Volume 361, 1 February 2016, Pages 194–207
Open Access
Highlights
- •
- Context and definitions for forest recovery following stand replacing disturbance.
- •
- Quantitative analysis of literature determines post-disturbance recovery trends.
- •
- On average, benchmarks for canopy cover and height were reached within 5–10 years.
- •
- Recovery rates varied by ecozone and disturbance type (wildfire, harvest).
- •
- Insights on time series of remotely sensed data for recovery characterization.
Abstract
The
recovery of forests following stand-replacing disturbance is of
widespread interest; however, there is both a lack of definitional
clarity for the term “recovery” and a dearth of empirical data on the
rates of forest recovery associated with different disturbance types. We
conducted a quantitative review of literature to determine recovery
times following wildfire and timber harvest and to evaluate variation in
recovery rates across Canada’s diverse forest ecosystems. Recovery was
assessed according to the rate of change associated with certain forest
structural attributes that have traditionally been used as indicators of
forest growth and productivity. The recovery of forest canopy cover,
tree height, and stand basal area varied at rates that depended on
disturbance type, forest biome, and ecozone. We found that, on average,
it took 5–10 years, depending on factors such as location and species,
for most forest ecosystems of Canada to attain a benchmark canopy cover
of 10% after wildfire or harvest. Similarly, regenerating stands in
Canada’s boreal forests were capable of attaining average heights of 5 m
within five to ten years after wildfire or harvest. Stands in the
Boreal Plains ecozone post-harvest reached stand basal area, benchmarked
at 10 m2 ha−1, faster than those in the Boreal
Shield, attributable to differences in tree species composition and the
rich mineral deposits of the Boreal Plains. Overall, recovery of canopy
cover, tree height, and stand basal area was similar or more rapid
following wildfire than harvest. Our review provides temporal benchmarks
for gauging recovery times after disturbance. Building upon these
temporal benchmarks, and conditioned by disturbance type, site
conditions, and location, we present opportunities for using dense time
series of remotely sensed data to inform on regional and national trends
in forest recovery following disturbance.
Keywords
- Forest;
- Disturbance;
- Recovery;
- Regeneration;
- Boreal;
- Canada;
- Remote sensing;
- Landsat
1. Introduction
Canada’s
forests are recognized globally for the important ecosystem services
that they provide; however, these forests are dynamic in nature and
subject to a wide range of natural and anthropogenic disturbances that
vary in severity, extent, and frequency (Bergeron et al., 2001, Stocks et al., 2002, Boucher et al., 2009 and Brandt et al., 2013). Disturbances such as wildfire and timber harvesting can directly impact forest structure and composition (Lavoie and Sirois, 1998, Brassard et al., 2008 and Fleming et al., 2014), and indirectly impact soil properties (Certini, 2005), thereby altering ecosystem productivity and function. The ongoing nature of disturbances to Canada’s forests (Brandt et al., 2013), combined with uncertainty related to climate change (Price et al., 2013), necessitates an improved understanding of forest dynamics and increasingly sophisticated and flexible management practices (Bergeron et al., 2004 and Burton et al., 2006).
Despite advances in knowledge and management practices, uncertainty
remains in the rates of forest recovery associated with different
disturbance types across the range of forest ecosystem conditions in
Canada (Sturtevant et al., 2014).
Forest recovery can also be understood from different perspectives, for
instance silvicultural and ecological, with different assessment
criteria and definitions present as a result.
Disturbances
are relatively discrete events that disrupt the forest ecosystem and
cause a change in the physical structure of vegetation, soil substrate,
and resource availability (White and Pickett, 1985 and Clark, 1990).
Although disturbed forests will recover if left long enough, of
interest is the extent and rate at which forests will return to
pre-disturbance condition. Early studies of post-disturbance recovery in
Canada’s forests have largely focused on general descriptions of
successional sequence (Black and Bliss, 1978 and Bergeron and Dubuc, 1989), recovery of net primary production (Amiro et al., 2000), and advance regeneration (Gradowski et al., 2010, Spence and MacLean, 2012 and Veilleux-Nolin and Payette, 2012). While these and other studies (Johnson, 1996, Greene et al., 1999 and Chen and Popadiouk, 2002)
have contributed to an increased understanding of post-disturbance
stand dynamics, there remains a paucity of quantitative information and
synthesis on rates of forest regrowth and the factors that influence the
forest recovery process.
The
nature and rate of forest recovery may depend on several factors
relating to the nature and severity of disturbance, presence of
biological legacies, and inherent productivity of the site (Johnson, 1996, Franklin et al., 2002, Chen et al., 2009 and Ilisson and Chen, 2009a). Different disturbances contrast markedly in terms of biological legacies (Franklin et al., 2007), and forests faced with repeated perturbations tend to be less resilient (Payette and Delwaide, 2003).
Rates of forest change following disturbances may ultimately depend on
multiple interacting factors, such as disturbance history,
pre-disturbance stand conditions, local site factors, regional species
pool, and species life histories, among others (Foster et al., 1998, Harper et al., 2005, Mansuy et al., 2012 and Girard et al., 2014).
However, it is unclear how these factors interact to explain variation
in rates of forest recovery. Some initial efforts have been made to
better understand regional level variability of forest recovery across
Canada’s forested ecosystems (Goetz et al., 2006 and Mansuy et al., 2012).
A better understanding of forest recovery rates and patterns in
different environmental and climatic conditions is necessary to
understand the overall dynamics of Canada’s forests and devise effective
strategies for sustainable forest management.
Among
the challenges encountered in characterizing rates of forest recovery
is the absence of a universal definition of what is meant by the term
recovery in a forest context. Because recovery involves the return of
vegetation cover, terminologies such as “revegetation,” “regeneration,”
and “regrowth” are often used, sometimes synonymously, to describe what
happens to forests following disturbance. Some consider recovery as the
reestablishment or redevelopment of forest biomass and canopy structure
characteristics after the impact of a particular disturbance (Frolking et al., 2009).
However, it is not entirely clear at what stage or condition a forest
that has experienced disturbance can be described as returning to its
function as a forest. In the context of these broader interpretations of
forest recovery, for the purposes of this study, we are interested in
the re-establishment and regeneration of vegetation at a site following a
stand-replacing disturbance, specifically wildfire and timber harvest.
The
Food and Agriculture Organization (FAO) defines a forest as an area of
land greater than 0.5 ha in size with greater than 10% tree canopy
cover, and trees that are capable of reaching a minimum height of 5 m (FAO, 2010).
This includes young stands or temporarily unstocked areas that have not
yet—but are expected to reach—a crown density of 10% and a tree height
of 5 m (FAO, 2010).
According to this definition of forest, it is possible to ascertain
from early indicators whether a disturbed forest has recovered or is
headed toward recovery. Therefore, the term recovery describes a
long-term process, whose endpoint ultimately depends on one’s interest
or point of view (i.e., ecological, economic). In the context of this
review, we consider a site to be regenerating or recovering if
vegetation is reoccupying a site, if trees capable of reaching a certain
height are re-establishing, and if there exists the potential of the
trees to reach a given canopy cover.
Disturbance processes are increasingly well understood and systematically captured through remote sensing approaches (Frolking et al., 2009).
The capacity of remotely sensed data to characterize vegetation
recovery post-disturbance is increasing with the widespread availability
of data and methods that enable dense time series analyses (Kennedy et al., 2014). Information on forest recovery is of interest from forest management, ecosystem services, and climate change perspectives (Anderson-Teixeira et al., 2013).
While plot-based studies focused on the site-specific return of
vegetation following disturbance have informed the forest management and
ecological understanding of forest recovery (e.g., Drever et al., 2006),
there is a need to bridge between the contexts offered by plot-based
measurements and associated knowledge with emerging opportunities for
large-area characterizations. Our objective in undertaking this review
was to evaluate trends, derived from plot-based studies, in rates of
forest regeneration following stand replacing disturbances, specifically
wildfire and harvest, across the forested ecosystems of Canada. We
focused specifically on quantifying the number of years required for
forests to reach the benchmark values specified in the FAO definition of
forests stated above, and the variability in regeneration among the
represented ecological regions of Canada. We close with an outlook on
opportunities to map and characterize recovery with time-series remotely
sensed data.
