Predicting ammonia volatilization from fertilized pastures used for grazing

https://doi.org/10.1016/j.agrformet.2020.107952Get rights and content

Highlights

  • Volatilization from urea applied to grazed dairy pastures was modelled with DairyMod.

  • The complex process of volatilization was simulated by a simple modelling approach.

  • DairyMod is versatile for use across a range of scenarios with overall low model errors.

  • Modelling is attractive considering the difficulty, expense and uncertainty in field studies.

Abstract

Ammonia (NH3) volatilization from fertilised agricultural soils is driven by complex interactions between edaphic, climatic and plant canopy factors that can be difficult to measure or predict. We developed a simplified approach using default parameters in the DairyMod model to predict daily NH3 volatilization from urea applied to grazed dairy pastures. Several published datasets were used to validate the reliability of the model to reproduce key related soil processes in a whole farming systems framework. For the sites where monitoring for the experimental duration occurred, DairyMod simulated the main features of the observed NH3 emissions, with an overall predicted median of 4.1 kg/ha or 7% of applied N, compared to the measured median of 6.1 kg/ha for 12% of applied N fertilizer. There was an overall root mean square error (RMSE) of 0.9 kg NH3single bondN/ha/d and an overall mean prediction error (MPE) of 0.5 kg NH3single bondN /ha/d. However, there was high uncertainty in several of the datasets used which made it difficult to be conclusive about the validation. The simulation accuracy was improved using daily wind speed (collected on-site in field campaigns) as input to the evapotranspiration calculations. In cases of high certainty in the volatilization data, it was concluded that the model was useful for the analysis of N cycling in situations used for dairy farming without the need for a more complex mechanistic method with difficult-to-obtain parameters. DairyMod presents a simple but readily reproducible prediction of NH3 volatilization from urea application on pasture in intensive livestock farming systems compatible with the certainty of the model inputs and scale of model application. However, the collective understanding of NH3 volatilization in pasture based dairy systems is currently based on a limited number of often uncertain, short term plot studies in the absence of animals.

Introduction

Globally, NH3 volatilization losses from all fertilized agricultural soils average 18% of the nitrogen (N) applied (Pan et al., 2016). The annual global use of synthetic fertilizer N on grasslands is somewhere around 4.3 million tons, with estimated loss rates via volatilization as 6% for industrialized countries (Bouwman et al., 2002). Widespread volatilization leads to increased concentrations in the atmosphere and increased concentrations of wet and dry deposition. Understanding and minimising volatilization are important from an agricultural perspective because of the economic imperative to increase the efficiency of nitrogen (N) use and the environmental imperative to reduce contamination of the air with oxides of N —an indirect result of NH3 emissions. In places like Australia and New Zealand where emissions are low overall (Bouwman et al., 2002) and air quality remains relatively unimpaired by volatilization, N fertilizer usage in intensive agriculture such as dairying remains high (Stott and Gourley, 2016), thereby increasing the potential for N losses. Being able to predict changes in N losses, due to modifications to landuse, management of agricultural systems and climate variations for example, are key issues for sustainable landuse.

In an agricultural context, NH3 volatilization mainly occurs from urea fertilizer shortly after application, from the bulk soil ammonium (NH4+) pool in the top soil layer, and from livestock urine. Better understanding of volatilization across a wide range of edaphic, geographic and agricultural settings is, however, restricted because measuring volatilization is costly, difficult and can be prone to errors. Ensuring that the process of volatilization is correctly understood and described in a model is of high importance as it prevents errors in N inputs that would otherwise cascade through the model.

Ammonia volatilization is a complex biophysical process regulated by a combination of soil and meteorological factors which control dissolution of the fertilizer prill, hydrolysis and ammonification, and the exchange of gases from the soil to plants and the air. The primary soil factors influencing NH3 volatilization are: soil water content, soil pH, cation exchange capacity, organic matter, clay content, urease activity, as well as fertilizer rate and crop type (Bouwman et al., 2002). The relative importance of each factor depends on different soil types and edaphic settings. Meteorological factors that regulate volatilization include solar radiation, the ambient air temperature, wind speed, precipitation, relative humidity (Bouwmeester et al., 1985) and structure of the plant canopy (Denmead et al., 1976). Both sets of factors affect the NH3 concentration gradient from a surface to the air above.

The links between soil hydrology and volatilization have often been noted, even if not directly (e.g. Bouwman et al., 2002; Denmead et al., 1976; Ellington, 1986; Ernst and Massey, 1960; Harper et al., 1983b; Martin and Chapman, 1951). Laboratory experiments have shown that NH3 volatilization from urea increases with the water content of the soil (Black et al., 1987; Bouwmeester et al., 1985; Ernst and Massey, 1960). Indeed, micro-meteorological techniques for the quantification of volatilization use measurements of wind speed, wet bulb and dry bulb air temperatures, net radiation, heat fluxes etc. (Denmead et al., 1998; Leuning et al., 1985). These same factors are involved in the prediction of evapotranspiration (ET, Monteith, 1986; Penman, 1948) using the Penman-Monteith equation (Allen et al., 1998), where a single exchange coefficient (typically used to represent the transport resistance) is considered to be dependant on soil structure and moisture, as well as on air temperature and wind speed. Apart from some limited studies, e.g. Denmead et al. (1976), the links between volatilization and the movement of water vapour from plants and soil have not been adequately defined.

The relative importance of the many soil and meteorological factors to NH3 volatilization loss is unclear, with differing studies showing disagreement in ranking the importance of each (e.g. Bouwman et al., 2002; Fillery and Khimashia, 2016; Macnack et al., 2013; Schwenke et al., 2014). For example, many experiments found that N application rate, soil organic carbon content and soil texture were important as first-order driving variables; but Bouwman et al. (2002) found a lack of consistent relationships from their global meta-analysis. In another study Burchill et al. (2016) found no significant difference in field volatilization across a range of soil types under similar environmental conditions, i.e. soil factors could not explain the variations in volatilization. Therefore, it seems that the importance of the first-order driving variables and their inter-relationships may change over time (e.g. due to seasons, in response to rainfall, with plant growth, soil moisture status) so a whole system perspective that can account for soil x environment x management interactions is required. The interacting complexity between soil and meteorological regulating factors means that predicting volatilization is well suited to biophysical systems modelling.

Prediction of the effects of changes to a farming system or the effectiveness of management interventions for mitigation would be greatly aided by a modelling framework that encapsulated the complex process of volatilization. Adequately describing volatilization is therefore central to ensuring that the addition of N to the system is properly accounted for, because errors and uncertainty in the size of additions to the soil N pools will cascade throughout the model and may indicate increased nitrate leaching and/or denitrification (Bussink and Oenema, 1998; Snow et al., 2014). The intended scale and use of the modelled outcomes determine the balance between complexity and parsimony. More comprehensively mechanistic models can involve numerous parameters that are difficult to measure or require local calibration. For routine analysis of a farming system including a grazed pasture of high variability, a robust, flexible and reliable modelling approach is required. Such a robust yet simple model would assist with farm-level diagnosis of N requirements, design and decision making for pasture-based grazing systems. An added benefit of a whole-of-system modelling approach is that it should show how intervention in one part may affect NH3 losses from other parts of the system.

Consequently, the objective of this paper was to present a simplified method to simulate NH3 volatilization from fertilized, grazed pastures. The conditions of the methods were it had to be suitable for routine use at the paddock and farm scale, utilise only readily available information, and not require specific, local calibration. As a sub-model to be incorporated into a whole-farm model, it needed to be parsimonious, but of comparable complexity with the other factors in the main model, called DairyMod (Johnson, 2016). The dataset for validation came from a wide range of climate zones with locations from temperate to subtropical from latitudes −38° to −17°, from irrigated and dryland sites, and from paddocks with temperate or subtropical pasture species across a range of seasonal conditions. Thus although the datasets came from Australia, due to the diversity in climatic and edaphic conditions this provided for rigorous testing of the model. A sensitivity analysis was also carried out to understand the effect on model outcomes of some of the main assumptions about parameter values and data input.

Section snippets

Representation of NH3 volatilisation in DairyMod

DairyMod (v. 5.8.2, Johnson, 2016) was used in this study. DairyMod is a farming systems model that mechanistically simulates the N, hydrological and daily energy cycles. A full N cycle is described that involves the breakdown of organic matter as well as ammonification and nitrification. Further details of NH4+ adsorption, nitrification and organic matter dynamics are given in Johnson (2016)http://imj.com.au/wp-content/uploads/2016/05/DM_SGS_documentation.pdf, however only the volatilization

Model parameters

For the selected parameters assessed, the choice of the reference ET (mm/day) had the most significant effect on volatilization (Fig. 2). The NH3 volatilized at reference soil E, the% urea inputs volatilized at reference ET and the decay rate of the labile organic matter had less, albeit a sizeable effect on volatilization. The combination of evaporation factor (reference evapotranspiration in Eq. (1)) and emission factor (% urea losses in Eq. (3)) produced different predictions of long-term

Discussion

For farming systems models to be useful, representation of the full N cycle is required (Snow et al., 2014). Errors and uncertainty in soil N additions can cascade throughout the model. Therefore, an adequate description of NH3 volatilization from applied N in fertiliser and urine is important to ensuring that the transformations of the added N are properly described. Differences between observed and model-predicted volatilization will relate to assumptions in the estimation of

Conclusion

The accurate simulation of NH3 volatilization from pastures where urea fertilizer has been applied is challenging, because the causal relationships with major edaphic and meteorological driving forces are often poorly defined in field data sets. As a result, the knowledge of NH3 emissions via volatilization remains limited for a range of farming systems where pastures fertilized with urea are used for grazing. this study demonstrated that a sub-model developed for DairyMod, although simplistic,

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The ‘More Profit from Nitrogen: enhancing the nutrient use efficiency of intensive cropping and pasture systems’ project was supported by funding from Dairy Australia and the Australian Government Department of Agriculture and Water Resources as part of its Rural R&D for Profit program. Author contributions: A.S. conceptualized the study, carried out the modelling and analysis, and wrote the paper; I. J. developed the model, contributed to the writing, M. B., S. K. L., G. S. and H. S. provided

References (45)

  • V.O. Snow

    The challenges – and some solutions – to process-based modelling of grazed agricultural systems

    Environ. Model. Softw.

    (2014)
  • K.J. Stott et al.

    Intensification, nitrogen use and recovery in grazing-based dairy systems

    Agric. Syst.

    (2016)
  • T. Al-Kanani et al.

    The influence of formula modifications and additives on ammonia loses from surface-applied urea-ammonium nitrate solutions

    Fertil. Res.

    (1990)
  • R.G. Allen et al.

    Crop evapotranspiration: Guidelines For Computing Crop Water Requirements

    (1998)
  • A.S. Black et al.

    Effect of timing and simulated rainfall on ammonia volatilization from urea, applied to soil of varying moisture-content

    J. Soil Sci.

    (1987)
  • A.F. Bouwman et al.

    Estimation of global NH3 volatilization loss from synthetic fertilizers and animal manure applied to arable lands and grasslands

    Global Biogeochem. Cycles

    (2002)
  • R.J.B. Bouwmeester et al.

    Effect of environmental-factors on ammonia volatilization from a urea-fertilized soil

    Soil Sci. Soc. Am.J.

    (1985)
  • W. Burchill

    A field-based comparison of ammonia emissions from six Irish soil types following urea fertiliser application

    Irish J. Agric. Food Res.

    (2016)
  • D.W. Bussink et al.

    Ammonia volatilization from dairy farming systems in temperate areas: a review

    Nutr. Cycl. Agroecosyst.

    (1998)
  • V. Catchpoole et al.

    Transformation and recovery of urea applied to a grass pasture in south-eastern Queensland

    Aust. J. Exp. Agric.

    (1983)
  • O.T. Denmead et al.

    Ammonia flux into the atmosphere from a grazed pasture

    Science

    (1974)
  • R.J. Eckard et al.

    Gaseous nitrogen loss from temperate perennial grass and clover dairy pastures in south-eastern Australia

    Aust. J. Agric. Res.

    (2003)
  • Cited by (8)

    • A semi-empirical semi-process model of ammonia volatilization from paddy fields under different irrigation modes and urea application regimes

      2022, Agricultural Water Management
      Citation Excerpt :

      In previous studies modeling the development of NH3 volatilization, the process-based model was commonly selected (Li, 2000; Dubache et al., 2019; Li et al., 2019). This is primarily because slurry and manure were used as fertilizer sources, and the NH3 volatilization process of these fertilizers was relatively complicated (Beuning et al., 2008; Huang et al., 2012; Balasubramanian et al., 2017; Hafner et al., 2019; Smith et al., 2020). The most critical process, the mineralization and decomposition of these fertilizers which produces NH4+-N, is further complicated by the various enzymes and microorganisms playing essential roles in this process (Xu et al., 2020).

    • Ammonia volatilisation from grazed, pasture based dairy farming systems

      2021, Agricultural Systems
      Citation Excerpt :

      Therefore, the conclusion from the above information, the analysis presented in Fig. S1. in the Supplementary Material, in addition to Smith et al. (2020), that NH3 volatilisation quantities from farm systems in the present study lie within the range of plausibility. Another strength of this study was that diet N is dynamically predicted on a daily basis which in turn affects excreeta N.

    • Simulated seasonal responses of grazed dairy pastures to nitrogen fertilizer in SE Australia: N loss and recovery

      2020, Agricultural Systems
      Citation Excerpt :

      Elliott has a cooler climate compared to Mt Gambier and thus we would have expected the model to estimate lower volatilization losses at Elliott. However, research from Smith et al. (2020) concluded that DairyMod presents a readily reproducible prediction of N volatilization from urea applications on pastures in intensive livestock farming systems. Thus the higher losses of N at Elliott was more likely due to the higher N application rate at the cooler site.

    View all citing articles on Scopus
    View full text