PUBLICATIONS
Geophysical Research Letters
RESEARCH LETTER
10.1002/2014GL060495
Key Points:
• Longer nonfrozen periods linked to
summer greenness decline across NA
• Drought indicator and precipitation
pattern confirm summer
drought signal
• Seasonal shift in hydrology is a leading
drought factor in northern ecosystems
Increasing summer drying in North American
ecosystems in response to longer
nonfrozen periods
Bikash R. Parida1 and Wolfgang Buermann2
1
Department of Civil Engineering, Shiv Nadar University, Dadri, India, 2Institute for Climate and Atmospheric Science,
School of Earth and Environment, University of Leeds, Leeds, UK
Abstract In snow-dominated northern ecosystems, spring warming is predicted to decrease water availability
Supporting Information:
• Readme
• Figures S1–S7
Correspondence to:
W. Buermann,
w.buermann@leeds.ac.uk
Citation:
Parida, B. R., and W. Buermann (2014),
Increasing summer drying in North
American ecosystems in response to
longer nonfrozen periods, Geophys. Res.
Lett., 41, doi:10.1002/2014GL060495.
Received 9 MAY 2014
Accepted 16 JUL 2014
Accepted article online 18 JUL 2014
later in the season and recent findings suggest that corresponding negative impacts on plant productivity
and wildfire frequency are already observable. Here we estimate the overall vulnerability of North American
ecosystems to warming-related seasonal shifts in hydrology through identifying robust interannual linkages
between nonfrozen periods, peak-to-late summer vegetation greenness, and an indicator of drought for
1982–2010. Our results show that longer nonfrozen periods earlier in the year are persistently associated with
declines in peak-to-late summer greenness and moisture availability across large portions of North America.
Hereby, vulnerabilities increase markedly across the dominant land covers with decreasing annual precipitation
rates, lowering contributions of summer rainfall, and increasing altitude. The implications are that in a warmer
world, seasonal hydrological shifts may emerge as a leading factor for summer drought in relatively dry
temperate-forested ecosystems and across the northern high latitudes.
1. Introduction
Across North American (NA) temperate and high-latitude regions, rising temperatures over roughly the last
five decades are characterized by a distinct seasonal pattern, with most of the warming being observed in
winter and spring [Karl et al., 1996; Bonsal et al., 2001; Cayan et al., 2001; Hengeveld et al., 2005]. This spring
warming has led to increases in the length of the nonfrozen period (inferred from temperature records) by
approximately 2–4 d/decade since 1950 [Easterling, 2002; Feng and Hu, 2004]. A newly available long-term
(1979–2010) freeze-thaw record based on satellite microwave measurements shows also a lengthening of
the average nonfrozen period in NA by 3 d/decade which is largely driven by earlier spring thawing
consistent with the seasonal temperature trends [Kim et al., 2012].
While generally a lengthening of the growing season across northern latitudes may be beneficial for
vegetation productivity [Xu et al., 2013] and carbon sequestration [Graven et al., 2013], earlier spring
onsets and snowmelt in conjunction with summer warming are thought to be the leading factors for a
precipitous increase in wildfires in the western U.S. and Canada in the last three decades [Stocks et al.,
2002; Westerling et al., 2006]. The earlier onset of spring alters seasonal hydrology and vegetation
phenology pattern through early runoff, increased evapotranspiration, and increased snow sublimation
[Barnett et al., 2005; Jepsen et al., 2012] with the combined effects leading to drier conditions around the
midst of the growing season.
Recent findings suggest that this mechanism for summer drought operates very efficiently also across NA
boreal forests, where over large portions consistent interannual covariations between earlier spring onset
and reduced peak-to-late summer plant growth have been observed [Buermann et al., 2013]. However, the
summer drought signal associated with an earlier spring onset has not been verified yet with soil moisture or
drought indicators. Further, we also posit that this spring onset-summer drought mechanism may operate at
continental scales across most of northern snow-dominated ecosystems rather than at regional scales or
specific to a particular land cover, since many of northern ecosystems rely on winter snowpacks as an
important source of water [Barnett et al., 2005].
In this study, we therefore examine the impact of longer nonfrozen periods during roughly the first half of the
year on subsequent summer ecohydrological conditions across all of NA temperate and high-latitude
nonmanaged ecosystems. Hereby, we exploit long-term (1982–2010) microwave and optical satellite records
PARIDA AND BUERMANN
©2014. American Geophysical Union. All Rights Reserved.
1
Geophysical Research Letters
10.1002/2014GL060495
to infer nonfrozen periods and vegetation greenness and also make use of an updated drought metric that
is driven by climate input data (see section 2). A particular focus is to assess the vulnerability of specific
land covers to summer drying associated with longer nonfrozen periods and how precipitation pattern
(annual totals and seasonal distribution) and elevation modulate this vulnerability.
2. Data and Methods
For this study, we analyzed a diverse set of ground- and satellite-based gridded data that represent vegetation
greenness and hydrological surface conditions for the study period 1982 to 2010. For satellite vegetation
greenness, we obtained an improved and updated version (3g) of the advanced very high resolution
radiometer-based bimonthly normalized difference vegetation index (NDVI) data at 8 km spatial resolution
[Pinzon and Tucker, 2014]. NDVI is calculated from optical near-infrared and red surface reflectances and
commonly used as a surrogate measurement of plant photosynthetic activity [Myneni et al., 1995]. Nonfrozen
periods of the land surface were estimated from a daily freeze-thaw record based on measurements from the
satellite scanning multichannel microwave radiometer (SMMR) and Special Sensor Microwave Imager (SSM/I) at
25 km spatial resolution [Kim et al., 2012]. The length of the nonfrozen period (NFP) during the roughly first half
of the year was determined through counting the nonfrozen days from January to August (denoted as
NFPJanAug). This seasonality indicator captures also relatively short-term freezing events even after spring
onset and thus may capture related effects on surface hydrology even better than the timing of spring
onset. In addition, we also obtained monthly relatively high-resolution (0.5°) gridded precipitation and
temperature fields from the Climatic Research Unit (CRU TS3.21) [Harris et al., 2013]. Further, land cover-specific
investigations are based on the MOD12C1 land cover classification at 5 km spatial resolution [Friedl et al., 2002].
Since our focus is snow-dominated regions, we included in our analyses only vegetated nonmanaged NA
ecosystems in the latitude band from 30°N to the northern limit (Figure S1 in the supporting information).
Last, for drought characterization we obtained monthly Palmer Drought Severity Index (PDSI) data at 0.5°
spatial resolution [Zhao and Running, 2010]. Input data for the PDSI include precipitation data from the Global
Historical Climatology Network [Chen et al., 2002], evaporation demand (National Centers for Environmental
Prediction/Department of Energy II), and soil water holding capacity as a function of soil texture. A lower
(higher) PDSI value is indicative of a drier (wetter) climate, respectively. It should be noted that this drought
index does not account explicitly for the effects of seasonal changes in snow melt and runoff as well as
vegetation activity [Dai, 2011] and thus may not fully capture the soil moisture status during the peak of the
growing season. Given these caveats, we evaluated the consistency of the PDSI against an ensemble of soil
moisture outputs from a set of more complex hydrological models that participated in the integrated project
water and global change (WATCH) [Weedon, 2011]. Corresponding results show the widespread robust
(positive) interannual covariations between summer soil moisture (from WATCH) and PDSI over our study
region suggesting the PDSI provides a credible estimate of moisture status (Figure S2 in the supporting
information). Furthermore, several studies have shown reasonably good agreement between the PDSI and
observed stream flow as well as soil moisture during warm seasons [Dai, 2011].
All satellite (NDVI, land cover) and climate (e.g., PDSI) fields were aggregated (pixel aggregate)/downscaled
(nearest neighbor) to a common 0.25° spatial grid on which all statistical analyses were performed. With a
focus on identifying interannual linkages, original data were detrended prior to computing (Pearson’s) grid
point correlations. No single optimal solution for removing trends in these relatively short and often complex
satellite and climate time series exists [Zhou et al., 2001]. We, therefore, used two different detrending
methods for assessing the robustness of the interannual covariation patterns: (i) removal of linear trends
(determined through least squares linear regression) over the entire study period 1982–2010 and (ii) first
differences determined through changes in successive years. Student’s t tests are used throughout to
evaluate statistical significance of correlations.
3. Results
In the first step, we explored patterns in the interannual covariations between nonfrozen periods during the
roughly first half of the year (NFPJanAug) and subsequent peak-to-late summer NDVI for 1982–2010.
Corresponding results document continental-scale negative correlation patterns over NA ecosystems
(Figures 1a and 1c), implying that longer nonfrozen periods during the first half of the year are consistently
PARIDA AND BUERMANN
©2014. American Geophysical Union. All Rights Reserved.
2
Geophysical Research Letters
10.1002/2014GL060495
Figure 1. Interannual covariations between the length of nonfrozen period during the roughly first half of the year and subsequent summer greenness as well as
drought. Maps show grid point correlations for the period 1982–2010 between the NFPJanAug and peak-to-late summer (a, c) NDVI and (b, d) PDSI. For the
northern high latitudes (Figures 1a and 1b) peak-to-late summer is defined as July through August, whereas for the temperate zones (Figures 1c and 1d) this epoch
is defined as August through September. At each grid point, original data have been detrended via removal of linear trends for 1982–2010 prior to correlations.
Absolute Pearson correlation values greater than 0.31 are statistically significant (p < 0.1).
associated with declines in peak-to-late summer vegetation greenness and vice versa. Increased
vulnerabilities (more robust negative correlation patterns) are evident over the climatologically drier regions
of western U.S. and the interior of Alaska as well as across western and interior Canada. In contrast,
positive correlations (peak-to-late summer greenness benefits from longer nonfrozen epochs and vice
versa) are spatially more confined in regions surrounding the Hudson Bay, the U.S. Great Plains as well
as along the U.S. East coast and the northern portion of the NA West coast.
Our working hypothesis is that these observed relationships between longer nonfrozen periods in the first
half of the year and decreased peak-to-late summer greenness can be explained by reduced soil moisture
resulting from increased evapotranspiration and early runoff earlier in the growing season. To further and
PARIDA AND BUERMANN
©2014. American Geophysical Union. All Rights Reserved.
3
Geophysical Research Letters
10.1002/2014GL060495
independently test this assertion, we also analyzed interannual covariations between the NFPJanAug and
peak-to-late summer PDSI for 1982–2010. Results also show the widespread negative correlation patterns
(Figures 1b and 1d) indicating that longer (shorter) nonfrozen periods during the first half of the year are
consistently linked to peak-to-late summer water deficits (surpluses). These negative patterns are broadly
colocated with those based on the covariations between the NFPJanAug period and peak-to-late summer NDVI
(Figures 1a and 1c). Further, across these coinciding patterns the PDSI and NDVI fields during peak-to-late
summer show also strong positive interannual covariations (Figure S3 in the supporting information) in line
with the notion that the PDSI does capture the soil moisture status in this season to some extent, which in
turn exerts a strong climatic constraint on plant growth. In contrast and as one may expect, over the
climatologically wetter more eastern NA regions where peak-to-late summer greenness benefits from longer
nonfrozen periods earlier in the year (Figures 1a and 1c), the PDSI does not serve as a robust predictor of
interannual NDVI variability during peak-to-late summer (Figure S3 in the supporting information).
A noteworthy and exceptional pattern emerges over portions of the U.S. Great Plains (~100°–110°W and 45°–50°N).
Over these relatively dry (total annual rainfall ~300 mm; Figure S4 in the supporting information) vast
grassland regions, the PDSI appears to be a reliable predictor of interannual variations in NDVI during
peak-to-late summer (Figure S3 in the supporting information), and the year-to-year covariations of
these two variables with the length of nonfrozen periods in the first half of the year are also both
positive and spatially colocated (Figures 1c and 1d). Further analyses over these regions (not shown)
show that longer nonfrozen periods in the early to middle growing season are associated with
increased summer rainfall and, hence, increased PDSI as well as NDVI (giving rise to the observed
positive correlation between NFPJanAug and NDVI as well as PDSI, respectively). These peculiar links
may be related to increased water cycling in years with longer nonfrozen periods early in the year
over the U.S. Great Plains, a region which is known for their strong coupling between soil moisture
and precipitation [Koster et al., 2004].
The robustness of these spatial correlation patterns with a focus on interannual timescales was assessed by
repeating the analysis with an alternative method for removing trends in the original data (see section 2).
Results show that the patterns are generally robust against method (Figure 1 and Figure S5 in the supporting
information), except for NFPJanAug-PDSI correlations in the northern high latitudes where some deviations
can be observed. However, across these more northern regions the PDSI is also known to be a less reliable
drought indicator, since processes that may be more important such as frozen soil, snow accumulation, and
melt are not considered [Heim, 2002; Dai et al., 2004].
In a next step, we explored the vulnerability of specific NA land covers to this “summer drought mechanism”
and how the annual total rainfall rates and seasonal distributions influence this vulnerability. Within
the NA boreal zone (>50°N), dominant land covers with relatively large vulnerable areal proportions
(Figures 2a–2d), include evergreen needleleaf forests (~20–24% areal proportions), mixed forests
(~22–27%), and woody savannas (~29–31%). Northern open shrublands (also known as tundra) show a
lower vulnerable areal proportion (~10%), but equal evergreen needleleaf forests in terms of total
vulnerable area because of their vast geographic distribution (Figures 2a and 2c). Within the
temperate regions, the most vulnerable land covers in regard to both areal proportions and total
vulnerable area (Figures 2e–2h) are evergreen needleleaf forests (~27–31% areal proportions) and
grasslands (~14–18%). Mixed forests (~7–12%) and open shrublands (~9–17%) showing a somewhat
lower vulnerability.
Scatterplots of summer (JJA) rainfall proportions against the annual total precipitation show that across most
of the vulnerable regions within the dominant NA land covers higher clustering occurs when the annual total
precipitation is typically below 500 mm and when the proportion of summer precipitation to total annual
rainfall is below 50% (Figure 2 and Figure S4 in the supporting information). In cases where these summer
rainfall proportions are comparably low (~25%), vulnerability for single land covers is also prevalent even in
relatively wet regions (e.g., boreal and temperate needleleaf forests as well as temperate mixed forests and
grasslands). At the other extreme, most boreal land covers in relatively dry regions (annual total rainfall
>250 mm) show high vulnerability despite receiving most of their annual rainfall in the summer months.
In the final step, we explored the influence of elevation on the efficacy of the summer drought mechanism
with linkage to longer nonfrozen periods, since soil storage capacities may be influenced by topographic
PARIDA AND BUERMANN
©2014. American Geophysical Union. All Rights Reserved.
4
10.1002/2014GL060495
100
a) ENF
19.7% (23.7%)
75
2.44 mill. km2
50
25
0
0
500
1000
1500
Fraction of summer PREC (%)
Fraction of summer PREC (%)
Geophysical Research Letters
100
b) MF
0.30 mill. km2
50
25
0
0
500
100
9.6% (9%)
75
4.69 mill. km2
50
25
0
0
500
1000
1500
d) WSA
0.93 mill. km2
25
0
500
1000
1500
2000
Fraction of summer PREC (%)
Fraction of summer PREC (%)
30.7% (26.6%)
50
0
75
0.60 mill. km2
50
25
0
0
500
2
50
25
0
500
1000
1500
Fraction of summer PREC (%)
Fraction of summer PREC (%)
9.3% (16.6%)
1.17 mill. km
0
1500
f) MF
7.2% (11.7%)
75
1.12 mill. km2
50
25
0
0
500
1000
1500
2000
Annual total PREC (mm)
100
75
1000
100
Annual total PREC (mm)
g) OSH
29% (31.2%)
Annual total PREC (mm)
100
75
1500
100
Annual total PREC (mm)
e) ENF
1000
Annual total PREC (mm)
Fraction of summer PREC (%)
Fraction of summer PREC (%)
Annual total PREC (mm)
c) OSH
21.6% (27%)
75
100
h) GRA
14.1% (17.7%)
75
2.07 mill. km2
50
25
0
0
Annual total PREC (mm)
500
1000
1500
Annual total PREC (mm)
Figure 2. Land cover and precipitation influence on vulnerability to summer drying linked to longer nonfrozen periods. For
each dominant land cover (Figure S1), panels depict scatterplots of the annual total precipitation (PREC) versus ratio of
summer (June-July-August (JJA)) precipitation over the annual total precipitation, based on the climatology for the period
1982–2010. (a–d) NA boreal land covers (50°N–75°N); and (e–h) temperate land covers (30°N–50°N). For each land cover,
total area and two estimates of vulnerable areal proportions (%) are also provided in the insets. Here vulnerable areas are
defined as those that exhibit significant (p < 0.1) negative interannual covariations between the NFPJanAug and peak-tolate summer NDVI (see Figures 1a and 1c, and Figures S5a and S5c in the supporting information). The two estimates of
vulnerable areal proportions are based on two methods of removing trends in the original data prior correlations analyses
(see section 2; values in brackets are based on the first differences method). Land covers are abbreviated as evergreen
needleleaf forest (ENF), mixed forests (MF), open shrublands (OSH), woody savannas (WSA), and grasslands (GRA).
Corresponding results for deciduous broadleaf forests (DBF) and WSA in the temperate zones are not shown since either
2
vulnerable areal proportions (1–6% for DBF) or total vulnerable area are less extensive (0.03–0.06 million km for WSA).
variations [Garcia and Tague, 2014]. The results show that vulnerable areal proportions as a function of
altitude reflect unimodal distributions for most dominant NA boreal land covers, but corresponding peak
vulnerabilities depend on land cover (Figure 3a). For example, evergreen needleleaf forests appear more
vulnerable at altitudinal ranges between 600 and 900 m, whereas for woody savannas (900–1200 m) and
PARIDA AND BUERMANN
©2014. American Geophysical Union. All Rights Reserved.
5
Geophysical Research Letters
10.1002/2014GL060495
70
a)
Vulnerable areal
proportions (%)
60
ENF
MF
OSH
WSA
50
40
30
20
10
0
0
300
600
900
1200
1500
1800
2100
2400
Elevation (m)
70
b)
Vulnerable areal
proportions (%)
60
GRA
50
40
30
20
10
0
0
300
600
900
1200 1500 1800 2100 2400 2700
Elevation (m)
Figure 3. Elevational influence on vulnerability to summer drying linked
to longer nonfrozen periods. Panels display vulnerable areal proportions,
defined as in Figure 2, for successive 300 m elevational intervals and
dominant NA land covers, located in the (a) boreal (50°N–75°N) and (b)
temperate (30°N–50°N) zones, respectively.
open shrublands (1200–1500 m), these
levels are shifted toward higher
elevations. Across vulnerable temperate
land covers, only evergreen needleleaf
forests show a similar elevational
dependency as observed in the boreal
zones, but peak vulnerability levels at
these lower latitudes are markedly
shifted toward higher elevations
(~1200–1500 m). For grasslands and
open shrublands, the efficiency of the
drought mechanism increases gradually
at altitudes above ~1200 m and more
sharply at relatively high altitudes
(>2100 m). Corresponding results for
very high elevations (>2400 m) should
be interpreted with caution since the
areal extent of land covers at these
altitudes is very small and both the
satellite optical (used to infer
vegetation greenness) and microwave
(for nonfrozen periods) data are less
reliable because of topographic
shading effects in very high mountain
regions [e.g., Kim et al., 2012].
4. Discussion and Conclusion
The overall good agreement in the spatial patterns pertaining to the interannual covariations between longer
nonfrozen periods during roughly the first half of the year and reduced peak-to-late summer vegetation
greenness as well as water availability may serve as the first line of hydrologic evidence that these joint large-scale
linkages are associated with a summer drought phenomena. This key finding is even more intriguing since the
exploited drought indicator may underestimate the true soil moisture status in summer substantially as it does not
encapsulate information on seasonal changes in snow melt, runoff, vegetation phenology, and root development
[Dai et al., 2004]. In line with these results, a few field-based studies (based on eddy covariance flux towers) have
reported declines in annual net ecosystem productivity owing to temporal shifts to earlier spring onsets and snow
melt that led to increases in early season runoff and evapotranspiration and decreasing water availability in the
later part of the growing season [Welp et al., 2007; Sacks et al., 2007; Hu et al., 2010].
Our results also suggest that forested ecosystems, including temperate and boreal evergreen needleleaf forests as
well as boreal mixed forests and woody savannas located predominantly in the western climatologically drier
portion of NA, represent the land covers that are more vulnerable to this summer drought mechanism. However,
corresponding vulnerabilities are also widespread among more open ecosystems including temperate grasslands
and northern tundra. Within these land covers, annual amount and seasonal distribution of rainfall rates are strong
determinants of vulnerability. We find that vulnerability increases with lower annual total rainfall and lower
proportions of summer rainfall as perhaps anticipated since these relations indicate overall drought vulnerability
and a greater reliance on winter snow packs as a source of water [Barnett et al., 2005]. In fact, the existence of such
dependencies provides another independent hydrologic-based evidence for this summer drought mechanism
with links to longer nonfrozen periods earlier in the year.
One may hypothesize that variations in summer rainfall have a greater influence on peak-to-late summer
greenness and moisture status than those associated with nonfrozen period variability during roughly the
first half of the year. A corresponding supplementary analysis shows that the (positive) influence of summer
rainfall on peak-to-late summer greenness status is particularly strong across the temperate grasslands and
open shrublands (Figure S6 in the supporting information). Over the northern high latitudes and temperate
more forested ecosystems, however, the influence of summer rainfall is substantially weaker (and even its
PARIDA AND BUERMANN
©2014. American Geophysical Union. All Rights Reserved.
6
Geophysical Research Letters
10.1002/2014GL060495
direction shifts from positive to negative) and is comparable to the one linked to early to middle season
nonfrozen periods (Figure 1 and Figure S6 in the supporting information). These perhaps not anticipated
results could be explained by a number of factors, one being related to the adverse influence of cloud cover
(that accompanies rainfall) on plant growth, a known colimiting factor in these regions [Nemani et al., 2003].
Another potential factor is the observation that over the northern high latitudes, the often abundant summer
rainfall (but also note that the NA high latitudes are climatologically relatively dry areas; see Figure S4 in the
supporting information) does aid in replenishing soil moisture near the surface but may be less effective in
recharging deeper soil layers (where trees generally have their roots) because a large fraction is lost to (re)
evaporation [Barichivich et al., 2014]. Summer rainfall exclusion experiments conducted in interior Alaska also
show that tree growth in well-drained upland forests is sustained primarily by snowmelt water and is not
substantially affected by summer rainfall deficits [Yarie, 2008].
Further, our results also show that elevation has a significant influence on the vulnerability of the dominant NA
land covers to this summer drought mechanism, consistent with earlier findings at more regional scales [Westerling
et al., 2006]. The general pattern is that vulnerability increases with increasing altitudes up to a certain level,
whereby for temperate latitudes, regions of peak vulnerability stretch over substantially higher elevational ranges.
These results are broadly consistent with reduced soil storage capacities due steeper slopes and well-drained
surfaces at higher elevations [Garcia and Tague, 2014], whereas at very high elevations (especially in the boreal
zone), low temperature constrains on plant growth may become a more dominant factor.
In an annual sense, vegetation growth in snow-dominated northern ecosystems is generally considered
temperature and radiation limited [Nemani et al., 2003]. As a result, rapid warming in northern mid- and highlatitudes throughout the last several decades appear to have led to longer growing seasons, increased
photosynthetic activity and carbon sequestration [Randerson et al., 1999; Tanja et al., 2003; Richardson et al., 2010;
Xu et al., 2013; Graven et al., 2013]. Our results show that summer drying and reduced photosynthetic activity in
response to longer nonfrozen periods earlier in the year can be robustly detected in observations covering the
last three decades. Because this summer drought mechanism acts at continental scales, it is not unlikely that this
factor has substantially modulated the general response of northern ecosystems to recent warming trends.
Increasing heat and drought stress on the world’s forests is presently emerging as global scale phenomena
[Allen et al., 2010], and these factors are also key in responses of ecosystems under future climate projections
[Sitch et al., 2008; Williams et al., 2013]. In this regard, the significance of summer drying with linkages to
longer nonfrozen periods specifically in the early to middle growing season across northern ecosystems
remains an open question. On one hand, there appears to be a direct link between spring warming,
associated longer nonfrozen periods (Figure S7 in the supporting information) and summer drying. One the other
hand, this drought mechanism acts specifically toward reducing peak-to-late summer plant growth and some
ecosystems may be more resilient to such relative short episodes of drought. Arguably the largest impact will
be on high-altitude needleleaf-dominated forests in climatologically relative dry temperate and boreal zones
where increasing peak-to-late summer drying may influence important disturbance regimes (e.g., fires and insects)
which in turn may lead to rapid declines in forest extent.
Acknowledgments
This research is funded by a grant from
the National Aeronautics and Space
Administration Carbon Cycle Science
Program (grant NNX11AD45G). We are
indebted to John Kimball group for
providing updated freeze/thaw record
from the SMMR and SSM/I sensors. We
thank Shiv Nadar University, India, for
providing computing facilities.
The Editor thanks two anonymous
reviewers for their assistance in
evaluating this paper.
PARIDA AND BUERMANN
In respect to carbon sequestration, northern forested ecosystems are credited as a sustained land-based carbon
sink for the contemporary period (1990–2007) [Pan et al., 2011]. However, longer nonfrozen periods owing largely
to warmer springs may pose an additional constraint on the future carbon sink in conjunction with projected
declines in summer precipitation over midlatitude regions of the Northern Hemisphere [Rowell, 2009; Christensen
et al., 2013]. The influence of this “longer nonfrozen period and summer drought” mechanism on the carbon
balance of northern ecosystems needs further study with coupled climate-carbon cycle models.
References
Allen, C. D., et al. (2010), A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests,
For. Ecol. Manage., 259, 660–84.
Barichivich, J., K. R. Briffa, R. Myneni, G. van der Schrier, W. Dorigo, C. J. Tucker, T. J. Osborn, and T. M. Melvin (2014), Temperature and snowmediated moisture controls of summer photosynthetic activity in northern terrestrial ecosystems between 1982 and 2011, Remote Sens.,
6, 1390–1431.
Barnett, T. P., J. C. Adam, and D. P. Lettenmaier (2005), Potential impacts of a warming climate on water availability in snow-dominated regions,
Nature, 438, 303–309, doi:10.1038/nature04141.
Bonsal, B. R., X. Zhang, L. A. Vincent, and W. D. Hogg (2001), Characteristics of daily and extreme temperatures over Canada, J. Clim., 14,
1959–1976.
©2014. American Geophysical Union. All Rights Reserved.
7
Geophysical Research Letters
10.1002/2014GL060495
Buermann, W., P. R. Bikash, M. Jung, D. H. Burn, and M. Reichstein (2013), Earlier springs decrease peak summer productivity in North American
boreal forests, Environ. Res. Lett., 8, 024027, doi:10.1088/1748-9326/8/2/024027.
Cayan, D. R., S. A. Kammerdiener, M. D. Dettinger, J. M. Caprio, and D. H. Peterson (2001), Changes in the onset of spring in the western United States,
Bull. Am. Meteorol. Soc., 82, 399–416.
Chen, M., P. Xie, J. E. Janowiak, and P. A. Arkin (2002), Global land precipitation: A 50-yr monthly analysis based on gauge observations,
J. Hydrometeorol., 3, 249–266.
Christensen, J. H., et al. (2013), Climate phenomena and their relevance for future regional climate change, in Climate Change 2013: The
Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change,
chap. 14, edited by T. F. Stocker et al., pp. 1217–1308, Cambridge Univ. Press, New York.
Dai, A. (2011), Characteristics and trends in various forms of the Palmer Drought Severity Index during 1900–2008, J. Geophys. Res., 116, D12115,
doi:10.1029/2010JD015541.
Dai, A., K. E. Trenberth, and T. Qian (2004), A global dataset of Palmer Drought Severity Index for 1870–2002: Relationship with soil moisture
and effects of surface warming, J. Hydrometeorol., 5, 1117–1130.
Easterling, D. R. (2002), Recent changes in frost days and the frost-free season in the United States, Bull. Am. Meteorol. Soc., 83, 1327–1332.
Feng, S., and Q. Hu (2004), Changes in agro-meteorological indicators in the contiguous United States: 1951–2000, Theor. Appl. Climatol., 78,
247–264.
Friedl, M., et al. (2002), Global land cover mapping from MODIS: Algorithms and early results, Remote Sens. Environ., 83, 287–302.
Garcia, E. S., and C. L. Tague (2014), Climate regime and soil storage capacity interact to effect evapotranspiration in western United States
mountain catchments, Hydrol. Earth Syst. Sci. Discuss., 11, 2277–2319.
Graven, H. D., et al. (2013), Enhanced seasonal exchange of CO2 by northern ecosystems since 1960, Science, 341, 1085–1089.
Harris, I., P. Jones, T. Osborn, and D. Lister (2013), Updated high-resolution grids of monthly climate observations—The CRU TS3.10 dataset,
Int. J. Climatol., 623–642, doi:10.1002/joc.3711.
Heim, R. R., Jr. (2002), A review of twentieth-century drought indices used in the United States, Bull. Am. Meteorol. Soc., 83, 1149–1165.
Hengeveld, H., B. Whitewood, and A. Fergusson (2005), An Introduction to Climate Change: A Canadian Perspective, 55 pp., Environment
Canada, Downsview, Ontario.
Hu, J., D. J. P. Moore, S. P. Burns, and R. K. Monson (2010), Longer growing seasons lead to less carbon sequestration by a subalpine forest,
Global Change Biol., 16, 771–783.
Jepsen, S. M., N. P. Molotch, M. W. Williams, K. E. Rittger, and J. O. Sickman (2012), Interannual variability of snowmelt in the Sierra Nevada and
Rocky Mountains, United States: Examples from two alpine watersheds, Water Resour. Res., 48, W02529, doi:10.1029/2011WR011006.
Karl, T., R. Knight, D. Easterling, and R. Quayle (1996), Indices of climate change for the United States, Bull. Am. Meteorol. Soc., 77, 279–292.
Kim, Y., J. S. Kimball, K. Zhang, and K. C. McDonald (2012), Satellite detection of increasing Northern Hemisphere non-frozen seasons from
1979 to 2008: Implications for regional vegetation growth, Remote Sens. Environ., 121, 472–487.
Koster, R. D., et al. (2004), Regions of strong coupling between soil moisture and precipitation, Science, 305, 1138–1140.
Myneni, R. B., F. G. Hall, P. J. Sellers, and A. L. Marshak (1995), The meaning of spectral vegetation indices, IEEE Trans. Geosci. Remote Sens., 33,
481–486.
Nemani, R. R., C. D. Keeling, H. Hashimoto, W. M. Jolly, S. C. Piper, C. J. Tucker, R. B. Myneni, and S. W. Running (2003), Climate-driven increases
in global terrestrial net primary production from 1982 to 1999, Science, 300, 1560–1563.
Pan, Y., et al. (2011), A large and persistent carbon sink in the world’s forests, Science, 333(6045), 988–993, doi:10.1126/science.1201609.
Pinzon, J. E., and C. J. Tucker (2014), A non-stationary 1981–2012 AVHRR NDVI3g time series, Remote Sens., 6(8), 6929–6960.
Randerson, J. T., C. B. Field, I. Y. Fung, and P. P. Tans (1999), Increases in early season ecosystem uptake explain recent changes in the seasonal
cycle of atmospheric CO2 at high northern latitudes, Geophys. Res. Lett., 26, 2765–2768, doi:10.1029/1999GL900500.
Richardson, A. D., et al. (2010), Influence of spring and autumn phonological transitions on forest ecosystem productivity, Philos. Trans. R. Soc. B,
365, 3227–3246, doi:10.1098/rstb.2010.0102.
Rowell, D. P. (2009), Projected midlatitude continental summer drying: North America versus Europe, J. Clim., 22, 2813–2833.
Sacks, W. J., D. S. Schimel, and R. K. Monson (2007), Coupling between carbon cycling and climate in a high-elevation, subalpine forest:
A model-data fusion analysis, Oecologia, 4, 54–68.
Sitch, S., et al. (2008), Evaluation of the terrestrial carbon cycle, future plant geography and climate-carbon cycle feedbacks using five
Dynamic Global Vegetation Models (DGVMs), Global Change Biol., 14, 1–25.
Stocks, B. J., et al. (2002), Large forest fires in Canada, 1959–1997, J. Geophys. Res., 107(D1), 8149, doi:10.1029/2001JD000484.
Tanja, S., et al. (2003), Air temperature triggers the recovery of evergreen boreal forest photosynthesis in spring, Global Change Biol., 9,
1410–1426, doi:10.1046/j.1365-2486.2003.00597.x.
Weedon, G. P. (2011), Creation of the WATCH 20th century ensemble product, WATCH Technical Report Number 37, UK Met Office.
Welp, L. R., J. T. Randerson, and H. P. Liu (2007), The sensitivity of carbon fluxes to spring warming and summer drought depends on plant
functional type in boreal forest ecosystems, Agric. For. Meteorol., 147, 172–85.
Westerling, A. L., H. G. Hidalgo, D. R. Cayan, and T. W. Swetnam (2006), Warming and earlier spring increase western U.S. forest wildfire activity,
Science, 313, 940–943.
Williams, A. P., et al. (2013), Temperature as a potent driver of regional forest drought stress and tree mortality, Nat. Clim. Change, 3(3), 292–297.
Xu, L., et al. (2013), Temperature and vegetation seasonality diminishment over northern lands, Nat. Clim. Change, 3, 581–586.
Yarie, J. (2008), Effects of moisture limitation on tree growth in upland and floodplain forest ecosystems in interior Alaska, For. Ecol. Manage.,
256, 1055–1063.
Zhao, M., and S. W. Running (2010), Drought-induced reduction in global terrestrial net primary production from 2000 through 2009, Science,
329(5994), 940–943.
Zhou, L. M., C. J. Tucker, R. K. Kaufmann, D. Slayback, N. V. Shabanov, and R. B. Myneni (2001), Variations in northern vegetation activity
inferred from satellite data of vegetation index during 1981 to 1999, J. Geophys. Res., 106, 20,069–20,083, doi:10.1029/2000JD000115.
PARIDA AND BUERMANN
©2014. American Geophysical Union. All Rights Reserved.
8