Soil Nitrogen Dynamics
231
of nitrate in percolating water at Prades. In zero-tension lysimeters, the nitrate flux below the forest floor was 22 kg NOrN ha- ' year-I, a figure of the
same order of magnitude as the release of N from the forest floor estimated
by the Jenny decomposition model (35 kg N ha- ' year-I).
Since the nitrate flux in throughfall was minor (Table 16.1), the above result means that more nitrate was leached from the forest floor than was produced in the H horizon through nitrification. Some of this extra nitrate may
be produced by nitrification in the F horizon, since, although freshly fallen
litter may act as a net sink for N (Berg and Ekbohm 1983), the forest floor
layers Land F may also act as a net source of N (Qualls et al. 1991; Casals
et al. 1995). However, comparisons between nitrification and nitrate leaching
fluxes must bear in mind that our nitrification estimate at Prades is based on
only four monthly incubations, and that small zero-tension lysimeters (and
throughflow collectors) may overestimate leaching fluxes through drawdown
oflateral water. Casals et al. (1995) and Cortina et al. (1995) measured similar
solution fluxes of N (14-20 kg N ha- ' year-I) under the forest floor in pine
forests near Montseny.
In the mineral soil, nitrate concentrations in percolating water decreased
at both Prades and Montseny (Fig. 16.4), probably due to plant uptake, microbial immobilization, and denitrification. Nitrate concentrations in
streamwater were very low at both sites, and were lower than in the water
percolating through the deep soil (Fig. 16.4). This was not due to nitrate retention in the streams (through uptake by the aquatic biota or other inchannel processes) since a seep under the permanent plot at Montseny already showed very low nitrate concentrations (Avila 1988). Therefore, the reduction in nitrate concentrations between the deep percolating water and the
streamwater arose either from spatial heterogeneity of soils within the
catchments, from streamwaters being a mixture of waters with a mean residence time longer than that of gravitational soil water (Neal et al. 1992), or
from terrestrial processes acting at depth such as uptake by holm oak roots
growing in rock fissures or deep denitrification.
The Prades lysimeters, placed on an upper slope site, measured a significant water flow at a soil depth of 30 cm. Thus, even though nitrate concentration in percolating water decreased with soil depth, the amount of nitrate
leached through the mineral soil at the bedrock contact was still quite high
(16.5 ± 3.0 kg N ha- ' year-I) compared with the output in streamwater (Table
16.1). In a holm oak forest on calcareous soil, Rapp (1990) and Lossaint and
Rapp (1978) found nitrate leaching losses of 2.6 and 10 kg N ha- ' year- I at 15and 30-cm depths, respectively. The fact that at Prades more nitrate was
leached from the soil than exported from the catchment, if not totally accounted for by the soil heterogeneity and the possible overestimation mentioned above, raises the interesting possibility that nitrate is lost from draining water as it moves downslope. This would translate into a fertility flux
from the upper to the lower parts of the catchment.
231
of nitrate in percolating water at Prades. In zero-tension lysimeters, the nitrate flux below the forest floor was 22 kg NOrN ha- ' year-I, a figure of the
same order of magnitude as the release of N from the forest floor estimated
by the Jenny decomposition model (35 kg N ha- ' year-I).
Since the nitrate flux in throughfall was minor (Table 16.1), the above result means that more nitrate was leached from the forest floor than was produced in the H horizon through nitrification. Some of this extra nitrate may
be produced by nitrification in the F horizon, since, although freshly fallen
litter may act as a net sink for N (Berg and Ekbohm 1983), the forest floor
layers Land F may also act as a net source of N (Qualls et al. 1991; Casals
et al. 1995). However, comparisons between nitrification and nitrate leaching
fluxes must bear in mind that our nitrification estimate at Prades is based on
only four monthly incubations, and that small zero-tension lysimeters (and
throughflow collectors) may overestimate leaching fluxes through drawdown
oflateral water. Casals et al. (1995) and Cortina et al. (1995) measured similar
solution fluxes of N (14-20 kg N ha- ' year-I) under the forest floor in pine
forests near Montseny.
In the mineral soil, nitrate concentrations in percolating water decreased
at both Prades and Montseny (Fig. 16.4), probably due to plant uptake, microbial immobilization, and denitrification. Nitrate concentrations in
streamwater were very low at both sites, and were lower than in the water
percolating through the deep soil (Fig. 16.4). This was not due to nitrate retention in the streams (through uptake by the aquatic biota or other inchannel processes) since a seep under the permanent plot at Montseny already showed very low nitrate concentrations (Avila 1988). Therefore, the reduction in nitrate concentrations between the deep percolating water and the
streamwater arose either from spatial heterogeneity of soils within the
catchments, from streamwaters being a mixture of waters with a mean residence time longer than that of gravitational soil water (Neal et al. 1992), or
from terrestrial processes acting at depth such as uptake by holm oak roots
growing in rock fissures or deep denitrification.
The Prades lysimeters, placed on an upper slope site, measured a significant water flow at a soil depth of 30 cm. Thus, even though nitrate concentration in percolating water decreased with soil depth, the amount of nitrate
leached through the mineral soil at the bedrock contact was still quite high
(16.5 ± 3.0 kg N ha- ' year-I) compared with the output in streamwater (Table
16.1). In a holm oak forest on calcareous soil, Rapp (1990) and Lossaint and
Rapp (1978) found nitrate leaching losses of 2.6 and 10 kg N ha- ' year- I at 15and 30-cm depths, respectively. The fact that at Prades more nitrate was
leached from the soil than exported from the catchment, if not totally accounted for by the soil heterogeneity and the possible overestimation mentioned above, raises the interesting possibility that nitrate is lost from draining water as it moves downslope. This would translate into a fertility flux
from the upper to the lower parts of the catchment.