Nitrate leaching from a subalpine forest ecosystem subjected to experimentally increased N deposition: patterns and mechanisms as functions of the considered space and time scales

Schleppi Patrick 1

1 Swiss Federal Institute for Forest, Snow and Landscape Research (WSL), CH-8903 Birmensdorf, Switzerland

4th International Symposium on Ecosystem Behaviour BIOGEOMON, Reading, UK, 17-21/08/2002


Nitrate leaching is usually regarded as the main symptom of nitrogen saturation in forests. At the subalpine research site Alptal,central Switzerland,we observed different mechanisms of NO3- leaching during 6 years of a nitrogen addition experiment in a paired-catchment design.
The site lies 1200m above see level and has a cool,wet climate (6°C; 2300 mm/year). The bulk deposition of NO3- + NH4+ is 12 kg N/ha/year. On a Flysch substratum,the soils are umbric Gleysols,with mor (raw humus) or anmoor (muck humus) depending on the micro-topography. The trees are predominantly Norway spruce (Picea abies),with 15% silver fir (Abies alba). The tree canopy is not very dense (LAI = 4.1) and the ground vegetation is well developped.
Two forested catchments (approx. 1500 m2 each) have been delimited by trenches. NH4NO3 was added to rainwater during precipitation events and applied by sprinklers to one of the catchments. The effects of the treatment (26 kg N/ha/year) were assessed after one year of pre-treatment measurements. During the first year of N addition,both NH4++ and NO3- were labelled with 15N. Precipitation and discharge-proportional runoff samples were analysed weekly for 6 years. Additionally,a replicated plot design was established to study N transformations in the soil.
In spite of needle analyses indicating a slight N-deficiency of the trees,NO3- was leaching from both catchments at 4 kg N/ha/year already during the pre-treatment year. In the following years,NO3- leaching varied between 2.5 and 5 kg N/ha in the control,half of which ocuring from March to May,i.e during the snowmelt period. The detailed analysis of rain or snowmelt events showed a general dilution of NO3- in runoff water during the events. Narrow concentration peaks occured at the maximum of broader discharge peaks and were indicative of NO3- flushing from the soil. Very short travel time of tracers spread on or injected into the soil showed the importance of preferential flow through soil macropores. This was confirmed by micro suction cups introduced between 1 and 5 cm depth into the soil: those collecting a dye tracer spread on the ground also captured more nitrate. At a larger scale,the micro-topography is another factor affecting the water fluxes,introducing lateral nutrient migrations from the mounds to the depressions. In the depressions,the water table rises sharply at the onset of rainfalls,forcing near-surface runoff. At even larger scales,drainage trenches and patterns of forest or grassland clearly affect the runoff patterns and nitrate concentrations.
Within weeks,the experimental N addition more than doubled the NO3- leaching compared to the control catchment. It also induced higher NO3- concentrations in the topsoil solution (with more denitrification after N-enriched rain events),but not in the soil solution collected below 10 cm. A close linkage between soil water from the topsoil and the catchment runoff indicated the dominance of near-surface and/or preferential flow. An end-member mixing analysis based on Ca2+,Cl-,SO42- and electrical conductivity showed that,during a rain event,a large proportion of the runoff water originates from the topsoil or directly from precipitation. The obtained mixing model gave a good prediction (r2 = 0.83) of NO3- concentrations. As a consequence,NO3- leaching from the Alptal Gleysols appeared to be mainly hydrologically driven and to originate directly from precipitation water rather than from production in the soil (by nitrification). In this case,NO3- leaching is therefore not a symptom of N saturation. This was corroborated by the 15N found in the supplementary NO3- leaching: it was close to that in the manipulated inputs,indicating few mixing with unlabelled soil N pools. While insufficient to retain all NO3-,the contact of precipitation water with the soil was enough to completely immobilise ammonium.
During the first year of treatment,the ecosystem retained approximately 90% of the added N,with 2/3 of the 15N remaining in the soil. Then,however,a larger proportion was leached as NO3-,increasing from 10% up to 30% of the added N in the last 2 years. In contrast to the short-time,hydrological reaction,this slowly evolving effect indicates that the manipulated catchment now moves towards true N saturation. The 15N signal in runoff nitrate,however,practically disappeared within 3 months from the end of the labelling. Leaching rates increasing from year to year can therefore be ascribed to a decreased retention of recent (unlabelled) N addition rather than to a progressive release of older (labelled) N first retained in the soil.
Our N addition experiment shows that different patterns arise from different mechanisms depending on which scales are condisered,both in space and time. Long-term ecosystem-scale experiments are necessary to cover the larger and longer scales considered here.