Spatiotemporal Patterns in the Biomass of Drift Macroalgae in the Indian River Lagoon, Florida, United States

Drift macroalgae plays key roles in the ecology of many coastal systems, including the Indian River Lagoon. In the lagoon, changes in the biomass of drift macroalgae may have interacted with an unprecedented bloom of phytoplankton in 2011. Patterns in the biomass of drift macroalgae were identified using new and original analyses of data from several sampling programs collected between 1997 and 2019. All available data show a relatively low biomass of drift macroalgae in 2010–2012, and surveys of fixed transects and seining as part of a fisheries independent monitoring program also recorded low biomass in 2016. Low light availability and potentially stressful temperatures appeared to be the main influences as indicated by the results of incubations in tanks to determine environmental tolerances and data on ambient conditions. Decreased biomass of drift macroalgae had implications for cycling of nutrients because carbon, nitrogen, and phosphorus not stored in the tissues of drift macroalgae became available for uptake by other primary producers, including phytoplankton. The estimated 14–18% increases in concentrations of these elements in the IRL could have promoted longer and more intense phytoplankton blooms, which would have reduced light availability and increased stress on algae and seagrasses. An improved understanding of such feedback and the ecological roles played by drift macroalgae will support more effective management of nutrient loads and the system by accounting for cycling of nutrients among primary producers.

Drift macroalgae plays key roles in the ecology of many coastal systems, including the Indian River Lagoon. In the lagoon, changes in the biomass of drift macroalgae may have interacted with an unprecedented bloom of phytoplankton in 2011. Patterns in the biomass of drift macroalgae were identified using new and original analyses of data from several sampling programs collected between 1997 and 2019. All available data show a relatively low biomass of drift macroalgae in 2010-2012, and surveys of fixed transects and seining as part of a fisheries independent monitoring program also recorded low biomass in 2016. Low light availability and potentially stressful temperatures appeared to be the main influences as indicated by the results of incubations in tanks to determine environmental tolerances and data on ambient conditions. Decreased biomass of drift macroalgae had implications for cycling of nutrients because carbon, nitrogen, and phosphorus not stored in the tissues of drift macroalgae became available for uptake by other primary producers, including phytoplankton. The estimated 14-18% increases in concentrations of these elements in the IRL could have promoted longer and more intense phytoplankton blooms, which would have reduced light availability and increased stress on algae and seagrasses. An improved understanding of such feedback and the ecological roles played by drift macroalgae will support more effective management of nutrient loads and the system by accounting for cycling of nutrients among primary producers.

INTRODUCTION
Interactions among three key primary producers, phytoplankton, macroalgae, and rooted macrophytes, represent influential processes in many estuaries (Duarte, 1995;Kinney and Roman, 1998;Viaroli et al., 2008). In the Indian River Lagoon (IRL), drift macroalgae (DMA) can dominate the total biomass of submersed aquatic vegetation, and even during years when seagrass is abundant, DMA contribute a significant proportion of the total primary productivity (Dawes et al., 1974;Josselyn, 1977;Thompson, 1978;Virnstein and Carbonara, 1985;Jensen and Gibson, 1986).
Beyond being an important primary producer, DMA also provide habitat for small animals and a place to forage for predators (Stoner, 1980;Kulczycki et al., 1981;Virnstein and Howard, 1987;Holmquist, 1997).
As key primary producers in the IRL, DMA play an important role in cycling of carbon, nitrogen, and phosphorus. The ability of macroalgae to take up and store nutrients makes them successful when nutrients are limiting or supplied in pulses, which allows them to compete with phytoplankton for access to elements in the water column (Marshall and Orr, 1949;Hanisak, 1983). However, DMA are less robust and persistent than rooted macrophytes, so their death or lack of growth can add or leave carbon, nitrogen, and phosphorus that become available for uptake by fast-growing phytoplankton (Hanisak, 1983(Hanisak, , 1993Lavery and McComb, 1991;Gao et al., 2013). In fact, shifts from dominance by benthic primary producers to dominance by phytoplankton have been observed in multiple systems with negative impacts on seagrass assemblages and their associated fauna (Jensen and Gibson, 1986;Duarte, 1995;Burkholder et al., 2007;Duarte et al., 2010). Such a shift may have occurred in the IRL because an unprecedented sequence of intense and longlasting blooms of phytoplankton has afflicted the system since 2011 (Phlips et al., 2010(Phlips et al., , 2011(Phlips et al., , 2015. Although important in the IRL and elsewhere, DMA are difficult to sample effectively and efficiently. The fact that DMA are moved by currents complicates the choice of locations to sample, and details regarding their movements in the IRL are poorly known (Kulczycki et al., 1981;Virnstein and Carbonara, 1985;Jensen and Gibson, 1986;Hanisak, 2021). Additionally, DMA occur across a wider range of depths than rooted macrophytes because they have lower requirements for light (Biber et al., 2004;Hily et al., 2004;McGlathery et al., 2007), so the need to survey more area and deeper water makes effective sampling more challenging. Drift macroalgae also are relatively ephemeral, and changes in their abundance within and between years, including 'boom and bust cycles' driven by environmental conditions make the choice of temporal intensity for sampling important (Virnstein and Carbonara, 1985;McGlathery et al., 2007).
In spite of these challenges, the objectives of this paper are to identify the presence, causes, and consequences of differences and changes in the biomass of DMA across space and through time in the IRL. Substantial and consistent patterns in the biomass of DMA are elucidated from data collected at multiple spatial and temporal scales, and those patterns are related to environmental drivers that reflect physiological tolerances identified in incubations. The implications of spatiotemporal variation in biomass of DMA are translated into effects on cycling of elements in the IRL via data on the carbon, nitrogen, and phosphorus content of several species of DMA.

Study Area
The IRL is a shallow, bar-built estuary along the east coast of Florida comprising three, interconnected sublagoons: Mosquito Lagoon, Banana River Lagoon, and Indian River Lagoon (Figure 1). The average water depth in the sublagoons is 1.9 m, and most of the system is microtidal (semidiurnal lunar tidal amplitude range 0.2-18.1 cm), with more substantial changes in water level driven by wind and the seasonal rise and fall of the coastal ocean (Smith, 1987;Pitts, 1989;Steward et al., 2005). Significant exchange with the coastal ocean is limited to five inlets located primarily in the southern portion of the IRL. This configuration contributes to residence times of over a year in the portions of the system that are far from inlets (Steward et al., 2005).
For this paper, the IRL was subdivided into reaches (Figure 1). The reaches were delineated by evaluating similarities in time series of water quality parameters (Lasi et al., this volume). The first six reaches were within the St. Johns River Water Management District. The remainder, reaches 7, 8, and 9, were within the boundary of the South Florida Water Management District.

Biomass of Drift Macroalgae
Data generated by visual surveys of fixed transects, seines of two sizes hauled as part of fisheries independent monitoring, and hydroacoustic surveys were analyzed independently because samples were taken at different frequencies, over different areal extents, and across depths that may or may not overlap (Supplementary Table 1). Given these variations, there was no attempt to intercalibrate the data, and interpretations focused on changes or differences that were consistent across methods.
A focus on consistent and substantial changes also resulted from a lack of details on the taxonomic composition of DMA. In surveys of transects, which focused on seagrass, and seining, which focused on fish and macroinvertebrates, DMA was considered bycatch; therefore, the biomasses of individual species typically were not measured. Furthermore, hydroacoustic surveys could not differentiate species. Nevertheless, widespread, consistent, and statistically significant spatiotemporal patterns provide valuable insights into the roles of DMA.

Fixed Transects
Fixed transects were surveyed at least twice a year (summer and winter) approximating times of annual maximum and minimum abundance of seagrasses (Figure 1; Virnstein and Morris, 1996). The location of each transect was marked with poles, and the path to be surveyed was delineated by a graduated line extending perpendicularly from the shoreline out to the deep end of the seagrass canopy. In summary, transects extended for 15-1,900 m across depths to 1.8 m.
At pre-designated points along transects, a suite of standardized, non-destructive measurements was taken within a 1-m 2 quadrat divided by strings into 100 cells, each 10 cm by 10 cm. The relevant measurement for DMA was a visual estimate of the biomass of all species expressed as an index, which was converted to biomass using empirically derived coefficients, i.e., 0 = 0 g dry weight (DW) m −2 , 1 = 0.35 g DW m −2 , 2 = 1.98 g DW m −2 , 3 = 21.37 g DW m −2 , 4 = 47.74 g DW m −2 , 5 = 141.94 g DW m −2 (Morris et al., 2001). For this study, we used data collected between 1998 and 2020.

Fisheries Independent Monitoring
Larger and in some cases deeper areas were sampled as part of long-term fisheries independent monitoring (FIM) conducted by the Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute. The program sampled sites selected without replacement using a stratified random sampling design (FWC-FWRI, 2016), and we used data from sites sampled between either 1997 or 1998 and 2019.
The FIM program employed multiple types of gear, and we used data from two seines, a 21.3-m × 1.8-m center-bag haul seine with 3.2-mm nylon mesh and a 183-m × 2.5-m center-bag haul seine with 38-mm stretched nylon mesh. Seines were deployed along estuarine shorelines to sample areas with emergent vegetation, fringing mangroves, seawalls, and beaches or on flats at least 5 m from the shoreline. The 21.3-m seine was pulled for 9.1 m across a 15.5-m wide strip of bottom, resulting in a sampling area of approximately 140 m 2 . Using a boat, the 183-m seine was set in a rectangular shape adjacent to the shoreline with a maximum depth of 2.5 m at the bag before being retrieved by hand. The area sampled by the net (approximately 40 m × 103 m = 4,120 m 2 ) was standardized by marking 40 m from each end to designate the corners of the rectangular set.
Bycatch of DMA was quantified to 0.1 gallons in the field, with reliable data being quantities of all species due to the challenges associated with identifying species visually. To convert these data to wet weights and dry weights, samples were collected during May and June 2012. Several species of DMA in quantities up to 5 gallons were collected at several sites, transferred to mesh bags, and spun to remove excess water. Mesh bags containing DMA were weighed to yield wet weights per gallon. The algae were dried at 80 • C and then weighed to yield dry weights per gallon. A conversion factor was generated by fitting a linear regression, forced through zero. This conversion factor and the area sampled by the seine were combined to convert gallons of DMA to g DW m −2 .

Hydroacoustic Surveys
To quantify the abundance and distribution of DMA found in deeper water, large-scale acoustic surveys were conducted between April and June in 2008, 2012, 2014 While the surveys covered up to 288 km 2 , we focused analysis of spatiotemporal variation on reaches 2, 3, and 4, which were completed in all 5 years.
Hydroacoustic data were acquired with a BioSonics DT-X echosounder and two multiplexed, single-beam digital transducers with full beamwidths of 10 • (38 kHz) and 6.4 • (418 kHz), operated at 5-Hz and 0.4 ms pulse duration (Riegl and Bushkirk, 2016;Foster et al., 2018). Surveys focused on water deeper than approximately 1.3 m. The east-west survey lines were spaced a minimum of 200 m and maximum of 400 m apart. Data were post-processed using BioSonics Visual Bottom Typer (VBT; Foster et al., 2018). To estimate percent cover of DMA, samples of algae and video were collected along the line of the hydroacoustic surveys, and wet weights were recorded. Based on these samples, the acoustic signals were grouped into a supervised training catalog for bare substrate, sparse algae, dense algae, and other vegetation. Using the catalog and the estimated conversion factor, counts of the acoustic signals in each class were used to estimate wet biomass, and the conversion factor calculated for samples from seining was used to estimate dry biomass (Riegl and Bushkirk, 2016;Foster et al., 2018).

Analysis of Data
To evaluate differences in biomass, dry weights of macroalgae generated by each of the methods were used in separate permutation analyses of variance (PERMANOVAs, Anderson et al., 2008). All the models included reach and years as fixed factors, and models applied to data from transects and seines included seasons as a random factor nested in years. Seasons were defined based on the results of agglomerative hierarchical clustering of water quality data, with a warm wet season running from May to October and a cool dry season running from November in 1 year through April of the following year (Lasi et al., this volume). The PERMANOVAs were based on Bray-Curtis distance measures calculated with a dummy variable of 1 to avoid undefined results.

Physiological Tolerances
To evaluate the effects of stress caused by extreme salinities, extreme temperatures, low light levels, and combinations of stressors, a series of incubations was conducted under controlled conditions in tanks holding 70 L of water (Hanisak, 2016). Incubations were conducted in batch cultures using approximately 80 g wet weight of Gracilaria tikvahiae and Crassiphycus secundus (formerly Hydropuntia secunda), which are two of the most widespread and abundant rhodophytes in the IRL (Hanisak, 2016(Hanisak, , 2021. Data were collected weekly over 35 days after 7 days of acclimation, and quantities of macroalgae were adjusted to maintain ∼80 g wet weight. The potential effects of macronutrient and trace metal limitation were obviated by midday additions of 39.0 µM dissolved inorganic nitrogen (i.e., 26 µM ammonium and 13 µM nitrate), 0.94 µM orthophosphate, and trace metals at concentrations found in F/2 medium. These additions achieved a nitrogen:phosphorus ratio of ∼41, which paralleled ratios measured in the IRL (Lasi unpublished data), and the ∼38 mg of nitrogen and ∼2 mg of phosphorus was sufficient to support growth of G. tikvahiae in past cultures (Hanisak, 1990). Salinities were maintained within 1 psu of the target by diluting water from a saltwater well. Temperatures were controlled within 1 • C of the targets by chillers and heaters. Light was provided by 100W SOL 1 -LED Grow Lights from 06:00 to 18:00.
Two types of incubations were conducted: single-factor range finding incubations for salinity, temperature, and light availability and incubations manipulating combinations of factors. The factors that were not manipulated were held at 25 psu, 25 • C, or 250 µmol m −2 s −1 , as appropriate. Incubations with varying salinity included 12, 18, 25, 35, 45, and 50 psu. Temperatures of 7, 10, 17, 25, and 33 • C were tested, and light levels of 0, 10, 30, 50, 100, 250, and 400 µmol m −2 s −1 were examined. Algae were harvested each week, and changes in wet weight were recorded. If incremental growth occurred, it was removed so the initial stocking density was returned to the tank. If biomass was lost, the decline was recorded. Wet weights were converted to dry weights using an empirical relationship derived from samples taken at the beginning and end of the incubations. In addition, samples taken each week were rinsed to remove excess salt, dried at 80 • C, ground to a fine powder with a mortar and pestle, and shipped to the University of Maryland's Nutrient Analytical Services Laboratory for analysis of carbon, nitrogen, and phosphorus content (University of Maryland, 2022) 1 . The resulting percentages were used to identify any data that were affected by nutrient limitation.
Data on accumulation of biomass (growth) were extracted from seven sets of incubations and analyzed with PERMANOVAs based on Bray-Curtis distance measures with the addition of a dummy variable set at 1. Where necessary, adding the minimum to all values eliminated negative numbers. The design accounted for the repeated measures nature of the incubations. Species and the appropriate stressor or combination of stressors were fixed factors. Tank was a random factor nested in the interaction of species and treatment, and week was a random factor nested in the interaction of species, treatment, and tank. To examine the most extensive set of treatments, some data were used in more than one analysis. 1 http://umces.edu/nutrient-analytical-services-laboratory

Timing and Extent of Suitable Conditions
Based on the results from the evaluation of stressors, the timing and extent of suitable conditions for growth of DMA were predicted. The data to characterize stressors were drawn from ongoing sampling of water quality in the lagoon (Lasi et al., this volume) and existing bathymetry.
Salinities, water temperatures, and photosynthetically active radiation (PAR) were measured monthly at 30 fixed stations in reaches 1 through 6 (Figure 1). Measurements were taken with multiparameter instruments that were calibrated and verified with standards and procedures prescribed by state guidelines (Florida Department of Environmental Protection, 2017). Light availability was characterized by coefficients of attenuation for PAR (K d ) as determined by applying Beer's Law (Kirk, 1983) to readings recorded simultaneously by three spherical quantum sensors. Sensors were held 0.2 and 0.5 m below the water's surface, and a third sensor was held 0.3 m from the bottom in water ≤ 1.8 m deep or 1.5 m below the surface in deeper water (Morris et al., 2001). The resulting K d values were combined with data on solar radiation from the Florida Automated Weather Network (FAWN 2 ) (University of Florida, Institute of Food and Agricultural Sciences, 2022) to calculate the depth that received light above the threshold causing stress. Data from FAWN were converted from Watts m −2 to µmol m −2 s −1 and daily means were calculated from daytime data for 1998 through 2019. The daily means were converted to irradiance (I 0 ) immediately below the surface of the water by reducing them by 6.6% to account for surface reflectance (Kirk, 1983), and the resulting values were combined with a threshold from the incubations (I z ) to determine the depth (z) receiving sufficient light using: The depths with sufficient light were translated to areal extent of suitable habitat for DMA by combining them with bathymetric data collected by Coastal Planning and Engineering (1997). Depth soundings were taken throughout the IRL at 15.2-m intervals along east-west transects that were separated by 150-300 m. This survey resulted in over 230,000 depth measurements that were referenced to North American Vertical Datum of 1988 (NAVD88) and adjusted to depth below mean water level (MWL). From these data, isobaths were generated in 0.1-m increments for reaches 1-6.
Stressful periods in each reach were identified by examining months with salinities, temperatures, or light availability that caused stress during the incubations. In addition, estimates of light penetration combined with bathymetry identified the extent of the lagoon with a suitable light regime.

Carbon, Nitrogen, and Phosphorus Stored in Drift Algae
The carbon, nitrogen, and phosphorus contents of DMA were determined from samples of common rhodophytes collected bimonthly from six sites in the Indian River and Banana River lagoons (reaches 2, 3, and 4) between July 2014 and August 2015. Three replicate samples of the two most abundant species were collected from beyond the deep edge of the seagrass beds, typically in 1.5-2.0 m of water. The samples were transported on ice to the laboratory where epiphytes and debris were removed before the algae were rinsed to remove excess salt and dried at 80 • C. Dried samples were ground to fine powder and analyzed for carbon, nitrogen, and phosphorus content at University of Maryland's Nutrient Analytical Services Laboratory (see Text Footnote 1).
The resulting data on carbon, nitrogen, and phosphorus content were used to gain insights into changes in the quantities of these elements stored in DMA. Median elemental contents were multiplied by DMA biomass per square meter from transects and surveys with seines to evaluate changes in cycling of key elements.

Patterns in Biomass of Drift Macroalgae
The biomass of the DMA assemblage varied significantly in space and through time according to analyses of data from surveys of transects, deployment of two different seines, and hydroacoustic surveys ( Table 1). Biomasses from the transects and seines varied among combinations of reach and season within year, and biomasses detected with hydroacoustics varied significantly among years.
Higher values for the biomass of DMA observed during surveys of transects or as bycatch from seining did not reveal consistent and ecologically meaningful patterns across reaches or among years (Figures 2A-D). However, data from surveys of transects indicated that 70% of the biomasses above the 90th percentile were recorded in the warmer, wetter months of summer, and monthly data from seining indicated that the biomass of drift algae tended to be higher from March through July (means ± standard error for 21.3-m seine = 1.7 ± 0.2 g m −2 and 183-m seine = 2.6 ± 0.2 g m −2 ) than in August-February (means ± standard error for 21.3-m seine = 1.0 ± 0.1 g m −2 and 183-m seine = 1.2 ± 0.1 g m −2 ).
In contrast, relatively low biomass was recorded by multiple methods during two periods. Low biomass was documented consistently along the transects and in the seines from late 2010 through 2011-2012 in reaches 1-7, but the same pattern was not obvious in reaches 8 and 9 where the abundance of drift algae was lower (Figures 2A,B). Additionally, biomass detected by hydroacoustic surveys was lowest in 2012 ( Figure 2E). Thus, multiple lines of evidence point to 2010-2012 as one period with little drift algae. In addition, drift algal biomass along transects and in the seines was consistently low in reaches 1-6 in 2016 (Figures 2A,C,D).

Physiological Tolerances
Evidence suggested that macroalgae were not limited by carbon, nitrogen, or phosphorus during the 35-day incubations. Cumulative growth of macroalgae over 35 days represented less than 2% of the 80 g incubated, so growth rates were unlikely to exhaust the nutrients, which were replenished daily. In addition, elemental contents from all incubations were consistent, which indicated that the algae received sufficient carbon, nitrogen, and phosphorus throughout the incubations. The mean content ± standard error was 28.169% ± 0.075% for carbon, 3.504% ± 0.026% for nitrogen, and 0.389% ± 0.004% for phosphorus.
Incubations with salinities that spanned 12-50 psu did not identify significant effects on cumulative growth or elemental contents ( Table 2). On average for each of the salinities, G. tikvahiae accumulated 5.1 ± 1.4 g to 10.2 ± 1.2 g of biomass over 35 days, and C. secundus accumulated 4.6 ± 0.8 g to 9.2 ± 1.4 g of biomass. Given these results, the effect of salinity was not explored further.
The two species had differing responses to the varying temperatures used in a range finding incubations ( Table 2). At all temperatures, G. tikvahiae accumulated more biomass over 35 days (Figure 3A). Both G. tikvahiae and C. secundus accumulated slightly more biomass at 25 • C, although growth at 33 • C was similar ( Figure 3A). On average, C. secundus lost biomass at 7 • C and only gained 0.8 ± 0.1 g at 10 • C.
Light availability characterized by PAR affected the growth of G. tikvahiae and C. secundus similarly as shown by the lack of a significant interaction between species and amount of PAR ( Table 2). Unlike C. secundus, G. tikvahiae did not survive in 0 µmol m −2 s −1 . Both species consistently accumulated biomass over 35 days only when provided with 250 or 400 µmol m −2 s −1 of PAR ( Figure 3B). On average, G. tikvahiae accumulated 2.5 ± 1.2 g of biomass, and C. secundus accumulated 3.7 ± 0.7 g of biomass. Given the greater and less variable accumulation of biomass, C. secundus appeared to cope with low light better than G. tikvahiae.
Two sets of results combined differing temperatures and differing amounts of light, with some data for G. tikvahiae used in two PERMANOVAs ( Table 2). In the first set of results, G. tikvahiae and C. secundus were incubated at 7, 10, 17, 25, and 33 • C under 10, 50, and 100 µmol m −2 s −1 of PAR, and in the second set, only G. tikvahiae was available to be incubated at 7, 10, 17, 25, and 33 • C under 10, 50, 100, and 250 µmol m −2 s −1 of PAR. The two species exhibited differing responses to combinations of temperature and PAR ( Table 2). For both species, there was loss or accumulation of <2 g of biomass over 35 days at 10 µmol m −2 s −1 of PAR (Figures 3C-E). At 50 µmol m −2 s −1 of PAR over 35 days, G. tikvahiae gained 2-3 g of biomass at all temperatures except 25 • C, and C. secundus gained about 2 g of biomass at 17 and 33 • C (Figures 3C-E). At 100 µmol m −2 s −1 of PAR over 35 days, G. tikvahiae gained 2-6 g of biomass, with the maximum gain at 25 • C (Figures 3C,E). At 100 µmol m −2 s −1 of PAR, C. secundus lost biomass at 7 and 10 • C and gained approximately 3 g over 35 days at 17, 25, and 33 • C (Figure 3D). Only G. tikvahiae was subjected to differing temperatures at 250 µmol m −2 s −1 , and it gained a maximum of 9.1 ± 0.7 g of biomass over 35 days at 25 • C, with gains decreasing to 3.4 ± 0.8 g of biomass at 7 • C and 7.5 ± 1.2 g of biomass at 33 • C ( Figure 3E). Both species tended to lose biomass when subjected to less than 50 µmol m −2 s −1 of PAR although they survived the 35-day incubations. Loss of biomass was more common at temperatures of 7 or 10 • C. Overall, the results indicated that light availability was a dominant influence on growth, with an effect of temperature appearing more strongly when the algae received over 100 µmol m −2 s −1 of PAR. In those conditions, optimum temperatures appeared to be 25 • C for G. tikvahiae and 17-33 • C for C. secundus.

Timing and Extent of Suitable Conditions
Salinity did not stress either G. tikvahiae or C. secundus, but incubations in the laboratory indicated that low temperatures (below 17 • C), high temperatures (above 25-33 • C), and reduced light availability (less than 100 µmol m −2 s −1 ) had the potential to reduce growth or cause mortality. Stressful periods for each reach were identified by coding monthly mean K d values higher than the 90th percentile of all values (>1.6 m −1 ), monthly mean temperatures above the 90th percentile (>30.3 • C), and monthly mean temperatures below the 10th percentile (<17.8 • C) as zero and all other values as one to produce a heatmap (Figure 4). Mean temperatures exceeded the 90th percentile in summers of all years, and temperatures dropped below the 10th percentile in multiple winters. The period with the most spatially and temporally extensive distribution of all stressors was 2009-2011, which corresponded with multiple records of low abundance for DMA (Figure 2). The incubations to identify tolerances highlighted PAR below 100 µmol m −2 s −1 as a key influence on the growth of DMA; therefore, depths receiving ≥100 µmol m −2 s −1 were associated with 0.1-m isobaths to identify the extent of suitable habitat in hectares (Figure 5 and Supplementary Figure 1). The extent of suitable habitat varied through time for each reach, and when compared to the background extent from 1998 to 2009, reaches 2-6 had more substantial losses of suitable habitat in 2010-2011 and in 2016-2018, which would translate to changes in  Table 1 for relevant statistics.
the amounts of carbon, nitrogen, and phosphorus stored in DMA (Figure 6 and Supplementary Table 2). These results were consistent with other evaluations of the abundance of DMA.

Carbon, Nitrogen, and Phosphorus Stored in Drift Algae
Five species of DMA were collected from the field, Acanthophora spicifera, Agardhiella subulata, G. tikvahiae, C. secundus, and Hypnea spinella. A series of PERMANOVAs using elemental compositions of the three species collected multiple times, A. subulata, G. tikvahiae, and C. secundus, identified significant variation among the combination of species and sampling event for mean percent carbon (p 10,212 = 0.001), nitrogen (p 10,212 = 0.001), and phosphorus (p 10,212 = 0.002). Mean carbon content ±SE varied from 16.6 ± 0.6% for A. subulata in July 2014 to 33.9 ± 0.4% for C. secundus in April 2015, mean nitrogen content ± SE varied from 1.5 ± 0.1% for A. subulata in August 2015 to 3.6 ± 0.1% for C. secundus in April 2015, and mean phosphorus content ± SE varied from 0.04 ± 0.003% for C. secundus in April 2015 to 0.2 ± 0.006% for G. tikvahiae in August 2014 (Supplementary Table 3). These results pointed to the value of expanded surveys to determine temporal, spatial, and interspecific variation in elemental compositions of DMA. These data can be used to make multiple estimates of changes in the amounts of carbon, nitrogen, and phosphorus stored in DMA, and as an example to illustrate the importance of such changes, the overall median percent compositions were multiplied by the relevant dry weights.
Carbon g = DMA biomass g DW m −2 × 0.2840 Nitrogen g = DMA biomass g DW m −2 × 0.0224 Phosphorus g = DMA biomass g DW m −2 × 0.0008 Based on mean annual biomass of DMA along transects, all reaches had less carbon, nitrogen, and phosphorus stored in DMA tissues in 2020 when compared to the amounts documented from 1998 to 2009 (Table 3). The losses of biomass ranged from 21% in reach 9 to 93% in reach 4 ( Table 3).
The decreases suggested that carbon, nitrogen, and phosphorus became available for uptake by other primary producers, such as phytoplankton.

DISCUSSION
Biomass of DMA varied significantly in space and through time, widespread decreases in biomass were related to thresholds for environmental tolerances as determined by incubations in the laboratory and field data, and less DMA translated into less carbon, nitrogen and phosphorus being stored in DMA. High biomass was recorded in different seasons, years, and reaches by the different methods, but low biomass was recorded by multiple sampling methods during 2010-2012 and in 2016, especially in reaches 1 through 6 where DMA was more common and abundant. Both these reductions in biomass of DMA followed periods of low light availability, and the 2010-2012 change also followed periods with potentially stressful low and high temperatures. Incubations indicated that light availability was a dominant influence, with less than ∼100 µmol m −2 s −1 leading to stress, and based on this threshold, there were widespread decreases in suitable habitat in 2010-2012, 2016, and 2018. During these periods, less carbon, nitrogen, and phosphorus were stored in DMA, with more of these elements likely to be available for uptake by other primary producers.
Although the biomass of DMA in the lagoon did not vary consistently between seasons in most years, it tended to be more abundant in March-July, and given the differences in sampling intensity and methodology, these results correspond well with previous surveys in the Indian River Lagoon (Benz et al., 1979;Kulczycki et al., 1981;Virnstein and Carbonara, 1985;Hanisak, 2021). Differences among the reports tended to occur in the cooler months, which was not unexpected given interannual variability in water temperatures. This temporal pattern was disrupted in 2010-2012 and 2016, with reduced light availability and stressful temperatures being apparent influences. The decreased availability of light coincided with phytoplankton blooms that increased light attenuation coefficients above the 90th percentiles for 6-8 months in all reaches during the  Table 2 for relevant statistics.
28 months from July 2010 to October 2012 and 6-12 months in reaches 1, 2, and 3 during the 18 months from July 2015 to December 2016 (Phlips et al., 2021;Lasi et al., this volume). A similar response to reductions in light availability caused by phytoplankton was noted in the coastal waters off Denmark (Nielsen et al., 2002).
Decreases in the biomass of DMA reduced storage of carbon, nitrogen, and phosphorus in this pool, and these elements should have become available to other primary producers. As an example, the amounts stored in DMA were calculated by multiplying the total amount of biomass in reaches 2-6 as estimated from the 2015 hydroacoustic survey (16,000 MT dry weight) by the conversion factors derived from samples of DMA. These calculations yielded estimates of 4,400 MT of carbon, 410 MT of nitrogen, and 13 MT of phosphorus stored in DMA. Applying the mean change in the biomass of DMA along transects in reaches 2-6 calculated for 2019 (−38%; Table 3) to these values indicated that 2,800 MT of carbon, 260 MT of nitrogen, and 8 metric tons of phosphorus would have become available. Given the volumes of reaches 2-6 and an assumption of no uptake, concentrations of carbon, nitrogen, and phosphorus in the water column would have increased by 1.419, 0.132, and 0.004 mg L −1 , respectively. These changes represented 14, 14, and 18% of the mean concentrations of dissolved organic carbon, dissolved nitrogen, and dissolved phosphorus in reaches 2-6 from 1997 to 2009, respectively (St. Johns River Water Management District, unpub. data). Thus, DMA played an important role in the cycling of elements in the lagoon that managers should consider.
Given that DMA play an important role in cycling of elements in the lagoon, managers could consider harvesting DMA to remove nitrogen and phosphorus and prevent shading of seagrass (Virnstein and Carbonara, 1985;Sfriso et al., 2020). Such an action demands careful consideration because removing DMA may increase the quantities of carbon, nitrogen, and phosphorus available to phytoplankton, but this issue could be addressed by timing the harvest to the period following maximum growth (Braun, 2020). In addition, less DMA translates into less structural habitat in the lagoon (Gore et al., 1981;Kulczycki et al., 1981;Virnstein and Howard, 1987;Holmquist, 1997;Monagail et al., 2017). For example, evidence has suggested that loss of DMA reduces refuge from predation for many macroinvertebrates (Heck, 1979;Heck and Thoman, 1981;Stoner, 1985;Stoner and Lewis, 1985;Edgar, 1987) and alters the composition of assemblages of benthic invertebrates due to species specific responses (Norkko et al., 2000).  Overall, management of the lagoon would be improved by a better understanding of the ecological roles played by DMA and attached macroalgae. Beyond elucidating its roles in cycling of elements and as a habitat, regular surveys that document the distribution, abundance, and diversity of DMA, experiments to elucidate the drivers of changes in the distribution and abundance of DMA more fully, and experiments to untangle the relationship between DMA and seagrass would supply valuable information to decision makers. In all cases, effort needs to be expanded to additional species of DMA and due consideration should be given to attached macroalgae. For example, DMA in reaches 2 and 4 during 2014-2015 were dominated by Chaetomorpha sp., large blooms of this green alga have blanketed seagrass in other locations (Pulich et al., 1997;Kennish et al., 2010;Gao et al., 2013), and approximately 35% of drift algae in parts of the lagoon are not Gracilaria (Hanisak, 2021). An increased likelihood of such blooms could be driven by climate change because green algae tolerated relatively high temperatures (Menéndez and Comín, 2000). Furthermore, blooms of DMA have been shown to shade seagrasses or self-shade, which increased concentrations of nutrients in the water column due to decomposition (Holmquist, 1994(Holmquist, , 1997Hauxwell et al., 2001;Cummins et al., 2004;Gao et al., 2013;Foster et al., 2018). Additionally, Fox et al. (2008) stated that Cladophora vagabunda and G. tikvahiae in locations with higher nitrogen loads from their watersheds stored up to 250% of the annual load. Similar roles may be played by attached algae. For example, Caulerpa prolifera has been documented as the dominant vegetation in the northern Indian River Lagoon and in deeper areas in Banana River Lagoon (White and Snodgrass, 1990;Provancha and Scheidt, 2000), and this species has demonstrated an ability to take up significant amounts of ammonium (Alexandre and Santos, 2020) so its role in cycling of nutrients should be considered.
In conclusion, reduced light availability and extreme temperatures in the IRL likely contributed to decreased growth or mortality of DMA. In turn, less DMA likely promoted phytoplankton blooms because more carbon, nitrogen, and phosphorus became available. The presence of DMA in the IRL has been recognized for decades (Thompson, 1978), but an increased understanding of its roles will support improved management of the system.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.