Knowledge gaps and challenges for conservation of Mediterranean wetlands: Evidence from a comprehensive inventory and literature analysis for Sardinia

1. Wetlands are some of the most important ecosystems on Earth. They play a key role in mitigating climate change-related events and filtering polluted water, and provide habitats for a wide range of species. Despite their importance, and numerous regulations that support their conservation, wetlands continue to be destroyed. Recent reports have indicated a progressive decline in ecological character of remaining wetlands. 2. This research reviews and extends the present status of wetland knowledge in Sardinia, the second largest Mediterranean island. For the first time, Sardinian wetlands were comprehensively mapped using satellite images and field validation. Impacts were also assessed. Trends in literature published about Sardinian wetlands since 1900 were then analysed, mainly according to the location(s) and topic(s) studied. 3. total, 2,501 Sardinian wetland sites were identified. The most common impacts observed in the field were vegetative degradation and water pollution. these wetlands, 2,274 have never subject of a research paper. Despite recent increases in publication rates, there was a lack of even basic knowledge about many wetlands, especially smaller ones. Larger wetlands have been studied from a range of viewpoints. effective and of in basic information, and to improvements in research and conservation, which might include multidisciplinary approaches in support of more comprehensive conservation management plans.


| INTRODUCTION
The Mediterranean Basin is one of the richest biodiversity hotspots, and the subject of several biodiversity assessments, management studies, and conservation actions (Cuttelod et al., 2009). Among Mediterranean ecosystems, wetlands are critical for biodiversity.
Although they occupy only 2-3% of the terrestrial surface, they support more than 30% of vertebrate species, as well as impressive numbers of microbes, plants, insects, and other species. Such high species richness is the result of the transitional nature of wetlands, meaning that they can support both terrestrial and aquatic species, from adjacent environments or further afield, at least transiently.
Mediterranean wetlands are also important for people. They offer nature-based solutions to problems such as climate change, provide ecosystem services that contribute to human well-being, and support many economic activities (Geijzendorffer et al., 2019a).  (Geijzendorffer et al., 2019a).
Mediterranean wetlands are threatened by anthropogenic factors such as construction of infrastructure and urban areas, conversion to agricultural land, irrigation, pollution, eutrophication, sediment loading, and hydrological alterations. Wetlands may be directly altered by these activities, or indirectly affected by activities in their catchments. In addition, Mediterranean wetlands are being affected, and will continue to be affected, by large-scale environmental changes such as warming temperatures, increased aridity and soil salinity, sea level rise and biological invasions (Pascual-Aguilar et al., 2015;Bolpagni, 2021).
Inventories, comprehensively describing the location, size, and other characteristics of wetlands in a defined area, are crucial for informed and wise decisions regarding wetland conservation (Bolpagni, 2020). However, many areas lack detailed inventories or directories and instead rely on information about wetlands generated as a by-product of other activities, such as bird surveys or land use cover appraisals (Finlayson et al., 1999). Existing inventories, both coarse and high resolutions and global or regional scales, are often based on satellite imagery and lacked consistent field validation. This has led to the neglect of several aspects, such as characterization of biological diversity and disturbances, and the detection of specific habitat types.
In the Mediterranean Basin, since the establishment of the Mediterranean Wetland Initiative (MedWet, https://medwet.org/) in 1991, a standardized method has been defined to provide knowledge about wetlands to policy makers, non-governmental organizations, and wetland site managers (Costa et al., 1996;Geijzendorffer et al., 2019b). Despite the compilation of much information on wetlands at global, Mediterranean, and national scales (Perennou et al., 2012;Geijzendorffer et al., 2019a), further improvements are needed. For instance, the interconnection of databases is irregular and, to include functional diversity, more data are necessary to bridge the remaining gaps and methodological variability (Cornwell et al., 2019). Moreover, national programmes to monitor trends and to improve the network of protected wetlands by including certain under-represented wetland types, such as temporary and small wetlands, are required in all countries, even where national inventories exist (Finlayson et al., 1999;Perennou et al., 2012).
This research is focused on wetlands of the second largest Mediterranean island (Sardinia, Italy). It is home to around 1.6 million people (approximately 68 inhabitants km À2 ) and attracts approximately three million tourists annually (Fois, Fenu & Bacchetta, 2019). Sardinia is of special biological interest because of its central geographical position in the western Mediterranean Sea. It hosts a great variety and number of wetland types and species related to both the African and European continents (Vogiatzakis, Mannion & Sarris, 2016).
Accordingly, eight Ramsar sites and numerous protected areas have been specifically designated for the protection of wetlands in Sardinia.
Despite this, efforts are needed to address knowledge, policy, and conservation gaps (Bagella, Caria & Filigheddu, 2013   sardegnageoportale.it). The presence or absence of impacts was also recorded for each wetland visited, following the MedWet scheme (Tomàs-Vives, 2008). The impacts were classified as follows: loss of aesthetic value(s), faunal changes, habitat degradation, decrease in wetland benefits, pollution, soil/land impacts, vegetative degradation, hydrological impacts.

| Literature analysis
Peer-reviewed research related to Sardinian wetlands published between January 1900 and December 2019 was reviewed (searched January 2020 in Web of Science, Scopus, and Google Scholar). The review of papers involved searches in English and Italian. It must be acknowledged that the search may have missed papers at scales larger than Sardinia (e.g. Italy, Europe, entire Mediterranean, or global) with relevant information on Sardinian wetlands.
From 8,530 initial results, publications that were not journal articles (e.g. books, book chapters, conference abstracts) and duplicates were removed, leaving 1,340 results. Subsequent screening of titles, abstracts, and full texts yielded a final list of 344 papers that carried out empirical research on one or more Sardinian wetlands (see Table S1).
All wetland study sites from each paper were extracted and georeferenced in QGIS and overlaid on wetland inventory and protection status layers. Log-linear models (Agresti, 2003) were fitted in R statistical software (version 3.6.1) (R Core Team, 2019) to test whether the following wetland attributes determined the number of papers per wetland: wetland type (16 categories), natural (yes/no), coastal (yes/no), protection (yes/no). In addition, all 344 relevant papers were categorized using 11 topic areas that are listed in Table 1. The maximum number of topics in a single paper was three.
Lastly, a network analysis was conducted, using the software Gephi 0.9.2 (Bastian, Heymann & Jacomy, 2009), to analyse links between the research topics in each paper.

| Wetland inventory
In total, 2,501 wetlands were found in Sardinia, covering a surface area of 494.2 km 2 (ca. 2% of the Sardinian territory). The full dataset is available online, in the framework of the Mediterranean Island Wetland project (https://italiaiswet.it).
The wetlands are highly variable in size, ranging from a minimum of 0.1 ha to the 6,566 ha of the Santa Gilla Lagoon. Artificial wetlands are more numerous (1,795) than natural wetlands (706; Table 2).
However, natural wetlands, especially those in coastal areas, are generally larger than the artificial ones, with a total surface area of 365.6 km 2 (compared with 128.6 km 2 for artificial wetlands). Natural wetlands are spread across the whole island, with approximately equal coastal (347) and inland (359) numbers. However, most (93.5%) of the wetland area is coastal. Seasonal freshwater ponds are the most common type of natural wetlands (193; 27.3% of natural wetlands), T A B L E 1 Description of the research topics to which papers were allocated, and some examples of papers included

Research topic Description Examples
History & Archaeology Exploitation of wetland resources in ancient times, archaeological excavations, and analysis of human interaction with local landscapes Marine and inland water resources in history (Carenti, 2013); recent industrialization (King, 1977) Hydrology Properties, distribution, and circulation of water  (Lugliè et al., 2001); analysis of miningderived pollutants (Cidu et al., 1997) Diseases Analysing risks related to human health; food-borne diseases and infections; organisms such as bacteria, parasites, viruses; biotoxins Malaria (Marchi & Munstermann, 1987); infection and viruses from wild birds (Kelvin et al., 2012) Sustainable economy & Tourism Natural resource management with a perspective of economic exploitation, but also considering maintenance of the functionality of natural systems. New economic models focusing on combined economic-environmental concerns in accordance with sustainable development policies Green infrastructure and ecological corridors (Cannas et al., 2018); development of low impact aquaculture (Chessa et al., 2005); sustainable water policies (Corsale, 2011); environment and tourism in fragile territories (Lai & Sistu, 2015) Dissemination policies Communication, education and awarenessraising activities, including how to enhance exchange of knowledge between the local population and stakeholders Overview of Italian Long-Term Ecosystem Research network (LTER-Italy) (Pugnetti et al., 2013); empowering local communities through collective grassroots actions (Ruiu & Ragnedda, 2017) Economic production Utilization of wetland resources for economic activities such as aquaculture, fishing and salt production; analysing the monetary value of wetland resources Floating upwelling system for breeding of Venerupis decussata juveniles (Chessa et al., 2013), eel fishing management (Rossi & Cannas, 1984) although a small number (22)

| Spatial and temporal trends of publications
Of the 2,501 identified wetlands, 227 (9.1%) are referenced in at least one of the 344 reviewed papers (Table 2). Studies have generally been concentrated in larger sites: the proportion of wetland area studied (89.7%) is far higher than the proportion of individual sites that have been studied (9.1%). This is also true for all wetland types other than temporary brackish ponds and marshes. Based on counts of papers alone (not their content), coastal natural wetlands were the most completely investigated, followed by inland natural wetlands in terms of number and artificial ones in terms of area. Among the 16 types, permanent wetlands were the most investigated (33% by number; 74% by area). This included coastal lagoons, estuaries and wetland systems and the inland permanent ponds and lakes. Among artificial wetlands, dam lakes/river barriers were most investigated (31% by number; 93% by area), with little published about other types.
The log-linear model results showed significant effects of wetland type and protection status on study effort (proxied by number of papers; Figure 1). For wetland type, temporary pools were significantly (P < 0.05) under-represented in the literature. There was also under-representation of temporary brackish/saltwater pools, although this was poorly statistically significant (P < 0.1), and significant over-representation (P < 0.001) of both wetland systems and dam/river barriers. Regarding protection status, legally protected sites were over-represented in literature (P < 0.001).
There was a general increasing trend in publication rates, rising particularly rapidly over the past two decades. Several national and international conventions and legislation, such as the Habitats Directive, could have contributed to such a trend (Figure 2).
T A B L E 2 Summary of the number and surface area of inventoried Sardinian wetlands, grouped according to macro-category and type  F I G U R E 1 Coefficient plots (made by 'coefplot' function in the 'coefplot' R package, Lander, 2018). Dots depict coefficient estimates derived by the log-linear model, while the whiskers indicate the related standard error around each estimate. Asterisks (P < 0.1; *: P < 0.05; ***: P < 0.001) are for the positive (light green) and negative (dark green) significant estimates. Consult Table 2

| Analysis by research topic
Among the 11 defined research topics (Table 1), more than half of the 344 research papers concerned flora and/or fauna (Figure 3).
Pollution was another common research topic (58 papers), followed by conservation assessment and management (32) and economic production (27).
Most papers (264; 76%) dealt with only one research topic.
Seventy-three papers (21%) were related to two research topics, and only nine papers (3%) addressed three research topics. This pattern varied among topics. For example, papers dealing with fauna were multi-topic (also addressing subjects such as pollution, economic production, diseases, and conservation assessment and management) significantly more often than papers dealing with flora ( Figure 3). Five research topics were more often considered in a multi-topic than a single-topic approach, including conservation assessment and management (mostly with flora and fauna), economic production (with sustainable economy, conservation assessment and management, pollution and diseases), and geology and palaeontology (with hydrology, history and archaeology, and flora).

| Wetland inventory, literature analysis, and conservation implications
The findings of this study have presented the most complete inventory to date of Sardinian wetlands, comprising 2,501 mapped sites. The inventory was founded on analyses of satellite imagery, which offers great potential for such work because of its repeated, homogeneous coverage of large regions (Rebelo, Finlayson & Nagabhatla, 2009). The inventory represents a valuable tool for the scientific community, policy makers, and wetland managers. For instance, recognizing that 2,274 of the 2,501 Sardinian wetlands have never been investigated is a crucial starting point and offers an important baseline for continuing monitoring.
As expected, the most commonly studied wetlands were the largest. These offer the most potential study sites, but are also the most affected by a long history of human presence, reflecting both historical and archaeological interests and recent economic interests (King, 1977). For example, the largest Sardinian wetland, Santa Gilla Lagoon, is adjacent to the island's capital, Cagliari.
Natural large coastal wetlands are the most protected for their high environmental value, but their conservation status is changing, especially owing to pollution problems and hydrological impacts related to past activities such as malaria eradication or agricultural expansion, or more recent impacts associated with human population growth (Marchi & Munstermann, 1987;Geijzendorffer et al., 2019a).
Conservation policies and management continue to be hampered by inadequate and/or obsolete scientific knowledge (Orlikowska et al., 2016). Even basic research was limited on small wetlands, temporary wetlands, and most artificial wetlands. For example, in Sardinia, other research on topics such as conservation assessment and management, diseases, and pollution, is largely absent in relation to small wetlands. However, these were recorded here and generally found as more widespread issues among all wetland types.  (Boix et al., 2012;Bagella et al., 2016). However, they are threatened by several human activities, mainly related to habitat fragmentation, vegetative degradation, and pollution. The high sensitivity of these environments is intrinsically the result of their limited extent, but also because they are often poorly identified and unappreciated, leaving them vulnerable even to unintentional destruction (Bagella, Caria & Filigheddu, 2013). To make significant progress on their conservation, improved basic information such as inventories is needed, but they should also be included as a specific water-body type under the European Water Framework Directive (Oertli et al., 2005;Bolpagni et al., 2019). Understanding has also increased for: (i) wetland ecosystem services, such as those related to economic production, hydrology, and human health; and (ii) archaeological aspects, justified by the presence of ancient settlements and activities.
This literature review has highlighted possible knowledge gaps: topics that have been rarely studied in the Sardinian context. First, Sardinian wetlands have been used by humans for thousands of years for example as a water resource, for fishing-hunting-gathering, (proto-) agriculture (Carenti, 2013)but surprisingly attitudes towards wetlands have not been more widely studied. The sustainable use of wetlands will enhance their long-term conservation, avoiding conflicts between human needs and maintenance of ecological functions (Brock, 2009 Fine-scale and multi-topic information, such as that contained in the inventory, might specifically support regional, provincial and municipal masterplans. This inventory might be included in the datasets of the Italian Institute for Environmental Protection and Research (ISPRA) and consulted by regional and provincial authorities when they are making decisions about environmental policies at various scales.
International research centres and non-governmental organizations might support and, in some cases, spur policy initiatives by promoting cooperation and knowledge sharing between local experts and civil associations.