Abstract
Large variations in vital rates are found among individuals in populations of Ascophyllum nodosum along the Swedish west coast due to temporal variation in the environment (years with or without ice). The population growth in these stochastic environments was analyzed for two populations using model simulations. A matrix model with a set of three population projection matrices for each population described the demography during three different types of years (environments), i.e., years with (1) no ice, (2) moderate damage due to ice, (3) large damage due to ice. The frequency of ice years is known for the area studied and was higher at one site (Goteborg) compared to the other (Tjarno). These frequencies were used as the stochastic process generating the sequence of matrices. The mean mortality rates decreased with increasing size in both populations. Simulating the growth for survival matrices (no reproduction included) resulted in an average population growth rate @ls of 0.84 at Tjarno and 0.76 at Goteborg. The average growth rate @ls was set to 1 by scaling the set of matrices for each population with a fertility function. To achieve @ls = 1, °3.2 times more fertility in each size class was required at Goteborg. Because of the large environmental variability, @ls varied from 0.3 to 1.3 during a long stochastic sample path. The maximum lifetime of A. nodosum individuals was estimated to be °50—60 yr, which probably is shorter than for areas with no ice. The population growth with the sets of scaled matrices (reproduction included) was simulated by following a large number of subpopulations for 300 yr. With extinction defined as the first passage time through the point N = 1, °25% of the subpopulations "died" during the simulation, and the mean extinction time for those who died was 163 yr. An elasticity analysis of @ls to changes in the matrix elements showed that at each time step is was most important for A. nodosum individuals to survive and grow to the next larger size class or to survive and remain in the same size class.

This publication has 0 references indexed in Scilit: