Monitoring of gully erosion in the Central Ebro
Basin by large-scale aerial photography taken
from a remotely controlled blimp
J.B. Ries*, I. Marzolff
Faculty of Geosciences/Geography, Institute of Physical Geography,
Johann Wolfgang Goethe University Frankfurt, Senckenberganlage 36, P.O. Box 11 19 32,
D-60054, Frankfurt, Germany
Received 13 October 2000; received in revised form 11 July 2001; accepted 20 July 2001
Abstract
Large deep gullies (Span. barrancos) are some of the most important sediment sources in the
semi-arid environment of the Central Ebro Basin. They are incised into the Quaternary valley
bottoms (Span. vales), which are characteristic landforms in this area. In the research project
EPRODESERT (Evaluation of Processes Leading to Land Degradation and Desertification under
Extensified Farming Systems), the development of a large barranco system is being investigated by
different methods, including documentation and monitoring by aerial photography.
Geomorphological forms and processes, such as sheet wash, rill and gully erosion, cannot be
documented sufficiently by conventional remote sensing methods. Spatial and temporal resolution of
satellite sensors as well as of conventional aerial photography do not correspond to the scale and
dynamics of geomorphological processes.
With a specially designed hot-air blimp as a sensor platform, large-scale aerial photographs were
obtained from the Barranco de las Lenas specifically aimed at the scientific demands (very high
spatial and temporal resolution). The development of the gully is documented and its dynamics are
evaluated by a sequence of six aerial photographs taken between 1995 and 1998.
D 2003 Elsevier Science B.V. All rights reserved.
Keywords: Gully erosion; Monitoring; Large-scale aerial photography; Blimp system; Central Ebro Basin
0341-8162/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved.
PII: S0341-8162(02)00133-9
* Corresponding author.
E-mail address: j.b.ries@em.uni-frankfurt.de (J.B. Ries).
www.elsevier.com/locate/catena
Catena 50 (2003) 309­328
1. Introduction
1.1. Problems and questions
Gullies present the most conspicuous erosion forms in the Mediterranean region.
Incising the Quaternary sediments of many valley bottoms, they frequently threaten
agricultural fields and plantations. In Spain, gullies (Span. barrancos) are considered as
the main sediment source responsible for the rapid siltation of reservoirs, which are of vital
importance for supply with drinking and irrigation water. Poesen et al. (1996, pp. 257f)
found a sediment rate of 9.7 m3
ha 1
year 1
and estimate that gully erosion contributes
80­83% of total sediment produced on rangelands in Southeast Spain (Poesen and Hooke,
1997, p. 172). Land use and land use change within the catchment have a clear impact on
gully erosion, and almond groves have been shown to be the location of the most active
gully heads (Oostwoud Wijdenes et al., 2000, p. 159).
In Northeast Spain, particularly in the almost completely cleared-out cultivated steppe
land of the Inner Ebro Basin, gullying is a characteristic and widespread phenomenon,
incising and dissecting the flat valley bottoms built up by Quaternary sediments which are
characteristic landforms in this area (Pellicer Corellano and Echeverría Arnedo, 1989, pp.
119ff).
Since the beginning of the 1990s, set-aside programmes of arable land have been
supported by subsidies of the EU with the result that the number of abandoned fields has
increased. This young fallow land shows very complex fluvial­geomorphologic pro-
cesses. Runoff is very high and erosion rates increase compared to those of arable land.
When the former arable land is no longer ploughed, harrowed and rolled regularly, as was
usual in the traditional dry farming system, soil crusts heavily reducing local infiltration
capacity develop on the sandy silts to silty loams. Runoff coefficients between 20% and
89% are the result (Ries et al., 2000, p. 101). Runoff magnitudes like these strongly favour
linear erosion forms like rills and gullies. The velocity of their development is generally
accepted to be determined by the activity at the head cut, and retreat is mostly controlled
by plunge pool undercutting and headwall break-off.
1.2. Aims of the study
Within the scope of the research project EPRODESERT (Evaluation of Processes
Leading to Land Degradation and Desertification under Extensified Farming Systems in
Northeast Spain, funded by the Deutsche Forschungsgemeinschaft), which investigates the
interaction of vegetation succession, geomorphodynamics and land use, large-scale aerial
photographic monitoring of such a gully head was carried out. In contrast to previous work
by Soriano (1993), Barrón et al. (1994) and Vandaele et al. (1997), no conventional aerial
photographs were used for this study, as their image scales would be too small. Changes at
the gully head occur in magnitudes of tens of centimeters. They cannot be mapped from
standard aerial photography like the Vuelo Nacional de Espan~a (1:20,000­30,000) with
sufficient precision, nor can process dynamics be sufficiently monitored due to the low
repeat cycle of these aerial surveys. Therefore, survey methods need to be developed
enabling the detection and documentation of changes at the gully head. The aim of this
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328310
paper is to present a remote sensing method using a remote controlled hot-air blimp as a
platfrom for photographic cameras which was designed to meet the needs of spatial and
temporal resolution for process monitoring. A detailed description of the survey system
and techniques is followed by the example of an image series at a gully head, which was
deliberately chosen for its little changes in order to show the high potential of this
monitoring method.
1.3. Study area
In the Central Ebro Basin, the study area María de Huerva is located in the Val de las
Lenas, an eastern tributary of the Huerva river, which flows into the Ebro at Zaragoza
about 15 km north (Fig. 1). Flat-lying Miocene gypsum, marl and clay series with
interbedded limestone and sandstone form the bedrock of the highly dissected slopes of
the mesa Plana de Zaragoza (ITGE, 1998). The series between 400 and 500 m a.s.l., which
are dominated by gypsum in the middle part, form an impressive erosion landscape with
mostly straight to slightly convex upper and mid slopes and provide the silty material
composing the Quaternary valley fillings. The slopes turn sharply into scree covered glacis
which change smoothly into the plane valley bottom. The valley bottom fillings are mostly
Holocene (Pen~a Monné et al., 1993; Sancho et al., 1991) but on the base partly Late
Pleistocene (Andres et al., 2000). They are up to 20 m thick and consist of interbedded
Fig. 1. Study area Val de las Lenas, María de Huerva, Central Inner Ebro Basin.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328 311
layers, 20­60 cm thick, of silty loam, sandy loamy silt, sandy silt and argillaceous loam,
interrupted by discontinuous stone and gravel layers with silty sandy matrix, which are 5­
20 cm thick (Table 1). In the margins, coarser material often prevails as fanglomerates,
which result from episodic removal from the slopes of the tributaries following intense
precipitation.
The flat valley bottoms are used for agriculture and today show a large percentage of
fallow land. On young fallow land, the slow vegetation succession with sparse therophyte
vegetation (Hordeum murinum) provides only low cover owing to the semi-arid climate
with 6 arid months and a total precipitation of 320 mm/year. After several years, the first
dwarf shrubs (Artemisia herba-alba) start to grow, but a vegetation cover of over 50% can
be observed only after 6­8 years (Marzolff, 1999, pp. 108ff). Elder fallow land, too, often
shows a vegetation cover under 60% and is covered with sparse steppe grass vegetation
(Lygeum spartum) and open shrub land (Salsola vermiculata, Rosmarinus officinalis,
Thymus vulgaris). Between them crusts of lichens, which seal the surface, are developed.
Although the pore volume is much higher than that of younger fallow land due to
bioturbation in the edaphon, and although this results in an increase of the average
infiltration rate and decrease of runoff and erosion rates, runoff coefficients of up to 76%
were also found on elder fallow land. In particular, the terrace edges on old fallow land are
starting points for incision and regressive erosion. Here, sheet wash basins and bank
gullies can be observed which act as collector and pathway for overland flow towards the
gully head.
2. The hot-air blimp monitoring method
Blimps and balloons as well as kites, model airplanes and helicopters have been used
before by scientists as unmanned platforms for photographic and video cameras (see for
example Batut, 1890; Bu¨rkert et al., 1996; Palacio-Prieto and López-Blanco, 1994; Walker
and de Vore, 1995). However, the combination of an open hot-air system with the shape of
a zeppelin (Fig. 2) is unusual and has been specially designed in order to combine the
advantages of both systems (Marzolff and Ries, 1997).
Table 1
Grain size distribution of the layers of the valley filling in the Val de las Lenas, María de Huerva, according to
DIN (German Industry Norm) in AG BODEN, BODENKUNDLICHE KARTIERANLEITUNG (KA4) (1994, p. 134), no
gypsum disaggregation
Situation below surface [cm] Sand [%] Silt [%] Clay [%] Solutes, mainly gypsum [%]
1 (crust) 12.9 80.2 6.9 30.0
28 33.1 58.8 8.1 35.6
52 12.1 79.9 8.0 25.2
76 12.4 76.8 10.8 26.0
85 17.4 70.3 12.3 39.0
150 5.5 86.9 7.6 27.0
310 21.5 70.4 8.1 26.6
380 23.9 67.2 8.9 28.5
420 7.5 88.2 4.3 31.5
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328312
2.1. Blimp envelope
The blimp body or envelope is made of rip-stop nylon and has a volume of 220 m3
.
When inflated, it measures approximately 11 m in length; it is just over 8 m high and
about 5 m wide. In contrast with conventional airships, the blimp body is not supported
by a framework: the shape of the envelope is given by the pressure of hot air only. The
blimp's three fins are supplied by an inflation system through a tunnel along the
blimp's back, ensuring a good distribution of hot air pressure throughout the blimp
body.
Where the envelope is open at the bottom, a scoop of fire-resistant material encloses the
burner system from three sides and guides front breezes from the blimp's nose towards the
burner and into the body, further increasing its inner stability.
2.2. Burner system
The burner system, electronic controls and cameras are stored in a metal storage frame
suspended from steel cables. Four spiral burners are mounted on top of the frame so they
are enclosed by the scoop from three sides. They are powered by liquid propane gas from
an 8-l aluminium gas bottle. Propane is used in many countries as cooking gas and can be
bought cheaply from gas supply stations.
Fig. 2. The hot-air monitoring blimp of the Institute of Physical Geography of Frankfurt University.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328 313
With the help of a small pilot flame, the main burner is ignited when the remotely
controlled electronical valves closing off the connection to the propane bottle are opened.
The liquid gas provides an extremely powerful quadruple flame. A gas filter is installed
between bottle and burner to prevent dirt and grit entering and blocking the valves.
2.3. Sensor system
The sensor system comprises two single-lens reflex cameras with motor drives, which
are mounted against each other with parallel camera bases onto an aluminium frame. The
cameras are suspended from a 360j electronic turntable by a damped cardan joint, which
ensures a vertical optical axis at any time. With focal distance set at infinity and shutter
speed set to 1/125 or faster, photographs are taken in automatic aperture mode using an
electronic trigger connected to a remotely controlled switch.
The twin camera systems allows photographs to be taken simultaneously with different
film types (normal or infrared colour) or with different focal lenghts (50 or 28 mm),
respectively. KODAK EKTACHROME Elite 100 films for transparency slides with a
skylight filter were used for the images appearing in this study. A wicker basket is used for
storage and protection of the camera system. As the cameras are completely protected
inside the basket and concealed from the view at the distance, two marker booms fixed to
the camera system protrude from the basket, indicating the direction of the oblong image
format to the blimp pilot on the ground.
The use of photographic cameras is, of course, not imperative, as any sensor could be
attached which can be remotely controlled and is not (like scanners) dependent on specific
flying velocities. The weight limit given by the blimp's load carrying capacity is
approximately 25 kg. With recent development in the field of digital photography, high-
end digital cameras as well as video cameras are now offering increasingly high resolution
and could in future provide a true alternative even to photographic films--particularly
when further digital processing and analysis is intended.
2.4. Remote control and electronics
Burner system, camera trigger and turntable are controlled from the ground by a
commercial remote control device. Separate channels are used to control electronic
switches opening pilot flame and main burner valves, triggering electronic ignition and
camera shutter and operating the camera turntable. A 12-V storage battery powers
electronic ignition, main burner valve and camera turntable, a second 5.6-V storage
battery supplies the receiver, pilot flame valve and camera trigger. High security is ensured
by the use of a PCM fail-safe system which automatically sets all valves and other
functions to OFF and thus cuts the gas supply in case of transmission faults.
2.5. Tether ropes
While burner and camera functions are remotely controlled, the blimp as a passive
unmanned airship has to be steered with tether ropes from the ground. It is guided with the
help of two special lightweight ropes (polyester coated dyneema, 2.5 mm) attached to the
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328314
blimp's nose. As wind pressure against the rope and blimp as well as slant rope angles
have to be taken into account, 500-m long tether ropes are required for a maximum flying
height of approximately 350 m. A third rope, suspended from the rear of the blimp near the
sensor system and marked in 5-m intervals, serves as a plumb line indicating ground
position and flying height.
2.6. Photographic survey
The inflation process is the most critical phase in terms of risk for the blimp; at least
three people are needed along the back of the blimp to fix the envelope to the ground and
prevent it from drifting into the burner flames. The air inside the body is heated with long
blasts from the main burner until the pressure is high enough for launching and keeping it
aloft. When the camera system is attached to the storage frame, the survey can be started.
The altitude of the blimp is controlled by short intermittent blasts; depending on their
frequency the airship will rise, float or sink, which requires some experience by the pilot.
One gas bottle filling is usually enough for about 50­60 min flying time (including
inflation); both films and gas bottles can be changed without deflating the blimp as the hot
air keeps long enough to park the airship for several minutes.
While the maximum altitude of the blimp is limited by the length of the tether ropes,
there is practically no lower limit to flying height. Depending on the altitude and focal
length, the photographs vary in scale between approximately 1:200 and 1:10,000, covering
areas from 35 m2
up to 10 ha (Table 2). The blimp can be positioned with the tether ropes
fairly precisely. Stereoscopic coverage can easily be accomplished by towing the blimp
along a straight line across the area.
The complete system, when packed in aluminium boxes and canvas bags, weighs
approximately 200 kg and can easily be transported by minivan. More details about the
technique are given by Marzolff (1999, pp. 34­49).
The hot-air blimp employed for this study is a further development of a smaller version
which belongs to the Institute of Physical Geography of Freiburg University (Marzolff and
Ries, 1997). This smaller system of 100 m3
could be carried by four people and has
successfully been employed in little accessible high mountain regions. In respect to load
carrying capacity and wind susceptibility, the size of this smaller blimp proves, however,
to be at a critical lower limit. Both blimp envelopes were designed by the balloon factory
Gefa-Flug (Aachen, Germany). As a direct result of field experience, earlier versions of the
Table 2
Image scales and areas covered depending on flying height and focal length
Flying height above 50-mm lens 28-mm lens
ground (hg) [m] scale area scale area
10 1:100 4.8 Â 7.2 m (35 m2
) 1:357 8.6 Â 12.9 m (111 m2
)
25 1:500 12 Â 18 m (216 m2
) 1:890 21 Â 32 m (0.7 ha)
50 1:1000 24 Â 36 m (864 m2
) 1:1785 43 Â 64 m (2.8 ha)
100 1:2000 48 Â 72 m (0.4 ha) 1:3570 86 Â 129 m (1.1 ha)
200 1:4000 96 Â 144 m (1.4 ha) 1:7140 171 Â 257 m (4.4 ha)
300 1:6000 144 Â 216 m (3.1 ha) 1:10,700 257 Â 386 m (9.9 ha)
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328 315
control system were greatly improved by the staff of the technical workshop of the Faculty
of Geosciences/Geography at Frankfurt University who constructed and assembled the
burner system, electronic controls and metal storage frame.
Fig. 3. Development of the gully head cut between April 1996 and August 1998. Large-scale topographic map
derived from stereoscopic photography taken in April 1996 (Photogrammetric plotting: R. Thormann, Karlsruhe
Polytechnic University 1997).
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328316
3. Results of gully monitoring at the Barranco de las Lenas head cut
The observed Barranco de las Lenas cuts into a terraced field abandoned approx. 60
years ago. The area covered by the photographs (Figs. 5­10) and map (Fig. 3)
encompasses three terraces separated by low, partly disintegrated stone walls. A sheet
wash basin is situated just upslope of the gully which guides overland flow from the
terrace towards the head cut and onto the first gully level 70 cm below the terrace area.
There are no signs of piping processes in the vicinity of the head cut which could indicate
a potential direction of head cut retreat.
Fig. 4. Daily precipitation 1995­1998 at the weather station of Botorrita, 1.5 km west of the study area. Aerial
photographic survey dates are marked with arrows.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328 317
Within the gully, the drainage line continues beyond a second 50-cm drop where the
gully narrows at the terrace edge of the second terrace level. In this part of the gully
bottom, several distinct features can be observed: A narrow arch connects a towering
outlier to the terrace surface level and links the middle part of the gully to a deeper branch
to the left. Further into the gully, another remnant of the former surface level is preserved
as an island. To the right of these remnants, a piping hole is situated which is partly filled
with sunken blocks. An isolated pillar remains in the lower right of the hole. Further
downstream from the mapped area, the gully continues into a deep gorge with up to 20-m
high walls which, together with various lateral gullies, forms a widely ramified system.
The large-scale aerial photographs (Figs. 5­10) were taken on October 6, 1995, April
6, 1996, August 18, 1996, April 2, 1997, April 8, 1998 and August 7, 1998, each between
7 and 9 a.m. The low sun casts long shadows from the upper right to the lower left in the
photographs where northeast is up. The images are full-size reproductions (not cropped) of
the uncorrected, i.e. relief-distorted original colour slides. Please note the similarity of
image scale, area covered and orientation, which can be achieved with the blimp system
and facilitates easy comparison of the individual dates. The following paragraphs will
describe the changes at some prominent places (head cut, arch, piping hole and pillar)
which were mapped from the aerial photographs.
The precipitation regime is characterised by high intra-annual and inter-annual
variability; the bimodal precipitation distribution typical for the Western Mediterranean
long-term climate (spring and autumn rainfalls with dry season in summer and a secondary
minimum in late winter) is not distinctive in individual years. Annual rainfall ranges
between 158 and 458 mm during the study period with daily sums reaching up to 40 mm.
Potentially erosive precipitation events (days with more than 10 mm; Ries, 2000, p. 184ff)
can be observed in all seasons (Fig. 4). The dates of the aerial photographic surveys are
marked in Fig. 4; precipitation sums of the monitoring periods are given in Table 3.
3.1. October 1995­April 1996
The photograph taken in October 1995 shows the situation of gully incision into
the former terraces fields (Fig. 5). While the lower and middle terraces exhibit a
Table 3
Summary of precipitation sum, potentially erosive precipitation, days with minimum temperature below 0 jC,
gully retreat and piping hole enlargement during the monitoring period
Monitoring
period
Precipitation
[mm]
Potentially
erosive
precipitation
[mm]
Days with
minimum
temperature
below 0 jC
Gully
retreat
[m2
]
Piping hole
enlargement
[m2
]
Oct. 1995­Apr. 1996 162 43 16 0.72 2.5
Apr. 1996­Aug. 1996 94 40 0 0.38 ­
Aug. 1996­Apr. 1997 245 118 16 1.32 4.0
Apr. 1997­Apr. 1998 396 236 18 0.76 ­
Apr. 1998­Aug. 1998 52 ­ 0 ­ ­
Total 949 437 50 3.18 6.5
Source of climate data: Instituto Nacional de Meteorología.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328318
medium to dense coverage with L. spartum and lichens, the upper terrace (transversed
diagonally by a sheep track) has very low cover due to heavy overgrazing. Within the
sheet wash basins retreating into this upper terrace, the slightly more dense vegetation
of dwarfed bushes and grass has been favoured by moister conditions. The actual head
cut is located 30 cm below the terrace wall and shows all typical characteristics with
vertical wall, undercut and small plunge pool. Below the head wall of the gully, a
broken bath tub dumped here some time earlier can be recognized. To the right, the
gully scarp shows small coves and bites.
Between October 1995 and April 1996 (Fig. 6), 162 mm of precipitation were recorded,
of which 43 mm were potentially erosive. The bath tub below the head wall was partly
filled with material broken off the wall to the right of the head cut. Further to the right,
gully scarp retreat of up to 30 cm can be recognized by the fresh coarse material crumbled
Fig. 5. Head cut of the Barranco de las Lenas, October 1995.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328 319
onto the grasses at the gully bottom. Total changes at the head wall amount to 0.72 m2
.
The piping hole has widened by approximately 2.5 m2
towards the upper left where its
edge, overgrown with L. spartum, has broken off due to undercutting. Fresh material has
piled up on the older sunken blocks within the hole.
The detailed topographic map shown in Fig. 3 was compiled from stereoscopic
photographs taken at this stage of April 1996.
3.2. April 1996­August 1996
Until August 1996, 94 mm of precipitation fell (40 mm potentially erosive) (Table 3).
The bath tub is now all but concealed by 0.38 m2
of material broken off the right head
wall (Fig. 7). Apart from this, no changes can be observed at the head cut, and the left
Fig. 6. Head cut of the Barranco de las Lenas, April 1996.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328320
part below is now overgrown by vegetation, which indicates head cut inactivity during
this last period. The arch as well as the pillar still remain as remnants of the former
terrace surface, and the size of the piping hole is unchanged.
3.3. August 1996­April 1997
While the bath tub now completely covered by material from the face of the wall, the
head cut itself remains unchanged (Fig. 8). The right gully wall has retreated approx-
imately 20 cm near the ground control point PP8 (see Fig. 3). The former outlier has
become an island by the disintegration of the arch which once connected it to the terrace
surface. Altogether, 1.32 m2
of surface was lost to the gully, the largest retreat recorded
during the study period. Precipitation during these 8 months amounts to 245 mm of which
118 were potentially erosive.
Fig. 7. Head cut of the Barranco de las Lenas, August 1996.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328 321
The piping hole has widened by approximately 4 m2
towards the left and upper right,
and the tip of the pillar has collapsed into the hole where it came to a rest on the material
accumulated here earlier.
3.4. April 1997­April 1998
The next 12 months show the highest precipitation sum (396 mm) and highest amount
of potentially erosive rain (236 mm). For the first time, the uppermost head cut has
retreated some centimetres and thus shows recurring activity (Fig. 9). To the right, the
head wall shows small changes as well: a protruding block of material has fallen off the
wall, and some centimetres of scarp retreat towards the eastern sheet wash basin can be
observed further to the right. The fallen tip of the pillar has broken apart. However, total
Fig. 8. Head cut of the Barranco de las Lenas, April 1997.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328322
gully retreat adds up to only 0.76 m2
and no changes could be observed at the piping
hole.
3.5. April 1998­August 1998
There are no further changes to be observed until August 1998: head cut and wall
remain unchanged and are inactive again in this period (Fig. 10). As no signs of flow
marks exist and even the material deposited during the last period shows no signs of sheet
wash, overland flow across the head cut and gully scarp can be excluded. A precipitation
of 52 mm, none of which is potentially erosive, was recorded.
During the monitoring period of 34 months, changes at the head cut and head wall add
up to 3.18 m2
, and a noticeable amount of material has broken off the face of the wall
Fig. 9. Head cut of the Barranco de las Lenas, April 1998.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328 323
itself. This figure represents the total gully enlargement, however, the actual head cut
retreat does not contribute substantially to the gully development. At the gully bottom, the
piping hole expanded by 6.5 m2
at the same time (see Fig. 3).
4. Discussion
The study presented here shows that large-scale aerial surveys with the hot-air blimp as
camera platform enable detailed monitoring of changes occurring at gully head cuts and
walls. The system provides an extremely stable and vibration-free platform for photo-
graphic surveys. This stands in contrast to other large-scale remote sensing systems, which
are usually less inert due to their dependence on motors (model airplanes and helicopters)
or wind (kites). Apart from the necessity of study areas being accessible on foot and
Fig. 10. Head cut of the Barranco de las Lenas, August 1998.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328324
largely devoid of trees, power lines or other objects which could provoke conflicts with the
tether ropes, the use of the blimp is not bound to specific launching places. A serious
problem is wind; although the zeppelin shape ensures a fair stability even in windy
conditions, anything more than a slight breeze will hamper inflation, and with more than
approximately 3 Beaufort, blimp navigation is not feasible. It should also be understood
that the system is not suitable for covering larger areas in a magnitude of square kilometres
with adjacent images.
The blimp system has proven to meet the demands regarding documentation for
spatially and temporally highly differentiated processes: head cut retreat as small as
centimetres can be recognized, effects of sheet wash discerned and loss of material from
the face of the wall can be verified by the presence of sediment deposits beneath even
where the gully scarp remains unchanged. Transport of material at the gully bottom can be
observed even in areas overcast by shadow from the wall owing to the good contrasts and
high resolution of the photographs. This is of considerable importance with respect to the
usually shaded gully bottoms. The temporal flexibility of the system allows high repeat
rates for the survey and, in the present case, enables to distinguish between phases of head
cut activity and inactive periods.
Although the cameras used for this study are small format, non-metric models and do
not have defined photogrammetric properties, they proved to be suitable for the
compilation of precise topographic maps (Fig. 3). Photogrammetric accuracy was con-
siderably improved by the use of the values of inner orientation, which were defined by
camera calibration at the Institute of Photogrammetry and Cartography of Karlsruhe
Polytechnic University (Thormann, 1997). Ground control points, which were marked
with red cardboard signals prior to the surveys and had their coordinates measured with a
total station, enable absolute orientation of stereo pairs as well as geometrical correction of
image time series. This also provides the prerequisite for further digital image processing,
the construction of digital elevation models of various test areas and GIS analysis
(Marzolff, 1999, pp. 76ff, 92ff, 103ff).
During the individual monitoring periods, no significant retreat due to linear flow at
the actual gully head cut could be observed even with above average precipitation
amounts of which up to 60% were potentially erosive. Instead, gravitative processes
dominate at the head wall. Tension cracks, drying or frost cracks might be considered
as a cause, and higher soil moisture can favour these gravitative processes. In
particular, the terrace southeast of the right head wall--where overland flow from
the sheet wash basin upslope (best seen in the centre of Fig. 8) infiltrates during
precipitation events like those recorded in the study period--is well moistened. Here,
the gully scarp retreats towards the east and is bound to tap into the direct flow path
from the sheet wash basin in the near future. Thus, apart from episodic flow at the
head cut, periodic gravitative break-offs at the gully walls have to be considered as an
important factor for the velocity and bearing of the gully development. However,
retreat rates cannot be easily correlated to precipitation amount, as can be seen from
the data summarized in Table 3: Although periods with little precipitation show the
smallest retreats, the peak retreat did not occur during the wettest period. The fact that
the three highest retreat values always occurred during a winter period suggests that
frost dynamics might be a major cause in material as silty as was found in this area.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328 325
No evidence of piping as a precursor for gully development could be found during the
study period.
Based on the observed retreat rate of 3.18 m2
in 34 months, an average annual retreat
rate in the order of 1 m2
year 1
can be estimated for the 60 m2
of the upper head cut area
(above the lower terrace edge) since abandonment 60 years ago. Taking field measure-
ments of gully depth into account, an average sediment yield of 26.7 kg m 2
year 1
can
be calculated, giving evidence of the high sediment contribution of the gully.
These figures do not include material loss within the gully due to piping processes:
within the gully, piping processes play the dominant role for material transport (Table 3).
The loss of material at the edge of the piping hole affects a noticeably larger area than
could be observed at the gully scarp.
From the exemplary study of gully head development presented here, it must be
deduced that sheet wash basins typically developing at terrace edges can be considered
predecessors of the actual gully head cut. The shape of the upper part of the gully head
suggests that it once originated from a similar situation: the narrowing at the edge of the
middle terrace testifies the location of a former sheet wash basin until today. Disregarding
the general similarity of gully development processes, gully shape and its development
thus are subject to specific conditions in terraced terrain.
5. Conclusion
With the hot-air blimp system, a survey tool for detailed monitoring with high spatial
and temporal resolution was developed. The authors feel that this study demonstrates the
as yet little exploited potential of large-scale aerial monitoring for geomorphological
process research. This method allows investigators to close the gap between terrestrial
photography and conventional aerial photographs.
Changes at the scarp of the observed Barranco de las Lenas are manifold and highly
variable. During the study period, they occurred mostly as lateral enlargement at the walls
due to break-offs. Although changes amount to only 3.18 m2
at the surface, this represents
a remarkable loss of material. However, the most important changes were observed at the
gully bottom as a result of piping processes.
Acknowledgements
The authors wish to express their thanks to Dra. M.T. Echeverría Arnedo and Dr.
Juan de la Riva Fernández, Zaragoza University, for introduction to the study area María
de Huerva and valuable support. Many thanks to the staff of the technical workshop of
the Faculty of Geoscience, Frankfurt University, for their tireless efforts to improve
technical details at the blimp system. Prof. Dr. J. Hooke and another anonymous
reviewer have contributed valuable suggestions, leading to the improvement of this
paper.
Research for this study was carried out as part of the project EPRODESERT funded by
the Deutsche Forschungsgemeinschaft (project code Ri-835/1-3).
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328326
References
Andres, W., Ries, J.B., Seeger, M., 2000. Indicators for high active fluvial dynamics during the late Pleistocene--
the sediments of the upper Barranco de las Lenas/Central Ebro Basin. INQUA Commission of the Holocene,
Environmental Changes in Holocene Sequences--Methods, Processes and Correlation. Seville/Spain 27.03.­
31.03.2000. Conference Proceedings, Sevilla.
Barrón, G., Echeverría, M.T., Ibarra, P., Marco, P., Pérez, F., 1994. Algunas consecuencias geomorfológicas del
uso del suelo agrícola en las últimas décadas. La actividad del piping en el bajo valle del Huerva (Zaragoza,
Espan~a). In: Arnáez, J., García-Ruiz, J.M., Gómez Villar, A. (Eds.), Estudios de Geomorfología en Espan~a.
Sociedad Espan~ola de Geomorfología, Logron~o, pp. 255­266.
Batut, A., 1890. La photographie aérienne par cerf-volant. Reprint in: Autha, D., Ne`gre, S., Beauffort, G. de,
Fosset, R. (Eds.), 1988. Labrugie`re-berceau de l'aérophotographie par cerf-volant/birthplace of kite aeropho-
tography. Arthur Batut, 1846­1918, 95­155.
Bu¨rkert, A., Mahler, F., Marschner, H., 1996. Soil productivity management and plant growth in the Sahel:
potential of an aerial monitoring technique. Plant Soil 180 (1), 29­38.
ITGE (Instituto Tecnológico Geominero de Espan~a) (Eds.), 1998. Mapa geológico 1:50,000, 383/27-13 Zarago-
za. Madrid.
Marzolff, I., 1999. Großmaßsta¨bige Fernerkundung mit einem unbemannten Heißluftzeppelin fu¨r GIS-gestu¨tztes
monitoring von Vegetationsentwicklung und Geomorphodynamik in Aragón (Spanien). Freibg. Geogr. Hefte
57. Freiburg i. Br., 226 pp. + 99 pp. map, appendix.
Marzolff, I., Ries, J., 1997. 35-mm photography taken from a hot-air blimp: monitoring land degradation in
Northern Spain. In: ASPRS (Eds.), Proceedings of the First North American Symposium on Small Format
Aerial Photography, Oct 14­17 1997, Cloquet, MN, USA, Bethesda, pp. 91­101.
Oostwoud Wijdenes, D., Poesen, J., Vandekerckhove, L., Ghesquiere, M., 2000. Spatial distribution of gully
head activity and sediment supply along an ephemeral channel in a Mediterranean environment. Catena 39,
147­167.
Palacio-Prieto, J.L., López-Blanco, J., 1994. Using video imagery for gully erosion evaluation. Z. Geomorphol.
N.F. 38 (1), 33­43.
Pellicer Corellano, F., Echeverría Arnedo, M.T., 1989. Formas de Relieve del Centro de la Depresíon Del Ebro
Institución Fernando et Católico (Publicación 1179), Zaragoza. 216 pp.
Pen~a Monné, J.L., Echeverría, M.T., Petit-Maire, N., Lafont, R., 1993. Cronología e interpretación de las
acumulaciones holocenas de la Val de las Lenas (Depresión del Ebro Zaragoza). Geographicalia 30, 321­332.
Poesen, J., Hooke, J., 1997. Erosion, flooding and channel management in Mediterranean environments of
southern Europe. Prog. Phys. Geogr. 21 (2), 157­199.
Poesen, J., Vandaele, K., Van Wesemael, B., 1996. Contribution of gully erosion to sediment production on
cultivated lands and rangelands. In: Walling, D.E., Webb, B.W. (Eds.), Erosion and Sediment Yield:
Global and Regional Perspectives (Proceedings of the Exeter Symposium, July 1996). IAHS Publ.
236, International Association of Hydrological Science, Wallingford, U.K, pp. 251­266.
Ries, J.B., 2000. Geomorphodynamik und Landdegradation auf Brachfla¨chen zwischen Ebrobecken und Pyre-
na¨en. Großmaßsta¨biges Monitoring zur Erfassung und Prognose des Prozessgeschehens im Landnutzungs-
wandel als Beitrag zur Methodenentwicklung. Habilitation thesis, Faculty of Geosciences/Geography,
University of Frankfurt. 598 pp.
Ries, J.B., Langer, M., Rehberg, Ch., 2000. Experimental investigations on water and wind erosion on aban-
doned fields and arable land in the Central Ebro Basin, Aragón/Spain. Z. Geomorphol. N.F. (Supplement
121), 91­108.
Sancho, C., Gutiérrez Elorza, M., Pen~a Monné, J.M., 1991. Erosion and sedimentation during the upper Holocene
in the Ebro depression: quantification and environmental significance. In: Sala, M., Rubio, J.L., García-Ruiz,
J.M. (Eds.), Soil Erosion Studies in Spain. Geoforma Editiones, Logron~o, pp. 219­228.
Soriano, M.A., 1993. Sequimiento fotográfico de la erosión en la Val de las Lenas (Zaragoza). Estudio preliminar.
Rev. Acad. Cienc. Zaragoza 48, 185­194.
Thormann, R., 1997. Analytisch­photogrammetrische Stereoauswertung von Kleinbilddias fu¨r ein digitales
Gela¨ndemodell und eine topographische Karte. Fachhochschule Karlsruhe, Fachbereich Geoinformationswe-
sen. Unpublished diploma thesis 1997. 85 pp.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328 327
Vandaele, K., Poesen, J., Marques de Silva, J.R., Govers, G., Desmet, P., 1997. Assessment of factors controlling
ephemeral gully erosion in southern Portugal and central Belgium using aerial photographs. Z. Geomorphol.
N.F. 41 (3), 273­287.
Walker, J.W., De Vore, S.L., 1995. Low Altitude Large Scale Reconnaissance: A Method of Obtaining High
Resolution Vertical Photographs for Small Areas Denver. Interagency Archaeological Services, Denver,
Colorado. 161 pp.
J.B. Ries, I. Marzolff / Catena 50 (2003) 309­328328