The Seascapes

The Seascapes

Thursday, February 28, 2013

OpenOcean 2013. Does our habitat paradigm cross the land-sea boundary? A call for the development of IOOS informed seascape ecology supporting ecosystem management

John P. Manderson1 and Josh T. Kohut2
The co-authors contributed equally to the development of this white paper submitted for 2012 National Workshops for Habitat Assessment (09/05/12) & Integrated Ocean Observation Systems (11/13/12)
1NOAA/NMFS/NEFSC James J. Howard Laboratory, Highlands, NJ 07732
2Rutgers, The State University of New Jersey, New Brunswick, NJ 08901

“..comparison of marine and terrestrial dynamics has more than theoretical interest. As we utilize marine and terrestrial environments, the consequences, deliberate or accidental, depend on [ecosystem] responses to physical and chemical change. The imposition of terrestrial standards for marine problems may produce too strict or too lax criteria--or most likely quite inappropriate ones” (John Steele, 1991. Can ecological theory cross the land-sea boundary?)

1. Introduction
            Ecosystem assessment and management in the sea is holistic, based upon interdisciplinary science that considers physical, chemical and biological processes, including feedbacks with human ecological systems, that structure and regulate marine ecosystems.  Space and time based tools for the management of human activities in the sea need to be informed by a regional scale habitat ecology that reflects the dynamic realities of the ocean.  Current spatial management strategies including marine spatial planning (MSP) and ocean zoning are based upon the patch-mosaic paradigm of terrestrial landscape ecology modified to consider principles of dispersal ecology, primarily for pelagic early life history stages of algae, fish and invertebrates.  This modification is not enough because fundamental differences in the role fluid properties and processes play in controlling ecological processes on land and in the sea makes blanket application of paradigms developed on land to the problems of ocean management fundamentally flawed.
  The rapid evolution of the Integrated Ocean Observation System (IOOS) made possible through interdisciplinary partnerships and networked data sharing provides descriptions of coastal ocean hydrography and hydrodynamics at fine scales of space and time and regional spatial extents.  This allows hydrography and hydrodynamics to be placed at the foundation of a seascape ecology in the way that geography and geophysics appropriately serve as the foundation of terrestrial landscape ecology.   IOOS not only provides the ocean data required to develop seascape ecology but also the infrastructure and expertise to operationalize it for regional ecosystem assessment and adaptive co-management. Finally, regional IOOS associations are cooperative partnerships of academic, government and private industry experts from diverse fields and interests. As a result the IOOS "culture" can foster the collaborative development of an interdisciplinary seascape science that is more likely to lead to effective and less adversarial strategies of regional ecosystem co-management that operate at space-time scales more closely matching those of the ecosystem itself.

2. Seascapes are not landscapes
            In 1984 Paul Risser and colleagues (Risser et al, 1984) summarized workshop deliberations to develop a modern framework for the science of landscape ecology using theoretical and empirical underpinnings of a broad scale spatially explicit ecology useful for terrestrial resource management.  The foundations of this synthesis rested on (1) developments in satellite remote sensing that allowed researchers to place fine scale ecological studies in broader spatial contexts; (2) advances in ecological theory that elucidated the role of dispersal and connectivity in determining regional community dynamics; and (3) the advent of modern computing that allowed researchers to store, analyze, and model large amounts of spatially and temporally explicit data and explore relationships between the changing landscape patterns and the processes potentially causing them.  Landscape ecology rests primarily upon the patch mosaic paradigm of habitat in which patches are defined by sharp gradients in vegetation and geomorphology (but see Cushman et al. 2010).  Geography and geological processes, particularly soil development, that control fundamental processes including primary productivity are the foundations of landscape classification. In terrestrial ecosystems, most organisms and processes are decoupled from the atmosphere by gravity and physiological adaptation to extreme variations in atmospheric properties, including temperature.  As a result, the primary features of terrestrial habitats and ecosystems are physical structures created by landform and plant communities that can be modified by disturbance. Community compositions are determined by climate (Wiens, 2007).  However, the role of the atmospheric fluid is of secondary importance and the space-time scales of terrestrial ecosystems [~ velocity, 0.1 cm sec-1] are orders of magnitude slower than the atmosphere [100 cm sec-1] and approximately the same speed as soil regeneration (Steele, 1991; Mamayev, 1996. Fig 1).

Figure 1. Ecosystems on land (above) operate at space-time scales orders of magnitude slower than turbulent features of the atmosphere while variability in marine ecosystems (below) matches the scales of variability of turbulence in the ocean fluid.  (From Steele, 1991 and Mamayev, 1996)

In contrast, the ocean is highly viscous and has a density close to that of living tissues.  Most marine organisms are therefore nearly buoyant in a fluid with dynamics that control their motions and those of other important particles including essential ecosystem building blocks.  Since the basic processes of cellular metabolism evolved in the sea, most living tissues are nearly isosmotic with seawater. This contrasts starkly with terrestrial organisms whose intracellular concentrations of solutes and water are dramatically different than the atmosphere.  Finally the specific heat capacity and thermal conductivity of seawater are about four and twenty-three times that of atmosphere by weight, respectively.  As a result, marine organisms experience much slower rates and ranges of temperature change than do terrestrial organisms. Temperature is tyrannical in the oceans where oxygen required for endothermic heat generation is limited and warm-blooded organisms are rare. Temperature regulates critical rates across all levels of ecological organization from the cell to marine ecosystems.
            Processes controlling primary productivity on land and the sea are also fundamentally different.  In the oceans nutrients required by plants constant fall out of sunlit surface waters where photosynthesis is possible.  As a result, tiny, fast living plants with high surface to volume ratios are entirely dependent on the oceans “plumbing” to deliver nutrients into the sunlit surface layers from sometimes remote land or deep waters sources. Phytoplankton have fast population dynamics to which other members of marine food webs must respond. In contrast, primary productivity on land depends on slow, local nutrient regeneration in soil at the interface with a nearly transparent atmosphere where sunlight is rarely in short supply.   As a result, plants at the base of terrestrial food web are often immobile, long lived, and have slow population dynamics to which higher trophic levels respond.
             Due to the tight coupling of physiology, movement of organisms and other critical ecological processes to the oceans fluid, the fluid is the primary driver structuring seascapes and regulating seascape processes. As a result, ecological processes in the ocean operate at approximately the same space-time scales (~velocities [~1 cm sec-1]) as ocean turbulence (Steele 1991, & Mamayev, 1996).  Bottom features are important to some marine organisms. However the functional importance of bottom features as surfaces concentrating advected materials, sites of energy acquisition and/or conservation in the contexts of  fluid flows, predation refugia in regions where preference for water properties such as temperature, salinity and oxygen are shared between predator and prey; are frequently primarily defined by fluid processes and properties.  
In summary, differences in the nature of the ocean and atmospheric fluid and adaptations of organisms to those fluids produce at least two critical differences in the characteristics of seascapes and landscapes.  Firstly, habitats in the sea have much faster spatial dynamics: their locations, volumes and quality change quickly at rates defined by the space-time scales of organisms responses to highly dynamic properties and processes of the oceans fluid that are driven in turn by atmospheric and planetary forcing in the form of percipiation, wind, temperature and tides. Secondly, because the ocean fluid is so viscous, horizontal and vertical currents driven by atmospheric and planetary forcing, transport essential habitat resources from sometimes remote sources and concentrate them in particular areas and times. In such cases habitats are not just locations in space supported by local resources, but nodes of networked resources and processes many of which are derived from distant "upstream" sources .  For these and other reasons relationships between habitat dynamics and processes regulating populations, including density dependent processes, are fundamentally different in the sea and on land.   Differences in the nature of habitat in the ocean and on land are in fact responsible for the order of magnitude differences in rates of change in species distribution and abundance in the sea and on land (~10 km yr-1 vs ~ 1 km yr-1) associated with recent rapid changes in climate (Chueng et al., 2009; Sorte et al., 2010).

3. The role of IOOS in seascape ecology 

            The presence of Integrated Ocean Observation Systems IOOS, including its infrastructure, data, models and the expertise of its diverse partners, now allows for the development of seascape ecology with a regional spatial scope that reflects the realities of the ocean.  Like landscape ecology's modern synthesis (Risser, 1984), an IOOS informed seascape ecology could provide the theoretical and empirical underpinnings for the broad scale spatially and temporally explicit ecology required for the regional assessment and management of ocean resources.   Seascape science will integrate the fields of fisheries oceanography, marine habitat ecology, and ecosystem science with hydrography and hydrodynamics at its foundation, just as terrestrial landscape ecology rests appropriately on the foundation of geography and geophysics.


4. Toward an IOOS informed seascape ecology

            With the support of the NOAA office of Science and Technology and North East Fisheries Science Center Cooperative Research Program we have taken advantage of the IOOS collaborative culture to form an interdisciplinary workgroup of habitat scientists, oceanographers, fishery managers, social scientists, and fishermen from academia, government and industry to develop ecologically informed habitat models for the purpose of addressing issues of related to the dynamics and assessment of habitats and populations of butterfish (Peprilus triacanthus) and longfin squid (Doryteuthis pealeii) fishery.  We held workshops to combine scientists’ and fishermen’s knowledge into a single model of butterfish habitat made using National Marine Fisheries Service (NOAA/NMFS) surveys of organisms and hydrography, and  satellite and high-frequency radar measurements of ocean properties and processes provided by the Mid-Atlantic Regional Association Coastal Ocean Observing System (MARACOOS), a regional component of IOOS (Manderson et al., 2011).  During the workshops scientists and fishermen each made “mental models” that included environmental variables each group considered important in defining habitat. Over lunch the two models were constructed and evaluated using cross-validation in a butterfish "smackdown".  Following lunch we discussed results of the “smackdown” and the ecological mechanisms potentially responsible for habitat associations described by scientists and fishermen.  Following the workshops we developed the model that combined the ecological expertise of fisherman and scientists.

            Once the combined model was complete, we worked with fishing industry partners to design an at-sea model evaluation using dynamic habitat model nowcasts provided by MARACOOS.  The combined model was adjusted slightly to include variables that has slow spatial dynamics (seabed rugosity) and fast dynamics (Sea surface temperature, water mass frontal boundaries) that could be delivered in real-time by the ocean observing system.
Figure 2.  Butterfish habitat preference predicted by the model during our 8-day evaluation.  The warmer colors indicate areas of preferred habitat.  The vessel track is shown in green.  

            During an 8-day trip on the F/V Karen Elizabeth, captained by Chris Roebuck, we transmitted updated dynamic butterfish habitat model nowcasts to the vessel (Fig 2). Our survey design involved sampling areas the nowcasts predicted “habitat suitability” would be  “high” and “low”. In each 3 station set we also included a site where the fisherman, Captain Roebuck, predicted butterfish would be abundant.  We sampled station sets for fish and the environment during the day and night in three canyon hotspots identified by the model along the edge of the Mid Atlantic Bight continental shelf. Using this approach we were able to formally incorporate fishermen’s knowledge into the design of our field evaluation survey. Throughout the evaluation the crew on board the Karen Elizabeth sent reports of preliminary results back to shore which we published on an online blog.

            The evaluation survey showed us that the combined model could be used to identify regions and times when butterfish concentrations were likely to be high at scales of 10s of kilometers.  We learned that fishermen understood species-habitat associations at scales much finer than could be described by the data used to construct the model and thus the model itself. Fishermen also knew locations and times where the animals were likely to occur that are not typically sampled on scientific surveys including those used in population assessments.

            We are further refining our mesoscale model with the help of the fishing industry for the recalibration of indices of population trend based upon the amount of habitat sampled in fisheries independent surveys.  We are also designing prototype adaptive industry based surveys of dynamic habitat guided by meso-scale habitat models that are intended to supplement information collected on regional scale fishery-independent population assessment surveys. These applications may prove especially useful for estimating population trends of ecosystem keystone species when rapid changes in climate are causing dramatic changes in fluid properties and processes and thus in the spatial dynamics of ocean habitats.

5. The Next Decade

Figure 3.  Thermal niche model based on metabolic theory parameterized for butterfish which we coupled to daily Regional Ocean Model (ROMS) hindcasts of bottom water temperature in 2006 for the North West Atlantic (Cape Hatteras [lower left] to Canada [top right])  We are currently using similar prototype models to assess changes in the proportion of habitat available in the ecosystem and sampled in each year during population assessment surveys, design adaptive fields surveys, and to better understand the relationships between habitat dynamics and population dynamics for mobile species that thermoregulate by tracking the thermal dynamics of the seascape.

            We are beginning to move beyond empirical ecological models based upon regional fisheries data and observations toward mechanistic ecological models that can be coupled to IOOS assimilative oceanographic models describing critical features of ocean habitats (Fig. 3).  Coupled mechanistic biophysical models will allow us to describe dynamic ocean habitats throughout the water column and avoid pitfalls associated with using correlative empirical models for forecasting. Using oceanographic models will also allow us to investigate the role of advection in delivering key habitat building blocks from sometimes remote sources to locations and times where/when ocean habitats form. Mechanistic seascape models that rest on the foundations of assimilative hydrodynamic models will be particularly useful if climate change produces ocean conditions we have never before observed.
We intend to continue to work with fishermen within the context of the IOOS collaborative culture. Integrating their practical ecological knowledge with academic knowledge of the sea should result the rapid development of accurate seascape models.  These models will first be considered hypotheses that can be adaptively tested within ocean observing systems.  Once vetted in this way they can be easily operationalized as tools for the space and time management of human activities in dynamic ocean ecosystems. We believe this adaptive, iterative, collaborative approach is the cost effective way to develop a seascape ecology with a scope broad enough to meet requirements for resource management in the sea.

6. Conclusion
Rapid changes in human demand and use patterns of marine resources combined with the profound effects climate change is having on species distributions and the structure of marine ecosystems have made the development of a regional scale seascape ecology reflecting the dynamic realities of the ocean urgent. The foundations of the landscape ecology synthesis in the early 1980s rested on developments in satellite remote sensing, advances in spatial ecology and the advent of modern computing. The recent development of operational ocean observing systems that integrate assimilative hydrodynamic models, and observations from remote sensing and insitu platforms along with important advances in our understanding of micro to macro-ecological process in the sea have made the time ripe for a similar synthesis and the development of a robust science of seascape ecology useful for the management of marine ecosystems.

Chueng, W. W. L., Lam V. W. Y., Sarmiento J. L., Kearney K., Watso R., and Pauly D. 2009. Projecting global marine biodiversity impacts under climate change scenarios. FISH and FISHERIES 2-14.

Cushman, S. A. E., Jeffrey S., McGarigal K., and Kiesecker J. M. 2010. Toward Gleasonian landscape ecology: From communities to species, from patches to pixels.  . 12 p.

Mamayev OI (1996) On Space Time Scales of Oceanic and Atmospheric Processes. Oceanology 36(6): 731-734

Manderson, J.M., L. Palamara, J. Kohut, and M. Oliver.  2011.  Ocean observatory data are useful for regional habitat modeling of species with different vertical habitat preferences.  MEPS.  Vol. 438: 1–17, doi: 10.3354/meps09308

Risser PG, Karr JR, Forman R (1984) Landscape ecology: directions and approaches. Illinois Natural History Survey Special Publication # 2. Champaign.  17pp.

Sorte, C. J. B., Williams S. L., and Carlton J. T. 2010. Marine range shifts and species introductions: comparative spread rates and community impacts. Global Ecology and Biogeography 19:303-316.

Steele JH (1991) Can ecological theory cross the land-sea boundary? Journal of Theoretical Biology 153:425-436

Weins John A. 2007 Foundation Papers in Landscape Ecology. ‪Columbia University Press‬, 2007