As more sampling is completed, e.g. baseline surveys and during pilot mining programs, there will be an opportunity to build a database for genetic information that can be linked to specific species. Accomplishing this aim may require a combined approach.Benthic organisms can be classified by size: megafauna , macrofauna , meiofauna , and microbial communities. Megafauna are the only group that can be readily identified with imagery generated by photo or video methods, which can also contribute to our understanding of organism behavior and interaction . On the other end of the spectrum, microbial biodiversity is identified via molecular techniques such as sequencing . Macro- and meiofauna fall somewhere in the middle, amenable to both approaches but perhaps both are not always simultaneously appropriate. Macrofauna are often used as environmental indicators , and there is work to develop similar deep-sea measures using meiofauna where they dominate the biological community . It has been shown that MBT and metabarcoding can perform equally as well for biomonitoring of freshwater systems and assessment of marine aquaculture impacts .
In the deep sea,vertical aeroponic tower garden the limited taxonomic expertise on deep-sea meiofauna , may make the latter approach more appropriate for rapid assessment of biodiversity and environmental impacts.Molecular methods are better suited for identifying rare and cryptic species than MBT. Because of their low abundance and patchy distribution, missing rare species because of under sampling is often a problem, especially in the abyssal deep sea. Additionally, small, soft-bodied protists often disappear in fixed sediment and only show up in eDNA . Although rare species do not have significant biomass, they can contribute unique functions , and new species are discovered frequently with few representative individuals . Choosing clustering cutoffs for OTUs can affect results, losing rare species if too conservative and retaining technical artefacts if too lenient . Cryptic species usually cannot be identified via morphology and, therefore, require molecular methods to distinguish them. They have been documented in deep-sea taxa, such as polychaetes and gastropods , and can significantly impact estimates of species richness .MBT can provide demographic information and material that allow for examination of life histories, lifestyles, and functional attributes. On a practical level, sequencing technology cannot yet reveal these kinds of data. Whether demographic information is relevant to mining environmental requirements depends on the goals and objectives set by the ISA . Metrics such as density, dispersal, size structure, sex ratio, and physiological responses to environmental changes are often used to assess ecosystem health .
Metabarcoding is sufficient to estimate relative abundance, dispersal, and community changes , but additional techniques are required for the remaining metrics . However, capacity to interpret molecular data continues to increase, such as experimental calibration to calculate abundance data , or computational techniques to identify differentially abundant taxa and community shifts .In order to establish identification of deep-sea species and their ecological roles, it is likely MBT and molecular approaches are needed in combination. Metabarcoding yields sequences of genes that are clustered into OTUs which can serve as a proxy for species. However, in order to assign an OTU to a species, that species must have been morphologically identified and sequenced previously, or belong to a taxonomic clade well-supported in a robust phylogeny . Once described morphologically and genetically, metabarcoding alone can then be used to identify it. However, a molecular approach may greatly benefit from the improvement of MBT-based analyses, such as collating phylogenetic histories with environmental data . The former is sufficient if species diversity is the only interest. However, ISA draft exploitation regulation calls for the application of “an ecosystem approach” which implies information on functional diversity and ecological interactions are also necessary .There could be scenarios in which rapid results are required in order to respond appropriately to environmental impacts, such as when an environmental threshold is surpassed or a trigger is set off indicating serious harm to the marine environment . An ecological threshold is a “tipping point” at which the system undergoes rapid and irreversible change that triggers an action .
As an example, a possible threshold to cessation of mining may be the loss of a specified amount of local biodiversity , implied by current mining guidelines for environmental management and monitoring plans , ISA, 2017; Annex VII , ISA, 2019. In a scenario using MBT, results can be obtained only after months of work, whereas metabarcoding results can be obtained within days after return to shore. This difference in response time may have significant impacts on the amount of environmental damage that ensues and, ultimately, the ability for an ecosystem to recover. The amount of time required to develop the expertise necessary for either approach differs greatly. Depending on taxon, it can take multiple years for a taxonomist to become accurate and efficient at species identification using MBT. In the deep sea, this expertise is especially lacking, particularly for small taxa, which can overwhelm experts’ workloads and increase turnaround time. Parataxonomy, which is the delegation of taks in an MBT workflow to non-experts, can alleviate some pressure on taxonomic experts. Although parataxonomy does not yield reproducible results, it can be useful when rapid biodiversity assessments are necessary . In contrast, popularity of molecular approaches has caused the proliferation of sequencing facilities, both academic and commercial, for turnaround time of approximately 2- 3 weeks from sample submission to data return. The more limited training required to process sequencing samples makes molecular approaches more accessible, and the automation that is actively being developed will further reduce processing times. Additionally, loss of taxonomic expertise over time can be mediated by genetic information which can be used as an immutable characteristic for taxonomic identification.One of the biggest challenges facing both MBT and molecular approaches is developing an adequate sampling design. There are still many gaps in our knowledge of deep-sea biodiversity , but it is unclear at what spatial and temporal scales scientists need to sample in order to accurately characterize it . Spatial heterogeneity is poorly defined and likely differs among habitat and location. In the CCZ, there can be unexpectedly high spatial heterogeneity . Additionally, adequate temporal resolution of sampling for accurate characterization of deep-sea habitats is unknown . MBT and metabarcoding often provide one snapshot in time and, given time constraints associated with MBT, a comprehensive monitoring program may incorporate both methodologies by employing metabarcoding more frequently and MBT at longer intervals. The study of deep-sea habitats is plagued by a lack of data, creating challenges that are exacerbated by the sheer vastness of the deep sea. The CCZ claim areas span almost 4500 km wide,vertical gardening in greenhouse larger than the continental U.S., which leads to high levels of uncertainty because of data limitation. A precautionary approach, wherein adequate measures are taken in order to manage risk, should be invoked . Therefore, collection of robust baseline data is critical for protection of the marine environment , as well as exhaustive monitoring, especially at the early stages of commercial exploitation . MBT and molecular approaches can be utilized together in these programs to facilitate future environmental assessment and monitoring.There are direct and indirect costs involved in collecting and processing sediment samples for MBT and metabarcoding. Here, they are separated into costs associated with consumables and labor . Other costs, such as laboratory equipment and bio-informatics pipelines, are held fixed. Choices within the decision networks are discussed in terms of how they affect the short-run economic cost. We assume that laboratories have basic amenities and access to people who are trained to do these tasks, i.e. identifying target taxa, preparing sediment samples for sequencing. Deep-sea samples, whether sediment, water, or individual organisms, are relatively expensive to collect because they often require research expeditions on global-class vessels equipped with specialized instruments.
One ship-day on a U.S. vessel can range from $42,000- 48,000 USD, which includes food, fuel, crew, and two technicians . The amount of time it would take to sail from the Scripps Institution of Oceanography pier to the eastern edge of the CCZ is approximately six days at full speed without stopping, resulting in over a quarter million dollars in transit time one-way. This cost estimate does not include the use of autonomous underwater vehicles , remotely-operated vehicles , or human-occupied vehicles , commonly used on deep-sea expeditions. In addition to collecting sediment samples for both MBT and sequencing, deep submergence vehicles can conduct visual surveys. The U.S. National Deep Submergence Facility operates their instruments at day-rates of $14,000 for AUV Sentry, $23,000 for ROV Jason, and $45,000 for HOV Alvin . Other sampling equipment includes box corers, multi-corers, epibenthic sleds, and CTD rosettes which can be deployed without special vehicles. Relevant results and sampling regimes from studies included in this paper are summarized in Table 3.2. Sediment samples were taken with multi-corers, box corers, or an epibenthic sled in one case, and the same sampling devices can be used for both approaches. The total number of samples used for analysis ranged from 5-41 and 15-42 per cruise for MBT and metabarcoding, respectively. Here, we assume number of samples equates to number of vertical fractions . Sampling costs are likely the biggest element but should not contribute significantly to cost differences between the two analytical approaches. However, the different methods used to process deep-sea sediment samples are a major factor.Consumables required for MBT are common to most labs and are readily available .Formalin is cheaper, easier to ship , and fixation results in little distortion of morphology. However, specimens fixed in ethanol can later be used for molecular studies which is advantageous and cannot be done with specimens fixed in formalin. The total cost of lab consumables for MBT ranges from $13.10-16.29 USD. Labor and expert time are the dominant costs of the MBT workflow . Based on a graduate student stipend in the U.S. of $29,500 USD annually, time alone ranges from $325-2,875 per sediment core slice. Additionally, the estimates shown here are only for the taxa listed for one vertical fraction of a sediment core although the top several centimeters are likely necessary for biodiversity characterization. The extensive time needed to manually sort and identify organisms, especially when using smaller sieve sizes that retain more sediment, is a large cost and a disadvantage of using MBT. Ludoxâ flotation, a method of concentrating animals in a liquid medium, can reduce picking and sorting times but adds additional consumable costs. The estimates we have provided are based on hand-sorting. There are many options when processing marine sediment for metabarcoding , influencing scientific outcomes and economic costs. A metabarcoding approach can also be used in parallel with other omics methods if relevant to objectives . The number of replicates in a study can refer to either sample replicates or PCR replicates. If there is enough sediment material in each sample, then pseudo-replicates can be added. PCR replicates are more often used and can help minimize errors due to PCR, such as unequal amplification and chimeras . Additional replicates yield more robust results and higher statistical power, but increase total cost . Sampling replicates and pseudoreplicates can also apply to MBT. Metabarcoding utilizes highly conserved fragments of DNA that all organisms have and within these fragments are markers, which are highly variable regions that allow for higher resolution taxonomic classification. Choosing markers, how many and which ones, can affect results and will depend on scientific objectives. 16S rRNA can be used to assess microbial diversity whereas eukaryote diversity is typically assessed with markers within the 18S rRNA or cytochrome oxidase I genes . More comprehensive diversity assessments may require multiple markers in order to completely capture the community. Advantages to working with widely-used markers include a more optimized protocol, the ability to compare to other studies, and a larger database. However, there is no truly universal marker that allows the retrieval of all taxa without bias. Additionally, a universal marker must be very conserved and therefore less useful for species-level taxonomic assignments. Development of markers targeting indicator taxa allows for better detection of them . These bioindicators can then be used for characterizing specific habitats, such as sediment plumes, and early detection of mining impacts .