Download An analysis of snow cover patterns as derived from oblique aerial

Snow, Hydrology and Forests in Hiai Alpine Areas (Proceedings of the Vienna
Symposium, August 1991). IAHSPubl. no. 205, 1991.
An analysis of snow cover patterns as derived from oblique
aerial photographs
R. KIRNBAUER & G. BLOSCHL
Institut fur Hydraulik, Gewâsserkunde und Wasserwirtschaft,
Technische Universitat Wien, Karlsplatz 13, A-1040 Vienna, Austria
P. WALDHÀUSL & F. HOCHSTÔGER
Institut fur Photogrammetrie und Fernerkundung, Technische
Universitat Wien, GuBhausstraBe 27-29, A-1040 Vienna, Austria
ABSTRACT Snow cover patterns in a 9.4 km2 basin in the Austrian
Alps were examined during the 1989 ablation period. Nine aerial
surveys were conducted at intervals of about two weeks. The oblique
aerial photographs were printed and snow boundary lines were
identified. The metric restitution of the images was performed by
digital monoplotting, making use of a digital terrain model with 25 m
grid spacing. The depletion process is documented and snow cover is
analyzed as a function of elevation, slope and aspect. The snow cover
is found to be better related to elevation and slope during winter
conditions than during the late ablation period. Differences in snow
cover associated with aspect are more pronounced later in the year.
INTRODUCTION
Evaluation of distributed hydrological models has received considerable attention in
the literature (e.g. Loague & Freeze, 1985; Obled, 1990; Moore & Grayson, 1991).
It is evident that distributed snowmelt models require distributed snow cover data
to form the basis of model calibration and evaluation. Runoff observations at the
outiet of a catchment alone are clearly not adequate for the assessment of model
performance because the numerous components of the snowmelt-runoff-process
cannot be disentangled from observations of their collective effect (Blôschl et al..
1991b).
Recently, Leavesley and Stannard (1990) reported on verifying a distributed
snowmelt model by comparing the spatial and temporal variations of simulated and
measured snow covered areas. The latter data were derived from satellite
measurements. In alpine terrain, however, satellite data do not meet the
requirements for modelling with respect to resolution in space and time (Rott,
1987). Snow cover mapping on the basis of terrestrial photogrammetry (e.g.
Rychetnik, 1984) is particularly suited for small scale investigations and requires
the position of the camera to be accessible and free of avalanche hazard.
In this paper the technique of mapping snow cover patterns from oblique aerial
photographs is described. Snow cover patterns are analyzed statistically and by
visual inspection, and relations between snow cover and terrain features are
examined. An application of these patterns for evaluating a distributed snowmelt
model is reported in Blôschl et al. (1991a).
STUDY SITE AND PERIOD
The study was conducted in the Lângental catchment near Kiihtai, Tirol at 47° 12'N,
91
R. Kimbauer et al.
92
11°E in the Austrian Alps (for a map see Fig. 1 in Blôschl et al.. 1991a,
this volume). The basin is 9.4 km2 in area and elevations range from 1900 to
3050 m a.s.l.
The topography is typical of cristalline bedrock (gneiss and micaschist) formed
by glaciers as an U-shaped valley with whaleback-forms and partly covered by
glacial sediments. The lower part of the valley consists of slopes, partly covered by
talus-fans with inclinations of 30-40° (Blôschl & Kirhbauer, 1991). The valley head
faces east, and three prominent cirques form the southeast corner of the basin.
Nearly the whole catchment lies above the timber line. Vegetation is
characterized by a few cembra-pines and larches, by Alpine roses and by Alpine
meadows in the flat areas. Fig. 1 gives an impression of the ruggedness of the
terrain. This aerial photograph was taken on 4 May 1989 and shows the eastfacing slope in the middle of the Langental.
Despite the lower-than-normal winter precipitation the depletion period in
1989 lasted from late April until late July. This was a consequence of frequent
snowfalls throughout spring and temperatures well below the long-term seasonal
average.
FIG. 1 Aerial photograph of the east-facing slopes and the valley
floor of the central part of the Langental, 4 May 1989. By permission
of BMLV, 13088/502-1.6/89.
93
Analysis of snow coverpatterns from oblique aerial photographs
MAPPING OF SNOW COVER PATTERNS
During spring and early summer of 1989 nine aerial surveys were conducted at
intervals of about two weeks depending on weather conditions. Due to the rugged
topography of the Lângental there is no single camera position covering the entire
basin. Therefore, each survey consisted of nine photographs. For ease of
interpretation a fixed flying route at an altitude of 3500 m a.s.l. was adhered to. The
camera was a non-metric Hasselblad 500 C/M, the interior orientation of which
was determined by off-line calibration.
All photographs were printed, the boundary lines of snow cover were identified
and marked manually. Manual identification was preferred to an automatic
procedure because of the potential inaccuracies of the latter method (Good &
Martinec, 1987). Under the specific weather conditions of spring 1989, these
inaccuracies were deemed to be large. Varying lighting conditions would have
impeded assigning fixed grey levels to snow covered or snow free areas and
caused misclassifications. Moreover, manual identification allows to discriminate
areas covered with a shallow layer of new snow from deep winter packs.
Ground control points were taken from 1 : 40 000 topographic aerial photography. Subsequently, ground control and tie points were identified on the oblique
photographs. Together with their coordinates they formed the basis for bundle
triangulation (Kager, 1980), which allowed to determine the exterior orientation
elements for all camera positions.
The boundary lines of snow covered patches were digitized counterclockwise.
The transformation of these lines from the photographs to the map scale was made
on the basis of digital monoplotting (Waldhâusl et ai. 1986; Hochstôger, 1989). In
this method, the image-rays from the camera to the snow cover boundary lines are
intersected point by point with a digital terrain model (DTM) to find the XYZ coordinates of the snow lines (Fig. 2). The inaccuracies of the monoplotting
technology due to intersection angle and DTM spacings were estimated to be in
the order of 5 m thus being negligible for the purposes of this study. In the map
scale individual pictures derived from single photographs were linked in a graphic
FIG. 2 As soon as the centre of projection (PRC) in the coordinate
system of the digital terrain model (DTM) and the orientation of the
bundle of rays are known, any ray S' -> PRC can be intersected with
the terrain (S) (Waldhâusl et ai, 1986).
R. Kirnbauer et al.
editor. Fig. 3 shows a map view of snow boundaries in the Lângental basin on
4 May 1989.
These vector data were rasterized to yield 25 x 25 m square ground elements
matching the grid of the digital terrain model. An element should be classified as
snow covered if more than 50% of its area were snow covered. This was
accomplished by first overlying the vector data by a 5 x 5 m grid. Each of these
elements was classified according to the relative position of its centres to the
boundary lines. A final ground element, consisting of 25 elements of the finer mesh
type, was deemed to be snow covered if it contained 13 or more snow covered
elements.
FIG. 3 Map of snow boundaries (thin lines) 4 May 1989. The
coverage of the aerial photograph (Fig. 1) is marked by thick lines
(redrawn from Hochstôger, 1989).
95
Analysis of snow cover patterns from oblique aerial photographs
RESULTS OF SNOW COVER ANALYSIS
In Fig. 4 the map views of the snow cover patterns for the nine surveys during the
ablation period are presented. Fig. 4 indicates that the maximum extent of the
snow cover (86%) occurred on 4 May 1989. Even on this day a relatively large
portion of the steep east-facing slopes on the left hand side of the Lângental
remained bare. Another prominent feature remained constant throughout the
ablation period: ridges, especially those between the cirques in the south-east of
the catchment, were generally bare, whereas in the cirques themselves the snow
cover did not disappear until late July.
A statistical analysis of the snow cover may be used for quantifying the
impressions gained from visual inspection and a-priori-knowledge. Fig. 5
FIG. 4 Snow cover patterns as mapped from nine surveys from
March to July 1989. White areas are snow covered, black areas are
bare.
96
R. Kirnbauer et al.
summarizes results of such an analysis performed for 20 March and 5 July, 1989.
The late winter distribution of snow (20 March, Fig. 5a,b) is characterized by
nearly 100% snow cover on areas inclined less than 35°, irrespective of aspect or
elevation. West and east-facing slopes exhibit a similar decrease in snow cover
above that limit. North and south-facing slopes, however, show a slight
asymmetry, particularly in the lower portion of the basin: the former ones are
completely covered up to an inclination of about 30°, whereas the percentage of
snow cover of the latter ones sharply decreases from 20° on. Slopes steeper than
60° were virtually never snow covered.
On 5 July (Fig. 5c,d) there were significant differences between areas below
and above 2400 m a.s.l. The lower portion of the basin was nearly bare whereas the
upper part was heavily snow covered. Again, west and east-facing slopes yield
similar results whereas south-facing slopes were significantly less snow covered
than north-facing slopes. These differences in aspect are more pronounced than
those on 20 March and apply to both the lower and the upper portion of the basin.
MARCH 20, 1989
100
Î?
80
Cd
^
o
o
g
JULY 5 , 1989
^
f/^'
o)
-i
J
60
40
60"
40"
y
(/J
20 , /
20"
BELOW 2400m
ABOVE 2400m
/
-40
NORTH
-20
0
20
SLOPE (DEC)
- ^ \
40
60
SOUTH
100
t
Î? °
CE:
-40
NORTH
-20
0
20
SLOPE (DEG)
40
60
SOUTH
100
8
£
o
c)
ao"
/
'
100
60
d
>
80"
"\
60"
\
O
g
40
in
20"1
-40
WEST
40"
\\
1
V
-20
0
20
SLOPE (DEG)
40
60
EAST
20"
/I——\
-40
WEST
-20
0
20
SLOPE (DEG)
40
60
EAST
FIG. 5 Statistical analysis of snow covered and snow free 25 x 25 m
ground elements in the Lângental on 20 March (a,b) and 5 July (c,d)
1989. Snow cover is plotted versus slope distinguishing between the
four cardinal points.
In Fig. 6 the joint distribution of snow cover as a function of slope and
elevation is presented for 4 May and 5 July. Fig. 6a shows an example of late
winter conditions. During this survey the snow cover appears to be closely related
to terrain features. This relation is represented by a snow cover near unity at
elevations lower than 2700 m and inclined less than 35° and an abrupt decrease
above these limits. Fig. 6b gives an example of the late ablation period. As
expected from Fig. 5c,d, the snow cover increased with increasing elevation and
decreasing slope. However, the relationship to terrain parameters is less
97
Analysis of snow cover patterns from oblique aerial photographs
FIG. 6 Percentage of snow cover in the Lângental basin as a function
of elevation and slope, a) 20 March, b) 5 July 1989.
pronounced than during late winter conditions.
DISCUSSION
It may be stated that steep slopes are generally snow free in the Lângental
catchment. Obviously, this may be attributed to sloughing, avalanching and the
combined effect of gravitation and wind. Differences in snow cover on slopes of
different aspects may be easily interpreted: during late winter conditions snow
distribution was basically controlled by accumulation. Only on the south facing
slopes in the lower portion of the basin some melt had occurred reducing the snow
covered area slightly. This may be explained by the combined effects of short wave
radiation and turbulent heat fluxes which, in lower elevations led to melting shallow
layers of new snow adhering to steep rocks. The cumulative effect of short wave
radiation during the ablation period clearly becomes more important in July
producing strikingly contrasting patterns on north and south-facing slopes. Many
authors have reported on this effect (e.g. Buttle & McDonnell, 1987; Elder et al..
1989).
Similarly as with steep slopes, areas at high elevations were generally snow
free. This is probably due to the enhanced exposure to wind compared to lower
elevations.
In contrast to the winter conditions, a weak relationship of snow cover to
terrain features was found for the ablation period. This may be due to both
variations in snow depths and the cumulative effects of melting and redistribution
processes (Blôschl & Kirnbauer, 1991). Similarly, Rychetnik (1987) only explained
a small portion of the spatial variations in the date of disappearance of the snow
cover by terrain features. He found maximum seasonal snow depths to be a better
index. For a small lower alpine catchment Meier & Schadler (1979) found
variations in global radiation to control depletion patterns.
R. Kimbauer et al.
98
CONCLUSIONS
The use of a non-metric medium format camera for periodical aerial photography
together with expert photogrammetric bundle adjustment, DTM-data base and
monoplotting technology proved to be an economic alternative to classical
terrestrial or aerial photogrammetry. Snow cover patterns mapped by these
methods were found to be quite complex. However, some basic structures
remained constant during the ablation period due to topographic effects. During
winter conditions the snow cover is found to be closely related to terrain features.
This relation is represented by a snow cover near unity at elevations lower than
2700 m and slopes inclined less than 35° and an abrupt decrease above these limits
being a consequence of redistribution processes. During the early summer the
cumulative effect of melting becomes more dominant. This results in a less
pronounced relation of snow cover to terrain features and in significantly less snow
cover on south facing slopes. Though attempts at correlating snow cover and
terrain features appear to be encouraging it is important to realize that terrain
features do not influence snow cover per se. They provide no more than a means for
indexing the highly complex redistribution and snowmelt processes. This fact will
limit the accuracy of predicting the snow cover distribution.
ACKNOWLEDGEMENTS We would like to acknowledge the contribution of the
Tyrolean Hydro-electric Power Company (TIWAG) who provided the aerial
photographs. This research was supported by a grant from the Fonds zur
Fôrderung der wissenschaftlichen Forschung under project Nos. P6387P and
P7002PHY.
REFERENCES
Blôschl, G. & Kirnbauer, R. (1991) An analysis of snow cover patterns in a small
alpine catchment. Submitted to J. Glaciol.
Blôschl, G., Kirnbauer, R. & Gutknecht, D. (1991a) A spatially distributed
snowmelt model for application in alpine terrain. In: Snow. Hydrology and
Forests in High Alpine Areas (Proc. Vienna Symp., August 1991). This
volume.
Blôschl, G., Kirnbauer, R. & Gutknecht, D. (1991b) Distributed snowmelt
simulations in an alpine catchment, 1. Model evaluation on the basis of snow
cover patterns. Submitted to Wat. Resour. Res.
Buttle, J. M. & McDonnell, J. J. (1987) Modelling the areal depletion of snowcover
in a forested catchment. J. Hydrol. 90.43-60.
Elder, K., Dozier, J. & Michaelsen, J. (1989) Spatial and temporal variation of net
snow accumulation in a small alpine watershed, Emerald Lake basin, Sierra
Nevada, California, U.S.A. Annals of Glaciology 13. 56-63.
Good, W. & Martinec, J. (1987) Pattern recognition of air photographs for
estimation of snow reserves. Annals of Glaciology 9.76-80.
Hochstôger, F. (1989) Ein Beitrag zur Anwendung und Visualisierung digitaler
Gelândemodelle (On the application and visualization of digital terrain
models). Geowissenschaftliche Mitteilungen 34. Verôffentlichungen des
Institutes fur Photogrammetrie und Fernerkundung, Technische Universitât
Wien.
Kager, H. (1980) Das interaktive Programmsystem ORIENT im Einsatz
(Application of the interactive program system ORIENT). Proc. XIV. Congress
of Photogrammetry and Remote Sensing, Comission V. Hamburg 1980.
99
Analysis of snow cover patterns from oblique aerial photographs
International Archives of Photogrammetry and Remote Sensing XXIII,
Comission V, 390-401.
Leavesley, G. H. & Stannard, L. G. (1990) Application of remotely sensed data in
a distributed-parameter watershed model. In: Proceedings of the Workshop on
Applications of Remote Sensing in Hydrology. Saskatoon, February 1990 (ed.
by G. W. Kite and A. Wankiewicz), 47-64.
Loague, K. M. & Freeze, R. A. (1985) A comparison of rainfall-runoff modeling
techniques on small upland catchments. Wat. Resour. Res. 21 (2), 229-248.
Meier, R. & Schâdler, B. (1979) Die Ausaperung der Schneedecke in Abhângigkeit
von Strahlung und Relief (Disappearance of snowcover depending on radiation
and relief). Archiv fur Météorologie, Geophysik und Bioklimatologie, Ser. B 27,
151-158.
Moore, I. D. & Grayson, R. B. (1991) Terrain-based catchment partitioning and
runoff prediction using vector elevation data. Wat. Resour. Res, (in press).
Obled, Ch. (1990) Hydrological modeling in regions of rugged relief. In: Hydrology
in Mountainous Regions. I - Hvdrological Measurements: the Water Cycle
(ed. by H. Lang and A. Musy) (Proc. Lausanne Symp., August 1990). IAHS
Publ. no. 193,599-613.
Rott, H. (1987) Remote sensing of snow. In: Large Scale Effects of Seasonal Snow
Cover (ed. by R. E. Goodison, R. G. Barry and J. Dozier) (Proc. Vancouver
Symp., August 1987). IAHS Publ. no. 166,279-290.
Rychetnik, J. (1984) Methoden der Erfassung und Auswertung der Ausaperung und
Lawinenaktivitât an einem Lawinenhang (Avalanche activities and snowmelt
on an avalanche slope. Methods, compilation and interpretation). In: Snow
Hydrologie Research in Central Europe (ed. by H. M. Brechtel), Mitteilungen
des Deutschen Verbandes fur Wasserwirtschaft und Kulturbau 7, 153-165.
Rychetnik, J. (1987) Snow cover disappearance as influenced by site conditions,
snow distribution, and avalanche activity. In: Avalanche Formation. Movement
and Effects (ed. by B. Salm and H. Gubler) (Proc. Davos Symp., September
1986). IAHS Publ. no. 162,355-361.
Waldhâusl, P., Enthofer, C. & Kager, H. (1986) Bildauswertung durch
Verschneiden des Strahlenbûndels mit dem digitalen Hôhenmodell
(Restitution of images by intersecting the bundle of rays with the digital terrain
model). Qsterreichische Zeitschrift fur Vermessungswesen und Photogrammetrie 74 (3). 155-166.