Premium
When should plant population models include age structure?
Author(s) -
Chu Chengjin,
Adler Peter B.
Publication year - 2014
Publication title -
journal of ecology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.452
H-Index - 181
eISSN - 1365-2745
pISSN - 0022-0477
DOI - 10.1111/1365-2745.12212
Subject(s) - survivorship curve , akaike information criterion , vital rates , deviance information criterion , population , statistics , age structure , mathematics , demography , population projection , biology , population size , regression , population growth , ecology , bayesian probability , sociology , bayesian inference
Summary Most population models for plants are structured exclusively by size or stage. However, the assumption that age can be safely ignored has not been rigorously tested. We used an unprecedented data set from western North America on age‐specific demography of 19 populations of perennial grasses to determine when age structure can alter model behaviour. Based on the Akaike Information Criterion, the best statistical models included both size and age for 14 and 18 of the 19 populations for the growth and survival function, respectively. The inclusion of age reduced the sums of squared residuals by 0.1–5.6% for growth rates and reduced the residual deviance by 0.06–8.2% for survival rates. When we compared observed per capita growth rates with predictions from individual‐based models, we found that in eleven populations, models with age explained 1.3–45.8% (with a mean of 12.8%) more variance than size‐only models. Density‐dependent integral projection models showed that the inclusion of age had strong effects on equilibrium cover for some species but not others. When equilibrium cover differed, it was almost always higher in models with age. The population consequences of incorporating age were best predicted by the shape of a species' survival curve: equilibrium cover was most sensitive to inclusion of age structure for species with very hollow (Type III ) survivorship curves. We speculate that age serves as a proxy for high individual‐level heterogeneity, which is known to generate Type III survivorship curves and can increase equilibrium abundance. Synthesis . Our results indicate that including age structure in models is most important for populations with strong Type III survivorship curves. For such populations, which are likely to contain high individual heterogeneity, models that ignore age could underestimate population growth rates and equilibrium cover. More research is needed to test whether our results, based on data from perennial grasses, also apply to other plant growth forms.