Scientific Assessment to Inform the Identification of Critical Habitat for Woodland Caribou (Rangifer tarandus caribou), Boreal Population, in Canada - 2011 Update
Previous Page | Table of Contents | Next Page
4. Discussion
The purpose of this assessment was to provide a scientific description of critical habitat for boreal caribou to inform critical habitat identification within the recovery strategy for this species at risk. A central premise of this assessment was that local populations and associated ranges are the appropriate biological and geographic units for critical habitat identification. Also important to framing this exercise is the goal of achieving self-sustaining populations, which has both a short-term (≤20 years) population trend component and a longer term (≥50 years) persistence component.
In the 2008 Scientific Review, the approach to critical habitat identification was focused on a probabilistic assessment of the adequacy of the current range conditions to support a self-sustaining population. A range assessment based on three lines of evidence (% total disturbance, population growth and population size) resulted in classification of critical habitat for each local population into one of three states: maintain current conditions, improve current conditions, or assess resilience to further disturbance. The present assessment extends the 2008 approach and addresses several key areas of uncertainty in the earlier assessment. However, it does not represent a fundamental shift from the premise that range is the appropriate geographic delineation. Further, the amount of total disturbance within a range remains the primary criteria for identifying critical habitat to meet a goal of self-sustaining local populations of caribou.
Significant advances were made in the conceptual and methodological design that underpins this 2011 Scientific Assessment. The first advance consists of reconsideration of the ecological representation of the objective of self-sustaining populations. This resulted in explicit recognition of two components encompassed by this population objective statement - stable and increasing population growth and long term persistence. The 2008 Scientific Review approach applied an integrated methodology that a priori combined these two components into one measure labeled as persistence, represented using three, equally-weighted lines of evidence. In the present assessment, indicators were identified that clearly distinguished these components of self-sustainability. Where discrepancies among indicators arose, decision rules were applied to help understand causes and resolve differences within the integrated assessment. This approach yielded better information to guide recommendations than simple averaging of contributing values.
The second major advance was the development of a conceptual framework and methodology to identify range-specific, disturbance-based management thresholds and support their interpretations. In the 2008 Scientific Review, the final range assessment led to designations of range self-sustaining (SS), not self-sustaining (NSS), or either (NSS/SS), based on available information. The critical habitat identification derived from these designations was limited to general statements about the condition of a given range relative to where it should be to support a self-sustaining population. In this 2011 Science Assessment, guidance is provided for use of a disturbance-based population growth function in conjunction with range specific information to extend the critical habitat description for the consideration of disturbance-based management thresholds. The present assessment articulates the clear dependence of the identification of disturbance-based management thresholds on specification of acceptable risk by decision makers. Once defined, these analyses could support action planning by informing decisions around several key questions, such as; “is habitat recovery needed?”; “how much habitat recovery is needed?”; “is there potential for further development in a range?”; and “how much more development could be tolerated?” at the specified level of risk.
As part of the 2011 Scientific Assessment, a significant update to the meta-analysis of recruitment in relation to disturbance was also undertaken, to assess the relative impacts of different disturbance types, and the effects of measures of habitat quality and configuration, on the underlying relationship. Substantial improvements in disturbance mapping were undertaken to support these analyses. Analyses examining the effects of different assumptions about the functional zone of influence of disturbances were used to clarify relationships between caribou recruitment and habitat condition. The updated, combined disturbance model explained nearly 70% of the variation in estimated calf recruitment across the study areas included in the analyses, most of which could be attributed to the negative effects of anthropogenic disturbance. Little statistical support was found for decomposing anthropogenic disturbances into finer classes to improve the predictive power of the recruitment model. However, the negative effect of linear disturbances on caribou demography was greater than the negative effect of polygonal disturbances, consistent with the results of the resource selection analyses examining caribou use. Unfortunately, the assessment of recruitment response to the hypothesized interaction between the amount of high quality habitat remaining in a range and total disturbance was limited by the very coarse habitat quality information available for the national resource selection function model. Several of the recruitment models evaluated suggest that better understanding of potential regional variation in response, including refined habitat selection models, could enhance range-specific applications. As well, further investigation of the ways that spatial configuration of different disturbance types could influence caribou demography is recommended. Regardless, the overall national relationship between total disturbance and calf recruitment was robust.
The methods for assessing self-sustainability were extended through the use of enhanced population modelling which enabled estimation of continuous probabilities, rather than discrete categories, and more explicit inclusion of uncertainty in both parameter estimation and outcomes. Probabilities describe the expected state of a criteria based upon the evidence gathered (e.g., statistical evidence from similar cases, modelling evidence, or expert opinion), but should not be interpreted as a prediction. A given probability that a local population is stable or increasing represents a relative likelihood in realizing a desired outcome that decision-makers can use to inform assessment of risk and determine management actions. Accounting for uncertainty offers the potential for more proactive, precautionary and innovative decision responses to assessing recovery than might otherwise be possible. Assessing uncertainty is also consistent with the application of an adaptive approach to caribou recovery.
A sensitivity analysis conducted as part of the population modelling demonstrated that population trend predictions were strongly influenced by adult mortality rates. There is little available information on adult mortality rates across boreal caribou ranges in Canada, and what is available is biased with respect to disturbance (i.e., there are generally more data available for highly disturbed ranges). In this assessment, a national average for female adult survival, derived from available data, was used. While expert opinion indicated that this estimate was reasonable, derivation of a disturbance-dependent female adult survival rate relationship would increase the certainty associated with population trend predictions. This highlights the need for more extensive monitoring programs that include adult female mortality assessments.
In contrast to the 2008 Scientific Review, the use of a set of hierarchical decision rules to combine different lines of evidence about the ability of ranges to maintain self-sustaining populations accounted for the types and quality of data available for each range, and dictated the relative contribution of each factor (including time scale) to the integrated risk assessment. The relative certainty associated with the integrated assessment can be used to inform monitoring needs, as well as the types of recovery actions that might be appropriate.
Similar to the 2008 Scientific Review, this assessment of the effects of range condition on self-sustainability was based on composite measures of disturbance and indicators from demographic modelling. Integration of new information or different sources of information is possible within the same framework, with the decision rules expanded to explicitly weight these additional sources of information.
Simple models of the key boreal ecosystem dynamics (regeneration or recovery of disturbed areas and new disturbance by fire) were developed and provided as information to support consideration of range-specific disturbance thresholds. The use of this approach is consistent with the need to consider species recovery in the context of the extent and rate at which critical habitat may change in response to environmental factors. This represents a key component of a precautionary approach to the goal of self-sustainability.
Critical Habitat Description
The final critical habitat descriptions derived by application of the critical habitat framework, and informed by the components described above, are provided in fact sheets for each of the 57 ranges assessed. The elements of the critical habitat description for each range are:
- range boundary and location;
- integrated risk assessment;
- information to support identification and interpretation of range-specific disturbance thresholds; and
- bio-physical attributes of habitat within a range.
Range boundary and location
Ranges used in this assessment were delineated by local jurisdictions based on a variety of methods and types of data. Some range delineations are more robust than others based on the type and quantity of data available. Most ranges in Canada have not been fully described owing to a lack of standardized animal location data and poor understanding of movement between adjacent or nearby ranges. Several ranges were delineated as conservation units, because animal location data were insufficient to support delineation of a local population. In these cases, the assessment represents an assessment of the condition of the conservation unit to support a self-sustaining local population. In a number of cases, the current range boundaries also do not consider trans-boundary movement of caribou between jurisdictions. In the case of Quebec and Labrador, updated range delineation and demographic data were not provided, therefore the 2011 assessment was completed using the 2008 information as the best available. As new information is obtained, the capacity of ranges to support self-sustaining populations of caribou will have to be re-assessed. Addressing the need for more animal location and movement information, as well as greater inter-jurisdictional collaboration, are important requirements to more fully describe ranges for local populations of boreal caribou and support continuous improvement of critical habitat description over time.
Integrated Risk Assessment
Consideration of different lines of evidence resulted in a statement of likelihood for each range as to its current ability to support a self-sustaining population. Of the 57 ranges evaluated, 17 were assessed as likely or very likely to be self-sustaining (SS), 33 as unlikely or very unlikely to be self-sustaining (NSS), and seven ranges to be as likely as not be self-sustaining (NSS/SS). These results differ from those presented in the 2008 Scientific Review for nine ranges.
In the present evaluation, the use of decision rules rather than averaging of probabilities across all indicators resulted in evidence with higher certainty carrying greater weight in the integrated assessment. As well, population size could override the population growth indicators, recognizing the additional risks associated with small populations. This consideration resulted in one range moving from SS (2008) to NSS/SS (2011). In the remaining eight cases, the habitat-based population growth indicator carried more weight than population information, either because no population information was available or a general population trend was reported that was inconsistent with the habitat-based indicator. Four of these ranges were in Saskatchewan, where total disturbance placed the ranges in NSS (as compared to NSS/SS in 2008). Population trend was not available. However, a large proportion of the disturbance was fire, thus it is possible that improved demographic data may suggest that the ranges are currently supporting self-sustaining populations. The weight of evidence of the remaining four ranges with different assessments from 2008 was placed on the habitat indicator of population growth due to the lack of population data; this resulted in a more optimistic assessment in two cases and a more pessimistic assessment in the other two cases. Again, improved demographic data may suggest a different outcome for these ranges, however in these last four cases total disturbance was not dominated by fire.
Additional changes to range assessment outcomes at the national level were a result of new range delineations including 1) Ontario, where eight new ranges were delineated, including the combination of three ranges from the 2008 assessment, and 2) NT, where two ranges are recognized for the 2011 assessment, as compared to six management units assessed in 2008.
A designation of “not self-sustaining” should not be interpreted as a conclusion that a range cannot be recovered, or that the local population cannot be maintained. The current range assessment does not explore the potential for recovery of a range through the application of management activities designed to mitigate disturbance or accelerate recovery, but rather it provides the likelihood of a range supporting a self-sustaining population given the current amount of total disturbance and the current condition of the population. The integrated risk assessment further supports critical habitat description by providing information used to locate each range within intervals of disturbance associated with varying levels of risk (see threshold discussion below).
Many ranges lack information on population trend and population size, which forced a reliance on estimated population trend based on the habitat indicator, and prevented the assessment of extinction risk where population size was unknown. There is an urgent need to implement monitoring and assessment programs for these ranges, as well as to continue monitoring programs where they currently exist.
Range Specific Disturbance Thresholds
For the lines of evidence approach used in this assessment, habitat condition was a primary indicator related to the recovery criteria of stable or positive population growth, and is presented as the starting point for considering range specific management thresholds related to critical habitat. The intervals associated with each likelihood statement reflect a range of indicator values, consistent with a probabilistic representation of risk relative to the information considered. The assignment of relative risk is informed by science but its qualitative interpretation reflects the acceptance of varying levels of certainty in desired outcome. Further, the probability intervals themselves could be altered to express different breakpoints in desired certainty. In recognition that level of acceptable risk must be specified by managers, range- specific disturbance thresholds were not identified for each of the 57 ranges. However, a consistent methodology for determining range-specific disturbance thresholds is presented and examples of the application of the approach provided.
The disturbance intervals associated with the generalized relationship between range condition and population growth reflect variability in expected outcomes based on patterns evident at a national scale. Consistent with the integrated risk assessment, other indicators related to recovery criteria, also expressed relative to risk or likelihood of desired outcome, are used to refine interpretation of thresholds at a range-level. This approach can be further extended to consider potential future conditions. To this end, future range condition projections are provided for each range for consideration in the interpretation of range specific management thresholds. The projections were restricted to simple models of effects of additional natural disturbances (fire only) and passive recovery from both natural and anthropogenic disturbances. Interpretation of the results of the future range condition projections supports the need for greater (or lesser) caution in the determination of range specific disturbance thresholds, and the urgency for management actions to offset the risk of local extinction. The future condition models applied here do not, however, integrate additional anthropogenic disturbance. The model framework could be used for this purpose but projecting future anthropogenic developments was beyond the scope of this assessment.
In general, the less information available, the less certainty there will be in outcomes, and the more precautious the management approach should be with respect to conservation. Multiple lines of evidence that suggest similar outcomes create greater certainty, as does high quality information. Certainty in outcome is the principal measure recommended for range-specific refinement of disturbance thresholds relative to acceptable risk.
Bio-physical Attributes
The general biophysical attributes of woodland caribou habitats have been reasonably well-studied in much of Canada. However, given that use of features can vary depending on their relative availability, and also across ecological regions, it is important to understand attributes of potential significance within each range. Clearly some generality is possible, for example, with respect to habitats supporting ground and arboreal lichens, but most attributes have been found to vary across caribou distribution (e.g., Thomas and Gray 2002). Certain habitat components and the response by caribou remain poorly understood, including forage availability and diet selection, while others such as the significance of large expanses of bogs and old upland conifer are well documented. The bio-physical attributes provided for each range in this science assessment should be considered as a starting point and should be augmented with more detailed and range specific information from jurisdictions and other information sources, such as Aboriginal Traditional Knowledge (ATK), at the recovery strategy and action planning stage.
Application of the Science Assessment to Boreal Caribou Recovery
The focus for this science assessment was the provision of a scientific description of critical habitat for each boreal caribou range to inform critical habitat identification in the National Recovery Strategy. However, the range assessment and associated modelling results will also be useful at the action planning stage, by providing an evaluation of the status of range conditions relative to critical habitat requirements, and therefore informing the need for, and urgency of, management actions. The conceptual approach and modelling tools can also be applied to the evaluation of effective protection of critical habitat, as part of recovery and action plan implementation.
The future condition modelling completed in this science assessment provides general trend information for future disturbance conditions given natural disturbance patterns and passive recovery rates for a range. The primary application of this information was to inform the threshold interpretation (as described above) as a component of critical habitat identification, as well as to indicate management urgency in cases where current conditions were not sufficient to support a self-sustaining population. There was no attempt to integrate prediction of future development or management activities designed to accelerate the time to recovery for specific disturbance types. However, the tools for modelling future conditions were developed as a flexible framework to support action planning through the assessment of management scenarios. The integration of the habitat model with the population model can be used as a decision support tool for exploring population response to additional development and/or habitat recovery activities, thus providing an assessment of the probability of maintaining or achieving range conditions required to support self-sustaining populations. The population model also allows manipulation of population parameters based on information (if available) that quantifies changes in, for example, recruitment and/or adult survival in response to management actions that are not targeted at habitat changes (e.g., predator control).
A National Assessment - Strengths and Limitations
Like the 2008 review, the 2011 assessment is a national assessment designed to ensure a consistent methodology is applied to boreal caribou ranges across Canada. For example, to ensure consistent data inputs for disturbance mapping for the meta-analysis and assessment of each of the 57 ranges, data sources were utilized that were available across all boreal caribou ranges. Analyses undertaken using different data sources may yield different total disturbance results. Similarly, the resource selection analysis utilized variables describing land cover that were standardized across the distribution of boreal caribou in Canada. A nationally consistent approach allows for direct comparison of results across all areas considered, and when evaluating relationships, permits inclusion of data covering a broader range of conditions than available at narrower extents, such as regional scales. As such, it improves the identification and understanding of general relationships, where they exist. The robustness of the national meta-analysis of the relationship between caribou calf recruitment and total disturbance supports the strength of this approach. However, while this provides a strong foundation for this assessment, additional demographic and habitat information can significantly augment current understanding, particularly where uncertainty is high.
Application of Adaptive Management
The related goals of assessing the self-sustainability of ranges, and establishment of management thresholds for disturbance, must both acknowledge uncertainties resulting from availability and reliability of information about current population condition, as well as from limited knowledge about how populations will respond to additional and often interacting stressors. The probabilistic approach taken in this assessment, together with the application of a set of decision rules that relate the contribution of the information in each indicator to the strength of evidence about whether ranges are likely to maintain a self-sustaining local population, explicitly incorporates effects of uncertainties and data quality in the assessment process. This approach is consistent with the concept of adaptive management, which expresses likelihood (certainty) of outcomes as hypotheses. Disturbance thresholds and associated management actions are then implemented as carefully-designed experiments to reduce uncertainty and improve knowledge over time. A particular focus is to identify and avoid actions that carry a high risk of unintended outcomes or irreversible harm. Under this approach, substantial gains in knowledge can accrue from coordinated management and monitoring activities, or learning by doing, if the commitment to those activities remains strong.
The process of adaptive management can be represented as a continuous learning cycle, involving the key components of planning, doing, evaluating and adjusting (Figure 19). Associated with each of these are a number of activities to support the identification and reduction of uncertainties, in order to improve decision-making (e.g., the application of management thresholds). Central to the concept is the close integration of management, research and monitoring, where systems are not only carefully monitored, but management responses are nimble enough to change course in response to the weight of evidence. Adaptive management involves controlled management experiments guided by current understanding of system dynamics. Highly uncertain outcomes associated with different policy options become strong candidates for experimentation. The results of experiments are then used to inform adjustments to management strategies as necessary, in light of improved understanding.
Implementation of disturbance thresholds for caribou through active adaptive management could yield the greatest knowledge gains by prescribing a range of recovery actions and development activities, at varying levels of risk, in a way that does not preclude future management options. The range of current conditions across boreal caribou local populations suggests that sufficient contrast exists to support this approach. Where certainty in the likelihood of outcomes is high, the policy options are clearer relative to risk. However, there exists an intermediate range of disturbance levels over which outcomes for caribou local populations are highly uncertain. Understanding what factors contribute, and why, to a more or less desirable outcome with respect to caribou conservation, would significantly improve management effectiveness and reduce risk.
In conclusion, the purpose of this 2011 Scientific Assessment was to inform the description of critical habitat for a federal recovery strategy for boreal caribou in Canada by assessing the ability of current boreal caribou ranges to support self-sustaining local populations. The approach and framework developed for this assessment built upon and extended that presented in Environment Canada’s 2008 Scientific Review for the Identification of Critical Habitat for Woodland Caribou (Rangifer tarandus caribou), Boreal Population, in Canada. While improved data and enhanced understanding would help address remaining uncertainties as a component of adaptive management, this report concludes that sufficient information exists to support a scientifically-grounded assessment of critical habitat for populations of boreal caribou across Canada, and provides a scientific description of critical habitat for each of the 57 identified ranges that comprise the full extent of occurrence of boreal caribou in Canada.
The breadth of information and knowledge compiled for this assessment exemplifies the comprehensive nature of, and interrelationships between, types of evidence available to provide a scientifically-based description of critical habitat for informing recovery planning for boreal caribou. Significant advances were made to the conceptual and methodological design during this assessment to address some key uncertainties or limitations identified in the 2008 Scientific Review. These advances improved the robustness of the results with respect to providing a scientific description of critical habitat for boreal caribou across Canada.
Figure 19. The adaptive management cycle (from Jones 2009).
