Calcification accretion units (CAUs): A standardized approach for quantifying recruitment and calcium carbonate accretion in marine habitats

Standardized metrics that quantify a component of ecosystem functioning are essential for evaluating the current status of coastal marine habitats and for monitoring how ecologically important ecosystems are changing in response to global and local environmental change. Calcification accretion units (CAUs) are a standardized tool for quantifying net calcium carbonate accretion, early successional community structure, recruitment of algae and sessile invertebrates and other response metrics that can be determined from image analyses in coastal marine habitats. CAUs are comprised of paired‐settlement tiles that are separated by a spacer. This design mimics the presence of different representative habitats that are common in most marine systems such as exposed benthic surfaces, cryptic spaces inaccessible to grazers and shaded overhangings. The protected space between the tiles facilitates recruitment and inclusion of cryptic taxa in community assemblage estimates. After a period of deployment, CAUs are photographed for image analysis and then decalcified to quantify calcium carbonate accretion rates. The CAU methodology provides a cost‐effective, standardized protocol for evaluating structure and function in marine benthic habitats. We illustrate how CAU data can be used to compare accretion rates and the relative proportion of carbonate polymorphs in ecosystems across the globe. Here we provide a comprehensive standard operating procedure for building, deploying and processing CAUs, to ensure that a consistent protocol is used for accurate data collection and cross‐system comparative studies.


| INTRODUC TI ON
Standardized approaches for evaluating ecosystem functioning are essential for characterizing the current status of coastal marine habitats. Such measurements allow for cross-system comparisons and provide a rigorous approach for understanding how ecologically valuable ecosystems are responding to rapid environmental change (Van Rein et al., 2009). Coastal degradation of marine habitats continues to accelerate due to local and global human impacts, and as a result the development and implementation of consistent protocols and strategies to monitor disparate habitats has become increasingly important (Jackson et al., 2001;McCauley et al., 2015). Evaluation of benthic community composition, such as tracking changes in cover or abundance of community members over time, is a commonly used metric in ecosystem monitoring. These data provide valuable insights into the structure of community assemblages (i.e. diversity, abundance), and although they can provide proxies for ecosystem health and function (Bellwood et al., 2019), they do not inherently scale up to ecosystem processes . Direct measurements of emergent ecosystem properties, such as biological production, biogeochemical cycling or calcium carbonate (CaCO 3 ) accretion, are therefore needed to determine spatial and temporal changes in marine ecosystem functions (Hatcher, 1997).
Emergent ecosystem properties represent the sum of all biotic, abiotic and interactive processes that shape communities (Hatcher, 1997) and provide a more holistic assessment of how an ecosystem is functioning . Calcium carbonate (CaCO 3 ) accretion is an example of an emergent ecosystem property in marine habitats that directly represents key ecosystem functions such as habitat vertical growth, stabilization, expansion and erosion (Hatcher, 1997). Although net accretion of carbonate is particularly important in calcifier-dominated habitats (e.g. coral and oyster reefs), it can be used as an indicator of ecosystem function in a suite of marine habitats across the globe.
Marine taxa ranging from bivalves and arthropods to corals and algae secrete external calcium carbonate shells and skeletons that persist even after the organism dies (Lowenstam & Weiner, 1989). The net accumulation of these carbonate structures over time, as well as other mechanisms of carbonate precipitation, is the product of abiotic and biotic constructional (i.e. CaCO 3 production) and erosional processes (Perry et al., 2008). The resulting carbonate accretion facilitates the growth and persistence of habitat frameworks that in turn support a diversity of associated species. Evaluating current and future rates of carbonate accretion is becoming increasingly important because ocean acidification impairs biogenic calcification and has the potential to decrease ecosystem accretion rates, which threatens the proper functioning of calcifier-dominated habitats. By monitoring net accretion, we can first establish baseline ecosystem rates and then develop a quantifiable and predictive understanding of the ecosystem-wide consequences of ocean acidification (Perry & Alvarez-Filip, 2019).
Calcification accretion units (CAUs) employ the commonly used 'settlement tile' methodology to collect a representative population of a marine community (Field et al., 2007) and build on it by including cryptic habitat space and incorporating the quantification of new carbonate accretion (Price et al., 2012;Vargas-Angel et al., 2015) ( Figure 1). In a typical study with settlement tiles, panels are deployed in a focal habitat where they are colonized by resident taxa for varying periods of time. Tiles are then analysed for community composition, diversity or recruitment success Thomason et al., 2002). However, a single standardized method for settlement tile deployments does not currently exist. Tiles are made of diverse materials, ranging from terracotta to plastic, and are deployed for different periods of time, which can influence the taxa that settle on the plates (Field et al., 2007). The substrate type, F I G U R E 1 CAU field deployments. Calcification accretion units (CAUs) are deployed in coastal marine habitats where they are used to evaluate early successional community structure and net calcium carbonate accretion as an indicator of ecosystem function. (a) CAUs are deployed by a SCUBA diver on a coral reef in the Eastern Tropical Pacific. (b) A new CAU ready to be swapped with a unit that was deployed for 1 year on a coral reef in Belize. (c) CAUs were deployed on PVC poles in soft bottom substrate on an oyster reef in the Indian River Lagoon in Florida, shown at low tide. Photo credit: Sean Mattson (a) and Maggie Johnson (b, c) duration of deployment, size of tile and sampling frequency often depends on the research question of interest, and because the approach varies by habitat and research group, the data are often not directly comparable (Field et al., 2007). Although a wide range of settlement tile data currently exist in the literature, much of it cannot be used to make direct spatial and temporal comparisons because of fundamental differences in methodology.
The CAU methodology offers a standardized protocol for constructing, deploying and processing settlement tiles, and builds on the classic single 'settlement tile' design by including a space that facilitates recruitment by cryptic taxa. Cryptic taxa contribute to biodiversity in marine habitats, but are often overlooked and underestimated by standard settlement tile protocols (Plaisance et al., 2011;Ransome et al., 2017). CAUs further elevate the typical settlement tile approach by incorporating a standardized technique to quantify the amount of CaCO 3 that accumulates on the tiles during the deployment through decalcification of tile communities and then calculation of the non-calcified (also referred to as fleshy) and calcified biomass of sessile taxa. Additionally, photographs of CAU tiles are used to evaluate community assemblages and can be used for any other image-based questions such as quantifying the recruitment of key taxa (i.e. corals) or invasive species.
The methodology of CAUs was introduced by Price (2010), as a modification of a design implemented in Raimondi and Morse (2000).
The first use of CAUs as paired-settlement tiles for quantifying The CAU technique is being adopted readily by different research groups and programs, in part because the method is elegant, cost-effective and less time-intensive than other approaches. Yet, a complete protocol is not widely available. To facilitate the broad usage of CAUs and to ensure that the protocol being implemented by different groups is indeed standardized, we describe a comprehensive standard operating procedure (SOP). Here we provide the instructions for constructing CAUs and the standard procedure for CAU deployment, retrieval and data collection.

| LIMITATI ON S OF C AUs
CAUs facilitate the collection of standardized data across habitats and provide insight to ecosystem functioning and early successional community structure. However, some caveats to the CAU methodology should be considered when implementing them in research and monitoring programs. CAUs are colonized by contemporary communities and represent early successional stages of sessile benthic communities; thus, accretion estimates from the CAUs represent 'new' accretion in the system. Furthermore, because the CAUs are not protected from biotic and abiotic pressures (e.g. herbivory, predation), any carbonate accumulation represents net accretion (i.e. the net sum of carbonate deposition and erosion). If a researcher is interested in exploring gross carbonate accretion metrics, the CAU methodology could be augmented with simple exclusion cages (Lewis, 2016;Smith et al., 2010). Another consideration is that CAUs represent a relatively small surface area relative to available benthic habitat. To ensure that CAUs are representative of the adjacent benthos, the number of replicates should be optimized per site.
The CAU methodology includes a protected space intended to facilitate settlement and inclusion of cryptic taxa. As a flat, homogenous surface (i.e. space between two sanded PVC tiles), this design is inherently limited in structural complexity. Microtopography and three-dimensional surface structure can influence settlement and recruitment processes (Edmunds et al., 2014;Nozawa et al., 2011), and thus the flattened, protected habitat of the CAU surfaces likely yield an under-representation of true cryptic diversity and may influence subsequent community dynamics and function (Brandl & Bellwood, 2016). Targeted studies that incorporate structural complexity into the CAU design, such as the different substrate types used in ARMs panels (e.g. plastic pond mesh), would shed light on the role of surface complexity in shaping early successional community dynamics. An additional consideration when using CAUs for estimations of biodiversity is that taxa are identified to coarse functional groups through image annotations, and image analyses underestimate biodiversity, particularly of cryptic taxa (Casey et al., 2021).
The utility of the CAU design is that it could easily be augmented to quantify biodiversity using molecular techniques, such as the DNA metabarcoding approach used with ARMS (Leray & Knowlton, 2015).
As with any methodology, appropriate implementation is contingent on the research question of interest. CAUs provide a standardized method for evaluating present-day net carbonate accretion as an indicator of ecosystem function and can be applied in virtually any benthic marine habitat. By using a closely standardized protocol, valuable data can be compared across different habitats, regions and ecosystems over time.

| CAU tile construction
CAUs are generally cost-effective because they are constructed from materials commonly available via local hardware stores or commercial suppliers (Figure 2a; Table 1), and the materials can be recycled for subsequent deployments after appropriate cleaning.
CAUs use polyvinyl chloride (PVC) as the standardized substrate for the tiles, a material that facilitates settlement of biotic assemblages similar to the surrounding substrata and is both inert and acid-tolerant (Hixon & Brostoff, 1985). Although pre-cut tiles may be ordered, large sheets of PVC plastic can be purchased and cut into the requisite 10 × 10 cm size using standard carpentry tools.
The steps for constructing and assembling CAUs are outlined in Figure 2, and a complete materials' list with example part numbers from a representative commercial supplier (e.g. McMaster Carr) is presented in Table 1. Each tile side, or PVC plastic sheet (prior to F I G U R E 2 CAU assembly and deployment. (a) Materials for constructing CAUs can be purchased at local hardware stores or from commercial hardware suppliers. (b) One CAU unit consists of two 10 × 10 cm PVC tiles stacked on top of each other and separated by a plastic spacer. (c) Stainless steel nuts, lock washers, tiles and spacers are threaded on to a stainless steel threaded rod. (d) One CAU unit recently deployed on a stainless steel stake on a coral reef in Belize. Upper case letters in images correspond to part numbers listed in Table 1 and the asterisk indicates an optional tag 10 cm cutting), is sanded to roughen the surface and facilitate recruitment (Hixon & Brostoff, 1985). Sanding is recommended with a rotary sander fitted with a medium grit sanding disc (e.g. size 80), and each tile should be sanded until surfaces are homogenous.
Sheets of PVC can then be cut to 10 × 10 cm tiles and a centre hole drilled with a drill press and ¼″ drill bit.  (Figure 1c). In mild-to-moderate disturbance environments, CAUs can be attached to stakes or PVC poles with at least three zip ties, cinched as tightly as possible. In higher energy habitats, CAUs should be secured to stakes with wire rope clamps that ensure CAUs maintain a level orientation during the deployment. Subtidal CAUs are deployed at each site by SCUBA divers, and sites are carefully recorded with GPS coordinates to facilitate relocation.

| Retrieval
The standard deployment duration for CAUs is 1 year, although this can be shortened to suit specific research questions or extended to longer periods for remote sites. For retrieval, a SCUBA diver or snorkeler removes the attachments for each CAU (i.e. rope clamps removed using wrenches or zip ties cut with snips) and places each CAU into a separate gallon Ziplock bag with seawater. Keeping organisms on the CAUs alive until processing is important to preserve organismal integrity (e.g. pigmentation) and aid species identification; direct sunlight and exposure to extreme temperature variability should be avoided. Units should be kept in well-flushed and oxygenated seawater until processing (5-gal TA B L E 1 Components for CAU assembly, including part numbers from a representative commercial supplier

| Immediate processing
CAUs Makeshift ovens can be built using heat lamps, aluminium foil and wooden or plastic boxes, but ensure fire danger is mitigated before use. Tiles should be stored in airtight bags or containers until subsequent processing (Figure 3b).

| Recruit enumeration and identification
While many taxa can be identified and enumerated using imagebased analyses, we strongly recommend using the actual dried tiles and a stereomicroscope to quantify the recruits of scleractinian cor-

| Decalcification
After the tiles are dried, they can be stored indefinitely in an airtight container away from light. If stored, tiles should be dried again briefly (1-2 hr) before subsequent processing. To begin processing, tiles should be removed from the drying oven, allowed to cool briefly (~1 hr) and then weighed to the nearest 0.01 g (total dry weight).
Commercially available HCl is sufficient for decalcification (laboratory grade is not necessary).

| Post-decalcification processing
Once decalcified, tiles are scraped with a straight edge razor blade to remove residual biomass from tiles, and tiles are rinsed with DI water.
The slurry (mix of neutralized HCl and fleshy biomass) is then vacuum filtered through pre-weighed 11μm cellulose filters (Whatman) using a ceramic funnel. In preparation, all filters should be numbered with a permanent marker, dried at 60°C for ~1 hr, cooled briefly and then weighed. This is the initial (bare) filter weight, which will be subtracted from final filter weights to calculate the actual non-calcified (hereafter referred to as fleshy) biomass.
Decalcification containers should be lightly rinsed with DI water to remove residual biomass, and the sides of the funnel should be lightly rinsed during filtration to allow all biomass to accumulate directly on the filter. The slurry is vacuum filtered until all liquid is removed (Figure 3d). The vacuum is then turned off and the vacuum seal is broken by gently lifting the corner of the filter with forceps.
The filter with biomass is gently removed with the forceps, folded in half and then wrapped in a foil envelope (as above for the tiles, leaving the top open to allow moisture to escape). If the acid slurry is not fully neutralized, the residual HCl in the filter can react with the foil, in which case wax paper should be placed between the filter and foil to protect the integrity of the filter and sample. Filters with fleshy biomass, as well as the scraped bare tiles, are then dried as described above. Bare tiles should be rinsed with DI water prior to drying and labelled appropriately.
After the drying period, bare tiles and filters with dried biomass are allowed to cool briefly and weighed as described above.
The bare tile weights and filter with dried biomass weights are both used in the calculation of accretion rates. The fleshy biomass of each tile is calculated by subtracting the bare filter weight from the weight of the dried filter with biomass (filter with dried biomass − bare filter = fleshy biomass). This represents the noncalcified, organic portion of the community on each tile. The calcified biomass is then calculated by subtracting the bare tile weight from the total dried tile weight and then subtracting the total fleshy biomass (total dry weight − bare tile − fleshy biomass = calcified biomass).
Net accretion (i.e. calcification) rate is determined by summing the calcified biomass of both tiles in a CAU unit and then normalizing to the surface area of the tiles in a CAU unit (400 cm 2 ) and duration of deployment in years. Net calcification per CAU unit is expressed as g CaCO 3 cm −2 year −1 . Rates can then be averaged across all CAUs within a site to calculate site-level net accretion rates.
The core data resulting from the decalcification process are as follows: calcified biomass, fleshy biomass (also referred to as non-calcified biomass) and total dried biomass. An example dataset of carbonate accretion rates and non-calcified biomass from coral and senescent reefs in the Atlantic and Pacific are presented in Table 2, illustrating the utility of CAUs for global comparisons.
Statistical analyses with these data should properly evaluate model assumptions, as some data will be prone to non-normal dis-

| IMAG E ANALYS IS
When possible, image analyses should be conducted prior to decalcification to allow the opportunity for cross-referencing image annotations with the actual tiles. At this step, recruits identified during the enumeration stage can be used to validate and inform image annotations. Image analysis can be conducted in any software that allows for random distribution of points, and identification of taxa underneath points, such as CPCe (Coral Point Count with Excel extensions) (Kohler & Gill, 2006) or the online platform CoralNet (https://coral net.ucsd.edu) (Beijbom et al., 2015). Images should be cropped and corrected where necessary to facilitate identifications. Image analysis is performed with 25-100 randomly stratified points overlaid on each tile image (actual number of points should be determined using a power analysis and will depend on the questions of interest and the variability that exists in the system). The taxon or substrate underneath each point is identified to the finest taxonomic resolution possible. Species can be later grouped by functional group and carbonate polymorph for subsequent community analyses (Price et al., 2012;Vargas-Angel et al., 2015). Specific identifications in the image analyses will vary by location. However, each taxon should be tagged with a minimum of the following designations: accretion group (calcified, non-calcified, abiotic) > functional group > carbonate polymorph > genus > species (when possible).
An example of functional group designations that have been used with CAUs deployed on coral reefs is presented in Table 3 ( Price et al., 2012).
From the image annotations, the per cent cover of individual taxa as well as the per cent cover of major functional groups and the proportion of taxa comprised of different CaCO 3 polymorphs (aragonite, calcite, high-mg calcite) can be calculated (see example data in Table 2). Abundance data are usually presented by 'habitat type' with the surface (top tile), cryptic (in between tiles) and overhang (underneath bottom tile) habitats reported separately. Alternatively, data can be reported as total per cent cover for the whole CAU unit.
The data can be mined later to answer additional research questions, such as recruitment of specific taxa (e.g. invasive species, coral recruits but see above). A schematic of the suggested workflow for the core metrics in the standardized CAU protocol is presented in Figure 4.

| CON CLUS ION
Habitat degradation continues to escalate in coastal marine habitats, and methods that implement standardized protocols to evaluate spatial and temporal changes in ecosystem functioning are becoming increasingly important. The increasing adoption and utilization of CAUs, as well as other standardized methods such as ARMS, further speaks to the need for a clear and consistent protocol that can be used by different research groups and programs. Although settlement tiles have long been used for recruitment studies in marine research, the CAU method we detail here provides an opportunity to elevate the data collected from settlement tiles beyond community assemblage and to incorporate a metric of ecosystem function. By increasing the availability and transparency of the standardized CAU F I G U R E 4 CAU protocol schematic. Suggested workflow for the standardized collection and processing of CAUs for community structure and carbonate accretion data protocol, we hope this tool can be adopted by more research groups and monitoring programs to enable large-scale cross-system studies and collaborative research to fill much needed gaps in our knowledge of coastal marine ecosystem performance now and in the future.

ACK N OWLED G EM ENTS
The authors thank Danny Merritt, Rusty Brainard, Christi Richards, Kevin Wong and Emily Donham for insight during the development stages of the CAU methodology.

CO N FLI C T O F I NTE R E S T
The authors declare no competing interests. All co-authors edited the manuscript and approved of the final submission.

PEER R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/2041-210X.13867.

DATA AVA I L A B I L I T Y S TAT E M E N T
This manuscript contains no data.