Appendix A. Sensitivity of the results of simulations to estimates of local population size, local population dynamics, and dispersal. All figures show, similarly to Fig. 2, effect of population destruction probability and habitat fraction left on survival probability of the whole metapopulation; the vertical axis indicates the proportion of realizations (out of 100) where Succisa pratensis persisted after 100 years. After local population destruction no living individual remains at the habitat patch. If not stated otherwise, we use transition matrices and initial population sizes as stated in the paper, distance-independent dispersal is 0.0001 and distance dependent-dispersal is 0.94 m.
|
| FIG. A1. Initial population sizes do not contain seeds and seedlings, i.e., test of the sensitivity to initial habitat occupancy. |
 |
| FIG. A2. Demographic matrices were constructed after bootstrapping the initial data on transition probabilities, i.e., test of the sensitivity to estimates of local population dynamics. |
 |
| FIG. A3. Distance-independent dispersal is 0.00001, i.e., test of the sensitivity to estimates of distance-independent dispersal. |
 |
| FIG. A4. Distance-independent dispersal is 0.001, i.e., test of the sensitivity to estimates of distance-independent dispersal. |
 |
| FIG. A5. Mean distance-dependent dispersal distance is 0.46 m, i.e., test of the sensitivity to estimates of distance-dependent dispersal. |
 |
| FIG. A6. Mean distance-dependent dispersal distance is 2.35 m, i.e., test of the sensitivity to estimates of distance-dependent dispersal. |