Soil processes drive seasonal variation in retention of ¹⁵N tracers in a deciduous forest catchment

Abstract

Seasonal patterns of stream nitrate concentration have long been interpreted as demonstrating the central role of plant uptake in regulating stream nitrogen loss from forested catchments. Soil processes are rarely considered as important drivers of these patterns. We examined seasonal variation in N retention in a deciduous forest using three whole-ecosystem 15N tracer additions: in late April (post-snowmelt, pre-leaf-out), late July (mid-growing- season), and late October (end of leaf-fall). We expected that plant 15N uptake would peak in late spring and midsummer, that immobilization in surface litter and soil would peak the following autumn leaf-fall, and that leaching losses would vary inversely with 15N retention. Similar to most other 15N tracer studies, we found that litter and soils dominated ecosystem retention of added 15N. However, 15N recovery in detrital pools varied tremendously by season, with > 90% retention in spring and autumn and sharply reduced 15N retention in late summer. During spring, over half of the 15N retained in soil occurred within one day in the heavy (mineral-associated) soil fraction. During summer, a large decrease in 15N retention one week after addition coincided with increased losses of 15NO3- to soil leachate and seasonal increases in soil and stream NO3- concentrations, although leaching accounted for only a small fraction of the lost 15N (< 0.2%). Uptake of 15N into roots did not vary by season and accounted for < 4% of each tracer addition. Denitrification or other processes that lead to N gas loss may have consumed the rest. These measurements of 15N movement provide strong evidence for the dominant role of soil processes in regulating seasonal N retention and losses in this catchment and perhaps others with similar soils.

Multiyear fate of a15N tracer in a mixed deciduous forest: retention, redistribution, and differences by mycorrhizal association

“…Recovery of the added 15 N tracer at the end of Year 0 (0–10 cm soil and roots only, from Goodale et al . ()), in above‐ and belowground (to 50 cm) pools in Years +1 and +5–6, and net changes between years +1 and +5–6.…”

“…All of the added 15 N was fully recovered in the O i layer and 0–10 cm soils in the days after each tracer addition, and by the end of the year of the tracer addition (Year 0), the O i layer formed a relatedly large 15 N sink with 25% of the added 15 N (Goodale et al ., ). This 15 N then had a net turnover time of ~4.0 years, similar to the O i pool itself.…”

“…Mechanisms for 15 N movement into the deeper soil include physical leaching and sorption and biological transport by roots, hyphae, and soil fauna. The amount of 15 N recovered in soil at 10–30 cm in Year +1 was >100 times larger than the estimated flux of 15 N‐ leaching below 10 cm during the period of tracer application, which amounted to only <0.1% of the 15 N addition (Goodale et al ., ). That lysimeter‐based estimate of 15 N‐ leaching was uncertain, but was also quantified relative to a paired bromide tracer, which showed that 15 N‐NO 3 leaching was <1% that of the recovered bromide.…”

“…Shallow soil (0–10 cm) formed the largest 15 N sink at the end of Year 0, with 48% of the applied 15 N; it too showed large changes, losing nearly one‐third of this initial 15 N recovery by Year +1. The rapidity of this loss was surprising, because other studies generally show greater persistence of soil 15 N recovery over this timescale (e.g., Buchman et al ., ; Seely & Lajtha, ; Perakis & Hedin, ; Hagedorn et al ., ), and because the majority of the tracer recovered in this pool 1–30 days after the first 15 N addition occurred in association with mineral‐associated (‘heavy’) soil material (Goodale et al ., ), a soil pool typically considered relatively old and resistant to change (Sollins et al ., ; Gaudinski et al ., ).…”

“…S1) selected to avoid plot edges. Previous work (Goodale et al ., ) measured 15 N‐ leaching losses and 15 N recovery in Oi, root and soil pools sampled frequently (14 times) to only 10 cm depth during year 0. O i material was collected from the surface of 7 cm diameter soil cores for pretreatment sampling and from the interior of a 20 cm diameter steel pipe in Years +1 and +5.…”