Capercaillie

TetrasPool 1.0.2

Introduction

Assumptions

Predictions

Model

FAQ

Download

Capercaillie

Hazel grouse

EXAMPLE 1. VIABILITY OF THE ENDANGERED CAPERCAILLIE IN THE JURA MOUNTAINS (SWITZERLAND)

Deterministic model

We ran the model for 100 years projection and 1000 replicates with the data value from a field study performed in the Jura montains from 1976 to 1999. The results of the deterministic model show no strong population trend. On average, capercaillie population is self replacing and extinction risk is null. The sensitivity analysis assess the viability of populations to small changes in one life history parameters, all others parameters being held constant. This analysis is fundamental to get a sufficient understanding of the process that determine population dynamics.

Sensitity analysis

We checked the sensitivities of life-history parameter for the deterministic population model with partial derivatice from Caswell (1978).

Capercaillie population growth rate is highly sensitive to adult survival (sij=1.0), juvenile survival (sij=1.0), proportion of reproductive females (sij=0.8) ; less sensitive to sex ratio (sij=0.4) and least sensitive to fecundity (sij=0.03), dispersal distance (sij=0) and migration rate (sij=0) suggesting that the three latter parameters have a small impact on metapopulation persistence.

Stochastic model

We ran the model for 100 years projection, 1000 replicates including demographic and environmental stochasticities. Although deterministic simulation shows no extinction, stochastic model exhibits a moderate risk (10.4%) that the metapopulation of Jura capercaillie will go extinct in the next 100 years. Variations in home range size strongly increase extinction risk as K simultaneously decrease for all populations given the larger home range size. Quasi-extinction threshold represents the probability that the populations will fall below a set of threshold population sizes (Ginsburg et al. 1982) and has a major influence on extinction risk. When no migration occurs, extinction risk is doubled but higher migration rates between population halfed extinction probability. Similarly, synchroneous environmental conditions between all populations doubled extinction risk but partial environmental correlations results in similar risks.

Sensitity analysis

Sensitivities are calculated for a 10% parameter change and indicate that capercaillie populations extinction risk is highly sensitive to adult survival, juvenile survival, proportion of reproductive females, fecundity and sex ratio. A +10% change in one of these parameters highly decrease extinction risk and increase median time to extinction, while a –10% change drastically increase extinction risk and lowers extinction time.

The lack of sensitivity to dispersal distance and migration rate demonstrate that these parameters do not significantly act on the demography of capercaillie and that we do not need to estimate them more precisely in the field. Therefore, we chose to simulate several management scenarios with different values of adult survival, juvenile survival and proportion of reproductive females; parameters that are most amenable to human alteration.

We examined the relationship between extinction risk and adult survival, juvenile survival and proportion of reproductive females. Extinction risk is most sensitive to changes in adult survival. The proportion of reproductive females and the juvenile survival have a strong, similar influence on extinction risk.

Ranking threats

A prospective PVA involving sensitivities on vital rates, changes in home range size, quasi-extinction threshold, migration rate and environmental variation allows to rank the threats according to their impact on the long-term persistence of the metapopulation.

1. A large impact on extinction probability is detected with small changes in adult survival, juvenile survival, proportion of reproductive females, fecundity, home range size and extinction threshold.
2. A moderate impact on extinction risk is observed with a high level of environmental correlation
3. A low impact on extinction risk is present with various changes in migration rate, in dispersal distance and small changes in levels of environmental correlation.

Our simulation results show that capercaillie metapopulation is highly sensitive to fecundity and mortality but also to the spatial requirements of the species and to the threshold at which a population is assume to be extinct. In contrast, small changes in dispersal parameters do not appear to have major impacts on population viability. Correlation of environmental variance has an inbetween effect, high levels of correlation having a medium impact on PVA and small levels having a minor impact.

Spatial population structure

We plot the mean immigrant number vs mean emigrant number in order to identify source and sink populations. The dispersal synthesis graph presents weak migrants numbers with a maximum value of 2.2/year for emigrants and 1.05/year for immigrants. Populations 21 and 22 have the largest size and are evident source populations, they yearly provide several emigrants whose dispersal maintains sink populations. Populations 24 and 27 mainly receive immigrants and are moderate sink populations that persist only through immigrant colonization. All others populations lie near the compensation axis and provide roughly as many immigrant as emigrant.

Nonetheless, many populations export or consume less than 0.5 individuals per year and are repectively low sources or sinks, with low sinks being much more numerous than low sources. Sink populations have smaller sizes than source populations.

Management scenarios

Capercaillie population are negatively affected by habitat degradation and human disturbance, both provoke habitat loss and a decrease in K. We tested five realistic management scenarios enhancing K for different populations. Present political context and financial resources made habitat enhancement of one third of the available area a reasonable and realistic measure. A first assessement with default parameters shows a very small impact of management on extinction risk but this is due to a lambda =1 meaning that populations are stable whatever the management is done.

We test the same scenario with with a +10% change in juvenile and in adult survival. Improvement of K by increasing available capercaillie habitat has a positive effect on the viability. Viability and median time to extinction are much higher for every scenario. Surprisingly, a +30% increase in K for small populations (e.g. populations 30 to 35) has a similar positive impact on extinction probability as +30% changes for larger population (e.g. populations 1 to 21). Nevertheless, median time to extinction is shorter for changes in small populations and scenarios 1 and 5 which reflect respectively changes in all French and Swiss populations largely increase median time to extinction. In scenario 2, a positive increase in K for source populations 21 and 22 give similar extinction risk than others scenarios.

Comparison of these different scenarios suggest that the most effective silvicultural strategy would be to improve K on most populations and not to restrict to new patches creation or act only on source populations.

Send your comments

	Deterministic model We ran the model for 100 years projection and 1000 replicates with the data value from a field study performed in the Jura montains from 1976 to 1999. The results of the deterministic model show no strong population trend. On average, capercaillie population is self replacing and extinction risk is null. The sensitivity analysis assess the viability of populations to small changes in one life history parameters, all others parameters being held constant. This analysis is fundamental to get a sufficient understanding of the process that determine population dynamics. Sensitity analysis We checked the sensitivities of life-history parameter for the deterministic population model with partial derivatice from Caswell (1978). Capercaillie population growth rate is highly sensitive to adult survival (sij=1.0), juvenile survival (sij=1.0), proportion of reproductive females (sij=0.8) ; less sensitive to sex ratio (sij=0.4) and least sensitive to fecundity (sij=0.03), dispersal distance (sij=0) and migration rate (sij=0) suggesting that the three latter parameters have a small impact on metapopulation persistence.
	Stochastic model We ran the model for 100 years projection, 1000 replicates including demographic and environmental stochasticities. Although deterministic simulation shows no extinction, stochastic model exhibits a moderate risk (10.4%) that the metapopulation of Jura capercaillie will go extinct in the next 100 years. Variations in home range size strongly increase extinction risk as K simultaneously decrease for all populations given the larger home range size. Quasi-extinction threshold represents the probability that the populations will fall below a set of threshold population sizes (Ginsburg et al. 1982) and has a major influence on extinction risk. When no migration occurs, extinction risk is doubled but higher migration rates between population halfed extinction probability. Similarly, synchroneous environmental conditions between all populations doubled extinction risk but partial environmental correlations results in similar risks. Sensitity analysis Sensitivities are calculated for a 10% parameter change and indicate that capercaillie populations extinction risk is highly sensitive to adult survival, juvenile survival, proportion of reproductive females, fecundity and sex ratio. A +10% change in one of these parameters highly decrease extinction risk and increase median time to extinction, while a –10% change drastically increase extinction risk and lowers extinction time. The lack of sensitivity to dispersal distance and migration rate demonstrate that these parameters do not significantly act on the demography of capercaillie and that we do not need to estimate them more precisely in the field. Therefore, we chose to simulate several management scenarios with different values of adult survival, juvenile survival and proportion of reproductive females; parameters that are most amenable to human alteration. We examined the relationship between extinction risk and adult survival, juvenile survival and proportion of reproductive females. Extinction risk is most sensitive to changes in adult survival. The proportion of reproductive females and the juvenile survival have a strong, similar influence on extinction risk.
	Ranking threats A prospective PVA involving sensitivities on vital rates, changes in home range size, quasi-extinction threshold, migration rate and environmental variation allows to rank the threats according to their impact on the long-term persistence of the metapopulation. 1. A large impact on extinction probability is detected with small changes in adult survival, juvenile survival, proportion of reproductive females, fecundity, home range size and extinction threshold. 2. A moderate impact on extinction risk is observed with a high level of environmental correlation 3. A low impact on extinction risk is present with various changes in migration rate, in dispersal distance and small changes in levels of environmental correlation. Our simulation results show that capercaillie metapopulation is highly sensitive to fecundity and mortality but also to the spatial requirements of the species and to the threshold at which a population is assume to be extinct. In contrast, small changes in dispersal parameters do not appear to have major impacts on population viability. Correlation of environmental variance has an inbetween effect, high levels of correlation having a medium impact on PVA and small levels having a minor impact.
	Spatial population structure We plot the mean immigrant number vs mean emigrant number in order to identify source and sink populations. The dispersal synthesis graph presents weak migrants numbers with a maximum value of 2.2/year for emigrants and 1.05/year for immigrants. Populations 21 and 22 have the largest size and are evident source populations, they yearly provide several emigrants whose dispersal maintains sink populations. Populations 24 and 27 mainly receive immigrants and are moderate sink populations that persist only through immigrant colonization. All others populations lie near the compensation axis and provide roughly as many immigrant as emigrant. Nonetheless, many populations export or consume less than 0.5 individuals per year and are repectively low sources or sinks, with low sinks being much more numerous than low sources. Sink populations have smaller sizes than source populations.
	Management scenarios Capercaillie population are negatively affected by habitat degradation and human disturbance, both provoke habitat loss and a decrease in K. We tested five realistic management scenarios enhancing K for different populations. Present political context and financial resources made habitat enhancement of one third of the available area a reasonable and realistic measure. A first assessement with default parameters shows a very small impact of management on extinction risk but this is due to a lambda =1 meaning that populations are stable whatever the management is done. We test the same scenario with with a +10% change in juvenile and in adult survival. Improvement of K by increasing available capercaillie habitat has a positive effect on the viability. Viability and median time to extinction are much higher for every scenario. Surprisingly, a +30% increase in K for small populations (e.g. populations 30 to 35) has a similar positive impact on extinction probability as +30% changes for larger population (e.g. populations 1 to 21). Nevertheless, median time to extinction is shorter for changes in small populations and scenarios 1 and 5 which reflect respectively changes in all French and Swiss populations largely increase median time to extinction. In scenario 2, a positive increase in K for source populations 21 and 22 give similar extinction risk than others scenarios. Comparison of these different scenarios suggest that the most effective silvicultural strategy would be to improve K on most populations and not to restrict to new patches creation or act only on source populations.