Thursday, August 7, 2008

PS 60-97: Aboveground production across a soil resource gradient: Allocation shift and potential calcium limitation

Tom W. Baribault, Richard K. Kobe, and David E. Rothstein. Michigan State University

Background/Question/Methods

Aboveground net primary production (ANPP) in temperate forests could be regulated by multiple soil resources, but most studies have focused on nitrogen limitation. We evaluated a broad suite of resources in three major categories—nitrogen, cations (calcium (Ca), magnesium, potassium, acidity), and water availability—as predictors of ANPP across a natural fertility-productivity gradient in the post-glacial landscape of northwestern lower Michigan. Our goals were to: 1) evaluate the strongest predictors of productivity within and among resource categories; 2) identify functional relationships between ANPP and resources; 3) test for soil resource effects on wood versus leaf production; and 4) assess species influences on ANPP. We measured annual ANPP (wood and leaf increments) and soil resource availability at 13 sites across the fertility gradient. We fit models of four functional forms to each combination of resource and ANPPwood, ANPPleaf, and ANPPtotal. We used Akaike’s Information Criterion to assess relative empirical support of different resources and functional forms as predictors of ANPP.

Results/Conclusions

Including all sites, ANPP was related to resources as a sigmoid function. Exchangeable Ca was the resource with the greatest empirical support for predicting ANPPtotal and ANPP wood. ANPPleaf was best predicted by a Michaelis-Menten function of soil nitrate. Species richness, diversity, and dominance were not related to production. In the best-supported relationships, ANPP reached an asymptote at relatively low resource levels. An outlying subset of sites with asymptotic productivity and elevated resource availability was responsible for driving the sigmoid relationships, prompting separate analysis of the majority of sites grouped at lower resource levels. Across these least fertile 9 sites, Ca again was the best predictor of ANPPtotal and ANPP wood as a linear function. Total nitrogen mineralization was the best predictor of leaf production, also as a linear relationship. In addition, under low fertility conditions ANPPleaf constitutes the majority of ANPPtotal, but ANPPwood becomes the major component as soil fertility increases. To assess direct Ca limitation versus Ca as integrated surrogate for other resources, we calculated an approximate annual calcium budget for each site. Current exchangeable Ca could supply ~3 to 10 years of Ca demand, suggesting the potential for direct calcium limitation at the lower fertility sites. These correlative findings are yet to be confirmed with experimental nutrient additions. Counter to our expectation that nitrogen would best predict ANPP, testing a wide array of soil resources identified Ca as the best predictor of ANPPtotal and ANPPwood.