ESRI Conservation Program: GIS Stories from the field

Point Reyes Bird Observatory

Point Reyes Bird Observatory: Over the last twenty five years, PRBO has conducted long-term ecological research using the life cycles of birds to understand ecosystem stresses and perturbations both natural and human induced. Birds provide "warning signals from nature;" within the PRBO area 411 of 600 identified north american bird species exist reflecting the areas unique geographical situation with diverse terrain, nearby nutrient upwellings, urban and agricultural areas. PRBO is also studying marine mammals along the central California coast, but PRBO studies are not limited to California; they extend from the arctic to the antarctic, and from the high Sierra crest to the open ocean.

Mapping Ocean Habitats (by Sarah Allen, PhD, from PRBO Observer Winter/Spring 1996) "When the Apex Houston freighter disgorged oil near the Golden Gate in 1986, one of the first questions asked was: "How and where in the ocean will marine animals be affected?" Unfortunately, this question was unanswerable at the time of the spill. As is often the case in the study of ocean creatures, we know much about their habits and habitats on shore but little of their marine requirements. Policy makers, resource managers, and research scientists universally acknowledge that information from most marine monitoring programs is insufficient for a genuine understanding of marine communities - or of the effects upon them of human activity. This is not surprising considering the ocean's vast expanse and dynamic nature, and the difficulty and expense of gathering data. Look at a satellite image of the ocean depicting sea-surface temperatures (see page 2), and you will see complexity, displayed in the swirls off promontories and of water colder or warmer than that of the larger ocean region. A geographic information system (gis) represents an excellent approach to understanding marine habitats - precisely because of the complexity and vast spatial area involved. Gis does not stand for "get in step" but might as well, since this sort of computer software is revolutionizing the way we access, analyze, and display data of all kinds. Put simply, it's a database management system that includes a spatial component. Picture a map depicting the spatial distribution of all the vegetation types on Point Reyes peninsula, then visualize several maps of the same area, each representing a separate variable (soil, roads, rivers, etc.). Imagine them printed on transparent film and stacked together: consider that all of this information is in digital form in a computer. You can take measurements between features of different maps, combine maps, create new maps from elements of each, or dump the data into a statistical program to perform analytical tests.

I was part of a prbo team, headed by David Ainley, that (with help from several cooperating scientists and agencies1) employed a geographic information system in a pioneering study of the Gulf of the Farallones. Our goal was to detect and define the persistent features of marine habitat important to seabirds, a highly visible group of species in the ecosystem. David Ainley saw the need for this information after observing the Farallon Island seabirds' dramatic change in diet (and possibly foraging locations) during the 1982-83 El Nio2 and later witnessing the Apex Houston spill. He initiated the study in 1985, before gis was widely available, and I joined the project in 1988, newly enthralled with the possibilities of applying gis to analyze spatial data.

We adopted three objectives: 1) identify persistent habitat variables that define seabird distribution and abundance at sea; 2) determine if there are temporal and spatial changes in distribution and abundance; and 3) based on the results of the first two, design spatial models of seabirds' habitat quality that are predictive. This would enable wildlife managers to anticipate the effects on seabird populations of a given activity in the Gulf of the Farallones.

The Gulf of the Farallones is one of only five eastern boundary currents in the world - highly productive ocean regions that support major seabird populations. Hydrographically and topographically complex, the Gulf's major large-scale marine features are plumes of upwelled water projecting from Point Reyes headland and Point Ano Nuevo, the San Francisco Bay freshwater plume extending seaward from the Golden Gate, and the shelf-break front, parallel to shore at 200 fathoms depth (the edge of the continental slope). Its primary terrestrial topographic features are Point Reyes headland, the Farallon Islands, the Golden Gate, and Ano Nuevo Island and Point Ano Nuevo. Major submarine features are Cordell Bank, the Monterey and Ascension undersea canyons, and the shelf break at 200 meters depth.

We spent ten years gathering data. During transects aboard the noaa vessel David Starr Jordan, we recorded seabird and marine mammal sightings from Monterey Bay to Bodega Bay every May and June from 1985 to 1994. (We conducted winter surveys during a few of these years, as well.) On the same cruises, we recorded environmental data every 15 minutes - sea surface temperature, salinity, depth, distance to land, and wind speed. We also measured a variable called conductivity-temperature-depth every 10 kilometers: this reveals the depth and slope of the thermocline, a subsurface feature that can concentrate prey.

The gis user becomes data-hungry, and I begged, borrowed, and bartered for data from many sources - bathymetry from the Environmental Protection Agency (epa), seabird nest sites from U.S. Fish and Wildlife Service, the shoreline from the State Lands Commission, and satellite images of sea surface temperature from the U.S. Weather Service. To detect correlations between environmental data and seabird distribution, I used advanced analytical tools. To detect spatial relationships and to develop habitat models, I used gis. Gis also enabled me to add variables to our analyses, based on comparing several data maps - for instance, the distance from a bird observed at sea to the nearest nesting colony, or to the shelf-break.

Findings: Only a few seabird species are predominant in the Gulf of the Farallones. These include the Sooty Shearwater, a non-breeding migrant, and breeding species such as Cassin's Auklet, Common Murre, Western Gull, Rhinoceros Auklet, and Ashy Storm-petrel. The Sooty Shearwater, which breeds in the southern hemisphere and migrates to central California to feed, was the most abundant species we found during upwelling years but was lower in abundance during El Nio years.

When analyzing the marine habitat relationships of seabirds, I selected 16 species (8 migrant and 8 nesting species) and found the most important variables correlated to their presence to be distance to land, distance to colony (for breeding species), and distance to shelf-break. Not surprisingly, seven of 8 nesting species were likelier to occur near colonies. Annual variation was also pronounced within our study. Compare 1986 with 1987 on Figures 1 and 2, for example: 1986 was an El Nino year, with lowered food productivity, and 1987 a good juvenile rockfish year and therefore a good food year for many seabirds. In 1986, more variables (including sea surface temperature) were correlated with the presence and abundance of species than in 1987, when the only significant variable was water depth.

(The best marine habitat (darkest squares) differs for the nearshore Common Murre (below) and the highly pelagic Ashy Storm-petrel (left). Paler gray to white squares indicate successively lower quality habitats.)

We also found strong differences between species in terms of presence and abundance. The highest densities of Common Murre, a piscivore, occurred in shelf waters near nest sites where there was a shallow thermocline and prey were concentrated, but Ashy Storm-petrels occurred mostly beyond the shelf in deep pelagic waters, where they forage on small organisms at the surface.

Even though these findings represent a breakthrough in describing features of seabirds' marine habitats, they relate nothing of the geographic distribution and abundance of species - and can even be misleading when taken alone. By themselves, our numbers showed Cassin's Auklets to occur in fairly shallow water, but if you look at the species' distribution map, the high density of sightings at Cordell Bank explains this preference for shallow water. Cordell Bank is an offshore seamount where upwelling concentrates food for pelagic birds and marine mammals.

With the addition of the spatial component, new relationships are discernible. Of several functions I performed with gis, two were most revealing: the simple overlay of sighting data with habitat variables and the development of predictive models of distribution. I began with simple distribution maps - where each seabird species was observed during each cruise - and overlaid these with the coastline and bathymetry data. Among the results: species segregate along the continental slope, the outer shelf, and nearshore; some species associate with certain topographic and hydrographic features (Cassin's Auklet with Cordell Bank, Brown Pelican with the warm water-plume of San Francisco Bay, Marbled Murrelet with nest sites at Ano Nuevo); and there were temporal and spatial changes in species distribution, particularly during El Nino years.

When using gis to model habitat quality, I drew upon variables that had proven significant in the foregoing steps - depth, distance to land, and distance to the nearest colony. Based on our at-sea data, I assigned each habitat variable a numeric value - high, medium, low, or no habitat suitability for each species. I then summed the variables and took the mean value to create a final map of habitat quality. In the case of the Common Murre, for example, I first classified ocean habitat as having high value if any part of the area was less than 10 km from any land mass. Land 20-80 km away gave the habitat low value. I repeated this step with the other two variables in the model, depth and distance to the nearest murre colony. I then combined the three data layers into a final map, showing the highest quality ocean habitat for Common Murre to be shallow water close to nesting colonies and also to any land mass.

These predictive models of habitat suitability can serve as working hypotheses against which to test future surveys. I was able to test our model against our own survey data from the years 1992-94, which were not included in the gis analysis. Overall, the model performed well: it proved 80% correct for the abundant species such as Cassin's Auklet and Common Murre. These are simple models, though, and with additional variables they likely will become even more accurate.

As they are, the models are important precedents for defining marine habitats for seabirds - and useful now for evaluating coastal zone management issues. For instance, the epa proposed four offshore sites for dumping dredge spoils from San Francisco Bay. A cursory look at the habitat models revealed which species might be affected by which dump site. The sites occur in mostly low-quality habitat for the Common Murre but in the heart of high-quality habitat for the more pelagic Ashy Storm-petrel. The disposal site was chosen on the basis of this and other information.3 The most exciting application of these models is to real-time events such as oil spills, fisheries conflicts (the location of gill nets), and water policy in the Bay/Delta that affects the San Francisco Bay plume.

The applications for gis are endless. Only the available datasets - and the boundaries of our imagination - limit what we can do with this powerful research and management tool.

Notes: 1. Several people, institutions and agencies supported the project, including the members of prbo. The National Marine Fisheries Service, in particular Bill Lenarz, donated weeks of ship time for surveys. Other major supporters of the project were R. Barrett and I. Timossi, U.C. Berkeley, who donated hours of computer time and advice; the Gulf of the Farallones National Marine Sanctuary; and epa, which funded two years of sea time and initial analyses. 2. See Observer 62, Summer 1983, for an explanation of El Nino, which disrupts the cycle of cold-water upwelling and productivity in the California Current. 3. Prbo recently completed an analysis of contaminants in the marine environment to serve as baseline for future studies of any effects from the dredge spoils dump site. See "Director's Report" in our 1995 Annual Report (included with this Observer) for a summary.

About the Author: Sarah Allen has long pursued interests in marine wildlife, beginning some 20 years ago with harbor seal studies at Point Reyes, as a PRBO field biologist. After working on the administrative staff at PRBO, she went on to advanced studies at U.C. Berkeley. For her doctorate, Sarah concentrated on the research explained in this article. A PRBO research associate, Sarah holds a scientific post (wildlife specialist) with the National Park Service. Her GIS expertise was of great value in assessing the October 1995 Mount Vision fire in Point Reyes National Seashore (see Observer 105, Fall 1995).

Text and graphics: Point Reyes Bird Observatory and Dr. Sarah Allen
January 2, 1997 , (Point Reyes Bird Observatory, 4990 Shoreline Highway, Stinson Beach, California 94970, Email:, tel:415-868-1221, fax:415-868-1946)

Web layout & design: ESRI Conservation Program, January 2, 1996

ESRI Conservation Program Home * ECP E-Mail (

|   Software - Data - Books - Consulting - GIS Store    |    News - Events - Careers - About ESRI   |
|   Support - GIS Education & Training - GIS for Your Specialty - Partner Solutions - Library   |

Copyright © Environmental Systems Research Institute, Inc.
Send your comments to:
ESRI Web Site Privacy Policy
July 19, 2000