PUBLICATIONS
Water Resources Research
RESEARCH ARTICLE
10.1002/2014WR016496
Key Points:
Canopy presence has a major effect
on snow distribution (average
reduction 49%)
With thicker snowpack, smaller snow
differences between open and
canopy areas
Snow difference (open-canopy) is
increased with time during the snow
season
Correspondence to:
J. Revuelto,
[email protected]
Citation:
Revuelto, J., J. I. L
opez-Moreno,
C. Azorin-Molina, and S. M. VicenteSerrano (2015), Canopy influence on
snow depth distribution in a pine
stand determined from terrestrial laser
data, Water Resour. Res., 51, 3476–3489,
doi:10.1002/2014WR016496.
Received 6 OCT 2014
Accepted 2 APR 2015
Accepted article online 6 APR 2015
Published online 11 MAY 2015
Canopy influence on snow depth distribution in a pine stand
determined from terrestrial laser data
pez-Moreno1, C. Azorin-Molina1, and S. M. Vicente-Serrano1
J. Revuelto1, J. I. Lo
1
Instituto Pirenaico de Ecologıa, Consejo Superior de Investigaciones Cientıficas (IPE-CSIC), Zaragoza, Spain
Abstract In this study, we analyzed the effects of the forest canopy and trunks of a pine stand in the central Spanish Pyrenees on the snow depth (SD) distribution. Using LiDAR technology with a terrestrial laser
scanner (TLS), high-resolution data on the SD distribution were acquired during the 2011–2012 and 2012–
2013 snow seasons, which were 2 years having very contrasting climatic and snow accumulation conditions.
Average SD evolution in open and canopy areas was characterized. Principal component analysis was
applied to identify days having similar spatial patterns of SD distribution. There was a clear contrast in the
temporal variability of the snowpack in different areas of the forest stand, corresponding generally to
beneath the canopy, and in open sites. The canopy and openings showed markedly different accumulation
and melting, with higher snow accumulation found in openings. Differences ranged from 14 to 80% reduction (average 49%) in the SD beneath the canopy relative to open sites. The difference in SD between open
and canopy areas increased throughout the snow season. The surveyed days were classified in terms of SD
distribution, and included days associated with: high SD, low SD, intense melting conditions and periods
when the SD distribution was driven by wind conditions. The SD increased with distance from the trunks to
a distance of 3.5–4.5 m, coinciding with the average size of the crown of individual trees.
1. Introduction
In mid and high latitude mountain areas, snow and forest comprise a resource of enormous economic and
€uchi et al., 2000]. Snow interacts with the forest in complex ways because
environmental importance [Kra
the canopy affects the snow cover distribution, and the physical properties and melting of snow. Furthermore, forest growth, and its health and survival, rely on the protective effect of snow cover under the
extreme weather conditions that occur in mountain areas, and snow supplies on soil moisture during the
growing season [Mellander et al., 2005, 2007].
C 2015. American Geophysical Union.
V
Research at many locations has investigated the effect of the forest canopy on the snow distribution and
pez-Moreno and Latron, 2008a;
the seasonal evolution of the snowpack [Hedstrom and Pomeroy, 1998; Lo
pez-Moreno and Sta
€hli, 2008; Lundquist et al., 2013; Musselman et al., 2008; Varhola et al., 2010]. This has
Lo
generally concluded that the forest canopy has a large effect on snow distribution and dynamics, but it can
vary greatly depending on the specific characteristics of the study site, the type of forest stand, and the
annual climatic conditions. The forest canopy reduces incoming solar radiation beneath the canopy but
enhances emission of long-wave radiation, and reduces the albedo because of the presence of plant litter
on the snow surface. Thus, differences in temperature among forest stands and differences in the exposure
of particular forests to solar radiation may lead to marked spatial and temporal differences in snow depth
pez-Moreno and Sta
€hli, 2008; Lundquist et al., 2013]. Moreover, depending on the
(SD) and snow melting [Lo
type of forest and the structure of particular stands, the capacity of the canopy to intercept snow and the
radiative fluxes beneath the canopy may show large spatial variability, even over very short distances
[Veatch et al., 2009]. The interception of snow by a particular forest stand may also vary among snowfall
events, depending on the preceding interception of snow by the tree branches, which tends to diminish
pez-Moreno and Latron, 2008b], and also
when the amount of previously intercepted snow is greater [Lo
depends on the temperature during the snowfall event [Lundquist et al., 2013]. The above implies that the
effect of the forest canopy on snowpack varies markedly at different spatial scales, and also exhibits great
inter and intra-annual variability, and indicates the need for research to enable better understanding of
how climate variability and change, and also forest management, might influence snow processes.
All Rights Reserved.
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3476
Water Resources Research
10.1002/2014WR016496
Most of the studies reported above were conducted at the stand scale (1–100 ha), with the aim of estimating reductions in the snow water equivalent (SWE) for areas beneath canopy compared with forest openpez-Moreno and Latron, 2008b; Varhola et al., 2010]. More detailed studies
ings or completely open areas [Lo
have assessed the effects of forest canopy characteristics, the variability of the various energy fluxes within
a given forest stand [Link and Marks, 1999; Mahat and Tarboton, 2013], and the influence of individual trees
by sampling the snow distribution around selected trunks [Faria et al., 2000; Musselman et al., 2008; Woo
and Steer, 1986].
In this study, we analyzed the spatiotemporal variability of SD within a 1000 m2 Pinus sylvestris stand in the
Spanish Pyrenees, using light detecting and ranging (LiDAR) technology, with a terrestrial laser scanner
(TLS), to assess small-scale relationships of the SD distribution to trees characteristics. The TLS was also used
to describe the geometry of the trunks and the canopy. A total of 20 field surveys were conducted during
two consecutive snow seasons (2011–2012 and 2012–2013). The TLS technique has been increasingly used
in snow studies [Deems et al., 2013], but it has not been applied for assessing the SD beneath forest
canopies.
The study generated a unique high-resolution data set of the SD distribution under forest canopy at a spatial scale not considered before, based on very frequent TLS data acquisition during the snow accumulation
and ablation periods. In order to keep undisturbed the SD distribution within the analyzed forest stand, we
did not survey the distribution of snow density. Nevertheless, the spatial variability of SD is high relative to
that of snow density [Marchand and Killingtveit, 2004; Mizukami and Perica, 2008; Pomeroy and Gray, 1995],
and the spatial variability of SWE is tightly related to SD variability. Differences on average SD values in
open and canopy areas were computed for the two analyzed snow seasons. Moreover, statistical analyses
(Principal Component Analysis) were used to identify days on which the spatial distribution of the snowpack
differed markedly. Such spatiotemporal patterns have been related to the distribution of the forest canopy
and the distance to the tree trunks. Pinus sylvestris is characteristic of Mediterranean mountain areas, but
there have been no previous studies of their effect on the snowpack distribution and dynamics. The
approach taken in this study enabled quantification at a very detailed spatial scale of how the forest affects
snow dynamics and how this is influenced by contrasting climatic conditions.
2. Study Area
The studied forest stand is located near the Balneario de Panticosa in the headwater of the Gallego River in
the central Spanish Pyrenees (Figure 1). In this part of the Pyrenees, 24% of the land surface is covered by
forests [Villar et al., 1997]. Pynus sylvestris is the dominant species from 1200 to 1700 m a.s.l. From 1700 to
1900 m a.s.l., there is a progressive transition to more sparse Pynus uncinata, which is the dominant tree
species from 1900 to 2300 m a.s.l (the approximate regional tree line) [Camarero et al., 1998].
The climatic conditions at Balneario de Panticosa (1630 m a.s.l.) reflect its location on the southern slopes of
the Pyrenees, which are affected by both Mediterranean and Atlantic Ocean climatic influences. Meteorological data are available from a weather station located in an open area of the forest 700 m away and
40 m lower than the study site. For the period 1994–2014, the average annual temperature was 7.28C
(winter average 1.58C; spring average 9.28C). The average annual precipitation is 1528 mm, and the average
pez-Moreno, 2005; Lo
pezannual SD is 4.34 m. The large interannual variability of climate in this region [Lo
Moreno and Vicente-Serrano, 2007] is exemplified by the stark contrast in climate and snow accumulation
between the two snow seasons considered in the study (Figure 2). The very low level of snow accumulation
recorded in the 2011–2012 season coincided with low temperatures in winter and a warm spring. In contrast, the 2012–2013 season was very wet, with temperatures close to the average in winter and cooler in
spring, leading to a thick and long lasting snowpack (Figure 2).
The study site is located at 1700 m a.s.l. (428450 N, 08040 W) and has an approximate area of 1000 m2. This
area was selected because of its relatively flat topography, making it easier to distinguish the effect of the
forest canopy on SD and assess the temporal evolution of the snowpack during the two snow seasons. The
analyzed stand of Pinus sylvestris comprised 17 trees of differing height and canopy radius. The tallest tree
was 14 m (average tree height 9.4 m), and the average canopy radius was 3.1 m (standard deviation 1.3 m).
The trees are heterogeneously distributed, resulting in alternating forest openings and densely covered
areas. The heterogeneous distribution enabled detailed comparison of the snow dynamics in a forested
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3477
Water Resources Research
10.1002/2014WR016496
Figure 1. Study site in the ‘‘Balneario de Panticosa.’’ (top) The location of the experimental site. (bottom) Panoramic views of the study site during (left) snow-covered and (right) snowfree periods.
area with that in forest openings, as well as analysis of how accumulation and melting patterns are affected
by canopies and trunks.
3. Data and Methods
3.1. SD Measurements
TLS is increasingly being applied in snow studies [Deems et al., 2013], having been tested against manual
sampling measurements [Prokop, 2008; Revuelto et al., 2014a] and applied at a diversity of sites [Egli et al.,
2012; Gr€
unewald et al., 2010; Mott et al., 2011; Revuelto et al., 2014b; Schirmer et al., 2011]. In forested mountain areas, airborne laser scanners (ALS) have also been applied to analysis of the SD distribution [Harpold
et al., 2014; Deems et al., 2006, 2008; Hopkinson et al., 2004; Trujillo et al., 2007, 2009]. However, TLS has not
been applied in the study of snow beneath forest canopies. This study is the first to use TLS to investigate
the SD distribution beneath the canopy of a pine forest stand and provides data on the snow distribution at
a very fine spatial resolution.
Field measurements were made on 20 survey days using a RIEGL LPM-321 laser scanner, which generated
3D point clouds for each sampling occasion; this provided highly distributed information (an average of 400
points/m2) on the snow surface characteristics within the forest stand. By assessing the topography of the
snow surface with that determined during the snow-free period, it was possible to determine the SD distribution for each survey occasion [Revuelto et al., 2014a]. In the present study, we established three scan positions to minimize shadows in the point clouds caused by trees. Eight cylindrical (0.12 m diameter and
height) reflective targets on poles of 2 m height were placed at fixed locations in the study area. These
were used to merge the point clouds acquired from each scan position, using a minimum of five common
targets visible from each pair of scan positions. Merging the information obtained from each scan position
resulted in at least 90% of the study area being scanned; this procedure substantially decreased the shadow
effect of trunks and canopies. Following processing and elimination of points corresponding to trees, the
point clouds were rasterized at a grid size resolution of 0.06 m2 (0.25 m 3 0.25 m); this generated >11,000
grid cells providing SD values per day. To avoid alteration to the snow pack in the study stand, manual
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3478
Water Resources Research
10.1002/2014WR016496
Figure 2. Evolution of temperature and snow depth measured at the nearest meteorological station for the snow seasons (left) 2011–2012 and (right) 2012–2013, and the 25th (P25)
and 75th (P75) percentiles of the historical data set (1994–2014). Vertical dashed lines indicate the survey dates.
measurements were not taken for validation. However, the deviation of TLS values for SD from measured
values for a TLS working distance <40 m was expected to be less than 0.08 m, based on the deviations
determined for distances of 200–800 m from the TLS device [Revuelto et al., 2014a]. In this study, recursive
comparisons between manual SD measurements and TLS SD measurements were done for several TLS surveys in a close location in the Pyrenees with the same device (RIEGL LPM-321). In such a way, Revuelto et al.
[2014a] stated that for a distance of about 200 m from the TLS the mean absolute error between both measurement methods was 0.06 m. This encourages accepting a maximum deviation between both methods
below this value at the Panticosa study site.
Figure 3 provides six SD distribution maps showing the effect of tree branch interception following snowfall
events (with lower SD values beneath canopies; 6 December 2012 and 27 February 2013) and the SD distribution following periods with no snow accumulation (20 December 2012, and 9 and 26 April 2013).
In the 2011–2012 season, only six survey days (20 December 2011; 2, 9, and 28 February 2012; 13 March 2012;
16 April 2012) were involved because the low levels of snow accumulation led to very rapid and early snowpack thaw. Between 2 and 9 February 2012, a strong northerly wind caused substantial snow redistribution
just prior to the commencement of the melt period. In contrast, during the 2012–2013 season the high SD
accumulation enabled exhaustive monitoring of SD evolution and involved a total of 14 days of field measurements. The first was conducted on 6 December 2012, followed by three subsequent surveys on 11, 20, and 27
December 2012. Road closure because of avalanche risk prevented further fieldwork until 27 February 2013.
Subsequent surveys occurred on 4, 8, 15, and 21 March; 9, 14, 18, and 26 April; and 2 May 2013.
In addition to scanning the snowpack, on each survey occasion two snow pits (one in a forest opening and
one beneath the forest canopy) were dug near the studied stand to measure the snow density in open
areas and beneath the canopy. The two density measurements were averaged and were used to provide a
measure of changes in snow density during the study period.
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3479
Water Resources Research
10.1002/2014WR016496
Figure 3. Snow depth distribution maps for 6 of the 20 survey days. The discrimination between open and canopy areas is shown on the right side of the figure. The soft white line represents the canopy limit.
To discriminate between canopy and opening areas, we used the TLS to reproduce the canopy structure,
i.e., the trees were scanned under snow-free conditions. The canopy presence mask (Figure 3, right) was
generated from several scans obtained from the three scanning positions with a specific TLS acquisition
mode for short distances and first signal rebound (near range first target). With a transition zone between
Open and Canopy zones of 0.25 m, SD in these two areas is considered herein after.
3.2. Statistical Analysis
Principal component analysis (PCA) was used to identify temporal SD distribution patterns within the study
forest stand. This selection is supported by the PCA capabilities for preserving the main characteristics of
the study area evolution and identifying local particularities of different zones within the area of interest
[Vicente-Serrano, 2005]. From the original variable data set, PCA reduces dimensionality obtaining new variables, which are the principal components. These components are linear combinations of the original variables. The coefficients of the linear combinations are the factorial scores, which represent the weight (i.e., the
correlation) of the original variables with the principal component [Hair et al., 1999]. The mathematical formulation of this PCA can be deeply consulted in Jollife [1990] and Baeriswyl and Rebetez [1997].
This type of statistical analysis has been used previously for studying snow accumulation patterns in forpez-Moreno and Latron, 2008a]. For reflecting the main spatial patested areas [Winkler and Moore, 2006; Lo
terns of the analyzed variables; in the PCA analysis, the grid cells were the cases and the SDs measured
during each survey were the variables. The intention of applying this cases/variables selection in the PCA is
to identify groups of days with contrasting SD configurations.
The number of selected components by the PCA was based on the percentage of explained variance [North
et al., 1982], in such a way components with an explained variance above 10% were retained for further
analysis. Varimax rotation [Kaiser, 1958] was applied to the components to obtain physically comprehensible
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3480
Water Resources Research
10.1002/2014WR016496
patterns [Richman, 1986]. PCA was
applied to the data set of 11,000 SD
grid cells for each of the 20 survey
days.
The components were related to the
open/canopy domains, focusing on
SD spatial patterns rather than on
general temporal patterns, which are
represented in the factorial score distributions. As an additional analysis,
average SD evolution in relation to
trunk distance was considered. SD
values of grid cells located at specific
distances from the trunk edge (0.25,
0.75, 1.25, 1.75, . . ., 5.75 m) were
averaged for each survey day. Subsequently, the obtained spatial evolution was represented grouping the
days depending on the PCA component that better describes the
observed SD spatial variability.
4. Results
4.1. Snow Depth Evolution
Patterns in Open/Canopy Zones
Figure 4 shows the average and
25th and 75th percentile values of
SD for all cells belonging to canopy
Figure 4. Evolution of SD in the 20 survey days for the open and canopy areas for
(C) and open (O) areas of the varithe (top) 2011–2012 and (bottom) 2012–2013 snow seasons. Vertical continuous
lines mark the 25th and 75th percentiles, and the filled geometric figures show the
ous survey days. Table 1 shows the
average values for each component.
average SD values for each area
and the differences (in %) between
open-canopy (O–C). Also it is shown in Table 1, the average snow density values for each survey day.
Depending on the survey occasion, the differences in SD between open zones and canopy zones
ranged between 14 and 80% (average difference 49%), with lower SD values always found for the canopy zones. Greater differences were found between open and canopy zones when the average snowpack in the basin was thinner; the difference increased throughout the snow season or with SD loss
(compaction and melting, or compaction alone) that occurred after snow accumulation events (e.g.,
from 6 to 27 December 2012). An exception was found for 16 April 2012, when only a few snow
patches were present, and the difference between open and canopy zones was only 14%. The average
density values (Table 1) increased throughout the snow season, with lower values compared with the
preceding survey day only being found following episodes in which a new fresh snow layer accumulated (e.g., 15 March 2013).
Table 2 shows the change in SD for open (O) and canopy (C) zones between consecutive surveys conducted
within a 3 week period, or for the first measurement in the year (in this case, the difference relative to bare
soil). Table 2 also shows the differences (in %) between the changes in SD found for the open and canopy.
The results show that during the only period of accumulation that satisfied the temporal threshold of 3
weeks (8–15 March 2013), and during the first survey day of each of the two seasons (when accumulation
occurred over bare soil), the open zones accumulated much more snow (52%, 36%, and 38% respectively;
average 42%) than occurred in the canopy zones. The other periods were characterized by no snow accumulation events. During winter, greater SD loss occurred in the open zones. However, during periods dominated by intense melting conditions (from the end of March until the snow disappeared), the reduction in
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3481
Water Resources Research
10.1002/2014WR016496
Table 1. Average SD Values for Open and Canopy Areas, % Differences Between Both Zones and Average Snow Density for Each Survey
Day
Exp. Campaign Date
Open (O) AvgSD (m)
Canopy (C) AvgSD (m)
O-C (%)
Snow density (kg/m3)
0.16
0.15
0.26
0.14
0.03
0.02
0.49
0.30
0.30
0.22
1.54
1.45
1.29
1.53
1.45
1.14
0.96
0.73
0.41
0.30
0.08
0.05
0.13
0.03
0.01
0.01
0.32
0.19
0.15
0.07
1.11
1.04
0.90
1.04
0.98
0.67
0.51
0.29
0.10
0.06
52
69
50
81
68
14
36
37
51
70
28
28
30
32
33
41
46
60
76
80
148
136
257
381
400
240
159
237
239
263
392
369
407
386
396
477
513
513
518
525
20/12/2011
02/02/2012
09/02/2012
28/02/2012
13/03/2012
16/04/2012
06/12/2012
11/12/2012
20/12/2012
27/12/2012
27/02/2013
04/03/2013
08/03/2013
15/03/2013
21/03/2013
09/04/2013
14/04/2013
18/04/2013
26/04/2013
02/05/2013
SD varied between the open and canopy zones, but for all considered periods the SD loss was higher in
open areas than beneath the canopy.
The PCA results showed four patterns of SD distribution that explained 87% of the total variance: component 1 (C1) explained 34%, component 2 (C2) explained 26%, component 3 (C3) explained 15%, and component 4 (C4) explained 12%. Figure 5 shows the spatial distribution of factorial scores for each component,
which represent the standardized anomalies of SD, that is, for positive (negative) factorial scores higher
(lower) snow accumulations are observed in comparison to the average SD accumulation of a specific day.
When a day is identified with a component, its SD spatial distribution is similar or equivalent to the spatial
distribution of the factorial scores of the component. Thereby, a day with high correlation with C1 would
have the highest SD values in the upper part of the map (dark red color) and the lower SD values in the bottom part of the map in both left and right sides (dark blue color). Similarly, a day with high correlation with
C4 would have the highest SD in the
lower central part of the map. Nevertheless, other characteristics of days repreTable 2. Snow Depth Changes in Open and Canopy Areas for Consecutive
Field Surveysa and for The First Sampling Occasion on Which Snow Accumusented by each component (high or low
lated Over Bare Soilb
average SD, wind action, etc.) must be
Open Avg
Canopy Avg
Open-Canopy
obtained from a separate analysis.
Analyzed Periods
Dif (m)
Dif (m)
Dif (%)
20/12/2011c
9/2/2012 to 28/2/2012
28/2/2012 to 13/3/2012
6/12/2012c
6/12/2012 to 11/12/2012
11/12/2012 to 20/12/2012
20/12/2012 to 27/12/2012
27/02/2013 to 4/3/2013
4/3/2013 to 8/3/2013
8/3/2013 to 15/3/2013c
15/3/2013 to 21/03/2013
21/03/2013 to 9/4/2013
9/04/2013 to 14/4/2013
14/4/2013 to 18/4/2013
18/4/2013 to 26/4/2013
26/4/2013 to 02/5/2013
10.16
20.122
20.106
10.49
20.193
20.002
20.080
20.084
20.170
10.237
20.071
20.311
20.182
20.234
20.320
20.111
10.08
20.103
20.016
10.32
20.128
20.004
20.083
20.070
20.145
10.146
20.065
20.310
20.155
20.224
20.192
20.041
52
16
85
36
34
NA
24
17
13
38
9
0
15
4
40
63
a
Consecutive surveys done within a 3 week period. The period 2–9 February 2012 is not included because of the strong winds that blew at that time.
b
In last column, the % difference between both changes is presented.
c
Indicates periods in which snow accumulation was recorded.
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
C1 represented days having high SD
(average SD in the forest stand ranged
from 0.5 to 1.3 m), with maximum correlations with snow distribution occurring
for 27 February; 4, 8, 15, and 21 March;
and 9, 14, and 18 April 2013. These dates
were characterized by high SD, especially in open areas having thinner snowpack beneath the canopy (Table 3). C2
represented days having low SD (average SD <0.5 m), with the maximum correlations with the distribution observed
for 20 December 2011; 2 February; and
6, 11, 20, and 27 December 2012. The
snow distribution represented by C2
indicated maximum accumulation in
3482
Water Resources Research
10.1002/2014WR016496
Figure 5. Factorial scores of grid cells of the four PCA components.
open zones, while thinner snowpack occurred in canopy zones. Despite there exist some similarities
between C1 and C2 snow distributions, clearly for C2 areas beneath individual trees exhibit very low SD
(see Figure 3), while for C1 snow depth beneath individual trees is not particularly low. This may be related
to the higher overall SD observed in C1 days. Comparison of the average SD in open and canopy zones
revealed higher differences between both zones for C2 than for C1. C3 represented the snow distribution
pattern during the surveys in 2012, on 9 and 28 February, 13 March and 16 April. Based on the field observations, this component represented the spatial distribution of snow resulting from very strong winds (gusts
>27 m/s were recorded at nearby meteorological stations; as at Izas Experimental Catchment Revuelto et al.
[2014b]) combined with snowfall during the period 6–8 February. Subsequently, snow was blown from
windward areas, and accumulated on the leeward side, behind the trees. Moreover, as shown in Figure 4,
the SD variability for those days included in C3 was relatively high because of wind redistribution (note the
low SD average values for 13 March and 16 April compared to 9 and 28 February, what difficult to observe
SD variability for the two first dates). Component C4 represents the snow distribution during 26 April and 2
May 2013, which were dates on which considerable snow remained in open areas, but had almost disappeared beneath the canopy.
Average SD values in open and canopy areas on days represented by each component are presented in
Table 3, ordered according to the PCA components for ease of comparison. The smallest differences
between open and canopy areas occurred on C1 days (high level of snow accumulation), with 27%, 60%,
and 37% being the minimum, maximum, and average differences, respectively. The greatest differences in
average SD value occurred for the 2 days associated with C4 (when there was almost no snow beneath the
canopy, but deep snowpack in the open areas), with 75% and 80% differences for these 2 days. The C2
component, which represents days having low SDs, generally exhibited greater percentage differences than
C1 (35%, 70%, and 52% for the minimum, maximum, and average differences, respectively). For days represented by C3 (wind redistribution), greater SD values were also found for open areas than beneath the canopy, but the distribution was very different from that found for the other three components with high
spatial variability for both zones (Figure 4).
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3483
Water Resources Research
10.1002/2014WR016496
Table 3. Average SD in Open and Canopy Areas for the Days Classified in the
Four Groups According to the PCA Components, and the % Difference
Between These Areas
PCA
Component
C1
C2
C3
Exp.
Campaign
Date
OpeningAvgSD
(m)
Canopy
AvgSD
(m)
Difference
Open-Canopy
(%)
27/2/2013
4/3/2013
8/3/2013
15/3/2013
21/3/2013
9/4/2013
14/4/2013
18/3/2013
20/12/2011
2/2/2012
6/12/2012
11/12/2012
20/12/2012
27/12/2012
9/2/2012
28/2/2012
13/3/2012
16/4/2012
26/4/2013
2/5/2013
1.53
1.45
1.29
1.53
1.45
1.14
0.96
0.73
0.16
0.15
0.49
0.30
0.30
0.22
0.26
0.14
0.03
0.02
0.41
0.30
1.11
1.04
0.90
1.04
0.98
0.67
0.51
0.29
0.08
0.05
0.32
0.19
0.15
0.07
0.13
0.03
0.01
0.01
0.10
0.06
28
28
30
32
33
41
46
60
52
69
36
37
51
70
50
81
68
14
76
80
In spite of the PCA survey day groups,
for both snow seasons (except for 16
April 2012), an increase of SD difference between canopy and open areas
is observed as snow season progresses.
This difference is only reduced when
one or several snow accumulation
events occur. This is quite evident
when the difference observed for 27
December 2013 (70%) is contrasted
(Table 3) to that observed on 27 February 2014 (27%) after several snow accumulation events (Figure 2). Besides,
after this day, it is clearly observed that
the reported SD difference increases
while the snow season progresses.
4.2. Snow Depth Changes as a
Function of Distance From Tree
Trunks
C4
Figure 6 shows the SD in the forest
stand against distance from tree trunks,
distinguishing survey days grouped
according to each of the four PCA groups. In general, and independently of the spatial patterns obtained,
the SD increased rapidly over the shortest distances to the trunks. This increase continues until 3.5–4.5 m
from the trunks for the days represented by C1, C2, and C4, where the maximum SD is generally observed.
This distance coincides with average canopy radius. In general, the edge effect of canopies was clearly evident, with a slight decrease in SD beyond the maximum depth in these zones. Days represented by C3
(wind redistribution) showed a very irregular pattern of SD change with distance from the tree trunks, with
the maximum SD generally observed 2–3 m from the trunk and marked variability with greater distance
between the observed SD values for the 4 days of this PCA group.
5. Discussion
This study represents a novel approach to assessing the SD distribution in forested areas using a TLS. This
device was also used to investigate the geometry of the forest canopy. The TLS enabled collection of data
at very high spatial resolution (0.25 m) during 20 field surveys conducted in 2 consecutive years having very
contrasting climatic and snow conditions that have shown an average SD reduction beneath canopies of a
49%. The PCA has facilitated the identification of those days when the pattern of snow distribution differed
in configuration. In addition, we also analyzed the change in average SD with distance from the tree trunks.
The procedures involved provided related but complementary information on how the forest canopy
affects the snowpack in a 1000 m2 Pinus sylvestris stand.
The TLS technique has proven very useful for obtaining detailed information on the snow distribution in forested areas and on the characterization of the forest canopy. In this study, we considered the forest canopy
to be a categorical variable, but the data set acquired may enable future distributed analyses to directly
~o et al.,
relate the density and structure of the forest canopy, derived from TLS or airborne LIDAR [e.g., Rian
2003; Seidel et al., 2012], to the variability of SD beneath the forest canopy. The acquired data also represent
a resource for improving modeling and validation of interception and melting processes in forest areas
[Mahat et al., 2013; Musselman et al., 2012]. In spite that TLS is a powerful information source that provides
high spatial resolution data, several particularities must be considered. Sometimes these particularities
might seem trivial, but they may lead to large errors and increase the needed time for data collection.
Hereby, TLS tripod and device stability must be extremely important and carefully monitored [Revuelto
et al., 2014a]. In such a way, it is suggested snow compaction around the tripod before scanning, with a regular revision of tribrach stability. Also study site selection must consider technical issues in addition to these
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3484
Water Resources Research
10.1002/2014WR016496
Figure 6. Average SD values plotted against trunk distances for the four PCA components.
of scientific interest. For TLS, survey in a forested study area is highly recommended to work with a maximum of three scan positions, what must enable to retrieve information of a high percentage of the study
area reducing trees shadows from the different scan position perspectives. Moreover, considering the study
site characteristics and the scan positions (which must be maintained during the study period) the reflective
targets locations must be selected. These locations must allow scanning at least three targets from any scan
positions (however, it is highly recommended to scan at least five from any scan stations), being appropriately fixed in bare soil to guarantee a quality SD acquisition. In addition to all these considerations, study
site accessibility in winter must consider that TLS equipment weights around 45 kg (TLS device, batteries,
laptop, tripod, etc.) and needs at least two people for the transport. Thereby, distance from the study site to
any place which can be reached by car must consider this.
The statistical analysis discriminated areas where the snow distribution patterns contrasted on different days.
The results showed that the forest canopy markedly reduced the depth of the snowpack relative to open
areas, which is consistent with previous findings in many different geographic areas [Jost et al., 2007; Lundberg
et al., 2004; Pomeroy and Gray, 1995; Winkler et al., 2005; Mahat and Tarboton, 2013]. The average reduction
found (49%) is similar to that reported by these previous studies. The average SD difference between open
and canopy areas observed in this study revealed very large variability (ranging from 14 to 80%, Table 1)
among the 20 survey days in the snow seasons 2011–2012 and 2012–2013. Differences in SD were greater
when the snowpack was thinner, which is a result consistent with that found in a beech and fir forest stand in
pez-Moreno and Latron, 2008a, 2008b]. A comparable result
a neighboring valley in the French Pyrenees [Lo
was also found by Connaughton [1933], who reported that the snow differences between open and forested
sites for scarce and average snow seasons were 27.5% and 4.3%, respectively. In British Columbia, the reduction of SWE coefficients of variation in various forest stands was observed to increase with average snow accumulation, and was correlated with year-to-year variations that explained 33% of the total SWE variance
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3485
Water Resources Research
10.1002/2014WR016496
[Winkler and Moore, 2006]. All these results are making evident that snow distribution in a forest stand in the
Pyrenees, a temperate mountain range, shows similar behavior to that observed in other environments in the
average values, but also in the observed variance. Moreover, the similarity between results presented here,
with those obtained with other sampling schemes and techniques, indicates that TLSs can be applied in snow
studies in forested areas providing high spatial resolution information.
The observed SD values in 2013, a snow-rich year, with smaller differences between open and canopy areas
following the main snow accumulation episodes (Table 3; surveys between 27 February 2013 and 21 March
2013), could be a consequence of the reduced snow interception capacity of trees when the snowpack is
deeper, because of the effect of the preceding snowfall load on the snowfall interception capacity. This phenomenon has been demonstrated experimentally [Hedstrom and Pomeroy, 1998] and reported based on
pez-Moreno and Latron, 2008b; McNay et al., 1988; Mellander
direct observations [Keller and Strobel, 1982; Lo
et al., 2005], supporting the hypothesis of our study.
The results also suggest that snow interception during periods with snow accumulation leads to greater differences between SD changes in open and canopy areas (superscript ‘‘c’’ periods in Table 2; average opencanopy difference of 42%), than the differences observed in periods in which no snow precipitation was
observed (periods without superscript ‘‘c’’ in Table 2; average open-canopy difference of 24%). Nevertheless,
in periods with no snow precipitation, the difference between SD changes in open and canopy areas varied
depending on the particular event, and the involvement of compaction or a combination of both compaction and melting. However, greater SD loss generally occurred in open areas than beneath the forest canopy, but the magnitude varied substantially among analysis periods. Despite the variability of SD losses in
open and canopy areas, it was clear that the difference in SD between these areas always increased during
the season, even when snowfall did not occur. In such a way, at the end of the snow season, snow presence
was observed in openings while it was not observed beneath canopies. The most plausible reason for this
behavior is the energy balance difference between both areas, because the increase of long-wave emission
and extinction of short-wave originated by the canopy [Pomeroy et al., 2009], enhanced as the season proceeds. Moreover, it must be considered that there was less snow accumulated below the canopy and therefore is less snow to melt and the melt out is earlier.
Lundquist et al. [2013] provided a very comprehensive explanation of how the forest canopy can accelerate
or delay snowmelt. They demonstrated that the effect of the forest canopy on melting rates depends on
the balance between short-wave and long-wave irradiance at a specific site or time of the year, which could
be related to the December–January–February (DJF) average temperature, with a 218C threshold. If average
temperature is above this value, snow ablation beneath canopies takes place 1–2 weeks before it is
observed in adjacent open areas. Panticosa experimental site has a DJF mean air temperature of 0.38C and
as observed, SD average values in Open areas are larger than in Canopy areas which are almost negligible
at the end of the snow season (Table 1 values for last surveys in 2012 and 2013); this result strongly suggests the validity of Lundquist et al. [2013] findings for our analyzed forest stand.
In spite that no direct measurement of snow interception by forest canopies neither of the emission of
long-wave radiation by trees was accomplished, their effect was evident through the analysis of the relationship between SD and distance to the tree trunks. The results clearly showed that SD increased very rapidly with distance from the trunk to the edge of the forest canopy (typically 3.5–4.5 m). This point of
maximum SD is the preferential deposition zone for snow falling from the canopy. At greater distances
from the trunk, the SD decreased slightly and then tended to remain stable, corresponding to open areas.
Similar patterns of SD with distance from tree trunks have been observed in studies focused on individual
trees [Faria et al., 2000; Musselman et al., 2008; Woo and Steer, 1986].
Some works have related SD evolution as a function of trunk distances analyzing solar radiation exposure
[Musselman et al., 2008]. In our study, the high degree of complexity of canopies in the surrounding area
did not allow to obtain a realistic representation of the canopies in the proximities of the study area by
means of the TLS. Such kind of information might enable to assess the distribution of solar radiation across
the study area. In this way, the combination of TLS information with other remote sensing techniques, such
as ALS or aerial photogrammetry, could help to improve such kind of analysis.
This study also demonstrated that isolated wind redistribution events could significantly alter the patterns
of snow distribution within the forest stand. This occurred in February 2012, and the pattern of contrasting
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3486
Water Resources Research
10.1002/2014WR016496
accumulation on the windward and leeward sides of the forest stand remained until the snowpack melted.
This is consistent with the findings of previous studies, which concluded that wind in forested areas can
markedly affect snow distribution [Lundquist et al., 2013; Mahat and Tarboton, 2013; Vajda et al., 2006].
Despite the action of strong winds has a large effect on SD distribution at the study site, the SD was generally greater in open areas than areas beneath canopy. Thus, it is still present the open/canopy effect on SD
distribution but with a higher spatial variability (Figure 4; surveys at February and March 2012).
In order to improve the available information to deeply analyze the relation between solar exposure and
SD, and also studying snow wind redistribution in forested areas, future work will combine different remote
sensing techniques with TLS (as aerial photogrammetry) and also more direct measurements of wind and
solar radiation within the study site.
6. Conclusions
Most similar studies of snow variability, including our research, have typically reported long-term average
values in the range 40–60% for the decrease in the snowpack beneath forest canopies relative to open
areas. However, the present study has shown how many factors could be involved in the accumulation and
melting processes in open and forested areas. Of major importance was the SD, with smaller differences
found between the forest canopy and open areas when a thicker snowpack was present. However, the role
of each factor at a specific site or under specific climatic conditions will probably vary geographically and
among years. Thereby, the main conclusions of this study are summarized as follows:
1. The usefulness of TLS for obtaining high-resolution distributed information on SD in a small forested area
was demonstrated, our results being comparable to other techniques and methodologies used in previous snow studies.
2. The presence of a canopy has a major effect on the SD distribution, with an observed average reduction
in canopy areas of 49% (range 14–80%), compared with open areas.
3. Smaller relative differences between SD values for open and canopy areas were observed with thicker
snowpack. Wind redistribution can have a large effect on the snow distribution in forested areas.
Acknowledgments
This research was supported by (i) the
FPU grant; (ii) the JCI-2011–10263
grant; and the research projects; (iii)
‘‘Hidrologıa nival en el Pirineo Central
Espa~
nol: Variabilidad espacial,
importancia hidrol
ogica y respuesta a
la variabilidad y cambio climatico
(CGL2011–27536/HID, Hidronieve),’’
financed by the Spanish Commission
of Science and Technology and FEDER;
(iv) ‘‘Creaci
on de un modelo de alta
resoluci
on espacial para cuantificar la
esquiabilidad y la afluencia turıstica en
el Pirineo bajo distintos escenarios de
cambio climatico (CTP1/12),’’ financed
by the Comunidad de Trabajo de los
Pirineos; and (v) CGL2014-52599-P:
‘‘Ibernieve: estudio del manto de nieve
en la monta~
na espa~
nola, y su
respuesta a la variabilidad y cambio
climatico’’. The database used in this
manuscript has been created by the
authors and can be accessed by
contacting them. The authors thank all
colleagues who helped in the
development of the field experiments.
The authors wish to acknowledge the
editor and the three anonymous
reviewers for their detailed and helpful
comments to the original manuscript.
REVUELTO ET AL.
4. There was an increase in the relative difference in SD between open and canopy areas with time during
the snow season only reduced when a snow accumulation event is observed.
5. Trunk proximity has revealed an influence on SD dynamics, with lower values at closest locations and a
progressive increase of SD for higher distances until the edge of the canopies is reached.
References
Baeriswyl, P. A., and M. Rebetez (1997), Regionalization of precipitation in Switzerland by means of principal component analysis, Theor.
Appl. Climatol., 58, 31–41.
Camarero, J. J., J. Guerrero-Campo, and E. Gutierrez (1998), Tree-ring growth and structure of Pinus uncinata and Pinus sylvestris in the Central Spanish Pyrenees, Arct. Alp. Res., 30, 1–10.
Connaughton, C. A. (1933), Effect of forest cover on snow melt, For. Work, 9(2), 9–10.
Deems, J. S., S. R. Fassnacht, and K. J. Elder (2006), Fractal distribution of snow depth from Lidar data, J. Hydrometeorol., 7(2), 285–297, doi:
10.1175/JHM487.1.
Deems, J. S., S. R. Fassnacht, and K. J. Elder (2008), Interannual consistency in fractal snow depth patterns at two Colorado mountain sites,
J. Hydrometeorol., 9, 977–988, doi:10.1175/2008jhm901.1.
Deems, J. S., T. H. Painter, and D. C. Finnegan (2013), Lidar measurement of snow depth: A review, J. Glaciol., 59(215), 467–479, doi:10.3189/
2013JoG12J154.
Egli, L., T. Jonas, T. Gr€
unewald, M. Schirmer, and P. Burlando (2012), Dynamics of snow ablation in a small Alpine catchment observed by
repeated terrestrial laser scans, Hydrol. Processes, 26(10), 1574–1585, doi:10.1002/hyp.8244.
Faria, D. A., J. W. Pomeroy, and R. L. H. Essery (2000), Effect of covariance between ablation and snow water equivalent on depletion of
snow-covered area in a forest, Hydrol. Processes, 14(15), 2683–2695, doi:10.1002/1099-1085(20001030)14:15<2683::AID-HYP86>3.0.CO;
2-N.
Gr€
unewald, T., M. Schirmer, R. Mott, and M. Lehning (2010), Spatial and temporal variability of snow depth and ablation rates in a small
mountain catchment, Cryosphere, 4, 215–225, doi:10.5194/tc-4-215-2010.
Hair, J. F., Jr., R. E. Anderson, R. L. Tatham, and W. C. Black (1999), An
alisis Multivariante, 5a edici
on., pp. 80–107, Prentice Hall, Iberia, Madrid,
ISBN 84–8322-035-0.
Harpold, A. A., et al. (2014), LiDAR-derived snowpack data sets from mixed conifer forests across the Western United States, Water Resour.
Res., 50, 2749–2755, doi:10.1002/2013WR013935.
Hedstrom, N. R., and J. W. Pomeroy (1998), Measurements and modelling of snow interception in the boreal forest, Hydrol. Processes,
12(10–11), 1611–1625, doi:10.1002/(SICI)1099-1085(199808/09)12:10/11<1611::AID-HYP684>3.0.CO;2–4.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3487
Water Resources Research
10.1002/2014WR016496
Hopkinson, C., M. Sitar, L. Chasmer, and P. Treitz (2004), Mapping snowpack depth beneath forest canopies using airborne lidar, Photogramm. Eng. Remote Sens., 70(3), 323–330, doi:10.14358/PERS.70.3.323.
Jollife, I. T. (1990), Principal component analysis: A beginner’s guide. Part I: Introduction and application, Weather, 45, 375–382.
Jost, G., M. Weiler, D. R. Gluns, and Y. Alila (2007), The influence of forest and topography on snow accumulation and melt at the watershed-scale, J. Hydrol., 347(1–2), 101–115, doi:10.1016/j.jhydrol.2007.09.006.
Kaiser, H. F. (1958), The varimax criterion for analytic rotation in factor analysis, Psychometrika, 23(3), 187–200, doi:10.1007/BF02289233.
Keller, H. M., and T. Strobel (1982), Water balance and nutrient budgets in subalpine basins of different forest cover, in Proceedings of Symposium on Hydrological Research Basins And Their Use In Water Resources Planning, Berne, Swizerland September 21–23, vol. 3, pp.
683–694, Bern, Landeshydrologie.
Kr€auchi, N., P. Brang, and W. Sch€
onenberger (2000), Forests of mountainous regions: Gaps in knowledge and research needs, For. Ecol.
Manage., 132(1), 73–82, doi:10.1016/S0378-1127(00)00382-0.
Link, T., and D. Marks (1999), Distributed simulation of snowcover mass- and energy-balance in the boreal forest, Hydrol. Processes, 13(14),
2439–2452, doi:10.1002/(SICI)1099-1085(199910)13:14/15<2439::AID-HYP866>3.0.CO;2-1.
L
opez-Moreno, J. I. (2005), Recent variations of snowpack depth in the Central Spanish Pyrenees, Arct. Antarct. Alp. Res., 37(2), 253–260, doi:
10.1657/1523-0430(2005)037[0253:RVOSDI]2.0.CO;2.
L
opez-Moreno, J. I., and J. Latron (2008a), Influence of canopy density on snow distribution in a temperate mountain range, Hydrol. Processes, 22(1), 117–126, doi:10.1002/hyp.6572.
L
opez-Moreno, J. I., and J. Latron (2008b), Spatial heterogeneity in snow water equivalent induced by forest canopy in a mixed beech–fir
stand in the Pyrenees, Ann. Glaciol., 49(1), 83–90, doi:10.3189/172756408787814951.
L
opez-Moreno, J. I., and M. St€ahli (2008), Statistical analysis of the snow cover variability in a subalpine watershed: Assessing the role of
topography and forest interactions, J. Hydrol., 348(3–4), 379–394, doi:10.1016/j.jhydrol.2007.10.018.
L
opez-Moreno, J. I., and S. M. Vicente-Serrano (2007), Atmospheric circulation influence on the interannual variability of snow pack in the
Spanish Pyrenees during the second half of the 20th century, Nord. Hydrol., 38(1), 33–44, doi:10.2166/nh.2007.030.
Lundberg, A., Y. Nakai, H. Thunehed, and S. Halldin (2004), Snow accumulation in forests from ground and remote-sensing data, Hydrol.
Processes, 18(10), 1941–1955, doi:10.1002/hyp.1459.
Lundquist, J. D., S. E. Dickerson-Lange, J. A. Lutz, and N. C. Cristea (2013), Lower forest density enhances snow retention in regions with
warmer winters: A global framework developed from plot-scale observations and modeling, Water Resour. Res., 49, 6356–6370, doi:
10.1002/wrcr.20504.
Mahat, V., and D. G. Tarboton (2013), Representation of canopy snow interception, unloading and melt in a parsimonious snowmelt model,
28(26), 6320–6336, doi:10.1002/hyp.10116.
Mahat, V., D. G. Tarboton, and N. P. Molotch (2013), Testing above- and below-canopy representations of turbulent fluxes in an energy balance snowmelt model. Water Resour. Res., 49, 1107–1122, doi:10.1002/wrcr.20073.
Marchand, W. D., and A. Killingtveit (2004), Statistical properties of spatial snowcover in mountainous catchments in Norway, Nord. Hydrol.,
35(2), 101–117.
McNay, R. S., L. D. Peterson, and J. B. Nyberg (1988), The influence of forest stand characteristics on snow interception in the coastal forests
of British Columbia, Can. J. For. Res., 18(5), 566–573, doi:10.1139/x88-082.
Mellander, P. E., H. Laudon, and K. Bishop (2005), Modelling variability of snow depths and soil temperatures in Scots pine stands, Agric.
For. Meteorol., 133(1–4), 109–118, doi:10.1016/j.agrformet.2005.08.008.
Mellander, P. E., M. O. L€
ofvenius, and H. Laudon (2007), Climate change impact on snow and soil temperature in boreal Scots pine stands,
Clim. Change, 85(1–2), 179–193, doi:10.1007/s10584-007-9254-3.
Mizukami, N., and S. Perica (2008), Spatiotemporal Characteristics of snowpack density in the mountainous regions of the Western United
States, J. Hydrometeorol., 9(6), 1416–1426, doi:10.1175/2008JHM981.1.
Mott, R., M. Schirmer, and M. Lehning (2011), Scaling properties of wind and snow depth distribution in an Alpine catchment, J. Geophys.
Res., 116, D06106, doi:10.1029/2010JD014886.
Musselman, K. N., N. P. Molotch, and P. D. Brooks (2008), Effects of vegetation on snow accumulation and ablation in a mid-latitude subalpine forest, Hydrol. Processes, 22(15), 2767–2776, doi:10.1002/hyp.7050.
Musselman, K. N., N. P. Molotch, S. A. Margulis, P. B. Kirchner, and R. C. Bales (2012), Influence of canopy structure and direct beam solar
irradiance on snowmelt rates in a mixed conifer forest, Agric. For. Meteorol., 161(15), 46–56, doi:10.1016/j.agrformet.2012.03.011.
North, G. R., T. L. Bell, R. F. Cahalan, and F. J. Moeng (1982), Sampling errors in the estimation of empirical orthogonal functions, Mon.
Weather Rev., 110(7), 699–706, doi:10.1175/1520-0493(1982)110<0699:SEITEO>2.0.CO;2.
Pomeroy, J. W., and Gray, D. M. (1995), Snowcover accumulation, relocation, and management, NHRI Sci. Rep. 7, 134 pp., Environ. Can., Saskatoon, Sask., ISBN 0–660-15816-7.
Pomeroy, J. W., D. Marks, T. Link, C. Ellis, J. Hardy, A. Rowlands, and R. Granger (2009), The impact of coniferous forest temperature on
incoming longwave radiation to melting snow, Hydrol. Processes, 23(17), 2513–2525, doi:10.1002/hyp.7325.
Prokop, A. (2008), Assessing the applicability of terrestrial laser scanning for spatial snow depth measurements, Cold Reg. Sci. Technol.,
54(3), 155–163, doi:10.1016/j.coldregions.2008.07.002.
Revuelto, J., J. I. L
opez-Moreno, C. Azorin-Molina, J. Zabalza, G. Arguedas, and S. M. Vicente-Serrano (2014a), Mapping the annual evolution
of snow depth in a small catchment in the Pyrenees using the long-range terrestrial laser scanning, J. Maps, 10(3), 379–393, doi:
10.1080/17445647.2013.869268.
Revuelto, J., J. I. L
opez-Moreno, C. Azorin-Molina, and S. M. Vicente-Serrano (2014b), Topographic control of snowpack distribution in a
small catchment in the central Spanish Pyrenees: Intra- and inter-annual persistence, Cryosphere Discuss., 8(2), 1937–1972, doi:10.5194/
tcd-8–1937-2014.
no, D., E. Meier, B. Allg€
ower, E. Chuvieco, and S. L. Ustin (2003), Modeling airborne laser scanning data for the spatial generation of critiRia~
cal forest parameters in fire behavior modeling, Remote Sens. Environ., 86(2), 177–186, doi:10.1016/S0034-4257(03)00098-1.
Richman, M. B. (1986), Rotation of principal components. J. Climatol., 6, 293–335.
Schirmer, M., V. Wirz, A. Clifton, and M. Lehning (2011), Persistence in intra-annual snow depth distribution. 1. Measurements and topographic control. Water Resour. Res., 47, W09516, doi:10.1029/2010WR009426.
Seidel, D., S. Fleck, and C. Leuschner (2012), Analyzing forest canopies with ground-based laser scanning: A comparison with hemispherical
photography, Agric. For. Meteorol., 154–155, 1–8, doi:10.1016/j.agrformet.2011.10.006.
Trujillo, E., J. A. Ramirez, and K. J. Elder (2007), Topographic, meteorologic, and canopy controls on the scaling characteristics of the spatial
distribution of snow depth fields, Water Resour. Res., 43, W07409, doi:10.1029/2006WR005317.
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3488
Water Resources Research
10.1002/2014WR016496
Trujillo, E., J. A. Ramirez, and K. J. Elder (2009), Scaling properties and spatial organization of snow depth fields in sub-alpine forest and
alpine tundra, Hydrol. Processes, 23(11), 1575–1590, doi:10.1002/hyp.7270.
Vajda, A., A. Ven€al€ainen, P. H€anninen, and R. Sutinen (2006), Effect of vegetation on snow cover at the northern timberline/a case study in
Finnish Lapland, Silva Fennica, 40(2), 195–207.
Varhola, A., N. C. Coops, M. Weiler, and R. D. Moore (2010), Forest canopy effects on snow accumulation and ablation: An integrative review
of empirical results, J. Hydrol., 392(3–4), 219–233, doi:10.1016/j.jhydrol.2010.08.009.
Veatch, W., P. D. Brooks, J. R. Gustafson, and N. P. Molotch (2009), Quantifying the effects of forest canopy cover on net snow accumulation
at a continental, mid-latitude site, Ecohydrology, 2(2), 115–128, doi:10.1002/eco.45.
Vicente-Serrano, S. M. (2005), Las Sequıas Clim
aticas en el Vale Medio del Ebro: Factores Atmosfericos, Evoluci
on Temporal y Variabilidad Especial, De la edici
on, pp. 65–68, Consejo de Protecci
on de la naturaliza de Arag
on, Zaragoza, Spain, ISBN: 84–89862-49-4.
Villar, L., J. B. Alonso, and M. P. Errea (1997), Los bosques del Pirineo centro-occidental como reserva ecologica y paisajıstica, in IRATI 97:
Montes del futuro: Respuestas ante un mundo en cambio: I Congreso forestal hispano-luso, II Congreso forestal espaû˘ol, Pamplona
23–27 de junio de 1997, vol. I–II, pp. 647–652.
Winkler, R. D., and R. D. Moore (2006), Variability in snow accumulation patterns within forest stands on the interior plateau of British
Columbia, Canada, Hydrol. Processes, 20(17), 3683–3695, doi:10.1002/hyp.6382.
Winkler, R. D., D. L. Spittlehouse, and D. L. Golding (2005), Measured differences in snow accumulation and melt among clearcut, juvenile,
and mature forests in southern British Columbia, Hydrol. Processes, 19(1), 51–62, doi:10.1002/hyp.5757.
Woo, M.-K., and P. Steer (1986), Monte Carlo simulation of snow depth in a forest, Water Resour. Res., 22(6), 864–868, doi:10.1029/
WR022i006p00864.
REVUELTO ET AL.
CANOPY INFLUENCE ON SNOW DEPTH DISTRIBUTION
3489