1. How does increasing deciduous
shrub dominance influence:
1. canopy phenology?
2. Growing and peak season
length?
3. vegetation-atmosphere carbon
exchange?
Results
References
Modeled LAI and CO2 Flux
We thank Rick Shory and Geneva Chong for help in
deploying equipment and with data analysis. We
thank Gus Shaver for advisement during the
preparation of the manuscript, and Jesse Krause,
Jessica Gersony, Heather Greaves, and TeamBird
2013 for field assistance. We also thank the Institute
of Arctic Biology and Toolik Field Station
(University of Alaska, Fairbanks) for support and
logistics. This project has been funded by
collaborative NSF grants from NSF’s Office of
Polar Programs (ARC 0908444 to N.T. Boelman,
0908602 to L. Gough, and ARC 0902030 to H.
Steltzer) and a Climate Center grant from Columbia
University (UR00801-07 60437 to K.L. Griffin).
Phenology and Season Length
AMAP (Arctic Monitoring and Assessment Programme) (2012) SWIPA Overview Report. In:
Arctic climate issues 2011: Changes in Arctic snow, water, ice and permafrost. 97pp. Oslo,
Norway.
Euskirchen ES et al. (2006) Importance of recent shifts in soil thermal dynamics on growing
season length, productivity, and carbon sequestration in terrestrial high-latitude ecosystems.
Glob Change Biol, 12:731–750.
Jia GJ et al. (2009) Vegetation greening in the Canadian arctic related to decadal warming. J
Environ Monitor, 11:2231-2238.
Myers-Smith IH et al. (2011) Shrub expansion in tundra ecosystems: dynamics, impacts and
research priorities. Environ Res Lett, 6:045509.
Shaver GR et al. (2013) Pan-Arctic modeling of net ecosystem exchange of CO2. Phil Trans R Soc
B, 368: 20120485.
Street LE et al. (2007) What is the relationship between changes in leaf area and changes in
photosynthetic CO2 flux in arctic ecosystems? J Ecol, 95:135-150.
Zeng H et al. (2011) Recent changes in phenology over the northern high latitudes detected from
multi-satellite data. Environ Res Lett, 6:045508
Canopy Phenology (Fig. 5)
u Compared to EG canopies, DS
canopies reached:
• onset of greening 2 days later
• onset of peak green 13 days
earlier (P < 0.001)
• onset of senescence 3 days earlier
u Compare to EG canopies, DS
canopies’:
• green season was 5 days shorter
• peak season was 10 days longer
(P < 0.01)
Deciduous Shrub Dominance (Fig. 6)
u greater deciduous shrub cover =
• faster green-up (P < 0.001)
• earlier peak onset (P < 0.001)
• longer peak season (P < 0.001)
Comparing Community Type (Fig. 7)
u Based on model estimates, DS canopies had:
• higher LAI
• greater net CO2 uptake
ü greater CO2 loss from RE
(P < 0.001)
ü greater CO2 uptake from GPP
(P < 0.001)
u During green season, DS canopies took up
~ 2x more carbon (an estimated 113 g C m-2
season-1) compared to EG canopies
(P < 0.001)
u During peak season, DS canopies took up ~
3x more carbon (an estimated 100 g C m-2
season-1) compared to EG canopies
(P < 0.001)
Deciduous Shrub Dominance
u greater deciduous shrub cover = greater NEE
during the green season (R2 = 0.8; P < 0.001)
and the peak season (R2 = 0.9; P < 0.001)
Changing Season Length (Table 1)
u Extending seasons increased C uptake, but
more significantly for peak season & more
for DS
~ 29% of additional C uptake by deciduous shrub communities during the
peak season was due to greater leaf area
~ 71% of additional C uptake was due to the longer peak season of deciduous
shrub communities compared to evergreen/graminoid communities
Shannan K. Sweet1, Natalie T. Boelman1, Kevin L. Griffin1, Heidi Steltzer2, Laura Gough3
1 Lamont-Doherty Earth Observatory, Department of Earth and Environmental Sciences, Columbia University; 2 Department of Biology, Fort Lewis College, Colorado; 3 Department of Biology, University of Texas, Arlington
Introduction
Greater deciduous shrub abundance extends tundra peak season and increases modeled net CO2 uptake * accepted at Global Change Biology
Acknowledgements
Fig. 5. Seasonal canopy greenness for deciduous shrub (DS) and evergreen/graminoid (EG) canopies. On the
respective NDVI curves, dates when canopy phenology metrics were reached are marked with red (DS) and blue
(EG) points, with arrows of matching colors extended to the x-axis. Canopy phenology parameters are indicated by
number boxes on the x-axis: [1] onset of greening; [2] onset of peak green; and [3] onset of senescence.
Green-uprate
(slope)
Onsetpeakgreen
(DOY)
Peakseasonlength
(#days)
Deciduous shrub cover (%)
Fig. 6.
Relationship
between
deciduous
shrub cover in
all 18 (9 DS +
9 EG) quadrats
and the (a) rate
of green-up, (b)
onset of peak
green, and (c)
peak green
season length.
175185195
R! = 0.8
17518519530405060
R! = 0.7
304050600.0050.0150.025
R! = 0.7
0.0050.0150.025
• Satellite studies suggest longer growing seasons and earlier
peak seasons due to increasing air temperature & earlier
spring snowmelt in some Arctic regions
• Longer growing season enhance carbon (C) uptake by
tundra vegetation, especially during peak season when C
uptake is at its maximum
(c)
(b)
(a)
Change in spring snow-cover
duration (1973 – 2009), days
Fig. 7. Seasonal (a) leaf area index and (b) net ecosystem exchange of
deciduous shrub-dominated (DS) and evergreen/graminoid-dominated (EG)
canopies. Dates when phenology metrics were reached are marked with red
(DS) and blue (EG) symbols {see Fig. 5}
Table 1. Seasonal average NEE (+/- 1 SEM) for deciduous shrub (DS) and evergreen/
graminoid (EG) canopies calculated under modeled season length scenarios: (a) DS green
season (63 days); EG green season (68 days); (b) DS peak season (44 days); and EG peak
season (34 days). Percentages indicate additional carbon uptake during the longer
seasons (i.e. 5-day longer green & 10-day longer peak season). {See Fig. 5 for phenology
dates}
(a)
DS Green
Season (63 d)
EG Green
Season (68 d)
%
Change
P
DS NEE
(grams C m-2
season-1
)
− 221 ± 10 − 228 ± 10 3% ns
EG NEE
(grams C m-2
season-1
)
− 103 ± 16 − 107 ± 16 4% ns
(b)
DS Peak
Season (44 d)
EG Peak
Season (34 d)
DS NEE
(grams C m-2
season-1
)
− 156 ± 5 − 85 ± 4 84% < 0.01
EG NEE
(grams C m-2
season-1
)
− 77 ± 12 − 47 ± 7 64% < 0.05
Change in air T (2005 – 2009,
baseline 1951 – 2000), deg-C
Change in deciduous shrub
cover (~ 1990s – 2010)
AMAP 2012
Increases No change
each day the growing
season is extended in
the arctic tundra
net carbon uptake
increases by
5.3 g C m-2 yr-1
Euskirchen et al. 2006
AMAP 2012
Zeng et al.
2011
Change in start of growing
season (2000-2010), days
Myers-Smith et
al. 2010
• Peak season phenology (i.e. peak season length) & how
indirect impacts of arctic warming (i.e. increasing
deciduous shrub cover) may be contributing to satellite
signals of changing regional phenology are less well
understood
Our Questions
We are interested in this length VS. this length
10 20 30 40 50 60
SITES:
u Data collected 1 Jun to 6 Sep 2013 at two
sites, North Slope of Alaska (Fig. 1)
u Each site had two plots (each plot had two
100-m transects):
• EG (dominated by evergreens &
graminoids)
• DS (dominated by deciduous shrubs)
• Chose 18 1-m2 quadrats along transects (n =
9 DS & 9 EG) based on functional group
dominance (Fig. 2) & because 18 equipment
assemblages were available (Fig. 3)
Fig. 1. Map of Alaska & North Slope of
Brooks Range (inset). Field sites are stars.
1
32
Day of year
NDVI
Fig. 4.
Model of
seasonal
canopy
phenology
B
Am
Fig. 3. Equipment
assemblage; collected
reflectance and
temperature data used
to calculate NDVI and
NEE (n = 18)
Fig. 2. Mean percent cover (%) of
deciduous shrubs and evergreens +
graminoids at each site & canopy type.
Different letters = significantly different
(P < 0.05).
Methods
NET ECOSYSTEM EXCHANGE MODEL:
u Modeled:
1) Leaf area index (LAI) {Eq. 2} using tundra specific model from Street et al. (2007)
2) Net ecosystem exchange (NEE) {Eq. 3}, using model from Shaver et al. (2013) &
low Arctic site parameters for respiration (RE) and gross primary production (GPP)
u Examined effect of changing LAI & season length across DS & EG canopies by
integrating NEE for green and peak seasons:
1) Using respective season lengths of each canopy
2) Imposing average season lengths of each canopy type on one another
CANOPY PHENOLOGY:
u Calculated normalized difference vegetation index (NDVI) {Eq. 1}
u Used piecewise regressions to determine canopy phenology (Fig. 4-circles):
① onset of greening (start rapid increase in NDVI)
② onset of peak green (level out of NDVI)
③ onset of senescence (start rapid decrease of NDVI)
u Calculated season lengths (Fig. 4-squares):
A peak season = onset peak green to onset senescence
B green season = onset greening to onset senescence
u Calculated green-up rates (Fig. 4-diamond):
m = slope à daily rate of change in NDVI
Satellite studies of the Arctic report longer growing and peak seasons since the 1980s, which lengthens the period of carbon uptake. These
trends are attributed to increasing air temperatures and reduced snow cover duration in spring and fall. Concurrently, deciduous shrubs are
becoming increasingly abundant in tundra landscapes, which may also impact canopy phenology and productivity.
To examine the impact of increasing deciduous shrub dominance on growing season phenology, season length, and carbon (C) flux we
monitored canopy phenology using the normalized difference vegetation index (NDVI) and modeled leaf area (LAI) and net ecosystem C
exchange (NEE) across a gradient of deciduous shrub- and evergreen/graminoid-dominated tundra . We found that deciduous shrub
canopies reached the onset of peak greenness 13 days earlier and the onset of senescence 3 days earlier compared to evergreen/graminoid
canopies, resulting in a 10-day extension of the peak season. The combined effect of the longer peak season and greater leaf area of
deciduous shrub canopies almost tripled C uptake of deciduous shrub communities compared to evergreen/graminoid communities, while
the longer peak season alone resulted in 84% greater C uptake in deciduous shrub communities. These results suggest that greater deciduous
shrub abundance increases C uptake not only due to greater leaf area, but also due to an extension of the period of peak greenness, which
extends the period of maximum C uptake.
Jia et al. 2009
Peak green date in two common arctic subzones
Key Findings
Abstract
{Eq. 1} NDVI = (NIR - R) / (NIR + R) {Eq. 2} LAI = a * eb*NDVI
{Eq. 3} NEE = RE – GPP