Morphostratigraphic framework of the Venice Lagoon (Italy) by very shallow water VHRS surveys: Evidence of radical changes triggered by human-induced river diversions



[1] This study is mainly based on a wide Very High Resolution Seismic (VHRS) survey that utilized an ad hoc technique designed for investigations in very shallow waters (about 1 m depth). This method allowed the acquisition of excellent images of the subsurface down to 15–20 m b.s.l. with a resolution of about 10 cm. Buried geomorphological features, such as fluvial channel-levee systems and tidal channels, were imaged for the first time in the shallows and provided new insight into the Holocene evolution of the southern lagoon basin. Furthermore, the new seismic data were used to reconstruct the morphostratigraphic framework of the Venice Lagoon. We provide an Upper Quaternary morphostratigraphic model of the Venice Lagoon and present some evidence of radical changes resulting from human-induced river diversion in the sedimentary regime and in the morphological setting of the southern basin that has occurred over the last millennium.

1. Introduction

[2] The Venice Lagoon is the largest lagoon in the Mediterranean basin area. It is oblong and arched in shape, 50 km long and 8–14 km wide, and stretches across an area of about 550 km2. Three inlets permit the exchange of water with the Adriatic Sea. The morphology of the lagoon consists of shallows, tidal flats and channels, salt marshes and islands, fish farm ponds and reclaimed areas.

[3] The lagoon originated during the Holocene transgression of the Adriatic Sea onto the northeastern part of the Late Pleistocene Po plain. Its evolution has been subjected to a complex combination of natural processes and human interventions, such as land subsidence, eustacy [Fontes and Bortolami, 1973; Carbognin and Tosi, 2002; Tosi et al., 2002; Brambati et al., 2003; Teatini et al., 2005], salt water intrusion [Rizzetto et al., 2003], sea floor erosion, river diversions, and inlet and channel dredging [Gatto and Carbognin, 1981].

[4] The Brenta River system was the major factor in the Late Pleistocene-Holocene evolution of the central and southern Venice Lagoon. At present, the distal margin of the Late Pleistocene Brenta River alluvial fan is exposed in the watershed NW of Venice [Tosi et al., 2007].

[5] Abandoned riverbeds recognized in the central and southern catchments mainly correspond to Late Holocene courses of the Brenta River, whereas some of the southernmost traces have been related to ancient fluvial systems of the Adige and Po Rivers (Figure 1).

Figure 1.

(a) ASTER satellite image of the southern Venice coastland. A shaded relief map of the lagoon bottom is superimposed on the image. Red lines: paleoriverbeds; orange lines: paleobeach ridges; dark violet and light violet areas: present lagoon channels and hydrographic network, respectively; green areas: salt marshes. White lines highlight the present courses of the Brenta and Bacchiglione rivers. Black dots: core locations. Inset: map of the seismic surveys. (b) Cores with paleoenvironments, stratigraphic descriptions, and chronological data [Serandrei Barbero et al., 2006; Tosi et al., 2007].

[6] Indications of paleobeach ridges, extending in the NE-SW direction, were identified by analyzing a digital elevation model (DEM) of the lagoon bottom (Figure 1). The westernmost paleobeach ridge is related to an ancient shoreline found in the coastal plain south of the Venice Lagoon, and it likely dates back to about 4.5 ka BP. In the last 4 ka, depositional activity of the rivers filled the back-barrier lagoon and the surrounding swamps, causing the eastward migration of the littoral zone. This is indicated by the clear evidence of a complex system of ancient beach ridges and sand dunes [Rizzetto et al., 2003] (Figure 1).

[7] Since Roman times, the lagoon has been considered a source of security against enemies because of its unique environmental setting, and several hydraulic projects have been carried out over the centuries with the aim of preserving the lagoon. The first major intervention was the diversion of its major fluvial tributaries towards the sea to avoid sediment input into the lagoon and the resulting risk of the conversion of the lagoon into marshland. This triggered an abrupt reversal in the natural evolution of the lagoon, in particular resulting in the deepening of the lagoon basin so that marine characteristics began to prevail. In the southern Venice watershed, these processes were triggered in the 1540 by the diversion of the lower course of the Brenta-Bacchiglione River system into the Adriatic Sea south of the lagoon (Figure 2).

Figure 2.

(a–d) Schematic evolution of the southern lagoon basin from 1556 to the present. Figures 2a–2c modified from Favero et al. [1988].

[8] The complexity of the natural and human-induced morpho-sedimentary processes is recorded in the subsoil of the lagoon shallows. To date, few studies have been performed to characterize buried geomorphological features and deposits; research has generally been performed using scattered cores that provided local information, due to the logistical and technical difficulties in investigating this shallow water basin. In addition, studies are complicated by the spatial and temporal variability of sediments.

[9] This study was carried out to understand the changes in the lagoon landscape over the Late Pleistocene and Holocene, as well as to distinguish between the natural and human-induced causes of its historical evolution. The results will be useful for other investigations that require knowledge of past hydrologic conditions, such as the effects and consequences of hydraulic works and human interventions or exceptional flooding events on lagoon morphology.

2. Data and Methods

[10] This study used a Very High Resolution Seismic (VHRS) system optimized to operate in water depths less than 1 m. This system consists of an impulsive energy source (boomer) and an electro-dynamic transducer UWAK05 mounted on a catamaran frame. The boomer produces a theoretical minimum phase wavelet with an amplitude spectrum between 200 and 9000 Hz. The maximum impulse rate was 8 pulses per second at 150 J. Different configurations a of single hydrophone and multiple arrays were tested to collected the signals. The best results were obtained with a pre-amplified oilfilled streamer composed of eight piezoelectric elements connected in series with a 2.8 m active array section. These hydrophones have a sensitivity of −63 dB/Volt/microbar and a bandwidth of 100–10,000 Hz.

[11] In order to avoid destructive interference between reflected signals and multiple events from the air/water, to minimize the spatial filtering of the array and to reduce dragging turbulence, the streamer was deployed parallel to the boomer and towed with a 2 m lateral offset as shallow as possible (about 0.3 m beneath the water surface).

[12] The seismic data were digitally recorded in SEG-Y format and were processed with a conventional processing sequence, including gain and spherical divergence, a time variant band-pass filter (300–7000 Hz), and muting. Positioning data were collected using an integrated DGPS navigation system.

[13] Since some seismic sections were badly affected by strong wave-motion effects, we applied a technique similar to the one normally applied to correct seismic sections on land for residual statics. This technique is based on the cross-correlation of each seismic trace with a pilot trace obtained from the data themselves within a designed window. When the procedure works correctly the shift of the cross correlation corresponds to the shift induced by the waves (Figure 3). Seismic unconformities, stratal reflection terminations (erosional truncations, onlaps, and downlaps), and stratigraphic configurations (i.e., acoustic facies) were used to identify and characterize meter-scale seismic units, and standard seismic stratigraphic techniques were used to distinguish seismic sequences and sequence boundaries. Two-way travel time was converted into meters, assuming a seismic velocity of 1500 m/s.

Figure 3.

(a) Comparison between a seismic section with strong wave motion effects and (b) the same section corrected with a residual static procedure.

[14] Reconstruction of the seismic-morpho-stratigraphic units was accomplished by integrating analysis of the seismic data with available geological information, i.e., sedimentological, stratigraphic, geotechnical, mineralogical, textural, and bathymetric data, 14C dating, satellite images, and historical maps, most of them collected by Serandrei Barbero et al. [2006] and Tosi et al. [2007]. Observed crosscutting relationships among stratigraphic elements provided a relative chronology of events (Figure 1), whereas geometries of mapped features gave clues about the processes responsible for their formation [Zecchin et al., 2008].

3. Results and Interpretations

3.1. Seismic-Stratigraphic Framework

[15] The depositional architecture of the Late Pleistocene sequence indicates the aggradation of the alluvial plain, which was locally incised by fluvial channels, during periods of relatively high rates of sediment supply [Zecchin et al., 2008]. The sequence pattern shows stacked channels and vertical successions of ancient buried surfaces, likely corresponding to sediment-starved phases of the plain with the resulting formation of paleosoils. The uppermost Pleistocene deposits date to about 18 ka BP, while the age of the base of the marine transgressive deposits spans from 10 to 6 ka BP (Figure 1) [Tosi et al., 2007].

[16] The seismic profiles show a rather complex Holocene sequence due to changes in relative sea level and sediment supply rates (Figure 4a). The lowest part, corresponding to the Transgressive Systems Tract (TST), has been called “Unit H1” and includes paralic and shallow-marine deposits (estuarine, lagoon, and local shoreface, shelf, and deltaic) that accumulated during the marine ingression that followed the Last Glacial Maximum. The seismic data indicate that H1 is a composite unit that was strongly influenced by basin physiography and the lateral variability of the depositional environments. The base of Unit H1, named S1, represents a sequence boundary amalgamated with the Transgressive Surface (TS), whereas unconformity S2, at the top of Unit H1, represents the maximum marine ingression and is locally linked to sea floor erosion from wave motion (Maximum Flooding Surface, MFS). In the innermost part of the lagoon, unconformity S2 marks the evolution from a back-barrier area (Unit H1) to a coastal/deltaic plain with frequent channel-levee deposits (Unit H2). Unit H2 is rather heterogeneous; towards the mainland, it is the product of deposition in a transitional environment that varied between a delta plain and a lagoon, whereas seaward Unit H2 is characterized by prograding clinoforms, as a result of the overall progradation of the deltaic and littoral systems. In the external area Unit H2 represents the shoreface-shelf prograding system [Donda et al., 2008; Zecchin et al., 2008]. Unit H3, bounded at the base by the S3 surface and at the top by the lagoon floor, corresponds to modern lagoonal sedimentation. It consists of tidal channel deposits that commonly show lateral accretion related to the migration of tidal point bars, as well as multistory fills. Channelized deposits are embedded in the previously accumulated Holocene sediments, especially near the modern inlets and the barrier islands, whereas they are shallower and less common landward, where they pass laterally into laminated sediments draping the structures of Unit H2.

Figure 4.

(a) Simplified architectural scheme of the Holocene deposits in the southern Venice Lagoon. Unit H1 represents the transgressive sequence, whereas Unit H2 is the regressive sequence. Unit H3 consists of sediments deposited during the recent human-induced transgression that followed delta abandonment. A-Active tidal channel; B-Lateral accretion; C-Buried tidal channel; D-Channel-levee system; E-Clinoforms; F-Early Holocene estuarine and fluvial channels; G-Pleistocene river; H-Pleistocene alluvial plain. (b) Late Pleistocene and Holocene complex channelized sequences. (c) Holocene inactive tidal channel fill system. (d) Holocene inactive channel-levee system. (e) Migration of a Holocene tidal channel complex system. Positions of seismic sections are shown in Figure 2.

3.2. Geomorphological Feature Anatomy

[17] Complex channelized sequences including vertical and lateral changes in seismic facies occur both in the Late Pleistocene and Holocene successions, and are signs of continual high hydrodynamics. In Figure 4b, seismic reflectors indicate the presence of two Pleistocene fluvial channels down to about 15 m depth. The channels are separated by a surface with an irregular morphology due to gas seepage and are filled with deposits characterized by thin, mainly horizontal, stacked layers; a point bar can be observed that corresponds to the right side of the larger paleoriverbed section. The Holocene channels in Unit H1 and Unit H2 have regular lateral accretion surfaces related to channel migration, and their basal scours cut the Pleistocene and H1 deposits, respectively. The channel network of Unit H2 became inactive after the human-induced diversion of the Brenta-Bacchiglione River system in the XVI century (Figure 2).

[18] Evidence of the presence of a Holocene deltaic environment prior to the human-induced river diversions is provided by the buried channel-levee systems (Unit H2) that are overtopped by thin layers of modern lagoon deposits (Unit H3).

[19] An example of the channel-levee systems is shown in Figure 4c. These systems are parts of ancient deltas located in the central Venetian catchments (NW of the study area) at least two thousand years ago, and were related to courses of the Brenta River that were active during the Roman Empire and until the Late Middle Ages [Tosi et al., 2007]. No evidence of the presence of these channels is shown in Figure 2; however, stratigraphic data and radiocarbon dating suggest that a fluvial/deltaic environment may have already been present here before the Roman era [Tosi et al., 2007]. Later, in 1457, the flow was partially reactivated by human activities, discharging freshwater and sediments into the nearby lagoon area until the beginning of the 17th century, when the Brenta River was diverted to the south. The depositional architecture of the Holocene deposits mainly indicates conditions of aggradation, while lateral migration appears to be absent. Over the last 40 years, the excavation of the Malamocco Canal to allow oil tanker access to the industrial area has increased the hydrodynamics and has triggered strong erosion processes, as shown by the thinness, or absence, of Unit H3.

[20] Other signs of the presence of Brenta River courses in the lagoon and the formation of deltaic systems are given in Figure 4d, where a buried, superficial Holocene channel-levee complex is recognizable within Unit H2. As shown in the sketch of the lagoon evolution (Figure 2), in the XVI century the area near this seismic section was characterized by the presence of exposed soils and channel-levee systems, which are relics of an ancient Brenta River delta flowing in a direction that was active in the Roman era. Recent studies have confirmed the presence of an ancient fluvial environment close to the inner south-western lagoon margin [Serandrei Barbero et al., 2006], with top sediments about 2 m b.s.l. dating back to about 2 ka BP. Later, the deltaic environment evolved into a salt marsh and then into tidal flats (Unit H3), as shown on the 1692 and 1780 maps, respectively, due to the 1540 Brenta River diversion (Figure 2). At present, since the Brenta River flows into the Adriatic Sea, subsidence and erosion prevail over deposition; as a result, a tidal flat persists and the channel network is more developed than in the past. Even so, remnants of the ancient emerging ridge recognizable on the eastern portion of the 1556 map (Figure 2a) can still be identified, corresponding to the eastern part of the seismic profile.

[21] The very shallow water VHRS profiles also provide images of the detailed architecture and evolution of the tidal channels. The example shown in Figure 4e shows two distinct phases of channel evolution. The first, characterized by a vertical sequence of channel-fill deposits, is representative of aggradation, whereas the second, more recent one (Unit H3) demonstrates the lateral migration of the channel. The latter, developed in a tidal flat after 1556, migrated westward from 1692 to the present (Figure 2).

4. Conclusions

[22] A very shallow water (less than 1 m depth) VHRS system was used to collect seismic profiles with a vertical resolution of about 10 cm. Seismic data integrated with available core analyses allowed the definition of seismic-morpho-stratigraphic model for the Venice Lagoon and for the first time provided detailed images of buried fluvial and tidal channels in the subsurface of lagoon shallows related to the Late Pleistocene and Holocene hydrology. Analysis of the anatomy of these shallows provided details of the architecture of the deposit, such as lateral migration clinoforms, vertical accretion layers, bars, and channel-levee systems. The genetic relationship between the geomorphological features in the lagoon basin and those of the Brenta-Bacchiglione River systems in the watershed was demonstrated by combining the seismic data with remote sensing data and the interpretation of historical maps and bathymetric measurements.

[23] Furthermore, the seismic sections showed the impact of human interventions in preventing the silting of the lagoon basin by sediment discharge from the Brenta-Bacchiglione Rivers. In particular, the human-induced diversion of their fluvial branches toward the sea caused the previous alluvial/deltaic geomorphological features to be buried by deposits and structures of the lagoon environment.

[24] Deltaic structures that were at the surface until five hundred years ago were identified in the inner part of the lagoon, where at present they are submerged under the lagoon shallows.

[25] Future quantitative geomorphologic investigations are required to understand the past hydrologic conditions of the drainage systems and to analyze the formative processes that control the morphology of lowland fluvial river and tidal creek systems.


[26] This study was performed within the framework of the following projects: Co.Ri.La. – R.L. 3.16; VECTOR, Action 3 - R.L. 5 (CLIVEN); CNR-RSTL 809.