High-resolution pollen as an indicator of fire ... - fs.fed.us

High-resolution pollen as an indicator of fire severity during the Populus Period, 2000 - 4000 cal yr BP Vachel Carter, Andrea Brunelle, Simon Brewer, Thomas Minckley

Department of Geography

Abstract

This study attempts to analyze fire severity using lake sediments from southeastern Wyoming, during a unique period of time coined the ‘Populus period’ (Carter et al., 2013). The Populus period (3800-2800 cal yr BP) was a time when vegetation composition and fire regimes changed from a Lodgepole pine dominated ecosystem to Lodgepole pine/Quaking Aspen mixed ecosystem. This study investigates fire events between 2000-4000 cal yr BP to determine the ecological response associated with fire events and to identify top-down or bottom-up drivers. In order to determine fire severity, this study compares high-resolution charcoal and pollen data to peak magnitude, which is an output from a statistical treatment program called CharAnalysis. Linear discriminant analysis (LDA) was used to set a threshold by which individual pollen taxa and grouped pollen taxa are either associated with low or high severity fires. Based on the LDA results, the severity threshold for individual pollen taxa are fire events with peak magnitudes greater than 100 particles/cm2/episode. The LDA results indicate severity thresholds of grouped pollen taxa ranging between 100-200 particles/cm2/episode. Superposed epoch analysis (SEA) is used to model pollen behavior both pre and post fire events to determine the ecological response associated with each of the fire events. Statistical analysis using LDA and SEA can potentially be used in combination to determine fire severity, which will be beneficial to land managers and policy makers in the 21st century.

Acknowledgements Funding was made possible by Rockies Express Pipeline LLC, Great Basin Earth

Science, Global Change and Sustainability Center, the Don Currey Scholarship, the David C. Williams Scholarship and the U.S. Forest Service.

Citations •  Blarquez, O., and Carcaillet, C., 2010. Fire, fuel composition and resilience threshold in subalpine ecosystem. PLoS ONE

5(8), e12480 •  Carter, V.A., Brunelle, A., Minckley, T.A., Dennison,P.E., Power, M.J., 2013. Regionalization of fire regimes in the Central

Rocky Mountains, USA. Quaternary Research 80, 406-416. •  Minckley, T.A., and Shriver, R.K., 2011. Vegetation responses to changing fire regimes in a Rocky Mountain forest. Fire

Ecology 7(2), 66-80. •  Shriver, R.K., and Minckley, T.A., 2013. Late-Holocene response of limber pine (Pinus flexilis) forests to fire disturbance in

the Pine Forest Range, Nevada, USA. Quaternary Research 78, 465-473.

Research Objectives

.

Previous Research

Preliminary Results

1. Quantify the changes in pollen abundance after 12 fire events that occurred between 2000-4000 cal yr BP to determine the severity of each fire event.

2. Determine the potential drivers of ecological change between 2000-4000 cal yr BP.

percentages of Poaceae (10%) and Amaranthaceae (~6.5%) pollen.Cyperaceae pollen percentages were low (3%). Initial forest constitu-ents include Abies bifolia, P. contorta-type and P. flexilis-type based onmacrofossils.

CharcoalFour fire episodes were identified with a mean fire return interval

(FRI) of 131 yr (Fig. 4). Peak magnitudes were variable between 0.1and 58.5 and averaged 18.0 particles/cm2/episode.

LL-II: the early Holocene (depth 385–258 cm, 11,800–9400 cal yr BP)

PollenPollen percentages for subalpine species, Abies (5%) and Picea (7%)

increased during LL-II. Pinus pollen percentages also increased to 50%,with Pinus flexilis-type and Pinus contorta-type pollen abundances aver-aging 5% and 10% respectively. Shrub and herbaceous pollen abun-dances decreased in this zone with Artemisia pollen percentagesdecreasing (18%), along with Cupressaceae (2%), Poaceae (4%), andAsteraceae (2%). Cyperaceae pollen percentages increased to 4%. Abiesbifolia, Picea engelmannii and P. flexilis-type needles were present.

CharcoalLL-II had 13 fire episodes with amean FRI of 246 yr. Peakmagnitudes

ranged between 0.3 and 144.0 and averaged 48.0 particles/cm2/episode.

LL-III: the middle Holocene (depth 258–160 cm, 9400–4000 cal yr BP)

PollenTotal Pinus pollen was the dominant pollen type during LL-III, aver-

aging 68%. Abies pollen percentages averaged a peak of 4% around9000 cal yr BP. Pinus flexilis-type (1%) and Pinus contorta-type (6%) pol-len percentages decreased in this zone. Picea pollen was initially

present (4%) but decreased to trace levels through the zone.Cupressaceae (3%) pollen percentages also decreased. Populus pollenincreased (1%) after 9000 cal yr BP. Artemisia pollen percentages (14%)along with Amaranthaceae (7%), Asteraceae (1%), Poaceae (2%) pollenpercentages were all lower than previous. Cyperaceae (1%) pollen per-centages, as well as riparian species, including Salix (b1%) and Alnus(b1%) increased. Short-term increases of Artemisia pollen percentagesand those of aquatic pollen types were notable in the 7-cm thick organiclayer (241–247 cm; 8600–8200 cal yr BP). Pinus flexilis-type increasedto 4% while Pinus contorta-type decreased to 2% during this shift. Abiespollen increased to 5% and Picea pollen decreased to 3%. Around9000 cal yr BP total Pinus pollen decreased to 36%,while Amaranthaceae(14%), Asteraceae (3%), Poaceae (2%), and Cyperaceae (4%) all increased.Aquatic and riparian species, including Salix (2%), and Alnus (b1%) alsoincreased.

CharcoalLL-III had 21 fire episodes with a mean FRI of 320 years, which

was the least frequent of the record (Fig. 3). Peak magnitudesfor the 21 events ranged between 0.1 and 871.0 and averaged124.0 particles/cm2/episode. Two out of the 21 fires (5500 and8230 cal yr BP) had peak magnitudes N500 particles/cm2/episode.

LL-IV: the Populus period (depth 160–140 cm, 4000–3100 cal yr BP)

PollenFrom 4000 to 3100 cal yr BP, Populus pollen percentages increased

from 1% to 31% and remained high for ~900 years. The increase inPopulus pollen abundance was anomalous in western North Americapollen diagrams, so these counts were verified by Carter, Brunelle andMinckley. All relative pollen abundances were lowered by the inclusionof Populus pollen percentages in the terrestrial sum. Pinus pollen

Figure 4. Pollen percentage and macrofossil data plotted against time for Long Lake, WY. Gray shading indicates 5× exaggeration of pollen percentage data. A ‘+’ symbol indicates thepresence of pollen at trace amounts. Black circles indicate identified macrofossils.

410 V.A. Carter et al. / Quaternary Research 80 (2013) 406–416

Climate  

Fire  Vegeta-on  

Climate  

Fire  Vegeta-on  

Site Location Long Lake, Wyoming

Methods

Pollen Charcoal

Grouped Taxa Canopy vs. Understory

& Conifers, successional, deciduous, herbs, and shrubs

200020502100215022002250230023502400245025002550260026502700275028002850290029503000305031003150320032503300335034003450350035503600365037003750380038503900395040004050

Age

(cal

yr B

P)

10000

Total P

ine (I)

50000 100000

Populu

s (I)

1000 2000 3000

Artemisi

a (I)

500 1000 1500

Amaranth

(I)

200 400 600 800

Poace

ae (I)

250 500

Asterac

eae (

I)

Future Research

•  Quantify the changes from a lodgepole pine

system to a quaking aspen mixed forest

•  Quantify the ecological response

to determine low severity vs. high

severity fire

•  Carter et al., 2013 identified a unique vegetation transition period from a logdepole pine dominated system to a lodgepole pine/ quaking aspen mixed forest between 3000-4000 cal yr BP. They coined this period the ‘Populus period.’ (see figures below)

Regionalization of fire regimes in the Central Rocky Mountains, USA

Vachel A. Carter a,⁎, Andrea Brunelle a, Thomas A. Minckley b, Philip E. Dennison c, Mitchell J. Power d

a RED Lab, Department of Geography, University of Utah, Salt Lake City, UT 84112, USAb Department of Geography and Program in Ecology, University of Wyoming, Laramie, WY 82071, USAc URSA Lab, Department of Geography, University of Utah, Salt Lake City, UT 84112, USAd Utah Museum of Natural History, Department of Geography, University of Utah, Salt Lake City, UT 84112, USA

a b s t r a c ta r t i c l e i n f o

Article history:Received 24 January 2013Available online 5 September 2013

Keywords:CharcoalFirePollenVegetationClimateClimatic boundaryLake sedimentsPaleoecology

Fire is one of the most important natural disturbances in the coniferous forests of the US Rocky Mountains. TheRocky Mountains are separated by a climatic boundary between 40° and 45° N, which we refer to as the centralRockyMountains (CRM). To determine whether the fire regime from the CRMwas more similar to the northernRocky Mountains (NRM) or southern Rocky Mountains (SRM) during the Holocene, a 12,539-yr-old sedimentcore from Long Lake, Wyoming, located in the CRM was analyzed for charcoal and pollen. These data werethen compared to charcoal records from the CRM, NRM and SRM. During the Younger Dryas chronozone, thefire regime was characterized as frequent at Long Lake. The early and middle Holocene fire regime was charac-terized as infrequent. A brief interval from 4000 to 3000 cal yr BP, termed the Populus period, had a frequentfire regime and remained frequent through the late Holocene at Long Lake. In comparison to sites from theNRM and SRM, the fire regime at Long Lake was most similar to the SRM during the past 12,539 cal yr BP.These results suggest the disturbance regime in the CRM has a greater affinity with those of the SRM.

© 2013 University of Washington. Published by Elsevier Inc. All rights reserved.

Introduction

Driven by factors such as temperature, precipitation, humidity, windand fuel availability (Westerling et al., 2003), fire is a dynamic forceshaping forest composition and is considered one of themost importantnatural disturbances in the coniferous forests of the western UnitedStates (US). Understanding these variables is important for determininghow fire regimes may vary in response to climate change (Dale et al.,2001). One of the fewwayswe can learn about the interactions betweenfire and climate is to look at past fire regimes as a baseline againstwhichto measure modern changes. Fire histories obtained through the exam-ination and quantification of charcoal preserved in lake sediments areparticularly useful because of their long temporal span. Unlike treerings, which offer annual resolution but are generally age-limited tothe past few hundred years, charcoal records preserved in lake sedi-ments have the ability to reconstruct a fire history over millennia(Long et al., 1998). Charcoal records are also useful in that they canidentify long-term shifts in fire regimes during periods of major climatechange (Brunelle and Whitlock, 2003). To understand the fire ecologyof a system, sediment-based fire reconstructions are compared todetermine how fire regimes respond as climate changes through time(Minckley and Shriver, 2011; Minckley et al., 2012).

The US Rocky Mountain region is normally divided into eitherthe northern Rocky Mountains Range (NRM) or southern Rocky Moun-tains Range (SRM). The central Rocky Mountains Range (CRM) are

defined here as the geographical location also known as the WyomingBasin that separates the NRM and SRM (Baker, 2009) (Fig. 1). In thiscontext, the CRM can be viewed as the transition zone between theGreat Basin, the Great Plains, the NRM, and the SRM (Brunelle et al.,2013). The CRM is of particular interest because currently precipitationin the CRM and NRM is influenced from westerly storms originatingfrom the northern Pacific Ocean in the winter, and both experiencesummers that are relatively warm and dry (Mock, 1996; Shinker,2010; Wise, 2010). However, it is unclear whether the CRM has beeninfluenced by these same precipitation patterns through time or howdifferent precipitation patterns may influence fire regimes in the CRM.

The SRM typically experience precipitation patterns out of phasewith the NRM; based on the observation that when the NRM are anom-alously wet, the SRM are anomalously dry (Dettinger et al., 1998;Wise,2010). Within this dipole, the CRM historically has followed the mois-ture patterns of the NRM (Mock, 1996; Baker, 2009; Shinker, 2010).The dipole fluctuation in precipitation between the NRM and SRM is as-sociated with El Nino–Southern Oscillation cycles (Wise, 2010), whichare known to influence wildfire occurrence and severity in particularregions in the United States (Westerling et al., 2003).

Past fire regimes in the NRM have been researched more heavilythan those of the SRM, with even fewer fire reconstructions along thetransition zone between the two regions (Minckley et al., 2007, 2012;Brunelle et al., 2013). Based on climatic association, it is not understoodwhether the CRMhas a distinct fire regime, or whether its fire regime ismore similar to the NRM or SRM. Dettinger et al. (1998) proposed aclimatic boundary that separates the NRM and SRM between 40° and45° N latitude, but this climatic boundary was likely not stationary

Quaternary Research 80 (2013) 406–416

⁎ Corresponding author.E-mail address: [email protected] (V.A. Carter).

0033-5894/$ – see front matter © 2013 University of Washington. Published by Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.yqres.2013.07.009

Contents lists available at ScienceDirect

Quaternary Research

j ourna l homepage: www.e lsev ie r .com/ locate /yqres

•  Pollen influx diagram showing the 6 most common pollen taxa and their responses to each individual fire (the horizontal lines)

•  Charcoal results showing charcoal accumulation and background (red line), 12 fire events (+ symbol), fire return interval (FRI) and peak magnitude.

•  LDA results indicating a threshold of 100 particles/cm2/episode using the 6 most common pollen taxa. Figures show pre fire samples and post fire samples (called the response). Group 0 indicates the 100 particles threshold. Group 1 indicates the 200 particles threshold.

•  When aspen pollen Influx increase, the FRI

increases and the largest peak magnitude

event occurs.

•  Examine grouped taxa responses to thresholds using an LDA approach. •  Analyze pre-fire and post-fire pollen taxa to examine the ecological responses after each fire event to determine the mechanism that altered the FRI. This will be done using a SEA approach. •  Blarquez and Carcaillet.(2010) & Shriver and Minckley (2013) both successfully used the SEA approach to examine pre-and-post fire responses. This study examines the ecological response at a much higher resolution than previously used.