Effects of 45 years of heavy road traffic and infrastructure on permafrost and tundra at Prudhoe Bay, Alaska
Ala s k a G e
ob
ot
an
y C e nter
AGU abstract:
GC23J-1215
Buchhorn, M.1,2 ([email protected]); M.K. Raynolds1; D.A. Walker1; M. Kanevskiy3; G. Matyshak4; Y. Shur3; and J. Peirce1
University of Alaska Fairbanks, Institute of Arctic Biology, Fairbanks, AK 99775 USA; 2University of Alaska Fairbanks, Geophysical Institute, Fairbanks, AK 99775 USA
3
University of Alaska Fairbanks, Department of Civil & Environmental Engineering, Fairbanks, AK 99775 USA; 4Moscow State University, Department of Soil Science, Moscow, Russia
1
I INTRODUCTION
Vegetation change analysis
III Time series of the Lake Colleen Site A from 1968 to 2013
Following the discovery of oil at Prudhoe Bay
in 1968, a series of environmental changes
occurred that were the result of natural longterm processes and changes caused directly
and indirectly by infrastructure. Here we examine the changes associated with a road at
Colleen Site A, Deadhorse, Alaska (Fig. 1).
The changes in non-infrastructure-related thermokarst are
thought to be due to a combination of a long-term upward trend
in summer warmth and exceptionally warm summers in 1989,
1998, and 2012 when the active layer (layer of soil that thaws
in summer) increased in thickness and melted the tops of ice
wedges (Fig. 5, 6).
An abrupt increase in the abundance of thermokarst features is evident in the time series analysis of aerial photos from 1968 to 2013 of the Lake Colleen Site A (Fig. 4).
This effect has also been noted in three studies in the Prudhoe Bay area. (Jorgenson et al. 2006; Raynolds et al. 2014, Jogenson et al. 2015).
1968
1972
1977
1984
1989
1995
2000
2005
2010
V Summary of findings
2013
time series analysis
• Prior to construction of the Spine Road in
1969, the Colleen site A study area had numerous scattered thermokarst pits indicating
that the area had some thawing ice wedges
at the intersections of polygon troughs. The
pattern of thermokarst changed very little
between 1949 and 1972.
Prior to road construction in 1969, the north side of the road
had more wet tundra and thermokarst pits than the south side of
the road (Fig. 7). After the road was built, extensive deep thermokarst developed along polygon troughs on the south side of
the road and aquatic vegetation increased. Other more subtle
changes occurred to the species composition of the major vegetation types (Fig. 8).
• The Spine Road was constructed in 1969,
altering drainage patterns and introduced
gravel and large quantities of dust to the
tundra adjacent to the road, such that over
the past 45 years the pattern of thermokarst
has changed dramatically.
Figure 8. Comparison of the main vegetation types between 1972 and 2013.
Non-infrastructure
thermokarst:
1.8x since
1990
Infrastructure
thermokarst:
3.5x since
1990
North-Side
Figure 1. The Lake Colleen region. False-color-infrared World View image
(July 9, 2010). Note: red tones show areas of highly productive vegetation.
Figure 5. Trends in thermokast for sample areas within the Prudhoe Bay oilfield 1970-2010.
Most changes in landscapes along roads are
caused by a combination of (a) flooding due
to the elevated road beds that interrupt natural
drainage flow patterns; (b) heavy road dust,
which smothers the vegetation and changes
the albedo of the tundra surface; and (c) snow
banks that form along the edges of the elevated roads. All of these factors tend to raise
soil temperatures, which in turn increase the
active layer thickness near roads, leading to
roadside thermokarst (Fig. 2).
35
1998
South-Side
2012
SWI (°C months)
30
July 15, 1972, U.S Army Cold Regions
Research and Engineering Laboratory
(CRREL), black & white, 1:6000
July 29, 1968, PB Alaska, color,
1’’ = 1000’
July 24, 1977, PB Alaska, B/W,
1’’ = 1500’
Aug 30, 1984, PB Alaska, color,
1’’ = 1500’
Aug 22, 1989, PB Alaska, CIR,
1’’ = 500’
Aug 8, 1995, PB Alaska, CIR,
1’’ = 600’
June 30, 2000, PB Alaska, color,
1’’ = 1500’
Aug 3, 2005, PB Alaska, color,
1’’ = 1500’
July 9, 2010, PB Alaska, color,
digital, 1 foot pixel
2013 BP Alaska, digital,
color, 0.75-foot resolution
25
20
15
• Trough-to-polygon topographic contrast of
Transect T2 is nearly double that of T1 (Fig.
14a).
10
Figure 6. Time series of summer warmth.
Summer warmth index (SWI) = sum of
monthly mean temperatures > 0ºC.
Summer warmth index
1970
1980
1990
2000
2010
Figure 7. Classification of the main vegetation types of the aerial photos from
1972 and 2013.
Interpretations of ice-wedges sections based on borehole profiles
IV Transect Analysis
A visualisation of all collected geo-ecological parameters along the transects T1 (north) and T2 (south) at 1 m intervals within 100 m of the road and at 5 m intervals from 100 to 200 m from the road can bee seen in Figure 9.
Planar view of transects with plot and
borehole locations
d
Jorgenson, M.T.; Shur, Y.L.; Pullman, E.R. (2006), Abrupt increase in permafrost degradation
in Arctic Alaska. Geophysical Research Letters, 25, L02503.
Jorgenson, M. T.; Kanevskiy, M.; Shur, Y.L. et al. (2015), Role of ground ice dynamics and
ecological feedbacks in recent ice-wedge degradation and stabilization, J. Geophys. Res.
Earth Surf., n/a–n/a, doi:10.1002/2015JF003602.
Raynolds, M. K.; Walker, D.A. et al. (2014), Cumulative geoecological effects of 62 years of
infrastructure and climate change in ice-rich permafrost landscapes, Prudhoe Bay Oilfield,
Alaska, Glob Chang Biol, 20(4), 1211–1224, doi:10.1111/gcb.12500.
Walker, D.A.; Everett, K.R.; Webber, P.J.; Brown, J. (eds.) (1980), Geobotanical Atlas of the
Prudhoe Bay Region, Alaska, CRREL Report 80-14. U.S. Army Corps of Engineers, Cold
Regions Research and Engineering Laboratory, Hanover.
Acknowledgements
NSF Arctic Science, Engineering and Education for Sustainability (ArcSEES) grant no. 1263854
with partial support from the Bureau of Ocean Energy Management (BOEM) and the Department of Interior (DOI) and NASA Land Cover Land Use Change (LCLUC) Program, Grant No.
NNX14AD90G. Moreover we thank BP Alaska Prudhoe Bay Unit and Aerometric Inc. for providing GIS, aerial-photo data, and analysis of the regional infrastructure.
0m
50m
150m
100m
200m
14.5 m
14.5 m
14.0 m
14.0 m
13.5 m
13.5 m
13.0 m
13.0 m
12.5 m
12.5 m
12.0 m
12.0 m
© M.Buchhorn
11.5 m
190 m
200 m
270 m
280 m
290 m
300 m
310 m
B
U4d
B
U4d
HC
M2d
R
WT
HC
U4d
E1d
M2d
WT
HC
M2d
HC RB Microrelief Type RB
Vegetation Type
50m
330 m
340 m
350 m
360 m
370 m
380 m
390 m
400 m
R
R
U4d
R
LC
U4d
R
WT
E1d
R
WT R
U4d E1d
LC
M2d
R DT R LC R R
LC
M2d
M2d
M2d
R
P
W1
R
FC
U4d
R WT R
U3d
E1d
LC
150m
R
U4d
WT R LC R
E1
11.5 m
410 m
NE
100m
M2d
E1d E3d
HC R
0m
E1d
B
P
320 m
Transect T1
U4d
E1d
U4d
260 m
WT
B
250 m
M2d
B
HC
DT
B
240 m
M2d
U4d
B
P
W1
B
U4d
HC DT
0m
E1d
B
HC
E1d
M2d
F
E1d
U4d
F
HC
M2d
F
U4d
F
DT
F HC F
HC
HC
U4d
M2d
WT
HC
M4d
R
U4d
M2d
DT
HC
U4d
HC
U4d
M2d
LC
HC
B
R
WT
U4d
R
M4d
LC
HC
R WT R
U4d
LC
U4d
R
50m
100m
M4d
R WT
230 m
Spine Road
E1d
150m
M4d
200m
220 m
U4d
Transect T2
SW
210 m
U3
LC
U3
R WT
E1
R
LC
M2
R
200m
DT
180 m
R
DT
R
WT R
LC
P
LC
R WT
E1
170 m
DT
LC
160 m
M2
150 m
DT
140 m
DT
130 m
U4
120 m
U4
110 m
U3
100 m
U4d
90 m
U4d
80 m
U3d
70 m
U3d
60 m
M2d
50 m
DT
40 m
LC
30 m
M2d
20 m
FC
10 m
Caaq
Salova
0m
U4
E1
U3
E1 U3
U4
E1
M2
U4
Relative profile (superelevation 1:10)
Surface height measurement (m)
Vegetation height measurement (m)
Thaw depth in August 2014 (m)
Water level (m)
Vegetation layer
Water / ponds
Unfrozen soil
Frozen soil
Frost boil
W1
E1
E3
M2
E1
100 cm
100 cm
LAI
2.0
1.5
75 cm
75 cm
1.5
M4
U3
U4
B
1.0
50 cm
50 cm
1.0
0.5
25 cm
25 cm
0.5
0
0 cm
0 cm
150m
50m
100m
0m
0m
50m
0
150m
100m
200m
15.0 m
15.0 m
14.5 m
14.5 m
14.0 m
14.0 m
13.5 m
13.5 m
13.0 m
13.0 m
Comparison of topo profile extracted from Lidar
data and archived by field survey
12.5 m
12.5 m
Topo profile archived by field survey (m)
Lidar profile (highest hit) - 2014 survey (m)
Lidar profile (bare earth model) - 2014 survey (m)
12.0 m
10 m
20 m
30 m
40 m
50 m
© M.Buchhorn
60 m
70 m
80 m
90 m
100 m
Microrelief type
Vegetation layer
Water / ponds
Unfrozen soil
Frozen soil
Frost boil
c
d
Surface and soil temperatures were
warmer on the south side (transect T2)
than the north side (transect T1), particularly in winter. (Fig. 13).
Figure 14. Differences in (a) trough-polygon center contrast, (b) thaw
depth, (c) vegetation height, and (d) Leaf Area Index (LAI) between the
transects T1 and T2.
Figure 12. Soil plug from center of an ice-wedge polygon. Note
the 13-cm thick mineral surface horizon, which is the dust layer
above the original organic surface horizon.
Figure 11. Estimates of percent live plant cover for different lifeforms on 1-m2 plots. Dark outline - polygon troughs (T); North
side on right.
Thaw depth - Dust - LAI Chart
Thaw depth in August 2014 (cm)
Dust layer in August 2014 (cm)
LAI (Leaf Area Index) in August 2014
A
Suerface temperature profiles
Temperature profiles
B
Soil temperature profiles (-40 cm)
Thermokast pit
Wet trough
Dry trough
Frost boil
110 m
120 m
Rim
High-centered polygon center
Low-centered polygon center
Flat-centered polygon center
Road side berm
130 m
140 m
150 m
160 m
170 m
180 m
190 m
200 m
Vegetation type
B
Dry Saxifraga oppositifolia, Juncus biglumis forb barren
Dry Carex aquatilis - Salix ovalifolia Barren
Barren
Disturbed sites (indicated by a ‘d’ in vegetation code)
210 m
W1
E1
E3
M2
220 m
230 m
240 m
250 m
Lakes & ponds
Aquatic Carex aquatilis sedge tundra
Aquatic Scorpidium scorpioides moss tundra
Wet Carex aquatilis, Drepanocladus brevifolius
sedge tundra
260 m
270 m
M4
U3
280 m
290 m
300 m
310 m
Wet Carex aquatilis, Scorpidium scorpioides
sedge tundra
Moist Eriophorum angustifolium, Dryas integrifolia,
Tomenthypnum nitens, Thamnolia subuliformis sedge,
dwarf shrub tundra
320 m
330 m
U4
340 m
350 m
360 m
370 m
Moist Carex aquatilis, Dryas integrifolia, Salix arctica,
Tomenthypnum nitens sedge, dwarf shrub tundra
380 m
390 m
400 m
Comparison of topo profile extracted from Lidar
data and archived by field survey
Topo profile archived by field survey (m)
Lidar profile (highest hit) - 2014 survey (m)
Lidar profile (bare earth model) - 2014 survey (m)
11.5 m
Thaw depth - Dust - LAI Graph
Thaw depth in August 2014 (cm)
Dust layer in August 2014 (cm)
LAI (Leaf Area Index) in August 2014
Sep14
1
Information from the Colleen Site A studies, combined with the rich record of historical aerial photographs provide an excellent basis for examining the changes to
this region.
2
There have been substantial changes to
the vegetation, soils, microtopography,
permafrost and ice wedges due to both climate change and infrastructure effects.
VI CONCLUSIONS
3
Vegetation composition studies on these
transects documented major reductions
in lichen cover over 24 years, reduction in
mosses, increase in shrub cover adjacent
to the road & overall decrease in diversity.
4
Ice-wedge degradation has halted on some
of the polygons on the south side of the
road. Enhanced productivity due to wetter, more nutrient-rich conditions has produced insulating vegetation and litter.
• Despite the effects of road construction and
heavy traffic on the stability of upper permafrost, ice-wedge degradation has halted
on some degraded wedges.
• Initial degradation of ice wedges along T2
was probably related to the flooding of the
south side of the Spine Road due to road
construction, but at present even the wedges under deep, water-filled troughs are more
stable than the wedges along T1, which
have not been affected by flooding.
• Species composition of the dominant moist
and wet vegetation changed due to dust
deposition.
410 m
Figure 9. (a) Overview of the transects T1 and T2. Note: red line mark the location of the transects NE and SW from the Spine Road. Position of center and trough permanent vegetation plots are marked as squares, and boreholes as white dots. (b) relative profiles of the transects T1 and T2. Note the 10x vertical exaggeration
(c) Microrelief and vegetation along the transects. (d) Graph showing the thaw depths, dust depths, and LAI measurements along the transects. (e) Comparison of the topo profile extracted from the LiDAR dataset (highest hit and bare earth model) and field survey.
• Most (35 of 43) boreholes drilled in icewedge troughs and surrounding rims encountered massive-ice bodies (mostly ice
wedges at various stages of degradation
and recovery).
• Major vegetation changes occurred over
45 years: aquatic vegetation and water increased on both sides of the road, but most
dramatically on the south side which is regularly flooded during spring snow melt.
12.0 m
11.5 m
0m
Compared to polygon centers, troughs
had almost no shrub cover, high graminoid cover and higher moss cover (often below water). Moss cover increased
with distance from road, and total plant
cover increased with distance from
road on north side (transect T1), but not
on south side (Transect T2) (Fig. 11).
Soils within about 50 m on both sides
of the road now have mineral surface
horizons composed largely of road dust
and gravel that overlie the original organic soil horizons (Fig. 12).
Vegetation and soil studies
Lakes & ponds
Aquatic Carex aquatilis sedge tundra
Aquatic Scorpidium scorpioides moss tundra
Wet Carex aquatilis, Drepanocladus brevifolius
sedge tundra
Wet Carex aquatilis, Scorpidium scorpioides
sedge tundra
Moist Eriophorum angustifolium, Dryas integrifolia,
Tomenthypnum nitens, Thamnolia subuliformis sedge,
dwarf shrub tundra
Moist Carex aquatilis, Dryas integrifolia, Salix arctica,
Tomenthypnum nitens sedge, dwarf shrub tundra
Dry Saxifraga oppositifolia, Juncus biglumis forb barren
Dry Carex aquatilis - Salix ovalifolia Barren
Barren
Disturbed sites (indicated by a ‘d’ in vegetation code)
• Vegetation heights are taller (Fig. 14c) and
leaf-area-index is greater (Fig. 14d) on the
south (wetter) side of the road.
b
Microrelief type
Thermokast pit
Wet trough
Dry trough
Frost boil
Rim
High-centered polygon center
Low-centered polygon center
Flat-centered polygon center
Road side berm
• Thaw depths are greater for all plant communities on the south (wetter) side of the
road (Fig. 14b).
a
Vegetation type
LAI
2.0
Surface height measurement (m)
Vegetation height measurement (m)
Thaw depth in August 2014 (m)
Water level (m)
References
0m
NE
15.0 m
Relative profile (superelevation 1:10)
Figure 3. Drilling of cores in ice-wedge polgon center and troughs.
50m
Transect T1
15.0 m
200m
e
Spine Road
100m
DT
c
150m
U4d
b
200m
WT
Vegetation plots along two transects were set
up in polygon centers and troughs at 5, 10,
25, 50, 100 and 200 m from both sides of the
heavily-traveled Spine Road at Colleen Site A
to analyze road effects. Differential GPS and
a robotic TotalStation were used to survey the
transects (Fig. 9). Soil and vegetation data
were collected at 1 m intervals within 100 m of
the road and at 5 m intervals from 100 to 200
m from the road (Fig. 9, 11, 12). Fifty-seven
boreholes were drilled in ice-wedge polygon
centers and troughs (Fig. 10). One hundred
and thirty Thermochron® Temperature Data
Loggers (iButtons) were installed to monitor
air and ground temperatures and determine
ice-wedge stability (Fig. 13).
Transect T2
SW
Thaw depth / Dust / LAI
Chart
This study, initiated in 2014, builds on baseline
data collected in the same area by Walker et
al. 1980. A time series of aerial photographs
of Colleen Site A was used to show the transition of the landscape between 1968 and 2013
(Fig. 4). Aerial photographs of 1972 and 2013
were used to produce a vegetation change
map (Fig. 7, 8) (aerial photograph of 1968 was
not available to the time of the analysis).
Relative profile (superelevation 1:10)
II METHODS
Microrelief &
Vegetation
Characterization
Figure 2. Roadside disturbances related to dust and thermokarst.
© BP, true color aerial photo of 2014
Comparison Lidar topo profile to survey
profile (superelevation 1:10)
Combined vegeta.on and ecosystem effects of dust and thermokarst a
Figure 10. Examples for drilling profiles for transect T1 and T2.
Borehole drilling encountered massive
ground ice, including wedge ice (WI),
thermokarst cave ice (TCI) and composite (ice/soil) wedges (CW) (Fig. 10).
In detail, aerial photography from 2014 is shown at the top of Fig. 9, with cross-sectional transect of surface elevation, vegetation height, thaw and water depth (b); surface geomorphology and vegetation type (c); probed thaw
depth, LAI, and depth of surface dust layer (d); and comparison of the field survey to LiDAR data (e). The interpretations of ice-wedge cross-sections based on borehole profiles is seen in Figure 10.
Climate-­‐related thermokarst away from infrastructure transects analysis
Exceptionally warm summers
occurred in 1989, 1998, and
2012.
5
Figure 4. Colleen Site A study area (as in Figure 1) time series 1968-2013, showing progression of change. Notes: The Spine Road was constructed in 1969 so does not appear on the 1968 image. Most of the thermokarst pits that are present in 1972 (shortly after construction of
the road) are also visible on the 1968 image. Greatly expanded thermokarst is seen beginning in 1989.
Roadside dust Roadside thermokarst 1989
• Thermokarst is now deepest and most extensive on the southwest side of the road,
which is periodically flooded. Historical climate data and photos indicate that between
1989 and 2012 a regional thawing of the
ice-wedges occurred, increasing the extent
of thermokarst on the both sides of the road.
Dec14
Mar15
Jun15
Sep14
Dec14
Mar15
Jun15
Figure 13. (a) Comparison of the average daily surface temperature at polygon centers (C) and troughs (T) at Colleen Site A for the time August 2014 to August 2015. (b) Comparison of the average daily soil temperature at -40 cm at polygon centers (C) and troughs (T) at Colleen Site A for the time August 2014 to August 2015.
5
There is a need to examine the consequences of these changes to other components of the ecosystem, including fish,
birds and insects.
6
Many of the recorded impacts are recent
and rapid. Continued monitoring will help
industry and regulators apply these findings to management and development issues.
• Moss cover decreased and bare soil cover
increased with road proximity.
• The effect of the road embankment is visible in the soil temperature regime, showing
warmer soils on the south-side transect.
• LiDAR data with >8 points per m2 can be
used for the production of high resolution
relative profiles (SD was ±16 cm compared
to our survey with a TotalStation).