Decline of Aquatic Plants in an Iconic European Protected Natural Area

Decline of aquatic plants in an iconic European protected natural area Pablo Garcia-Murillo a , Carmen Díaz-Paniagua b,* , Rocío Fern´ andez-Zamudio b a Department of Plant Biology and Ecology, University of Sevilla, 41004, Sevilla, Spain b Do˜ nana Biological Station-CSIC, Avda Am´ erico Vespucio 26, 41013 Sevilla, Spain ARTICLE INFO Keywords: Aquatic macrophytes Do˜ nana Threatened flora Extinctions Wetlands Mediterranean ponds ABSTRACT We examined occurrence patterns over time for aquatic vascular plant species in a major European natural protected area: Do˜ nana National Park (southwestern Spain). We used all available records (1965 to the present) for aquatic vascular plants found in the park’s pond network. Information was available for 38 species across 263 sites, including more than 100 ponds that were intensively monitored between 2001 and 2023. Our results show that aquatic vascular plant species began to decline in the last third of the 20th century. Two phases were apparent: 1) an early phase (late 20th century), during which declines in aquatic vascular plant populations were largely provoked by the arrival of an invasive species, Procambarus clarkii, and 2) a more recent phase during which the overexploitation of the aquifer has resulted in many ponds drying up and in shorter flooding periods in the remaining ponds. At present, nine species have disappeared, and 72% of the remaining species have smaller ranges. Our findings suggest that the dramatic local extinction and decline of Do˜ nana’s various aquatic plant populations is a consequence of the rapid habitat loss that has occurred in recent decades. Historically, the park’s pond network acted as a refuge for many endangered and unique species facing major threats to their survival. Such no longer appears to be the case. 1. Introduction Habitat loss has been described as the greatest challenge of our times (Lambdon & Cronk, 2020) and is leading to species extinction. Aquatic habitats rank among the world’s most threatened ecosystems: approximately 87 % of natural wetlands have been lost since the early 18th century (Davidson, 2014; Dudgeon et al., 2006). An enormous amount of water is currently required to meet agricultural, urban, and industrial demands (Acreman et al., 2022), which is leading to the alteration and destruction of aquatic habitats. The result is a worldwide loss of native aquatic species, especially in places well suited to agricultural production and urban development (Kolp et al., 2021). The earliest records signaling declines in aquatic vegetation date back to 1900, with the pattern being seen in the Northern and Southern Hemispheres and at high and low latitudes; this decline has accelerated over recent decades (Zhang et al., 2017). In Europe over the last century, aquatic habitat loss and deterioration have reduced the richness of freshwater vascular plant species (Sand-Jensen et al., 2018). Additionally, a recent assessment conducted as part of Europe’s Water Framework Directive (WFD) revealed that around 60 % of water bodies in the EU do not have a healthy ecological status and pointed to hydromorphological changes as a key pressure on surface water bodies (Kristensen et al., 2018). The most effective response to this challenge has been provided by protected natural areas, in which the preservation of optimal, unaltered habitats fosters the persistence of plant and animal populations, especially those of endangered species (Cazzolla et al., 2023; Pimm et al., 2018; Sergio et al., 2022). In recent years, the concerns of conservation biologists about managing threats to prevent species extinction have resulted in considerable changes to conservation policies; notably, the number and quality of reserves have increased, and management practices have improved. Much more attention is also being paid to species losses and habitat deterioration, which serve as early warning signs that protected areas are no longer serving their functional purpose (Sergio et al., 2022). However, stakeholders in Mediterranean ecosystems vary in how they interpret this information (Moreira et al., 2019; Salmer´ on et al., 2021). Located in southwestern Spain, Do˜ nana National Park is one of Europe’s best known natural protected areas, and it has long served as an example of how to preserve species and biological communities. For example, the park has helped many species escape extinction, including animals such as the Iberian lynx (Pacín et al., 2024), the crested coot * Corresponding author. E-mail address: [email protected] (C. Díaz-Paniagua). Contents lists available at ScienceDirect Journal for Nature Conservation journal homepage: www.elsevier.com/locate/jnc https://doi.org/10.1016/j.jnc.2024.126814 Received 1 October 2024; Received in revised form 18 December 2024; Accepted 18 December 2024 Journal for Nature Conservation 84 (2025) 126814 Available online 19 December 2024 1617-1381/© 2024 The Author(s). Published by Elsevier GmbH. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). (Hortas et al., 2023), and the marbled teal (Ortego et al., 2024) and plants such as Armeria gaditana, Linaria tursica, and Vulpia fontquerana, among many others (Asensi and Diez-Garretas, 2017; Fern´ andezZamudio et al., 2019; García-Murillo et al., 2019). Do˜ nana is one of Europe’s most important wetlands—it is a major overwintering site for waterfowl and contains the continent’s largest temporary pond network (Díaz-Paniagua et al., 2015; Green et al. 2018; Scheffer et al., 2015). It has received several international designations of note (Ramsar Site, World Heritage Site, Special Protection Area for Birds [European Birds Directive]) and is also a site of community importance (SCI; European Habitats Directive), all of which underscores the park’s important role in biodiversity conservation in the western Mediterranean Basin and western Europe. Most of Do˜ nana’s wetlands are temporary, and those located on sandy substrates are groundwater dependent (Custodio et al., 2009; Díaz–Paniagua et al., 2015), and their inundation cycle requires adequate proximity to the water table to ensure that, after autumn or winter rains, they can flood and persist long enough to permit the successful development of the organisms they host. The national park has been granted Spain’s highest possible level of legal protection. Unfortunately, the aquifer under Do˜ nana extends far beyond the park’s boundaries, and agricultural irrigation and residential usage (local villages and tourist resorts) in neighboring areas is resulting in intensive groundwater abstraction. Consequently, the groundwater held in the Do˜ nana aquifer has been gradually declining over recent decades (Custodio et al., 2009; Dimitrou et al., 2017). These dynamics, combined with the effects of climate change (i.e., reduced annual rainfall over the last 10 years), have led to shorter inundation cycles in approximately 80 % of ponds and the complete absence of inundation in 59 % of ponds (de Felipe et al., 2023; Díaz–Paniagua et al., 2023; 2024). However, the magnitude of the impact produced by the overexploitation of the aquifer on the aquatic organisms of this protected area has not yet been documented in detail, which we urgently require if we want to make evidence-based decisions to better protect the Do˜ nana ecosystems. Do˜ nana represents an important hotspot for aquatic plants (GarcíaMurillo et al., 2006; García-Murillo & Fern´ andez–Zamudio, 2015). The flora of a given area is affected by internal and external forces, whose impacts are particularly pronounced in habitats experiencing anthropogenic disturbances. Modern ecologists are using plant species records to better understand historical conditions, landscape ecology, and conservation (MacKenzie et al., 2019). Floristic studies can provide a reference point for gauging the success of conservation efforts. To understand the degradation of inland wetlands over time, it is crucial to characterize macrophyte occurrence and community composition since alterations can lead to dramatic decreases in habitat structural complexity and biodiversity (Cantonati et al., 2020). Macrophytes strongly influence their local physicochemical environments (Fern´ andez–Zamudio et al., 2016). Changes in aquatic plant community abundance or composition are often obvious signals that the ecological condition of a wetland has changed (Birk et al., 2012) and are good indicators of eutrophication or organic pollution and hydrological or morphological changes (Birk et al., 2012). By assessing which species occur within a given region over time, it is possible to determine whether certain plant species have disappeared or experienced shifts in distribution or abundance, which provides insight into patterns of occupancy and fragmentation and can yield findings that inform conservation actions (Büttner et al., 2022). In this sense, the extinction of plant species and thus the loss of biodiversity in target regions can be shown by comparing historical data (herbarium specimens and/or old floristic publications, historical datasets) with the present-day situation. This approach has been used to examine country-scale extinction patterns in southern Europe (Bornand et al., 2016; Kempel et al., 2020). Since its first consideration as a protected area in 1964, several botanical sampling surveys were conducted around that time, including some that collected aquatic plants (Cabezudo, 1974; Castroviejo, Vald´ es-Bermejo, Rivas-Martínez, & Costa, 1980; Rivas-Martínez, Costa, Castroviejo, & Valdes, 1980); these efforts have provided extensive historical information that can be used to assess the intensity of recent park deterioration. In this manuscript, we have compiled all available data on the aquatic vascular plants of Do˜ nana’s pond network with the aim of quantifying the changes in this group’s biodiversity. Indeed, aquatic vascular plants are particularly sensitive to the environmental transformations that have been occurring in the vicinity of Do˜ nana National Park. Here, we examine the loss of aquatic vascular plant species over time in the wetlands of the park’s sandy area. Our main goal was to quantify variation in this group’s biodiversity. More precisely, the objectives were to: 1) identify how many aquatic vascular plant species that were present 50 years ago were no longer observed after 2020; and 2) assess the changes in aquatic vascular plant occurrence over time, evidence that can inform urgent decision-making aimed at preserving Do˜ nana’s unique aquatic habitats. 2. Materials and methods In this study, we used all available records—from 1965 to the present—for aquatic vascular plants in the freshwater wetlands in the sandy area of Do˜ nana National Park. We also gathered data on plant presence from general studies, herbaria, personal collections, and field notes made by botanists and the authors. 2.1. Study area Do˜ nana National Park (around 36◦59 ′ N, 6◦27 ′ W) is located in Fig. 1. Location of the ponds surveyed for aquatic plants in the sandy area of Do˜ nana National Park. P. Garcia-Murillo et al. Journal for Nature Conservation 84 (2025) 126814 2 southwestern Spain (Fig. 1) and spans an area of approximately 54,300 ha. It is on the right bank of the Guadalquivir River, in the estuary zone where the river encounters the Atlantic Ocean. It experiences a Mediterranean climate—mild winters (mean winter temperature =10.9 ◦C), hot dry summers (mean summer temperature =23.5 ◦C), and mean annual rainfall of around 550 L/m 2 (data recorded from 1978 to 2023 in a meteorological station located in the central area of the park). This region hosts a diverse range of ecosystems and is widely acknowledged as one of Europe’s most important wetlands (Green et al., 2018). The park contains two well-differentiated geomorphological zones: an extensive marsh on a clay substrate and a sandy area. The marsh forms a broad floodplain that is fed by water from various tributary streams and the accumulated rainfall from autumn to early spring. The sandy area is formed of moving and stable dunes, interspersed with a multitude of irregular and moderately sized depressions that form groundwater–dependent ponds after heavy rains. The transition zones between the marshes and the stable dune area and between the moving dunes and stable dunes (the peridune area) contain the largest abundance of ponds in general, as well as the ponds with the largest hydroperiods (i.e., inundation period duration); consequently, the transition zones contain the park’s highest levels of biodiversity (Díaz–Paniagua et al., 2015). In this study, we focus on the aquatic vascular plants found in the pond network, which we have been intensively surveying since 2003. The marshes were not considered within the scope of our study. 2.2. General description of the pond network Most of Do˜ nana’s ponds have been classified as Mediterranean temporary ponds (EU priority habitat code 3170). They are shallow groundwater-dependent wetlands that flood after rainy periods and commonly dry out in summer. Pond flooding may occur in autumn, winter, or spring (exceptionally), depending on the timing or quantity of annual rains, which are highly unpredictable in the Mediterranean. Therefore, pond hydroperiods can vary widely from year to year. In a wet year (2004), researchers observed more than 3,000 temporary ponds in the park using remote-sensing data (G´ omez-Rodríguez et al., 2011). In contrast, natural permanent ponds are uncommon. We consider that there are only four: three big ponds in the sandy area and a fourth along the border of the marsh; they retain water year round, except during severely dry periods. Recently, the hydroperiods of some of these permanent ponds have changed. Since 2012, one pond has dried out every summer, another has dried out in 3 of the last 12 years, and a third, the biggest and deepest one, dried out during the summers of 2022 and 2023. At present, only the pond located along the border with the marsh can be said to permanently retain a shallow level of water. Groundwater discharge originating beneath the moving dunes forms small permanent streams (called retuertas by locals) that run between the marshes and the dunes. There are also intermittent streams that, in rainy periods, flow through the area and into the neighboring marsh. In previous decades, the streams would form ponds (ca˜ nos) at their mouths that could hold water all year. However, in the last decade, most have dried out in the summer. In addition, in 2005, two large artificial ponds were built for sediment deposition and the partial treatment of waters arriving from agricultural areas around the park. Likewise, Do˜ nana contains around 200 ponds that were artificially transformed into permanent water bodies. Named zacallones by the locals, they are usually small and deep (approximately 15 x 4 m in length and 1.5–2 m in depth) and were created by deepening a small area of natural temporary ponds to make groundwater available for different traditional uses (e.g., watering animals, irrigating small farms). They are no longer used for these purposes, but they remain important permanent aquatic habitats for many organisms, with aquatic plants near the top of the list. The physicochemical characteristics of Do˜ nana ponds have been described in previous studies (Fern´ andez-Zamudio et al., 2016; G´ omezRodríguez et al. 2009; Serrano et al., 2006). They are freshwater ponds, with pH commonly ranging from 5.7 to 9.3, low electrical conductivity (0.24–4.8 mS/cm), high concentrations of Fe 2+ , and poor carbonate content. Only in exceptional ponds, we have recorded high conductivity values, as recorded in a pond located in the southern park, in which we registered 12.87 mS/cm in 2021. 2.3. Species nomenclature and selection criteria for aquatic plants Defining an “aquatic plant” can be complicated. Here, we applied the criteria of Cirujano et al. (2014), for whom aquatic plants are plants that complete their life cycles as submerged or floating organisms in aquatic environments. Unlike Cirujano et al. (2014), but in agreement with Bolpagni et al. (2020), Fois et al. (2024), and Murphy et al. (2019), we defined Avellara fistulosa, Eryngium corniculatum, and Juncus heterophyllus as aquatic plants because, to successfully develop, they must live in water during at least the early stages of development. Evidence for this requirement can be seen in stem and leaf anatomy for juveniles of these species. For plant nomenclature and taxonomy, we used Flora Iberica (Castroviejo et al., 1986–2020) criteria, even if some nomenclatural and/or taxonomic changes occurred after the publication of that work. Ranunculus peltatus is a problematic taxon because there is a high degree of plasticity in certain diagnostic traits; we thus focused on the species level, avoiding the different criteria used by different authors to identify subspecies (Cook, 1986; Pizarro, 1995; Velayos, 1988; Wiegleb et al., 2017). The diminutive Elatine species (E. brochonii, E. hexandra, and E. macropoda) display strong phenotypic plasticity (Łysko et al., 2022; Razifard et al., 2017) and can only be easily distinguished using reproductive traits. Because we frequently observed them in their vegetative state, we refer to them as the Elatine minute group (Elatine sp.). In this study, we only discuss autochthonous species. We are aware that Azolla filiculoides was first detected in the park in 2001 (García–Murillo et al., 2007) and was seen covering a large percentage of the marsh borders between 2001 and 2009, including some ponds in close proximity to the sandy area (Díaz-Delgado et al., 2011). The species’ presence in the Do˜ nana pond network is now limited to a few isolated points near the marshes (see the present-day map in Appendix A). 2.4. Threatened species For the species considered in this study, we have taken into account threat levels and protection levels, based on regional, national, and European regulations, as well as the information contained in red lists and red books. These sources of reference are indicated in Table 1. 2.5. Sources of aquatic plant data Because Do˜ nana is an area of great natural value, it has been relatively well prospected from a floristic point of view. However, surveys largely focused on terrestrial plants, generally ignoring aquatic plants. From 1965 to 2000, few intensive data were collected across the park; most of the information available relates to species occurrence. These data are nonetheless important because they indicate the presence of singular or threatened species that we have not observed in more recent surveys. The first data on aquatic vegetation were collected by Marazanof (1967), during one of the initial expeditions to Do˜ nana and were a complementary part of descriptions of the area’s biodiversity, including wetland biodiversity. The first publications specifically focused on the flora in the protected area were Cabezudo (1974; 1975; 1978) and Galiano & Cabezudo (1976), which were followed by Castroviejo et al. (1980) and Rivas-Martínez et al. (1980). This body of work established the basis for what is known about Do˜ nana’s flora and vegetation. However, it is important to reiterate that these studies were especially interested in terrestrial plants and provide sparse information about aquatic plants. Data on the park’s aquatic plants generally come from García-Murillo et al. (2006; 2014), Garcia–Murillo & P. Garcia-Murillo et al. Journal for Nature Conservation 84 (2025) 126814 3 Table 1 List of the aquatic plant species historically found in the sandy area of Do˜ nana National Park, indicating biotype, habitat (permanent or temporary), year of the last observation, and presence in the park. Total species occupancy (%) was estimated by dividing the total number of grid cells in which each species was detected by the total number of grid cells surveyed historically (n =135). Species occupancy across the park was also examined by looking at the number of grid cells occupied by the species in the northern park (N), central peridune area (P), and southern park (S). Results are shown for the GLMMs conducted using the number of grid cells occupied by each species between 2001 and 2010 (P2000) and between 2021 and 2023 (P2020) (significant differences in bold), only for those species for which we have data in these two periods. Also indicated are the red-list categories in which each species is classified (and the related references). Species Red-list category Biotype Habitat Last year observed % total % N % S % P P2000 P2020 GLMM results Apium inundatum DD 4 AM temporary 2023 37.04 68.00 12.00 20.00 26 11 χ 2 ¼7.7121, p ¼ 0.0055 Avellara fistulosa EN 10, 11 CR 1, 2, 4, 7 AM temporary 2023 13.33 90.00 5.00 5.00 10 4 χ 2 ¼12.394, p ¼ 0.0004 Callitriche brutia LC 1 SA temporary 2023 44.44 78.33 8.33 13.33 37 23 χ 2 ¼6.6601, p ¼ 0.010 Callitriche lusitanica NT 1 EN 2 DD 4 SA temporary 2023 13.33 72.22 11.11 16.67 6 2 χ 2 =1.923, p = 0.1655 Callitriche obtusangula DD 4 SA temporary 2023 28.89 66.67 12.82 20.51 20 7 χ 2 ¼16.794, p < 0.0005 Callitriche stagnalis LC 1 AM temporary 2023 32.59 70.45 11.36 18.18 25 10 χ 2 ¼9.5807, p ¼ 0.0020 Callitriche truncata subsp. occidentalis DD 1 SA temporary 2023 17.78 58.33 12.50 29.17 4 9 χ 2 =2.7359, p = 0.0981 Ceratophyllum demersum DD 4 SP permanent 1978 2.22 66.67 0.00 33.33 0 0  Elatine alsinastrum NT 1, 2 SA temporary 2023 17.78 87.50 0.00 12.50 9 6 χ 2 =0.8637, p = 0.3527 Elatine sp. NT 1, 2 * DD 4 SA temporary 2023 16.30 54.55 9.09 36.36 7 7 χ 2 =0.8637, p = 0.3527 Eryngium corniculatum VU 4 AM temporary 2023 25.93 71.43 8.57 20.00 24 2 χ 2 ¼10.445, p ¼ 0.0012 Hydrocharis morsus-ranae CR 2, 4, 7 EN 10, 11 VU 12 FP permanent 2006 1.48 0.00 100.00 0.00 0 0  Illecebrum verticillatum LC 1 AM temporary 2023 40.74 78.18 0.00 21.82 31 17 χ 2 ¼11.103, p ¼ 0.0009 Isoetes velata NT 1 DD 4 SA temporary 2021 8.89 83.33 0.00 16.67 10 1 χ 2 ¼4.4664, p ¼ 0.0346 Isolepis fluitans DD 4 FA temporary 2021 16.30 77.27 0.00 22.73 16 2 χ 2 ¼27.498, p < 0.0005 Juncus heterophyllus NT 1 AM temporary 2022 43.70 72.88 6.78 20.34 29 28 χ 2 =2.2013, p = 0.1379 Lemna gibba LC 1 FA permanent 2023 34.81 48.94 44.68 6.38 27 8 χ 2 ¼11.599, p ¼ 0.0007 Lemna minor LC 1 FA permanent 2023 51.85 51.43 34.29 14.29 35 20 χ 2 ¼8.0529, p ¼ 0.0045 Lemna trisulca CR 2, 3 DD 4 FP permanent 2022 7.41 0.00 100.00 0.00 5 4 χ 2 =0.4399, p = 0.5072 Marsilea strigosa VU 1, 2, 4, 5, 6, 8, 10, 12, 13 AM temporary 2023 1.48 100.00 0.00 0.00 0 1 χ 2 ¼6.2389, p ¼ 0.0125 Myriophyllum alterniflorum LC 1 SA temporary 2023 48.15 75.38 4.62 20.00 39 20 χ 2 ¼9.9806, p ¼ 0.0016 Najas marina LC 1 SA permanent 2018 0.74 0 0 100 0 1  Nuphar luteum CR 4 VU 10 FP permanent 1981 0.74 100.00 0.00 0.00 0 0  Nymphaea alba CR 4 VU 10 FP permanent 2008 2.22 100.00 0.00 0.00 0 0  Potamogeton crispus LC 1 SP permanent 1978 2.22 66.67 0.00 33.33 0 0  Potamogeton lucens DD 4 SP permanent 2023 10.37 7.14 64.29 28.57 8 7 χ 2 ¼8.9623, p ¼ 0.0028 Potamogeton natans DD 4 SP permanent 2023 15.56 90.48 0.00 9.52 6 6 χ 2 =0.3116, p = 0.5767 Potamogeton pectinatus LC 1 SP permanent 2023 24.44 42.42 33.33 24.24 14 15 χ 2 =0, p =1 Potamogeton polygonifolius DD 4 AM permanent 2010 5.93 87.50 12.50 0.00 5 0 χ 2 ¼122706, p < 0.00001 Potamogeton trichoides LC 1 SA temporary 2023 23.70 46.88 28.13 25.00 8 10 χ 2 =1.1898, p = 0.2754 Ranunculus peltatus LC 1 SA temporary 2023 70.37 61.05 25.26 13.68 44 48 χ 2 =0.0308, p = 0.8606 Ranunculus tripartitus LC 1 SA temporary 2021 20.00 81.48 3.70 14.81 13 2 χ 2 ¼22.342, p ¼ 0.00001 Ruppia drepanensis DD 4 SA temporary 2022 5.19 0.00 57.14 42.86 3 1 Х 2 =2.9929, p = 0.0836 Spirodela polyrhiza DD 4 FP permanent 2013 2.22 33.33 66.67 0.00 0 0  Utricularia australis EX 4 , EN 10 FP permanent 1977 4.44 83.33 16.67 0.00 0 0  Utricularia gibba NT 1 , CR 2, 4, 7 VU 10, 12 FP permanent 1977 0.74 100.00 0.00 0.00 0 0  Wolffia arrhiza EN 2,4, VU 10, 12 FP permanent 2023 11.11 6.67 93.33 0.00 9 1 χ 2 ¼323323, p < 0.00001 Zannichellia obtusifolia NT 1 , VU 4 SA temporary 2023 25.19 26.47 47.06 26.47 8 13 Х 2 =2.4203, p = 0.1198 P. Garcia-Murillo et al. Journal for Nature Conservation 84 (2025) 126814 4 Fernandez–Zamudio (2015), and Fern´ andez-Zamudio et al. (2016), in addition to various monographs on hydrophyte genera (Garcia–Murillo et al., 2000; Talavera et al., 1986) and chorological studies (Fern´ andez–Zamudio et al., 2006; García–Murillo et al., 1990). Although charophytes are important components of aquatic ecosystems, there is not abundant historical information about these macrophytes in Do˜ nana’s ponds. Therefore, we have not included this group in this study, although we are currently assessing their conservation status for a future study. Overall, for the period from 1965 to 2023, there are vegetation data for 263 sites across the park (Fig. 1), including 127 temporary ponds (of which 25 have dried up in recent years: 2021–2023); 21 ponds formed by intermittent streams (ca˜ nos); 4 permanent natural ponds (of which 3 have dried out during at least one summer over the last decade); 2 artificial ponds; 106 zacallones (of which 16 have disappeared and 32 now dry out in the summer); and 3 retuertas (Table 2). In previous studies describing the dynamics of animal and plant assemblages in the Do˜ nana pond network (Fern´ andez-Zamudio et al., 2016; Florencio et al., 2014), we intensively monitored 167 sites from 2001 to 2010, noting all the aquatic plant species occurring at each site. To evaluate changes in the park’s plant assemblages, we conducted new surveys from 2021 to 2023, during which we revisited 114 of the above sites as well as 31 other sites that had hosted ponds in previous decades (but that have dried up in more recent years). We also consulted the following herbaria for this work: SEV (University of Seville, Seville, Spain), MA (Royal Botanical Gardens-CSIC, Madrid, Spain), UNEX (University of Extremadura, Badajoz, Spain), and HSS (Research Center of La Orden-Valdesequera, M´ erida, Spain). 2.6. Data analyses We constructed a data matrix comprising the species for which we had records and the years they were observed (from 1965 to 2023). The data were grouped into four periods: a) before 2000, b) 2001–2010 (hereafter, P2000), c) 2011–2020, and d) 2021–2023 (hereafter, P2020). We divided the entire park into Universal Transverse Mercator (UTM) grid cells (1 km x 1 km) onto which we mapped the periodspecific presence of each species using Qgis (v. 3.16). While we only had sporadic observations for the period before 2000, the information available to us was extremely valuable in detecting the local extinction of certain species. In total, we had information for 135 grid cells. Data from 2003 onwards included periods of intensive and systematic surveys that employed similar methodologies; they were useful for evaluating changes in species occupancy (i.e., the presence of a species in the grid cells) throughout the park. Specifically, we had intensively surveyed ponds (n =164) in 73 grid cells over two periods (P2000 and P2020), allowing us to conduct statistical comparisons. The park’s ponds are heterogeneously distributed. Pond density is high in the northern park, even higher in the central peridune area, and lower in the southern park (G´ omez-Rodríguez et al., 2011). Because comparing individual ponds could have given rise to pseudoreplication, we used the grid cells as the units in which the presence of aquatic plant species was quantified. We estimated the number of grid cells occupied by each species and then divided this number by the total number (n = 73) of cells surveyed during both periods (P2000 and P2020) to arrive at relative species abundance. To conduct comparisons, we used generalized linear mixed models (GLMMs) with a binomial error distribution; period was the predictor variable, and grid cell identity was a random variable. We also estimated the number of species recorded in each grid cell during the two periods and then performed comparisons using a GLMM with a Poisson error distribution (grid cell identity =random factor). For each species, we estimated the number of grid cells occupied during each period. We then used the difference in grid cells occupied between periods as the response variable in a generalized linear model (GLM) with a Poisson error distribution. In this model, the number of grid cells occupied in P2000 was a covariable, and the predictor variables were species area of distribution (northern, central, and southern); biotype (AM: amphibious annual; SA: submerged annual; FA: floating annual; SP: submerged perennial; FP: floating perennial) and habitat type (permanent or temporary in P2000). Post-hoc Tukey tests were used for paired comparisons between different areas of the park. 3. Results 3.1. Disappearance and decline of Dona˜ na’s aquatic vascular plant species We were able to obtain records from 1965 to 2023 for 38 species of autochthonous aquatic vascular plants that occur in the wetlands of Do˜ nana’s sandy area (Table 1). These species comprised 8 amphibious plants, 11 floating plants, and 19 submerged plants. In terms of habitat, 21 species were found in temporary wetlands, and 17 occurred in permanent wetlands. All were members of Spermatophyta, with the exception of Marsilea strigosa, a member of the aquatic Pteridophyta, and Isoetes velata as a Lycopodiophyta. We mapped species occurrence in space and time (Appendix A). It was apparent from the records we gathered that nine aquatic vascular plant species—Ceratophyllum demersum, Potamogeton crispus, Nuphar luteum, Nymphaea alba, Utricularia gibba, Utricularia australis, Potamogeton polygonifolius, Hydrocharis morsus-ranae, and Spirodela polyrhiza—have disappeared from the park. All nine of these species tended to occur in permanent water bodies; six were floating perennials, two were submerged perennials, and one was amphibious (Fig. 2). Ceratophyllum demersum and P. crispus were relatively abundant in the Iberian Peninsula until the 1980 s. They are species typical of permanent backwaters and are both easy to identify. Their first records in the study area are found in Marazanof (1967), who collected C. demersum in 1965 in a permanent pond in the central peridune area; he found P. crispus in this same pond, as well as in a ca˜ no along the marsh border. Both species were last observed in 1978, when C. demersum was collected from a different permanent pond, and P. crispus was collected from a different ca˜ no at the marsh border (Castroviejo et al., 1980). Do˜ nana used to contain two water lilies—N. luteum and N. alba—that are currently protected by regional law (Decreto 23/2012 in Table 1). AM: amphibious; SP: submerged perennial; SA: submerged annual; FP: floating perennial; FA: floating annual. 1: IUCN (2024): (https://www.iucnredlist.org/); 2: Moreno (2010); 3: Ba˜ nares et al. (2006); 4: Cabezudo et al. (2005); 5: Moreno et al. (2019); 6: Ba˜ nares et al. (2010); 7: Ba˜ nares et al. (2004); 8: Habitats Directive 92/43/EEC; 9: Real Decreto 139/2011 (https://www.boe.es/eli/es/rd/2011/02/04/139); 10: Decreto 23/2012 (https:// www.juntadeandalucia.es/boja/2012/60/6); 11: Spanish Catalogue of Endangered Species (https://www.miteco.gob.es/es/biodiversidad/temas/conservacion-de-es pecies/especies-proteccion-especial/ce-proteccion-listado.html); 12: Blanca et al. (1999), 13: List of Wildlife Species under Special Protection Regimen (https://www. miteco.gob.es/es/biodiversidad/temas/conservacion-de-. Table 2 Number of different aquatic habitats in Do˜ nana surveyed for aquatic plants between 1965 and 2023. Also shown are the number of the permanent and temporary ponds surveyed that have dried up and the number of formerly permanent sites that now dry out in the summer. N Dried up Dry in summer Temporary ponds 127 25 102 Permanent natural ponds 4 0 3 Streams (including ca˜ nos) 21 0 10 Permanent artificial ponds 2 0 0 Retuertas (permanent dune streams) 3 0 0 Zacallones (artificially deepened ponds) 106 16 32 P. Garcia-Murillo et al. Journal for Nature Conservation 84 (2025) 126814 5 Nuphar luteum was first recorded in the park in 1981, most frequently in the large, deep backwaters of a permanent stream. However, it was not found in later years. Until the 1980 s, it was also relatively common in permanent backwaters in the Spanish regions of Andalusia, Extremadura, and Castilla La Mancha (Cirujano et al., 2014; Anthos, 2024). It then began to disappear from many places and is currently an extremely rare species in Andalusia, where it is listed in the Red List of Andalusian Flora, as critically endangered (Cabezudo et al., 2005). To thrive, it needs slow, medium-depth waters in euthrophic, but scarcely polluted environments (Cirujano et al., 2014; Melendo et al., 2003). Nymphaea alba also began to disappear from the southern Iberian Peninsula in the 1980 s and was observed in Do˜ nana national park until 2008. At present, it is a rare species found in the southern Iberian Peninsula. It still persists in the area around the national park, at one location towards the beginning of La Rocina stream, considered to be nearly the last population in western Andalusia. To thrive, it needs conditions similar to those of the previous species (Cirujano et al., 2014; Melendo et al., 2003). Utricularia australis and U. gibba are species typically found in still, dystrophic waters that are poor in dissolved nutrients and rich in humic acids resulting from plant decomposition, which contribute to environmental acidification (Cirujano et al., 2014). Until the 1970 s, U. australis was found in the study area in the permanent parts of two ca˜ nos along the marsh border. Utricularia gibba was observed in Do˜ nana until the 1980 s, in permanent backwaters of intermittent and peaty watercourses that flowed into the marsh. Both species were also present in some of the permanent streams that flowed into the northern park. Utricularia australis is protected by regional law (Decreto 23/2012 in Table 1) but is classified as extinct in the Red List of Andalusian Flora, which has labelled U. gibba as critically endangered (Cabezudo et al., 2005). Potamogeton polygonifolius is also a species typically found in rivers, streams, peat bogs, and peaty wetlands, which have acidic waters that are poor in salts and nutrients (Cirujano et al., 2014). In the study area, it occurred in the permanent waters of intermittent streams along the marsh border, where it coexisted with U. australis and U. gibba. It was still found in a single peaty pond up until 2010. Although no longer occurring in the park, it is present in the surrounding area. Historically, Hydrocharis morsus-ranae occurred in the retuertas and one of the permanent ponds along the marsh border. The last herbarium specimen was collected in 1985 from a retuerta, the only place it remained; this unique habitat is formed by groundwater discharge emerging from beneath the moving dunes, on the edge of the marsh. This site was the last location where the species was found on the Iberian Peninsula (Garcia-Murillo et al., 2000). It is protected by Spanish national law (Real Decreto 139/2011 in Table 1, amended April 7, 2023) and regional law (Decreto 23/2012 in Table 1). Likewise, H. morsusranae has been classified as critically endangered by the 2010 Red List of Spanish Vascular Flora (Moreno, 2010), the Red List of Andalusian Vascular Flora (Cabezudo et al., 2005), and the Atlas and Red Book of Spain’s Threatened Vascular Flora (Ba˜ nares et al., 2004). It has been the focus of various reintroduction and repopulation programs in Do˜ nana, where its presence has been intensively monitored. The species has not been observed naturally in the wild since 2010, but individuals of the reintroduced population were present from 2011 to 2019 (Cobo, personal communication). New reintroduction efforts are underway. Spirodela polyrhiza, or greater duckweed, was rare but observed between 1988 and 2013 in a permanent zacall´ on that has now dried up. After 2013, we never again found it at that site. Other species appear to be rare because they were present across less than 8 % of the study area in the period from 2000 to 2010. These species were Potamogeton natans, Lemna trisulca, Callitriche lusitanica. C. truncata subsp. occidentalis, and Ruppia drepanensis. Between 2021 and 2023, we observed low levels of occupancy (<4 %) for Avellara fistulosa, Isolepis fluitans, Ranunculus tripartitus, and Isoetes velata. Furthermore, in recent years, Marsilea strigosa has only been seen in an ephemeral pond, and Wolffia arrhiza has only been observed in a retuerta. Of the 38 aquatic plant species for which we had records, all them appear in a IUCN red-list category, based on the various red lists and red books we used. Twelve species (31.6 %) are classified in categories considered to be threatened with extinction (vulnerable, endangered or critically endangered). One species is extinct at the regional level: Utricularia australis; 6 are critically endangered (CR): Avellara fistulosa, Hydrocharis morsus-ranae, Lemna trisulca, Nuphar luteum, Nymphaea alba, and Utricularia gibba; 2 are endangered (EN): Callitriche lusitanica and Wolffia arrhiza; 3 are vulnerable (VU): Eryngium corniculatum, Marsilea strigosa and Zannichellia obtusifolia; 3 are near threatened (NT): Elatine alsinastrum, Juncus heterophyllus and Isoetes velata; 10 are data deficient (DD): Apium inundatum, Callitriche obtusangula, C. truncata, Ceratophyllum demersum, Isolepis fluitans, Potamogeton lucens, P. natans, P. polygonifolius, Ruppia drepanensis and Spirodela polyrhiza; and the remaining ones are classified as least concern (LC) (Table 1). Eight (21.2 %) are protected under European, Spanish, or Andalusian law. Notably, M. strigosa is protected at all three levels; A. fistulosa and H. morsus-ranae are protected at national and regional levels; and N. luteum, N. alba, U. australis, U. gibba and W. arrhiza are protected at the regional level (Table 1). 3.2. Changes in species occupancy (2003–2007 to 2021–2023) In addition to the extinct species, 21 species displayed a decrease in occupancy over time (Fig. 3). During P2000, 11 species were very abundant and occupied more than 30 % of the grid cells. During P2020, only three species had high levels of occupancy (>30 %); of these, only Ranunculus peltatus, the most widely distributed species (Fig. 4), had a higher occupancy level than previously, while Callitriche brutia and Juncus heterophyllus had a lower occupancy (Fig. 3). All other frequent species in P2000 exhibited, in P2020, a reduction of 20–30 %, as in the case of Apium inundatum (Fig. 3 and Fig. 4). Two other species (Zannichellia obtusifolia and Potamogeton pectinatus) displayed small, non- –significant increases in occupancy levels: they went from being among the species with lower occupancy in P2000 to the group of common Fig. 2. Number of vascular aquatic plant species (extant vs. locally extinct) observed in the wetlands of the Do˜ nana sandy area, classified by biotype and habitat (permanent vs. temporary). P. Garcia-Murillo et al. Journal for Nature Conservation 84 (2025) 126814 6 species in P2020. Of the 28 species for which we had data for both P2000 and P2020, 15 underwent a significant decrease in occupancy between periods, with an average reduction of 17.8 % (range: 1.36–30.1 %, Table 1, Fig. 3). Area of distribution ( χ 2 =19.378, df =2, p =0.006) and biotype ( χ 2 =11.169, df =4, p =0.0245) significantly influenced species occupancy between the two periods. The species found in the northern park differed significantly from the species found across the entire study area (Tukey test: p =0.012), and the submerged annuals differed from the floating annuals (Tukey test: p <0.019). Occupancy decreased the most for the submerged perennials, a pattern that seemed attributable to the significant decrease in Potamogeton lucens, which disappeared from 4 of the 9 grid cells it had previously occupied. We observed that both P. lucens and P. natans now only occur in ponds in a reduced number of vegetative forms that have no flowers or seeds (Fig. 5). In the case of P. natans, the species occupied a similar number of ponds in both periods, but that presence was limited to isolated individuals in P2020, contrasting with its high degree of pond coverage and flower production in P2000 (Fig. 5). The number of species per grid cell differed among the three main areas of the park ( χ 2 =31.41, df =2, p <0.0005). The central peridune area had the greatest species richness during both periods (Table 3). The number of species per grid cell also differed between the two periods ( χ 2 =61.00, df =1, p <0.0005). In P2000, the average number of species per grid cell was 7.25 (+0.61) and decreased to 4.15 (+0.34) in P2020. The decrease was highest in the northern park (49.5 %) and less pronounced in the central peridune area and southern park (29.5 and 29.7 %, respectively; Table 3 and Fig. 6). 4. Discussion Our work assembled existing records for 38 species of vascular Fig. 3. Relative abundance (%) of aquatic plants observed in the Do˜ nana pond network (number of grid cells in which a species was present divided by the number of grid cells surveyed) for 2001–2010 (P2000) and 2021–2023 (P2020). Fig. 4. Maps of Ranunculus peltatus and Apium inundatum presence in the park. Grid cell color indicates the year of the last observation: before 2000 (red); between 2001 and 2010 (blue); between 2011 and 2020 (gray); and between 2021 and 2023 (black). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) P. Garcia-Murillo et al. Journal for Nature Conservation 84 (2025) 126814 7 aquatic plants found in the ponds of Do˜ nana National Park. Our results revealed the important role played by this protected area in maintaining Iberian and European biodiversity (Diaz–Paniagua et al., 2015; Florencio et al., 2014; 2016). Indeed, more than 60 % of these species are classified into one of the IUCN’s red list categories, and nearly 25 % are protected under various laws. Thus, Do˜ nana has played a crucial role in protecting species that are highly sensitive to the anthropogenic changes that have occurred over recent decades in this region. For most of these species, existence outside the park’s borders is impossible, given the ways the environment is being aggressively modified by anthropogenic activities, which include intensive farming, heavy tourism, and petrochemical operations. In addition, even Do˜ nana is no longer a safe haven, as agriculture and tourism outside the park’s borders are profoundly transforming the habitats within the protected area, as we discuss below. Aquatic plant populations play a key role in aquatic habitats (Fern´ andez-Zamudio et al., 2016; García Murillo & Fern´ andez-Zamudio, 2015) by providing vital structural and functional ecosystem services (O’Hare et al., 2018). Therefore, the conservation of the aquatic plants in Do˜ nana’s ponds would help guarantee proper ecosystem functioning. Unfortunately, the park has not been spared from the widespread deterioration of aquatic ecosystems occurring elsewhere. In fact, the park is currently facing several conservation challenges, which are consequences of both climate change and anthropogenic disturbances. Based on all the records we gathered, we were able to identify two distinct phases of species decline. There was an early phase, around the end of the 20th century, which coincided with the introduction and spread of the red swamp crayfish (Procambarus clarkii), an exotic species that can act as an ecosystem engineer because of its dramatic impacts on the physical and chemical characteristics of aquatic habitats, leading to biodiversity losses (Geiger et al., 2005; Gherardi, 2007; Souty-Grosset et al., 2016). There was also a more recent phase related to the overexploitation of the regional aquifer, one of the main factors, which together with the increase in temperature and lower rainfall, has led many ponds to dry up completely, with the remaining ponds Fig. 5. Potamogeton natans was commonly seen covering ponds with large numbers of reproductive plants (2007 on left) but now only occurs as sparse vegetative forms (2023 on right). Table 3 Mean number of vascular aquatic plant species per grid cell in the northern park, the central peridune area, and the southern park, as well as the total number of ponds surveyed. The relative number of species lost (%) is also indicated (mean number of species present during the second focal survey period [2021–2023] versus the first focal survey period [2001–2010]). Northern park Peridune area Southern park Total 2001–2010 8.33 (±0.79) 9.54 (±1.11) 3.70 (±0.95) 7.25 (±0.607) 2021–2023 4.21 (±0.39) 6.82 (±0.84) 2.60 (±0.65) 4.15 (±0.34) % species loss 49.46 29.45 29.73 42.76 Fig. 6. Comparison of mean species number per grid cell between the two focal survey periods: 2001–2010 (P2000) and 2021–2023 (P2020). The grid colors indicate species loss (red) or gain (blue) or no difference between periods (white). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) P. Garcia-Murillo et al. Journal for Nature Conservation 84 (2025) 126814 8 experiencing reduced hydroperiods, particularly in the northwestern park (de Felipe et al., 2023). 4.1. Loss of aquatic vascular plant species Approximately 24 % of the aquatic vascular plants ever recorded in the Do˜ nana pond network have disappeared. Five species (Utricularia australis, U. gibba, Nymphaea alba, Ceratophyllum demersum, Potamogeton crispus) were last observed before 2000, and their disappearance can be linked to the expansion of the exotic red swamp crayfish (P. clarkii). This invasive species was introduced into the Do˜ nana marsh in 1974 and spread into the ponds in the sandy area around 1983 (Díaz-Paniagua et al., 2014). Research observed the disappearance of the abundant biomass produced by submerged and floating macrophytes in the marsh and its adjacent streams before the 1990 s. This pattern was attributed to the crayfish (Duarte et al., 1990), which strongly prefers macrophytes over other potential food sources (Cirujano et al., 2004). The crayfish’s impact on pond structure has been studied experimentally (Arribas et al., 2014)—it has been found to increase turbidity, boost nutrient concentrations, reduce oxygen levels, and devour all macrophyte biomass. All five of the plant species mentioned above were inhabitants of permanent waters (e.g., natural ponds or zacallones), which is also the habitat that allows the crayfish’s populations to persist (Díaz-Paniagua et al., 2014). Additionally, the explosion in agricultural activity around the park has led to a verified influx of nutrients in Do˜ nana’s aquatic habitats, which have arrived via entry waters and the increase in urban wastewaters from the increasingly populated villages adjacent to the park (Paredes et al., 2019; 2021; Peris et al., 2024). Livestock farming, high livestock densities, fertilizer usage combined with irrigation agriculture, and urban wastewaters in areas surrounding Do˜ nana have contributed to the eutrophication of aquatic habitats in the northern park, where the five above species used to grow. These factors, interacted with the spread of the exotic crayfish, had a significant impact on the populations of Utricularia australis, U. gibba, and Potamogeton polygonifolius, which need dystrophic waters poor in salts and nutrients (Cirujano et al., 2014). Most of the other submerged species were affected as well. Three species resisted the crayfish’s invasion somewhat longer. They persisted in isolated permanent waters before finally disappearing during the next decade, when eutrophication and the broad-scale reduction in pond hydroperiods made the species’ permanent habitats temporary. Indeed, Potamogeton polygonifolius, and Spirodela polyrhiza were found in isolated ponds and ca˜ nos up until 2010 and 2013, respectively, and Hydrocharis morsus-ranae was observed in one pond in which it had naturally occurred in previous decades. The disappearance of aquatic plant species from Do˜ nana represents more than a local loss of biodiversity; it has impacts at larger scales, affecting biodiversity on the Iberian Peninsula and even in Europe. The persistence of these species has special importance, since most are part of peripheral populations, and their disappearance means that the species’ extent of occurrence (EOO) is greatly reduced (Joppa et al., 2016). When EOO significantly decreases over a short period of time, it is a recognized indicator that a species’ conservation status is deteriorating (Izco, 2015). For H. morsus-ranae, the closest verified populations are found in France (García Murillo et al., 2000; 2004); for U. gibba, it is Morocco. Algeria and Tunisia, in northern Africa (de B´ elair & Rhazi, 2010). In the case of U. australis, it is now northern Spain (Laínz, 1979; Mayor & Alvarez, 1978; Robinson et al., 2009), as the last time the species was seen in the southern half of the country was in 1979, in a stream in the Guadamellato Valley in the province of C´ ordoba (Arenas et al., 1983); it is now considered to be extinct in Andalusia (Cabezudo et al., 2005). For N. luteum, the nearest populations are found in the province of Badajoz in western Spain (UNEX herbarium). The situation is similar for N. alba: the next closest population is located around 280 km to the northwest, also in the province of Badajoz (Bautista et al., 2012). The same is true for S. polyrrhiza (specimens in HSS herbarium), P. polygonifolius (García Río, 2006), and P. crispus (T´ ellez et al., 2008). 4.2. Recent decrease in species ranges Of the aquatic vascular plant species still found in the Do˜ nana pond system, 78.9 % have experienced a decline or even extinction of their populations. This reality is startling given that Do˜ nana is one of the most important national parks in Europe and that several of these species are also protected by national or regional laws. The effects on aquatic plant communities may be related to the general deterioration of aquatic habitats that has been occurring in the region over the last four decades. The greatest threat to the conservation of Do˜ nana’s wetlands is anthropogenic pressure on water resources: Do˜ nana ponds must have access to adequate levels of groundwater to maintain their inundation and desiccation cycles. Recent research has described the deterioration of the park’s pond network, characterized by a gradual reduction in pond hydroperiod and surface area, as well as the drying up of around 60 % of ponds (de Felipe et al., 2023). The region’s aquifer is being overexploited to irrigate agricultural fields around the park and supply water to neighboring vacation resorts and residences. Local depression cones around the pumping area are causing declining water levels and reducing natural discharge from the aquifer to certain ponds (Custodio et al., 2009; Manzano et al., 2005; Serrano & Serrano, 1996). A similar effect is being caused by the increased temperatures and lower rainfall levels in recent decades (de Felipe et al., 2023). Our study also confirms that aquatic habitats are being lost, given that nearly 20 % of the sites we surveyed between the 2000 s and 2020 s have dried up. These habitats include Mediterranean temporary ponds, an EU priority habitat (code 3170), considered close to be a disappearing ecosystem (Bagella, 2023; Zacharias and Zamparas, 2010). Additionally, large permanent ponds have been transformed into seasonal ponds. Unsurprisingly, the loss of ponds is affecting aquatic plant populations, as we have underscored in this study. The greatest loss of vascular aquatic plant species has occurred in the part of the park with the greatest loss of ponds and largest decrease in pond hydroperiod (see Fig. 6 in de Felipe et al., 2023), namely the northwestern park, which has been the most affected by aquifer overexploitation. In fact, some species commonly found throughout the park have been infrequent in this area because there are no longer any ponds to inhabit, and the species that are restricted to the northern park or to the central peridune area are those whose ranges have decreased the most. Some species that have historically been rare in Do˜ nana still persist, even if their presence is now even more limited. Sometimes, their populations are restricted to a few isolated locations, where only a small number of individuals can develop. Given these circumstances, it seems likely that they will eventually disappear. This trend is particularly alarming for A. fistulosa, a monospecific genus whose last remaining populations are those in Do˜ nana (Ortiz Herrera et al., 2008). It inhabits wet areas covered by hygroturbous grasslands in the ecotone between the sandy area and marshland. The species has been severely affected by habitat loss (due to the lack of water) and intense herbivory (GarciaMurillo & Fern´ andez-Zamudio, 2015). For certain other species, the disappearance of their populations from Do˜ nana would considerably reduce their EOO. Such is the case for Potamogeton natans (MartínBlanco & Carrasco, 2005; Melendo et al., 2003), P. lucens (Martín-Blanco and Carrasco, 2005), and Lemna trisulca (Cirujano et al., 2006; García–Murillo et al., 1990), for which their nearest populations are hundreds of kilometres to the North. Our data about the presence of these Potamogeton species do not still detect a reduction of their range in the area, but it is alarming the transformation observed from dense cover of reproductive individuals in the 2000 ′ s to the presence of a scarce number of non-reproductive individuals still persisting at the same localities. They are submerged perennial plants that still persist in some zacallones, P. natans from the northern area, and P. lucens from the central and southern area. Their main threat is the temporalization or even P. Garcia-Murillo et al. Journal for Nature Conservation 84 (2025) 126814 9