Denudation, restoration and forest succession: a view from the North Cascades

What happens when a forest ecosystem is removed very thoroughly, but in small pieces? The surrounding matrix is left relatively in tact, but small plots are reduced to bare mineral soil. All vegetation is removed from within these small areas as the is organic soil and litter that harbors seeds, soil fungi and microbes that facilitate growth. Now imagine that there are abundant overstory trees on the edges of this disturbance. I don't mean logging a couple hundred acres and leaving thousands around them, but more like small house foundations scraped out of the soil and rock with trees all around.

Now imagine that you're tasked with restoring such an ecosystem. What environmental and biological limitations may exist that filter out establishing vegetation? Are there sufficient propagules to revegetate the area? Can the native seed rain be enhanced by active management? Is there a low-intensity, cost-effective approach to this active management?

Rodney Pond, restoration ecologist, native Floridian, swell guy and a giant among the alder.

Friend and former UW colleague Rodney Pond and I have spent significant time working with one another over the last couple years, including stints in the Center for Urban Horticulture greenhouse, Union Bay Natural Area, Stillaguamish Watershed, and most recently within North Cascades National Park (NOCA). In a scenario exactly like the ambiguous micro-disturbance situation described above within NOCA, Rodney undertook a pair of forest restoration projects, and the resulting systems have proceeded quite nicely over time. Rodney, my partner and collaborator Lexine Long, friend Dan Owen and I have been making trips to follow up on the data and see if the treatments have changed the forest's successional trajectory since initial measurements six years ago.

Today, we're going to focus on preliminary six-year data from a campsite retirement that coincided with the construction of what is now the North Cascades Institute's Environmental Learning Center (ELC). The site was initially a fishing camp that included several small cabins and their associated footprints (e.g. trails, decks, parking spots, etc). It was nestled in and around a relatively in tact forest overstory in some places, comprised of both early successional species and a few later successional ones as well. When the ELC was constructed, the structures were leveled and the bare mineral substrate beneath the cabins, decks and pathways was left exposed. The predominant biological legacy at the site was an early-successional, mixed hardwood-conifer forest comprised of Douglas fir, Douglas maple, Big-leaf, lodgepole pine, paper birch and alder.

Restoration sites relative to the Environmental Learning Center at NOCA.

In addition to the visible campsite scars that stood out within the National park landscape, it was thought that very little native vegetation would grow on such a heavily disturbed site with such low quality soil. Restoration was thought to be required to increase the substrate quality and local availability of propagules. Rodney set out to test the effectiveness of two low-intensity treatments at establishing native forest species. He hoped to observe differences in the pioneer tree assemblages over time and elucidate the subtle differences in forest succession over time.
Alder heaven in the mineral substrate plots!

We used three treatments to push the forest community development forward:

Mineral soil: existing compacted soils, devoid of organic matter.

Mineral soil + forest litter: locally collected organic layer soil; mostly leaf/needle litter and twigs with some lichens, mosses, was raked into the mineral soil. This treatment was intended to change soil texture to a more amenable texture for seed germination/growth and introduce additional propagules to the site.

Mineral soil + wood chips: Wood chips comprised of leftover wood waste from the ELC construction. This treatment was intended to provide an insulating organic layer that retained soil moisture late into the growing season, and maintain a strong carbon source for mycorrhizal colonization.

2011 follow-up vegetation data was collected in one-square-meter plots using a completely randomized block design. We measured absolute abundance of all woody plant species, counting all stems and measuring their heights in August 2010. We used PERMANOVA to test the hypothesis that the multivariate groups differed between treatments. We included block in the model, to correct for any siting differences in propagule rain. We also used non-metric multidimensional scaling to examine how the plots differed in 3-d space and indicator species analysis to see what species drove the vegetation assemblages. For the sake of this discussion, we'll stop the fine print here and get to the fun part, the results.

We found that soil treatment did significantly drive the vegetation assemblages within each treatment (F=1.88, R2=0.083, P=0.034 using 10k permutations). NMDS showed this pattern clearly (Stress = 9.167, P = 0.003, 10k permutations) in a 3-dimensional solution.
NMDS ouput for the first two axes of the three-dimensional solution. Blue vectors are significant species within the ordination at P <0.01>

Indicator species analysis (ISA) returned four species:

Mineral substrate: Alnus rubra (Red alder)

Mineral and litter: Tsuga heterophylla (western redcedar) and Thuja plicata (western hemlock)

Mineral and wood chips: Pseudotsuga menziesii (Douglas fir)

So, what does all of this mean? Is there any ecological explanation for the observed differences between the treatments? Given the biology of the indicator species, I say yes, there almost certainly is. Alder is a wind-dispersed species that showed up both within the NMDS ordination and the ISA, and was most commonly correlated to no treatment. Alder is a well-known pioneer species that colonizes after disturbance, and it appears it filled that ecological role very well when soils were not amended. Similarly, hemlock and cedar have wind dispersed cones that accumulate well in forest duff, germinating when appropriate light and moisture conditions occur. We should also note that cedar and hemlock were absent from the adjacent forest stands and that the litter treatment was the likely source of introduction; this was also corroborated by greenhouse germination data. Douglas fir, a drought-tolerant, early-successional conifer was the driving species behind the mulched plots, which may be a function of the soil texture left by the mulching treatment. Douglas fir cones are relatively large, and the initial seedlings are somewhat drought tolerant relative to the other adjacent overstory species (paper birch, alder, etc).

So the results make sense, and the hypothesis that low-intensity soil treatments can affect the developmental trajectory of a denuded forest site is supported. What does this mean for restoration practice though? How does the experiment at hand inform the restoration of denuded forest patches?

  • Using mere forest litter without planting or seeding can potentially shift the forest assemblage towards conifer species.
  • Unmanipulated, disturbed soils will attract highly competitive species such as alder and exclude species unable to cope with compacted soils.
  • Using relatively unskilled labor to gather readily available soil litter may be a fast, cost-effective way to restore small patches such as retired campsites, parking lots, etc when the surrounding matrix is in good ecological health.
  • An additional recommendation: Because this study took place in a relatively pristine portion of a national park, weeds were not an issue. In the "real world," weed infestations may limit forest species colonization.
*Please note that this is preliminary data and that additional analyses are being performed in support of the project. Hopefully these results will be coming to an applied ecological restoration journal near you in 2011. If you have any questions, please get in touch

Greatest Hits