Atmospheric ammonia at a moorland site. II: Long-term surface-atmosphere micrometeorological flux measurements

Long term micrometeorological measurements of the net vertical exchange of atmospheric ammonia (NH3) over moorland vegetation at a site in southern Scotland from February 1995 through to February 1996 are reported. The measurements, made using a combination of eddy covariance and aerodynamic flux-gradient methods, provided satisfactory 30-minute average fluxes approximately 50% of the time using continuous annular wet denuders for NH3 sampling at three heights between 0.4 and 3.4 m above the ground. The moorland vegetation was a net sink for atmospheric NH3 as 93.4% of all half-hourly measured fluxes indicated dry deposition to the surface with a mean flux of -5.2 ng m(-2) s(-1) and a mean deposition velocity V-d of 11.7 mm s(-1) at (z - d) = 1 m. Despite the dominance of deposition to the moorland, emission fluxes were observed approximately 6.6% of the time with a median of +1.1 ng m(-2) s(-1), and most commonly resulting from drying water films on foliar surfaces. The mean aerodynamic resistance at a reference height of (z - d) = 1 m and the mean viscous sub-layer resistance were 36.6 s m(-1) and 11.6 s m(-1) respectively, and of a similar magnitude to the mean canopy or surface resistance R-c of 37 s m(-1). When the moorland surfaces were wetted by rain or dew, canopy resistances to NH3 deposition were generally smaller (averaging 23 s m(-1)) than in dry conditions (when R-c was 61 s m(-1)) or when surfaces were covered in snow (R-c = 56 s m(-1)). Therefore, moorland surfaces cannot be treated as perfect sinks for NH3 deposition. Diurnal and seasonal patterns in rates of deposition and net fluxes to the moorland surface area were determined mainly by a combination of wind speed, surface wetness and ambient NH3 concentration. Thus, during winter, deposition rates expressed as a flux or deposition velocity were generally much larger than in summer by factors of 2 or 3 respectively, as a result of stronger winds and the presence of surface water for extensive periods which kept canopy resistances small except in frozen conditions. At temperatures below 0 degrees C, the surface became progressively dry and canopy resistances were large (50-100 s m(-1)). The measurements provided direct estimates of the net annual gaseous NH3 input to the moorland, which was 2.5 kg NH3-N ha(-1) a(-1) and was almost identical to the measured wet deposition input at the site of 2.4 kg NH4+-N ha(-1) a(-1).