Modeling Contemporary Range

Contraction in Great Basin Pikas

Jennifer Wilkening

Boundary Peak, White Mountains, Nevada

Rising temperatures,

changing precipitation

patterns, increased frequency

of extreme weather events

Global Climate ChangeCrane Mountain,

Warner Mountains, Oregon

Alpine species particularly vulnerable,

what about pikas ?

Located only in mountains surrounded by sagebrush areas

Relictual populations from cooler climates

Dispersal is difficult

“Habitat Islands” susceptible to biogeographic patterns

of extinction

Mount Jefferson, Toquima Range, Nevada

Several in the Great Basin have become extirpated during the 20th century (Beever et al. 2003)

Occurred at lower elevation sites, hotter and drier

Other factors◦ Amount of talus habitat

◦ Distance to a primary road

However, single strongest determinant of persistence was elevation of nearby habitat

Direct thermal stress

◦ Hyperthermia and death from high temperatures in the 25.5-29.4°C range

◦ Survival declines during extremely cold winters with less insulating snow cover, increased freeze-thaw events

Indirect thermal stress◦ Higher temperatures

limits activity during the day

◦ Changes in vegetation plant communities

Photo by Chris Ray

Lava Beds National Monument

Acute or chronic heat stress: pikas no longer occur where mean daily summer temperature and the amount of time above a temperature threshold are high

Cold stress: pikas no longer occur where amount of time below a temperature threshold is high

Vegetation: pikas no longer occur where there are less forbs and graminoids, more xeric adapted species

Combinations: ◦ Pikas no longer occur where high mean summer

temperatures combined with low forbs/grams◦ Pikas no longer occur where high amounts of time below

a temperature threshold combined with low forb cover

25 Field sites located in Nevada, Oregon, California Site Classification

Sites of persistence vs. sites of extirpation(n = 20)

Site level analysis

Transitional (n = 5), lower occupancy limit > than 200 meters upslope

Sub-site level analysis

Desatoya Range, Nevada

Thermochron ibuttons◦ 5-8 per site

◦ Placed adjacent to hay piles, scat (extirpation sites), GPS locations (transitional sites)

◦ Recorded temperature readings every 4 hours

◦ Placed inside the talus at a depth between 0.5-1 meters

◦ Multiple localities within each site (n = 191)

◦ Varying aspects, elevations (none below historical site elevation

Temperature Methods

Line-point-intercept method

4-5 Vegetation Surveys per site randomly selected

A 50 meter long transect with the data logger location as center point, one additional 50 meter long transect both above and below

Vegetation was placed into 1 of 6 life form categories: Forbs (herbaceous, flowering plants, excluding cushion plants), Graminoids (grasses and grass-like plants such as sedges and rushes), Shrubs (woody plants), Trees, Cushion Plants (low, mat forming plants), and Non-Vascular Plants (including lichen)

Toiyabe Range, Nevada

Toiyabe Range, NevadaDesatoya Range, Nevada

Predictors based on temperature

Mean summer temperature

Number of days above 26 C

Number of days above 28 C

Number of days below -5 C

Number of days below -10 C

Predictors based on vegetation

Relative cover of forbs

Relative cover of graminoids

Sources: 1) Beever et al. 2010 2) Smith and Ivins 1983 3) Smith 1978 4) Hafner 1993

5) Tapper 1973 6) Smith 1974a 7) MacArthur and Wang 1973 8) Dearing 1995, 1996, 1997a

9) Huntly et al. 1986 10) Sundby 2002 11) Ray and Beever 2007 12) Dearing 1995, 1996

13) Kreuzer and Huntly 2003

7 Predictor Variables Logistic Regression Various models composed

of different combinations of temperature and vegetation factors

R 2.6.2. was used to model pika persistence as a function of predictor variables

Relative support for each model and predictor were calculated using AICc

Welch two-sample t-test used to compare mean value of each predictor variable

Steens Mountain, Oregon

Model: Predictor (effect sign) AICC ΔAICC Akaike weight

SITE LEVEL

DaysBelow-10 C (+),

MeanSummerTemp (-)

17.381 0 0.529

MeanSummerTemp/RelForbCov

er (-)

19.473 2.093 0.186

Mean SummerTemp (-) 19.843 2.463 0.154

SUB-SITE LEVEL

Null model (intercept only) 16.363 0 0.282

DaysBelow-10 C/RelForbCover

(-)

17.976 1.613 0.126

MeanSummerTemp/RelGramCo

ver (-)

18.109 1.746 0.118

MeanSummerTemp/RelForbCov

er (-)

18.185 1.822 0.113

RelGramCover (+) 18.433 2.070 0.100

RelForbCover (+) 19.281 2.918 0.066

Photo by Shana Weber

Predictor Akaike

weight

Mean Akaike

wt/model

Sign of effect

SITE LEVEL

MeanSummerTemp 0.722 0.241 Neg (3)

MeanSummerTemp/RelForbCover 0.186 0.186 Neg (1)

DaysBelow-10 C 0.536 0.179 Pos (2), Neg (1)

RelForbCover 0.066 0.033 Pos (2)

DaysAbove28 C 0.058 0.019 Neg (2), Pos (1)

RelGramCover 0.016 0.008 Neg (2)

DaysBelow-10 C/RelForbCover 0.004 0.004 Neg (1)

MeanSummerTemp/RelGramCover 0.000 0.000 Pos (1)

SUB-SITE LEVEL

DaysBelow-10 C/RelForbCover 0.126 0.126 Neg (1)

MeanSummerTemp/RelGramCover 0.118 0.118 Neg (1)

MeanSummerTemp/RelForbCover 0.113 0.113 Neg (1)

RelGramCover 0.100 0.050 Pos (2)

RelForbCover 0.066 0.033 Pos (2)

DaysAbove28 C 0.075 0.019 Pos (4)

MeanSummerTemp 0.071 0.018 Neg (4)

DaysBelow-10 C 0.071 0.018 Neg (4)

The best model (ΔAICc = 0) of persistence at the site level

-5 0 5 10

0.0

0.2

0.4

0.6

0.8

1.0

Linear predictor: f(Mean summer temp., Days below -10 C)

Pers

iste

nce o

bserv

ed (

dots

) and m

odele

d (

line)

C H

D

(a)

Duffer Peak, Pine Forest Range, Nevada

Hays Canyon Range, Nevada

Disease Relationship

between intestinal bacteria and other endoparasites

Species interactions Reduced dispersal Reduction in the

amount of available forage time

Night time activity?

Photo by Chris Ray

Sites of persistence mean

summer temp (17.04°C)

vs. sites of extirpation (11.74°C),

p = 0.00

Mean number of days above 28°C at sites of persistence was 1.23, at sites of extirpation it was 10.42

( p = 0.02)

Hyperthermia and death can occur from even brief exposure to moderately high temperatures (25.5-29.4°C; MacArthur and Wang, 1973, 1974; Smith, 1974)

Behavioral thermoregulation, access to cooler temperatures ? Hart Mountain, Oregon

Pika survival declines during harsh winters; shallow snowpack, ice barrier, etc.

However, number of days below -10°C was positively correlated to persistence, possibly due to: Two year time series

represents a limited period of time.

Pika mortality resulting from harsh winters may occur only periodically.

Long term pika persistence may be less affected by colder winter temperatures.

White Pine Range, Nevada

Ruby Mountains, Nevada

Highly variable

Consume grasses immediately (smaller, less toxic)

Store forbs (herbaceous, flowering plants) for winter consumption (larger, toxic secondary compounds)

Hart Mountain, OregonToiyabe Range, Nevada

Hay piles sometimes contain more than enough plant material for winter survival (Dearing 1997)

OR Hay piles do not always contain

sufficient quantities of vegetation to provide an exclusive food source for the winter (Millar and Zwickel 1972)

OR Pikas may forage outside of the

hay pile (Conner 1983)OR

Hay piles are not always necessary (Simpson 2001)

HOWEVER Most likely hay piles function as an

adaptive response to environmental unpredictability

Photo by Chris Ray

Kiger Gorge, Oregon

Hay pile presence may be more important in the Great Basin

Haying food resources may be more important than grazing food resources

Essential nutrients found only in forbs

Forbs contain preservatives (Dearing 1997)

Moisture content of forbs

Sites of persistence relative forb

cover (28.79) vs. sites of

extirpation (8.61),

p = 0.00

Future of pikas in the Great Basin?

Arc Dome, Toiyabe Range, Nevada