Seagrass (Posidonia oceanica) monitoring in western Mediterranean ...

Environ Monit Assess (2010) 171:365–380 DOI 10.1007/s10661-009-1284-z

Seagrass (Posidonia oceanica) monitoring in western Mediterranean: implications for management and conservation Cecilia Lopez y Royo · Gérard Pergent · Christine Pergent-Martini · Gianna Casazza

Received: 23 April 2009 / Accepted: 2 December 2009 / Published online: 21 January 2010 © Springer Science+Business Media B.V. 2009

Abstract The seagrass Posidonia oceanica is extensively monitored in Mediterranean coastal waters and is an ideal candidate for an eco-regional assessment of the coastal ecosystem. The aim of this study is to evaluate the potential of P. oceanica as eco-regional indicator for its assessment at the scale of Mediterranean basin. For this purpose, regional and national P. oceanica monitoring programmes are identified, and their data and metadata are collected and compared in terms of objectives, strategies, sampling designs and sampling methods. The analysis identifies a number of issues concerning data quality, reliability and comparability. In particular, the adoption of different sampling designs and methods may introduce relevant errors when comparing data. The results of

C. Lopez y Royo · G. Casazza Inland and Marine Waters Department, APAT, Agency for Environmental Protection and TS, via Brancati 48, 00144 Rome, Italy C. Lopez y Royo (B) · G. Pergent Faculty of Sciences, University of Corsica, EqEL, 20250 Corte, France e-mail: [email protected] C. Pergent-Martini Regional Activity Centre for Specially Protected Areas, United Nations Environmental Programme, Tunis, Tunisia

this study stress the necessity of carefully planning monitoring programmes. Moreover, it highlights that the adoption of a number of common tools would facilitate all Mediterranean monitoring activities and allows an optimisation of management efforts at an eco-regional scale. Keywords Monitoring · Bioindicator · Quality assurance · Data comparability · Seagrass · Posidonia oceanica · Mediterranean

Introduction Seagrass monitoring is an issue of increasing interest in research, management and policies, due to seagrass ecological role (Costanza et al. 1997; Orth et al. 2006), their global decline (Orth et al. 2006; Short et al. 2006) and their associated indicator characteristics (Pergent et al. 1995; Short and Wyllie-Echeverria 1996). Policies have recently highlighted the use of seagrass monitoring, among other elements, through the adoption of the ecosystem based approach (CBD 2000; EC 2000; US EPA 2001; UNEP 2005; EC 2008). An ecosystem approach has a dual component in aquatic environments: (1) water quality is to be determined at ecosystem level by using the appropriate ecological indicators integrated with physical and chemical characteristics, (2) management needs to be carried out at eco-regional or river

366

basin level, across political boundaries if necessary, as this ensures a greater emphasis on ecological results (EC 2000; US EPA 2001). In Europe, following the adoption of the EU Water Framework Directive (WFD), monitoring angiosperm (therefore, seagrass in coastal waters) coherently at eco-regional scale became a legal requirement (EC 2000). Additionally, the newly adopted EU Marine Strategy Directive (EC 2008), which also requires seagrass monitoring, lays further emphasis on these aspects. Thus, in this context, it is important that seagrass monitoring be associated with a clear understanding of both the species’ indicator characteristics and its responses to disturbance at eco-regional level. In the Mediterranean, monitoring of the endemic seagrass Posidonia oceanica has been ongoing since the 1980s, when the “Réseau de Surveillance Posidonies” was set up in the French Riviera (Boudouresque et al. 2000). Since then, P. oceanica monitoring has expanded to other Mediterranean regions (Boudouresque et al. 2006; Pergent et al. 2007), covering extensive areas of the Mediterranean basin. The historical database created by these monitoring programmes, associated with the specie’s widespread distribution in the Mediterranean (Procaccini et al. 2003) and its recognised ecological indicator characteristics (Pergent et al. 1995; Pergent-Martini et al. 2005; Romero et al. 2007), makes P. oceanica an ideal candidate for a comprehensive assessment at Mediterranean eco-regional level, in line with the ecosystem based approach. Considering the geographical extent of existing monitoring programmes, the integration of the different datasets would determine a Mediterranean baseline and allow the detection of changes coherently over time and over space. Pooling of data would (1) broaden the geographical scale of the evaluation, (2) allow to develop more accurate regional reference conditions, (3) allow to understand threats and loss mechanisms at a regional scale, thereby increasing overall accuracy and efficiency of management and conservation efforts (Houston et al. 2002; Back et al. 2002; Kabuta and Laane 2003). However, data integration may be problematic when dealing with different monitoring programmes, due to the possible differences in strategies, designs, methods and data management (Kirkman

Environ Monit Assess (2010) 171:365–380

1996; Astin 2006); which may create a “border effect”, raising important data comparability or quality assurance issues. The aim of this study is to evaluate the potential of the seagrass P. oceanica as ecological indicator at eco-regional level, for an assessment of its ecological status at the scale of Mediterranean basin. The possibility of integrating existing P. oceanica monitoring programmes is investigated by (1) identifying regional and/or national P. oceanica monitoring programmes, (2) collecting and organising their data and metadata, (3) analysing and comparing these monitoring programmes, in terms of objectives, strategies, sampling designs and sampling methods and any other relevant information, (4) assessing the effects and issues which may arise when acquiring data according to different strategies and methods.

Materials and methods Study area All national or regional P. oceanica monitoring programmes identified and available in the western Mediterranean basin are considered in this study (Fig. 1). The Eastern Mediterranean basin is not included as, to our knowledge, at present in that area most P. oceanica monitoring programmes are carried out locally and in very few cases regionally (Turk and Lipej 2006; MATT 2001a). In certain areas of the Western Mediterranean, however, no regional monitoring programmes were identified such as in southern Spain or relevant parts of the Algerian coastline or were not available as in the Balearic Islands (Fig. 1). Data collection and analysis P. oceanica monitoring reports are collected for all regions in which they were made available by national or regional administrations and researchers (Table 1). Data and metadata are extracted from the reports and re-organised. Data of each monitoring programme is provisionally organised in individual relational databases (MS Access 2000). The data range from


367

Fig. 1 Sites, regions and countries for which monitoring reports were available for this study

descriptors of sites (e.g. coastal morphology, pressures present) to descriptors of P. oceanica (population, individual and physiological descriptors).

These data will eventually be merged and integrated according to the outcome of the metadata analysis.

Table 1 P. oceanica monitoring programmes that were available for this study Country Algeria France

Italy

Malta Spain Tunisia

Region Corsica Provence-Alpes-Côte d’Azur Languedoc Roussillon Provence-Alpes-Côte d’Azur & Languedoc Roussillon All coastal regions Liguria Liguria, Tuscany, Latium Calabria and Campania Sardinia Sicily All Catalonia Valencian Community

Programme

Reference

Abbreviation

Monitoring set-up Regional monitoring Regional monitoring Site monitoring WFD monitoring

Boumaza and Semroud (2000) Pergent et al. (2005) Cadiou et al. (2004) Ballesta et al. (2005) Gobert et al. (2007)

Algeria RSP Corse RSP PACA CBM FrWFD

National monitoring Regional monitoring Mapping follow up P. oceanica mapping P. oceanica mapping P. oceanica mapping Baseline Survey WFD monitoring WFD monitoring Different studies

MATT (2001a) ARPAL (2005) RIPO (2002) MATT (2004) MATT (2001b) SINPOS (2001) MEPA (2002) Romero et al. (2005) Ramos Esplà et al. (2005) Vela (2006); Sghaier (2006); Djeloulli (2004)

ItNational ARPAL RIPO CCMap SaMap SINPOS Malta Catalan Valencian Tunisia

Ref. no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

PACA Provence-Alpes-Côte d’Azur, RSP Réseau de Surveillance Posidonies, CBM Cerbères-Banyuls-sur-mer, WFD Water Framerwork Directive, ARPAL Agenzia Regionale per la Protezione dell’Ambiente Ligure, RIPO RIvisitazione di praterie di POsidonia, SINPOS Sistema INformativo per la Posidonia in Sicilia

368

Systematic data collection is chosen to organise metadata, and a standard set of questions is used to investigate the reports (Weller and Romney 1988). The questions cover all aspects of monitoring programmes: aim, specific objectives, monitoring strategy (site selection, monitoring frequency, descriptors chosen, sampling depth), sampling design (area sampled, type of area, spatial scale, statistical design) and methods used to sample and analyse descriptors (sampling protocol, type of estimation, instruments used, size of samples, replicates, reference material). “Practical and structural” questions are used to identify monitoring strategies, sampling designs and methods, “guiding” questions are used to identify specific and complementary elements (Strauss and Corbin 1998). The answers to the questions may be dichotomous or multiple choice, although free answers are also included for additional information (Weller and Romney 1988). Results were analysed using a simple matrix (Strauss and Corbin 1998). Henceforth, monitoring programmes will be cited using their abbreviation or reference number (Table 1).

Results General characteristics of the P. oceanica monitoring programmes The general aim of allP. oceanica monitoring programmes is to assess the state of the seagrass meadows and to observe trends in this state over time and/or over space (Pergent et al. 1995; Boudouresque et al. 2000; Ballesta et al. 2005; Romero et al. 2007). Monitoring programmes, however, differ in specific objectives, and the programmes under study are carried out: (1) for a survey of the area in the case of mapping or baseline surveys (33%), (2) for the protection and conservation of the meadow in which the potential degradation of the meadow is under observation (33%), (3) for an evaluation of the water quality through the assessment of ecosystem status (33%; Table 1).


Monitoring strategy Important aspects of monitoring strategies include site selection and the subdivision of the total monitoring area (Vos et al. 2000; Astin 2006). In the monitoring programmes under study, 53% select their sites according to a census strategy and 47% according to a targeted one. In the census strategy, every unit is assessed for the entire defined region, whereas in the targeted strategy, reference sites and sites subject to specific pressures are identified. Sampling depths selected by monitoring programmes cover most of P. oceanica’s habitat and range from 1 to 38 m (Table 2). Monitoring programmes have selected different sampling depths in terms of: (1) bathymetric depths (i.e. specific depth in metres) and of (2) types of depth (i.e. limits of the meadow or general depth ranges). Concerning types of depth, 46% of the programmes have selected the lower limit of the meadow, 26% selected the upper limit of the meadow, 46% a superficial depth other than the upper limit of the meadow, 33% an intermediate depth, and 7% a deep range other than the lower limit. Certain programmes have chosen a combination of depths, such as both the upper and lower limit of the meadow (13%), or an intermediate, upper and lower limit depth (13%), or both a superficial and deep range (7%). Sampling season of the different monitoring programmes are spread over all seasons (Table 2). However summer is the most common sampling season (73%), spring is shared by a number of programmes (47%), whereas autumn and winter are seldom considered.

Sampling design Sampling design, in this study, covers the general sampling procedure: type of sampling, area sampled, spatial scale, replicates and statistical design. Three types of sampling design are used in the monitoring programmes under study: (1) fixed transect design (chosen by 33% of programmes), (2) nested or hierarchical design (33% of pro-


369

Table 2 Characteristics of the different P. oceanica monitoring programmes Country

Programme

Sampling season

Sampling type of depth (and depth)

Sampling design

No. sites

Sampling frequency

Algeria France

Algeria RSP Corse

Summer Spring–summer

Transect Transect

1 30

Every 3 years Every 3 years

RSP PACA

All (not seasonally)

Transect

33

Every 3 years

CBM WFD National

Spring Spring Winter–spring

Transect Random Transect

1 17 23

Irregular – Annually

ARPAL

Spring–summer

Random

6

RIPO

Summer

CC map Sa Map

Summer Summer

Sinpos Malta Catalan Valencian All

Summer Summer Autumn Summer–autumn All (not seasonally)

Lower limit (8 m) Upper limit (5–11 m) Lower limit (24–38 m) Upper limit (2–15 m) Lower limit (14–38 m) Lower limit (19 m) Between 4 and 19 m Lower limit (17–37 m) Superficial (4–12 m) Upper limit (3–10 m) Intermediate (7–19 m) Lower limit (15–28 m) Upper limit (3–12 m) Intermediate (10–23 m) Lower limit (14–35 m) Superficial (10 ± 1 m) Superficial (5–7 m) Deep (25 ± 1 m) Superficial (1–15 m) Superficial (10 ± 1 m) Intermediate (11–18 m) Intermediate (13–22 m) Between 1 and 13 m

Italy

Malta Spain Tunisia

grammes), and (3) random sampling (33% of programmes) (Table 2). The fixed transect design relies on the use of a balisage (Boudouresque et al. 2000), which consists in placing along a portion of the meadow limit 10 to 12 ballasts (i.e. 20 kg cement blocks) at 5 m distances from each other. All fixed transects are placed on the limit of the meadows; however, not all programmes that sample on meadow limits have selected a transect design (71%). Opera-

Once

Random

12 (×5 × 3)

Every 10 years

Nested Random

60 60 (×2)

Once Once

Nested Nested Nested Nested Random

60 15 27 17 6

Once Once Annually Annually Once

tionally, the number of samples collected in this design varies between 11 and 36. The ballasts are fixed; therefore, sampling over time is carried out exactly in the same place. The nested or hierarchical design includes the selection of different areas and stations within a site; this design has been chosen by 50% of programmes that sample at homogeneous depths (superficial, intermediate or deep). Each of these monitoring programmes has selected a different

Fig. 2 Nested sampling design of P. oceanica monitoring programmes applying it

370

specific ‘design’ (Fig. 2) that range from three areas within a site (Valencian), to five stations sampled in each of the five areas within a site (SINPOS). Replicates taken in each of the stations or areas sampled vary between one and six, according to the programme and the descriptors measured. Finally, the random design is performed by sampling at random in one selected site and has been chosen by the other 50% of programmes that sample at homogeneous depths and by 29% of programmes that sample on the limits of the meadow. Both the nested and the random design for the return on site rely mainly on GPS coordinates and not on fixed marks; therefore, monitoring over time will sample approximately (GPS usually have a nominal precision of about 10 m) but not exactly the same area. Descriptors measured in the different monitoring programmes Each P. oceanica monitoring programmes under study is based on a specific selection of descriptors (Table 3). Only one descriptor is common to all programmes, i.e. ‘shoot density’; other than this, the set of descriptors selected varies considerably between programmes. To compare the use of descriptors in the different monitoring programmes, a selection of the most common ones was performed, i.e. descriptors that are present in more than 50% of the monitoring programmes under study were selected for further analysis. This resulted in the selection of 14 descriptors: lower limit depth, limit type, shoot density, cover, plagiotropic rhizomes, shoot baring, number of leaves, coefficient A, shoot foliar surface, leaf area index, leaf production, rhizome elongation, rhizome production and epiphyte biomass. An essential condition for integration and comparison of data is that descriptors correspond to the same definition in different monitoring programmes. The metadata analysis highlighted that four descriptors show differences in definition (Table 4). Cover is measured in 73% of programmes under study. In 91% of these, it is defined as the percent substrate covered by P. oceanica leaves


compared to bare substrate (dead matte, sand, rock). One programme (i.e. 9%), however, defines Cover as the percent substrate “occupied” by P. oceanica, in which occupation of substrate is determined by the shoot base and not by the leaves of the plant (Catalonia). This measure is taken in a quadrat, in which distance between shoot bases is estimated, and when below 10 cm, it is considered continuous. Shoot baring is measured by 53% of the monitoring programmes under study. In 75% of these, it is defined as the distance between the sediment and the base of leaves in orthotropic rhizomes or between the sediment and the lower part of the rhizome in plagiotropic rhizomes, as defined by Boudouresque et al. (1984). However, in 25% of the cases, shoot baring is defined as the percentage of bared rhizomes in relation to shoots that show no baring (ItNational, ARPAL). Coefficient A (coeff A) is measured by 73% of the monitoring programmes under study. All of these (100%) refer to the definition given by Giraud (1979) in which the coeff A is the percentage of leaves that have lost their apex, either broken or grazed. Due to the hybrid nature of this descriptor, a cross between grazing and hydrodynamic activity, some monitoring programmes measuring this descriptor have chosen to specify their measures: 18% defined it as the percentage of broken leaves, 9% as the percentage of grazing and 36% as the percentage of broken and grazed leaves. The remaining 27% of programmes only refer to Giraud (1979) without any further specifications. Leaf area index (LAI) is measured by 87% of the monitoring programmes under study; it is generally defined as the surface area of leaves in a defined area unit (square metres) (Giraud 1979). In 92% of the programmes, the LAI is measured by relating the shoot foliar surface (expressed in cm2 shoot−1 ) to the shoot density. However, one monitoring programme (i.e. 7%) measures the LAI by considering only the photosynthesising surface area, which is measured by considering only the green tissue of leaves (given by the descriptor B/G/W tissue SA in Table 3). Only one descriptor, epiphyte biomass, although having the same definition, is expressed in different units: 75% of monitoring programmes


371

Table 3 Descriptors measured in the different P. oceanica monitoring systems Descriptors

Algeria France Italy Malta Spain Tunisia RSP RSP CBM WFD National ARPAL RIPO Mappinga Catalonia Valencia Corse PACA

Lower limit depth Limit type Shoot density Global shoot density Cover Dead matte % plagiotropes Shoot baring Canopy height No. leaves SFS LAI B/G/W tissue SA % Necrosis Coefficient A Shoot biomass Leaf production Rhizome elongation Rhizome production Epiphytes biomass Epiphyte assemblages Sediment N and P in rhizomes C in rhizomes 15 N in rhizomes 34 S in rhizomes N in epiphytes Trace metals Genetics Balisage

X

X

X

X

X

X

X

X

/

X X

X X

X X

X X

X X

X X X

X X

X X X

/ X

X

X

X X

X

X

X

/

X X

X / X X

/ /

X X

X X

X X X

X X X

/ / /

X

X

/

X X X X X

X

/

X

X

X

X

X

X X

X

X X

X X

X X X

X X X

X X X

X

X X X X

X X X X

/ X X X X

X X X X

X X

X X

X /

X

X X X

X

X

X

/

X

X

X

X

X

X

X

X

/

X

X

X

X

X

X

X

X

/

X

X

X

X

X

X

X

X

X

/

X

/

X

X

X

X

/

X

X X X X X X X

/ X

X

X

X

X

/

“X” parameter measured, “/” parameter not always measured, “blank” parameter not measured, SFS shoot foliar surface, LAI leaf area index, B/G/W tissue SA surface area of the brown, the green and the white tissue a Mapping: the three regional mapping programmes have been grouped, as descriptors have been selected according to the same rationale (CCmap, SaMap, SINPOS)

372


Table 4 Definitions and units of measure of the most common descriptors in the different monitoring programmes Descriptor

Definition

Lower limit depth Limit type Shoot density Cover

Bathymetric depth reached by the meadow As defined by Meinesz and Laurent (1978) Number of shoots per surface area unit % substrate covered by leaves Or % substrate covered by shoot base % plagiotropes % shoots with plagiotropic growth Shoot baring Distance sediment—base of leaves/rhizome Or Percentage of bared rhizomes No. leaves Number of leaves per shoot Coeff A Indicator hydrodynamic activity and grazing Or Indicator hydrodynamic activity Or Indicator of grazing SFS Shoot foliar surface LAI Foliar surface per area unit Or Photosynthetic leaf surface per area unit Leaf production Number of leaves produced in a year Rhizome elongation Growth (length) of the rhizome in year Rhizome production Biomass production of rhizome in a year Epiphytes biomass Biomass of epiphytes present on leaves

Unit

Programmes

Metres n/a Shoots m−2 %

All that measure it All that measure it All that measure it 1, 2, 3, 4, 6, 7, 8, 9, 10, 11, 14, 15

% % cm Or % Leaves shoot−1 %

13 All that measure it 1, 2, 3, 13, 14, 15 6, 7 All that measure it 1, 2, 4, 5, 6, 7, 8, 9, 10, 13, 14, 15

%

11, 12

cm2 shoot−1 m m−2

3 All that measure it 1, 2, 3, 5, 6, 8, 9, 11, 12, 13, 14, 15

m m−2 leaves shoot−1 year−1 mm shoot−1 year−1 mg shoot−1 year−1 mg shoot−1 mg m−2

10 All that measure it All that measure it All that measure it 1, 5, 10, 11, 12, 15 13, 14

Descriptors with different definitions are underlined

relate epiphyte biomass to the shoot, whereas 25% relate it to a defined surface area unit (square metres). Differences in unit, however, can be resolved by converting values from one unit to the other by using shoot density, which is measured by all monitoring programmes. Thus, to facilitate the comparison of methods used to measure descriptors, those having different definitions are distinguished accordingly (i.e. leaf cover and shoot cover, length of shoot baring and percent shoot baring, coeff A and mechanical damage and grazing, LAI and photosynthetic LAI). Methods used to measure descriptors Descriptors may be divided into different types of levels, where population descriptors correspond to the meadow characteristics (e.g. cover, shoot density etc) and individual descriptors to the plant characteristics (e.g. leaf surface area). Additional levels can be identified, such as for instance

physiological descriptors (e.g. N and P contents); however, none of these were included in the selection of descriptors to be analysed (see Section “Descriptors measured in the different monitoring programmes”). Population level descriptors Population descriptors generally require measures to be taken, at least in part, in the field (Table 5). Lower limit depth is measured by 73% of monitoring programmes under study and is estimated either directly in the field using a depth-meter (82%) or in laboratory using the results given by the Side Scan Sonar (18%). Measures in the field are taken at the level of each ballast along the balisage (44%), at each extremity of the balisage (33%) or randomly at the level of the limit (22%). In the same way, the limit type, which is measured by 66% of monitoring programmes, is estimated either directly in the field (80%) or in the labo-


373

Table 5 Methods used to measure population descriptors in the different monitoring programmes under study Descriptor

Estimation

Collection

Sampling

Sample size

Replicates

Programme

Lower limit depth

Field (M)

Depth-meter

Balisage

Limit type

Laboratory (M) Field (E)

Side scan sonar Visual

Random Transect Balisage

Laboratory (E)

Video ROV Quadrat

Random Balisage Transect Balisage

– – – 200 m 55 m 5m – 5m 200 m 0.04 m2

Random

0.16 m2 0.16 m2

Balisage Circle Transect Balisage Balisage

0.16 m2 /4 0.25 m2 – 10 m diameter 20 m 1 m2 0.09 m2 at 3 m

Balisage Transect Transect Balisage Transect Random Balisage Balisage Balisage Random Random Balisage

Photo at 3 m – 0.25 m2 /4 – 50 m 0.16 m2 /4 0.04 m2 – 0.16 m2 – – 50 m

11–12 2 1 5 1 11–12 1 11–12 5 33–36 11–12 5 9 5 12 25 11 2 9 60 33–36 11 11 1 27 (9 × 3) 22–24 1 12 11 22–24 11 72 30 1

1, 2, 3, 4 5, 6 7, 8, 13 10, 11 1, 6 2, 3, 4, 5 7, 8 2 10, 11 1, 3 2, 4 6, 15 10, 12, 14 5, 7, 8 13 9, 11 2, 3, 4 6, 7, 8 14 1 3 2 2 9, 10, 11 13 1, 3 6, 7 13 2, 5, 15 3, 4 2 1, 15 14 6, 7

Shoot density

Cover (leaf)

Field (C)

Field (E)

Visual

Field (C)

Quadrat Grid

Laboratory (E) Cover (shoot) % Plagiotropes

Field (C) Field (E)

Shoot baring (cm)

Field (C) Field (E)

Shoot baring (%)

Field (M) Field (E)

video ROV Quadrat Visual Quadrat Quadrat Visual Quadrat Shoot Shoot Visual

Descriptors that have different definitions are distinguished accordingly to facilitate method comparison M measured, E estimated, C counted, ROV remotely operated vehicle

ratory by processing images collected in the field (30%). Estimates in the field are carried out visually on the balisage transect (37%) or between each ballast of the balisage (50%) or randomly (12%). Laboratory estimates of the limit type are made by processing images given either by video taken over the balisage transect by an operator (33%) or by remotely operated vehicles (ROV; 67%). Shoot density is measured by all monitoring programmes in the field by shoot counts in quadrats (100%). Main differences are in quadrat size and type of sampling: 60% of monitoring pro-

grammes use a 0.16 m2 quadrat, of which 6% have divided the quadrat in four equal parts: 27% use a 0.04-m2 quadrat, and 13% use a 0.25-m2 quadrat. Programmes using a 0.04-m2 quadrat always count the number of shoots at the level of ballasts on the balisage transect, whereas those that use the 0.16-m2 quadrat either take counts randomly along the balisage transect (11%) or randomly at the chosen depth (89%). Leaf cover is evaluated in the field (73%) or in laboratory (36%) using a variety of instruments (Table 5). Evaluations in the field are carried out either by visual estimation (100%) or

374

by counts (25%). Counts are carried out using a 0.09-m2 grid divided into nine squares, in which the operator counts the covered squares at 3 m from the bottom, and the overall projected surface of estimation is 4 m2 , according to the method of Gravez et al. (1995). Visual estimates in the field are made without any support instruments (87%) at the level of the balises (43%) along a transect (14%) or randomly (43%) or are made with the help of a 1-m2 quadrat at the level of the balises (13%). Laboratory estimates are made by processing images given either by video taken over the balisage transect by an operator (25%) or by ROV (75%). Concerning shoot cover no differences in method exist as it is measured only by one monitoring programme. Percent plagiotropes is measured by 53% of the monitoring programmes under study; it is only measured in the field, either by visual estimation (62%) or by counts (37%). Visual estimates are made without any support instruments along a transect (80%) or with the help of the quadrat in which shoot density is measured (20%). Counts are carried out by counting the number of plagiotropic shoots vs orthotropic shoots in the 0.04-m2 quadrat, used to evaluate shoot density. Shoot baring (centimetres) is evaluated only in the field either by visual estimation (83%) or by measuring the length of the baring on individual shoots (17%). Visual estimation of shoot baring is made without any support instruments along the balisage transect (40%) or on shoots within the quadrat used to measure shoot density (20%) or randomly on individual shoots (40%). Percent shoot baring is evaluated in the same way in both monitoring programmes measuring it by visual estimation. Individual level descriptors The most common descriptors at individual level concern mainly leaf biometry (number of leaves, shoot foliar surface, leaf area index, photosynthetic leaf area index and coefficient A) and lepidochronological measures (leaf production, rhizome production and rhizome elongation). Individual level descriptors are generally measured in the laboratory, after shoot collection in the field. Shoot collection is carried out according to


the sampling design (see paragraph 3.2.1; Table 2), and the number of replicates (individual shoots) varies between 10 (15%), 15 (15%), 20 (31%), 25 (15%) and 30 (23%). Leaf biometry descriptors are measured according to the protocol defined by Giraud (1979) by all monitoring programmes concerned. Leaves are separated respecting their distich order of insertion, types of leaves (adult, intermediate or juvenile) is recorded, as well as number, width and length of leaves and length of sheath. These measures are used to calculate shoot foliar surface, leaf area index and coefficient A. Lepidochronological descriptors are measured according to the protocol defined by Pergent (1990), by all monitoring programmes concerned. Scales (sheaths that remain attached to the rhizome after leaves have fallent) are detached following their distich insertion, ordered respecting the lepidochronological cycle and counted; rhizomes are sectioned in correspondence with the end of each cycle, measured and weighed (dry weight) (Pergent 1990). These measures are used to calculate leaf production, rhizome production and rhizome elongation and refer to years prior to sampling.

Discussion The potential of P. oceanica as ecological indicator at eco-regional level is based on the possibility of integrating data from the existing P. oceanica monitoring programmes. And a scientifically sound integration of data may be compromised by heterogeneity of data collection and interpretation. The results of this study identify consistent differences in monitoring strategies, sampling design and sampling methods. These differences may lead essentially to two types of limitations in data integration: (1) direct data comparability issues and (2) data quality issues. Implications for direct data comparability Differences that affect direct data comparability are mainly due to differences in descriptor definition, type of sampling depth and season. In the first case, descriptors with different


definitions are incompatible in terms of data integration and must be distinguished. For the type of sampling depth (i.e. lower limit, upper limit, deep, intermediate or superficial depths), given the depth-dependence of most P. oceanica descriptors (Pergent-Martini et al. 1994; Zupo et al. 1997), differences affect direct data comparability of most descriptors. Indeed, monitoring the limit of the meadow allows only replicability of measures over time at the expense of replicability over space, due to the variability of the bathymetric depth of the limit, whereas sampling at homogeneous depth ranges will also favour a replicability of measures over space. Finally, the selection of the sampling season will also affect comparability of season-dependent descriptors. Thus, differences in monitoring programmes may affect direct data comparability; however, if these are acknowledged—and data is considered in relation to depth, season or definition when relevant—they do not compromise the validity of data integration.

Implications for data quality and reliability Sampling design based on the appropriate spatial scale is essential for the effectiveness of monitoring programmes (Vos et al. 2000; Burdick and Kendrick 2001). Natural sources of variability, such as hydrodynamic activity, substrate, light, temperature and patchiness, highly affect P. oceanica meadow structure and morphology on all spatial scales (Panayotidis et al. 1981; Alcoverro et al. 1995; Marbà and Duarte 1997; Greve and Binzer 2004). P. oceanica natural variability is particularly high at shallow depths (Alcoverro et al. 1995; Marbà and Duarte 1997; Balestri et al. 2003), whereas its variability is more clearly linked to environmental quality in deep meadows (Alcoverro et al. 1995; Marbà and Duarte 1997). Therefore, different depth selection affects data quality further than direct comparability of depthdependent parameters. If the aim is to detect changes in the P. oceanica ecosystem that indicate disturbance or environmental degradation, the choice of superficial or upper limit depths may introduce a type-II error (i.e. accepting the null hypothesis when it is false) by detecting changes

375

that are linked to natural variability rather than to a response to disturbance. The selection of different spatial scales for sampling may also affect data quality and reliability. Incorrect generalisations can be made if smallscale patchiness is not adequately taken into consideration when assessing large spatial scales such as seagrass meadows (Benedetti-Cecchi 2001; Burdick and Kendrick 2001; Balestri et al. 2003). Balestri et al. (2003) showed that significantly different evaluations of P. oceanica descriptors can be obtained according to the spatial scale used (10s of metres, 100s of metres, 10s of kilometres); moreover, they showed that by simulating different sampling designs with the same set of data, opposite results could be obtained. The selection of inappropriate spatial scales may, therefore, lead to both type-I and type-II errors (Benedetti-Cecchi 2001; Balestri et al. 2003). The nested design has been suggested as the most appropriate to provide a representative estimate of P. oceanica status (Benedetti-Cecchi 2001; Burdick and Kendrick 2001; Balestri et al. 2003). The adoption of a design different from the nested one (66% of monitoring programmes), by lowering data reliability, may affect data integration both within and between programmes. Sampling methods used to measure the most common descriptors differ between monitoring programmes, particularly for population descriptors. In order to maximise reliability of the results, laboratory measures are preferred to field measures (Duarte and Kirkman 2001); moreover, methods should include as few subjective elements as possible, preferring counts to estimates (Vos et al. 2000; Duarte and Kirkman 2001). However, on average in more than 60% of monitoring programmes, population descriptors are measured by visual estimates in the field. Field visual estimates in P. oceanica are highly affected by subjectivity of and variability between operators and in most cases are not standardised methods (Pergent-Martini et al. 2005). The adoption of a method with low reliability may result in considerable data variability even within a same meadow. Furthermore, the adoption of different sampling methods for a same descriptor, associated to the different levels of data reliability provided by these methods, may create additional data

376

integration issues. For instance, in the estimation of leaf cover, the method of Gravez et al. (1995) based on counts (see Section “Population level descriptors”) relies on a large projected surface (4 m2 ) and determines meadow structure (Leriche et al. 2006). Whereas in the method based on the use of a standard quadrat, in which estimates are made within quadrats layed directly on the meadow, the microscale (metres) patchiness of the meadow will be taken into consideration and will be thus a more representative population estimate (Panayotidis et al. 1981). Moreover, leaf cover estimates using photographed quadrats have been suggested as the most reliable method (Duarte and Kirkman 2001; McDonald et al. 2006), as they provide higher standardisation and lower bias; however, none of the monitoring programmes under study have adopted it. Concerning shoot density, the minimum size of a quadrat to obtain acceptable levels of standard error is 0.16 m2 (Panayotidis et al. 1981) with at least 10 replicates (Panayotidis et al. 1981; Pergent et al. 1995). In the monitoring programmes under study, 27% use a quadrat smaller than 0.16 m2 , and another 53% use less than 10 replicates. Therefore, 80% of monitoring programmes potentially introduce important errors in the estimation of shoot density. Thus, the adoption of a different sampling designs, spatial scales and sampling methods can (1) increase data variability and (2) affect data reliability and interpretation, within and between monitoring programmes. It has not been possible at this stage to assess the magnitude of the errors introduced. However, it is clear that it involves quality assurance issues, which may limit the potential of P. oceanica as ecological indicator at eco-regional level. Implications for data analysis at eco-regional scale For a scientifically sound integration of data at eco-regional scale, both (1) direct comparability and (2) data quality and reliability issues need to be addressed. Data comparability issues are overcome by relating data to its sampling characteristics (depth, season, definition). Data quality and reliability issues would require a quality


assurance procedure. Data quality assurance may be addressed through intercalibration procedures or international quality standards (e.g. CEN and ISO), as these ensure the provision of data of an equivalent scientific quality and comparability. However, no intercalibration procedures or international quality standards are available at present for P. oceanica sampling methods (CIS-WFD 2003). At a broader scale, the verification of the comparability of the overall output of monitoring programmes could support data quality assurance. The assessment of comparability of the overall monitoring evaluation obtained by different monitoring programmes is strongly supported by policy (EC 2000) and management (Houston et al. 2002). However, of the monitoring programmes under study, only one is associated to an overall evaluation index of P. oceanica status (Romero et al. 2007). It is, therefore, not possible to adopt any of the suggested intercalibration procedures. Thus, at this stage, it is not possible to assess data quality any further. Several authors, however, have determined that integration of data derived from different monitoring and sampling designs allows the emergence of information that cannot be obtained from independent efforts (Astin 2006) and can lead to a more comprehensive management effort (Houston et al. 2002; Astin 2006). Data from monitoring programmes with different strategies and designs, although potentially subject to data comparability issues, are, therefore, considered representative of the range of conditions that can be found in the western Mediterranean. Implications for monitoring efficiency The results of this study have also highlighted a number of issues which may affect monitoring and management efficiency. Clearly defined monitoring objectives are essential in order to design effective and efficient monitoring programmes (Vos et al. 2000). P. oceanica monitoring programmes under study may be related to three broad management objectives, i.e. conservation objectives, water quality objectives and mapping objectives. However, these programmes would greatly benefit from additional specifications which address goals to be


reached, amounts and type of change to be detected, desired level of significance and detection power. The choice of descriptors has to be strongly related to the objectives and to the desired outputs, for an optimisation of efficiency and effectiveness of the monitoring (Vos et al. 2000; de Jonge et al. 2006). Monitoring programmes under study have generally selected descriptors that are agreed indicators of P. oceanica status (Pergent et al. 1995; Ruiz and Romero 2003; Pergent-Martini et al. 2005; Leoni et al. 2006). However, only one, the Catalan monitoring programme, has selected descriptors according to an index for the evaluation of ecological status (Romero et al. 2005, 2007). The adoption of an index allows (1) to efficiently select descriptors accordingly and (2) to report and present monitoring results effectively.

Conclusions Data collected through monitoring programmes should enable optimal management practices (Vos et al. 2000; Back et al. 2002). However, the differences found in P. oceanica monitoring programmes in the western Mediterranean seem to limit a basin-wide eco-regional approach. The results of this study highlight the necessity, when planning a P. oceanica monitoring programme, to carefully consider: (1) the definition of clear and precise objectives, (2) the definition of a monitoring strategy according to objectives, (3) the selection of sampling design on the basis of its reliability, representativeness and efficiency, (4) the selection of descriptors on the basis of their sensitivity to disturbance and to their relevance with objectives and outputs, (5) the selection of sampling methods on the basis of their reliability, efficiency and cost-effectiveness. Quality assurance and data integration issues pose a limit to an eco-regional assessment, emphasising the importance of working towards a “P. oceanica Mediterranean monitoring network”. This would improve the ecosystem approach, as well as respond to the often stressed necessity of global seagrass monitoring network (Kirkman 1996; Duarte 1999, 2002) A common tool box, which includes a number of agreed monitoring

377

criteria, would create the conditions for an optimisation of P. oceanica monitoring and conservation at a Mediterranean scale, in terms of reliability, effectiveness and cost. It would moreover facilitate managers in setting up reliable and effective monitoring programmes and would thereby support decision-takers efficiently. Furthermore, an overall Mediterranean assessment would provide the necessary tools in order to acquire an integrated view, in which causes of deterioration may be related to ecosystem-wide effects. Acknowledgements We would like to thank all the national or regional administrations and researchers that have made their monitoring reports available for this study, in particular, the GIS Posidonie, the Italian Ministry of Environment, the Maltese Environmental Planning Authority, the Catalan Water Agency, the University of Barcelona and the University of Alicante.

References Alcoverro, T., Duarte, C. M., & Romero, J. (1995). Annual growth dynamics of Posidonia oceanica: Contribution of large-scale versus local factors to seasonality. Marine Ecology Progress Series, 120, 203–210. ARPAL (2005). Completamento della rete di controllo sulle praterie a Posidonia oceanica. Genova: Agenzia Regionale per la Protezione dell’Ambiente Ligure. Astin, L. E. (2006). Data synthesis and bioindicator development for nontidal streams in the interstate Potomac River basin, USA. Ecological Indicators, 6, 664–685. Back, S., Ekebom, J., & Kangas, P. (2002). A proposal for a long-term baseline phytobenthos monitoring programme for the Finnish Baltic coastal waters: Monitoring submerged rocky shore vegetation. Environmental Monitoring and Assessment, 79, 13–27. Balestri, E., Cinelli, F., & Lardicci, C. (2003). Spatial variation in Posidonia oceanica structural, morphological and dynamic features in a northwestern Mediterranean coastal area: A multi-scale analysis. Marine Ecology Progress Series, 250, 51–60. Ballesta, L., Pergent, G., Pasqualini, V., & Pergent-Martini, C. (2005). Distribution and dymanics of Posidonia oceanica beds along the Albères coastline. Comptes Rendus de l’Académie des Sciences, Serie III—Sciences de la Vie, 323, 407–414. Benedetti-Cecchi, L. (2001). Variability in abundance of algae and invertebrates at different spatial scales on rocky sea shores. Marine Ecology Progress Series, 215, 79–92. Boudouresque, C. F., Charbonnel, E., Meinesz, A., Pergent, G., Pergent-Martini, C., Cadiou, G., et al. (2000). A monitoring network based on the seagrass Posidonia oceanica in the northwestern

378 Mediterranean Sea. Biologia Marina Mediterranea, 7, 328–331. Boudouresque, C. F., Charbonnel, E., Coppo, S., Le Direach, L., & Ruitton, S. (2006). Les méthodes de surveillances des herbiers à Posidonia oceanica. In C. F. Boudouresque, G. Bernard, P. Bonhomme, E. Charbonnel, G. Diviacco, A. Meinesz, et al. (Eds.), Préservation et conservation des herbiers à Posidonia oceanica (pp. 141–155). France: Ramoge. Boudouresque, C. F., Jeudy de Grissac, A., & Meinesz, A. (1984). Relation entre la sédimentation et l’allongement des rhizomes orthotropes de Posidonia oceanica dans la baie d’Elbu (Corse). In C. F. Boudouresque, A. Jeudy de Grissac, & J. Olivier (Eds.), International workshop on Posidonia oceanica Beds (pp. 185–191). Paris: GIS. Boumaza, S., & Semroud, R. (2000). Surveillance de l’herbier à Posidonia oceanica d’El Djamila (Algérie). In Proceedings of the f irst Mediterranean symposium on marine vegetation (pp. 110–114). Tunis: RAC/SPA. Burdick, D. M., & Kendrick, G. A. (2001). Standards for seagrass collection, identification and sample design. In F. T. Short & R. G. Coles (Eds.), Global seagrass research methods (pp. 79–100). Amsterdam: Elsevier. Cadiou, G., Charbonnel, E., Boudouresque, C. F., Meinesz, A., Bonhomme, P., Bernard, G., et al. (2004). Le réseau de Surveillance Posidonies de la région Provence-Alpes-Cote d’Azur:résultats du suivi 2003. Région PACA/GIS Posidonie/SM13/DDE83/Conseil Général 06. Marseille, France: GIS Posidonie publ. CBD (2000). The ecosystem approach. Decision V/6, fifth conference of the parties (COP) to the convention on biodiversity, Nairobi, Kenya, May 2000. Available from: http://www.cbd.int/decisions. CIS-WFD (2003). Common implementation strategy for the water framework directive (2000/60/EC). Guidance Document no. 7: Monitoring under the water framerwork directive. Working Group 2.7 – http://forum.europa.eu.int/Public/irc/env/wfd/library. Costanza, R., d’Arge, R., de Groot, R., Farber, S., Grasso, M., Hannon, B., et al. (1997). The value of the world’s ecosystem services and natural capital. Nature, 387, 253–260. de Jonge, V., Elliott, M., & Brauer, V. (2006). Marine monitoring: Its shortcomings and mismatch with the EU water framework directive’s objectives. Marine Pollution Bulletin, 53, 5–19. Djellouli, Z. (2004). Ef fets des herbiers à Posidonia oceanica sur la dynamique marine et la sédimentologie littorale dans la baie de Monastir (Tunisie orientale). Thèse de Doctorat Université de Corse, France. Duarte, C. (1999). Seagrass ecology at the turn of the millennium: Challenges for the new century. Aquatic Botany, 65, 7–20. Duarte, C. (2002). The future of seagrass meadows. Environmental Conservation, 29, 192–206. Duarte, C., & Kirkman, H. (2001). Methods for the measurement of seagrass abundance and depth distribution. In F. T. Short & R. G. Coles (Eds.), Global seagrass research methods (pp. 141–153). Amsterdam: Elsevier.

Environ Monit Assess (2010) 171:365–380 EC (2000). DIRECTIVE 2000/60/EC of the European parliament and of the council, of 23 October 2000, establishing a framework for Community action in the field of water policy. Of f icial Journal of the European Commission, G.U.C.E. 22/12/2000, L 327. EC (2008). DIRECTIVE 2008/56/EC of the European parliament and of the council, of 17June 2008, establishing a framework for Community action in the field of marine environmental policy (Marine Strategy Framework Directive). Of f icial Journal of the European Commission, G.U.C.E. 25/6/2008, L 164/19. Giraud, G. (1979). Sur une méthode de mesure et de comptage des structures foliaires de Posidonia oceanica (Linnaeus) Delile. Bulletin du Museum d’Histoire Naturelle de Marseille, 39, 33–39. Gobert, S., Andral, B., Chery, P., Lejeune, P., RicoRaimondino, V., Sartoretto, S., et al. (2007). National f inal report on the intercalibration of coastal waters in France as part of the requirements of the water framework directive (2000/60/EEC). Gravez, V., Gelin, A., Charbonnel, E., Francour, P., Abellard, O., & Remonnay, L. (1995). Surveillance de l’herbier de Posidonie de la baie du Prado (Marseille) - Suivi 1995. Marseille, Fr.: Ville de Marseille & GIS Posidonie. Greve, T. M., & Binzer, T. (2004). Which factors regulate seagrass growth and distribution? In J. Borum, C. M. Duarte, D. Krause-Jensen, & T. M. Greve (Eds.), European seagrasses – an introduction to monitoring and management (pp. 19–23). EU project EVK3-CT2000–00044. Houston, L., Barbour, M. T., Lenat, D., & Penrose, D. (2002). A multi-agency comparison of aquatic macroinvertebrate-based stream bioassessment methodologies. Ecological Indicators, 1, 279–292. Kabuta, S. H., & Laane, R. W. (2003). Ecological performance indicators in the North Sea: Development and application. Ocean and Coastal Management, 46, 277– 297. Kirkman, H. (1996). Baseline and monitoring methods for Seagrass meadows. Journal of Environmental Management, 47, 191–201. Leoni, V., Pasqualini, V., Pergent-Martini, C., Vela, A., & Pergent, G. (2006). Morphological responses of Posidonia oceanica to experimental nutrient enrichment of the canopy water. Journal of Experimental Marine Biology and Ecology, 339, 1–14. Leriche, A., Boudouresque, C. F., Gravez, V., & Mayot, N. (2006). Does coverage matte rat mesoscale within a Posidonia oceanica meadow? Comptes Rendus Biologie, 329, 733–741. Marbà, N., & Duarte, C. M. (1997). Interrannual changes in seagrass (Posidonia oceanica) growth and environmental change in the Spanish Mediterranean littoral zone. Limnology and Oceanography, 42, 800–810. MATT (2001a). Metodologie Analitiche di Riferimento: Programma di monitoraggio per il controllo dell’ambiente marino-costiero (triennio 2001–2003). In A. M. Cicero & I. Di Girolamo (Eds.), Ministero dell’ Ambiente e della Tutela del Territorio. Rome: ICRAM.

Environ Monit Assess (2010) 171:365–380 MATT (2001b). Mappatura delle praterie di Posidonia oceanica lungo le coste della Sardegna e delle piccole isole circostanti—Relazione f inale. Rome: Ministero dell’Ambiente, Ispettorato Centrale per la Difesa del mare. MATT (2004). Mappatura delle praterie di Posidonia oceanica e di altre fanerogame marine lungo le coste della Campania e della Calabria e delle isole minori circostanti—Relazione f inale. Rome: Ministero dell’Ambiente e della Tutela del Territorio, Direzione per la protezione della Natura. McDonald, J. I., Coupland, G. T., & Kendrick, G. A. (2006). Underwater video as a monitoring tool to detect change in seagrass cover. Journal of Environmental Management, 80, 148–155. Meinesz, A., & Laurent, R. (1978). Cartographie et etat de la limite inférieure de l’herbier de Posidonia oceanica dans les Alpes-maritimes (France). Botanica Marina, 21, 513–526. MEPA (2002). Baseline survey of the extent and character of Posidonia oceanica (L.) Delile Meadows in the territorial waters of the Maltese Islands. Malta: Maltese Environmental Planning Authority. Orth, R., Carruthers, T., Dennison, W., Duarte, C., Fourqurean, J., Heck, K., et al. (2006). A global crisis for seagrass ecosystems. BioScience, 56, 987–996. Panayotidis, P., Boudouresque, C. F., & MarcotCoqueugniot, F. (1981). Microstructure de l’herbier de Posidonia oceanica (Linnaeus) Delile. Botanica Marina, 24, 115–124. Pergent, G. (1990). Lepidochronological analysis of the seagrass Posidonia oceanica (L.) Delile: A standardised approach. Aquatic Botany, 57, 39–54. Pergent, G., Clabaut, P., Mimault, B., Pasqualini, V., & Pergent-Martini, C. (2005). Mise en œuvre d’un réseau de surveillance Posidonies le long du littoral de la Corse. Deuxième phase: Porto-Vecchio/Ajaccio. Corte: Contract Office de l’Environnement de la Corse et GIS Posidonie Centre de Corse, GIS. Pergent, G., Pergent-Martini, C., & Boudouresque, C. F. (1995). Utilisation de l’herbier à Posidonia oceanica comme indicateur biologique de la qualité du milieu littoral en Méditerranée: Etat des connaissances. Mesogée, 54, 3–29. Pergent, G., Pergent-Martini, C., Casalta, B., Lopez y Royo, C., Mimault, B., Salivas-Decaux, M., et al. (2007). Comparison of three seagrass monitoring systems: Seagrassnet, Posidonia programme and RSP. Proceedings of the 3rd Mediterranean Symposium on Marine vegetation (pp. 141–150). Tunis: RAC/SPA. Pergent-Martini, C., Leoni, V., Pasqualini, V., Ardizzone, G. D., Balestri, E., Bedini, R., et al. (2005). Descriptors of Posidonia oceanica meadows: Use and application. Ecological Indicators, 5, 213–230. Pergent-Martini, C., Rico-Raimondino, V., & Pergent, G. (1994). Primary production of Posidonia oceanica in the Mediterranean Basin. Marine Biology, 120, 9–15. Procaccini, G., Buia, M. C., Gambi, M. C., Perez, M., Pergent, G., Pergent-Martini, C., et al (2003). Seagrass status and extent along the Mediterranean coasts of Italy, France and Spain. In E. P. Green, F. T. Short, &

379 M. D. Spalding (Eds.), World Atlas of seagrass: Present status and future conservation (pp 56–66). Berkeley: University of California Press. Ramos Esplà, A. A., Diaz-Valdez, M., Fernández Torquemada, Y., Giménez, F., Abellán, E., Izquierdo, A., et al (2005). Estudio de la Franja litoral de la Comunidad Valenciana dentro de la Directiva Marco del Agua: Macrof itos e invertebrados bentonicos (Substrato rocoso). Alicante: Universidad de Alicante. RIPO (2002). Rivisitazione di alcune praterie di Posidonia oceanica lungo le coste delle regioni: Liguria – Toscana – Lazio – Basilicata – Puglia. MATT, Ministero dell’Ambiente e della Tutela del Territorio – CoNISMA, Consorzio Nazionale Interuniversitario per le Scienze del Mare. Romero, J., Alcoverro, T., Martinez-Crego, B., Escartin, S., & Perez, M. (2005). Implementaciò d’una xarxa de vigilància de la qualitat de las aigues litorals a Catalunya, basada en indicadors biològics d’herbeis de Posidonia oceanica. Barcelona: Agència Catalana de l’Aigua– Universitat de Barcelona. Romero, J., Martinez-Crego, B., Alcoverro, T., & Perez, M. (2007). A multivariate index based on the seagrass Posidonia oceanica (POMI) to assess ecological status of coastal water under the Water Framework Directive (WFD). Marine Pollution Bulletin, 55, 196–204. Ruiz, J. M., & Romero, J. (2003). Effects of disturbances caused by coastal constructions on spatial structure, growth dynamics and photosynthesis of the seagrass Posidonia oceanica. Marine Pollution Bulletin, 46, 1523–1533. Sghaier, Y. (2006). Etude comparative de deux herbiers de Posidonia oceanica des côtes orientales de la Tunisie: Phénologie, lépidochronologie et Amphipodes associés. In Mémoire de Mastère en Aquaculture et Biotechnologie Marine (pp. 1–140). Université de Tunis. Short, F. T., Koch, E. W., Creed, J. C., Magalhaes, K. M., Fernandez, E., & Gaeckle, J. L. (2006). SeagrassNet monitoring across the Americas: Case studies of seagrass decline. Marine Ecology, 27, 277–289. Short, F. T., & Wyllie-Echeverria, S. (1996). Natural and human-induced disturbances of seagrasses. Environmental Conservation, 23, 17–27. SINPOS (2001). Mappatura delle praterie di Posidonia oceanica lungo le coste della Sicilia e delle isole minori circostanti – Relazione f inale. Contratto, reperterio no. 85 del 27 Dicembre 1999, del Ministero dell’Ambiente, Ispettorato Centrale per la Difesa del mare con il raggrupamento temporaneo d’Imprese (CEOM, GAS, GEOLAB, THETIS, Universita’di Pisa e Universita’ di Palermo). Strauss, A., & Corbin, J. (1998). Basics of qualitative research: Grounded theory procedures and techniques (2nd ed.). Newbury Park: SAGE. Turk, R., & Lipej, L. (2006). Research on seagrasses off the Slovenian coast (Northern Adriatic)—state of the art. Biologia Marina Mediterranea, 13, 282–286. UNEP (2005). Application of the ecosystem approach to the management of human activities in the marine environment of the Mediterranean sea. Definition, implications and tentative road

380 map. UNEP(DEC)/MED/WG.264/Inf.3, 5 May 2005. US EPA (2001). The watershed approach. Watersheds. US Environmental Protection Agency, Office of Wetlands, Oceans, and Watersheds. http://www. epa.gov/owow/watershed/. Vela, A. (2006). Fonctionnement et production primaire de l’herbier à Posidonia oceanica (L) Delile en Méditerranée occidentale. Thèse de Doctorat Université de Corse, France.

Environ Monit Assess (2010) 171:365–380 Vos, P., Meelis, E., & Ter Kreus, W. J. (2000). A framework for the design of ecological monitoring programs as a tool for environmental and nature management. Environmental Monitoring and Assessment, 61, 317– 344. Weller, S. C., & Romney, A. K. (1988). Systematic data collection (p. 132). Newbury Park: SAGE. Zupo, V., Buia, M. C., & Mazzella, L. (1997). A production model for Posidonia oceanica based on temperature. Estuarine and Coastal Shelf Science, 1, 483–492.