¹⁵N study of the reactivity of atmospheric nitrogen in four mountain forest catchments (Czech Republic, central Europe)

Highlights

• Runoff generation models based on δ¹⁸O–H₂O data were used to quantify contributions of groundwater and rainfall to catchment discharge.

• Proceeding reactions were estimated from catchment hydrochemistry data plus δ¹⁵N analyses of NH₄⁺ and NO₃⁻ in precipitation, soil water, and stream discharge.

• Nitrogen release from catchments was controlled by the seasonality of soil processes: high rate during the growing season, low rate during the dormant season.

• The apparent ¹⁵N fractionation was lower than published data because of a groundwater-dilution effect in discharge and depletion of the ammonium pool.

• Calculated N loss via denitrification varied from 0.3 to 1.3 kg N ha⁻¹ y⁻¹, corresponding to 3 to 23% of total N deposition.

Abstract

Reactivity of atmospherically deposited nitrate (NO₃⁻) and ammonium (NH₄⁺) was investigated in three mountain forest catchments and one ombrotrophic peat bog located in the northern Czech Republic. The study sites are characterized by moderate to high N pollution rates that are currently decreasing. Monitoring of hydrodynamic data and catchment hydrochemistry (precipitation, soil solutions and runoff) was complemented by δ¹⁵N analyses of dissolved inorganic N (NO₃⁻, NH₄⁺). Measured δ ¹⁵N data were used to calculate the extent of biogeochemical reactions and to construct N mass balances. Two-component models of runoff generation were developed using hydrological and δ¹⁸O–H₂O data on local precipitation, runoff, and soil waters collected by lysimeters. Catchment discharge was formed by groundwater (60–80%) and storm precipitation (40–20%), with minimal time lags. Groundwater had a diluting effect on reactions proceeding in the soil zone, including nitrification and denitrification. The rates of nitrification were estimated by comparing the δ¹⁵N values of atmospheric input and soil water, while the rates of denitrification were derived from the differences in the δ¹⁵N values of soil water and stream discharge. The Ν isotope effect of mineralisation of organic N was assessed by comparing δ¹⁵N values of soil and soil extracts. Nitrogen release from the catchments was controlled by temperature-dependent seasonality of soil reactions. Atmospheric N entered the runoff directly only during a short late winter - early spring period, accounting for approximately 10% of the total N input. Important inputs of mineralised organically cycled soil N were observed at the beginning and the end of the growing season, with measurable denitrification occurring during the same time periods. The measured apparent N isotope fractionations were significantly lower than previously published values (−3 to −14‰ vs. −30‰ for nitrification, and −3 vs. −20‰ for denitrification). Lower ¹⁵N fractionations originated both from the dilution effect of groundwater on stream discharge and from the depletion of available ammonium during nitrification reactions. Denitrification proceeding during recharge of the groundwater body was estimated from the difference in the δ¹⁵N values of NO₃⁻ in precipitation and groundwater.

Introduction

Anthropogenic emissions of reactive nitrogen (N_r), dominated by ammonium (NH₄⁺) and nitrate (NO₃⁻) from fossil fuel combustion, traffic, and agriculture, result in high N_r deposition rates. Anthropogenic N inputs into forested catchments often result in elevated N export in both dissolved (NO₃-) and gaseous (N₂O, N₂) forms (Aber et al., 2003; Tang et al., 2018). Denitrification and nitrification are the main biological processes producing N₂O, one of the major greenhouse gases (Stevens et al., 1997). According to the IPCC report (2001), soils represent approximately 60% of the total natural sources of N₂O, and emissions from temperate forests account for approximately one-sixth of these emissions. The N₂O emissions are significantly higher from the deciduous soils than from the coniferous soils (Ambus et al., 2006). Losses of N due to nitrification and the microbial reduction of aqueous nitrate to N₂O or N₂ are difficult to quantify because of difficulties associated with direct flux measurements under field conditions (Groffman et al., 2006). However, N isotope methods have been successfully used to differentiate between N₂O production from nitrification and denitrification reactions (Mathieu et al., 2006). Nitrogen isotope methods can also follow the production and consumption of NH₄⁺ in soils (Braun et al., 2018). Transport and transformation of N species in groundwater in agricultural areas are another typical application of nitrogen isotope methods (Nikolenko et al., 2018). The measured fractionation factors associated with nitrification and denitrification, however, strongly depend on the reaction conditions (Mariotti et al., 1982; Hubner, 1986; Yoshida, 1988; Lewicka-Szcerbak et al., 2014; Denk et al., 2017; Nikolenko et al., 2018). Fractionating processes exhibit positive discrimination when the substrate is plentiful, but N isotope discrimination will not be observed when the reaction is substrate-limited and all of the substrate is converted to the product (Evans, 2007).

This study aims at:

• estimating the direct contribution of atmospheric N and mineralised organic N to catchment runoff;

• examining the seasonality in biogeochemical N reactions in the studied catchments.

We developed a runoff generation model for each of the studied catchments based on measured precipitation and runoff water fluxes and δ¹⁸O–H₂O values. The catchment hydrochemistry (precipitation and runoff) and hydrodynamic data were complemented by ¹⁵N analyses of the dissolved inorganic N (NH₄⁺ and NO₃⁻) in all components of the water cycle, including soil water. The δ¹⁵N data were used to evaluate N cycling and to construct N mass balances. We verified that the temporal variability in δ¹⁵N values of atmospheric N is significantly different from simultaneous δ¹⁵N values of soil water and discharge and their contribution is thus traceable. Measurements of the dual isotopic composition of NO₃⁻ (Kendall and Caldwell, 1998) can strongly improved identification of NO₃⁻ sources and their loss by denitrification. Unfortunately, such data are not available for our study sites.