A new approach for surveying submerged aquatic vegetation in shallow coastal waters: Application of digital echosounder technique for ecosystem management.

A new approach for surveying submerged aquatic vegetation in shallow coastal waters: Application of digital echosounder technique for ecosystem management Patrick Schneider a, Janusz Burczynski b, Agustín Monteoliva c a Infraestructura & Ecología, S.L., C./San Antonio Ma Claret, 186, 4º-2ª, 08025 Barcelona, Spain b BioSonics, Inc., 4027 Leary Way NW, Seattle, WA 98107, USA c Infraestructura & Ecología, S.L., Urb. Los Llanos, 52 bajo, 39011 Peñacastillo, Spain Corresponding author: P. Schneider (bcn@infraeco.es) 1. INTRODUCTION For over 50 years the application of hydroacoustic techniques has been limited mainly to the evaluation of fish and bottom. Although the recognition of certain features in the bottom signal has been used in some applications for obtaining an indication of the presence of a plant layer, this has basically remained a marginal application. A new methodology, developed by Bruce Sabol, USACE Waterways Experiment Station, Vicksburg, under the name SAVEWS (Submersed Aquatic Vegetation Early Warning System) (Sabol and Melton, 1995) and adapted to the Windows® environment through BioSonics® Inc., Seattle, under the name EcoSAV, now allows to convert a digital echosounder into a dedicated tool for assessing submerged aquatic vegetation (SAV). Through this expansion of possibilities, the echosounder technique becomes a comprehensive and unique tool for ecosystem management: The assessment of abundance and distribution of fish and plankton, SAV and seabed classification, besides bathymetric information, provides an almost complete set of data on the principal constituents of aquatic systems. This also makes it a very interesting system especially for environmental consulting, as comprehensive data sets can be obtained at a comparably low cost due to minimised effort (same system, same operator and one single field trip). As all data acquired is georeferenced by default, the easy integration of results into a GIS system is a further advantage. This paper focuses on the application of the new SAV evaluation methodology and the obtained results, as well as its integration into other data sets obtained with the same echosounder system from surveys in two different kinds of aquatic ecosystems. 2. MATERIAL AND METHODS During August and September 2000 the River Asón Estuary (Northern Spain), a marshland of international importance and a Natural Reserve under the authority of the National Park Administration., included 1994 in the RAMSAR listing, has been subject to a high resolution survey in order to asses and map the distribution and abundance of two seagrass species, Zostera marina (eelgrass) and Zostera noltii. With a transect 1 spacing of 20m and one sample each 0,31m (~1 ft) on average, a total area of 387ha has been surveyed, corresponding to about 800.000 samples. The unusually high transect density (Fig. 1) has been chosen for two reasons: The highest possible resolution requested by the direction of the Park Administration on one hand and the time-cost limitations on the other hand. A simple calculation, based on total available time, time needed for loading and unloading the ship as well as the time required to access the harbour as well as the area to be surveyed and finally the ship speed (usually about 5 kn) , resulted in a total length of the cruise track of about 250km. Based on the surface, the approximate distance between transects was then determined by dividing the total surface of the area to be surveyed by the total length of the cruise track (MacLennan and Simmonds, 1992). The field work was performed between end of August and beginning of September, on a total of 11 days. The required scale for the resulting map was 1:500, based on the requirements of the Natural Reserves management, whose decisions and restoration projects emerge from the definition of ecological niches present within the estuary. It is the believe of the Reserves management, that the delimitations of these niches are lost if a resolution corresponding to a scale of 1:500 is not achieved. We considered that the resolution of 20m between transects is adequate for this scale, also taking into account the very high density of sampling units within each transect, as stated above. A second objective of the survey was to establish bathymetric maps and to determine bottom types of the access channels of two local fishing ports, Santoña and Colindres, in order to allow the corresponding administration to take decisions on the need of dredging these areas. Transect spacing here was between 20 and 30m (Figs. 2 and 3). In July 2001 a small survey in the strait of Gibraltar, Southern Spain, on Cymodocea nodosa, known to be present in the area, was performed. Here, transect spacing was considerably higher (175m on average), due to different requirements. The aim of this survey was to assess the general distribution of Cymodocea nodosa, while avoiding to confuse detection of this plant with detections of macro algae (Laminaria sp.), also present in the area, as well as to determine bottom substrate. The area where SAV was expected was restricted to a band parallel to the shore line, up to a bottom depth of about 15-20m, corresponding to an area of about 272ha, while the transects performed in order to determine bottom substrate and bathymetry extended to bottom depths of 80100m, corresponding to a total area of about 1980ha. The area where SAV was expected is the shallow coastal part of this area (Fig. 4). This survey was performed in a total of two days. The equipment used in both surveys is a BioSonics® DT 6000 208kHz split beam digital echosounder. The transducer was mounted in a towed body (BioSonics® BioFin 4’), beaming vertically, and was towed on the port side at about the ships centre, at a mostly constant depth of 52cm (measuring from the transducer surface to the water surface). This difference was accounted for in the post processing. In the Santoña survey, position information was provided by an external DGPS (Trimble 4000 Si base and rover system), connected to the echosounder and providing 2 an accuracy of about 2m. The system was set up with a local base station (GPS receiver 1), transmitting correction signals through a UHF transmitter. These signals were received on board by a UHF receiver that passed the RTCM corrections to the rover (GPS receiver 2). In case of the Tarifa survey, the internal GPS of the echosounder was used, as no particularly high precision was required. The internal GPS currently produces an error of about 20m (with SA off). In both cases, position data is directly associated with the acoustic data and stored on a laptop computer by the echosounder systems acquisition software (BioSonics® VisualAcquisition). Acoustic samples were taken in both surveys at a rate of 8 pings/sec. The software used for analysing the data on SAV (BioSonics® EcoSAV) requires acquisition of data with an extremely low threshold setting. This setting is not suitable for other uses of the acoustic data than SAV evaluation. Thus, data for bottom type evaluation had to be acquired separately with different threshold settings. Parallel to the laptop computer running the acquisition software, a second laptop computer, also connected to the GPS (or DGPS) receiver, was used for real time navigation. For this purpose, navigation software (WINGPS® Pro 3.0, Stentec® Software and GPS TrackMaker 11 by Odilon Ferreira Jr.) was used. The set-up with two laptop computers has the advantage that each individual (the ships captain and the scientist or technician) is provided with his own display, this way allowing for optimal visibility. The ship used during the Santoña survey is a small fibreglass boat with very low draught (15cm), equipped with a 40HP outboard engine. For the survey in Tarifa a 15m diving boat, was used. This ship was especially adapted before the survey in order to provide the means for installing an iron pole needed to tow the body at a safe distance to the ships hull. The EcoSAV software was used later in the lab in order to extract the SAV data from the acquired acoustical raw data. The software uses specific algorithms (Sabol and Melton, 1995), enabling the user to extract data on coverage and height of such vegetation, together with its exact centre position (a mean over 8 measurements is built during the process), from digital echosounder data. After post processing, the georeferenced coverage and height data of the observed vegetation was available in ASCII format for direct inclusion in a GIS system or for further analysis. Further processing and interpolation of the data was performed in order to obtain maps presenting the different data sets in a geographical context. Interpolation of the data was performed with different interpolation methods, depending on the data to be represented. For all interpolations, as well as for the elaboration of the maps, Surfer® 7 (Golden Software, Inc.) was used. Data on SAV coverage from Santoña was interpolated using the “Inverse Distance to a Power” gridding method. With this method, data are weighted during interpolation such 3 that the influence of one point relative to another declines with distance from the grid node (Surfer 7 Help, 1999, Golden Software, Inc.). Omitting the use of a smoothing factor, the method acts as an exact interpolator, i.e., the values of the original data points are not modified. Further, a high weighting power (4) was used to reduce the influence of neighbour points. The data on SAV height in the Tarifa study were interpolated with the same method, but using a weighting power of 2 and a smoothing factor of 3. Data from bottom typing for the ports of Santoña and Colindres were interpolated with “Minimum curvature”, while data from bottom typing in Tarifa was interpolated using “Inverse distance to a power” (with a weighting power of 1 and smoothing factor of 3). Tide levels were measured in regular intervals, and the obtained data was used later in order to correct the depth data from acoustics correspondingly. For both surveys, the SAV data set was complemented with an assessment of bottom types present in the area. This requires different echosounder settings, and correspondingly a separate prospection of the area where these data are to be acquired. Processing of the latter data was also performed in the lab, using BioSonics® VBT (Visual Bottom Typing) software. This software provides several different methods for analysing the data. Data from both surveys were analysed using the Fractal Dimension method. We determined empirically that this method produced the results that corresponded best with the data from ground truthing. Classified data output from VBT is georeferenced by default. The SAV data has been ground truthed in both surveys, although to a different extent and in different ways. Given that the Santoña area is an estuary, subject to important changes in water level during tides, most of the samples in the survey area were taken at low tide by foot. Only in some occasions samples had to be taken by divers. A total of 73 physical samples were taken here, out of which 67 samples have been evaluated. The evaluation consisted in measuring the length of the blades, determining the number of sheaves and the wet and dry weight per area. All samples were taken within a 50x50cm frame. The sampling was performed between 2-3 months after the acoustic survey. This obvious problem will be addressed further in the discussion chapter. In the Tarifa survey, only two samples have been taken immediately after the acoustic survey. Based on certain echogram features related to SAV detections, several positions were marked upon detection within the navigation software. After completing the survey, the ship revisited some of these locations and was centred as precisely as possible above these locations. The position was corrected until presence of SAV could be confirmed through readings from the actual echogram (the echosounder was running during this operation). Then a diver confirmed the presence of Cymodocea nodosa. At the first position, this was done visually, at the second position, two divers took physical samples. The presence of Cymodocea nodosa (Fig. 5) was successfully confirmed in both cases. The algorithms used in BioSonics’ VBT software use specific characteristics in the bottom signal for the categorisation of each acoustic sample (ping). In order to assign a 4 specific bottom type to each ping, the characteristic information from each ping is compared to a pre-established library of bottom classes. To establish this library, the user should acquire acoustic data (i.e. record bottom echoes) and at the same time take the physical core samples of that bottom (Anonymous, 1998). For this reason, grab samples were taken at several locations that previously had been sampled acoustically in Santoña as well as in Tarifa. A total of 12 grab samples was taken in the access channel of Santoña harbour, and another 12 grab samples in the access channel of Colindres harbour (Figs. 2 and 3). In Tarifa, a total of 33 grab samples, representative for the different substrate types that can be encountered in the area, were taken. In this survey, the sampling locations were chosen based on results from previous studies. 3. RESULTS Although applied to different species of SAV (Zostera marina and Zostera noltii in Santoña and Cymodocea nodosa in Tarifa) and under different conditions, especially at different depths, the software was able to detect successfully the existence of submerged aquatic vegetation, at least in all cases that have been subject to ground truthing. In Santoña, the general distribution of Zostera sp. was in most cases obvious, as the plant coverage of major parts of the surveyed area is directly visible on low tides. However, a part of the vegetation remains constantly under water. Detections of this part of the vegetation cover remained mostly unverified, though the signal when plants are present could clearly be distinguished (from echogram readings) from signal without plants. However, based on both observations, the encountered distribution of Zostera from interpolated data as shown in the map (Fig. 6), was found to be in good correspondence with the visually observed distribution. Furthermore, all physical samples taken from within the surveyed area confirmed the presence of SAV. In both surveys, patches of SAV were successfully detected, up to depths of 15m, the plant height being only about 15-20cm in this case (Cymodocea nodosa in the Tarifa survey). These detections as well as the canopy height measured by the EcoSAV software have been confirmed at two locations through ground truthing (divers) in Tarifa. Here, Cymodocea nodosa has been confirmed by divers at 10m depth, but identical acoustical signals were observed up to a depth of 15m. Occurrences of Laminaria sp. in the Tarifa survey could be easily distinguished from occurrences of Cymodocea due to their obvious difference in height as detected by the EcoSAV software. In this case, the particular location of Laminaria growth has been indirectly confirmed by local divers who know the area well and visit the spot regularly with diving tourists. Laminaria height was about 1m and more, while Cymodocea had a maximum detected height of 30cm. 5 Table 1a shows the composition of the grab samples in the access channel of Santoña harbour, table 1b the composition of grab samples from the access channel of Colindres harbour. Sample BT1 BT2 BT3 BT4 BT5 BT6 BT7 BT8 BT9 BT10 BT11 BT12 UTM-X 462591.2 462542.9 462520.5 462600.1 462589.6 462532.4 462669.9 462629.1 462546.2 462811.7 462684.6 462603.4 UTM-Y % Stones % Gravel 4810398.5 0 0.53 4810405.0 0 0.03 4810407.7 0 3.78 4810241.8 0 1.22 4810239.7 0 0.48 4810254.4 0 1.44 4810027.1 0 3.52 4810001.8 0 0.29 4810014.4 0 0.49 4809875.2 0 1.83 4809867.1 0 0.13 4809853.1 0 2.95 % Sand 84.25 99.43 86.27 97.07 96.62 98.82 94.28 98.55 97.69 87.85 99.54 96.77 % Mud 14.71 0.23 10.00 1.88 2.88 0.27 2.44 0.87 1.14 10.84 0.19 0.32 Table 1a: Composition of the grab samples in the access channel of Santoña harbour. Sample BT13 BT14 BT15 BT16 BT17 BT18 BT19 BT20 BT21 BT22 BT23 BT24 UTM-X 462502.2 462392.5 462344.0 462518.0 462430.1 462344.1 462544.3 462458.3 462330.0 462564.0 462476.8 462370.0 UTM-Y % Stones % Gravel 4804898.1 0 4.17 4804886.0 100 0.00 4804886.0 100 0.00 4805082.4 0 0.65 4805047.0 0 2.25 4805093.0 100 0.00 4805232.2 0 4.28 4805238.2 100 0.00 4805225.0 100 0.00 4805379.8 0 14.86 4805368.0 0 25.38 4805390.8 0 4.96 % Sand 95.13 0.00 0.00 68.44 97.33 0.00 70.23 0.00 0.00 84.41 73.72 80.30 % Mud 0.92 0.00 0.00 30.96 0.47 0.00 25.47 0.00 0.00 1.75 0.90 15.00 Table 1b: Composition of the grab samples in the access channel of Colindres harbour. 6 Table 2 shows the sediment types determined from the grab samples in Tarifa. Sample A1 A2 A3 A4 B1 B2 B3 B4 C1 D1 D2 D3 D4 E1 E2 E3 E4 G1 G2 G3 G4 H1 H2 H3 H4 I1 I2 I3 I4 J1 J2 J3 J4 UTM-X 260298 260314 260348 260390 260522 260512 260545 260545 260359 260623 260815 260825 260825 260895 260918 260918 260918 260986 260990 260992 260994 261200 261246 261246 261245 262728 262729 262728 262728 262913 262928 262929 262929 UTM-Y 3989314 3989303 3989278 3989241 3989618 3989602 3989583 3989584 3989459 3989993 3989967 3989955 3989955 3990103 3990097 3990097 3990097 3990372 3990379 3990385 3990392 3990506 3990499 3990498 3990495 3991034 3991034 3991034 3991034 3991070 3991055 3991055 3991055 Depth [m] 45.6 46.0 46.8 47.1 36.4 36.6 36.6 36.6 40.6 26.4 26.4 26.4 26.4 25.1 25.1 25.1 25.1 21.4 21.2 21.1 21.1 15.5 15.2 14.9 15.0 10.5 10.2 10.2 10.2 5.0 4.8 4.8 4.8 Sediment type Hard bottom Gravel Hard bottom Hard bottom Coarse sand Hard bottom Hard bottom Hard bottom Stones Hard bottom Fine sand Hard bottom Hard bottom Fine sand Hard bottom Hard bottom Hard bottom Medium/fine sand Hard bottom Hard bottom Hard bottom Hard bottom Hard bottom Hard bottom Fine sand Hard bottom Fine sand Hard bottom Hard bottom Fine sand Hard bottom Hard bottom Hard bottom Table 2: Results of the grab samples in Tarifa. As a distinction between Zostera noltii and Zostera marina was not possible based on the data from acoustics, only coverage with Zostera sp. in the estuary of Santoña was presented in a map (Fig. 6). The results for the bottom classification in the access channels of the ports of Santoña and Colindres, obtained with the BioSonics® VBT Seabed Classifier software, are presented in Figs. 7 and 8. Fig. 9 presents the results for bottom classification from the Tarifa survey, the distribution and delimitation of SAV as well as bathymetric information. 7 4. DISCUSSION 4.1. SAV It has to be pointed out that the original SAVEWS algorithms (and the algorithms used in EcoSAV) have been developed and tested (Sabol et al., 2002) for a specific echosounder system, but above all, for a specific transducer and frequency. It has not been tested for the 208kHz split beam transducer used in this work. With respect to this, the most important deficiency of the Santoña data set is that the physical sampling was done between two to three months after the acoustical survey, i.e., in late autumn and at the beginning of winter. Due to the dynamic nature of the estuary, but moreover due to the natural decline in seagrass coverage after the end of summer, the samples have been of very limited use for testing the performance and reliability of the system. There is few doubt, however, that the presence of SAV is detected correctly, as the signal seen when plants are present, could clearly be distinguished (from echogram readings) from the signal when plants were absent, i.e., characteristic patterns observed on the echogram screen only in areas where presence of SAV could be verified through direct observation, have been used as a basis to verify SAV detections through the software’s algorithms during post processing in the lab, in cases where no direct field observations were available. This is especially true for Zostera marina; the plant is quite robust and therefor produces a clearly recognisable signal. Such echogram verifications have been performed extensively and were also used to eliminate occasional wrong detections of plants at larger depths. It must be pointed out that this procedure is only an indirect and insufficient verification of the obtained data, and that it can only be considered as a preliminary solution, until more suitable data is available for thoroughly testing the systems performance. Correspondingly, the plant height determined by the software did not correspond well to the measured length of the blades. The correlation coefficient between the measured length of blades from 50 physical samplings of Zostera marina and the residual values from interpolation of plant height at the same positions where the samples were taken was 0,36. The main reason for this must be attributed to the fact that physical samples were taken several months after the acoustical sampling, as stated above. But, even if the physical sampling had been done immediately after or before the acoustic sampling, the comparison between interpolated values and physical samples would include the system error plus the error from interpolation. For this reason, it is important to obtain both, acoustic and physical sampling data, at the exactly same positions and at the same time, and then compare the actually measured data at these positions. Currents are a further factor that impacts height measurement through the software. For this reason, actual blade length can not be compared directly with the computed canopy height, as seagrass (and other SAV) is usually inclined due to currents (Sabol et al., 1997). Instead, ground truthing for testing system performance has to include the determination of canopy height in situ, i.e. the actual height of the plant cover above the 8 bottom must be determined (through divers, e.g.) and compared to the software’s output, rather than the length of the blades as measured in the lab. The reason for the very small number of samples taken during the Tarifa survey is the fact that the survey had to be finished prematurely due to the forecast of extremely adverse weather conditions, which then lasted several days. However, both samples have been taken following indications only based on echogram readings, and subsequently SAV was correctly detected through the EcoSAV software at these exact locations and the same height that was measured in situ was determined by the soft, too. 4.2. Bottom typing Although sampling of the different bottom types present in the area was performed in both locations as described above, a bottom library could not be established in a reliable fashion in both cases, as strong currents displaced the grab to an unpredictable extent. Hence, it could not be determined if the grab sample corresponded to the acoustic samples taken at the same surface location. Especially the Gibraltar strait is known for its strong currents. In Santoña the problem arose out of the fact that strong tides cause changing currents, and the virtual current-free moment at high tide in practice is almost not appreciable and moreover provides a window of opportunity too small to perform the necessary grabs in a reasonable way. However, due to the relatively small depth, the calibration data from Santoña and Colindres are more reliable, as the possible derivation of the position is limited, but nevertheless it must be taken into account that a highly patchy bottom can still affect the quality of the calibration data. Under the assumption that the derivation of the grab was inside acceptable limits, bottom classes have been established and applied to the entire data set. This has been done in two separate and independent procedures for the Santoña and the Tarifa survey (for details, see Anonymous, 1998). 4.3. Interpolation Due to the high resolution of the original data on SAV from the Santoña estuary, the intention was to present this data “as is”, only using the interpolation in order to cover the spaces between data points, but without introducing trends as for example with certain configurations of the “Kriging” method. Further, the use of a high power (4) restricted the influence of neighbours to a high extent. Although Guan et al. (1999) recommend “Kriging” with particular configurations as the method that renders the best results in case of height and coverage data of SAV, the spatial resolution of original data in the present study has been considerably higher than the resolution of the data set used for testing. Further, Guan et al. (1999) did not perform a test with the specific configuration (power of 4) for the “Inverse Distance to a Power” method used here, so that results from that study could not directly be applied. On the other hand, an in-depth study of the most adequate interpolation method for the data presented here was not possible within the project presented here. However, the authors believe that the applied interpolation method, as far as this can be derived from the very valuable information given by Guan et al. (1999), is adequate in 9 this case and does not generally produce misinterpretations of the data, although it might not be optimal. Due to the considerably larger distance between transects in the case of the Tarifa survey, parameters for interpolation of the SAV data were modified. In order to admit more influence of neighbouring points, a power of only 2 was used and a smoothing factor of 3 together with a search ellipse of 80m in order to reduce the effect of missing data between transects (mean distance 175m). This was done intentionally, being aware of the fact that this may introduce certain trends into the presented data that do not fully reflect the real situation. The objective in this case was to get a general idea of the areas where Cymodocea nodosa is present for protection purposes, rather than mapping the exact dimension of individual patches, so a certain exaggeration of the found distribution was considered acceptable, rather than to “miss out” existing (but not detected) patches between transects due to the relatively low resolution of the survey. Again, the differing configuration and methods were chosen due to the different resolution of the original data sets (high resolution of the data in Santoña and relatively low resolution in Tarifa). Although most of the important procedures necessary to verify the obtained data and test system performance have not been accomplished in both surveys mainly due to adverse conditions and other limitations, it remains clear that under normal conditions these requirements are not difficult to fulfil. Once these standard procedures required by the methods described here are completed successfully, both methods – the EcoSAV software for detecting SAV as well as the VBT Seabed Classifier software – can be operated in a fast and uncomplicated way, allowing a comprehensive study of entire ecosystems. The availability of several complementary data sets within short time and using the same equipment provides a clear advantage over current methods, such as aerial photography or direct sampling, currently used in ecosystem assessments. Although it might be argued that aerial photography could overcome many of the difficulties detailed here, and moreover in a quicker and possibly even cheaper way, it remains a method with important disadvantages. First, aerial photography is very dependant on optimal weather conditions, and is easily affected by waves and increased levels of turbidity, which in turn is not the case for the echosounder approach, at least not at the same levels. Second, accurate measurement of canopy height of submerged vegetation is only possible with the echosounder system presented here. Once the system is tested for performance and “calibrated”, no further field sampling is necessary or only to a small extent, as long as the same species are observed. This way, large areas can be assessed in short time, obtaining detailed information including also bathymetric information by default. The same applies for the use of the system as a bottom classification tool. Once a library for categorising bottom substrate type is established in a reliable fashion and for the 10 particular system, it can be used to determine bottom types in large areas with little effort, and especially without the need for further ground truthing. As both methods use the same echosounder system, the advantages through minimised efforts are obvious. In the current set-up, the system can be operated by one single person, if needed. If we add to these already important advantages the fact that the same system, with the same configuration can be further used to acquire data on fish and plankton abundance and biomass (using the same data as used for bottom classification, e.g.), the versatility of the system and its usefulness as a tool for the assessment of ecosystems becomes even more clear. 11 References Anonymous, Guide to Using VBT – Seabed Classifier, 1998, BioSonics, Inc. Anonymous, Surfer 7 Help, 1999, Golden Software, Inc. Guan, W., Chamberlain, R.H., Sabol, B.M., and Doering, P.H. 1999. Mapping submerged aquatic vegetation with GIS in the Caloosahatchee Estuary: Evaluation of different interpolation methods. Marine Geodesy, 22: 69-91. MacLennan, D.N., Simmonds, E.J., 1992. Fisheries acoustics, Fish and Fisheries Series, 5. Chapman & Hall, London. Sabol, B.M., and Melton jr., R.E. 1995. Development of an automated system for detection and mapping of submersed aquatic vegetation with hydroacoustic and global positioning system technologies, report I: the Submersed Aquatic Vegetation Early Warning System (SAVEWS) - system description and user's guide (version 1.0), USACE Waterways Experiment Station, Vicksburg, MS. Sabol, B., McCarthy, E., and Rocha, K. 1997. Hydroacoustic basis for detection and characterization of eelgrass (Zostera marina). Proceedings of the 4th Conference on Remote Sensing of Marine Environments, pp I-679-693. Sabol, B., Melton, R.E., Chamberlain, R., Doering, P., and Haunert, K. 2002. Evaluation of a digital echo sounder for detection of submersed aquatic vegetation. Estuaries 25(1):133-141. 12 Figures Fig. 1: Map of the Santoña survey area with track log (blue lines). In the top right corner of the image the estuary opens to the Atlantic ocean, the right arm at the bottom is the river Asón that enters the estuary. 13 Fig. 2: Survey track, location and bottom type of physical samples, port of Santoña. 14 Fig. 3: Survey track, location and bottom type of physical samples, port of Colindres. 15 Fig. 4: Survey area in the Strait of Gibraltar (Tarifa, Playa de los Lances) with bathymetry obtained from the survey data. Green: Area where SAV was surveyed. Inset: Location of survey area at the Southern tip of Spain. Fig. 5: Cymodocea nodosa. 16 Fig. 6: Coverage with Zostera sp. of the Santoña estuary. Locations of physical samples are shown with their corresponding codes. 17 Fig. 7: Results from bottom typing in the access channel of Santoña harbour. 18 Fig. 8: Results from bottom typing in the access channel of Colindres harbour. 19 Fig. 9: 3D map showing results from Tarifa: Bottom types, bathymetry (in meters) and delimitation of Cymodocea nodosa (light green) and Laminaria sp. (dark green). A ltu ra P re s e n ta c ió n 3 D a ltu ra Z o s te ra s p . M a p a d e c o n to rn o c o b e rtu ra Z o s te ra s p . (D is ta n c ia e n tre lín e a s = 1 0 % ) Fig. 10: Example for the presentation of data obtained from EcoSAV (Santoña survey). SAV height (above, in meters) and percent coverage (below), which can further be combined with a bathymetric map (not presented here). 20