You are here:
THE INFLUENCE OF MODEL TIME STEP ON THE RELATIVE SENSITIVIY OF POPULATION GROWTH RATE TO REPRODUCTION
Citation:
Gleason, T R., W R. Munns Jr., AND D E. Nacci. THE INFLUENCE OF MODEL TIME STEP ON THE RELATIVE SENSITIVIY OF POPULATION GROWTH RATE TO REPRODUCTION. Presented at Association Of Applied Biologists, Conference On Challenges Of Applied Population Biology, London England, July 8-9, 1999.
Description:
In recent years there has been an increasing interest in using population models in environmental assessments. Matrix population models represent a valuable tool for extrapolating from life stage-specific stressor effects on survival and reproduction to effects on finite population growth rate ( ). Elasticity analysis of a matrix model allows an evaluation of the relative sensitivity of population growth rate to each element of the matrix. Since demographic elasticities sum to one, elasticity analysis can determine the relative importance of each life history stage a priori. One of the attractive features of stage-classified matrix models is that the projection interval or model time step can be chosen in a manner that allows for explicit incorporation of the demographic rates measured in laboratory bioassays. The tacit assumption has been that model time step does not influence model results, however, to date there has not been a critical analysis of this assumption. We developed three models with the same one year stage durations and altered only the projection intervals: one age-structured model with a projection interval of one year, and two stage-structured models with projection intervals of six months and three months, respectively. Daily survival rates were used to adjust survival to each projection interval. Stage-specific reproductive output was held constant across the three models. Annualized was the same for each of the three models. Total reproductive elasticity, summed across all stages, declined markedly with the reductions in projection intervals (12 mos = 0.4,6 mos = 0.31, 3 mos = 0.21). The elasticity analyses indicated that three presumably reasonable and equivalent model representations produced very different values for the importance of various life history stages, particularly a decline in the importance of reproductive output. In this example, an anthropogenic stress that impaired reproduction would be projected to have very different population-level effects depending on the selected model construct. This analysis demonstrates that the choices made during model construction, particularly the time step or projection interval, can have serious and largely unrecognized implications for model results.