Skip to main content
Dryad logo

Datasets for: Primary production and habitat stability organize marine communities

Citation

Cramer, Alli; Katz, Steve (2021), Datasets for: Primary production and habitat stability organize marine communities, Dryad, Dataset, https://doi.org/10.5061/dryad.wpzgmsbjv

Abstract

Aim: The emergence of pattern in the natural world can carry important messages about underlying processes. For example, collections of broadly similar terrestrial ecosystems have historically been categorized as biomes – groupings of systems which sort along energetic and structural process axes. In marine systems however, a similar classification of biomes has not emerged. The aim here is to develop an effective classification scheme for marine biomes and communities.

Approach: Candidate predictor variables that could explain pattern and process in differentiating marine communities, such as light, nutrients, depth, etc., were collected from the existing literature with a systematic review. The candidate predictors were then evaluated in an inductive process, allowing community level observations to demonstrate patterns across marine biomes. Marine biomes and communities were identified a priori, and emergent patterns were evaluated quantitatively, via Principal Component ordination based on the vectors of explanatory predictors, and qualitatively via mapping.

Conclusions: Gross primary production and substrate mobility not only effectively sort marine biomes, but also work on finer scales discriminating communities within biomes. As a result, these predictors were more effective in classifying marine communities than other scales, such as available light and nutrients. The richness of this classification is also demonstrated in revealing other patterns, such as the distribution of human impacts. The effectiveness of this mapping provides support for, but is not a test of, the hypothesis that primary production and substrate mobility are important underlying processes that interact to structure marine ecosystems.

Methods

To develop classification schemes we performed a systematic review of the literature to compile representative values for a set of candidate predictor variables from a range of marine communities and biomes. Marine biomes included coral reefs, hydrothermal vents, hard bottom, soft bottom, vascular vegetation, open ocean or pelagic, and ice. Hard bottom biome communities span a range from gravel beds to continuous rock. Similarly, soft bottom biome communities range from small gravels to clay communities and are characterized by infaunal sessile organisms. Open ocean communities are characterized by their lack of stable substrate and pelagic organisms. The vascular vegetation category includes any communities dominated by vascular plants such as sea grasses and mangroves, and were largely at the ocean surface. Comparable literature data for other community types, including cold seeps, deep water corals and ephemeral habitats such as whale falls, were desired, but unavailable.

Predictor Variables

For each community, literature values were obtained for depth, light, nutrient flux, Gross Primary Production (GPP), substrate grain size and substrate mobility.

Except where noted, the values of the all variables are drawn from single sources and are therefore correlated for each observation rather than being independent, but representative for a given community. For comparison, GPP and nutrient flux are normalized to unit of bottom area of that community type. Light, depth, grain size, and substrate mobility, being uniform with respect to an observation, were not normalized to area.

Depth

When possible, depth values represent the maximum reported depth for a given GPP observation. If depth values were not reported in a given reference for GPP, we determined depth based on site descriptions, location, and community type. For temperate and tropical seagrass communities we determined depth values by taking the mean depth values of temperate and tropical species from the World Atlas of Seagrasses (Green & Short, 2003). For shore or intertidal communities, a depth of 0 m was used. The depths of hydrothermal vents were calculated using the average depth from the InterRidge Global Database of Active Submarine Hydrothermal Vent Fields Version 3.3 (Beaulieu, 2015). For open ocean observations we used a maximum depth of 200 m. 

Light

Light availability, or intensity within various communities was determined using the Lamber-Beer equation for the attenuation of light in water, ID = I0*e-kD , where I0 is the intensity of light at the surface, D is depth, and k is the extinction coefficient of light in water (Dennison, 1987). The extinction coefficient expresses the rate of absorption and scattering of light as it travels through water. Although the true penetration of light into marine habitats varies widely with salinity, suspended sediments, and biota, an extinction coefficient of pure water (0.035) was used to allow a general comparison (Röttgers, McKee, & Utschig, 2014; Valiela, 1995). Using the attenuation equation, the relative light intensity for any given depth of water was determined by setting available light intensity at the surface, I0, to 1.

Nutrients

Nutrient availability in each community was expressed as characteristic nitrogen flux in units of g N y-1 as reported by Gruber and Galloway (2008), who reported yearly flux for nearshore and offshore compartments of the ocean. They estimate natural (i.e. unperturbed) nearshore nitrogen flux to be 30 Tg N y-1 and natural offshore flux, a combination of nitrogen fixation and atmospheric deposition, to be 150 Tg N y-1. Nearshore and offshore compartment fluxes were converted to community-specific net annual fluxes by scaling with the fraction of compartment composed by each community. Total global area of various communities was determined using available habitat maps and databases. We summed the areas of communities within the nearshore and offshore habitats to obtain area estimates for nearshore and offshore compartments. We then multiplied by the proportion of respective compartments taken up by candidate communities (Eq 1).

Eq. 1. Community g N m-2 y-1 = Total g y-1 N compartment area m2* Proportion of compartment

For hydrothermal vents, total nitrogen flux consisted of the sum of nitrogen species (i.e. NO3-, NO2- & NH4+) released from a characteristic hydrothermal vent chimney. Concentrations of nmol l-1 [N2O], umol l-1 [NO3- + NO2-], and umol l-1 [NH4+] were totaled for each location (Bourbonnais et al., 2012) and converted to total g l-1 N. If a nitrogen species wasn’t measured for a particular location, we assumed a minimum concentration equal to the lowest observation reported for that chemical species at all locations. We then averaged observations to obtain a characteristic concentration of 5.4 × 10-4 g l-1 N. An annual nitrogen flux rate was estimated by multiplying the characteristic concentration by annual vent flow rate, then normalizing by our hydrothermal vent community area (see above).

Gross Primary Production

Sources that could contribute values for GPP were obtained through a systematic review of online literature using the search terms: “primary production”, “g C m-2”, “community metabolism”, “photosynthesis “, and “chemosynthesis” within the various communities and biomes of interest. To account for high seasonal and daily variation in primary production, values are expressed on a yearly basis (g C·m-2y-1); values reported in other units in the literature were converted to g C·m-2y-1.

Values for GPP at hydrothermal vents were adjusted to account for both the monitoring methods and the fact that the “halo”, or area of habitat influenced by the point source of vent flow may not easily correlate with size of vent chimneys. Annual primary production was estimated as the product of vent fluid carbon concentration and the characteristic vent flow. Vent carbon concentration was 2 mg C l-1 (McCollom, 2000; Winn, Cowen, & Karl, 1995). Characteristic vent flow was determined by averaging flow rates from focused hydrothermal vents (1,953 cm3 s-1; Germanovich et al. 2015). Literature values for area of habitat varied widely, in part as they differed in the unit of habitat from single vent “chimneys”, cracks, mounds or vent systems (e.g. Copley, Tyler, Murton, & Van Dover, 1997; Fustec, Desbruyères, & Juniper, 1987; Gebruk, Chevaldonné, Shank, Lutz, & Vrijenhoek, 2000). We estimated a characteristic hydrothermal vent community size of 10 m x 10 m, or 100 m2 based on the range of literature values.

Within sea ice communities there are multiple types of ice, with primary production occurring in various layers and estimates ranging widely. We combined estimates from both surface ice (Arrigo, Worthen, Lizotte, Dixon, & Dieckmann, 2009; Lizotte & Sullivan, 1991) and platelet ice (Grossi, Kottmeier, Moe, Taylor, & Sullivan, 1987). Ice is transient. For comparison with permanent habitats, productivity rate under ice was normalized by the estimate of ice-on duration (approximately 66 days or 0.18 of a year, Grossi et al. 1987) to obtain a “yearly” primary productivity rate value.

For seagrass communities gross primary production values came from Duarte et al’s (2010) estimates of seagrass community metabolism. Those authors aggregated seagrass community production from 52 sources and reported GPP for temperate and tropical seagrass communities’ separately. Because the seagrass data stems from a review study, rather than individual studies, we reported the mean GPP values for temperate and tropical seagrass communities aggregating across the 52 studies. To reflect the range of GPP within seagrass we also reported the aggregate GPP value one standard deviation above the mean, resulting in two values for each seagrass community type.  

With the exception of coastal ocean, and some values for salt marshes (Knox, 2000) and kelp forests (Reed, Rassweiler, Arkema, & Barbara, 2008; Rodgers & Shears, 2016), all values are for gross primary production. Net primary production values were used when GPP values were not found in the literature search and descriptions of net primary productivity calculations did not allow for back calculation to GPP values. Based on published values, this should result in a 33% underestimate at worst (Matsumoto, Abe, Fujiki, & Sukigara, 2016; Whittaker, 1975).

Grain Size

Characteristic grain sizes within habitats were plotted using a modified Phi (φ) scale (-8 for boulders to 10 for fine clays; Krumbein 1938), with additions of -9 for precipitated solids (i.e. corals and hydrothermal vent substrates) and -10 for solid rocks. If a given reference for GPP did not contain descriptions of local substrate, published grain sizes were used. In cases of mixed bottom types, grain size values were averaged to obtain a single grain size, φ, for each GPP observation.

Substrate Mobility

Sediment mobility is documented well for a range of unconsolidated substrates (e.g. Dalyander et al., 2015), but it is not reported broadly for the full range of marine substrates considered here. For this analysis we have extended and generalized the definition of substrate mobility, and populated values for the communities considered.

Here, substrate mobility was calculated as the difference between the critical shear stress (τcr) of sediments within systems of interest and the median bottom shear stress (τmed) experienced in that biome. Substrate mobility in this sense is a function expressing sediment mobility applied to a broader range of substrates that includes consolidated and non-solid materials.

Critical shear stress is specific for substrate type and is based on the substrate particle size, density differences between the substrate material and water, the kinematic viscosity of water, and then scaled with empirically developed coefficients (Dalyander, Butman, Sherwood, & Signell, 2015). Literature values for critical stress of granular sediments are available (Grabowski, Droppo, & Wharton, 2010; Julien, 1998). For larger sediments and consolidated sediments, values are based on estimates of erosion thresholds (Shan, Shen, Kilgore, & Kerenyi, 2015). Critical stress values for precipitate substrates were assumed to be similar to rough rocky substrates, unless fractured into smaller pieces (such as coral rubble) where critical stress was estimated based on grain size. Sea ice critical stress depends on thickness, configuration, and temperature, but ice is frequently cracked at values well below laboratory threshold tests (Mellor 1986). Therefore, critical stress for sea ice fell between boulder and cobble substrate values, but this value reflects only one estimate of expected critical stress.

Median stress levels were obtained through the U.S. Geological Survey Sea Floor Stress and Sediment Mobility Database (Dalyander et al. 2012, woodshole.er.usgs.gov/project-pages/mobility). Sediment characteristics for median stress values were taken from the USGS East-Coast Sediment Texture database (McMullen, Paskevich, & Poppe, 2014). Sediment data was available as a point layer within GIS, while Median Bottom Stress was a polygon shapefile.

As reported in the literature (e.g., Halley 1996), the frequency distributions for bed stress excitation is commonly expressed on an annual basis (i.e. number of events of a given size per year), but this has limited relevance in the intertidal or wave dominated systems. Maximum bed stress has been shown to be nearly three times higher in shoreline systems than offshore (Yüksel, Çevik, & Kapdanşli, 1998), thus multiple aspects of the distribution of wave forces are larger in the wave swept shore. It has been shown that for a given wave height distributed along a gradient in wave exposure, the greatest forces at the most wave exposed sites are approximately an order of magnitude higher (11.4 to 19.7 fold) than the forces as wave exposure approaches zero (Helmuth & Denny, 2003). Therefore, in wave exposed habitats we reduced the difference between bed stress and critical stress by an order of magnitude, or by one log unit.

Covariates: Human Impact

Human impact values were taken from mean cumulative impact scores (Halpern et al., 2008), and normalized by the fraction of the global ocean composed of that habitat type. The result, IEj , is a relative measure of net human impacts on each marine habitat type, j, in all locations E across the global ocean (see Appendix S4 of Cramer and Katz for index development).

Calculating Community Areas (Appendix S3)

To determine community areas, we used published values of global community distributions along with publicly available datasets. For communities characterized by their substrate, rather than fauna, we used the Carte Sédimentaire Mondiale v7.1 (Garlan, Gabelotaud, Lucas, & Marchès, 2018), a global map of seabed sediment using classifications analogous to the USGS Sediment Database. The Carte Sedimentaire Mondiale does not have a one to one relationship with sediment classifications – each polygon can have multiple sediment types. To assign each polygon a single sediment we tallied the sediment subtypes (i.e. sand, silt, clay) represented in each polygon. We assigned a polygon as “hard” if the number of represented hard subcategories were equal to or greater than the number of soft subcategories (see R script “AreaCalculations.R”).

For shoreline communities we assumed an area of 1 km2 for all intertidal zones. We used Dhanjal-Adams et al. 2016 as a case study to determine the ratio of intertidal habitats - mudflats, rocky shore, sandy shore - to overall coastline. We then determined the total global shoreline length using the Global Self-Consistent, Hierarchical, High-Resolution Geography Database to determine total global shoreline Version 2.3.7 (GSHHG Database) (Wessel & Smith, 1996).

We treated open water communities as slices, rather than cubes, of ocean habitat to allow for reasonable comparisons.

References

Arrigo, K. R., Worthen, D. L., Lizotte, M. P., Dixon, P., & Dieckmann, G. (2009). Primary Production in Antarctic Sea Ice. The Genome of the Diatom Thalassiosira Pseudonana: Ecology, Evolution, and Metabolism, 394(1997), 394–397. https://doi.org/10.1126/science.276.5311.394

Beaulieu, S. E. (2015). InterRidge Global Database of Active Submarine Hydrothermal Vent Fields: prepared for InterRidge, Version 3.3. Retrieved May 28, 2018, from http://vents-data.interridge.org/

Bourbonnais, A., Juniper, S. K., Butterfield, D. A., Devol, A. H., Kuypers, M. M. M., Lavik, G., … Lehmann, M. F. (2012). Activity and abundance of denitrifying bacteria in the subsurface biosphere of diffuse hydrothermal vents of the Juan de Fuca Ridge. Biogeosciences, 9(11), 4661–4678. https://doi.org/10.5194/bg-9-4661-2012

Copley, J. T. P., Tyler, P. A., Murton, B. J., & Van Dover, C. L. (1997). Spatial and interannual variation in the faunal distribution at Broken Spur vent field (29°N, Mid-Atlantic Ridge). Marine Biology, 129(4), 723–733. https://doi.org/10.1007/s002270050215

Dalyander, P. S., Butman, B., Sherwood, C. R., & Signell, R. P. (2015). Documentation of the U.S. Geological Survey Stress and Sediment Mobility Database (ver. 1.1, May 2015): U.S. Geological Survey Open-File Report 2012–1137. Retrieved from http://pubs.usgs.gov/of/2012/1137/

Dalyander, P. S., Butman, B., Sherwood, C. R., & Signell, R. P. (2012). U.S. Geological Survey Sea Floor Stress and Sediment Mobility Database. Retrieved from http://woodshole.er.usgs.gov/project-pages/mobility

Dennison, W. C. (1987). Effects of light on seagrass photosynthesis, growth and depth distribution. Aquatic Botany, 27(1), 15–26.

Duarte, C. M., Marbà, N., Gacia, E., Fourqurean, J. W., Beggins, J., Barrón, C., & Apostolaki, E. T. (2010). Seagrass community metabolism: Assessing the carbon sink capacity of seagrass meadows. Global Biogeochemical Cycles, 24(4), 1–8. https://doi.org/10.1029/2010GB003793

Fustec, A., Desbruyères, D., & Juniper, S. K. (1987). Deep-Sea hydrothermal vent communities at 13°N on the East Pacific Rise : microdistribution and temporal variations. Biological Oceanagraphy, 4(February), 37–41. https://doi.org/10.1080/01965581.1987.10749487

Garlan, T., Gabelotaud, I., Lucas, S., & Marchès, E. (2018). A World Map of Seabed Sediment Based on 50 Years of Knowledge. International Journal of Geological and Environmental Engineering, 12(6), 403–413.

Gebruk, A. V., Chevaldonné, P., Shank, T., Lutz, R. A., & Vrijenhoek, R. C. (2000). Deep-sea hydrothermal vent communities of the Logatchev area (14°45′N, Mid-Atlantic Ridge): diverse biotopes and high biomass. Journal of the Marine Biological Association of the United Kingdom, 80(3), 383–393. https://doi.org/10.1017/s0025315499002088

Germanovich, L. N., Hurt, R. S., Smith, J. E., Genc, G., & Lowell, R. P. (2015). Measuring fluid flow and heat output in seafloor hydrothermal environments. Journal of Geophysical Research: Solid Earth, 120(2), 8031–8055. https://doi.org/10.1002/2015JB012245.

Grabowski, R. C., Droppo, I. G., & Wharton, G. (2010). Estimation of critical shear stress from cohesive strength meter-. Limnology and Oceanography: Methods, 8, 678–685. https://doi.org/DOI 10:4319/lom.2010.8.678 678

Green, E. P., & Short, F. (2003). World atlas of seagrasses. Prepared by the UIMEP World Conservation Monitoring Centre. University of California Press, Berkeley, USA. (Vol. 47). https://doi.org/10.1515/BOT.2004.029

Grossi, S. M., Kottmeier, S. T., Moe, R. L., Taylor, G. T., & Sullivan, C. W. (1987). Sea Ice Microbial Communities .6. Growth and Primary Production in Bottom Ice Under Graded Snow Cover. Marine Ecology Progress Series, 35(1–2), 153–164. https://doi.org/10.3354/meps035153

Gruber, N., & Galloway, J. N. (2008). An Earth-system perspective of the global nitrogen cycle. Nature, 451(7176), 293–296. https://doi.org/10.1038/nature06592

Halley, J. M. (1996). Ecology, evolution and 1 f -noise. Trends in Ecology & Evolution, 11(1), 33–37. https://doi.org/10.1016/0169-5347(96)81067-6

Halpern, B. S., Walbridge, S., Selkoe, K., Kappel, C. V, Micheli, F., D’Agrosa, C., … Watson, R. (2008). A Global Map of Human Impact on Marine Ecosystems. Science, 319(February), 948–953.

Helmuth, B., & Denny, M. W. (2003). Predicting wave exposure in the rocky intertidal zone : Do bigger waves always lead to larger forces ? Limnology and Oceanography, 48(3), 1338–1345.

Knox, G. A. (2000). The Ecology of Seashores. Boca Raton: CRC Press.

Krumbein, W. C. (1938). Size frequency distributions of sediments and the normal phi curve. Journal of Sedimentary Petrology, 8(3), 84–90. https://doi.org/10.1046/j.1365-2745.2003.00770.x

Lizotte, M. P., & Sullivan, C. W. (1991). Photosynthesis-irradiance relationships in microalgae associated with Antarctic pack ice : evidence for in situ activity. Marine Ecology Progress Series, 71, 175–184.

Matsumoto, K., Abe, O., Fujiki, T., & Sukigara, C. (2016). Primary productivity at the time ‑ series stations in the northwestern Pacific Ocean : is the subtropical station unproductive ? Journal of Oceanography, 72(3), 359–371. https://doi.org/10.1007/s10872-016-0354-4

McCollom, T. M. (2000). Geochemical constraints on primary productivity in submarine hydrothermal vent plumes. Deep Sea Research Part I: Oceanographic Research Papers, 47(1), 85–101. https://doi.org/10.1016/S0967-0637(99)00048-5

McMullen, K., Paskevich, V., & Poppe, L. (2014). U.S. Geological Survey East-Coast Sediment Texture Database (ver. 3.0 November 2014). (L. Poppe, K. McMullen, S. Williams, & V. Paskevich, Eds.), USGS east-coast sediment analysis: Procedures, database, and GIS data. Retrieved from http://pubs.usgs.gov/of/2005/1001/.

Mellor, Malcolm. 1986. “Mechanical behavior of sea ice.” In Geophysics of Sea Ice, edited by N. Untersteiner, 146:165–281. New York: Plenum Press. https://doi.org/10.1017/CBO9781107415324.004.

Reed, D. C., Rassweiler, A., Arkema, K. K., & Barbara, S. (2008). Biomass rather than growth rate determines variation in net primary production by giant kelp. Ecology, 89(9), 2493–2505. https://doi.org/10.1890/07-1106.1

Rodgers, K. L., & Shears, N. T. (2016). Modelling kelp forest primary production using in situ photosynthesis, biomass and light measurements. Marine Ecology Progress Series, 553, 67–79.

Röttgers, R., McKee, D., & Utschig, C. (2014). Temperature and salinity correction coefficients for light absorption by water in the visible to infrared spectral region. Optics Express, 22(21), 25093. https://doi.org/10.1364/OE.22.025093

Shan, H., Shen, J., Kilgore, R., & Kerenyi, K. (2015). Scour in Cohesive Soils.

Wessel, P., & Smith, W. H. F. (1996). A global, self-consistent, hierarchical, high-resolution shoreline database. Journal of Geophysical Research, 101, 8741–8743. https://doi.org/10.1029/96jb00104

Winn, C. D., Cowen, J. P., & Karl, D. M. (1995). Microbes in deep-sea hydrothermal vent plumes. In D. M. Karl (Ed.), The Microbiology of Deep-Sea Hydrothermal Vents (pp. 255–273). Boca Raton, FL: CRC Press.

Yüksel, Y., Çevik, E. özkan, & Kapdanşli, S. (1998). Bed Shear Stress Distribution over Beach Profiles. Journal of Coastal Research, 14(3), 1044–1053.