Slurry field management and ammonia emissions under Mediterranean conditions

In Spain, farmers are interested in applying pig (Sus scrofa domesticus) slurry (PS) to their fields throughout the year. During the spring and summer months, ammonia (NH3) volatilization may be high. We studied the potential range of NH3 losses under a warm and a hot period of the year, using available field practices, and two strategies: PS directly incorporated into the soil, in spring (I‐spring); and PS applied by splash‐plate, in summer time (SP‐summer), both to bare soil. Measurements were conducted, after PS application, using the micrometeorological mass balance integrated horizontal flux method. The cumulative NH3‐N volatilization was 35% (I‐spring) and 60% (SP‐summer) of total ammonium nitrogen applied, and half of the total NH3‐N losses happened by 17 h and 8 h, respectively, after application. Incorporation strategy was less effective in avoiding NH3 losses than is described in the literature. This fact has important consequences for the implementation of NH3 mitigation measures in Mediterranean agricultural systems.


Introduction
Spain is the second largest pig producer in Europe, and nearly 50% of Spanish pig production is concentrated in the Ebro river valley (MARM, 2013).
Pig slurry (PS) is mainly applied to maize (long cycle) at sowing (April-May) and to winter cereals at sowing (October-November) or as a side-dressing (February-March), leaving a gap of 5-6 months when applications are normally avoided. In irrigated areas, a second crop after the cereal harvest or the extension of the maize (short cycle) sowing period can cover this gap.
Despite the importance of quantifying European NH 3 emissions, little work done under Mediterranean conditions has been published on the topic (G enermont & Cellier, 1997;Sanz et al., 2010), and existing articles do not cover the whole annual period, nor the range of typical conditions.
Our objective was to quantify NH 3 volatilization losses from PS applied according to two different strategies: (i) direct incorporation in spring (I-spring), assumed to approximate to minimum likely losses, and (ii) splash-plate application in summer time (SP-summer), taken to approximate to maximum likely losses.

Materials and methods
Two experiments were conducted in a representative area of the Ebro valley (41°44′N, 0°49′W, altitude 225 m) on bare ploughed soil (Table 1) before cereal establishment. The first was established on 16-17 May 2007. An incorporation machine was employed. Pig slurry incorporation was by a tube divided into three hoses (12 outlets), with a total application width of 4.80 m. Each outlet was located between two shares, the first one opened a slot in the soil and the one located at the back buried the applied PS at a depth of about 0.15-0.20 m. The second experiment was established on 2-3 August 2007 where surface PS spreading was by a tank fitted with a splash-plate; it was spread over the soil without incorporation. For each strategy (Table 2), three replicates were set up plus a control.
Weather in the experimental period has two limitations, in May with minimum average temperature (T) and maximum Soil Use and Management, September 2013, 29, 397-400 doi: 10.1111/sum.12061

SoilUse and Management
average relative humidity (RH), and in August with maximum T and minimum RH ( Figure 1).
The micrometeorological mass balance integrated horizontal flux method was used, following the description and procedure given by Wood et al. (2000). Each rotating mast supported three passive NH 3 flux samplers mounted at three heights (0.375, 0.75 and 1.50 m), with the greatest height being 10% of fetch length, in agreement with estimates by Itier & Perrier (1976), which were confirmed by a previous field test. The ammonium solution obtained in passive samples was analysed with a continuous flow analyser (AA3-Bran + Luebbe). Sampling started immediately after application, and periodicity for the first day was from 1 to 2 h, at 3 h, from 5 to 8 h, at 12 h and before 24 h after application. Later on, the intensity of sampling decreased with time according to the declining of the intensity of NH 3 flux losses ( Figure 2).
Differences between strategies in NH 3 -N cumulative emissions and also as a percentage of applied NH 4 + -N were analysed using analysis of variance and LSMEANS test (P = 0.05). The statistical package SAS v8.2 was used for all statistical analysis.

Results and discussion
The average for cumulative NH 3 emissions, measured during the experiments, was 35% of total ammonium nitrogen (TAN) applied (99 kg NH 3 -N/ha, 25% of total N) in the I-spring strategy and 60% of applied TAN (122 kg NH 3 -N/ ha, 42% of total N) in the SP-summer strategy ( Table 2). Half of the maximum NH 3 -N loss was estimated to occur by 17 h and 8 h after application in I-spring and SP-summer strategies, respectively (Figure 2). For the SP-summer strategy, the emission was in the upper ranges found in other studies (e.g. Sommer et al., 2003;Rochette et al., 2009), but according to Misselbrook et al. (2005), the slurry dry matter content in the SP-summer strategy can explain losses equivalent to 60% of applied TAN. Incorporation did not reduce volatilization as much as it was expected to, according to the results from other experiments (Huijsmans et al., 2003), probably because dry soil conditions (Table 1) made it difficult to fully bury the PS, favouring NH 3 gas diffusion through it.
In the context of Mediterranean agricultural systems, the official advice to bury PS spread over land not later than 24 h after application would not be fully effective in reducing NH 3 emissions, as more than 50% of applied TAN could already be lost during the first 24 h after application. In addition, the low moisture content in the soil, in the hottest months of the year, limits the effectiveness of PS incorporation. New methods must be investigated, and they should be orientated to the reduction in slurry-air contact (e.g. trail hoses) or/and to slurry infiltration enhancement (e.g. light irrigation). Nevertheless, their effectiveness will be influenced by soil carbonate content, and this aspect needs further practical evaluation. This research is also necessary in the framework of models such as ALFAM (Søgaard et al., 2002) that do not fully cover the special aspects that need to be considered in Mediterranean environments.