Title: Range of Natural Variability: What does it mean for sustainable forest management?
Publication Type: Journal Article
Year of Publication: 2002
Authors: Drever R, Wong C
Journal: Ecoforestry
Volume: 17
Issue: 2
Start Page: 11-17

Introduction:

What is RONV?
 
The concept of the range of natural variability[1] is increasingly recognized as an important element of strategies for sustainable forest management. Generally, the range of natural variability (RONV) refers to the spectrum of natural conditions possible in ecosystem structure, composition, and function, when considering both temporal and spatial scales (Swanson et al. 1994) (Figure 1). This concept is useful for evaluating how management interventions affect the structures, composition, and functions of forest ecosystems. Here, our objectives are to review the concept of RONV, discuss its merits and challenges and provide a checklist with which to evaluate a defensible determination of RONV.

Figure 1 – Hypothetical range of natural variability for a given attribute of interest in a forest. This could refer to stand-level
attributes such as snag density or amount of tall old trees as well as landscape-level attributes such as the amount of old forest or
size of stand-replacing fires. The outside lines indicate the full range of possible values over time, whereas the inside dotted lines
indicate the range within which lie 90 percent of the values for the attribute of interest. The box at the right side of the figure indicates
a hypothetical “apparent range of variability” – the variability that can be observed by a given method of observation, e.g., tree
cores.

Why is it important for sustainable forest management?


Holling and Meffe (1996) defined the “Golden Rule” of sustainable natural resource management as: “management should strive to
retain critical types and ranges of natural variation in resource systems in order to maintain their resiliency.” Retaining ecosystem
 structure, composition, and function in managed forests within their RONV increases the likelihood of maintaining ecosystem integrity
and biodiversity. Ecosystem integrity and biodiversity are thought to be two characteristics of ecosystems that allow them to be
resilient, i.e., capable of absorbing disturbances without reorganizing from one ecosystem to another stable state (Holling 1973; Folke
et al. 1996). Forest ecosystems that are resilient are capable, theoretically, of providing in perpetuity ecological services and products
such as water filtration and timber (Scheffer et al. 2001).
 
What are the challenges in using RONV?

Using RONV to guide planning and implementation of forest practices is not simple. The use of RONV is inherently challenging
because it attempts to characterize ecological processes and patterns that can vary greatly across time and space and are stochastic
rather than strictly deterministic (Swanson et al. 1994; Landres et al. 1999). Moreover, little consensus exists in the scientific literature
about clear definitions (e.g., just what is “natural”?) and what methods are appropriate for measuring and applying the RONV concept.
Several of these issues are related to scale, in particular about how to adequately determine the ‘bounds’ of an ecosystem that remain
relatively consistent over time and space (Morgan et al. 1994).

Simply put, how far back in time and how big of an area is necessary
to capture the full range of variability? In general, the scales of analysis for forest research and management depend intimately on the
ecosystem attribute or landscape pattern in consideration, i.e., the nature of the disturbance regimes that create the attribute or pattern,
their climatic and topographic “drivers,” and the lifespan of trees (Allen and Hoekstra 1991; Swetnam et al. 1999; Wong and Iverson
2001). For instance, measuring the variation in the recruitment of large downed cedar material in coastal rainforests involves time
scales on the order of several hundred to a thousand years, whereas to determine the RONV in the frequency of fires in interior
 Ponderosa pine forests involves measuring intervals between fires over at least a hundred years.

Temporal Scale


No single, widely applicable optimal period exists for determining RONV, and “relevance is lost if too long a time period is used,
because conditions such as climate and species composition may have changed dramatically” (Landres et al. 1999). For forest
ecosystems, an appropriate time scale should generally encompass several generations of trees or disturbance events in order to
estimate the variation between events (Morgan et al. 1994). Thus, selecting an appropriate time period becomes a trade-off between a
period long enough to fully capture natural variability but not so long that it encompasses significant changes like climatic shifts.
A temporal extent of 2,000 years – as, for example, is stipulated in the Forest Stewardship Council’s definition of RONV – is likely
broad enough to allow achieving this tradeoff in most ecosystems for the various ecosystem attributes considered in the Standards
(FSC-BC Standards Team 2001). However, the temporal extent must consider RONV before the advent of European settlement to
estimate the variation between events (Morgan et al. 1994). Thus, selecting an appropriate time period becomes a trade-off between a
period long enough to fully capture natural variability but not so long that it encompasses significant changes like climatic shifts.

A temporal extent of 2,000 years – as, for example, is stipulated in the Forest Stewardship Council’s definition of RONV – is likely
broad enough to allow achieving this tradeoff in most ecosystems for the various ecosystem attributes considered in the Standards
(FSC-BC Standards Team 2001). However, the temporal extent must consider RONV before the advent of European settlement to
avoid capturing our relatively rapid (from the perspective of ecosystem change) impacts on forest structure and composition from
 clear cutting, road building, igniting fires for mineral exploration, introduced species, fire suppression, livestock grazing (Wallin et al.
1996; Delong and Tanner 1996). Although 2,000 years from present encapsulates a time period during which temperate climates,
ecosystems, and their disturbance regimes were relatively stable (Hann et al. 1997; FSC-BC Standards Team 2001), the Little Ice Age
(1450-1850) has been documented in BC (Smith and Laroque 1996) and estimating RONV within its context is challenging (Miller and
Woolfenden 1999).

Unfortunately, the temporal scale will most often be limited by the availability of data. For most disturbance regimes, very little evidence
extends back for 2,000 years in time. For example, most fire history studies in the interior dry Douglas-fir and Ponderosa pine forests
 are based on the record of fire-scarred trees that does not extend back further than the 15th century and most periods of reliability[2]
begin from the early 1700s onwards (e.g., Gray et al. 1998, Riccius 1998, Wong 2000, Heyerdahl and Lertzman unpublished).
Disturbance history is also difficult to reconstruct in the coastal forests in southern BC because much of it has been logged.
The use of simulation modeling (i.e., repeated projections through long time periods) and substituting space for time in sampling (i.e.,
sampling many stands to capture the variability in stand development that may occur throughout time in one stand; Lertzman et al.
1998) may allow us to overcome some of the limitations of the historical record.

Spatial Scale

The relevant spatial scale of analysis depends fundamentally on the scale at which the attribute in consideration functions, as well as on
the relevant management scale. For example, if RONV in landscape pattern is estimated from simulated disturbance regimes, the
model extent (i.e., the size of a simulated landscape) should be at least 10 times larger than the largest expected disturbance
(Cumming and Burton 1996). In general, at least a landscape scale is necessary for setting retention targets for landscape pattern and
other goals for forest management and health, according to the RONV concept (Dellasalla et al. 1996).

Moore et al. (1994) provide more specific direction by recommending that RONV be assessed over areas that are consistent in terms
of their edaphic, topographic, and bio-geographic conditions. For example, Gavin (2000) found the time-since-fire varied greatly with
topography in Clayoquot Sound − the median time-since-fire was significantly six times greater on terraces (ca. 4,500 years) than 
hill slopes (ca. 750 years). Thus, there are specific “terrain units” or “process domains” (Montgomery 1999) – physical units over which
disturbance regimes act relatively uniformly – which can form the spatial resolution of RONV estimates; i.e., RONV estimates on
terraces should guide management on terraces and RONV estimates on hill slopes should guide management on hill slopes. This is a
clear example of where averages over larger areas are unlikely to be ecologically meaningful without consideration of process
domains (Lertzman et al. 1998).
 
Checklist for managers to evaluate how RONV was determined
In most areas in BC, data on RONV are scarce or absent. This is due to a lack of research, lack of remaining evidence of past
structure and function and/or lack of methods (e.g., spatially explicit simulation models) to derive RONV. Moreover, even where data
 are present, they may only represent the portion of RONV they were able to capture, rather than the entire range of variability. The term
”apparent range of variability” has been suggested to convey the dependence of the definition on the temporal and spatial extent of
available data and information (Figure 1; Wong 1999, Wong and Iverson 2001) – e.g., the fire history of an area could be that apparent
from 200- 400 years of tree ring records sampled over 100 ha on a north-facing aspect.
Forest managers and auditors will often be faced with evaluating whether the temporal and spatial scales over which the RONV was
determined are adequate. For example, the field testing for Draft 2 of the FSC Regional Certification Standards for British Columbia
indicated that simpler guidance in the RONV concept is required for auditors to consistently use RONV as a verifier (Moore
2001).

Below we suggest a checklist for managers and auditors to determine whether RONV has been adequately determined.
 
When evaluating RONV for stand- and landscape-level attributes, was RONV estimated by:

1. Using data from stands or landscape with sufficient pre-settlement structure?

2. Sampling stands or landscapes representative of pre-settlement conditions? For instance, remaining old-growth stands are often
a function of previous logging patterns and may not represent the entire range of pre-settlement stand conditions.

3. Sampling stands or landscape representative of the operating area? Or, if landscape pattern was simulated, was RONV
estimated by using locally determined disturbance parameters for the spatially explicit model and simulating a landscape similar
in topography and vegetation as the operating area?

4. Sampling at a resolution appropriate for describing the variation of the attribute in question?


5. Methods that account for post-settlement changes? For example, stand reconstruction of logged areas must examine whether
patterns of stand development in a given area are “natural” or due to fire exclusion from fire suppression and fuel reduction from
grazing.

6. Considering temporal and spatial trends in the data?

7. Considering the uncertainty and assumptions in data by using sensitivity analyses? This includes age errors in the Forest
Inventory Planning data, assumptions about the original composition and structure of logged stands, assumptions about the role
of insects.

8. Using appropriate methods for determining RONV? For example, fire return intervals determined from fire scars must be from
cross-dated samples[3] whereas fire return intervals determined from time-since-fire maps (in BC, most studies have used the
forest cover maps as proxies) require different assumptions depending on the method used (see review in Wong et al. 2002).


9. Defining clearly, consistently and a priori what part of the distribution will define the “range”?

10. Using a process that included independent peer-review?

How is RONV used in current forest management?

At the time of writing, the concept of RONV is incorporated into forest management in British Columbia (BC) through the Biodiversity
Guidebook of the Forest Practices Code (B.C. Ministry of Forests and B.C. Ministry of Environment, Lands, and Parks 1995). The
Biodiversity Guidebook classifies BC into five different natural disturbance types, based on BC’s established system of ecosystem
classification. From this classification flow management recommendations for targets of landscape-level attributes such as the
distribution of seral stages and connectivity, as well as targets for stand-level retention of forest structure and wildlife
trees. Unfortunately, most of these recommended targets are not based on empirical data of natural disturbance regimes – rather, they
use informed professional judgment and “represent an attempt to integrate society’s desire both to generate commercial forest
products and to ensure the conservation of biological diversity in managed forests” (B.C. Ministry of Forests and B.C. Ministry of
Environment, Lands, and Parks 1995). In some cases, the targets are inconsistent with recent research on natural disturbance
regimes. For example, based on a mean return interval for disturbances of 250 years, the Guidebook recommends a target for less
than 30 percent of a given forest landscape in the coastal rainforest to be in early seral stage; this contrasts with findings that coastal
rainforests are dominated by old-growth and have mean disturbance return intervals between ca. 750 and 6,000 years (Gavin 2000).

General recommendations for appropriate use of RONV


1. Require that data sources and scales of analyses used to estimate RONV and any resulting guidelines for management be
explicitly stated – this allows certifiers and other interested parties to make their own judgement about the scientific validity of the
 RONV assessments used to justify a given management prescription.

2. Use a per-cautionary approach to the interpretation of RONV.

This means:


a) Avoid setting levels of retention for stand structural attributes and landscape pattern based solely on estimates of the
mathematical range. The minimum or maximum value of given attributes over long-time scales are typically defined by rare,
extreme events (Landres et al. 1999). Moreover, uncertainty about ecosystem attributes is typically highest at their spatial and
temporal limits (Landres et al. 1999) – i.e., the sampled record “fades” in quality the further back in time. Focus instead on the
distributions around the central tendency for the attribute of interest, e.g., values within the 90 percent confidence intervals or the
 80 percent percentile of the distribution, as well as the shape of the curve that describes the attribute, e.g., normal, negative
exponential, etc (Figure 1).


b) Where large uncertainty exists in estimates of RONV, resulting descriptions should err on the side of favoring the persistence
of attributes of conservation concern. For example, landscape-level simulations of landscape pattern in the Arrow Timber Supply
Area were greatly affected by assumptions about what age stands were before harvest (Dorner and Lertzman 2001).
Disturbance intervals – the average time between disturbances − were shorter if it were assumed that logging historically
targeted all stands greater than 120 years equally than if it were assumed that logging more likely targeted older stands (Dorner
and Lertzman 2001). In this case, if late successional forests are of conservation concern, it is necessary to use the second
assumption to derive RONV estimates of disturbance intervals in the Arrow TSA.

3. Make explicit that watershed- or landscape-level analyses are necessary to quantify RONV for structures and patterns that occur
at this scale and that extrapolating information determined in a different location but in a similar bio-geo climatic sub-zone/variant is
not an adequate estimate of local RONV (although it may do in the interim in the absence of data). Process domains are likely
better units upon which to base RONV estimates (Cissel et al. 1999; Montgomery 1999).

4. Commit to long-term adaptive management – monitor how well various management activities maintain the RONV in specific
 attributes and affect biodiversity at the stand and landscape levels, as well as the capacity of ecosystems to resist and be
resilient to disturbances.

5. Recognize that determining RONV for stand and landscape level attributes can be an intensive and expensive process posing
 specific challenges for small forest management operations. Innovative ways to overcome this challenge should be considered in 
future discussions on RONV. Where there is an absence of data, an interim definition of RONV can be based on inferences from
 similar forests or from expert opinion from local forest ecologists (Swetnam et al. 1999).


Conclusion


RONV is essentially a “coarse-filter” approach to sustainable forest management, aimed at conserving structural and biological
diversity through maintaining forest heterogeneity. It, however, is not an easy concept to operationalize because it has various
expressions at different temporal and spatial scales. Judicious use of the RONV concept requires its careful definition both temporally
and spatially, careful sampling/modeling, and consistent and precautionary interpretation in forestry and land use planning.


C. Ronnie Drever, when he wrote this article, was with the David Suzuki Foundation, Suite 219, 2211 West 4th Ave., Vancouver BC
 V6K 4S2 t: 604-732- 4228


Carmen Wong, Independent Consultant, 651 Commonage Rd., Vernon, BC, Canada, V1H 1G3 t: 250-558-5292

Footnotes

[1] Other terms used to describe the same or a similar concept include: “historic or historical range of variability,” “natural range of
variation,” “reference variability” or simply “natural variability or variation.”

[2] A period of time with sufficient replication in the tree-ring record that fire regimes can be confidently reconstructed.
[3] Ring widths on fire-scarred samples are cross-dated to master chronologies of climatic patterns to obtain annual accuracy of fire
dates.

Acknowledgements


Helpful comments on this paper were provided by Brian Starzomski.

Literature Cited


Allen, T. F. H. and T. W. Hoekstra. 1991. Role of heterogeneity in scaling of ecological systems under analysis. In J. Kolasa and
 S. T. A. Pickett (eds.) Ecological Heterogeneity. Springer-Verlag, New York.
B.C. Ministry of Forests and B.C. Ministry of Environment, Lands, and Parks. 1995. Forest Practices Code of British Columbia:
 Biodiversity Guidebook. Province of British Columbia, Victoria.
 Cissel, J.H., F.J. Swanson, P.J. Weisberg. 1999. Landscape management using historical fire regimes: Blue River, Oregon.
Ecological Applications 9(4): 1217-1231.
 Cumming, S. G. and P. J. Burton. 1996. Boreal mix wood may have no “representative” areas: some implications for reserve
design. Ecography. 19:162-180.
 Dellasalla, D.A., J.R. Strittholt, R.F. Noss, and D.M. Olson. 1996. A critical role for core reserves in managing Inland Northwest 
landscapes for natural resources and biodiversity. Wildlife Society Bulletin 24(2):209-221 
DeLong, S.C., and D. Tanner. 1996. Managing the pattern of forest harvest: lessons from wildfire. Biodiversity and Conservation 
5:1191-1205.
 Dorner, B., and K. Lertzman. 2001. Disturbance history in the Arrow TSA: an explicit reconstruction of historical landscape age
structure from modern-day aerial photography. Report to the Arrow IFPA.
Folke, C., C.S. Holling, and C. Perrings. 1996. Biological Diversity, Ecosystems, and the Human Scale. Ecological Applications
6: 1018-1024.
 FSC-BC Standards Team 2001. Forest Stewardship Council (FSC) Regional Certification Standards for British Columbia Draft
2 – MAY 22, 2001. Prepared by the FSC-BC Standards Team for the FSC-BC Regional Initiative Steering Committee.
Gavin, D. 2000. Holocene Fire History of a Coastal Temperate Rain Forest, Vancouver Island, British Columbia, Canada.
Unpubl., PhD. Thesis. Univ. of Washington. Seattle, WA.
Gray, R.W., E.H. Riccius, and C. Wong. 1998. Comparison of current and historic stand structure in two IDFdm2 sites in the
Rocky Mountain Trench. R.W. Gray Consulting, Ltd., Chilliwack, BC.
Hann, W.J., J.L. Jones, M.G. Karl, P.F. Hessburg, R.E. Keane, D.G. Long, J.P. Menakis, C.H. McNicoll, S.G. Leonard, R.A.
Gravenmier, and B.G. Smith. 1997. Landscape dynamics of the Basin. In an assessment of ecosystem components in the
Interior Columbia Basin and portions of the Klamath and Great Basins, Volume II (T.M. Quigley, S.J. Arbelbide, technical
editors). USDA Forest Service General Technical Report PNW-GTR-405, Pacific Northwest Research Station, Portland, OR.
Heyerdahl, E.K., and K.P. Lertzman. unpubl. Historical mixed severity fire regimes in the southwestern interior of British
Columbia. Ecological Applications.
Holling, C.S. 1973. Resilience and stability of ecological systems. Annual Review of Ecology and Systematics 4: 1-23.
Holling, C.S. and G. K. Meffe. 1996. Command and control and the pathology of natural resource management. Conservation
Biology 10(2): 328-337.
Landres, P.B., P. Morgan, and F.J. Swanson. 1999. Overview of the use of natural variability concepts in managing ecological
systems. Ecological Applications 9(4):1179-1188.
Lertzman, K.P., J. Fall, and B. Dorner. 1998. Three kinds of heterogeneity in fire regimes: At the crossroads of fire history and
landscape ecology. Northwest Science 72:4-23.
Miller, C. I. and W. B. Woolfenden. 1999. The role of climate change in interpreting historical variability. Ecological Applications.
9:1207-1216.
Montgomery, D. R. 1999. Process Domains and the River Continuum. Journal of the Water Resources Association 35(2):1-14.
Moore, K. 2001. Field-Testing Draft 2 of the FSC Regional Certification Standards for BC Report 2 – Findings and
Recommendations.
Morgan, P., G.H. Aplet, J.B. Haufler, H.C. Humphries, M..M. Moore, and W.D. Wilson. 1994. Historical range of variability: a
useful tool for evaluating ecosystem change. Journal of Sustainable Forestry 2:87-111.
Riccius, E.H. 1998. Scale issues in the fire history of a fine grained landscape. Master’s thesis, Simon Fraser University,
Burnaby, BC.
Scheffer, M., S. Carpenter, J.A. Foley, C. Folke, and B. Walker. 2001. Catastrophic shifts in ecosystems. Nature 413: 591 – 596
Smith, D. J. and Laroque, C. J. 1996. Dendroglaciological dating of a Little Ice Age glacial advance at Moving Glacier,
Vancouver Island, BC. Geographie physique et Quaternaire. 50: 47-55.
 Swanson, F.J., J.A. Jones, D.A. Wailin, and J.H. Cissel. 1994. Natural Variability – Implications for Ecosystem Management.
Pages 85-99 in USDA Forest Service General Technical Report PNW-GTR- 318.
Swetnam, T.W., C.D. Allen, and J.L. Betancourt. 1999. Applied historical ecology: using the past to manage for the future.
Ecological Applications 9(4): 1189-1206.
Wallin, D. O., F. J. Swanson, B. Marks, J. H. Cissel and J. Kertis. 1996. Comparison of managed and pre-settlement landscape
dynamics in forests of the Pacific Northwest. Forest Ecology and Management. 85:291-309.
Wong, C. M. 1999. Memories of natural disturbances in ponderosa pine-Douglas-fir age structure, southwestern British
Columbia. Master’s Thesis, Simon Fraser University, Burnaby, BC.
Wong, C.M. 2000. Natural disturbance regimes in the Cariboo region: what is known to guide forest management? Report for
Lignum Ltd., Williams Lake, BC.
Wong, C. and Iverson, K. 2001. Range of natural variability: Understanding and applying the concept to forest management in
central British Columbia. Report for Lignum Ltd., Williams Lake, BC.
Wong, C. M., Dorner, B. and H. Sandmann. 2002. Estimating historical variability of natural disturbances in BC. Part 2: A
 Catalogue of Methods. Report for Research Branch, BC Ministry of Forests, Victoria, BC.

drever_wong_range_var_graphic.jpg