Burn severity and heterogeneity mediate avian response to wildfire in a hemiboreal forest

https://doi.org/10.1016/j.foreco.2019.02.043Get rights and content

Highlights

  • Bird community patterns were best described by variations in burn severity.

  • High-severity burns had reduced abundance and richness of birds following fire.

  • Fire responses by most species were mediated by burn severity.

  • Burn heterogeneity and burn status were important for some species’ responses.

  • Results were similar when counts were either adjusted for detectability or not.

Abstract

Recent studies have shown the importance of accounting for burn severity when assessing the effects of forest fires on avian communities. We add to this growing literature with one of the first studies to assess these effects in boreal and hemiboreal regions of North America. We conducted point counts in control and treatment areas for 2–3 years before (2009–2011) and 4–5 years after (2012–2016) the Pagami Creek Fire in northern Minnesota, USA. Our primary objectives were to (1) assess the effects of burn status, burn severity and burn heterogeneity on avian communities and patterns of abundance in individual bird species and (2) compare these results when counts were either restricted to 100 m count radii or corrected for potential detection heterogeneity. Species richness and total abundance of birds were best described by a burn severity by year interaction, with lower-severity burns having similar patterns to control areas and higher-severity burns showing reduced abundance and richness following the fire. Nevertheless, values associated with higher-severity burns returned to near-control levels after five years. Of 43 species analyzed, 27 showed detectable responses to the fire through shifts in abundance. Nearly equal proportions of these species had generally positive versus negative effects of the fire. Sixteen (59%) of these fire responses were mediated by burn severity, while four (15%) and seven (26%) were best predicted by burn heterogeneity or burn status, respectively. Of the species whose response to the fire was mediated by burn severity, the majority (81%) had decreased abundance as burn severity increased. However, all species responding to burn heterogeneity or burn status showed positive relationships to these metrics. Results were similar between analytical methods that adjusted for species detectability vs those that did not, with 78% of species having the same interpretation of fire effects. Burn severity and heterogeneity, in addition to the sole distinction of whether an area was burned, dictate the response of birds to forest fire in hemiboreal forests. Mixed-severity fires, which include a range of burn severities and associated vegetation change, will likely benefit the most bird species in this region. When possible, management should focus on retaining this spectrum of post-fire conditions, though of primary importance is the inclusion of wildfire in forest planning and conservation efforts.

Introduction

A recent paradigm in fire ecology identifies burn severity and burn heterogeneity as dominant drivers mediating the effects of forest fires on birds and bird communities (Fontaine and Kennedy, 2012, Hutto and Patterson, 2016, Taillie et al., 2018, Tingley et al., 2016). Studies have generally found that bird species respond uniquely to burn severity levels as opposed to simply the presence of fire. Once severity is accounted for, the response of many bird species becomes clear and often a positive relationship with fire is detected (Smucker et al., 2005). This is especially true when post-fire time series are long and the time-since-fire is taken into account (Hutto and Patterson, 2016). Based on these results and those relating to other ecosystem processes, scientists are recommending the use of severe fires and associated mixed-severity impacts as a means of maintaining biodiversity (Bond et al., 2012, DellaSala et al., 2017, Hutto et al., 2015, Kelly et al., 2017).

Mixed-severity fires create a patchwork of vegetation, comprising areas that are unburned, lightly burned and areas where complete stand turnover has occurred (DellaSala and Hanson, 2015). This landscape heterogeneity provides habitat for bird species exhibiting a range of responses to fire (Hutto et al., 2015, Hutto and Patterson, 2016). Benefits extend not only to archetypal species like Black-backed Woodpecker (Picoides arcticus; Tremblay et al., 2016), but also species that utilize residual live trees and edges like Olive-sided Flycatcher (Contopus cooperi; Baker et al., 2016) and species showing delayed responses to post-fire succession such as shrub nesters (Stephens et al., 2015).

To date, studies assessing the effects of burn severity and heterogeneity on birds have primarily been conducted in western forests (Hutto and Patterson, 2016, Kotliar et al., 2007, Smucker et al., 2005, Stephens et al., 2015, Tingley et al., 2016) as well as Appalachian forests (Greenberg et al., 2013, Rose et al., 2016, Rush et al., 2012). Although boreal and hemiboreal forests have received much attention regarding effects of fire on birds (e.g., Haney et al., 2008, Hannon and Drapeau, 2005, Niemi, 1978, Schieck and Song, 2006, Schulte and Niemi, 1998), the effects of burn severity and heterogeneity have rarely been tested across the bird community (but see Azeria et al., 2011, Knaggs, 2018).

In this study, we compare the effects of fire, burn severity and burn heterogeneity on bird species and community patterns before and after the Pagami Creek Fire (2011) in the hemiboreal forests of northern MN, USA. We used a before-after control-impact study design with point counts collected for 2–3 years pre-fire and 4–5 years post-fire at unburned controls and across a continuum of burn severities. We predicted that most individual bird species would show responses to the fire, with this effect generally depending on burn severity or burn heterogeneity, as opposed to solely the presence of fire disturbance. We expected community patterns in the highest burn severities to be most divergent from unburned controls. In addition, we compare the effects of the above parameters when bird data have been corrected with density estimates that account for detectability (e.g., Stephens et al., 2015) versus raw point-count data restricted to 100 m from observers (e.g., Smucker et al., 2005).

Section snippets

Study area

The study was conducted near the southern edge of the boreal forest in the Superior National Forest (SNF; 48°N 92°W) of northeastern Minnesota (Fig. 1). The SNF is approximately 1.6 million ha and is primarily managed for multiple uses such as timber harvest and recreation. However, it also contains a 440,000 ha wilderness area, the Boundary Waters Canoe Area Wilderness, where motorized travel is limited and timber harvesting is prohibited. Wildfire was historically the dominant disturbance

Fire effects on bird community

In total, we had 11,284 bird detections of 98 species on 655 counts from 2009 to 2016 (Appendix A). On a per-point-count level, both species richness and total abundance of birds were best predicted by the severity interaction model (S*; Table 3). The variation in the numbers of species and individual birds detected at counts varied by burn severity, with lower severity locations having similar counts to controls and higher severity locations showing lower richness and abundances (Fig. 2). This

Discussion

Wildfire has long been known to shape the ecological communities of boreal and hemiboreal forests (Heinselman, 1973, Pastor et al., 1996). Yet, the positive effects of mixed-severity fires on biodiversity, and especially avian communities, have only more recently come into focus (Hutto et al., 2015). The varying and positive effects of mixed-severity fire on bird communities have been observed in a variety of forested regions including the Pacific Northwest and Interior west (Smucker et al.,

Conclusions

We found about two thirds of bird species analyzed responded to the fire with changes in abundance. Half of these increased in burned areas, which is similar to many forested regions of North America (Kotliar et al., 2007, Smucker et al., 2005, Stephens et al., 2015, Taillie et al., 2018), though somewhat lower than others (Hutto and Patterson, 2016, Knaggs, 2018, Rose and Simons, 2016). Likewise, we also found that many of the species-specific and bird-community responses to the fire were

Acknowledgements

The authors thank Josh Bednar, Carly Lapin and Steve Kolbe for assistance with field work. Jan Green was integral to the formation of this research project. Richard Hutto, Wayne Thogmartin, Matt Etterson, Nathan Pollesch, Alexis Grinde, Katie Zlonis and two anonymous reviewers provided valuable suggestions that greatly improved the manuscript. Kim Rewinkel provided editorial support.

Funding

This work was supported by the USDA Forest Service, Northern Research Station, Rhinelander, WI. Contributions by B. Sturtevant were supported by the National Fire Plan.

References (58)

  • Bates, D., Mächler, M., Bolker, B., Walker, S., 2014. Fitting Linear Mixed-Effects Models using lme4. ArXiv14065823...
  • M.L. Bond et al.

    A new forest fire paradigm: the need for high-severity fires

    Wildl. Prof.

    (2012)
  • S.T. Buckland et al.

    Estimating bird abundance: making methods work

    Bird Conserv. Int.

    (2008)
  • K.P. Burnham et al.

    Model Selection and Multimodel Inference: A practical Information-theoretic Approach

    (2002)
  • R. Campos-Ruiz et al.

    Temporal patterns of wildfire activity in areas of contrasting human influence in the canadian boreal forest

    Forests

    (2018)
  • D.A. DellaSala et al.

    The Ecological Importance of Mixed-Severity Fires: Nature’s Phoenix

    (2015)
  • D.A. DellaSala et al.

    Accommodating mixed-severity fire to restore and maintain ecosystem integrity with a focus on the sierra Nevada of California, USA

    Fire Ecol.

    (2017)
  • M.A. Etterson et al.

    Estimating the effects of detection heterogeneity and overdispersion on trends estimated from avian point counts

    Ecol. Appl.

    (2009)
  • J.B. Fontaine et al.

    Meta-analysis of avian and small-mammal response to fire severity and fire surrogate treatments in US fire-prone forests

    Ecol. Appl.

    (2012)
  • L.E. Frelich et al.

    Spatial Patterns and Succession in a Minnesota Southern-Boreal Forest

    Ecol. Monogr.

    (1995)
  • A. Haney et al.

    Thirty years of post-fire succession in a southern boreal forest bird community

    Am. Midl. Nat.

    (2008)
  • S.J. Hannon et al.

    Bird responses to burning and logging in the boreal forest of Canada

    Stud. Avian Biol.

    (2005)
  • J.M. Hanowski et al.

    A comparison of on- and off-road bird counts: do you need to go off road to count birds accurately? (Una Comparación de Conteos Dentro-de y Fuera-de Caminos: ¿Hay que Alejarse de los Caminos Para Contar Aves con Exactitud?)

    J. Field Ornithol.

    (1995)
  • M.C. Hansen et al.

    High-resolution global maps of 21st-century forest cover change

    Science

    (2013)
  • M.L. Heinselman

    The Boundary Waters Wilderness Ecosystem

    (1996)
  • R.L. Hutto

    Should scientists be required to use a model-based solution to adjust for possible distance-based detectability bias?

    Ecol. Appl.

    (2016)
  • Hutto, R.L., Bond, M.L., DellaSala, D.A., 2015. Using bird ecology to learn about the benefits of severe fire. In: The...
  • R.L. Hutto et al.

    Positive effects of fire on birds may appear only under narrow combinations of fire severity and time-since-fire

    Int. J. Wildland Fire

    (2016)
  • Jain, T.B., Graham, R.T., 2007. The relation between tree burn severity and forest structure in the Rocky Mountains....
  • Cited by (7)

    • Influence of lake levels on water extent, interspersion, and marsh birds in Great Lakes coastal wetlands

      2021, Journal of Great Lakes Research
      Citation Excerpt :

      We used the aictab function to compare 39 combinations of these variables, including the full model (all 5 variables plus the lake × lake level interaction), and models with 1–4 of the variables. We assumed that variation in species-specific detection was random with respect to the predictor variables used in our models (Hutto, 2016; Johnson, 2008; Zlonis et al., 2019); therefore, we included no adjustment for bird detectability among lakes, lake-level year, and the GIS wetland attributes (water extent and interspersion). We calculated model-predicted values ± 95% confidence intervals for our mixed effects models using bootstrapping with the “ciTools” package in R; in doing so, we ignored random site effects to obtain population level predictions (i.e., includeRanef = FALSE; Hamen and Avery 2020).

    View all citing articles on Scopus
    View full text