[an error occurred while processing this directive][an error occurred while processing this directive]
Mapping Benthic Habitats; The Marine GIS Challenge
Mapping the seafloor is getting easier, but finding the data is still sometimes like hunting for an octopus in a coral reef. Hard-bottom and reef regions of the marine ecosystem provide essential habitat for a wide variety of marine species, including sea turtles, lobsters, and an abundant array of fish and invertebrates. The effects of commercial and recreational fishing, as well as other anthropogenic pressures, can threaten and change the long-term viability of fish populations. Study and management of these effects are problematic because of the difficulties involved in assessing fish population size and flux.
The influence of human activity, including the damaging effect of anchors, trawling, the use of explosives for fishing, and large amounts of runoff and coastal pollution, has measurable negative effects. Knowledge of the location and extent of critical habitat allows researchers and managers to track more accurately the effects of fishing and, thus, more effectively protect essential areas. This knowledge is crucial to the protection of reef-type habitats and the practice of sustainable fish harvesting. But how is critical habitat determined? And once the assessment has been made that a marine habitat supports enough biodiversity to be considered critical, how can the seafloor be mapped using GIS?
Assessing Benthic Complexity
Many studies have tackled the question: What makes a seafloor livable? One of these studies from the work of Jeff Ardron of the Living Oceans Society, Sointula, British Columbia, Canada, shows that the physical complexity of the seafloor is a determining factor. A complex seafloor rich with features that contribute to the terrain is more likely to harbor life-forms of all shapes and sizes. These heterogeneous habitats that support multiple life-forms are often associated with species richness. The slope of the terrain; its relief; and, most important, its complexity are all contributing factors to measuring the likelihood of multiple species habitation. Raster calculation tools in the ArcGIS Spatial Analyst extension in conjunction with ArcInfo are used to perform analysis on ecological distribution, and the resulting layers are new information derived from the data of the original bathymetry.
The bathymetric layer in GIS usually starts as a large set of points. These points are collected most often using multibeam echo soundings or side-scan sonar devices. To the marine GIS user, the resulting data generated by soundings is simply a set of x,y,z coordinates, and rather simple ones at that, because they only have one associated z value for each x,y and usually occur at one instance in time. From the table of x,y,z values, a set of points can be created; with this set of points, the interpolation begins.
With databases now storing tens of millions of points from echo sounding surveys, the challenge begins with the need for mass point data storage, a rapid update of newly surveyed points, and the visual display and representation of the surface created from the points. The points are usually a dense set of dispersed known values. The interpolation tools that can be found in the ArcGIS Spatial Analyst extension give the user the ability to predict values that are then assigned to all other locations on a cell-by-cell basis. Input points can be either randomly or regularly spaced or based on some sampling scheme; in the case of echo sounding data, it is most often a track line created by the sounding device, or fish.
[Editor's note: Since the publication of this article, the ArcGIS 3D Analyst extension has released the concept of Terrain Datasets. This new functionality supports the storage and maintenance of billions of elevation points and breaklines, and exposes the data as a TIN-based, multi-resolution surface. Multibeam sonar data is an excellent candidate for being managed in the geodatabase inside a Terrain Dataset.]
Creating the Seafloor
Surface interpolation functions create a continuous, or prediction, surface from sampled point locations, generating a raster data set with values for all the cells whether a measurement has been taken at the location or not. There are many ways to derive a prediction surface using a variety of calculations and different assumptions that are made of the data. Inverse Distance Weighted and Spline are two of the methods used to assign values to locations based on the surrounding measured values, using mathematical formulas that determine the smoothness of the resulting surface. There are also geostatistical methods found in the ArcGIS Geostatistical Analyst extension, such as kriging, which is based on statistical autocorrelation or the statistical relationship between the points. By kriging, users have the capacity to produce not only the prediction surface but also a measure of confidence in the accuracy of the predictions. The resulting bathymetric surface is often only the beginning of the layers that are added to the marine GIS.
Important to the overall process of creating a valid "seascape" is having a well-designed data model. The Bottom-Mapping Workgroup of the Florida Marine Research Institute (FMRI) developed a format for survey data covering the South Atlantic Bight (SAB) database, prior to beginning data collection by the National Marine Fisheries Service and South Atlantic Fisheries Management Council. Historical sources were then identified, analyzed, and considered for inclusion in the SAB database. Protocols for determining positional accuracy and deducing bottom type, based on the type of gear and the method used in collecting each sample, were established. Gear types and methods used in these historical surveys included side-scan sonar, vibra core, aerial photography, dredge, trawl, and trap. The work group also established protocols for tabulating the various data sources and relating them to the primary bottom records.
This process resulted in an effective protocol for determining the locations of hard-bottom habitat: examination of trawl and trap data for the presence of those fish species considered to be indicators of the presence of reef-type habitat. To further complicate the issue, several species are known to use different habitats at different life stages. For example, many species will use grass beds or pelagic habitats as nursery areas but are considered reef species as adults.
Marine data and the resulting ArcGIS 3D Analyst (ArcScene and ArcGlobe) visual representation can range from the standard point, line, and polygon models to a more complex combination of images and seafloor rasters and a corresponding wire frame or "fishnet" mesh. Each transect and the related data collected in a marine survey, in addition to the sonar, and other observation data can be used to assign a bottom type to the grid cell. A reef or hard-bottom assemblage may intersect several raster grid cells while each transect, or area, maintains its identification as a single entity in the database. The result is an expansive grid of one-minute cells covering SAB, with each cell recording the number of data records and bottom type that occurred within the cell and, through the database, tying back to the original data source. Once the modeling of a database for seafloor parameters has been achieved, the architecture and design of a solution to the continual updating of the database with dynamic multidimensional data are the next challenges.
For more information, contact Joe Breman, Esri (e-mail: email@example.com). This article is derived from chapters by Douglas Wilder, Henry Norris, and Jeff Ardron in the Esri Press book Marine Geography: GIS for the Oceans and Seas (ISBN: 1-58948-045-7) edited by Joe Breman. Esri Press books are available at better bookstores, online (www.esri.com/esripress), or by calling 1-800-447-9778. Outside the United States, contact your local Esri distributor.