total AE energy

all days

normalized AE energy

NC

NM

% o

ccur

renc

e un

der I

ndic

ated

dai

ly te

mpe

ratu

re a

nom

aly

Eppes

EP41C-0942: Real-time Observations of Rock Cracking and Weather Provide Insights into Thermal Stress-Related Processes of Mechanical Weathering.

Eppes, M.C., Magi, B., (Geography & Earth Sciences) Keanini, R., (Mechanical Engineering and Engineering Sci.) EAR-0844335, 844401, 0705277

Mot

ivat

ing

Fiel

d O

bser

vatio

ns

Southwestern USMcFadden et al., 2005

Owens Valley, CAD’Arcy et al., 2014

Gobi DesertEppes et al., 2010

Mars , Spirit TraverseEppes et al., 2015

Central EgyptAdelsberger et al., 2009

North Carolinadebris flow

3500 cal yr bp

Pennsylvaniaperiglacial

boulder field

0°

90°

180°

270°

10%8%6%4%2%

0°

90°

180°

270°

8%6%4%

2%

Mot

ivat

ing

Inst

rum

enta

tion

Obs

erva

tions

New

Num

eric

al M

odel

ing

Of

Sola

r-In

duce

d Cr

ack

Grow

th

0

5

10

15

20

25

30

35

40

45

50

0

500

1000

1500

2000

2500

S-Top S-NE-Top S-NS-E S-S S-WS-SW-Bottom S-Bottom EventsT-Ambient T-Max T-MIN

Hour of Day0 09 06 03 12 15 18 21 24

A.

max

imum

prin

cipa

l stra

in (u

s)Ev

entR

ate

(Eve

nts/

Min

)

Tem

pera

ture

(ºC)

sunrise sunset

1

New

Ana

lysis

of A

E-m

easu

red

……

……

……

.c

rack

ing

Data

led to

which led to

Conc

lusio

ns a

nd Im

plic

atio

ns

and and

Cracks in mid-latitude desert rocks exhibit preferred orientations,

as do cracks in rocks on Mars and in humid-temperate boulder fields,

Eastern USA, Aldred et al., 2015

leading to the inference that insolation plays a key role in rock breakdown under a broad range of temperature and humidity conditions.

Research QuestionWhat is the role of those conditions in driving or limiting mechanical weathering?

All data presented in this panel are crack strike orientations that were measured on randomly chosen non-bedrock boulders. Raoand/or Raleigh P-Values for all plots are <0.05, indicating non-random orientations.

at both sites, the vast majority of observed cracking occurred during thermal events, and recorded strain shows evidence of thermal fatigue,

AE

temp

moisture

meteorology

strain

We directly monitored rock cracking using acoustic emission (AE) sensors, and rock-environmental conditions in

two granite boulders (NC – 1 year; NM – 3years),

and found thatNC

NCNM

NM

We developed a model for temp- & moisture-dependent subcritical crack growth to inform our interpretation of measured cracking and field data.

Left: example days in which the highest numbers of AE occurred in the NC rock. Virtually all event clusters for NM & NC occur during relatively rapid and complex rock surface temperature changes that coincide with peak solar-induced thermal stresses at sunset or mid-day. Above: A typical day in which cracking occurred, where rock surface strain tracks with rock surface temperature, and cracking occurs during relatively rapid temperature and strain changes, in this case brought on by a sudden storm.

and

but cracking rates do not correlate with temperature, temperature gradient, or rate of temperature change.

Data from Eppes et al., in review

Eppes and Keanini, in preparation

Led to

Left: All three panels depict temperature conditions during individual minutes when AE were recorded in NC (n = 1201) versus the AE rate, a measure that is proportional to rock fracture. We observe no trends in these temperature metrics and rock cracking throughout the entire period of record for the NC rock. (Gradient = Max rock surface temp. – min. rock surface temp)

A Ca

se S

tudy

of R

ock

Fall

Crac

king

Rat

e (A

E/m

in)

Crac

king

Rat

e (A

E/m

in)

Crac

king

Rat

e (A

E/m

in)

Avg. Rock Temp (C) Max Rock Temp Gradient (C) Avg. Rock ΙΔT/minΙ (C)

1-14-112

Inso

latio

n (k

W/m

2 )Tm

ax-T

min

Nor

mal

ized

Even

ts/m

in

1000

100

10

1

1

0.5

0

Insolation

Tmax-Tmin

Events/min

00:00 12:00 23:59

Key

Date#AE

Model Parameters1) Solar-Induced Temperature Field

--conductive heat transfer of a sinusoidal heating function (from Holzhousen, 1989) 1) Solar-induced thermal stress field (from Holzhousen, 1989)

-- intergranular interactions only-- f(diurnal temp. range) --f(differences in Coefficient of thermal expansion of adjacent minerals.

3) Crack growth modeled as a novel combination of Paris’s Law (fatigue) & Charles’ Law

𝑎𝑎 𝑁𝑁 = 𝑎𝑎𝑜𝑜𝛽𝛽 + 𝛽𝛽𝐶𝐶1𝑁𝑁

1/𝛽𝛽

• 𝑎𝑎 𝑁𝑁 =crack length at diurnal cycle, 𝑁𝑁• 𝑎𝑎0 = 𝑎𝑎 0 = initial crack length• 𝛽𝛽 =1-m/2• 𝐶𝐶1 = 𝐶𝐶∆𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚𝜋𝜋𝑚𝑚/2 , where 𝐶𝐶 & m are material- and environment-dependent Paris law

coefficients available in the literature, and ∆𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚=∆𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚(𝑧𝑧) is the depth-dependent maximum thermal stress

Model results show that simple, insolation-induced crack growth rates are strongly dependent on 1) temperature range and 2) moisture

Model validation

1

100

10000

1000000

100000000

0 50 100Relative Humidity (%)

Tim

e (y

ears

) 19.5C

21C

20C20.5C

21.5C

Preliminary model results show the strong dependence that both diurnal temperature range and relative humidity have on rock lifetimes (calculated as spalling of a 1 m boulder).Different lines represent different average annual diurnal temperature ranges.

Modeled rock lifetimes

Diurnal temperature range by itself is not a strong predictor of AE-measured cracking rate s(left), but cracking disproportionately occurs under relatively high ΔT and humidity conditions (bottom left).

Cracking also disproportionately occurs on days with climatologically extreme temperatures (right).

Diurnal Rock Surface ΔT (C)

Two NC outliers

mea

n 24

hr A

E en

ergy NC

NM Stock et al., 2013Using Stock et al.’s (2013) detailed ~100 year record of rock falls from Yosemite National Park, we examined the percentage of all falls occurring on days with anonymously high or low temperatures compared to climate data from Yosemite Valley. About 10% more rock falls than expected occur on these days overall.

>1

>1

<1

<1

%

%-1<ta<1

cracking with respect to daily relative humidity

cracking with respect to diurnal temp. range

cracking with respect to temp. anomalies

NC NMWe examined the fraction of the entire period of record characterized by various humidity conditions (A), compared to the fraction of AE energy measured under those conditions (B). Generally, cracking rates are higher than expected at high humidity (circled data).

A.

B.

Total AE Energy

We compared average daily temperatures for days in which AE occurred, to the average daily temperatures for those dates derived from 1951-1980 climate data from nearby weather stations in order to calculate the temperature anomaly of any given “cracking day” with respect to climate data. We then normalized the data to account for any over-or under-sampling of extreme temperature days in the year of observation.

As such, any absolute temperature anomaly values greater than one represent climatically extreme temperatures (> 1σ) for days when cracking occurred.

Model results show that insolation-induced crack growth rates astrdependent on 1) temperature range and 2) moisture

Model results show that crack growth rates in rock, regardless of loading mechanism, are strongly dependent on moisture.

Crack growth rates under solar-induced intergranular stresses are strongly dependent on the variance in rock surface

temperature from far field temperature (ΔT)(approximated in the model by the diurnal temperature

range).

Days with extreme temperatures represent the days when there is the highest likelihood of reaching maximums in ΔT ,

thus approaching critical stresses necessary to rapidly propagate cracks.

Therefore when other stress-inducing triggers occur (freezing conditions, weather-induced thermal stress), the rock is

“primed” for failure.

Thus, when considering drivers or limits to all stress-loading mechanisms, both moisture and insolation-related stresses

should be considered.

Model results show that insolation-induced crack growth rates are strongly dependent on 1) temperature range and 2)

moisture

Rock fall data from a complex field site where other stress loading is undoubtedly at play (e.g. topography, higher occurrence of freezing

temperatures) are consistent with our AE observations:

More than expected rock falls occur under extreme temperatures.

Occurrences (%) of indicated temp anomaly

>1

<1-1<t

a<1

all days

rock fall days

16%

24%21%

19%

We examined the fraction of the entire period of record characterized by various temperature conditions (A), compared to the fraction of rock falls observed under those conditions (B). Generally, rock fall rates are higher than expected at high temperatures (C).

A. B. C.