Role of vegetated coastal ecosystems as nitrogen and phosphorous filters and sinks in the coasts of Saudi Arabia

Vegetated coastal ecosystems along the Red Sea and Arabian Gulf coasts of Saudi Arabia thrive in an extremely arid and oligotrophic environment, with high seawater temperatures and salinity. Mangrove, seagrass and saltmarsh ecosystems have been shown to act as efficient sinks of sediment organic carbon, earning these vegetated ecosystems the moniker ‘blue carbon’ ecosystems. However, their role as nitrogen and phosphorus (N and P) sinks remains poorly understood. In this study, we examine the capacity of blue carbon ecosystems to trap and store nitrogen and phosphorous in their sediments in the central Red Sea and Arabian Gulf. We estimated the N and P stocks (in 0.2 m thick-sediments) and accumulation rates (for the last century based on 210Pb and for the last millennia based on 14C) in mangrove, seagrass and saltmarsh sediments from eight locations along the coast of Saudi Arabia (81 cores in total). The N and P stocks contained in the top 20 cm sediments ranged from 61 g N m−2 in Red Sea seagrass to 265 g N m−2 in the Gulf saltmarshes and from 70 g P m−2 in Red Sea seagrass meadows and mangroves to 58 g P m−2 in the Gulf saltmarshes. The short-term N and P accumulation rates ranged from 0.09 mg N cm−2 yr−1 in Red Sea seagrass to 0.38 mg N cm−2 yr−1 in Gulf mangrove, and from 0.027 mg P cm−2 yr−1 in the Gulf seagrass to 0.092 mg P cm−2 yr−1 in Red Sea mangroves. Short-term N and P accumulation rates were up to 10-fold higher than long-term accumulation rates, highlighting increasing sequestration of N and P over the past century, likely due to anthropogenic activities such as coastal development and wastewater inputs.


Introduction
Vegetated coastal habitats offer myriad vital services that support coastal communities. Mangroves, seagrass meadows and saltmarshes can protect coasts from storm surges and rising sea levels (Duarte et al 2013) and provide essential habitats for fisheries and wildlife (Mumby et al 2004, Aburto-Oropeza et al 2008. They are also very effective at trapping and storing sediment, organic matter, anthropogenic pollutants and nutrients from surrounding waters (Jordan et al 2003, Alongi and McKinnon 2005, Rabaoui et al 2019. Special attention has been paid in recent years to the capacity of these vegetated coastal habitats to sequester carbon in their sediments efficiently. The high productivity and sedimentation rates, and their ability to vertically accrete sediment combined with the low oxygen availability, means these habitats can store carbon for millennia in their sediments (Donato et al 2011, Mcleod et al 2011, Duarte et al 2013, Atwood et al 2017. The capacity of mangroves, seagrasses and saltmarshes, collectively known as 'blue carbon' habitats, to trap and bury organic carbon (C org ) may also make them effective at removing excess nutrients from the surrounding environment. Nutrients are stoichiometrically linked to carbon in organic matter sequestered in sediments (Redfield ratio). Nutrient densities (mg m −3 ) and stocks (mg m −2 over a certain sediment depth) of nitrogen (N) and phosphorus (P) have been already assessed in seagrass (e.g. . Data on accumulation rates (mg N or P cm −2 yr −1 ) in those ecosystems, however, remain scarce (e.g. Breithaupt et al 2014, Noll et al 2019 as the measurement of nutrient densities and overall sediment accretion rates (cm yr −1 , or mass accretion rates in g cm yr −1 ) is required. The most traditional way to obtain sediment accretion rates is through radiometric dating of the sediment, by measuring the activity of 137 Cs for the most recent decades, 210 Pb for the last 100 years, and 14 C for centennial-to millennial-scale (Breithaupt et al 2018, Arias-Ortiz et al 2018a, 2018b. Here, we assess N and P stocks and accumulation rates in the sediments of seagrass meadows and mangroves along the Saudi coast in the Red Sea, as well as in those of seagrasses, mangroves and saltmarshes along the Saudi coast in the Gulf. We measured the N and P stocks in 0.2 m sediment cores, and, calculated the short-and long-term accumulation rates using the 210 Pb and 14 C sediment accretion rates reported for the same cores by Cusack et al (2018) and Saderne et al (2018). We use the changes in accumulation rates to hypothesize on changes in ecosystem functioning in the coastal ecosystems over the past millennia to centuries.

Ecosystem descriptions
The Arabian Peninsula is an especially arid region of the globe, with most of the peninsula receiving very little rainfall on an annual basis (Almazroui et al 2012). This low precipitation results in minimal fluvial input of sediments anywhere along the coasts of the Red Sea or the Gulf coast. The Red Sea does not have any permanent rivers. The Arabian Gulf, on the other hand, receives water from the Shatt al Arab River in the north, which itself is a result of the confluence of the Tigris and Euphrates Rivers. The sediments of the Red Sea and Arabian Gulf originate mainly from carbonate deposits from seawater and deposited airborne sediment of terrestrial origin, which is reported to comprise up to one-third of Gulf sediments (Sugden 1963, Saderne et al 2018. Dust deposition from frequent dust storms originating in the Sahara and the Arabian Peninsula is believed to be one of the primary sources of nutrients to the Red Sea and the Gulf (Acosta et al 2013, Engelbrecht et al 2017).
The Red Sea is oligotrophic, with P and N sea surface concentrations estimated, in its central part, at 0.05-0.1 μmol kg −1 and 0.03-0.2 μmol kg −1 , respectively (Weikert 1987, Wafar et al 2016. Saderne et al (2018) measured surface water concentrations of P and N, in a seagrass (Khor Almasena'a) and mangrove area, from the same study zone, and reported annual mean (±SD) records of 0.8±0.4 μmol P kg −1 and 1.3±0.5 μmol N kg −1 , and 2.3±0.5 μmol P kg −1 and 0.9±0.2 μmol N kg −1 , respectively. The chemical oceanography of the Arabian Gulf has been largely understudied, with only five sampling expeditions since 1965 (although substantial gray literature exists) (Al-Yamani and Naqvi 2019). The most recent published data for the Saudi coast of the Arabian Gulf are from 1993 to 1994 (Hashimoto et al 1998). Concentrations of phosphates between 0.2 and 0.4 μM and of nitrates from <0.09 to 0.2 μM were reported (Brewer andDyrssen 1985, Hashimoto et al 1998). Annual mean (±SE) concentrations of nitrates and phosphates measured in 2011-2012 in seagrass beds of the Saudi coasts of the Gulf ranged between 0.4±0.14 μM and 0.3±0.03 μM, respectively (KFUPM 2011).
This study was conducted in the Central Red Sea and along the Gulf coast of Saudi Arabia (figure 1). The study area of the Red Sea comprised four sites encompassing about 80 km of the coastline: Al Taweelah Island (22°16′ N, 39°05′ E), in front of the traditional fisherman urbanization of Thuwal, whose coastline has been modified drastically over the past ten years; Khor Almesena'a lagoon (22°22′ N, 39°07′ E), which has seen the development of King Abdullah Economic City to its north and King Abdullah University of Science and Technology to its south in the past decade; Khor Al-Baqila lagoon, whose entire southern side was converted into a large petrochemical terminal in 1981; and Khor Al-Kharrar lagoon (22°57′ N, 38°51′ E), which is relatively undeveloped.
The study area in the Saudi Gulf coast spanned 400 km along the Saudi coast. The southernmost site, Uqair (25°43′ N, 50°13′ E), has undergone negligible development. The Ras Tanura/Safwa site (26°41′ N, 50°00′ E) is an embayment north of the city of Dammam. Now a natural park, this area have been subjected to significant urban and industrial development since 1950. The site of Abu Ali Island (27°17 N, 49°33′ E) is now also part of a natural park, but has seen development due to the petrochemical industry and extensive modification of the area's hydrography due to the construction of dikes. The northernmost site, Ras Safanya (27°58 N, 48°46′ E), is a site of extensive petrochemical and oil extraction activities on the Safaniya oil field, one of the largest offshore oil fields in the world.
The mangrove ecosystems sampled on both coastlines were very similar, formed by mono-specific stands of Avicennia marina of less than 2 m height (Anton et al submitted). Seagrass species found in the Red Sea include Halophila stipulacea, Thalassia hemprichii, Enhalus acoroides, Thalassodendrum ciliatum, and Halodule uninervis. In the Gulf, the dominant seagrass species are H. uninervis (most widely distributed), H. stipulacea and H. ovalis (less common). Seagrass meadows across the Gulf occur at 46% of the coastal areas and 30% of offshore sites of the Gulf (Erftemeijer and Shuail 2012), and are under threat due to local dredging activities, land reclamation and associated changes in hydrology, together with pressures derived from climate change and marine pollution (Almahasheer et al 2013, Almahasheer 2018. Finally, saltmarshes of the Gulf coast are dominated by salttolerant or halophytic perennial grasses such as Aleuropus lagopoides and Bienertia cycloptera, conforming an ecosystem known locally as Sabkahs.

Sediment sampling
The sediments were sampled using manual percussion and rotation of PVC pipes (Red Sea samples: internal diameter 60 mm. Gulf samples: 110 mm). A total of 27 and 29 cores were collected from seagrass and mangrove ecosystems in the Red Sea, respectively, while in the Gulf, a total of 12, 7, and 6 cores were collected in seagrass, mangrove and saltmarsh ecosystems, respectively. At each site, cores were sampled within 1-10 km 2 . The sediment cores in the present study are the same as those described in Cusack et al (2018) and Saderne et al (2018).
In the Red Sea, two types of cores were sampled; plain PVC pipes and PVC pipes with pre-drilled sampling ports (Howard et al 2014) of 3 cm diameter at 6 cm apart along the length of the core. The sediments of the port cores were sampled in the field using modified plastic syringes (i.e. cut open to allow subsampling the ports). The regular cores were opened lengthwise and cut into 1 cm thick slices in the laboratory. All samples were dried at 60°C until constant weight was reached to estimate dry bulk density (DBD), and subsamples were milled for nutrient analysis. In the Red Sea cores, N concentrations were measured in every slice (1 cm thick) down to a depth of 20 cm (compressed) and then every second cm down to the bottom of the core in mangroves, and every cm from top to bottom in the seagrass cores. In the port cores, N concentrations were measured in every sample, and P concentrations were measured in every port sample down to 27 cm. In the Red Sea cores, P concentrations were measured every 5 cm in the whole core down to 20 cm (compressed), while in the Arabian Gulf cores, N and P were measured every 5 cm down to approx. 20 cm (compressed). All data presented here are corrected for compression effects during coring, following the procedure in Howard et al (2014). The difference between the length of the core barrel inserted in the sediment and the length of the core retrieved was measured and the difference distributed uniformly throughout the length of the core.
N concentrations in each sediment sample were determined from 200 mg samples contained in a precombusted aluminum case using a FLASH 2000 CHNS Elemental Analyzer. P concentrations were determined by inductively coupled plasma-optical emission spectrometry (Varian Inc. model 720-ES). 200 mg of sediment were acid digested in 5 ml of concentrated HCl and a few drops of H 2 O 2 in a Digi PREP digestion system in three temperature steps: 30°C for 30 min, 50°C for 30 min and 75°C for 45 min. The digested samples were then cooled, diluted to 25 ml with Mili-Q water and analyzed for P concentrations. Standard reference materials from the US National Institute of Standards and Technology were employed to verify the accuracy of P measurements. For N concentrations, a standard reference material (BBOT) and calibration standard (Sulfanilamide) were used to ensure the accuracy of results.
The N and P contents were calculated in each slice by multiplying the sediment DBD (g cm −3 ) by the nutrient concentration. In each core, the nutrient stocks in g m −2 integrated over the top 20 cm was estimated by building a continuous depth profile of N and P content by linear interpolation. The nutrient accumulation rates (mg m −2 yr −1 ) were calculated by multiplying the nutrient content by the sediment accumulation rates (mass m −2 yr −1 ) derived from 14 C and 210 Pb age-depth models (where successfully measured) and published in Cusack et al (2018) and Saderne et al (2018) for the same cores used in this study. The changes in accumulation rates of N and P were calculated for different periods by averaging the burial rates across each habitat type for the periods provided by 210 Pb dating. Nutrient ratios are reported as atomic ratios (mol/mol). All standard errors presented include error propagation for the calculation of accumulation rates from elemental contents and sediment accretion rates, and the calculation of averages within sites, ecosystems and regions.
General linear models (GLMs) were used to test differences in the variables studied between regions, among habitat types, and among habitat types within regions. Differences in soil P and N accumulation rates among long-term periods (based on 14 C), and three periods over the last century (1850( -1930( , 1930( -1980( and Modern (1980; based on 210 Pb) for Red Sea mangrove habitats, and Arabian Gulf seagrass, mangrove and saltmarsh habitats were also tested with a GLM. Region (Red Sea versus Arabian Gulf), habitat type (saltmarsh, seagrass and mangrove) and periods (1850( -1930( , 1930( -1980( and Modern: 1980 were treated as fixed factors in all statistical models (probability distribution: normal; link function: identity).

Results
The average (±SE) N concentration in mangrove and seagrass sediments from the Red Sea were similar, at 0.03±0.09% DW and 0.03±0.02% DW, respectively. In the Gulf, N concentrations for mangrove, seagrass, and saltmarsh sediments were at least twice those of Red Sea sediments at 0.017±0.003%, 0.09±0.03% and 0.10±0.08%, respectively (table 1). The highest sediment N density was in Gulf saltmarsh sediments, specifically in Ras Tanura, Arabian Gulf (1.68±1.13 mg N cm −3 ) and the lowest N density was recorded in seagrass sediments of Al-Baqila, Red Sea (0.015±0.03 mg N cm −3 ). The average P concentrations were 0.043±0.007 and 0.03±0.02 %DW in Red Sea seagrass and mangrove sediments (table 1). They were 0.019±0.004, 0.076±0.015 and 0.022±0.003 %DW in the Gulf seagrass, mangrove and saltmarsh sediments (table 1). The highest sediment P density was measured in seagrass sediments of Al Taweelah Island, Red Sea (0.6±0.05 mg P cm −3 ), while the lowest was recorded in seagrass sediments of Ras Tanura, Arabian Gulf (0.19±0.02 mg P cm −3 ). The variability of N and P %DW with sediment depth for Red Sea and Arabian Gulf habitats are provided in the supplementary material. A marked increasing trend in N concentrations (%DW) towards surface sediments is visible across all Red Sea habitats (supp. figure 1, available online at stacks.iop.org/ERL/15/034058/mmedia) except for seagrass sediments in Al-Baqila, Red Sea (supp. figure 1). In contrast, P concentrations (% DW) do not appear to undergo any discernible trend between shallow and deeper sediments, except in Khor Al-Kharrar mangrove sediments (sup. figure 1). Measurements of N and P for Gulf sediments were performed to a maximum depth of 20 cm, and the variability of N and P %DW with sediment depth are shown in supp. figure 1. Concentrations in the top 20 cm do not show any obvious increasing trend toward shallower sediments.
Stocks of N were 61±32 and 99±47 mg N cm −2 in seagrass and mangrove ecosystems of the Red Sea. Stocks of N in each habitat were at least three-fold higher for the Arabian Gulf, with 223±110, 209±96 and 265±93 mg N cm −2 in seagrass, mangrove and saltmarsh sediments (figure 2(a); table 2). In contrast, stocks of P were more similar across the Red Sea and Gulf habitats with 70±57 and 70±15 mg P cm −2 in seagrass and mangrove sediments of the Red Sea, and 33±11, 47±16 and 58±23 mg P cm −2 in seagrass, mangrove and saltmarsh sediments of the Gulf (figure 2(a); table 2). The highest N stocks were found in Gulf saltmarsh sediments at 265±93 mg N cm −2 and the highest P stocks were recorded in Red Sea ecosystems at 70±57 and 70±15 mg P cm −2 in seagrass and mangroves, respectively, closely followed by Gulf saltmarshes (58±23 mg P cm −2 ; table 2).
Long-term sediment accretion rates, as determined by 14 C dating, varied from a minimum of 0.04±0.04 and 0.04±0.01 cm yr −1 in Arabian Gulf mangroves and saltmarshes to a maximum of 0.2±0.1 cm yr −1 in Arabian Gulf seagrass meadows. The elevated accretion rate in Gulf seagrasses resulted in a very high long-term N accumulation rate of 0.3±0.2 mg N cm −2 yr −1 (table 2; figure 2(b)). However, although sediment accretion rates in Gulf seagrass sediments were high, long-term accumulation rates of P remained low (0.036±0.027 mg P m −2 yr −1 ), comparable to that in seagrass sediments in the Red Sea (0.052±0.036 mg P cm −2 yr −1 ; figure 2(b)).
The short-term N accumulation rates in mangroves averaged 0.093±0.062 mg N cm −2 yr −1 in the Red Sea and 0.38±0.16 mg N cm −2 yr −1 in the Gulf (figure 2(b)), with remarkably high N accumulation rates in mangrove stands at Ras Tanura in the Gulf Table 1. The mean (± SE) sediment N and P density (mg cm −3 ) and % dry weight (DW) and atomic stoichiometric ratios for each site and habitat for the Central Red Sea and the Gulf. (table 2). The short-term P and N accumulation rates in Red Sea mangroves were found to be similar (0.092±0.073 mg P cm −2 yr −1 and 0.093±0.062 mg N cm −2 yr −1 ). In contrast, in the Gulf ecosystems the P accumulation rates were one order of magnitude lower than the N accumulation rates in (figure 2(c)).
Changes in N and P accumulation over specific periods are presented in figure 3 and supplementary table 1: 1850-1930 (pre-oil discovery), 1930-1980 (oil discovery and exploitation) and Modern (1980-present-day). Mean N accumulation rates across all habitats for the pre-oil discovery period (0.104 mg N cm −2 yr −1 ) were significantly lower than in modern times (0.147 mg N cm −2 yr −1 ; P<0.001; figure 3, supplementary table 1). N accumulation in Red Sea mangroves doubled between the pre-oil era and the present day (0.074-0.144 mg N cm −2 yr −1 ). However, N accumulation rates in the Gulf peaked in all three habitat types during 1930-1980, followed by a slight decrease in the modern era. The highest accumulation rates across all habitats were for Gulf mangroves, which also underwent the largest increase in accumulation rates from the pre-oil era (0.13 mg N cm −2 yr −1 ) to the post-oil era (0.24 mg N cm −2 yr −1 ). P accumulation rates were relatively flat across the three periods in both the Red Sea and Gulf, except for Gulf mangroves, which did show an increasing trend in P accumulation rates from the pre-oil era (0.03 mg P cm −2 yr −1 ) to the modern era (0.06 mg P cm −2 yr −1 ).
Nutrient ratios varied widely not only geographically but also within habitats. The contrast in N and P distribution is reflected in the nutrient atomic ratios in the sediments, with a cross-habitat average N: P of 1:1 in the Red Sea, whereas in the Arabian Gulf this ratio was 5:1 (table 1). Furthermore, C:N ratios in mangroves sediments of the Red Sea were more than twice those of Gulf (table 1). We measured N:P ratios in seagrass sediments in the Red Sea ranging five-fold from 0.3:1 to 1.4:1, and in Red Sea mangrove sediments ranging three-fold from 0.5:1 to 1.6:1. In the Arabian Gulf sediments, N was much more abundant relative to P, with ratios ranging from a minimum of 4:1 in mangrove to a maximum of 8:1 in seagrass sediments in Abu Ali Island (table 1).

Discussion
The N and P stocks were quite comparable within blue carbon ecosystems of the Red Sea and the Gulf ( figure 2(a)). On average, N stocks were at least twofold higher in the Gulf compared to the Red Sea, while P stocks were only a maximum of 1.2 times higher in Red Sea vegetated coastal habitats.
Owing to the negligible rainfall and minimal fluvial inputs of terrestrial material to the Red Sea and Gulf, desert dust deposition is believed to be one of the main sources of nutrients N, P and Fe to both marine environments (Acosta et al 2013). Engelbrecht et al (2017) reported that deposited desert dust for the same region of the Red Sea coastline as this study, contained on average 0.8% NO 3 − and 0.3% PO 4 3− . Assuming that 6 Mt and 5.5 Mt of mineral dust is deposited into the Red Sea and the Gulf every year, respectively (Hamza et al 2011, Jish Prakash et al 2015, which would be equivalent to ∼21 000 and ∼11 000 tonnes of N and P  Table 2. Sediment stocks of N and P (g m −2 ) to 0.2 m sediment depth, and short-term sediment accretion and N and P accumulation rates as determined by 210 Pb analysis, and long-term accumulation rates as determined by 14 C from Cusack et al (2018) and Saderne et al (2018). Values are mean±SE of each site. in both the Red Sea and the Gulf, we estimate that blue carbon habitats of Saudi Arabia could potentially accumulate between 0.3% and 0.8% of annually deposited nutrients by dust deposition. The ratios of N:P in our sediments, ranging from 0.3:1 to 8:1 (table 1), highlight that vegetated coastal habitats along the Saudi coast are severely N limited, based on the Redfield ratio, which states that molar ratios of N:P <16:1 are N limited (Redfield 1960, Downing 1997). This hypothesis is further reinforced if we consider the ratio of 30:1 for vegetated benthic ecosystems of Atkinson and Smith (1983). Red Sea sediments are more severely N limited relative to Gulf ecosystems (table 1). This low ratio in the benthic habitats of the Gulf contrasts with the ratio in open waters that approaches the Redfield ratio (Hashimoto et al 1998, Al-Yamani andNaqvi 2019). This ratio is, however, consistent with the low ratio found in the water column above seagrass beds of the Saudi coast of the Gulf, at 1.6±2.1 (mean±SD) (KFUPM 2011). A higher ratio has been reported in a seagrass and mangrove ecosystem of the central Red Sea with 8.1±9.6 and 4.9±5.6, respectively (mean±SD; calculated from Saderne et al 2019).
In contrast to P, N accumulation rates exhibit higher spatial and temporal variability, respectively. Short-term N accumulation rates (figure 2(c)) have increased between 1.3 and 15-fold relative to longterm accumulation rates ( figure 2(b)). Comparing modern-day N accumulation rates to those of the preoil discovery period (figure 3), there has been a significant rise in accumulation rates in Red Sea and Gulf mangroves, which is possibly a result of large-scale N mobilization from anthropogenic activities. In the case of Red Sea mangroves, N accumulation rates have increased steadily since the pre-oil discovery period until the modern-day (figure 3). Almahasheer et al (2016b), who also described sediment accretion rates and C org accumulation for the same locations of the Red Sea as this study (Almahasheer et al 2017), suggested that high values of δ 15 N in mangrove sediments of the Central Red Sea was indicative of sewage and fertilizer sources, thus indicating a possible local source of anthropogenic N. Unfortunately, the Red Sea and the Gulf suffer from a lack of data on seawater chemistry, spatial and temporal, that would allow to correlate the increase of nutrients in the sediments to a potential increase in the water column (Al-Yamani and Naqvi 2019). We note however, that the apparent reduction of N and P accumulation rates in deep sediments could also be due to remobilization of labile N and P from deep to surface sediments.
The capacity of vegetated coastal ecosystems of the Arabian Peninsula to sequester nutrients is significant. Whereas the area coverage of seagrass meadows in the Red Sea and the Gulf remains poorly constrained, mangrove coverage has been accurately mapped for the Red Sea (135 km 2 ; Almahasheer et al 2016a) and the Gulf (165 km 2 ; Almahasheer 2018). The extrapolation of mangrove N and P stocks measured in this study to the entire Red Sea and Gulf, suggest that mangrove sediments hold 13 365±6345 tons of N and 9450±2025 tons of P in the Red Sea, and 34 485±15 840 tons of N and 7755±2640 tons of P in the Gulf within the top 0.2 m of sediments. Based on our average short-term sediment accumulation rates of N and P (table 2), mangrove sequester 125±83 Mg yr −1 of N and 124±98 Mg yr −1 of P in the Red Sea, and 518±218 Mg yr −1 of N and 115±42 Mg yr −1 of P in the Gulf. Furthermore, at characteristic concentrations of about 2 mmol m −3 nitrate and 0.2 mmol m −3 phosphate in surface waters of the highly oligotrophic Red Sea, the average stock in one square meter of the top 0.2 m of mangrove sediments contain the equivalent in N and P of 13 700 m 3 and 102 000 m 3 of Red Sea surface waters, respectively. For Figure 3. Accumulation rates of P and N (g m −2 yr −1 ±SE) for long-term ( 14 C dating) and periods 1850-1930, 1930-1980and Modern (1980 from 210 Pb for Red Sea mangrove habitats, and Arabian Gulf seagrass, mangrove and saltmarsh habitats. The 210 Pb and 14 C sediment accretion rates used for calculations are from Cusack et al (2018) and Saderne et al (2018). seagrass sediments, this represents the equivalent of 8200 m 3 and 102 000 m 3 of Red Sea surface waters for N and P, respectively, which highlights the role of vegetated coastal ecosystems as massive sinks and reservoirs for nutrients in the oligotrophic Red Sea.
Published data on nutrient accumulation rates in blue carbon ecosystems are rather scarce. Breithaupt et al (2014) presented a short review of the available accumulation rates for mangroves, mostly obtained using the 210 Pb dating technique, and found a global mean/median of 1.25/0.89 mg N cm −2 yr −1 and 0.65/0.07 mg P cm −2 yr −1 . In comparison, the accumulation rates of N in the Red Sea and Gulf mangroves (table 2) range among the lowest, while the accumulation rates of P are within the global median (with the mean being strongly influenced by high accumulation rates reported in the Jiulongjiang Estuary in China). The low short-term N accumulation rates found in our study are mainly due to low N densities because sediment accumulation rates in our Red Sea and Gulf cores are within the global average (see discussion in Saderne et al 2018. For example, Breithaupt et al 2014 found N and P densities in mangrove sediments of 2.5±0.9 and 0.21±0.06 mg cm −3 in the Everglades (Florida, USA), which is higher than our N densities (0.33±0.26 and 1.04±0.22 mg N cm −3 in the Red Sea and Gulf), but comparable in terms of P densities 0.34±0.18 and 0.24±0.05 mg P cm −3 for the Red Sea and Gulf) (table 1).
Regarding seagrass, Eyre et al (2016) reported short-term N accumulation rates ranging from 0.82 to 0.13 mg N cm −2 yr −1 in three estuarine systems of New South Wales (Australia), which are similar to those reported here for the Gulf (0.22 mg N cm −2 yr −1 ; table 2). Similarly, in the coastal bays of the Delmarva Peninsula (Virginia, USA) Aoki et al (2019) found N accumulation rates of 0.35 mg N cm −2 yr −1 . We could not find other direct reports of N and P accumulation rates in seagrass meadows. However, we estimated N and P accumulation rates in Florida Bay (Cheng et al 2012) and Shark Bay (Western Australia, Arias-Ortiz et al 2018a, 2018b) based on separate reports of nutrients densities and 210 Pb sediment accretion rates (0.08 mg P cm −2 yr −1 and 0.025 mg P cm −2 yr −1 for Florida Bay and Shark Bay, respectively), which are comparable to the rates estimated for the Gulf (0.027 mg P cm −2 yr −1 ). For Shark Bay, we estimated an N accumulation rate of 0.29 mg N cm −2 yr −1 , which is comparable to the rate found in the Gulf.
To our knowledge, our study is the first report of N and P accumulation rates in sediments of arid saltmarshes (short-term: 0.195±0.107 and 0.045±0.014 mg cm −2 yr −1 respectively). Overall, the only study reporting such data was from Cape Fear River in North Carolina (USA), with mean (±SE) N and P accumulation rates based on 210 Pb of 0.49±0.22 mg N cm −2 yr −1 and 0.05±0.03 mg P cm −2 yr −1 (Noll et al 2019). For comparative reasons, we estimated short-term N and P accumulation rates separately from four different studies. In the tropical saltmarshes of Mexico and El Salvador, we calculated rates ranging from 0.08 mg N cm −2 yr −1 to 1.4 mg N cm −2 yr −1 , with an estimated mean of 0.46 mg N cm −2 yr −1 from the data in Ruiz-Fernández et al (2018). In a North Mediterranean saltmarsh (Marano lagoon, Italy, Adratic Sea). Covelli et al (2012) reported N accumulation rates ranging from 0.3 to 1 g N cm −2 yr −1 . Regarding P, short-term accumulation rates ranging from 4.5 to 6 mg P cm −2 yr −1 were reported for temperate saltmarshes of the Scheldt estuary (the Netherlands; Zwolsman et al 1993), i.e. 100 times higher than the rates in the Gulf. These high burial rates are, however, explained by sediment accretion rates 9-13 times higher than in our cores, rather than differences in nutrient concentrations. In comparison, Spencer et al (2003) found P accumulation rates of 0.12 mg P cm −2 yr −1 in a saltmarsh of the Medway river estuary (UK). Overall, although comparisons between such different types of saltmarshes are somewhat difficult, it seems that the rates found in the Gulf range among the lowest reported anywhere.

Conclusion
Vegetated coastal ecosystems such as mangroves, seagrass meadows and saltmarshes provide a range of essential services to coastal communities, such as protection against sea-level rise, fish nurseries, and trapping of carbon and sediment. However, their role as possible sinks for nutrients has received comparatively little attention, despite the massive mobilization of nutrients, especially N and P, into the environment on a massive scale over the past century. Here, we measured the below-ground sediment stocks of N and P at several locations in the central Red Sea and along the Gulf coast of Saudi Arabia to determine that these habitats are major sinks for N and P. Indeed, the accumulation rate of these nutrients has increased over time, quite possibly because of human activities leading to the major mobilization of these nutrients. Our results provide evidence for a hitherto understudied but valuable ecosystem service that is, on a local scale, as important as their capacity to sequester carbon, and provides further impetus for the conservation of these habitats.
the Coastal and Marine Resources core lab at KAUST for help with sampling. We thank I Schulz, N Geraldi, K Rowe, S Roth, M Ennasri, D Prabowo and I Mendia for help with laboratory analyses. We wish to thank the two anonymous reviewers, as well as the editor for their precious comments/suggestions that improved the manuscript.
Any data that support the findings of this study are included within the article.