Woodland caribou scientific review to identify critical habitat: chapter 6
6.1 Interpretation of Proposed Critical Habitat Outcomes
The application of the Critical Habitat Framework and associated Decision Analysis provided an assessment of all local populations or units of analysis within the current distribution of boreal caribou in Canada. Like habitat selection by caribou, Critical Habitat identi. cation is a hierarchical process that must consider needs across multiple spatial and temporal scales. The national analysis focused on the scale most appropriate for considering the persistence of local populations - the local population range. Factors operating at this scale act as constraints on population dynamics, and determine whether or not a population is likely to be sustained. It has been previously demonstrated, and is implicit in this evaluation, that predation acts as a limiting factor for boreal caribou populations. Conditions present in the range of a local population determine the type, amount and distribution of habitat for caribou and other prey species with shared predators on caribou, and hence the abundance and distribution of these predators within the range. As a result, the premise of this evaluation
- that Critical Habitat is most appropriately identi.ed at the scale of the local population range - is not equivalent to saying that every element within the range is critical to support a self-sustaining boreal caribou population, in all instances. However, it does provide a spatial delineation of the area of consideration when assessing the current conditions and quantifying risk relative to the recovery goal of maintaining or restoring self-sustaining local populations, for assigning potential Critical Habitat outcomes, and for planning for the management of the habitat conditions necessary to support population persistence (e.g. maintaining the functional attributes of the range). Re.nement of needs at .ner spatial scales over speci.c timeframes is possible within the constraints of the range level designation. Guidance on important considerations is provided in the Habitat Narrative (Appendix 6.3). General parameters associated with Critical Habitat outcomes are described below.
For each local population or unit of analysis, proposed Critical Habitat was expressed as one of three outcomes, based on weight of evidence from the integrated assessment (Range Self-Sustaining or Range Not Self-Sustaining; Section 2.6.5), and application of decision rules (Section 2.6.6). These outcomes included: Current Range, Current Range and Improved Conditions, or Current Range and Consider Resilience. An interpretation of each is provided below.
Current Range: Current range condition and extent are required to maintain potential for self-sustaining population. Further degradation of the current range may compromise the ability to meet the recovery goal. Five scenarios occurred under this outcome.
1) Local populations or units of analysis for several large and relatively continuous areas within the current distribution of boreal caribou have yet to be been delineated. The present assessment considered the extent of occurrence within the relevant jurisdiction as a single unit of analysis. In some cases, this indicated a moderate to high probability of the area supporting a self-sustaining population (P=0.6). However, caribou in the area may consist of more than one local population. As a result, the mean condition among these populations could be masking important variation across the extent of the area considered, with implications for population sustainability and critical habitat evaluation. Population units should be identi.ed and assessed, which could lead to alternative outcomes.
2) The uncertainty around population condition (trend unknown) in combination with moderate disturbance did not provide a clear indication of whether the current range is adequate to support a self-sustaining population (P=0.5). The .rst priority is to address the information gaps, then to re-assess the local population.
3) An integrated probability of P=0.5 when all parameters were known was interpreted as a marginal situation. The criteria assigned the greatest risk (lowest individual probability) should be examined, and additional local information considered.
4) Weight of evidence supported Range Not Self-Sustaining (P=0.4) for a number of local populations, but improvements to range condition were not clearly indicated, because either (a) disturbance was very low or low, or (b) population trend was stable. Maintenance of current range in conjunction with (a) investigation of other factors negatively affecting the population, or (b) close monitoring of trend for possible lag effects is recommended. Situations falling under (b) should also be examined to better understand potential resilience to different forms of disturbance.
5) In several cases, weight of evidence supported Range Not Self-Sustaining (P=0.4), but the total disturbance was comprised primarily of .re (e.g., the amount of anthropogenic disturbance was low or very low), and population trend was unknown. Improvements to range condition were not clearly indicated given that percent range burned explained little variation in the relationship underlying the disturbance categories, at least up to upper end of the moderate disturbance level. A better understanding of the differential effects of .re and anthropogenic disturbances on caribou demography was identi.ed as an area for further study.
Current Range and Improved Conditions: Current range conditions and/or extent would need to be improved to restore the potential to support a self-sustaining population. Further degradation of the range may have serious consequences for local population persistence. Three scenarios occurred under this outcome.
1) For most local populations or units of analysis with weight of evidence supporting Range Not Self-Sustaining (P=0.4), levels of anthropogenic disturbance in conjunction with population trend suggest that recovery efforts are required to restore conditions that support persistence (e.g., a reduction in anthropogenic disturbance and recovery of disturbed habitat is necessary). The nature and magnitude of restoration could be determined through spatial population modeling combined with dynamic landscape simulation.
2) For several local populations or units of analysis, a high level of total disturbance was comprised primarily of .re, with low levels of anthropogenic disturbance, but was associated with a declining population trend. The percent area burned fell outside the range of values included in the meta-analysis (Appendix 6.5), thus inference based on the documented relationship was weak. Natural recovery may be suf.cient to improve range condition, but additional stressors on the population should be considered, including potential interactions between .re and anthropogenic disturbance at high levels of .re, and non-habitat factors (e.g., mortality sources).
3) In two cases, the total measured disturbance levels were low or very low, but a negative population trend indicated the need for improved range conditions and/or extent. Therefore, aspects of habitat condition other than disturbance may be affecting the local population. Non-habitat factors such as poaching, reduction in habitat quality for example low .ying aircraft or other forms of disturbance not included here, and population health (disease and parasites) should also be considered. It is also possible that the current range has been reduced in extent such that it is insuf.cient to support a self-sustaining local population, and restoration of adjacent habitat is required to enable the population to persist. Population isolation and the need to restore connectivity should be examined.
Current Range and Consider Resilience: Current range condition and extent may be suf.cient to absorb additional disturbance while maintaining capacity to support a self-sustaining population. Two scenarios occurred under this outcome.
1) Local populations or units of analysis with large or very large population size (e.g., above critical based on the non-spatial population viability analysis), stable or increasing population trend, and levels of total disturbance that were moderate, low or very low. This situation presents the least risk with respect to meeting the population objective of the recovery goal, and represents the greatest potential to apply active adaptive management to evaluate resilience (e.g., experimental management to test alternate hypotheses regarding population responses to different types and levels of disturbance).
2) Local populations or units of analysis with small population size, stable or increasing trends, and low or very low levels of total disturbance. This situation also represents a relatively high probability of achieving the recovery goal. However, the inherent risks associated with a small population size warrant a cautious approach when considering potential resilience to any additional disturbance. Nevertheless, this situation may also present an opportunity for active adaptive management.
One of the guiding principles of the science review was to recognize and address the dynamic nature of boreal systems and resultant effects on boreal caribou habitat in time and over space. Boreal landscapes are naturally dynamic, driven by processes such as . re and other disturbances and resultant forest succession. Similar landscape dynamics may be associated with certain types of anthropogenic disturbances. Recognition of such dynamics is commensurate with the scale of consideration for Critical Habitat identi.cation - the local population range - which re.ects multi-decadal dynamics of the system and species response. However, neither the spatial nor temporal dynamics within a local population range were directly addressed by this evaluation.
The non-spatial population viability analysis considered temporal components of persistence associated with demographic, and to some extent, environmental stochasticity. As well, the 50-year window for area burned considered by the meta-analysis recognized in a limited way the dynamic properties of disturbance by .re, relative to habitat recovery and response by caribou. Nonetheless, the present evaluation represents a point-in-time assessment of the current range relative to the recovery goal of self-sustaining local populations.
Further elaboration of Critical Habitat outcomes for local populations can be achieved through spatial population viability analysis linked with dynamic landscape modelling (see Section 2.6.6 and Appendix 6.7). Incorporation of landscape dynamics is necessary to understand the conditions and management options associated with recovery (Current Range and Improved Conditions) and resilience (Current Range and Consider Resilience), as well as additional risks associated with present conditions (Current Range). Such evaluations may be undertaken with varying levels of complexity and concomitant requirements for data. It is clear from the present review that minimum data requirements could be met for most areas within the current distribution of boreal caribou in Canada, particularly when viewed in the context of adaptive management.
6.2 Decision Analysis and Adaptive Management
The Decision Tree provided a structured and transparent method to evaluate individual local populations and determine prior probabilities of alternative hypotheses regarding de.nition of Critical Habitat, through consideration of measurable criteria assigned to categorical states based on available quantitative data and published scienti.c information. The prior probabilities indicated the most plausible outcome, relative to probability of persistence, for each local population or unit of analysis. At each step in the Decision Tree, any assumptions made were explicitly described, and uncertainties were identi.ed that could be addressed through a Schedule of Studies to improve understanding.
The approach to Critical Habitat identi.cation applied here follows established methodologies for decision-analysis in operations research and management science. In this case, the objective function is population persistence, expressed as the set of conditions necessary to support self-sustaining local populations. Syntheses of existing information, evaluation of likely outcomes, and re.nement of understanding are also fundamental components of the adaptive management framework. While a more detailed Decision Tree could be developed to elucidate the relationships among criteria (variables) and identify underlying mechanisms, the simple model considered here is a "white box" that can be easily applied, evaluated, and communicated with available information, and supports a science based component of the process leading to the potential . nal identi.cation of proposed Critical Habitat across a spectrum of local population conditions.
The assignment of prior probabilities and their use in the identi.cation of Critical Habitat represents a starting point in an adaptive management cycle (Figure 4). As uncertainties are addressed through the Schedule of Studies, and new information becomes available, local population assignments can be updated. The Decision Tree can also be interpreted as a Bayesian Decision Network (BDN). The assessment criteria are equivalent to nodes in a BDN, representing variables that can assume multiple states. Associated with each node is a probability table that expresses the likelihood of each state, conditional on the state of nodes that feed into it. Weightings could be assigned to nodes to represent the relative importance of the variable on the outcome. The current process does not address interactions among the criteria or their relative in.uence on outcomes, so no weightings were applied to the assessment criteria (population trend, population size, and range disturbance), nor were conditional probabilities assigned to individual criteria. However, estimation methods for generating these probabilities exist, and can be incorporated over time through the adaptive management process. Development of a more comprehensive BDN is recommended as part of the Schedule of Studies, to enhance understanding and provide a formal process for updating the prior probability distribution for the recovery goal of self-sustaining local populations.
6.3 Transition to Action Planning/Recovery Implementation
As previously noted, this national analysis and proposed identi.cation of critical habitat was conducted at a spatial scale appropriate to addressing persistence of local populations, as per the recovery goal and objectives for this species. However, habitat selection by boreal caribou is hierarchical, and where/if deemed necessary, assessments may be further re.ned within local population ranges to identify the habitat necessary for the recovery of the species at .ner temporal and spatial scales.
A variety of approaches could be applied at the local population level to de.ne the degree of change required in range condition and/or extent to support persistence, the appropriate management strategies for maintaining conditions where range is currently self-suf.cient, and the amount of additional disturbance that might be absorbed by local populations with potential resilience. For example, the probability of persistence over speci.ed time frames can be further quanti.ed using spatially explicit population viability analysis to model the fate of populations relative to changing habitat conditions, and to identify probable outcomes under a range of habitat scenarios. By linking spatially explicit population and landscape simulation models, dynamic elements of the system can be incorporated (see Appendix 6.7 - spatial PVA). Further meta-analyses could be applied across multiple populations to link current conditions (e.g., vegetation composition and structure), created by natural and anthropogenic factors, to population status, and predict future trends. Similarly, a retrospective approach could be used to explore conditions for persistence, by quantifying historic variation in natural systems and examining circumstances that have contributed to persistence, recognizing the uncertainty among persistence, historical disturbances, and habitat change. Such investigations could also yield insights into the differential effects of .re and anthropogenic disturbance on caribou demography; an important distinction when considering the application of such approaches to caribou management.
The Boreal Caribou Critical Habitat Science Review performed by EC was undertaken with the support of an independent Science Advisory Group that provided continuous peer-review throughout the process. Development of a Critical Habitat Framework and Decision Tree provided a formal structure for assembling and analyzing data relevant to Critical Habitat identi.cation, and the foundation for continuous improvement of knowledge through the process of adaptive management. A weight of evidence approach was used to identify the most plausible outcome of combinations of population and habitat conditions relative to the recovery goal of self-sustaining local populations.
This report contains a proposed Critical Habitat identi.cation, based on empirical science and inherent assumptions associated with the methodology used, for each of the spatial analytical units associated with each local population. Other factors such as the incorporation of Aboriginal and traditional knowledge (ATK), and the extent to which the assumptions taken in this report align with Critical Habitat policy directives, may in.uence any potential .nal identi.cation of Critical Habitat in the Recovery Strategy.
General conclusions from the review include:
1) Critical Habitat for boreal caribou is most appropriately identi.ed at the scale of local population range, and expressed relative to the probability of the range supporting a self-sustaining local population;
2) Range is a function of the extent and condition of habitat, where habitat includes the suite of resources and environmental conditions that determine the presence, survival and reproduction of a population;
3) Application of the Critical Habitat Identi.cation Framework, for the 57 recognized local populations or units of analysis for Boreal caribou in Canada, yielded 3 proposed outcomes: Current Range, Current Range and Improved Conditions, or Current Range and Consider Resilience;
4) Like habitat selection by caribou, Critical Habitat identi.cation for Boreal caribou is a hierarchical process with considerations across multiple spatial and temporal scales. Further elaboration of Critical Habitat outcomes at spatial scales .ner than range, over speci.ed time frames, may be achieved through spatial population viability analysis linked with dynamic landscape modelling;
5) Acknowledging that current knowledge and the dynamic nature of landscapes impart uncertainty, present .ndings should be monitored and assessed for the purposes of re.nement and adjustment over time, as new knowledge becomes available (e.g., a Schedule of Studies as part of Adaptive Management).
This science based review was framed as one of transparent decision-analysis and adaptive management. Thus, the Schedule of Studies produced is a key requirement of the process that is designed to produce continuously improving results over time. Aboriginal and Traditional Knowledge was not included in the present review, nor are needs speci.c to this body of knowledge identi.ed in the Schedule of Studies.
6.5 Addressing Uncertainty - A Schedule of Studies
All readily available information, including peer-reviewed and grey literature, caribou population and location data, and biophysical and land-use data was reviewed to support the Critical Habitat Decision Analysis. A Schedule of Studies is required by SARA (S. 41(1) (c.1)) if suf.cient information is not available to complete the identi.cation of Critical Habitat. Thus, a Schedule of Studies remains a requirement of the process, as described throughout this document. The Schedule of Studies is an outline of activities (e.g., survey work, mapping, population viability analysis) designed to address knowledge gaps and uncertainties to improve the Critical Habitat identi.cation process. These activities include new studies, improvement or continuation of existing studies, and collection of standardized data through monitoring and assessment. Aboriginal traditional knowledge was not considered in the present Science Review, except where accessible in published documents, nor are needs speci.c to this body of knowledge addressed in the Schedule of Studies. Aboriginal and traditional knowledge provides important information that could augment this review and improve understanding of critical habitat for boreal caribou.
The following Schedule of Studies is designed to address uncertainties identi.ed at each step in the Decision Tree (see Figure 4).
Table 7: Schedule of Studies
|Identify Current Distribution:
The current distribution of boreal caribou across Canada is described and mapped in order to defi ne the
national scope of Critical Habitat Identifi cation.
|Environmental Nich Analysis||The Environmental Niche Analysis (Appendix 6.4) should be further
developed and applied to identify areas of uncertainty based on available
abiotic and biotic data, and therefore guide sampling efforts to refi ne
understanding (model-based sampling) of the drivers of current distribution,
as well as patterns of occupancy within the distribution. This method could
also be used to identify areas with high restoration potential, and areas for
enhancing population connectivity, where necessary.
|Identify Unit of Analysis:
The ranges of local populations are the unit of analysis for Critical Habitat Identifi cation
|Develop a Local Population
Range Mapping Standard
|Develop a standardized approach to delineating local population ranges
(units of analysis) that can be applied across Canada by jurisdictions
responsible for the management of Boreal Caribou.
|Determine Local Populations||Determine and/or update local population ranges using standardized criteria
Note: Delineation of local populations is a high priority for large continuous
distribution areas currently lacking this information.
|Population and Habitat Assessment :
Application of a systematic process for evaluating the probability of persistence of a local population given
observed states of population and range condition.A
|Develop a comprehensive
Bayesian Decision Network
|Identify and incorporate measurable parameters (variables) that infl uence
population persistence into a comprehensive BDN that specifi es the
conditional probabilities among variables, and provides a formal method for
updating Critical Habitat assignments with new knowledge. This activity
will be informed by results from additional meta-analyses and non-spatial
and spatial population viability analyses.
|Conduct additional metaanalyses
demography and range
|Extend analyses of national data to incorporate additional measures of
population and range condition (e.g., adult survival, habitat fragmentation,
forest composition), understand variation in relationships attributable to
different disturbance types, other habitat measures, or regional contexts, and
augment or refi ne criteria used to assess range condition for identifi cation
of Critical Habitat.
|Refine population size
thresholds in relation to
probability of persistence
|Further develop the Non-Spatial PVA by:
■ Incorporating maximum age and senescence
■ Evaluating interactions between selected demographic parameters,
and the infl uence of population size on these relationships, relative to
risk of extinction and expected time to extinction
|Develop survey standards||Develop standardized criteria and methods for boreal caribou population
assessments, including local population size and trend information.
|Determine local population
|Population demographic data are required to calculate lambda and evaluate
trends of local populations, including more detailed demographic data (from
survival analyses, population composition and recruitment surveys).
|Determine local population
|Population census data are required to determine current population size.|
|Critical Habitat Identifi cation:
Determining the quantity, quality and spatial confi guration of habitat required for persistence of boreal caribou
populations throughout their current distribution in Canada.
|Refine Quantity, Quality and
Spatial Confi guration of Critical
Habitat for local populations
|Identifi cation and completion of case studies using spatially-explicit
population modeling to explore a range of population and habitat conditions,
and management scenarios, to improve understanding of habitat-based
constraints on population persistence (quantity, quality and spatial
confi guration) and inform development of the Bayesian Decision Network.
A variety of modeling approaches should be explored, to inform Critical
Habitat identifi cation and recovery planning (e.g. effective protection and
Alternative analytical approaches, such as additional meta-analyses, can
also support this activity.
|Develop and/or apply methods
for determining needs and
conditions to support population
|Critical Habitat has been identifi ed at the scale of the range of local
populations, with the assumption that local populations experience
limited exchange of individuals with other groups. Enhanced population
connectivity may be necessary to support persistence of small populations,
and maintenance of existing connectivity an important element of Critical
Habitat for large populations. Development and/or application of methods to
evaluate population connectivity and its relationship to habitat or landscape
attributes is necessary. This work could be undertaken in conjunction with
spatially-explicit population modeling.
|Identify opportunities for active
|Uncertainties regarding the potential resilience of local populations to
different levels and types of disturbance may be most effectively addressed
through active adaptive management designed to test alternate hypotheses
regarding population response. Parameters to support this could be
identifi ed through spatially-explicit population modelling.
- Date modified: