Content uploaded by Ashish Sharma
Author content
All content in this area was uploaded by Ashish Sharma on Nov 14, 2017
Content may be subject to copyright.
Reduced spatial extent of extreme storms
at higher temperatures
Conrad Wasko
1
, Ashish Sharma
1
, and Seth Westra
2
1
School of Civil and Environmental Engineering, University of New South Wales, Sydney, New South Wales, Australia,
2
School of Civil, Environmental and Mining Engineering, University of Adelaide, Adelaide, South Australia, Australia
Abstract Extreme precipitation intensity is expected to increase in proportion to the water-holding
capacity of the atmosphere. However, increases beyond this expectation have been observed, implying
that changes in storm dynamics may be occurring alongside changes in moisture availability. Such changes
imply shifts in the spatial organization of storms, and we test this by analyzing present-day sensitivities
between storm spatial organization and near-surface atmospheric temperature. We show that both the total
precipitation depth and the peak precipitation intensity increases with temperature, while the storm’s spatial
extent decreases. This suggests that storm cells intensify at warmer temperatures, with a greater total
amount of moisture in the storm, as well as a redistribution of moisture toward the storm center. The results
have significant implications for the severity of flooding, as precipitation may become both more intense and
spatially concentrated in a warming climate.
1. Introduction
Short-duration extreme precipitation is predicted to intensify as a result of increases in atmospheric tempera-
ture in most locations globally [Kirtman et al., 2013]. Investigation of the historical sensitivity of precipitation
to temperature is an important source of evidence to understand how extreme precipitation might change in
the future [Collins et al., 2013]. In the absence of changes to circulation patterns and relative humidity, ther-
modynamic factors suggest that extreme precipitation intensity should scale at about 7%/°C as governed by
the Clausius-Clapeyron (C-C) relationship, which describes the capacity of the atmosphere to hold moisture
[Trenberth et al., 2003; Westra et al., 2014]. Using present-day climate data, short-duration precipitation has
been found to scale at rates ranging from C-C to double C-C for temperatures below 24°C [Lenderink and
van Meijgaard, 2008; Hardwick Jones et al., 2010; Lenderink et al., 2011; Mishra et al., 2012; Berg et al., 2013],
with the scaling differing with storm duration [Panthou et al., 2014; Wasko et al., 2015]. Similar rates of scaling
have been observed in long-term historical trends of extreme precipitation, supporting the hypothesis that
these relationships may be indicative of future climatic conditions [Westra and Sisson, 2011; Fujibe, 2013;
Westra et al., 2013, 2014].
A mechanism that has been put forward to explain departures from the thermodynamic (C-C) scaling rate is a
change in storm dynamics. This has been attributed to the release of latent heat, which can result in the storm
increasing the area over which it sources moisture [Trenberth et al., 2003]. Observation of scaling rates above
C-C has led to conjecture that there may be changes in dominant precipitation types as convective events have
greater scaling rates than nonconvective events [Haerter and Berg,2009;Berg et al., 2013; Collins et al., 2013;
Molnar et al., 2015] and also that both the thermodynamic and dynamical mechanisms may operate jointly
[Lenderink and van Meijgaard,2008;Collins et al., 2013; Wasko and Sharma, 2015] enhancing the convergence
of moisture possibly from a larger “moisture source”area [Trenberth et al., 2003; Westra et al., 2014].
The specific causes of changes in precipitation intensity with temperature can therefore be expected to be
associated with distinct temporal and spatial signatures. For example, if thermodynamic factors dominate
and the precipitation intensifies in proportion to the water-holding capacity of the atmosphere, a consistent
increase in precipitation intensity across a storm could be expected. In contrast, if dynamic factors dominate,
the spatial extent of each storm cell and the organization of rainfall within the storm cell should change
[Westra et al., 2014]. Although evidence exists of intensifying temporal patterns at higher temperatures
[Wasko and Sharma, 2015], studies investigating possible changes to the spatial signature of storms with
climate change remain limited having focused on areal reduction factors for engineering applications
[Li et al., 2015] and spatial structures in regional climate models [Guinard et al., 2015].
WASKO ET AL. REDUCED SPATIAL EXTENT OF EXTREME STORMS 4026
PUBLICATION
S
Geophysical Research Letters
RESEARCH LETTER
10.1002/2016GL068509
Key Points:
•Spatial extent of storms reduces as
temperatures increase
•Storm patterns are less uniform at
higher temperatures
•Moisture is redistributed from the
storm boundaries to the storm center
Supporting Information:
•Supporting Information S1
Correspondence to:
A. Sharma,
a.sharma@unsw.edu.au
Citation:
Wasko, C., A. Sharma, and S. Westra
(2016), Reduced spatial extent of
extreme storms at higher temperatures,
Geophys. Res. Lett.,43, 4026–4032,
doi:10.1002/2016GL068509.
Received 1 MAR 2016
Accepted 5 APR 2016
Accepted article online 7 APR 2016
Published online 25 APR 2016
©2016. American Geophysical Union.
All Rights Reserved.
Here we investigate the relationship between the spatial organization of moisture within a storm and the
near-surface dry-bulb temperature. We first examine the hypothesis that the distribution of precipitation with
distance from the storm centers differs when storms are stratified by temperature. Subsequently, we examine
the relationship with temperature of a range of statistics that describe the organization of moisture within the
storm. Finally, using the derived relationships, we develop projections of precipitation distribution with dis-
tance from the storm center for higher temperatures.
2. Data and Methods
Subdaily precipitation and 1.2 m dry-bulb temperature were obtained from the Australian Bureau of
Meteorology weather station data set and have been used extensively in previous studies [Hardwick Jones
et al., 2010; Westra and Sisson, 2011; Westra et al., 2012; Wasko and Sharma, 2015]. Precipitation is measured
using a tipping bucket rain gauge or pluviograph and reported every 6 min. In this study the reported
precipitation was accumulated to resolutions of 1 and 3 h. The temperature data resolution is typically 3 to
12 h, depending on the station. The data set contains over 1300 precipitation stations and approximately
1700 temperature stations across Australia.
To analyze the spatial distribution of moisture within a storm, point observations of precipitation need to
be transformed to spatial fields. First, at each gauge location, independent precipitation events were
identified. Storm events were defined as independent if they were separated by 5h of zero precipitation
for the hourly rainfall series and 15 h of zero precipitation for the 3-hourly rainfall series [Wasko and
Sharma, 2014]. The precipitation chosen for further analysis was the maximum 1 h or 3 h precipitation burst
with each storm. Each precipitation maximum was matched to the coincident temperature, which was cal-
culated as a 24 h moving average centered on the time of the maximum precipitation. The precipitation-
temperature pairs were extended to a spatial field by finding the precipitation that occurred at the same
time as the maximum from the central gauge at neighboring gauges up to a radius of 50 km. A spatial field
was considered for analysis if it had a minimum of five data points and the maximum precipitation in the
spatial field occurred at the central gauge. All other spatial fields were discarded. For a station to be
included in the analysis it had to have a minimum of 10 years of record length and a minimum of 100
observed independent spatial fields. In total, 93 stations were considered for the 1 h duration and 78 sta-
tions for the 3 h duration.
To isolate extreme events, a 90th percentile exponential quantile regression was fitted to the peak
precipitation-temperature pairs [Wasko and Sharma, 2014] and only spatial field events that exceeded the
regression line were analyzed for their sensitivity to temperature. The final data set for each site contains
j= 1..mspatial fields matched to their coincident temperature T
j
. Each spatial field jcontains i= 1..nprecipita-
tion observations p
i
at distance d
i
from the center gauge; thus, p
ij
is the precipitation observation iof spatial
field j.
A set of statistics was chosen to represent the spatial organization of moisture within the spatial fields and
calculated on the final mextreme sets of spatial fields. The statistics are as follows:
1. Peak precipitation (PP), which is the maximum precipitation within the spatial field (therefore by defini-
tion located at the storm center);
2. Total precipitation (PT), which is the two-dimensional integration of the precipitation plotted against the
radial distance from the storm center;
3. Fraction of zero rainfall observations (PZ) in the spatial field;
4. Coefficient of variation (CV), which is the variance of the observations within the spatial field divided by
their mean;
5. Effective radius (RE), defined at the centroid of the precipitation where p
ij
are the nprecipitation observations
within the spatial field jwith distance d
ij
from the central gauge:
REj¼
X
n
i¼1
pijdij
X
n
i¼1
pij
; (1)
Geophysical Research Letters 10.1002/2016GL068509
WASKO ET AL. REDUCED SPATIAL EXTENT OF EXTREME STORMS 4027
6. Parameters Aand Bof an expo-
nential curve that describes
how the storm decays with
distance from its center [von
Hardenberg et al., 2003; Rebora
and Ferraris, 2006] where the
parameters to be estimated for
each spatial field j:
pij dij
¼AjeBjdij
:(2)
Here Aand Bwere fitted by mini-
mizing the absolute residuals using
a shuffled complex evolution opti-
mization algorithm [Andrews et al.,
2011]. A schematic of PP, PT, and
RE is presented in Figure S1 in the
supporting information.
An exponential regression was
then used to find the relationship
between each desired statistic set
S
j
= {PP
j
,PT
j
,CV
j
,RE
j
,A
j
,B
j
} with
temperature T
j
(equation (3)), with
the exception of the fraction of
zero statistic where a linear regression was used. For a ΔTchange in the temperature, the fitted relationship
is as follows [Hardwick Jones et al., 2010; Utsumi et al., 2011], where α
s
is the rate at which the statistic scales
per degree temperature change:
STþΔTðÞ¼STðÞ1þαS
ðÞ
ΔT
:(3)
The data represent a wide range of climatic zones and precipitation-temperature sensitives [Hardwick
Jones et al., 2010; Utsumi et al., 2011; Wasko and Sharma, 2014]. The Australian climate can be split into
three main zones (Figure 2): tropical in the north, temperate in the east, and arid in the southwest. The
precipitation is summer dominant in the tropical north and winter dominant in the subtropical south
mainland. Temperate zones in the east have slightly winter-dominant precipitation, with precipitation
seasonality becoming uniform toward the south. The scaling of the statistics is aggregated across these
climatic zones.
Finally, the spatial fields were projected for a warmer climate on the basis of fitted exponential curves
(equation (2)). To represent the current (base) climate, parameters Aand Bwere calculated at a temperature
of 20°C. The curves were then projected in one degree increments using the calculated scaling above (equa-
tion (3)). It was found that the scaling of the parameters Aand Bwas less stable than the peak precipitation
(PP) and effective radius (RE) scaling. Hence, the latter were used to scale the fitted curves. The peak precipi-
tation scaling is equivalent to the parameter Ascaling and the inverse of the effective radius scaling is equiva-
lent to the parameter scaling B.
3. Results
3.1. Spatial Organization as a Function of Temperature
In order to confirm the hypothesis that there is a different spatial organization of moisture within storm cells
at differing temperature, we begin by splitting all the spatial fields across Australia into two groups: the first
for temperatures above 25°C and the second for temperatures below 18°C. If we are able to detect differences
in the organization of moisture within the largest storms in these groups, it would suggest that temperature
does in fact covary with the spatial signature of storm cells. For both groups the greatest 1000 events by
precipitation depth were chosen.
Figure 1. Fitted exponential curves for the greatest 1000 hourly storm bursts
by precipitation depth below 18°C (blue) and above 25°C (red). The vertical
dashed lines are the corresponding effective radii of the storm data defined
as the centroid of the storm by precipitation volume. Three-dimensional
curves are also presented emphasizing the increase in the storm intensity at
the center of the storm and reduced spatial extent at higher temperatures.
Geophysical Research Letters 10.1002/2016GL068509
WASKO ET AL. REDUCED SPATIAL EXTENT OF EXTREME STORMS 4028
Figure 1 presents the calculated
exponential curves and the
effective radii (equation (2)) of the
two groups. The confidence inter-
vals for the parameters Aand B
were calculated using 1000 boot-
strapped replicates of the precipi-
tation data. As can be seen, the
intensity at the storm center is
greater at high temperatures and
decays more rapidly with distance
from the center. This indicates that
the moisture becomes more con-
centrated near the storm center
for the warmer storms. This is con-
firmed by considering the effective
storm radius: at the higher tem-
perature, the effective storm radius
is located closer to the center.
Consistent results were found
when the analysis was repeated
with spatial field aggregated on a
regional basis (not shown).
3.2. Scaling of Spatial
Field Statistics
Having found that temperature
affects the spatial organization of
moisture within storm cells, we
now focus on the statistics
describing the spatial organization
for the most extreme events, that
is, those above the 90th percentile
exponential quantile regression.
The scaling of peak precipitation
and effective storm radius with
temperature for the extreme spa-
tial fields is presented in Figure 2.
The peak precipitation increases
with temperature at 95% of the sites (Figure 2a), and the effective radius decreases with temperature at
82% of the sites (Figure 2b). An example of the scaling calculations is presented in Figure S2. Consistent
increases in the peak precipitation and decreases in the effective radius were found for differing threshold
percentiles; however, the magnitude of the scaling was reduced for less extreme threshold percentiles.
The scaling presented in Figure 2 can be grouped by climatic zone and has been presented as box plot dis-
tributions in Figure 3. The median scaling of the peak precipitation in the arid zone is approximately 5%°C
1
;
however, the total storm precipitation scales at a slightly lower rate of approximately 4%°C
1
. As the effective
radius reduces at higher temperatures this suggests that the precipitation is greater at the storm center and
lesser at the storm extents.
This effect is more pronounced for the temperate zone results. The peak precipitation scaling is approxi-
mately 7%°C
1
and the effective radius has a negative scaling at a median rate of 3%°C
1
. Consequently,
the scaling in the total precipitation is less than the peak precipitation. This confirms that moisture is being
redistributed from the storm boundaries to the storm center, resulting in less precipitation at the storm
boundaries and a reduced storm size.
Figure 2. Scaling of the peak precipitation and effective radius for 1 h
duration events. Positive scaling is shown in red and negative scaling in
blue. (a) Circles show the scaling of peak precipitation, while the background
shading denotes the Koppen climate classification [Peel et al., 2007]. General
climatic zones are also shown. All the gauge sites used in the analysis are
shown as grey dots. (b) Circles show the scaling of effective radius.
Geophysical Research Letters 10.1002/2016GL068509
WASKO ET AL. REDUCED SPATIAL EXTENT OF EXTREME STORMS 4029
The scaling of the peak precipitation
and storm size is stronger at short
durations as shown by comparing
the 1 h scaling to the 3 h scaling
(Figure S3). For 3 h duration bursts,
97% of the sites have positive peak
precipitation scaling and 74% of the
sitesnegativeeffectiveradiusscal-
ing. However, the median scaling of
the peak precipitation is reduced to
5%°C
1
for the temperate zone
and 3%°C
1
for the arid zone
(Figure S4). Although the median
remained similar for the effective
radius scaling for the arid zone, in
the temperate zone it decreased in
magnitude from 3%°C
1
for the
1 h duration to 2%°C
1
for the 3 h
duration storms, suggesting that the
redistribution of moisture from the
storm boundaries to the storm cen-
ter is less at longer storm durations.
The scaling of the fraction of obser-
vations with zero precipitation (PZ)
and the coefficient of variation (CV)
in each spatial field also supports
the observed trends for the 1 h
(Figure S5) and 3 h durations
(Figure S6). Throughout Australia,
there is a positive scaling in the
fraction of zeros and coefficient of
variation at 87% and 92% of the
sites, respectively, for a 1 h duration.
Similar statistics are obtained for a
3 h duration. These results point to
a less uniform spatial distribution of
precipitation occurring over a smal-
lerareaathighertemperatures.
3.3. Projection for Higher
Temperatures
Using the scaling relationships
developed above, projections were
developed to illustrate how the
average storm shape for 1 h and 3
durations might change with
temperature increases of up to
3°C (Figure 4). The scaling applied
is the mean scaling for all the
stations within the climatic zone
of interest. The blue curve repre-
sents the storm shape for a base
climate of 20°C and the red curve
Figure 3. Box plots of peak precipitation scaling, effective radius scaling, and
total precipitation scaling for hourly precipitation grouped by arid and
temperate zones.
Figure 4. Predicted precipitation distribution due to temperature increase.
Each panel consists of four curves. The blue curve is fitted to a temperature
of 20°C. Each subsequent curve changes from blue to red in 1°C increments, up
to 23°C. The precipitation is standardized by the peak precipitation depth for
the baseline climate of 20°C to ensure the results are independent to the
baseline temperature choice.
Geophysical Research Letters 10.1002/2016GL068509
WASKO ET AL. REDUCED SPATIAL EXTENT OF EXTREME STORMS 4030
represents the shape for a projected increase of 3°C (i.e., 23°C); the intermediate curves represent increases in
1° increments. The precipitation in Figure 4 is standardized by the peak precipitation depth for the baseline
climate so the results presented are independent of the baseline temperature choice.
There is minimal change in the storm shape for the arid zone, with an increase in overall precipitation depth
occurring at the same rate as the scaling of the peak precipitation regardless of the storm duration. In
contrast, in the temperate zone, peak precipitation at the storm center increases at a greater rate than the
precipitation at the storm boundaries. In fact there is a decrease in precipitation at the storm boundaries
for both 1 h and 3 h duration. This presents the possibility of a smaller storm extent with redistribution of
moisture to the storm center at higher temperatures. The redistribution of moisture from the storm boundary
to the storm center is less for longer durations (e.g., 3 h), suggesting thatthe convergence of moisture is more
prominent for shorter durations.
4. Conclusions
It is increasingly understood that extreme rainfall scaleswith atmospheric temperature at or above the Clausius-
Clapeyron rate, but how this relates to the spatial organization of events is not well known. The results here
show that within a storm burst, precipitation scaling is not spatially uniform, with moisture being redistributed
from the storm boundaries to the storm center as temperature increases. Whereas the positive scaling in the
total precipitation supports the assertion that the thermodynamic mechanism of increasing moisture capacity
dominates increases in precipitation intensity, the redistribution of moisture from the storm boundaries and
nonconstant precipitation scaling with radial distance from the storm center suggests that storm dynamics
may also be changing. The results therefore suggest that both thermodynamic and dynamic factors result in
a storm with greater precipitation intensity with more moisture concentrated at the center. This is consistent
with trends from regional climate modeling which predict precipitation structures with larger volume and
greater heterogeneity in the future [Guinard et al., 2015]. If the identified historical relationships were to be
maintained as global temperatures increase, more concentrated spatial storm events could be expected with
higher temperatures. This redistribution could have significant implications for flood severity, as precipitation
may become both more intense and spatially concentrated in a warming climate. Future studies will focus
on replicating this work using observations from precipitation radars and linking synoptic weather patterns
to extreme precipitation patterns to better understand the drivers of precipitation extremes.
References
Andrews, F. T., B. F. W. Croke, and A. J. Jakeman (2011), An open software environment for hydrologi cal model assessment and development,
Environ. Modell. Software,26(10), 1171–1185, doi:10.1016/j.envsoft.2011.04.006.
Berg, P., C. Moseley, and J. O. Haerter (2013), Strong increase in convective precipitation in response to higher temperatures, Nat. Geosci.,
6(3), 181–185, doi:10.1038/ngeo1731.
Collins, M., et al. (2013), Long-term Climate Change: Projections, Commitments and Irreversibility, 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,edited
by T. Stocker et al., pp. 1029–1136, Cambridge Univ. Press, Cambridge U. K., and New York.
Fujibe, F. (2013), Clausius-Clapeyron-like relationship in multidecadal changes of extreme short-term precipitation and temperature in
Japan, Atmos. Sci. Lett.,14(3), 127–132, doi:10.1002/asl2.428.
Guinard,K.,A.Mailhot,andD.Caya(2015),Projectedchangesincharacteristics of precipitation spatial structures over North America,
Int. J. Climatol.,35(4), 596–612, doi:10.1002/joc.4006.
Haerter, J. O., and P. Berg (2009), Unexpected rise in extreme precipitation caused by a shift in rain type?, Nat. Geosci.,2(6), 372–373,
doi:10.1038/ngeo523.
Hardwick Jones, R., S. Westra, and A. Sharma (2010), Observed relationships between extreme sub-daily precipitation, surface temperature,
and relative humidity, Geophys. Res. Lett.,37, L22805, doi:10.1029/2010GL045081.
Kirtman, B., et al. (2013), Near-term climate change: Projections and predictability, 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, edited by T. Stocker et al.,
pp. 953–1028, Cambridge Univ. Press, Cambridge, U. K., and New York.
Lenderink, G., and E. van Meijgaard (2008), Increase in hourly precipitation extremes beyond expectations from temperature changes,
Nat. Geosci.,1(8), 511–514, doi:10.1038/ngeo262.
Lenderink, G., H. Y. Mok, T. C. Lee, and G. J. van Oldenborgh (2011), Scaling and trends of hourly precipitation extremes in two different
climate zones—Hong Kong and the Netherlands, Hydrol. Earth Syst. Sci.,15(9), 3033–3041, doi:10.5194/hess-15-3033-2011.
Li, J., A. Sharma, F. Johnson, and J. Evans (2015), Evaluating the effect of climate change on areal reduction factors using regional climate
model projections, J. Hydrol.,528, 419–434, doi:10.1016/j.jhydrol.2015.06.067.
Mishra, V., J. M. Wallace, and D. P. Lettenmaier (2012), Relationship between hourly extreme precipitation and local air temperature in the
United States, Geophys. Res. Lett.,39, L16403, doi:10.1029/2012GL052790.
Molnar, P., S. Fatichi, L. Gaál, J. Szolgay, and P. Burlando (2015), Storm type effects on super Clausius–Clapeyron scaling of intense rainstorm
properties with air temperature, Hydrol. Earth Syst. Sci.,19(4), 1753–1766, doi:10.5194/hess-19-1753-2015.
Geophysical Research Letters 10.1002/2016GL068509
WASKO ET AL. REDUCED SPATIAL EXTENT OF EXTREME STORMS 4031
Acknowledgments
The authors are grateful for the funding
support from the Australian Research
Council for this project. Westra was
supported by ARC Discovery project
DP150100411. The data set can be
obtained from http://www.bom.gov.au/
climate/data/stations/.
Panthou, G., A. Mailhot, E. Laurence, and G. Talbot (2014), Relationship between surface temperature and extreme rainfalls: A multi-time-scale
and event-based analysis, J. Hydrometeorol.,15(5), 1999–2011, doi:10.1175/JHM-D-14-0020.1.
Peel, M., B. Finlayson, and T. McMahon (2007), Updated world map of the Köppen-Geiger climate classification, Hydrol. Earth Syst. Sci.,11,
1633–1644.
Rebora, N., and L. Ferraris (2006), The structure of convective rain cells at mid-latitudes, Adv. Geosci.,7,31–35.
Trenberth, K. E., A. Dai, R. M. Rasmussen, and D. B. Parsons (2003), The changing character of precipitation, Bull. Am. Meteorol. Soc.,84(9),
1205–1217, doi:10.1175/BAMS-84-9-1205.
Utsumi, N., S. Seto, S. Kanae, E. E. Maeda, and T. Oki (2011), Does higher surface temperature intensify extreme precipitation?, Geophys. Res.
Lett.,38, L16708, doi:10.1029/2011GL048426.
von Hardenberg, J., L. Ferraris, and A. Provenzale (2003), The shape of convective rain cells, Geophys. Res. Lett.,30(24), 2280, doi:10.1029/
2003GL018539.
Wasko, C., and A. Sharma (2014), Quantile regression for investigating scaling of extreme precipitati on with temperature, Water Resour. Res.,
50, 3608–3614, doi:10.1002/2013WR015194.
Wasko, C., and A. Sharma (2015), Steeper temporal distribution of rain intensity at higher temperatures within Australian storms, Nat. Geosci.,
8(7), 527–529, doi:10.1038/ngeo2456.
Wasko, C., A. Sharma, and F. Johnson (2015), Does storm duration modulate the extreme precipitation-temperature scaling relationship?,
Geophys. Res. Lett.,42, 8783–8790, doi:10.1002/2015GL066274.
Westra, S., and S. A. Sisson (2011), Detection of non-stationarity in precipitation extremes using a max-stable process model, J. Hydrol.,
406(1–2), 119–128, doi:10.1016/j.jhydrol.2011.06.014.
Westra, S., R. Mehrotra, A. Sharma, and R. Srikanthan (2012), Continuous rainfall simulation: 1. A regionalized subdaily disaggregation
approach, Water Resour. Res.,48, W01535, doi:10.1029/2011WR010489.
Westra, S., L. Alexander, and F. Zwiers (2013), Global increasing trends in annual maximum daily precipitation, J. Clim.,26, 3904–3918,
doi:10.1175/JCLI-D-12-00502.1.
Westra, S., H. J. Fowler, J. P. Evans, L. Alexander, P. Berg, F. Johnson, E. J. Kendon, G. Lenderink, and N. M. Roberts (2014), Future changes to the
intensity and frequency of short-duration extreme rainfall, Rev. Geophys.,52, 522–555, doi:10.1002/2014RG000464.
Geophysical Research Letters 10.1002/2016GL068509
WASKO ET AL. REDUCED SPATIAL EXTENT OF EXTREME STORMS 4032
