Effects of hunting regulations, population density, and weather on moose harvest in Ontario: Developing tools to guide moose management decisions

Background/Question/Methods

Moose (Alces alces) are synonymous with Ontario’s northern environment and bring ecological, cultural, and economic value to the province. The large demand for moose hunting opportunities in Ontario has necessitated strict harvest management approaches to maintain viable populations. One of the major factors that drives harvest management decisions is harvest success, the ratio of harvested moose to the number of hunters. We sought to identify the major factors influencing harvest success and develop forecasting tools to guide management decisions. We hypothesized that harvest success depends on characteristics of the environment that affect hunter efficiency and effort, the number of moose available during the hunt, and the number of hunters competing for limited resources. Using discrete-time survival models in a Bayesian hierarchical framework, we first investigated the effects of hunt characteristics (e.g. firearm type, season timing), weather conditions, landscape characteristics, the number of hunting tags issued, and moose density on per-day harvest success across provincial management units over a twenty year period (2000 – 2020). Next, we incorporated influential variables into Bayesian structural time series of harvest success and assessed model predictive ability using cross-validation on a rolling basis.

Results/Conclusions

The most influential variables across years and management units were moose density and the number of tags issued. Harvest success was positively related to moose density across management units and years, while the number of tags issued was negatively related to harvest success. These results lend support to our hypotheses and suggest that low numbers of moose may act as a limited resource to hunters even when hunting opportunities are regulated. Harvest success was also larger in management units with high road densities, suggesting that roads have a positive effect on hunter efficiency and forest accessibility. Despite the impact of moose density and number of tags across years in our hierarchical models, the best one-step ahead predictive models involved structural time series without additional covariates. Our predictive models achieved greater point-based accuracies for harvest success when compared to conventional rolling average approaches. Our forecasting tool will also allow managers to incorporate prediction uncertainty into their harvest allocation decisions, which can allow for more conservative harvests for regions with sensitive populations.