Recruitment or Regeneration Studies

We have undertaken several studies to better understand the processes controlling recruitment in northern forests. We have found that the key factors influencing seedling recruitment after disturbance in northern temperate, sub-boreal and boreal forests are the availability and favourability of seedbed substrates, the abundance and proximity of parent trees, and canopy structure. Unfavourable seedbed substrates can over ride all other factors. 

To develop regeneration behaviours for use in the SORTIE-ND model we have used the inverse modeling approach pioneered by Ribbens et al. (1994). Maximum likelihood methods and inverse modeling have proven effective for determining which processes control tree establishment in forests. We review these methods and the pros and cons of three commonly used empirical seed/seedling dispersal functions for trees in Greene et al. (2004), see below.  

Our core recruitment studies are LePage et al. (2000) for northern temperate interior cedar hemlock forests and Astrup et al. (2008) for mountain pine beetle impacted sub-boreal spruce forests, see below. The other studies summarized provide context for understanding the processes controlling regeneration dynamics in northern forests. 

Evaluation of stand dynamics in old mountain pine beetle attacked stands
Coates, K.D., Astrup, A., Amoroso, A.M., Burleigh, J. 2008.

This Forest Science Program funded study (M085196) of regeneration and growth dynamics after mountain pine beetle (MPB) attack was undertaken in 2007 in the Flathead area which was impacted by the MPB between 1978 and 1982. A paper is in preparation. The regeneration component of the study was designed to complement the Astrup et al. 2008 study of recruitment dynamics, see below, where we were able to examine regeneration success up to 10 years after attack in the current epidemic hitting northern BC. The growth component of the study was to gain a better understanding of the growth potential of secondary structure (see Coates et al. 2006) that survives a MPB disturbance.  

The extent of MPB mortality, regeneration delay, release of advance regeneration and other surviving residual trees (secondary structure) and post-beetle stand growth was determined. Five plots were established in each of 22 stands representing a range of MPB attack intensity. Pre-attack basal area varied between 29 and 58 m2 ha-1. The percent of basal area killed by beetles varied from 42 to 100% with most stands between 60 and 80%. Six stands exceeded 80% basal area mortality and 3 stands were below 60% mortality. In 2007, twenty-five to 30 years after attack basal area varied between 4 and 51 m2 ha-1. Five stands, ranging in mortality at time of attack from 51 to 79%, had fully recovered their pre-attack stand basal area. Four stands, ranging in mortality at time of attack from 68 to 100%, had poor basal area recovery. Growth release on surviving trees exhibited species and size variability. Release of surviving lodgepole pine trees was often dramatic. Recruitment of new regeneration post-beetle attack was often, but not always, delayed by 5-10 years. Densities of post-beetle regeneration were often high in 2007. Based on age of understory trees in 2007, there was little advance regeneration in these forests at the time of the beetle attack.   

Recruitment limitation in forests:  lessons from an unprecedented mountain pine beetle epidemic
Astrup R., K.D. Coates, E. Hall. 2008.

The processes controlling recruitment from seed in post-MPB disturbed stands were explored using an inverse-modeling approach. Specifically, the influences of seed-source availability, substrate type, overstory structure and time since MPB attack on post-MPB recruitment were examined. Because the presence of a seedling bank may be important in the regeneration success of these forests, we also examine the relationship between local seed-source availability and overstory structure on the abundance and pattern of the existing seedling bank.  We found recruitment up to 10 years post-MPB attack was sparse. Subalpine fir comprised the majority of the post-MPB recruitment. It increased with local parent tree basal area and increased strongly with proximity to a major seed source. This resulted in a patchy distribution for subalpine fir post-MPB regeneration. Lodgepole pine post-MPB recruitment was limited by overstory shading. Recruitment of pine decreased as the total overstory basal area increased. Interior spruce post-MPB recruitment was similarly limited by total overstory basal area.

Seedbed substrates were uniform and dominated by moss. Substrate type distribution changed little with time since MPB disturbance (up to 10 years). The overall low post-MPB recruitment observed was likely due to a lack of disturbance to the moss-dominated forest floor. Moss is known to be a poor substrate in northern forests (see LePage et al. 2000; Wright et al. 1998). The distribution of the seedling bank showed the same trends as the post-MPB regeneration. Post-MPB seedling recruitment dynamics of these forests was not substantially changed from conditions prior to MPB disturbance. There was no pulse of regeneration up to 10 years post-MPB disturbance. Unless this changes, future stand structure will be dominated by the seedling bank established prior to the MPB epidemic. Subalpine fir dominated the seedling bank (68%) and post-MPB recruitment (94%). This suggests that MPB-disturbed forests are undergoing a substantial shift in landscape-level species composition.

An evaluation of alternative dispersal functions for trees
Greene, D.F., C.D. Canham, K.D. Coates, P.T. LePage. 2004.

This study examined the pros and cons of three commonly used empirical seed/seedling dispersal functions for trees (lognormal, 2Dt, and two-parameter Weibull). As part of this analysis, the recruitment data from Date Creek (LePage et al. 2000) was re-analyized. The three functions were also compared by the analysis of other published studies where the location of the source was known, and by new inverse modeling within an old growth hardwood forest in southern Quebec. Almost all the species examined in the published studies and new study were wind-dispersed. For discrete source studies, the lognormal was clearly superior, while for the inverse modeling the performance of the three dispersal functions was somewhat more even. Collisions with boles may have spuriously enhanced the likelihood of the 2Dt and the Weibull with inverse modeling, as both these functions assume that the greatest seed/seedling density will occur at the base of the maternal parent bole. In general, the lognormal function was preferred because, as well as providing a framework for mechanistic interpretation, it tends to provide a closer approximation to observed dispersal curves. The analysis also suggests that the mean distances travelled by seed crops are far more extensive than indicated by previous studies that used the Weibull function.

Tree recruitment in gaps of various size, clearcuts and undisturbed mixed forest of interior British Columbia, Canada.
Coates, K.D. 2002.

This paper presents results from the broad natural regeneration studies at the Date Creek Experiment. Seedling recruitment success was a function of mature tree canopy cover, gap size and position in a gap. Recruitment was abundant within canopy gaps across a wide range of gap sizes (20 to 5000 m2), but recruit numbers dropped off rapidly under the closed forest canopy and in the open conditions of clearcuts. Inside canopy gaps, recruitment was similar by gap position in small gaps (<300 m2) but exhibited a trend of increasing density from the sunny north to shady south end of larger gaps. This was true for all tree species regardless of their shade tolerance ranking. There was no evidence of gap partitioning by any of the tree species during the regeneration phase, suggesting that adaptation to the subtleties of gap size during early recruitment are not well developed in these tree species. Favourable locations for emergence and early establishment of germinants were less favourable for growth and survival of established seedlings, that is, the regeneration niches in these forests were discordant. Tree abundance and species diversity appears to be controlled more by differentiation among growth and survival niches than by the regeneration niches. From the perspective of forest management, abundant natural regeneration of all the dominant tree species of these mixed-species forests can be obtained after partial cutting.

Seed abundance versus substrate limitation of seedling recruitment in northern temperate forests of British Columbia
LePage, P.T., C.D. Canham, K.D. Coates, P. Bartemucci. 2000.

This is the recruitment study undertaken in the interior cedar hemlock forests (ICH) at Date Creek as part of our overall SORTIE-related studies. This study builds on the work of Ribbens et al. (1994) but expanded the approach to include the influences of seedbed substrate and overstory canopy conditions. This study uses model selection to understsnd the influence of (i) the spatial distribution and abundance of parent trees (as seed sources), and, (ii) the abundance and favourability of seedbed substrates, on seedling recruitment under four levels of canopy openness (full canopy, partial canopy, large gap, and clearcut). Under undisturbed canopies, the lack of suitable seedbed substrate appears to be the immediate factor limiting seedling recruitment. The forest floor disturbance associated with partial cutting created a diversity of favourable substrates for seedling establishment. Effective seedling dispersion distances, however, are still relatively limited in partially cut stands; therefore, a good distribution of seed trees will be required to ensure a mix of species in natural seedling regeneration.

Cutting more trees to form distinct gaps removed physical barriers to seed dispersal from those trees bordering the gap, resulting in higher effective seedling dispersion distances but, at the same time, slightly reducing the favourability of seedbeds for establishment of individual seedlings. Clear-cutting appears to shift recruitment control away from seed availability and substrate favourability and brings microclimate effects into play. Sampling in clearcuts was done relatively close to the clearcut edges (<100 m); however, despite this close proximity to a seed  source and a diverse mix of substrates, seedling densities were extremely low. A silvicultural system that combines partial harvesting with the creation of large (0.1–0.5 ha) gaps appears to hold the greatest promise for maintaining the diversity of both seedling and canopy recruitment in these forests.

Regeneration from seed of six tree species in the interior cedar???????hemlock forests of British Columbia as affected by substrate and canopy gap position.
Wright, E.F., K.D. Coates, P. Bartemucci. 1998.

This paper also presents results from the broad natural regeneration studies at the Date Creek Experiment. After establishing the Date Creek study in 1992 and observing natural regeneration success for a few years it became clear to us that seedbed substrate played a major role in seedling establishment. This experiment was specifically designed to provide better information on regeneration success of six tree species when seeded onto two substrates (organic and undisturbed moss) in three positions along a north–south gradient within eight replicate 600-m2 canopy gaps and in the intact forest south of each gap. Gap positions were monitored in 1995 and 1996.  The position a seed falls inside or outside a canopy gap and the substrate on which it lands profoundly influenced emergence success, subsequent establishment, and early survival of the major tree species in northwestern British Columbia. Successful regeneration from seed increased with forest floor disturbance (organic seedbed) and soil moisture and decreased as a function of increasing light availability. The only exception to these trends was subalpine fir, which showed little preference for seedbed substrate, although germination was enhanced by increased shade.