|Cross-site comparison of herbivore impact on nitrogen availability in grasslands: the role of plant nitrogen concentration
|Year of Publication
|Bakker, ES, Knops, JMH, Milchunas, DG, Ritchie, ME, Olff, H
Herbivores may influence the nitrogen (N) recycling rates and consequently increase or decrease the productivity of grasslands. Plant N concentration emerged as a critical parameter to explain herbivore effects from several conceptual models, which predict that herbivores decrease soil N availability when plant N concentration is low whereas they increase it when plant N concentration is high (Hobbs 1996, Ritchie et al. 1998, Pastor et al. 2006). However, a broader cross-site comparison among published studies to test these predictions is hampered by the different methodologies used to measure soil N availability or a proxy thereof, and a lack of measurements of plant N concentration. Therefore it remains unclear whether these model predictions are generally valid across a range of grasslands. We tested whether there is a relationship between plant N concentration and herbivore impact on soil N availability (measured with resin bags) with a study of replicate 6–8 year old exclosures (with an unfenced control) of vertebrate herbivores (>1 kg) established at each of seven grassland sites in North America and Europe. Contrary to model predictions, we found a negative relationship between the effect of herbivores on resin bag soil N availability and plant N concentration. Our study confirms the importance of plant N concentration as a predictor of herbivore effect on soil N availability across grasslands, but contradicts the models. A possible explanation may be that the results represent a transient situation as the exclosures were relatively young whereas the models may refer to an equilibrium state. Simultaneous measurements of both plant N concentration and herbivore effect on soil N availability from more grassland sites, preferably with contrasting plant N concentrations and including exclosures of different ages, should resolve the contrast between model predictions and our field measurements.