Outputs

Monitoring Assessment

The assessment of Appropriateness territory was based on four phases:

1. the Data Product Specification (DPS);
2. the Upstream Data (UD) description;
3. the Targeted Data Product (TPD) description;
4. the calculation of indicators.

ISO Quality Elements was selected to describe DPS, TPD and UD, consisting of nine ranks plus one expert evaluation on the usability of the product. These Quality Elements corresponded to nine quality measures for appropriateness specifically defined within the MedSea Checkpoint.

Quality Errors was computed for each Targeted Product component from DPS and TPD quality measures and for each UD associated to the Targeted Product as well. “Errors” for Quality Elements were defined as the differences between what was realized and what had been “expected” or “required”.

The appropriateness indicators were related to “errors” in the Quality Elements just defined. Appropriateness corresponded then to “low” errors in the specific quality element.

In the case of appropriateness, it was less immediate than for availability to provide a simple characterization of the indicators at a high level of aggregation. Some simplifying assumptions were applied, allowing a non-expert to easily assess the appropriateness indicators without looking at the metadata and reports.

Appropriateness indicator values can have negative or positive values. The former score was an under-fitting score, representing lower than expected quality while the latter was an over-fitting score. Both scores were saturated at ±100%.

In order to associate a range of indicator values to a synthetic indicator score, it was necessary to establish “thresholds” for the values. It was decided that products with ‘errors’ within -10% and +10% with respect to DPS are ‘appropriate’ or at least partly adequate. Values smaller than -10% were under-fitting and not adequate while values larger than +10% were over-fitting or totally adequate, no need for further development.

Each Targeted Product was associated with its Upstream Data sets to dynamically perform the data adequacy evaluation

Analysis Results

The Checkpoint meta-database contained 266 input dataset descriptors, which identifed and assessed potentially usable information for the construction of the products. 45 Targeted Products were generated following the customer requests:

• 3 for Challenge 1 (Wind Farm Siting);
• 6 for Challenge 2 (Marine Protected Areas);
• 2 for Challenge 3 (Oil Platform Leak);
• 13 for Challenge 4 (Climate and Coastal Protection);
• 8 for Challenge 5 (Fishery Management);
• 4 for Challenge 6 (Marine Environment);
• 9 for Challenge 7(River Input).

Only 90 over the 266 input dataset were used for the realization of the 45 Targeted Products and they were referred to as Upstream Data.

The number of input data sets potentially used by the Challenges to generate their products is described in the figure below where it is evident that Wind Farm Siting, MPAs, Oil Platform Leaks and River Inputs requested more data sets than the others.

Number of input data sets for each Challenge

The availability indicators were used to carry out the first part of the monitoring system assessment. The indicators for the 266 input data sets were analysed by means of a distribution histograms of the scores. The results were summarized below, where the availability indicator scores were summed considering all the input data sets. 
Above 60% of the input data sets, contributing to the monitoring landscape, were totally and partly inadequate for the INSPIRE Catalogue. Above 40% of the input data sets were partly and totally inadequate for Policy Visibility, Delivery mechanism, Data Policy and Responsiveness.

The appropriateness indicators were computed on the Upstream Datasets selected to be used in the Challenge Targeted Products.

Availability indicators scores in percentage over the total number of input data sets (266) present in the meta-database

Only 90 of the potential 266 data sets entered the analysis determining not robust statistics. The results appear noisier than ones obtained from availability indicators suggesting that more challenges application should be implemented in order to have more significant statistics on fitness for purpose of the monitoring landscape.

Over a total of 29 characteristics used in the generation of Targeted Products, 17 appear not adequate at the scale of the Mediterranean Sea monitoring and the most frequent quality elements that score “not adequate” were: 1) horizontal coverage; 2) temporal coverage; 3) horizontal resolution, and 4) temporal validity.

The analysis of Targeted Products quality from appropriateness indicators showed that most of the products had consistent quality with respect to the DPS requirements and that the largest TDP errors were linked to inadequate horizontal coverage, horizontal resolution, and temporal validity.

In order to provide an expert evaluation of the quality of each Targeted Product, the experts assigned an overall product quality score with respect to scope (fitness for purpose) and explained why.

System Gaps

Was the Mediterranean Sea monitoring system adequate? If not, for which characteristics (essential marine variables)? In order to answer this question, the availability and the appropriateness scores had been organized by characteristics across all the challenges and products in the tables below.

Synthesis for availability indicators: characteristic categories that were inadequate in order of inadequacy

Synthesis for appropriateness indicators: characteristic categories that were inadequate in order of inadequacy

The conclusion of the Mediterranean Checkpoint assessment derived from the comparison between the inadequate characteristics for availability and appropriateness indicators. Gaps in the Mediterranean Sea monitoring system were identified by the common inadequate characteristics in the two territories.

The monitoring system gaps at the basin scale in view of the 7 prescribed Challenges were related to:

1. the sediment mass balance monitoring data. The requested Targeted Product (spatial data layers of sediment mass balance at the coast for the past 10, 50 and 100 years, Challenge 4) could not be realized because data are only available in the literature and none INSPIRE catalogue and database exists containing the data collected within the framework of EU funded projects.
2. the fishery management data, such as fish catch and by-catch. Their availability was totally inadequate (visibility, EU INSPIRE catalogue, data policy visibility, readiness, data delivery and data policy) and they failed to cover the Targeted Products requirements (tables of mass and number of landings of fish by species and year, mass and number of discards and bycatch of fish, mammals, reptiles and seabirds by species and year, Challenge 5)
 due to both horizontal and temporal coverage and temporal validity. The gap in the spatial coverage was mainly due to the absence of data from non-EU countries, while for the temporal coverage the reason was the lack of data before 2002. Data were not updated frequently and they were usually made available after 1 to 2 years from the monitoring. The major points were the scarcity of the data collected in the past 2 years of search and the lack of standardization in the data collection process.
3. the habitat extent input data sets, such as Posidonia Oceanica, Coralligenous, Maerl and seabed sensible habitats. Their availability was totally inadequate in terms of data policy and responsiveness. They failed to meet the scope of the Targeted Products in Challenge 1 (suitability of sites for wind farm development), Challenge 2 (representativeness and coherency of existing European network of marine protected areas), Challenge 3 (assessment of oil spill impact on sensitive coastal habitats) and Challenge 5 (bottom habitat disturbance by bottom trawling, change in level of disturbance over past ten years, the extent of fisheries impact on the sea floor) mainly due to their inadequate vertical and horizontal coverage, temporal and horizontal resolution.
4. the wave height, period, direction and spectral parameters input data sets. Their availability was totally inadequate because of negative scores for visibility, INSPIRE Catalogue, Data Policy, Pricing and responsiveness. They did not fully meet the scope of the Targeted Products of Challenge 1 (suitability of sites for wind farm development) and Challenge 3 (real time prediction of the likely oil leak trajectory) in terms of temporal coverage, horizontal and temporal resolution.
5. the maritime traffic (platform movement) input data sets were totally inadequate because of negative availability scores for visibility, INSPIRE Catalogue, responsiveness. A main gap depended on the difficulty of obtaining data from the competent authorities, which meant they had to be purchased from private providers. Another important gap was related to the unavailability of a few required characteristics due to confidentiality concerns related to the European laws on privacy, which made access to raw data extremely difficult in most cases. In terms of appropriateness, they failed to meet the scope of Targeted Products (bottom habitat disturbance by bottom trawling, change in level of disturbance over past ten years, the extent of fisheries impact on the sea floor, Challenge 5) in terms of horizontal and temporal coverage, temporal validity. In fact, the VMS system was implemented in the EU Mediterranean countries in different years, while the fishing fleets of most Mediterranean countries had only recently adopted the AIS system and it was still to be implemented by most North-African vessels. Limited geographical coverage was also related to the limited range of VHF signals and management of the system by the vessel crew.