Analysis of Decadal Time Series in Wet N Concentrations at Five Rural Sites in NE Spain

Abstract

Nitrogen emissions have grown in Spain during the last 15 years. As precipitation scavenges gases and aerosols from the atmosphere, an effect on rainwater concentrations can be expected. However, time-series studies on wet N concentrations in the Iberian Peninsula are very scarce. This paper aims to fill this gap by analysing weekly rainfall N concentrations at a set of rural sites in Catalonia (NE Spain) from 1995/1996 to 2007 and a forest site monitored from 1983 to 2007. The sites encompass a range of rural environments and climate conditions, from the inland pre-Pyrenees (Sort) to the Mediterranean coast (Begur) and from north (Sort and Begur) to central (Palautordera and La Castanya) and south Catalonia (La Senia). We found a 1-year cycle for concentrations of NH ⁺₄ and NO ⁻₃ whereby higher values were reached at the end of spring–early summer, except at the easternmost coastal site of Begur. Weekly NH ⁺₄ concentrations decreased with time at all sites (except at La Senia) whilst NO ⁻₃ concentrations increased at all sites during the same period. Rainfall SO ²⁻₄ concentrations decreased with time at all sites. The opposite trends in NO ⁻₃ and SO ²⁻₄ concentrations determined a shift in the relative acid contribution of those anions during the 12–13-year period. To interpret the increasing trend, mean annual NO ⁻₃ concentrations were regressed against NO₂ Spanish emissions and to some indicators of local anthropogenic activity. The increase at Sort and Palautordera showed good correlation with local anthropogenic indicators. Wet inorganic N deposition ranged between 4.2 and 6.7 kg ha⁻¹ year⁻¹. When including estimates of dry deposition, total annual deposition rose up to 10–20 kg ha⁻¹ year⁻¹, values that have been found to initiate adverse effects on Mediterranean-type forest ecosystems.

Appendix

The use of logarithm of concentrations instead of concentrations in the regression approach above precludes the direct identification of coefficient b as a trend term of the untransformed concentration time series. However, it is easy to show that the so-calculated coefficient b is, in fact, a good approximation to the annual change rate of the concentrations. Let y the natural logarithm of the concentration of a compound z such that y = Ln (z). Then, taking differences:

$$ d{\text{Ln}}(z) = \frac{{d{\text{z}}}}{z} = a\; \times \;d\left( {\cos \left( {\frac{{2\pi }}{365}t - \phi } \right)} \right) + b\; \times \;dt + c\; \times \;dP $$

where each parameter is defined as in Eq. 4 above. The term dz/z is the instantaneous change rate in the concentration of compound z and is a function of the summation of three terms: a trigonometric function which cancels if integrated in a 1-year interval, a constant term b and a term that depends on the precipitation. We can now readily identify:

$$ b\; \times \;dt = \left( {\frac{dz}{z}} \right)_{\text{trend}} $$

as the instantaneous change rate in the concentration of compound z due to the presence of a trend in the time series. Given that, in general, z » dz and that changes in concentration due to the trend are very small during 1 year, we can conclude that:

$$ b \cdot \int\limits_{\text{1 year}} {dt} = b = \int {\left( {\frac{dz}{z}} \right)_{\text{trend}} } \approx \frac{1}{{\overline{z} }}\; \times \;\Delta z $$

where ∆z is the total change of the concentration in a year and $\overline{z}$ is the yearly average of the concentration. Thus, coefficient b clearly represents the approximate annual change rate in concentration.