as sinks because methanotrophic bacteria oxidize and therefore consume CH 4 in aerobic environments.When considering that methane monooxygenase can also oxidize NH 4 + [8], increasing soil NH 4 + concentration through nitrogen fertilization may induce an increase in the net soil CH 4 emission into the atmosphere.Managed grasslands are one of the most extensive agricultural systems in the world, occupying more than 25% of the global land surface [9].Depending on edaphoclimatic conditions fertilization, and the type of management, grasslands can be significant sources of CO 2 , CH 4 , and N 2 O emissions [1].Thus, the intensification of these managed productive systems with large inputs of nitrogen (N) fertilizers can lead to serious environmental risks, mainly N 2 O emissions.Soil tillage is another management practice that is used to increase grassland yield by introducing improved grass species.It can affect, amongst other parameters, the size and distribution of soil organic matter aggregates, as well as soil porosity, and therefore its aeration and drainage.Accordingly, as a result, it can influence the N 2 O and CO 2 fluxes [10].
Regarding grassland species, perennial ryegrass is one of the main forage grasses sown in the world due to its rapid recovery after grazing, its forage quality, its adaptability to different soil conditions and its productivity over long growing seasons [11,12].Forage quality is based on parameters such as nutritional composition, crude protein (CP) content and dry matter digestibility [11].Dry matter digestibility is related to the proportion of insoluble fraction of the fiber that can be assimilated by livestock.Both CP content and dry matter digestibility are primarily determined by the genetic characteristics of the grass and its maturity, while the application of N fertilizers is the most widely used strategy to increase CP content.
According to the National Administrative Department of Statistics, in 2014, around 80% of the agricultural surface in Colombia was assigned to grasslands and forage production [13], and Colombia constitutes one of the major consumers of fertilizers in the Latin America and Caribbean region.While the most commonly used N fertilizer in the world is urea, which in 2013 and 2014 represented more than half (57%) of the global N fertilizer demand [14], in Colombia, in 2016, ca.70% of the imports of N fertilizers were urea [15].Thus, the objective of this study was to evaluate the use of ammonium sulfate nitrate (ASN) in combination with DMPP as an alternative to the conventional application of urea in a tilled grassland in the Bogotá savanna.The effect of the different fertilization management strategies was measured in terms of their environmental impact (GHGs emissions) and forage yield and quality.

Study Area
The field experiment was carried out in 2013 at the Colombian Agricultural Research Corporation (CORPOICA) in a long-term permanent grassland (eight years old) that was located in the Bogotá savanna (4 • 42 N, 74 • 12 W), with an average annual temperature of 15 • C and an average annual Agronomy 2019, 9, 102 3 of 13 rainfall of 660 mm. Figure 1F shows the rainfall during the study period.The soil had a loamy texture and it was classified as a Typic Haplustands (Table 1).
depth) on the 18th September.Perennial ryegrass (Lolium perenne L. var.Samson) was sown at a density of 50 kg ha −1 .On the 3rd October, a randomized block design was established with an individual plot size of 10 m 2 .In all treatments, N-fertilizer was applied at a rate of 80 kg N ha −1 on 11th October.Three different fertilizers were used and four treatments were then evaluated: urea, ammonium sulfate nitrate 26% (ASN), a combination of ASN with DMPP (ASN+DMPP) that was marketed under the name of Entec ® 26 (trademark registered by EuroChem Agro GmbH, Mannheim, Germany), and an untreated control.The N content in ASN was 7.5% as nitrate and 18.5% as ammonium, and the DMPP content in Entec ® 26 was 0.8% of the total ammonium content.The experiment was initiated with mechanical tillage (ploughing and harrowing) (0-10 cm depth) on the 18th September.Perennial ryegrass (Lolium perenne L. var.Samson) was sown at a density of 50 kg ha −1 .On the 3rd October, a randomized block design was established with an individual plot size of 10 m 2 .In all treatments, N-fertilizer was applied at a rate of 80 kg N ha −1 on 11th October.Three different fertilizers were used and four treatments were then evaluated: urea, ammonium sulfate nitrate 26% (ASN), a combination of ASN with DMPP (ASN+DMPP) that was marketed under the name of Entec ® 26 (trademark registered by EuroChem Agro GmbH, Mannheim, Germany), and an untreated control.The N content in ASN was 7.5% as nitrate and 18.5% as ammonium, and the DMPP content in Entec ® 26 was 0.8% of the total ammonium content.

GHGs Emissions
GHGs emissions were measured using dark closed chambers (headspace of 0.01 m 3 ) and the technique that was described by Chadwick et al. [16].The samples were taken between 9.00 and 12.00 am to avoid the effects of daytime variation when determining daily emission rates.By previously measuring gas accumulation over a period of 30 min, the linearity of GHGs emissions was validated.Four chambers were placed randomly in each plot and the fluxes from two chambers were alternatively measured every sampling day.The initial gas concentration was determined by sampling 30% of the chambers.The concentration of each of the emitted gases (N 2 O, CO 2 , and CH 4 ) was determined using a gas chromatograph (Agilent, 7890A; Agilent Tech., Wilmington, DE, USA) that was equipped with an electron capture detector (ECD) for N 2 O detection, a flame ionization detector (FID) for CH 4 , and a methanizer to convert CO 2 to CH 4 .A capillary column (IA KRCIAES 6017: 240 • C, 30 m × 320 µm; Agilent Tech., Wilmington, DE, USA) was used and the samples were injected by means of a headspace autosampler (Teledyne Tekmar HT3; Teledyne Tech., Manson, OH, USA).The standards for the three gases were stored and then analyzed at the same time as the samples.Soil temperature and water content (0-10 cm depth) were also measured each time that gas samples were collected.
Cumulative GHGs emissions were estimated throughout the crop cycle using the trapezoidal integration method (linear interpolation and numerical integration between sampling times) [17].In the fertilized treatments, the N 2 O emission factor (EF) was calculated as a percentage of the N fertilizer that was applied by subtracting the cumulative N 2 O-N emission of the unfertilized treatment from that of the fertilized treatment and dividing the result by the total amount of N applied.Cumulative N 2 O and CH 4 emissions were converted to CO 2 equivalents following the recommendations of the Intergovernmental Panel on Climate Change (IPCC) [18], using global warming potential (GWP) factors of 298 for N 2 O and 25 for CH 4 .

Soil Mineral N and Soil Water Content
Soil mineral N content was calculated from a sample that was taken from a mixture of three subsamples of soil per plot (0-10 cm depth and 2.5 cm diameter) before and after the application of the fertilizer, and at a frequency of one or two weeks after the application.100 g of soil were extracted with 200 mL of KCl (1 M).The extracts were filtered through Sep-Pak C18 filters (Waters; Waters Corp., Milford, MA, USA) and then frozen until the day of measurement.The concentrations of nitrate (NO 3 − ) and ammonium (NH 4 + ) in the extracts were determined by colorimetry, using the method that was described by Cawse [19] and the Berthelot reaction (indophenol method), respectively.Soil water content was gravimetrically determined and then expressed as the percentage of water-filled pore space (WFPS), which was calculated as Linn and Doran [20]: where: θv = volumetric percent of water content = (% gravimetric water content) × soil bulk density TP = percent of total soil porosity = [1 − (soil particle density/soil bulk density)] × 100 The standard value of 2.65 Mg/m 3 was assumed as soil particle density.The used value of soil bulk density was 0.96 Mg/m 3 , which was determined using an intact core soil sampler (0-10 cm depth and determining its weight after drying.

Ryegrass Yield and Quality, and Nitrogen Use Efficiency
Ryegrass was harvested on 10th December.Two subsamples of 2 m 2 per plot were randomly selected and cut to a height of 2 cm above the soil.The samples were dried at 70 • C for 48 h before measuring the dry matter (DM) yield.A segment of the dried samples was used to determine the plant N content by the Kjeldahl method.CP content was calculated, multiplying the empirical coefficient of 6.25 to the N Kjeldahl content [21].Forage digestibility was estimated by determining the acid detergent fiber (ADF) content using the standardized method [21].
The apparent nitrogen recovery (ANR) rate is a means of determining N fertilization efficiency.It was calculated from the yield (DM) and plant N content based on the following formula: where: N F = N uptake in plants from fertilized treatments N U = N uptake in plants from the unfertilized control treatment N Ap = amount of N applied.Nitrogen use efficiency (NUE) was also calculated, as follows: where: DM F = dry matter obtained from fertilized treatments DM U = dry matter obtained from the unfertilized control treatment

Statistical Analysis
The means and standard errors are presented for each parameter.One-way ANOVA was performed using IBM SPSS 21.0 (IBM Corp., Armonk, NY, USA).Duncan's multiple-range test and the least significant difference (LSD) were used to determine significantly different means.Pearson's correlation coefficient was used to determine the relationship between the CO 2 fluxes and soil parameters.

N 2 O Emissions
Under the conditions of the study, daily N 2 O fluxes did not exceed 32 g N 2 O-N ha −1 day −1 from the unfertilized treatment, whereas the maximum emission rate was up to seven times higher from the fertilized treatments, reaching a value of 237 g N 2 O-N ha −1 day −1 in the treatment with urea (Figure 1A).For the fertilized treatments, the highest emission rates were obtained 30 and 42 days after sowing, coinciding with WFPS values of 95% and 74%, respectively (Figure 1F).
Cumulative N 2 O emissions in the whole experiment (60 days) (Table 2) for the ASN+DMPP treatment were 4.0 and 2.2 times lower than for the urea and ASN treatments, respectively.Thus, DMPP efficiency mitigating N 2 O emissions could be verified.Additionally, the emission factor (EF), which represents the percentage of the applied N-fertilizer that has been lost in the form of N 2 O, was 86% and 71% lower with the use of ASN+DMPP with respect to the use of urea and ASN (Table 2).Daily CO 2 fluxes ranged from 25 to 145 kg CO 2 -C ha −1 day −1 (Figure 1B).Under the conditions of this study, there was a significant positive correlation (r = 0.526; p < 0.01) between CO 2 flux and WFPS (Figure 2A).We observed that even small temperature changes could enhance CO 2 emissions, because the small oscillations in soil temperature between 14.9 and 17.4 • C showed a positive correlation with CO 2 emission rates (r = 0.638; p < 0.01) (Figure 2B).Despite a slight increase in the daily CO 2 emission rates for the fertilized treatments (Figure 1B), the cumulative CO 2 emissions did not present significant differences between the treatments (Table 2), with the average value for the four treatments being 3555 kg CO 2 -C ha −1 .

CO2 and CH4 Emissions
Daily CO2 fluxes ranged from 25 to 145 kg CO2-C ha −1 day −1 (Figure 1B).Under the conditions of this study, there was a significant positive correlation (r = 0.526; p < 0.01) between CO2 flux and WFPS (Figure 2A).We observed that even small temperature changes could enhance CO2 emissions, because the small oscillations in soil temperature between 14.9 and 17.4 °C showed a positive correlation with CO2 emission rates (r = 0.638; p < 0.01) (Figure 2B).Despite a slight increase in the daily CO2 emission rates for the fertilized treatments (Figure 1B), the cumulative CO2 emissions did not present significant differences between the treatments (Table 2), with the average value for the four treatments being 3555 kg CO2-C ha −1 .
Regarding CH4 fluxes, maximum uptake rates of around −5 g CH4-C ha −1 day −1 and a maximum emission rate to the atmosphere of 4 g CH4-C ha −1 day −1 (Figure 1C) were observed.

Ryegrass Yield and Quality
The ryegrass yields were close to 2000 kg DM ha −1 in the fertilized treatments sixty days after fertilization (Table 3).
With respect to forage quality, the CP content can widely vary in ryegrass species, from 8% to 27% [22].In our study, this content was already as high as 26% in the unfertilized treatment (Table 3) and a significant increment of this content was obtained with the use of urea.Another key factor in forage quality is digestibility, which is evaluated through ADF content.Neither fertilization nor DMPP application affected the concentration of ADF (Table 3).
Regarding the efficiency of N uptake (ANR), as DM yield was higher in the urea and ASN+DMPP treatments, and the CP content was higher in the urea treatment; the ANR was higher in both urea and ASN+DMPP treatments, with values of ca.34% (Table 3).When ASN was applied without an inhibitor, the ANR was as low as 17%.Regarding CH 4 fluxes, maximum uptake rates of around −5 g CH 4 -C ha −1 day −1 and a maximum emission rate to the atmosphere of 4 g CH 4 -C ha −1 day −1 (Figure 1C) were observed.

Ryegrass Yield and Quality
The ryegrass yields were close to 2000 kg DM ha −1 in the fertilized treatments sixty days after fertilization (Table 3).
With respect to forage quality, the CP content can widely vary in ryegrass species, from 8% to 27% [22].In our study, this content was already as high as 26% in the unfertilized treatment (Table 3) and a significant increment of this content was obtained with the use of urea.Another key factor in forage quality is digestibility, which is evaluated through ADF content.Neither fertilization nor DMPP application affected the concentration of ADF (Table 3).
Regarding the efficiency of N uptake (ANR), as DM yield was higher in the urea and ASN+DMPP treatments, and the CP content was higher in the urea treatment; the ANR was higher in both urea and ASN+DMPP treatments, with values of ca.34% (Table 3).When ASN was applied without an inhibitor, the ANR was as low as 17%.

N 2 O Emissions
We observed that N 2 O emissions from the urea treatment were 1.8 times higher than from ASN (Table 2), as reported by Clayton et al. [23], who also observed 2.0 times higher cumulative N 2 O emissions in urea-fertilized plots when compared to ASN-fertilized plots.Davidson [24] reported that denitrification is a predominant process when soil moisture remains at a WFPS of above 70% because of the lower availability of O 2 .Moreover, under high soil NO 3 − content conditions (Figure 1E), the competitive behavior of NO 3 − versus N 2 O as a terminal electron acceptor during denitrification can inhibit the reduction of N 2 O to N 2 [25][26][27], generating the high daily N 2 O emission rates that were observed in our study for the fertilized treatments.Another factor that could have also induced the observed N 2 O emissions is the high soil organic carbon content of 9%.For example, in our unfertilized treatment, N 2 O cumulative emissions were up to 38 times higher than those that were reported by Misselbrook [28] in different United Kingdom (UK) soil types with soil organic carbon contents <5%.In this regard, Bouwman [29] measured a significant increase in N 2 O emissions in soils with high organic carbon contents (>3%), indicating that this organic carbon effect is favored by a fine soil texture and a neutral to slightly acidic pH, as was the case of the present study.In fact, under the same climatic conditions and a similar N fertilizer dose (70 kg ha −1 ), but with an organic carbon content of 4.2%, Huérfano et al. [30] recorded N 2 O emissions rates for fertilized treatments that were four times lower than those that were found in the present study.Even under high soil water content conditions (90% WFPS), Ruser et al. [27] observed peak N 2 O emission rates.They also attributed this to the influence of the availability of soil organic carbon (1.63%), since this availability enhances the consumption of O 2 , thus favoring the formation of anoxic zones that promote the emission of N 2 O, especially when NO 3 − is readily available [31], as was the case in our experiment.
With respect to the effect of temperature and WFPS on the efficiency of DMPP, in Atlantic grasslands, Merino et al. [32] reported that DMPP was more efficient in reducing N 2 O emissions at temperatures ranging between 6 to 11 • C than at temperatures above 16 • C, and Menéndez et al. [33] reported that, at 80% WFPS, DMPP's capacity to reduce N 2 O emissions decreased when soil temperature increased from 10 to 20 • C. In the present study, we have confirmed that DMPP efficiently reduces N 2 O emissions at soil temperatures close to 16 • C and prevalent conditions of WFPS >70%.
It is also worth highlighting that the efficiency that was observed after DMPP application was the highest which could possibly be obtained, based on the assumption that DMPP could reduce N 2 O emissions down to the level of the unfertilized control.
When calculating N 2 O emission due to fertilization in terms of emission factor (EF), we must bear in mind that the N 2 O emission factors that were obtained in this study should not be directly compared to the default value that was proposed by the IPCC [34], because they were calculated for a short period of 60 days, corresponding to one ryegrass cut.Regardless, when considering the non-seasonality and therefore the low climatic variation of the study area (the Bogotá savanna can produce up to six cuts per year), we can assume that the emission factor that was obtained for a single cut is very similar to what would be obtained for one year when applying fertilizer after each cut.Our results show that the N 2 O emission factors for the urea and ASN treatments were 8.5 and 4.1 times higher, respectively, than the 1% default value that was proposed by the IPCC [34].However, the ASN+DMPP treatment, whose EF was 86% and 71% lower than in the urea and ASN treatments, respectively, gave an EF of 1.2% (Table 2).This value is within the range that was proposed by the IPCC, thereby highlighting DMPP's capacity to reduce high N 2 O emissions that are induced by fertilization.

CO 2 and CH 4 Emissions
Kiese and Butterbach-Bahl [35] and Davidson et al. [36] reported a negative correlation between CO 2 emissions and soil moisture content that is above 60% WFPS, which they attributed to the anoxic conditions prevailing at high moisture content.By contrast, other authors, such as Ruser et al. [27] and Frank et al. [37], did not observe a clear influence of soil moisture content on CO 2 emissions.In the evaluated edaphoclimatic conditions, the positive correlation between CO 2 flux and WFPS (Figure 2A) indicates that the soil respiration rate was stimulated by the increase of soil water content in the range of 58% < WFPS < 94%.In this sense, Doran et al. [38] proposed a model to describe the relationship between WFPS and respiration in soils with different textures, defining that the highest respiration rates occur in the range of 40% < WFPS < 70%; therefore, the positive correlation that we observed could be explained, because most of the measurements were determined in that interval.Temperature is also a determining factor of CO 2 fluxes, because it favors the microbial activity of the soil [39].In this experiment, we have been able to determine that even small changes in this variable can favor the increase of CO 2 emissions, since the small oscillations of temperature of the soil (Figure 1F) presented a significant positive correlation with the emissions of this gas (r = 0.638) (Figure 2B).
Agricultural soils have demonstrated the capacity to oxidize CH 4 at non-flooded aerobic conditions [40,41].In our study, WFPS was in the range of 70-95% during most of the study time, therefore the O 2 availability was low but it did not generate anaerobic conditions.Hence, it was not either enough to favor CH 4 consumption [42].Thus, the daily emissions were close to zero, fluctuating between −5 to 4 g CH 4 ha −1 day −1 (Figure 1C).The cumulative CH 4 emissions were similar for all treatment (Table 2).Accordingly, under the edaphoclimatic conditions of this study, and according to Huérfano et al. [40], we did not observe any effect of N fertilization or DMPP application on CH 4 emissions.

Effect on Global GHGs Emissions
Regarding environmental implications, the GWP has been established as a relative measure that reflects, on average, how long a GHG remains in the atmosphere and how strongly it absorbs energy.For conducting national inventories of GHGs emissions, the GWP is calculated in CO 2 equivalents and, when establishing a horizon of time of 100 years, the GWP of CO 2 is 1, of CH 4 is 25, and of N 2 O is 298 [18].In the present study, the GWP was determined based on the total emission of the three GHGs (Table 2).Despite that the GWP (expresed in CO 2 equivalents) of N 2 O is considerably larger than the GWP of CO 2 , the small amount of grams of N 2 O emitted with respect to CO 2 , makes the emission of N 2 O only contribute 4% and 6-18% of the total GWP in the unfertilized and the fertilized treatments, respectively.Therefore, CO 2 emissions were the determining factor in the magnitude of the GWP.When considering that the evaluated treatments did not exert any effect on CO 2 emissions, neither did they exert any efect on the GWP.

Effect on Yield and Quality
The efficiency of fertilization was measured in terms of NUE, where 1 kg of N fertilizer induced an increase of ca.6 kg DM ha −1 in every treatment.Thus, no significant differences in NUE were observed, which was probably due to the high soil nitrate content (Figure 1E) prior to fertilization.On the other hand, the ANR values were low (Table 3) when compared to a two-year study with eight varieties of ryegrass and fertilizer additions above 100 kg N ha −1 , where ANR values of ca.70% were obtained [43].As was mentioned before, ANR was higher in the urea and ASN+DMPP treatments.When comparing these two treatments, the use of DMPP was shown to compensate for the decrease in CP content in ASN and ASN+DMPP treatments through an increase in yield (kg DM ha −1 ).
On the other hand, given that, in our trial, NH 4 + remains available for the first 15 days after fertilization (Figure 1D), our results confirm the tolerance of ryegrass to ammonium nutrition that was reported by Belastegui-Macadam et al. [44].As a conclusion, N fertilization that was based on urea or ASN+DMPP could be good alternatives in terms of yield and nitrogen uptake (ANR) under our experimental conditions.

Efficiency of the N Fertilization Managements
The high pre-fertilization soil NO 3 − content that was observed in the unfertilized treatment (Figure 1E) could be the result of tillage that oxygenated the soil and would have subsequently induced the mineralization of organic matter, as observed by Pinto et al. [45] after tilling a permanent pasture.Despite the high soil NO 3 − content that was measured before fertilization, we still observed yield response to fertilization, and the urea and ASN+DMPP treatments presented significantly higher yields, even in comparison to the ASN treatment.
Van Groenigen et al. [46] established a parameter called "N-yield-scaled N 2 O emission" based on the expression of N 2 O emissions in relation to crop productivity (above-ground N uptake).When considering that the fertilizer treatments in our experiment had different effects on yield and CP content, we also calculated the relation between N 2 O emissions and the total DM that was harvested as the "yield-scaled N 2 O emission".Each treatment that was assessed in our study was observed to have a similar effect on both N-yield-scaled N 2 O emission and yield-scaled N 2 O emission (Table 4).While for the urea treatment the N-yield-scaled N 2 O and yield-scaled N 2 O emissions presented values of 89 g N 2 O-N kg N harvested −1 and 4 g N 2 O-N kg DM harvested −1 , respectively, these values were reduced by 36% and 42% when ASN was applied.Moreover, in the ASN+DMPP treatment, they were further reduced to the level of the unfertilized treatment (17-23 g N 2 O-N kg N harvested −1 and 0.7-1 g N 2 O-N kg DM harvested −1 ).Provided that the N-yield-scaled N 2 O emission is a relevant factor when assessing the environmental cost (in terms of N 2 O) of food production [47], the use of DMPP in the Bogotá savanna grasslands constitutes an effective management practice to increase the efficiency of N fertilization, while simultaneously reducing its negative environmental impact.Conventional nitrogen fertilization management with urea and the use of ASN+DMPP generated the highest ryegrass yields and apparent nitrogen recoveries.However, in the case of the application of ASN+DMPP, the significant reduction of N-yield-scaled N 2 O emissions down to the unfertilized control level is evidence that the use of DMPP can effectively increase grassland yield while also reducing the environmental impact of N fertilization.

Conclusions
This study provides the first data on GHGs emissions from grasslands that were grown under the climatic conditions of the Bogotá savanna.This is an important starting point for the measurement and assessment of GHGs mitigation practices to be implemented in the most important productive system in this region.Both the conventional nitrogen fertilization management with urea and the use of ASN+DMPP generated the highest ryegrass yield and apparent nitrogen recovery.However, the significant reduction of N-yield-scaled N 2 O emissions by ASN+DMPP down to the unfertilized control level support the effectiveness regarding the use of DMPP to increase grassland yield while reducing the environmental impact of N fertilization.Regarding forage quality, neither the application of nitrogen nor the use of DMPP affected the digestibility of the forage in ADF content terms.

Table 4 .
N-yield-scaled N 2 O emissions and yield-scaled N 2 O emissions.Values in cells with different letters are significantly different (p < 0.05; n = 4).
* Organic carbon, ** Effective cation exchange capacity (Sum of exchangeable basic and acidic cations).

Table 2 .
CO 2 , CH 4 , and N 2 O cumulative emissions, global warming potential (GWP), and N 2 O emission factor (EF).In brackets: percentage of reduction of EF by ASN+DMPP.Values in cells with different letters are significantly different (p < 0.05; n = 4).