In ecology, a common form of statistical analysis relates a biological variable to variables that delineate the physical environment, typically by fitting a regression model or one of its extensions. Unfortunately, the biological data and the physical data are frequently obtained from eparate sources of data. In such cases there is no guarantee that the biological and physical data are co-located and the regression model cannot be used. A common and pragmatic solution is to predict the physical variables at the locations of the biological variables and then to use the predictions as if they were observations.We show that this procedure can cause potentially misleading inferences and we use generalized linear models as an example. We propose a Berkson error model which overcomes the limitations. The differences between using predicted covariates and the Berkson error model are illustrated by using data from the marine environment, and a simulation study based on these data.

A number of terms used in this book are derived from the fields of biogeography and benthic ecology and these are defined in the glossary; the reader is also referred to the works cited at the end of this chapter for further information. Many of the case studies presented in this book refer to habitat classification schemes that have been developed based on principles of biogeography and ecology. For these reasons a brief overview is provided here to explain the concepts of biodiversity, biogeography and benthic ecology that are most relevant to habitat mapping and classification. Of particular relevance is that these concepts underpin classification schemes employed by GeoHab scientists in mapping habitats and other bioregions. A selection of published schemes, from both deep and shallow water environments, are reviewed and their similarities and differences are examined.

The World Summit on Sustainable Development implementation plan requires, by 2012, a representative system of marine protected areas (RSMPA) for the purposes of long-term conservation of marine biodiversity. A great challenge for meeting this goal, particularly in data-poor regions, is to avoid inadvertant failure while giving science the time and resources to provide better knowledge. A staged process is needed for identifying areas in data-poor regions that would enable the objectives to be achieved in the long term. We elaborate a procedure that would satisfy the first stage of identifying a RSMPA, including areas suitable as climate change refugia and as reference areas for monitoring change without direct interference of human activities. The procedure is based on the principles of systematic conservation planning. The first step involves the identification of ecologically-separated provinces along with the physical heterogeneity of habitats within those provinces. Ecological theory is then used to identify the scale and placement of MPAs, aiming to be the minimum spatial requirements that would satisfy the principles for a representative system: comprehensiveness, adequacy and representativeness (CAR). We apply the procedure to eastern Antarctica, a region with spatially-restricted sampling of most biota. We use widely available satellite and model data to identify a number of large areas that are likely to encompass important areas for inclusion in a RSMPA. Three large areas are identified for their pelagic and benthic values as well as their suitability as climate change refugia and reference areas. Four other areas are identified specifically for their benthic values. These areas would need to be managed to maintain these values but we would expect them to be refined over time as more knowledge becomes available on the specific location and spatial extent of those values.

Physical sedimentological processes such as the mobilisation and transport of shelf sediments during extreme storm events give rise to disturbances that characterise many shelf ecosystems. The intermediate disturbance hypothesis predicts that biodiversity is controlled by the frequency of disturbance events, their spatial extent and the amount of time required for ecological succession. A review of available literature suggests that periods of ecological succession in shelf environments range from 1 to over 10 years. Physical sedimentological processes operating on continental shelves having this same return frequency include synoptic storms, eddies shed from intruding ocean currents and extreme storm events (cyclones, typhoons and hurricanes). Modelling studies that characterise the Australian continental shelf in terms of bed stress due to tides, waves and ocean currents were used here to create a map of ecological disturbance, defined as occurring when the Shield's parameter exceeds a threshold of 0.25. We also define a dimensionless ecological disturbance ratio (ED) as the rate of ecological succession divided by the recurrence interval of disturbance events. The results illustrate that on the outer part of Australia's southern, wave-dominated shelf the mean number of days between threshold events that the Shield's parameter exceeds 0.25 is several hundred days.

Marine physical and geochemical data can be valuable in predicting the potential distributions and assemblages of marine species, acting as surrogate measures of biodiversity. The results of surrogacy analysis can also be useful for identifying ecological processes that link physical environmental attributes to the distribution of seabed biota. This paper reports the results of a surrogacy study in Jervis Bay, a shallow-water, sandy marine embayment in south-eastern Australia. A wide range of high-resolution co-located physical and biological data were employed, including multibeam bathymetry and backscatter data and their derivatives, parameters that describe seabed sediment and water column physical characteristics, seabed exposure, and infauna species. The study applied three decision tree models and a robust model selection process. The results show that the model performance for three diversity indices and seven out of eight infauna species range from acceptable to good. Important surrogates for infauna diversity and species distributions within the mapped area are broad-scale habitat type, seabed exposure, sediment nutrient status, and seabed rugosity and heterogeneity. The results demonstrate that abiotic environmental parameters of a sandy embayment can be used to effectively predict infauna species distributions and biodiversity patterns. International Journal of Geographical Information Science

Submerged relict reef systems and modern coral communities discovered around the Balls Pyramid shelf are presented as new evidence of extensive carbonate production at the boundary of reef-forming seas. Balls Pyramid is the southernmost island in a chain of island-reefs in the southwest Pacific Ocean, 24 km south of the southernmost known coral reef in the Pacific Ocean at Lord Howe Island. This paper explores the detailed geomorphic structure of the shelf through the production of a high resolution bathymetric model from multibeam echosounder data and depth estimates from satellite imagery. Key seafloor features identified include a large, mid shelf reef dominating the shelf landscape in 20 - 60 m water depth, mid shelf basins and channels, and shelf margin terrace sequences in 50 - 100 m depth. Sub-bottom profiles, backscatter, drill core and vibro-core data are used to investigate the seafloor composition. Drill cores extracted from the submerged reef surface confirm coral, coralline algae and cemented sands composition, and vibro-core material extracted from unconsolidated areas demonstrate substantial accumulation of carbonates shed from the reef surface. Underwater video imagery reveals abundant modern mesophotic reef communities, including hard corals, colonising the relict reef surface. This paper reveals prolific past reef growth and abundant modern coral growth on what was previously considered to be a planated volcanic shelf outside of reef-forming seas, thus extending understanding of reef evolution at, and beyond, the limits of growth.

This dataset contains sediment and geochemistry information for the Oceanic Shoals Commonwealth Marine Reserve (CMR) in the Timor Sea collected by Geoscience Australia during September and October 2012 on RV Solander (survey GA0339/SOL5650). Further information on the survey is available in the post-survey report published as Geoscience Australia Record 2013/38: Nichol, S.L., Howard, F.J.F., Kool, J., Stowar, M., Bouchet, P., Radke, L., Siwabessy, J., Przeslawski, R., Picard, K., Alvarez de Glasby, B., Colquhoun, J., Letessier, T. & Heyward, A. 2013. Oceanic Shoals Commonwealth Marine Reserve (Timor Sea) Biodiversity Survey: GA0339/SOL5650 - Post Survey Report. Record 2013/38. Geoscience Australia: Canberra. (GEOCAT #76658).

This study presents new information on the regional geochemical characteristics of deep-sea floor sediments (1300 - 2423 m water depth) on the Lord Howe Rise (deep-sea plateau) and Gifford Guyot (seamount/tablemount), remote areas off eastern Australia. The aim was to provide a coherent synthesis for a suite of geochemical data that can be used to make habitat inferences and to develop surrogates of biodiversity. Sediment characteristics analysed were mineralogy, organic carbon and nitrogen concentrations and isotopic compositions, and concentrations of major and trace elements. We also measured parameters that convey information about the reactivity of organic matter and on the bio-availability of bioactive trace elements (e.g. chlorin indices and acid-extractable elements). Surface sediments from the region were calcareous oozes that were carbon-lean (0.26±0.1%) and had moderate to high chlorin indices (0.62 - 0.97)..

This study tested the performance of 16 species models in predicting the distribution of sponges on the Australian continental shelf using a common set of environmental variables. The models included traditional regression and more recently developed machine learning models. The results demonstrate that the spatial distributions of sponge as a species group can be successfully predicted. A new method of deriving pseudo-absence data (weighted pseudo-absence) was compared with random pseudo-absence data - the new data were able to improve modelling performance for all the models both in terms of statistics (~10%) and in the predicted spatial distributions. Overall, machine learning models achieved the best prediction performance. The direct variable of bottom water temperature and the resource variables that describe bottom water nutrient status were found to be useful surrogates for sponge distribution at the broad regional scale. This study demonstrates that predictive modelling techniques can enhance our understanding of processes that influence spatial patterns of benthic marine biodiversity. Ecological Informatics

Marine benthic biodiversity can be quantified using a range of sampling methods, including benthic sleds or trawls, grabs, and imaging systems, each of which targets a particular community or habitat. Research studies often incorporate only one of these sampling methods in published results, and the generality of marine biodiversity patterns identified among different sampling methods remains unknown. In this study we use three biological collections obtained during a collaborative survey between Geoscience Australian and the Australian Institute of Marine Science to the Van Diemen Rise in northern Australia: 1) Infauna sampled from a Smith-McIntyre grab, 2) Epifauna sampled from a benthic sled, and 3) Biological communities identified from video. For each dataset, we investigated potential patterns of species richness and community structure in relation to depth, geomorphology, and study area, as well as the relationships between datasets. No gear type yielded data that was strongly correlated with depth, but different patterns were evident among gear types based on study area and geomorphology. Comparisons among datasets indicate that species richness from sleds and grabs are more strongly correlated with each other than with richness from video. Further research is planned to incorporate datasets from other regions and habitats in order to provide a general assessment of sampling methods used in the quantification of benthic marine biodiversity in Australasia.