Supplement 4. Complete AICc results for three experiments at Konza Prairie LTER site (Manhattan, KS, USA).
Supplement4.csv (MD5: 9c2a02597ef066d593745194c0a92bb0)
Model selection results from analysis of the relationship between a priori 1- and 2-variable weather models and total species (A) richness and (B) Shannon diversity in each fire frequency treatment (Experiment = FRI, 1984–2011, N = 28), grazing treatment (Experiment = Graze, 1994–2011, N = 18), and season of burn treatment (Experiment = SB, 1994–2011, N = 18) at Konza Prairie LTER. R² = least square means regression coefficient, K = # of model parameters including intercept and first order autoregressive parameter, LL = log likelihood, AICc = Akaike Information Criterion corrected for small samples size, DAICC = difference in AICc value between given model and model with lowest AICc, wi = Akaike weights (measure of relative importance of a given model compared to all other models in the model set). Treatments are as follows: 1 = annually burned, 4 = burned once every 4 years, 20 = burned once every 20 years, GR = bison grazed, UG = ungrazed, FAL = fall-burned, WIN = winter-burned, SPR = spring-burned, SUM = summer burned. CV = coefficient of variation, sum = summer, spr = spring, win = winter, PPT = precipitation, TEMP = temperature, t_1 = season preceding growing season (spring and summer) when plants were sampled. Ecologically meaningful models (EMMs) were those with DAICC < 2 and R² > 0.30. Model # corresponds to a priori models as described in Appendix B.