Coastal ecosystems produce and sequester significant amounts of carbon (‘blue carbon’), which has been well documented in humid and semi-humid regions of temperate and tropical climates but less so in arid regions where mangroves, marshes, and seagrasses exist near the limit of their tolerance for extreme temperature and salinity. To better understand these unique systems, we measured whole-ecosystem carbon stocks (above- and belowground biomass and soil) in 58 sites across the United Arab Emirates in natural and planted mangroves, salt marshes, seagrass beds, microbial mats, and coastal sabkha (inter- and supratidal unvegetated salt flats) and were funded by the Abu Dhabi Global Environmental Data Initiative.

Plant carbon for mature mangrove trees were measured following methodology from Kauffman and Donato (2012). Due to their smaller structure, all trees with stems > 3 cm diameter at breast height (DBH; 1.3 m in height) within the 7 m plot were measured instead of restricting measurement to trees with DBH > 10 cm. Standing dead trees and downed woody debris were found at some Gulf of Oman sites and pools were quantified appropriately (Kauffman and Donato 2012). In the planted mangrove sites, five 2-m radius plots were established at 10-m intervals along a 40-m transect. When stands contained individuals < 1.3 m tall, we measured the crown diameter and main stem diameter at 30-50 cm in height. In the 3, 5, and 10 year old planted mangrove sites in Abu al Abyad, trees were planted in an evenly-spaced grid; therefore, the plant density was calculated by measuring the average plant spacing and main stem and crown diameter of 50-75 trees.

Salt marsh transect length and plot spacing were the same as with the mature mangroves, although the plot radius ranged from 1-4 m depending on plant density. The height and elliptical crown area (perpendicular crown widths centered on the canopy) were measured on every plant rooted in each plot.

Biomass for A. marina and A. macrostachyum were calculated using allometric equations. Global tree carbon percentages of 48 and 39% for above- and belowground biomass, respectively, were applied (Kauffman and Donato 2012). To examine differences in average annual carbon sequestered in planted mangrove trees, we divided total biomass for each stand by the number of years since plantation establishment. We developed allometric equations for A. macrostachyum aboveground biomass, as none previously had been published. Twenty-four plants were collected from three sites and measured for crown dimensions and succulent woody tissue fresh weight. A subsample of each tissue type from every plant was weighed fresh and dry (constant weight at 50°C) to calculate a wet-to-dry mass conversion factor for the entire plant. Simple linear regression of natural log-transformed oven-dry biomass and plant volume (height x elliptical crown area) were calculated, producing two different relationships depending on plant size. Tissue samples were analyzed for percent carbon and nitrogen; carbon content of woody (n = 12) and succulent (n = 10) tissue averaged 45.5 and 34.0%, respectively (SE ± 0.7% for both). We used the proportion of 52% woody tissue reported in Neves et al. (2010) to calculate carbon content for aboveground biomass as 40.3 ± 1.4%C. The reported root to shoot ratio of A. macrostachyum measured in Portugal (Neves et al. 2010) is likely a conservative estimate of root biomass, as soil cores were taken to a depth of 15 cm instead of 20-30 cm as in other studies (reviewed in Curcó et al. 2002).

Soil Carbon Pools

At mangrove and salt marsh plots, undisturbed soil samples were collected following methodology from Kauffman and Donato (2012) using a 1 m-long gouge auger with an open-face, semi-cylindrical chamber of 5.1 cm radius. Soils were cored to 3 m or until coarse marine sands or coral rubble representing the parent material was encountered. The soil core was divided into depth intervals of 0-15, 15-30, 30-50, 50-100, and >100 cm, or until refusal. Subsamples collected from center of each interval were analyzed for bulk density (dry mass per unit volume) and carbon concentration (organic and inorganic). If encountered, unique soil layers were sampled separately. The same soil sampling methodology was used within microbial mats and coastal sabkha; the number of plots sampled per transect varied from 3-6 plots spaced at 20 m intervals along a transect. We determined soil carbon stocks following methods outlined in Fourqurean et al. (2012a), which are designed to account for soils containing carbonates.

A variety of soil biogeochemical measurements, soil respiration, elevation, and tidal data were collected in selected mangroves, salt marsh, microbial mats, and sabkha sampled in 2013; mangroves sampled in 2014 were not sampled. Redox potential (Eh) was estimated from five replicate platinum-tipped electrodes inserted 10 cm into the soil for a period of 1-20 minutes and corrected for the potential of the calomel reference electrode by adding 244 mV (Megonigal and Rabenhorst 2013). At sites that had a shallow water table, soil pore-water was collected from corer boreholes at 5-10 cm below the surface and analyzed by the standard methods described in Keller et al. (2009). Salinity was determined either by refractometer (values < 160) or calculated from [Cl-] (values > 160). pH was measured in the field with a portable electrode. Porewater dissolved methane (CH4) was measured by headspace equilibration following Keller et al. (2009) and stored in evacuated Exetainer vials until analysis by Varian 450-GC gas chromatography. Porewater [SO42-] and [Cl-] were determined by Dionex ICS-2000 ion chromatography on filter-sterilized (0.22 µm), HCl-acidified samples, and the sulfate depletion ratio was calculated per Keller et al. (2009). Gas and filtered porewater samples were stored for four weeks before analysis, which is well within the capacity of the storage methods used. We quantified instantaneous CO2 gas exchange rates as a simple index of activity to assist with comparisons across the ecosystems in this study. Soil surface CO2 emissions were measured with a LICOR 6400 soil respiration analyzer, with a range of 2-18 measurements per site, depending on time spent at each site and the rate of soil respiration; more measurements were possible at higher respiration rates.