International Journal of Remote Sensing Vol. 26, No. 12, 20 June 2005, 2631–2656 Review article Thirty years of analysing and modelling avian habitat relationships using satellite imagery data: a review T. K. GOTTSCHALK*{, F. HUETTMANN{ and M. EHLERS§ {Justus-Liebig-University Giessen, Research Centre for Bio Systems, Land Use and Nutrition, Department of Animal Ecology, Heinrich-Buff-Ring 26–32, 35398 Giessen, Germany {University of Alaska Fairbanks, EWAHALE Laboratory, Biology and Wildlife Department, Institute of Arctic Biology, Alaska 99775-7000, USA §University of Vechta, Institute for Environmental Sciences, P.O. Box 1553, 49364 Vechta, Germany (Received 25 July 2003; in final form 13 October 2004 ) The application of remote sensing and Geographic Information Systems (GIS) technologies provides powerful tools when used to investigate wildlife and its habitat for an analysis or modelling approach. In this context, birds have been of great and progressive value as biological and environmental indicators. In order to learn about the common approaches used—its methods, processing steps, trends, advantages and challenges—over 120 representative publications of the last 30 years that made use of satellite images for avian applications have been reviewed. The reviewed studies have shown that GIS-based analyses of satellite and bird data have been well established for efficient ecosystem descriptions and species modelling within a large range of scales and habitats. In order to improve the quality of inference and for comparative analyses, it is recommended that further studies are documented in detail. Also, in order to verify and improve the obtained results, additional ground data on the main structure of the vegetation relevant to the bird species in question are usually necessary. Satellite-based remote sensing applications in ornithology could be used increasingly for assisting in habitat evaluation, habitat modelling and monitoring programmes and in achieving overall wildlife conservation and management objectives effectively. This is especially true for remote regions of the world that are difficult to access where few habitat studies have been undertaken to date but whose study is urgently needed. 1. Introduction The conservation of biodiversity of the 21st century faces two main challenges: habitat loss and species extinction (see, for example, IUCN (2000) and BirdLife International (2000)). It is widely accepted that the provisioning of habitat, and managing it for fauna and flora, is crucial for the conservation and protection of biodiversity. Current data on the distribution, status and ecology of wildlife species are generally not available for many countries and for many habitats of the world. *Corresponding author. Email: thomasgottschalk@surfeu.de International Journal of Remote Sensing ISSN 0143-1161 print/ISSN 1366-5901 online # 2005 Taylor & Francis Group Ltd http://www.tandf.co.uk/journals DOI: 10.1080/01431160512331338041 2632 T. Gottschalk et al. This is especially true for the tropics, which present a habitat that covers about 10% of the world. However, other large knowledge gaps still exist for the world’s oceans that cover even two-thirds of the Earth’s surface. Given the vast number of ecosystems in the world still unknown to us, conventional ground-based survey and mapping methods cannot always deliver the necessary information in a timely and cost-effective fashion. Satellite images generally present the sole source of habitat data for many regions of the world important to wildlife. With satellite-based remote sensing technology, we may be able to make real progress in understanding why more species occur in some places than in others and in identifying the most critical places (Bibby et al. 1992), allowing for effective habitat conservation and management. While habitat conservation focuses on general biodiversity and wildlife, birds are traditionally among the species that have received much of the attention and investigations. Reasons for this ‘bias’ are simple: birds are relatively easy to identify, their taxonomy is established, they are ubiquitous and they are sensitive or vulnerable to environmental changes (Steele et al. 1984, Morrison 1986, Baillie 1991, Furness et al. 1993). Birds are prominent representatives of wildlife and have been used effectively as bio-indicators of the state of complex ecosystems; often, their research presents rolemodels for general wildlife research. Several examples of effects on birds related to human-induced change to their habitat, such as land management practises, introduction of exotic plants and animals, and hunting were given in Morrison (1986) and Furness and Greenwood (1993). Further, Brooks et al. (2001) presented an example of how birds were used in a land conservation planning programme to represent the majority of other terrestrial vertebrate diversity. By reviewing 109 representative studies published in 122 publications of the last 30 years, we attempt to give a state-of-art overview for a wide range of approaches to remote sensing techniques that involve bird–habitat linkages. We further suggest directions for applications, studies and improvements. 2. Methods To identify papers that focus on bird and satellite remote sensing we used the terms ‘satellite’, ‘Landsat’, ‘SPOT’, ‘AVHRR’, ‘IRS’, ‘Quickbird’, ‘Ikonos’ or ‘Orbview’ in combination with the words ‘avian’ or ‘bird’ in the ISI Web of Science, which searches for these words in the abstract, title and keywords of papers published mainly during the last 10 years. However, most papers were found according to the ‘snowball system’: papers focusing on birds and satellite remote sensing were searched for additional references of the same topics. Our search was limited to studies published since the 1970s, the beginning of the public use of satellite images, up to August 2004. Although this method might have missed publications we assume that we have covered a representative part of the spectrum on birds in conjunction with satellite remote sensing topics. 3. 3.1 Literature review On the general use of satellite images and birds Since 1972, when satellite images became widely available for public use, extensive research has been conducted in which satellite images and animals have been investigated, increasingly supported by Geographic Information Systems (GIS) (see, for example Martinko 1981, Saxon 1983, Ormsby and Lunetta 1987, Pearce 1991, Hansen et al. 2001, Kerr et al. 2001). 2633 Avian habitat relationships Table 1. Classification scheme of avian habitat studies. Analysing possibilities Possible output Map type 1. 2. 3. 4. Maps Maps Maps Maps Binary scaled Ordinal scaled Binary/ordinal scaled Binary/ordinal scaled Identifying Characterizing Monitoring Predictive modelling of of of of distribution habitat use change potential distribution The objective of most of these studies was to find and to describe relationships between habitat categorization, based on spectral reflectance pattern and georeferenced bird species records. Herr and Queen (1993) distinguish three different ways of applying GIS in order to assess wildlife habitat. This classification is used in table 1 and a fourth category of ‘predictive modelling’ was added. The table was supplemented by ‘possible map outputs’. It is worthwhile emphasizing that many of the mentioned analysing possibilities are interrelated and often can be combined. For instance, in order to model bird distributions the main characteristics of its habitat need to be identified and mapped first, at least on a minimum level. In order to monitor habitats effectively they have to be identified. According to the classification scheme of table 1, habitats of the species were identified in 50 out of 109 analysed studies. Seventy-seven studies dealt with characterizing and classifying habitats, whereas modelling approaches to predict species distribution or abundance were found in 48 studies. Monitoring of bird species habitats was the objective in only 10 studies. We found that most of the 109 analysed studies were performed in North America (58 in USA, eight in Canada), three were conducted in Central America (two in Costa Rica, one northern Central America), two in South America (Argentina, Ecuador), 22 were carried out in Europe (nine in Great Britain, five in Spain, two in Norway and Sweden and one in Austria, Belgium, France and Italy), eight in Africa (Botswana, Senegal, Uganda, Tanzania, two in East Africa and two Crozet Archipelago) and just four in Asia (China, India, Russia, Turkey), four in Australia and two in Antarctica (table 2). Only 10 published studies that used satellite remote sensing techniques have been conducted in tropical regions. In the reviewed studies there is a huge variation in habitats and sizes of the study areas with the smallest covering 13 km2 in Kansas, USA (Palmeirim 1988) and the largest with 5 393 343 km2 in Australia (Roshier et al. 2001) or the whole United States with about 8 000 000 km2 (O’Connor et al. 1996). Other large-scale studies comprise three African countries (Wallin et al. 1992, Johnson et al. 1998) or the north-west Atlantic (Huettmann and Diamond 2001). The median of the size of the study areas was 1580 km2 (n575). As computer technology and worldwide data availability has developed considerably there is a steady increase in the size of the study areas during the 30 year period publication were analysed for that review. Except for nine studies (Priede 1983, Haney and McGillivary 1985a, Briggs et al. 1987, Haney 1989b, Jouventin et al. 1994, Guinet et al. 1995, Rodhouse et al. 1996, Huettmann and Diamond 2001, Yen et al. 2004), the use of remote sensing imagery for bird habitat investigations deals with terrestrial habitats. Most studies do not provide an exact percentage of the different types of habitat that are investigated. Therefore, we can report here only our own estimations. While remote sensing applications of open habitats such as agricultural land, heath, pasture, grass-, wetand moorland were predominant, forests as the main habitats were listed in more than a third of the studies. 2634 T. Gottschalk et al. Table 2. Location and area of the studies reviewed. ID No. Reference Study area (study size, location, country) 422.4 km2, Mendocino County, north coast of California, USA 41 km2, western Oklahoma, USA 109 km2, Sauvie Island, Oregon, USA 6340 km2, north-west Missouri, USA California, USA 1 Katibah and Graves (1978) 2 3 4 5 9 10 11 12 13 14 15 16 17 18 Cannon et al. (1982) Lyon (1983) Cary et al. (1983) Priede (1983) Briggs et al. (1984) Palmeirim (1985) Haney and McGillivary (1985a) Haney and McGillivary (1985b) Haney (1985) Haney (1986) Haney (1989a) Schwaller et al. (1984) Schwaller et al. (1986) Schwaller et al. (1989) Haney (1989b) Young et al. (1987) Briggs et al. (1987) Hodgson et al. (1987) Shaw and Atkinson (1988) Hodgson et al. (1988) Palmeirim (1988) Miller and Conroy (1990) Sader et al. (1991) Lauga and Joachim (1992) 19 20 21 22 23 24 Stoms et al. (1992) Wallin et al. (1992) Aspinall and Veitch (1993) Herr and Queen (1993) Homer et al. (1993) Griffiths et al. (1993) 25 26 27 Griffiths et al. (1993) Green and Griffiths (1994) Gustafson et al. (1994) 28 29 Andries et al. (1994) Jouventin et al. (1994) 30 31 Roseberry et al. (1994) Guinet et al. (1995) 32 Knick and Rotenberry (1995) 33 34 Warkentin et al. (1995) Avery and Haines-Young (1990) 35 Lavers and Haines-Young (1996a) Lavers et al. (1996) Lavers and Haines-Young (1996b) Lavers and Haines-Young (1997) Yates et al. (1996) Wash and Essex coast, east UK 6 7 8 52 km2, north-east Kansas, USA Southern South Atlantic, USA Ross Island, Antarctic Gulf of Alaska and South Atlantic Bight, USA Washington, USA Entire Californian coast, USA 1519 km2, Birdsville, south-east Georgia, USA 54 695 km2, central Texas, USA 1519 km2, Birdsville, south-east Georgia, USA 13 km2, Kansas, USA 25 km2, Eleuthera Island, Bahamas, USA 570 km2, Sarapique region, north-east Costa Rica 2327 km2, north-west of Toulouse, south-west France 25 000 km2, southern California, USA Somalia, Kenya, Tanzania 1500 km2, north-east Scotland North-west Minnesota, USA 2548 km2, Utah, USA 28861 km2 squares in southern Upland/northern Pennines, UK 15261 km2, Grampian region, UK 600 km2, Breckland, UK 3620 100 ha, west-central and north-central Indiana, USA 288 km2, Hageland, Belgium Possesion Island, Crozet Archipelago, south-west Indian Ocean, France 1127 km2, Hamilton County, Illinois, USA Ile aux Cochons, south-west Indian Ocean, France 2000 km2, Prey National Conservation Area, south-western Idaho, USA 732 km2, Selva Lacandona, Mexico 79 sites, 1850 km2, Flow Country, northern Scotland, UK Avian habitat relationships 2635 Table 2. (Continued). ID No. Reference Study area (study size, location, country) 36 37 38 39 40 O’Connor et al. (1996) Rodhouse et al. (1996) Neave et al. (1996) Austin et al. (1996) Gratto-Trevor (1996) 41 Jørgensen and Nøhr (1996) Nøhr and Jørgensen (1997) Donovan et al. (1997) White et al. (1997) Morrison (1997) 8000 000 km2, entire USA South Georgia, UK 7581 km2, 165 field sites, south-east Australia 2000 km2, Argyll, Scotland, UK 7650 km2, Mackenzie Delta, Northwest Territories, Canada 35 000 km2, Ferlo region, northern Senegal 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 Illinois, Indiana, Missouri, USA 1580 km2, Monroe County, Pennsylvania, USA 9 948 km2, Prince Charles Island, Foxe Basin, Northwest Territories, Canada Mack et al. (1997) 151 sample woodlands (200 m2 to 0.3 km2), area of 75 km645 km, eastern UK Tucker et al. (1997) 2903 km2, Catchment of River Tyne, northern UK Verlinden and Masogo (1997) Southern Kalahari, Botswana Hepinstall and Sader (1997) 85 000 km2, Maine, USA Thibault et al. (1998) Northern Quebec, Lacs des Loups Marins area, Canada Johnson et al. (1998) East Africa Berry et al. (1998) Boulder, Colorado, USA Roseberry and Sudkamp (1998) .145 900 km2, Illinois, USA Fuller et al. (1998) 2497 km2, six study sites, Sango Bay, Uganda Dettmers and Bart (1999) 560.9 km2 and 3260.3 km2, Wayne National Forest, south-eastern Ohio, USA Debinski et al. (1999) 324 km2, Greater Yellowstone Ecosystem, USA Debinski et al. (2000) Gallatin National Forest, USA Burke and Nol (2000) Ontario, Canada Curnutt et al. (2000) 10 000 km2, Florida, USA Miller and Cale (2000) 1680 km2, Kellerberrin district, Western Australia Vander Haegen et al. (2000) 78 study plots, Columbia Basin, eastern Washington, USA Grillmayer (2000) Five study plots in Burgenland, Lower Austria Howell et al. (2000) 10 study sites (3.4–301.25 km2) in Missouri, USA Haire et al. (2000) Boulder, Colorado, USA Fauth et al. (2000) Northern Indiana, USA Mörtberg and Wallentinus (2000) 2500 km2, Sweden Pino et al. (2000) 480 km2, Catalonia, Spain Cully and Winter (2000) 14.2 km2 Saline County, Kansas, USA Jones et al. (2000) Mid-Atlantic region, USA Jones et al. (2001) Mid-Atlantic region, USA Klaus et al. (2001) 1200 km2, province Gansu, China Bajema and Lima (2001) 249.8 km2, 19 study sites, south-western Indiana, USA Pearce et al. (2001) North-east New South Wales, Australia Daily et al. (2001) 21 study sites (ca 1.1 km2) within a 707 km2 circle, San Vito, southern Costa Rica Saveraid et al. (2001) 5060.01 km2, Greater Yellowstone Ecosystem, USA Osborne et al. (2001) ca 1266132 km, Madrid province, Spain Roshier et al. (2001) 5393 343 km2, inland Australia Huettmann and Diamond (2001) North Atlantic, Canada Groom and Grubb (2002) 48 km river in Franklin, Madison and Union Counties, Ohio, USA 2636 T. Gottschalk et al. Table 2. (Continued). ID No. Reference 79 80 81 Shriner et al. (2002) Shapiro et al. (2002) Pearson and Simons (2002) 82 83 Hepinstall et al. (2002) Hunsaker et al. (2002) 84 85 Bani et al. (2002) Pearman (2002) 86 87 88 89 Reese et al. (2002) Suarez-Seone et al. (2002) Pearlstine et al. (2002) Jepsen et al. (2002) 90 Debinski et al. (2002) 91 Borboroglu et al. (2002) 92 93 Meyer et al. (2002) May et al. (2002) 94 95 96 97 Gottschalk (2003) Kastdalen et al. (2003) Jenkins et al. (2003a) Jenkins et al. (2003b) Pidgeon et al. (2003) 98 Fearer and Stauffer (2003) 99 Hennings and Edge (2003) 100 101 102 103 104 105 106 107 108 109 3.2 H-Acevedo and Currie (2003) Mayer and Cameron (2003) Jeganathan et al. (2004) Bustamante and Seoane (2004) Yen et al. (2004). Venier et al. (2004) Hawkins (2004) Berberoglu et al. (2004) Seoane et al. (2004) Fujita et al. (2004) Study area (study size, location, country) Great Smoky Mountains National Park, USA 861.6 km2, Fort Hood, Texas, USA 3125 km2, Gulf coast of Texas, Louisana and Mississippi, USA Maine, USA 606 km2, southern Sierra Nevada, California, USA 2500 km2, Lombardy, Italy 2360.5 ha plots within 10 000 km2, Jatun Sacha Preserve, Upper Amazon Basin, Ecuador Dalarna county, Sweden 493 486 km2, peninsular Spain 176 187.56 km2, Florida, USA 62 700 km2, Svalbard archipelago, Spitsbergen, Norway Grand Teton National Park, Bridger-Teton National Forest in Wyoming and Gallatin National Forest in Montana, USA 340 km coastline and 45 islands in northern San Jorge Gulf, Argentina Southern Oregon and northern California, USA 2538 km2, 98 study plots, western South Dakota, USA 1100 km2, Serengeti Plains, Tanzania 2000 km2, Forollhogna National Park, Norway 500 km2, Everglades of South Florida, USA 661.08 km2 plots within 2406.76 km2, Fort Bliss Military Reserve, northern Chihuahuan Desert in south-central New Mexico, USA 103.4 km2, Clinch Mountain Wildlife Management Area, south-western Virginia, USA 956.48 km2, 54 riparian sites, north-west Oregon, USA 37 000 km2, North and Central America 644.53 km2, Ohio, USA Caddapah district, Andhra Pradesh, India 37 700 km2, Andalusia, southern Spain British Columbia, Canada 800 000 km2, Great Lake Basin, Canada and USA 48 400 km2, Eastern North America Cukurova delta, Turkey 9 800 km2, western Andalusia, southern Spain 20 km bands along 2188–2679 km long migration routes, Russia, China Methodologies used in reviewed studies The detailed methodologies of the studies depended mostly on bird census techniques, remote sensing data, the use of additional environmental data and statistical analyses. However, the typical studies using satellite-based remote sensing and bird data follow a common pattern (see figure 1). 1. Most of the studies involve satellite image classification using pre-defined ecological classes believed to be important. The importance of this Avian habitat relationships 2. 3. 4. 5. 2637 classification scheme was either derived from expert knowledge, from the literature, or habitat class separability according to its specific reflectance and the characteristics of the remote sensing system. The inherent assumption is that the produced map has ecological relevance, and predictive power to the wildlife species of interest. Bird data were collected, roughly for the same time period when the imagery was taken. Additional data were used, such as terrestrially mapped vegetation structures, soil maps or aerial photographs. The GIS analyses involved a logical or arithmetic map overlay operation of the satellite imagery with the bird data and other environmental data, and were followed by statistical analyses of relationships between the bird data and occurring classes in the satellite imagery and the other environmental data, respectively. In some studies, these relationships were then modelled by predicting the likelihood of occurrence in an area with unknown bird data but known habitat data derived from a satellite. 3.2.1 Satellite imagery. The most frequently used satellite sensor was Landsat (see table 3). While more than half of the studies were based on Landsat Thematic Mapper (TM) images, Landsat Multi-Spectral Scanner (MSS) scenes were used in 11 studies, sometimes along with Landsat TM scenes. Analyses of the higher spatial Figure 1. Flow chart of the methodology typically utilized in studies using satellite-based remote sensing and bird data. 2638 T. Gottschalk et al. resolution French SPOT satellite were performed nine times (Miller and Conroy 1990, Andries et al. 1994, Green and Griffiths 1994, Guinet et al. 1995, Thibault et al. 1998, Debinski et al. 2000, 2002, Klaus et al. 2001, Saveraid et al. 2001) and the newer Indian Remote Sensing (IRS) two times (Debinski et al. 2002, Kastdalen et al. 2003). The Advanced Very High Resolution Radiometer (AVHRR) from the National Oceanic and Atmospheric Administration (NOAA) was applied in 22 studies; six times it was used together with other satellite sensors. These coarseresolution imageries were often used in large-scale studies covering entire or large parts of countries (Wallin et al. 1992, Suarez-Seoane et al. 2002, Venier et al. 2004), continents (O’Connor et al. 1996, Johnson et al. 1998, Roshier et al. 2001, HAcevedo and Currie 2003) or the ocean (Haney and McGillivary 1985a, Jouventin et al. 1994, Huettmann and Diamond 2001). With the help of NOAA AVHRR imagery often the actual or integrated normalized difference vegetation index (NDVI) was calculated to produce maps of primary production or to detect green biomass (Wallin et al. 1992, Andries et al. 1994, Hepinstall and Sader 1997, Verlinden and Masogo 1997, Osborne et al. 2001, Saveraid et al. 2001), in order to detect habitat change (Sader et al. 1991, Debinski et al. 2000), to map wetland distribution (Roshier et al. 2001) or to explain the distribution of centres of bird endemism (Johnson et al. 1998). The NOAA AVHRR Climate Data Set incorporating values of atmospheric temperatures and sea surface temperatures in the Atlantic and the Pacific, were used for instance by Haney (1989b), Jouventin et al. (1994), Rodhouse et al. (1996), Huettmann and Diamond (2001), Meyer et al. (2002) and Yen et al. (2004) to analyse seabird habitats or to model their distribution. A radar ERS image was applied by Thibault et al. (1998) to investigate ice-cover on a river. The ice-cover was used to determine open water areas, which are the habitat of the endangered harlequin duck Histrionicus histrionicus. Detailed information on the processing infrastructure such as operating system, software and hardware as well as the spectral bands used for image classification were often lacking. Forty-nine out of the 109 reviewed studies did not provide information about geo-coding of the remote sensing imagery. However, in 34 studies images were geometrically corrected. In terms of thematic resolution, the number of habitat classes derived from remote sensing analyses ranged from 2 to 71 with a mean number of approximately 9.3 habitat (or vegetation/land use) classes (n568). Only 24 of the reviewed studies provided details on the overall accuracy of the classification process, which varied between 60% and 99% (mean value of about 85%). In 59 studies accuracy assessments were not performed or at least not mentioned. Some studies utilized an existing land cover classification scheme, such as, for example, the Centre for Ecology and Hydrology Landsat cover map of Great Britain, which was used by Mack et al. (1997) and Dettmers and Bart (1999) or the CORINE land-cover map of Europe, used by Seoane et al. (2004). When existing land-use maps were used usually no details of the classification process or the accuracy of the used map were mentioned. Further, 14 studies used raw data, without classifying the image or calculating, for instance, the NDVI. In terms of temporal scale, for many of the investigated studies the bird census period does not fit with the exact time when the satellite image was taken. The difference between time of satellite image acquisition and bird census period spans up to 78 years (mean difference of 3.5 years). In only 42 out of 81 studies which provided details this difference was less than 1 year. 2639 Avian habitat relationships Most analyses were not solely based on satellite imagery but used also additional data and covariates. In 30% of the studies aerial photographs—mainly colour infrared pictures—were included. In the majority of these studies they were used to verify image classification. Only a few studies used actual terrestrial estimations of per cent cover as additional vegetation type information. In 20 of the 109 analysed studies, a digital elevation model (DEM) was used. With the help of DEM data, slope, aspect or land ruggedness could be calculated and then used for further bird–habitat investigations. Other additional data sources were information on soils, nitrogen deposition, water depths, climate and different vegetation measurements such as vegetation height or cover value of a specific vegetation type, and man-made features such as roads and buildings. Those terrestrial sampled data were used in 36% of all studies. In several instances spatial texture measures, i.e. patchiness of a specific habitat, spatial configuration, fragmentation, area and boundary length of landscape elements and vegetation types were calculated. For example, Fauth et al. (2000), Haire et al. (2000), Saveraid et al. (2001) and Meyer et al. (2002) used FRAGSTATS, a spatial pattern analysis program for quantifying landscape structure (McGarigal and Marks 1994). Hepinstall and Sader (1997) successfully applied spatial texture measures to determine species preferences. Also, distances between bird records and, for example, agricultural land, coast, shelf edge, roads or buildings were calculated. These additional data were used for further refinements of the birds’ habitat analysis or its model. 3.2.2 Bird data. In order to link the habitat information with birds, and depending on the habitat or species being targeted, three main bird census techniques were used (see table 4). Mostly, simple counts of sightings of bird species or their nests were conducted. This was carried out in systematic and non-systematic ways, and sometimes with the help of tools such as aircraft (Briggs et al. 1984, Hodgson et al. 1987, 1988, Haney 1989b), helicopter (Wallin et al. 1992, Herr and Queen 1993, Morrison 1997, Jenkins et al. 2003a), boat (Briggs et al. 1984, Haney and McGillivary 1985a, Briggs et al. 1987, Haney 1989b, Huettmann and Diamond 2001, Yen et al. 2004), canoe (Groom and Grubb 2002), radio telemetry (Young et al. 1987, Hunsaker et al. 2002, Fearer and Stauffer 2003) or satellite telemetry (Guinet et al. 1995, Jouventin et al. 1994, Rodhouse et al. 1996). A direct link between bird breeding sites and satellite imagery was applied by Guinet et al. (1995) and Schwaller et al. (1984, 1986, 1989). They identified nesting rookeries of penguins on the ground by their spectral reflectance using Landsat TM and SPOT imagery. Transect counts and Point-Stop Counts (PSC) were performed in 21 and 29 studies, respectively. Furthermore, bird data derived from the literature was used in 32 studies. Twenty-two studies used ‘presence/absence’ data and 16 of the Table 3. Satellite images in 109 analysed studies. In some studies more than one satellite type was used. In six studies the source of satellite sensor data was not given. Satellite type NOAA Advanced Very High Resolution Radiometer (AVHRR) European Remote Sensing Satellite (ERS) radar image Indian Remote Sensing (IRS) Landsat Multi-Spectral Scanner (MSS) Landsat Thematic Mapper (TM) Meteosat High Resolution Radiometer (HRR) Système Probatoire pour l’Observation de la Terre (SPOT) No. of studies 22 1 2 11 66 1 9 2640 T. Gottschalk et al. Table 4. Census techniques and bird data structure in 109 analysed studies. In some studies more than one data source was used. In 35 studies detailed information on bird census data was not given. Number of studies Avian census technique/ Source of bird data Counts of individuals/ nests/colonies (systematic and random) PSC Transect counts Relative abundance True abundance (index counts) (densities) Presence data only Presence/ absence data 14 10 4 1 1 1 8 4 20 15 1 investigated studies were based on ‘presence only’ data. Census techniques that use counts of bird detection as an index of relative abundance (Rosenstock et al. 2002) were used 39 times. Distance Sampling (Buckland et al. 1993), an empirical modelling technique for survey data that directly estimates bird densities (absolute abundance) by taking the variation in species detectability into consideration, were only used in the study by Kastdalen et al. (2003). In eight studies the source and census methodology for the bird data was not given. Little research has been carried out on multi-species and avian community analyses. Most of the research using remote sensing was carried out to find habitat use, preferences or to predict species distribution either for a single species (41%) or for two to five bird species (11%). More than 20 bird species were used in 23% of all studies investigated. 3.2.3 Statistical analyses of wildlife–environment relationships. For the study of relationships between bird species distribution patterns and environmental variables the 109 studies used a wide array of statistical methods. Some of them have been specifically designed for these studies, for example Aspinall and Veitch (1993) and Tucker et al. (1997) who used specific Bayesian approaches. Other studies applied common statistical techniques. Many papers did mention statistical tests, but were not really specific about it. Mostly univariate and bivariate tests as well as multivariate statistics were carried out. Bayesian models were used in Aspinall and Veitch (1993), Griffiths et al. (1993), Tucker et al. (1997), Hepinstall and Sader (1997) and expert opinion techniques were applied by Pearce et al. (2001). Chisquare test, T-test and Mann–Whitney U-test were commonly carried out to compare single habitat variables between study plots with and without the species, and between two study areas containing species. The use of logistic and multiple regressions to predict species distribution has strongly increased since the 1990s (e.g. Avery and Haines-Young 1990, Lauga and Joachim 1992, Griffiths et al. 1993, Austin et al. 1996, Nøhr and Jørgensen 1997, Saveraid et al. 2001, Huettmann and Diamond 2001, Osborne et al. 2001, Meyer et al. 2002, Jepsen et al. 2002, SuarezSeoane et al. 2002, Jeganathan et al. 2004, Venier et al. 2004, Yen et al. 2004). 4. 4.1 Discussion Topics related to the remote sensing imagery The selection of the ‘right’ satellite sensor is a crucial decision which affects the outcome of the habitat description and the wildlife–habitat relationship. So far, no Avian habitat relationships 2641 methods are known to us that provide guidance for such decisions in detail. The reviewed studies did not show many considerations in regards to spatial scale and choice of appropriate satellite sensor. Topics like mismatch of scales between remote sensing data and avian data were discussed in only few studies. Thibault et al. (1998) and Venier et al. (2004) compared different satellite sensors and showed how the choice of satellite sensor may affect results in a bird species–habitat study. However, the preferred usage of Landsat in almost 80% of all analysed studies is probably due to the fact that the Landsat programme was the first major satellite programme for public use. In addition, the high spectral resolution of this sensor, its robust data structure, and a good price–performance ratio contributed to its success. O’Neill et al. (1996) recommend that the spatial resolution should be two to five times smaller than the spatial features of interest. That involves knowing the landscape features relevant to the species of interest beforehand, which is often impossible. Additionally, studies containing more fine-grained habitat information need adequate species knowledge or are fraught with many assumptions about the species’ ecology than on studies that have fewer details and are more general in design (Pearson and Simons 2002). Morrison et al. (1992) suggest a hierarchical approach viewing habitat analysis first from the broadest scale, and then working down to the finest level of scale necessary to answer the question of interest. For some bird species the habitat classification based on satellite images can be of insufficient resolution (see examples in table 5). Depending on the spatial resolution of the sensor, specific ‘hard edge’ features such as narrow linear grooves, hedgerows, fences and small patches of land or single trees cannot be (precisely) identified in many satellite images, although, they may be of major importance to specific wildlife species (see Mack et al. 1997). For these species the features are suggested to be identified and be mapped by ground surveys or by specific image analysis techniques such as contextual analysis or edge detection (Andries et al. 1994). Additionally, satellite imagery may not capture important events affecting the habitat (Jenkins et al. 2003b). Several non-habitat factors, like historical reason, intraspecific and interspecific interaction, microhabitat characteristics or the three-dimensional architecture of vegetation are beyond the scope of satellite remote sensing technology. Table 5. Examples of limitations of bird species–habitat relationship studies based on satellite images. Reference Bird species Palmeirim (1985) Ruffed grouse (Bonasa umbellus) Satellite sensor Main limitations mentioned Landsat TM Certain pertinent landscape features and local variables could not be detected Certain pertinent landscape features and local variables could not be detected Edge habitats could not be identified Satellite-based classes extended over a range of habitats Herr and Greater sandhill crane Landsat TM Queen (1993) (Grus canadensis tabida) Gratto-Trevor (1996) Tucker et al. (1997) Saveraid et al. (2001) 10 shorebird species Landsat TM Landsat TM Coal tit (Parus major), golden plover (Pluvialis apricaria), snipe (Gallinago gallinago) 11 passerine bird species Landsat TM and SPOT Spatial resolution was too coarse and did not contain enough information 2642 T. Gottschalk et al. Generally, the finer a satellite-derived habitat map is classified, the more bird species–habitat relationships can potentially be revealed. The new generation of high resolution satellite systems such as Ikonos, Quickbird and Orbview may advance possibilities to achieve a detailed habitat map, as they provide multi-spectral imagery of 2.8–4 m spatial resolution with an effective revisit capability of just a few days. Their characteristics offer the possibility of doing research that was previously only possible by ground-based studies or using airborne sensors (Clark et al. 2004, Ehlers 2004). Although their low spectral resolution (lacking of short-wave infrared) was identified by Mehner et al. (2004) as the main weakness of the data quality, they found that the high resolution satellite systems were capable of higher accuracy than achieved with previous approaches. However, usually the accuracy of the classification increases with a decrease in the number of classes used (Rees 1990). A higher-resolution satellite image could improve habitat separability as they introduce a substantial amount of extra detail and spatial variation into the classification process, and thus provide a higher number of habitats on a finer scale. Working at a too detailed level of scale could mean that the number of points per habitat type with bird records would significantly decrease thereby diminishing the power of the statistical tests to distinguish differences between habitat use and availability (Garshelis 2000). Stoms et al. (1992) tested the effects of coarser spatial resolution of a satellite-based habitat map. They eliminated polygons of a habitat map that are less than 20 ha and observed no significant effects on a habitat suitability model for the California condor Gymnogyps californianus. Venier et al. (2004) demonstrated that the finer resolution Landsat MSS land cover were not more useful than the coarser AVHRR land cover at identifying local habitat associations of 10 forest songbirds. Seoane et al. (2004) compared the predictive ability of bird distribution models derived from 30 m and 250 m resolution satellitebased maps. As well, they found no significant differences and suggested that improving vegetation maps above a certain limit has no effect in their power to predict bird distribution. Many sources of bias stem from the image processing algorithms employed, such as atmospheric correction, supervised or unsupervised classification techniques, and image transformations. These choices can be driven by the processing software available to the user. Additionally high cloud cover and shadow can restrict the choice of imagery available to the research project. As a consequence, this restriction can produce a significant temporal difference between bird and satellite sensor data. For example, Justice and Hiernaux (1986) stated that the 18 day orbital period of the Landsat platform is insufficient to provide regular cloud-free coverage during the 2 month growing and breeding season of the Sahelian Zone. Assigning habitat types to pixels can cause errors (White et al. 1997) as habitat classes determined from spectral signatures of the image may not directly correspond to specific and relevant avian units or habitats for birds. Certain bird species are likely to be better represented than others by the habitat classes. A difference between specialist and generalists species was found by Jones et al. (2000), as landscape characteristics derived from Landsat TM explained a greater proportion of the specialist species richness than the generalist species richness. Generally, it is recommended to partition the habitat types in terms of relevant features according to the habitat preferences (Garshelis 2000). In order to map habitats under species specific criteria a post-processing of the classified image is sometimes necessary. Unfortunately, this post-processing is highly dependent on the Avian habitat relationships 2643 prior knowledge of the interpreter and can lead to subjective biases. Thus, extracting land cover information from remote sensing images through a classifying process involves much human intervention and judgement, rather than being driven by objective biological requirements of the species. Drawing boundaries in natural vegetation areas can be highly problematic (see, for example, Fortin et al. 2000, Janssen 2000). Habitat types can display gradual change in community type. This is characterized in the satellite imagery through high frequency variation in spectral values and in a high proportion of mixed pixels (see Aspinall and Veitch 1993). Obviously, the level of variation has a limiting effect on the correlation between habitat and species (Nagendra 2001). Although sometimes a separation of transitional stages might be necessary, this separation can result in ‘too many’ habitat classes. That again would increase bird census efforts, as a higher density of bird records would be necessary to achieve significant species–environment relationships. If only few habitat classes are defined, each is likely to have considerable variation in plant species composition and other attributes (Cully and Winter 2000) and thus reduce the possibility to detect significant bird species–habitat relationships. In a species–habitat study, map accuracy can usually be confirmed efficiently by using the ground data from the bird sample plots (Berry et al. 1998, Seoane et al. 2004). The importance of a high accuracy of image classification is illustrated by Lyon et al. (1987). The authors showed that even a 5% change in classification accuracy of a land-cover map caused significant differences in levels of a habitat quality index. However the study did not focus on general sources of noise. Hunsacker et al. (2002) compared relationships between occurrence of California spotted owl Strix occidentalis and proportion of canopy-cover derived from Landsat TM imagery and derived from aerial photography. They found that differences in canopy-cover composition of less than 10% can significantly affect occupancy by the owl. In several studies computer-assisted image classification was not undertaken; instead raw reflectance data were used, or images were hand-traced. The disadvantage of this approach might be that relationships between reflectance data and species distribution can be identified but habitats are not described and stratified in ecologically meaningful terms. In addition, species–reflectance studies tend not to provide reliable answers when applied to areas other than the ones in which they were developed for because season, weather and soil condition may adversely affect the reliability of extrapolation to another image (Nagendra 2001). Furthermore, exact relationships between bird occurrence and satellite-derived land cover can fluctuate. Avery and Haines-Young (1990) and Lavers et al. (1996) found correlations between dunlin (Calidiris alpina) abundance and near-infrared reflectance of the Landsat image in Caithness and Sutherland, Britain, but not for the nearby Shetland Islands as on that island the easily detectable pools were missing (Lavers and Haines-Young 1997). Further, Morrison (1997) concluded that mapping wildlife habitats of the Arctic requires constant ground truthing and recalibration of habitat classes in different regions. The importance of a large number of environmental variables increases with the complexity of the animals and their habitats, and with the degree of accuracy of the species–environment relationships to be studied. Vertical habitat structure features can severely affect wildlife distribution and abundance. Severe limitations exist in woodlands and forests, as most satellites, except for microwave imagery and Synthetic Aperture Radar (SAR), are unable to detect or characterize the structure of lower layers, which are likely to be of importance for some species (see, for 2644 T. Gottschalk et al. instance, Palmeirim 1985). In some of the reviewed studies, the additional groundbased data used along with satellite-based data were biased by the fact that the data were chosen because they were available. However, to find explanatory and meaningful variables determining wildlife occurrence is a difficult endeavour and depends much on the biotope of interest. In order to measure for instance forest habitat structures, Morrison et al. (1992) list 29 different variables, such as woody foliage profile density or tree stump size. To analyse habitats or to build models of species habitat relationships using remote sensing data and GIS methods, a species must be either common enough and/or habitat-specific enough to exhibit a significant relationship with one or more remotely sensed habitat types (Debinski et al. 1999). Significant species–environment relationship depends on the bird species itself, and whether it tends to be ubiquitous or rare and to the degree of habitat map detail. The main habitat features, which are assumed to be closely related to the objectives, should be mapped, as far as time and financial resources are available. Such specific habitat features can usually indicate habitat preferences (e.g. Aebischer and Roberston 1992, Jones 2001, Manly et al. 2002). Some of these habitat features are helpful as reference data for digital image classification. Roseberry and Sudkamp (1998) selected variables that were theoretically meaningful, predictive, and consistent with the resolution and accuracy of the land-cover data. Pearlstine et al. (2002) suggested that predicted species models only based on land cover classes fail to incorporate many of the basic ecological characteristics of the species. By applying ground-based habitat data, Mack et al. (1997) found that species–area relationships described the bird species numbers better than those from remotely sensed data. However, they also showed that remotely sensed data are of sufficient spatial resolution for coarse scale estimates of species–area relationships. To assess the relative contribution of ground-based data and satellite sensor data in explaining differences in data of 62 bird species, Gottschalk (2003) used the coefficients of determination (R2) to compare different regression models. The best fit of the models was achieved by combining satellite- and ground-based variables (R250.26) compared to the sole use of satellite sensor data (R250.18). Nagendra (2001) pointed out that ancillary data may best to predict species distribution in natural areas. Keeping in mind that for large scale studies exhaustive terrestrial data are often not available, Venier et al. (2004) showed that significant improvements to model bird species distribution can also be made by adding worldwide available NOAA AVHRR climate data to land cover maps. 4.2 Temporal scale and seasonality In many ecological studies a source of bias is presented by the relatively short census period, e.g. within a season and rarely more than over a period of 3 years. In addition, an analysis based on multi-temporal imagery has greater discriminatory power than the analysis of a single-date image. Firstly, habitat separability and identification can be improved by multi-date image based classifications. Secondly, it is unlikely that short duration studies adequately reveal the variability inherent in habitat use by animals. Temporal shifts of habitat use and habitat suitability can occur within time of day, and within and between seasons and years. These shifts can result from numerous factors such as weather, food abundance, and densities of conspecifics, competitors and predators (Block and Brennan 1993). For example, extreme seasonal weather events, such as El Niño, can force animals into habitats not used during typical years (Homer et al. 1993) and thus bias wildlife–habitat Avian habitat relationships 2645 investigations. Therefore multiyear replication is necessary to accurately identify most of the species present and the image should normally be closely matched in time to the period(s) of the bird survey. By contrast, some ground cover characteristics may change little from year to year (see, for example, Lavers and Haines-Young 1997). 4.3 Bird census In many of the reviewed studies, non-systematic and non-standardized censuses and surveys were used. These techniques, however, have the disadvantages that they are not transparent and that results can be biased and based on incomplete and not representative data. Of the standardized census techniques, geo-referenced PSC and further line transect counts seem to be the most appropriate methods, particularly for studies in remote or difficult to access areas. This is because they are less walking intensive and thus less time-consuming and more effective. These census techniques are useful for a broad spectrum of both, natural and non-natural habitats. Georeferenced PSC and transect counts can be spatially adjusted to the required scale of the remote sensing imagery. If data on distributions and true abundances (densities) are needed, especially for accurate modelling approaches, for many applications, distance sampling (see Buckland et al. 1993, Bibby et al. 2000) is usually recommended. Its use allows correction for the possibility of detecting birds as it takes into account that some birds are detectable over much greater distances than others, and that a species may be more easily spotted in one habitat than another. General sources of errors caused by bird census techniques are discussed, for example, in Ralph and Scott (1981), Bibby et al. (2000) and Farnsworth et al. (2002). The number of survey points should be adjusted, as well as the number of required repeats of surveys in order to obtain robust estimates. No study was found that tested different efforts on bird census to show the effect on species–environment relationships. However, Bibby et al. (1998) recommended a minimum of about 50 points per habitat. For rare species, a higher bird census effort, e.g. density and repeats of surveys, might be required in order to obtain robust estimates. If a model is based on relative or absolute densities, as well as on presence– absence data, distortion of concern can be caused by autocorrelation (Schneider 1994). Double recordings of the same bird at two or more (neighbouring) points can be possible for high, mobile vagrant or large birds visible from two or more points. Although the number of papers that involve satellite tracking techniques have increased (e.g. Ginati et al. 1995, Hatch et al. 2000, Berthold et al. 2001, Hashmi et al. 2001, Watzke et al. 2001), only three of the papers reviewed by us referred to these techniques (Jouventin et al. 1994, Guinet et al. 1995, Rodhouse et al. 1996). This may be due to current inaccuracies in location determination, which actually lies at best around ¡30–150 m and would only allow broad-scale analyses. It is to be expected, however, that future developments will improve satellite tracking techniques and will allow a cheaper and more precise determination of animal locations. Linking of precise bird locations derived by satellite tracking, with habitat maps derived by space-borne sensors, presents a future challenge that needs to be investigated further. 4.4 Analysing approaches to investigate bird–habitat relationships and model performance Remote sensing imagery offers great potential to investigate the subject of ‘habitat availability’ which is crucial to address true preferences from habitat use, and for 2646 T. Gottschalk et al. valid model inferences (e.g. Aebischer and Robertson 1994, Manly et al. 2002). The reviewed papers did not discuss these issues, and thus appeared to be biased towards showing statistical relationships but not inferred biologically meaningful habitat links or true habitat preferences. An alternative can be the use of resource selection functions (RSF) (e.g. Aebischer and Robertson 1992, Boyce and McDonald 1999, Manly et al. 2002, Kastdalen et al. 2003), which offer the opportunity that only a small but representative fraction of the study area needs to be sampled and surveyed for the wildlife species. Once a robust relationship has been established this relationship can be modelled and extrapolated to the entire area that is recorded, classified and inventoried by remote sensing. Additionally, software developments like Biomapper (for presence only data) and GARP (Genetic Algorithm for Rule-set Prediction) have received much attention in recent times for spatial analysis and predicting species occurrences (Hirzel et al. 2002, Stockwell and Peters 1999). Furthermore, programmes such as FRAGSTATS or the Patchanalyst are well-used tools in remote sensing applications for quantifying landscapes/habitats and studying wildlife–habitat relationships. One utility of models is to guide further efforts and studies (Garshelis 2000). The methodology, however, needs to be adjusted by a refinement process that is based on the specific abiotic and biotic circumstances. Model verification refers to a broad spectrum of performance standards and criteria including model credibility and acceptability, realism, generality, precision, breadth, and depth (Marcot et al. 1983). Key concepts such as threshold-based and non-threshold-based accuracy checking for models of species habitat or species distribution are given for instance by Fielding and Bell (1997) and Pearce and Ferrier (2000). So far, however, they have not been applied very often. The strength of the species–habitat relationship is fundamental in a model and also in a monitoring programme based on bioindicators. It should therefore be improved through rigorous testing. The value of ground truthing the relationship between remotely sensed habitat and species on the ground cannot be overestimated (Debinski and Humphrey 1997). 5. Conclusions and suggestions for future studies Thirty years of using satellite-based remote sensing and bird data in a GIS environment has greatly increased our ability to assess causal effects in species– environment relationships. Many of the findings depicted from more than 100 avian studies using satellite remote sensing are relevant for general wildlife studies and can therefore used as a template. These techniques allow the convenient presentation, efficient prediction and monitoring of the distribution of species/communities and the estimation of the bird population sizes for large and remote areas. Its highest potential might exist in its application in ecosystems where access is limited and coarse and quick but quantitative estimates with statistical confidence limits on biodiversity are urgently needed. In terrestrial environments, combining satellite sensor data with terrestrially derived measurements seems to be the most effective way to determine species–environment relationships. For most study sites, the only alternative to remote sensing is ground data collection, which makes the cost of covering large areas prohibitively high. For issues that deal with best habitat classification, the optimum number of habitat features and the most appropriate scales need to be addressed. This has to be specified according to the ecology of the target species and the local landscape configuration. To Avian habitat relationships 2647 better understand and to improve the quality of the inference and for comparative analyses, it is highly recommended that these studies are standardized. Therefore, bird census techniques, terrestrial measurements, satellite image sources and classification, accuracy assessment and statistical analyses should be well thought out and documented in further investigations. This detailed documentation should be given for data collected by the investigator as well as by others. More multidisciplinary studies and approaches are needed in order to properly advance the benefits that remote sensing methods can bring to the research and conservation on wildlife and its habitats. To date, not all remote sensing applications in ornithology are mature, and cost–benefit analyses of remote sensing applications for precise avian–habitat applications are somewhat lacking. Nevertheless, the method has been well established within the last 30 years and is likely to play an increasing role in a huge range of applications, especially those determining the essential spatial elements of species habitats, modelling the distribution of birds and habitats, and monitoring their changes at a global level. Acknowledgements We would like to thank D. Gann, P. Lurz, S. Franklin, M. Wulder, J. Linke, G. McDermid, Lisa Strecker, and two anonymous reviewers for their input and suggestions to improve earlier versions of the manuscript. F.H. was supported by the University of Alaska–Fairbanks, and with a Postdoctoral Killam Fund, held with the Geography Department at the University of Calgary, Canada. References AEBISCHER, N.J. and ROBERTSON, P.A., 1992, Practical aspects of compositional analysis as applied to pheasant habitat utilization. In Wildlife Telemetry: Remote Monitoring and Tracking of Animals, I.G. Priede and S.M. Swift (Eds), (New York: Ellis Horwood) pp. 285–293. AEBISCHER, N.J. and ROBERTSON, P.A., 1994, Testing for resource use and selection by marine birds: a comment. Journal of Field Ornithology, 65, pp. 210–213. ANDRIES, A.M., GULINCK, H. and HERREMANNS, M., 1994, Spatial modelling of the Barn Owl Tyto alba habitat using landscape characteristics derived from SPOT data. Ecography, 17, pp. 278–287. ASPINALL, R. and VEITCH, N., 1993, Habitat mapping from satellite imagery and wildlife survey data using a Bayesian modelling procedure in a GIS. Photogrammetric Engineering and Remote Sensing, 59, pp. 537–543. AUSTIN, G.E., THOMAS, C.J., HOUSTON, D.C. and THOMPSON, D.B.A., 1996, Predicting the spatial distribution of buzzard Buteo buteo nesting areas using a Geographical Information System and remote sensing. Journal of Applied Ecology, 33, pp. 1541–1550. AVERY, M.I. and HAINES-YOUNG, R.H., 1990, Population estimates for the dunlin Calidris alpina derived from remotely sensed satellite imagery of the Flow Country of northern Scotland. Nature, 344, pp. 860–862. BAILLIE, S.R., 1991, Monitoring terrestrial breeding bird populations. In Monitoring for Conservation and Ecology, B. Goldsmith (Ed.) (London: Chapman & Hall), pp. 112–132. BAJEMA, R.A. and LIMA, S.L., 2001, Landscape-level analyses of Henslow’s Sparrow (Ammodramus henslowii) abundance in reclaimed coal mine grasslands. American Midland Naturalist, 145, pp. 288–298. BANI, L., MARCO, B., LUCIANA, B. and RENATO, M., 2002, The use of focal species in designing a habitat network for a lowland area of Lombardy, Italy. Conservation Biology, 16, pp. 826–831. 2648 T. Gottschalk et al. BERBEROGLU, S., YILMAZ, K.T. and OZKAN, C., 2004, Mapping and monitoring of coastal wetlands of Cukurova Delta in the Eastern Mediterranean region. Biodiversity and Conservation, 13, pp. 615–633. BERRY, M.E., BOCK, C.E. and HAIR, S.L., 1998, Abundance of diurnal raptors on open space grasslands in an urbanized landscape. Condor, 100, pp. 601–608. BERTHOLD, P., VAN DEN BOSSCHE, W., FIEDLER, W., KAATZ, C., KAATZ, M., LESHEM, Y., NOWAK, E. and QUERNER, U., 2001, Detection of a new important staging and wintering area of the White Stork Ciconia ciconia by satellite tracking. Ibis, 143, pp. 450–455. BIBBY, C.J., COLLAR, N.J., CROSBY, M.J., HEATH, M.F., IMBODEN, C., JOHNSON, T.H., LONG, A.J., STATTERSFIELD, A.J. and THIRGOOD, S.J., 1992, Putting Biodiversity on the Map, International Council for Bird Preservation, Cambridge. BIBBY, C.J., JONES, M. and MARSDEN, S., 1998, Expedition Field Techniques, Bird Surveys. Expedition Advisory Centre, Royal Geographical Society, London. BIBBY, C.J., BURGESS, N.D., HILL, D.A. and MUSTOE, S., 2000, Bird Census Techniques, 2nd edn (London: Academic Press). BIRDLIFE INTERNATIONAL 2000, Threatened Birds of the World (Barcelona and Cambridge: Lynx Edicions and BirdLife International). BLOCK, W.M. and BRENNAN, L.A., 1993, The habitat concept in ornithology: theory and applications. Current Ornithology, 11, pp. 35–91. BORBOROGLU, P.G., YORIO, P., BOERSMA, P.D., VALLE, H.D. and BERTELLOTTI, M., 2002, Habitat use and breeding distribution of Magellanic Penguins in the northern San Jorge Gulf, Patagonia, Argentina. Auk, 119, pp. 233–239. BOYCE, M.S. and MCDONALD, L.L., 1999, Relating populations to habitats using resource selection functions. Trends in Ecology and Evolution, 14, pp. 268–272. BRIGGS, K.T., DETTMAN, K.F., LEWIS, D.B. and TYLER, W.B., 1984, Phalarope feeding in relation to autumn upwelling off California. In Marine Birds: Their Feeding Ecology and Commercial Fisheries Relationships, D.N. Nettleship, G.A. Sanger and P.F. Springer (Eds), pp. 51–62 (Ottawa: Canadian Wildlife Service). BRIGGS, K.T., TYLER, W.B., LEWIS, D.B. and CARLSON, D.R., 1987, Bird communities at sea off California: 1975 to 1983. Studies in Avian Biology, 11, pp. 1–74. BROOKS, T., BALMFORD, A., BURGESS, N., HANSEN, L.A., MOORE, J., RAHBEK, C., WILLIAMS, P., BENNUN, L.A., BYARUHANGA, A., KASOMA, P., NJOROGE, P., POMEROY, D. and WONDAFRASH, M., 2001, Conservation priorities for birds and biodiversity: do East African Important Bird Areas represent species diversity in other terrestrial vertebrate groups? Ostrich, (Suppl.), 15, pp. 3–12. BUCKLAND, S.T., ANDERSON, D.R., BURNHAM, K.P. and LAAKE, J.T., 1993, Distance Sampling: Estimating Abundance of Biological Populations (London: Chapman & Hall). BURKE, D.M. and NOL, E., 2000, Landscape and fragment size effects on reproductive success of forest-breeding birds in Ontario. Ecological Applications, 10, pp. 1749–1761. BUSTAMANTE, J. and SEOANE, J., 2004, Predicting the distribution of four species of raptors (Aves: Accipitridae) in southern Spain: statistical models work better than existing maps. Journal of Biogeography, 31, pp. 295–306. CANNON, R.W., KNOPF, F.L. and PETTINGER, L.R., 1982, Use of Landsat data to evaluate Lesser Prairie-Chicken habitats in western Oklahoma. Journal of Wildlife Management, 46, pp. 915–922. CARY, D.L., COOK, E.A. and BARNEY, T.W., 1983, Interpretation of pheasant habitat using satellite imagery. Perdix, 3, pp. 148–157. CLARK, D.B., READ, J.M., CLARK, M., MURILLO CRUZ, A., FALLAS DOTTI, M. and CLARK, D.A., 2004, Application of 1-m and 4-m resolution satellite data to ecological studies of tropical rain forest. Ecological Applications, 14, pp. 61–74. CULLY, J.F. and WINTER, S.L., 2000, Evaluation of land condition trend analysis for birds on a Kansas military training site. Environmental Management, 25, pp. 625–633. Avian habitat relationships 2649 CURNUTT, J.L., COMISKEY, J., NOTT, M.P. and GROSS, L.J., 2000, Landscape-based spatially explicit species index models for Everglades restoration. Ecological Applications, 10, pp. 1840–1860. DAILY, G.C., EHRLICH, P.R. and SANCHEZ-AZOFEIFA, G.A., 2001, Countryside biogeography: use of human-dominated habitats by the avifauna of southern Costa Rica. Ecological Applications, 11, pp. 1–13. DEBINSKI, D.M. and HUMPHREY, P.S., 1997, An integrated approach to biological diversity assessment. Natural Areas Journal, 17, pp. 355–365. DEBINSKI, D.M., KINDSCHER, K. and JAKUBAUSKAS, M.E., 1999, A remote sensing and GISbased model of habitats and biodiversity in the Greater Yellowstone Ecosystem. International Journal of Remote Sensing, 20, pp. 3281–3291. DEBINSKI, D.M., JAKUBAUSKAS, M.E. and KINDSCHER, K., 2000, Montane Meadows as indicators of environmental change. Environmental Monitoring and Assessment, 64, pp. 213–225. DEBINSKI, D.M., JAKUBAUSKAS, M.E., KINDSCHER, K., SAVERAID, E.H. and BORGOGNONE, M.G., 2002, Predicting meadow communities and species occurences in the Greater Yellowstone Ecosystem. In Predicting Species Occurence Issues of Accuracy and Scale, J.M. Scott, P.J. Heglund and M.L. Morrison (Eds) (Washington, DC: Island Press), pp. 499–506. DETTMERS, R. and BART, J., 1999, A GIS modelling method applied to predicting forest songbird habitat. Ecological Applications, 9, pp. 152–163. DONOVAN, T.M., JONES, P.W., ANNAND, E.M. and THOMPSON, F.R., 1997, Variation in local-scale edge effects: mechanisms and landscape context. Ecology, 78, pp. 2064–2075. EHLERS, M., 2004, Remote sensing for GIS applications: new sensors and analysis methods. In Remote Sensing for Environmental Monitoring, GIS Applications, and Geology III, Proceedings of SPIE Vol. 5239, M. Ehlers, H.J. Kaufmann and U. Michel (Eds), pp. 1–13 (Bellingham, WA: The International Society for Optical Engineering). FARNSWORTH, G.L., POLLOCK, K.H., NICHOLS, J.D., SIMONS, T.R., HINES, J.E. and SAUER, J.R., 2002, A removal model for estimating detection probabilities from point count surveys. Auk, 119, pp. 414–425. FAUTH, P.T., GUSTAFSON, E.J. and RABENOLD, K.N., 2000, Using landscape metrics to model source habitat for Neotropical migrants in the midwestern U.S. Landscape Ecology, 15, pp. 621–631. FEARER, T.M. and STAUFFER, D.F., 2003, Relationship of Ruffed Grouse (Bonasa umbellus) home range size to landscape characteristics. American Midland Naturalist, 150, pp. 104–114. FIELDING, A.H. and BELL, J.F., 1997, A review of methods for the assessment of prediction errors in conservation presence/absence models. Environmental Conservation, 24, pp. 38–49. FORTIN, M.-J., OLSON, R.J., FERSON, S., IVERSON, L., HUNSACKER, C., EDWARDS, C., LEVINE, D., BOTERA, K. and CLEMAS, W.V., 2000, Issues related to the detection of boundaries. Landscape Ecology, 15, pp. 453–466. FUJITA, G., HONG-LIANG, G., UETA, M., GOROSHKO, O., KREVER, V., OZAKI, K., MITA, N. and HIGUCHI, H., 2004, Comparing areas of suitable habitats along travelled and possible shortest routes in migration of White-naped Cranes Grus vipio in East Asia. Ibis, 146, pp. 461–474. FULLER, R.M., GROOM, G.B., MUGISHA, S., IPULET, P., POMEROY, D., KATENDE, A., BAILEY, R. and OGUTU-OHWAYO, R., 1998, The integration of field survey and remote sensing for biodiversity assessment: a case study in the tropical forests and wetlands of Sango Bay, Uganda. Biological Conservation, 86, pp. 379–391. FURNESS, R.W. and GREENWOOD, J.J.D., 1993, Birds As Monitors of Environmental Change, (Eds), (London: Chapman & Hall), 368 pp. 2650 T. Gottschalk et al. FURNESS, R.W., GREENWOOD, J.J.D. and JARVAIS, P.J., 1993, Can birds be used to monitor the environment? In Birds As Monitors of Environmental Change, R.W. Furness and J.J.D. Greenwood (Eds), pp. 1–41 (London: Chapman & Hall). GARSHELIS, D.L., 2000, Delusions in habitat evaluation: measuring use, selection, and importance. In Research Techniques in Animal Ecology, L. Boitani and T.K. Fuller (Eds) (New York: Columbia University Press), pp. 111–164. GINATI, A., LEHMANN, G. and SCHULZ, U., 1995, New trends and capabilities of satellites for bird tracking and monitoring Israel Journal of Zoology, 41, pp. 253–259. GOTTSCHALK, T.K., 2003, A remote sensing and GIS-based model of avian species habitat. In 30th International Symposium on Remote Sensing of Environment. Information for Risk Management and Sustainable Development, Honolulu, Hawaii (Tucson: ISRSE), pp. TS46.1. GRATTO-TREVOR, C.L., 1996, Use of Landsat TM imagery in determining important shorebird habitat in the Outer Mackenzie Delta, Northwest territories. Arctic, 49, pp. 11–22. GREEN, R.E. and GRIFFITHS, G.H., 1994, Use of preferred nesting by Stone Curlews Burhinusoedicnemus in relation to vegetation structure. Journal of Zoology, 233, pp. 457–471. GRIFFITHS, G.H., SMITH, J.M., VEITCH, N. and ASPINALL, R., 1993, The ecological interpretation of satellite imagery with special reference to bird habitats. In Landscape Ecology and Geographic Information Systems, R.H. Haines-Young and S.H. Cousins (Eds), pp. 255–272 (London: Taylor & Francis). GRILLMAYER, R., 2000, Satellitenbildbasierte Analyse der Landschaftsstruktur für die Habitatbewertung des Steinkauzes. In Angewandte Geographische Informationsverarbeitung, XII J. Strobl, T. Blaschke, and G. Griesebner (Eds), pp. 202–209 (Heidelberg: Wichmann). GROOM, J.D. and GRUBB, T.C., 2002, Bird species associated with riparian woodland in fragmented, temperate–deciduous forest. Conservation Biology, 16, pp. 832–836. GUINET, C., JOUVENTIN, P. and MALACAMP, J., 1995, Satellite remote sensing in monitoring change of seabirds: use of SPOT image in king penguin population increase at Ile aux Cochons, Crozet Archipelago. Polar Biology, 15, pp. 511–515. GUSTAFSON, E.J., PARKER, G.R. and BACKS, S.E., 1994, Evaluating spatial pattern of wildlife habitat: a case study of the Wild Turkey (Meleagris gallopavo). American Midland Naturalist, 131, pp. 24–33. H-ACEVEDO, D. and CURRIE, D.J., 2003, Does climate determine broad-scale patterns of species richness? A test of the causal link by natural experiment. Global Ecology and Biogeography, 12, pp. 461–473. HAIRE, S.L., BOCK, C.E., CADE, B.S. and BENNETT, B.C., 2000, The role of landscape and habitat characteristics in limiting abundance of grassland nesting songbirds in an urban open space. Landscape and Urban Planning, 48, pp. 65–82. HANEY, J.C., 1985, Wintering Phalaropes off the south-eastern United States: application of remote sensing imagery to seabird habitat analysis at oceanic fronts. Journal of Field Ornithology, 56, pp. 321–333. HANEY, J.C., 1986, Remote sensing of seabird environments. American Birds, 40, pp. 396–400. HANEY, J.C., 1989a, Remote characterization of marine bird habitats with satellite imagery. Colonial Waterbirds, 12, pp. 67–77. HANEY, J.C., 1989b, Iterative techniques for characterizing marine bird habitats with timeseries of satellite images. Colonial Waterbirds, 12, pp. 78–89. HANEY, J.C. and MCGILLIVARY, P.A., 1985a, Midshelf fronts in the South Atlantic Bight and their influence on seabird distribution and seasonal abundance. Biological Oceanography, 3, pp. 401–430. HANEY, J.C. and MCGILLIVARY, P.A., 1985b, Aggregations of Cory’s Shearwaters (Calonectris diomedea) at Gulf Stream fronts. Wilson Bulletin, 97, pp. 191–200. Avian habitat relationships 2651 HANSEN, M.J., FRANKLIN, S.E., WOUDSMA, C.G. and PETERSON, M., 2001, Caribou habitat mapping and fragmentation analysis using Landsat MSS, TM, and GIS data in the North Columbia Mountains, British Columbia, Canada. Remote Sensing of Environment, 77, pp. 50–65. HASHMI, D., RISTOW, D., BERTHOLD, P. and QUERNER, U., 2001, Day–night variation in spatial activity of wintering Cory’s Shearwaters Calonectris diomedea recorded by satellite tracking. Vogelwelt, 122, pp. 273–277. HATCH, S.A., MEYERS, P.M., MULCAHY, D.M. and DOUGLAS, D.C., 2000, Seasonal movements and pelagic habitat use of murres and puffins determined by satellite telemetry. Condor, 102, pp. 145–154. HAWKINS, B.A., 2004, Summer vegetation, deglaciation and the anomalous bird diversity gradient in eastern North America. Global Ecology and Biogeography, 13, pp. 321–325. HENNINGS, L.A. and EDGE, W.D., 2003, Riparian bird community structure in Portland, Oregon: habitat, urbanization, and spatial scale patterns. Condor, 105, pp. 288–302. HEPINSTALL, J.A. and SADER, S.A., 1997, Using Bayesian Statistics, Thematic Mapper Satellite Imagery, and breeding bird survey data to model bird species probability of occurrence in Maine. Photogrammetric Engineering and Remote Sensing, 63, pp. 1231–1237. HEPINSTALL, J.A., KROHN, W.B. and SADER, S.A., 2002, Effects of niche on the performance and agreement of avian habitat models. In Predicting Species Occurrence: Issues of Accuracy and Scale, J.M. Scott, P.J. Heglund and M.L. Morrison (Eds), pp. 593–606 (Washington, DC: Island Press). HERR, A.M. and QUEEN, L.P., 1993, Crane habitat evaluation using GIS and remote sensing. Photogrammetric Engineering and Remote Sensing, 59, pp. 1531–1538. HIRZEL, A.H., HAUSSER, J., CHESSEL, D. and PERRIN, N., 2002, Ecological-niche factor analysis: how to compute habitat-suitability maps without absence data? Ecology, 83, pp. 2027–2036. HODGSON, M., JENSEN, J., MACKEY, H. and COULTER, M., 1987, Remote sensing of wetland habitat: a Wood Stork example. Photogrammetric Engineering and Remote Sensing, 53, pp. 1075–1080. HODGSON, M., JENSEN, J., MACKEY, H. and COULTER, M., 1988, Monitoring Wood Stork foraging habitat using remote sensing and Geographic Information Systems. Photogrammetric Engineering and Remote Sensing, 54, pp. 1601–1607. HOMER, C.G., EDWARDS, T.C., RAMSEY, R.D. and PRINCE, K.P., 1993, Use of remote sensing methods in modelling sage grouse winter habitat. Journal of Wildlife Management, 57, pp. 78–84. HOWELL, C.A., LATTA, S.C., DONOVAN, T.M., PORNELUZI, P.A., PARKS, G.R. and FAABORG, J., 2000, Landscape effects mediate breeding bird abundance in midwestern forests. Landscape Ecology, 15, pp. 547–562. HUETTMANN, F. and DIAMOND, A.W., 2001, Seabird colony locations and environmental determination of seabird distribution: a spatially explicit breeding seabird model for the Northwest Atlantic. Ecology Modelling, 14, pp. 261–298. HUNSAKER, C.T., BOROSKI, B.B. and STEGER, G.N., 2002, Relations between canopy cover and the occurrence and productivity of California Spotted Owls. In Predicting Species Occurrence: Issues of Accuracy and Scale, M.L. Morrison (Ed.), pp. 687–700 (Washington, DC: Island Press). IUCN 2000, IUCN Red List of threatened species, URL: http://www.redlist.org/. JANSSEN, L.L.F. (Ed.), 2000, Principles of Remote Sensing, ITC Educational Textbook Series 2, ITC, Enschede, The Netherlands. JEGANATHAN, P., GREEN, R.E., NORRIS, K., VOGIATZAKIS, I.N., BARTSCH, A., WOTTON, S.R., BOWDEN, C.G.R., GRIFFITHS, G.H., PAIN, D. and RAHMANI, A.R., 2004, Modelling habitat selection and distribution of the critically endangered Jerdon’s courser Rhinoptilus bitorquatus in scrub jungle: an application of a new tracking method. Journal of Applied Ecology, 41, pp. 224–237. 2652 T. Gottschalk et al. JENKINS, C.N., POWELL, R.D., BASS, O.L. and PIMM, S.L., 2003a, Demonstrating the destruction of the habitat of the Cape Sable seaside sparrow. Animal Conservation, 6, pp. 29–38. JENKINS, C.N., POWELL, R.D., BASS, O.L. and PIMM, S.L., 2003b, Why sparrow distributions do not match model predictions. Animal Conservation, 6, pp. 39–46. JEPSEN, J.U., EIDE, N.E., PRESTRUD, P. and JACOBSEN, L.B., 2002, The importance of prey distribution in habitat use by arctic foxes (Alopex lagopus). Canadian Journal of Zoology, 80, pp. 418–429. JOHNSON, D.D.P., HAY, S.I. and ROGERS, D.J., 1998, Contemporary environmental correlates of endemic bird areas derived from meteorological satellite sensors. Proceedings of the Royal Society of London. Series B, 265, pp. 951–959. JONES, J., 2001, Habitat selection studies in avian ecology: a critical review. Auk, 118, pp. 557–562. JONES, K.B., NEALE, A.C., NASH, M.S., RIITTERS, K.H., WICKHAM, J.D., O’NEILL, R.V. and VAN REMORTEL, R.D., 2000, Landscape correlates of breeding bird richness across the United States Mid-Atlantic region. Environmental Monitoring and Assessment, 63, pp. 159–174. JONES, K.B., NEALE, A.C., WADE, T.G., WICKHAM, J.D., CROSS, C.L., EDMONDS, C.M., LOVELAND, T.R., NASH, M.S., RIITTERS, K.H. and SMITH, E.R., 2001, The consequences of landscape change on ecological resources: an assessment of the United States mid-Atlantic region, 1973–1993. Ecosystem Health, 7, pp. 229–242. JØRGENSEN, A.F. and NØHR, H., 1996, The use of satellite images for mapping of landscape and biological diversity in the Sahel. International. Journal of Remote Sensing, 17, pp. 91–109. JOUVENTIN, P., CAPDEVILLE, D., CUENOT-CHAILLET, F. and BOITEAU, C., 1994, Exploitation of pelagic resources by a non-flying seabird: satellite tracking of the king penguin throughout the breeding cycle. Marine Ecology Progress Series, 106, pp. 11–19. JUSTICE, C.O. and HIERNAUX, P.H.Y., 1986, Monitoring the grasslands of the Sahel using NOAA AVHRR data: Niger 1983. International. Journal of Remote Sensing, 7, pp. 1475–1497. KASTDALEN, L., PEDERSEN, H.C., FJONE, G. and ANDREASSEN, H.P., 2003, Combining resource selection functions and distance sampling: an example with Willow Ptarmigan. In Proceedings of 1st International Conference on Resource Selection by Animals, Laramie, Wyoming, 13–15 January 2003 (Madison, WI: Omnipress), pp. 52–59. KATIBAH, E.F. and GRAVES, W.C., 1978, Remote sensing-aided assessment of wildlife turkey habitat. Proceedings of the Annual William T. Pecora Memorial Symposium Held at Pecora IV, the Application of Remote Sensing Data to Wildlife Management, Souix Falls, 10–12 October 1978, pp. 78–81. KERR, J.T., SOUTHWOOD, T.R.E. and CIHLAR, J., 2001, Remotely sensed habitat diversity predicts butterfly species richness and community similarity in Canada Proceedings of the National Academy of Sciences, USA, 98, pp. 11365–11370. KLAUS, S., SELSAM, P., SUN, Y.-H. and FANG, Y., 2001, Habitat mapping from SPOT images as tool for the conservation of endangered species—Chinese Grouse Bonasa sewerzowi as an example. Naturschutz und Landschaftsplanung, 33, pp. 281–285. KNICK, S.T. and ROTENBERRY, J.T., 1995, Landscape characteristics of fragmented shrub steppe habitats and breeding passerine birds. Conservation Biology, 9, pp. 1059–1071. LAUGA, J. and JOACHIM, J., 1992, Modelling the effects of forest fragmentation on certain species of forest-breeding birds. Landscape Ecology, 6, pp. 183–193. LAVERS, C.P. and HAINES-YOUNG, R.H., 1996a, Using models of bird abundance to predict the impact of current land-use and conservation policies in the Flow Country of Caithness and Sutherland, Northern Scotland. Biological Conservation, 75, pp. 71–77. LAVERS, C.P. and HAINES-YOUNG, R.H., 1996b, The pattern of Dunlin Calidris alpina distribution and abundance in relation to habitat variation in the Flow Country of northern Scotland. Bird Study, 43, pp. 231–239. Avian habitat relationships 2653 LAVERS, C.P. and HAINES-YOUNG, R.H., 1997, The use of satellite imagery to estimate Dunlin Calidirs alpina abundance in Caithness and Sutherland and in the Shetland Islands. Bird Study, 44, pp. 220–226. LAVERS, C.P., HAINES-YOUNG, R.H. and AVERY, M.I., 1996, The habitat associations of Dunlin (Calidris alpina) in the Flow Country of northern Scotland and an improved model for predicting habitat quality. Journal of Applied Ecology, 33, pp. 279–290. LYON, J.G., 1983, Landsat-derived land-cover classifications for locating potential Kestrel nesting habitat. Photogrammetric Engineering and Remote Sensing, 43, pp. 245–250. LYON, J.G., ASCE, A.M., HEINEN, J.T., MEAD, R.A. and ROLLER, N.E.G., 1987, Spatial data for modeling wildlife habitat. Journal of Surveying Engineering, 113, pp. 88–100. MACk, E.L., FIRBANK, L.G., BELLAMY, P.E., HINSLEY, S.A. and VEITCH, N., 1997, The comparison of remotely sensed and ground-based habitat area data using species-area models. Journal of Applied Ecology, 34, pp. 1222–1228. MANLY, B.F.J., MCDONALD, L.L., THOMAS, D.L., MCDONALD, T.L. and ERICKSON, W.P., 2002, Resource Selection by Animals: Statistical Design and Analysis for Field Studies, 2nd edn (Dordrecht: Kluwer Academic). MARCOT, B.G., RAPHAEL, M.G. and BERRY, K.H., 1983, Monitoring wildlife habitat and validation of wildlife–habitat relationships models. Transactions of the North American Wildlife and Natural Resources Conference, 48, pp. 315–329. MARTINKO, E.A., 1981, Monitoring agricultural growth in pronghorn antelope habitat. Proceedings Pecora 7 Memorial Symposium, Sioux Falls, South Dakota, pp. 210–216. MAY, S.M., NAUGLE, D.E. and HIGGINS, K.F., 2002, Effects of land use on nongame wetland birds in Western South Dakota stock ponds, USA. Waterbirds, 25, pp. 51–55. MAYER, A.L. and CAMERON, G.N., 2003, Landscape characteristics, spatial extent, and breeding bird diversity in Ohio, USA. Diversity and Distributions, 9, pp. 297–311. MCGARIGAL, K. and MARKS, B.J., 1994, FRAGSTATS—Spatial Pattern Analysis Program for Quantifying Landscape Structure, Version 2.0, Forest Science Department, Oregon State University, Corvallis. MEYER, C.B., MILLER, S.L. and RALPH, C.J., 2002, Multi-scale landscape and seascape patterns associated with Marbled Murrelet nesting areas on the U.S. west coast. Landscape Ecology, 17, pp. 95–115. MEHNER, H., CUTLER, M., FAIRBAIRN, D. and THOMPSON, G., 2004, Remote sensing of upland vegetation: the potential of high spatial resolution satellite sensors. Global Ecology and Biogeography, 13, pp. 359–369. MILLER, J.R. and CALE, P., 2000, Behavioral mechanisms and habitat use by birds in a fragmented agricultural landscape. Ecological Applications, 10, pp. 1732–1748. MILLER, K.V. and CONROY, M.J., 1990, Spot satellite imagery for mapping Kirtland’s Warbler wintering habitat in the Bahamas. Wildlife Society Bulletin, 18, pp. 252–257. MORRISON, M.L., 1986, Birds populations as indicators of environmental change. Current Ornithology, 3, pp. 429–451. MORRISON, M.L., MARCOT, B.G. and MANNAN, R.W., 1992, Wildlife–habitat Relationships: Concepts and Applications (Madison, WI: University of Wisconsin Press). MORRISON, R.I.G., 1997, The use of remote sensing to evaluate shorebird habitats and populations on Prince Charles Island, Foxe Basin, Canada. Arctic, 50, pp. 55–75. MÖRTBERG, U. and WALLENTINUS, H.-G., 2000, Red-listed forest bird species in an urban environment: assessment of green space corridors. Landscape and Urban Planning, 50, pp. 215–226. NAGENDRA, H., 2001, Using remote sensing to assess biodiversity. International Journal of Remote Sensing, 22, pp. 2377–2400. NEAVE, H.M., CUNNINGHAM, R.B., NORTON, T.W. and NIX, H.A., 1996, Biological inventory for conservation evaluation III. Relationships between birds, vegetation and environmental attributes in southern Australia. Forest Ecology and Management, 85, pp. 197–218. 2654 T. Gottschalk et al. NØHR, H. and JØRGENSEN, A.F., 1997, Mapping of biological diversity in Sahel by means of satellite image analyses and ornithological surveys. Biodiversity and Conservation, 6, pp. 545–566. O’CONNOR, R.J., JONES, M.T., WHITE, D., HUNSAKER, C., LOVELAND, T., JONES, B. and PRESTON, E., 1996, Spatial partitioning of environmental correlates of avian biodiversity in the conterminous United States. Biodiversity Letters, 3, pp. 97–110. O’NEILL, R.V., HUNSAKER, C.T., TIMMINS, S.P., JACKSON, B.L., JONES, K.B., RIITTERS, K.H. and WICKHAM, J.D., 1996, Scale problems in reporting landscape pattern at the regional scale. Landscape Ecology, 11, pp. 169–180. ORMSBY, J.P. and LUNETTA, R.S., 1987, Whitetail deer food availability maps from thematic mapper data. Photogrammetric Engineering and Remote Sensing, 53, pp. 1081–1085. OSBORNE, P.E., ALONSO, J.C. and BRYANT, R.G., 2001, Modelling landscape-scale habitat use using GIS and remote sensing: a case study with Great Bustards. Journal of Applied Ecology, 38, pp. 458–471. PALMEIRIM, J.M., 1985, Using Landsat TM imagery and spatial modelling in automatic habitat evaluation and release site selection for the Ruffed Grouse (Galliformes: Tetraonidae). In Proceedings of the 19th International. Symposium on Remote Sensing of Environment, University of Michigan, Ann Arbor, MI, 2, pp. 729–738. PALMEIRIM, J.M., 1988, Automatic mapping of avian species habitat using satellite imagery. Oikos, 52, pp. 59–68. PEARCE, C.M., 1991, Mapping muskox habitat in the Canadian high arctic with SPOT satellite data. Arctic, 44(suppl. 1), pp. 49–57. PEARCE, J.L. and FERRIER, S., 2000, Evaluating the predictive performance of habitat models developed using logistic regression. Ecology Modelling, 133, pp. 225–245. PEARCE, J.L., CHERRY, K., DRIELSMA, M., FERRIER, S. and WHISH, G., 2001, Incorporating expert opinion and fine-scale vegetation mapping into statistical models of faunal distribution. Journal of Applied Ecology, 38, pp. 412–424. PEARLSTINE, L.G., SMITH, S.E., BRANDT, L.A., ALLEN, C.R., KITCHENS, W.M. and STENBERG, J., 2002, Assessing state-wide biodiversity in the Florida Gap analysis project. Journal of Environmental Management, 66, pp. 127–144. PEARMAN, P.B., 2002, The scale of community structure: habitat variation and avian guilds in tropical forest understory. Ecological Monographs, 72, pp. 19–39. PEARSON, S.M. and SIMONS, T.R., 2002, Spatial analysis of stopover habitats of neotropical migrant birds. In Predicting Species Occurrence: Issues of Accuracy and Scale, J.M. Scott, P.J. Heglund and M.L. Morrison (Eds), pp. 581–592 (Washington, DC: Island Press). PIDGEON, A.M., RADELOFF, V.C. and MATHEWS, N.E., 2003, Landscape-scale patterns of black-throated sparrow (Amphispiza bilineata) abundance and nest success. Ecological Applications, 13, pp. 530–542. PINO, J., RODA, F., RIBAS, J. and PONS, X., 2000, Landscape structure and bird species richness: implications for conservation in rural areas between natural parks. Landscape and Urban Planning, 49, pp. 35–48. PRIEDE, I.G., 1983, Use of satellites in marine biology. In Experimental Biology at Sea, A.G. McDonald and I.G. Priede (Eds), pp. 3–50 (London: Academic Press). RALPH, C.J. and SCOTT, J.M., 1981, Estimating numbers of terrestrial birds. Studies in Avian Biology, No. 6. REES, W.G., 1990, Physical Principles of Remote Sensing (Cambridge: Cambridge University Press). REESE, H., NILSSON, M., SANDSTROM, P. and OLSSON, H., 2002, Applications using estimates of forest parameters derived from satellite and forest inventory data. Computers and Electronics in Agriculture, 37, pp. 37–55. RODHOUSE, P.G., PRINCE, P.A., TRATHAN, P.N., HATFIELD, E.M.C., WATKINS, J.L., BONE, D.G., MURPHY, E.J. and WHITE, M.G., 1996, Cephalopods and mesoscale Avian habitat relationships 2655 oceanography at the Antarctic Polar Front: satellite tracked predators locate pelagic trophic interactions. Marine Ecology Progress Series, 136, pp. 37–50. ROSEBERRY, J.L., RICHARDS, B.J. and HOLLENHORST, T.P., 1994, Assessing the potential impact of Conservation Reserve Program on Bobwhite habitat using remote sensing, GIS, and habitat modelling. Photogrammetric Engineering and Remote Sensing, 60, pp. 1139–1143. ROSEBERRY, J.L. and SUDKAMP, S.D., 1998, Assessing the suitability of landscapes for Northern Bobwhite. Journal of Wildlife Management, 62, pp. 895–902. ROSENSTOCK, S.S., ANDERSON, D.R., GIESEN, K.M., LEUKERING, T. and CARTER, M.F., 2002, Landbird counting techniques: current practices and an alternative. Auk, 19, pp. 46–53. ROSHIER, D.A., ROBERTSON, A.I., KINGSFORD, R.T. and GREEN, D.G., 2001, Continentalscale interactions with temporary resources may explain the paradox of large populations of desert waterbirds in Australia. Landscape Ecology, 16, pp. 547–556. SADER, S.A., POWELL, G.V.N. and RAPPOLE, J.H., 1991, Migratory bird habitat monitoring through remote sensing. International Journal of Remote Sensing, 12, pp. 363–372. SAVERAID, E.H., DEBINSKI, D.M., KINDSCHER, K. and JAKUBAUSKAS, M.E., 2001, A comparison of satellite data and landscape variables in predicting bird species occurrences in the Greater Yellowstone Ecosystem, USA. Landscape Ecology, 16, pp. 71–83. SAXON, E.C., 1983, Mapping the habitats of rare animals in the Tanami Wildlife Sanctuary (central Australia): An application of satellite imagery. Biological Conservation, 27, pp. 243–257. SCHNEIDER, D., 1994, Quantitative Ecology, Spatial and Temporal Scaling (New York: Academic Press). SCHWALLER, M.R., BENNINGHOFF, W.S. and OLSON, C.E., 1984, Prospects for satellite remote sensing of Adelie penguin rookeries. International Journal of Remote Sensing, 5, pp. 849–853. SCHWALLER, M.R., OLSON, C.E., BENNINGHOFF, W.S. and WEHNES, K.M., 1986, Preliminary report on satellite remote sensing of adélie penguin rookeries. Antarctic Journal, 21, pp. 205–206. SCHWALLER, M.R., OLSON, C.E., MA, Z., ZHILIANG, Z. and DAHMER, P., 1989, A Remote sensing analysis of Adélie Penguin rookeries. Remote Sensing of Environment, 28, pp. 199–206. SEOANE, J., BUSTAMANTE, J. and DÍAZ-DELGADO, R., 2004, Are existing vegetation maps adequate to predict bird distributions? Ecological Modelling, 175, pp. 137–149. SHAPIRO, A.-M., HARPER, S.J. and WESTERVELT, J., 2002, Modeling Cowbird occurrences and parasitism rates: statistical and individual-based approaches. In Predicting Species Occurrence: Issues of Accuracy and Scale, J.M. Scott, P.J. Heglund and M.L. Morrison (Eds), pp. 547–558 (Washington, DC: Island Press). SHAW, D.M. and ATKINSON, S.F., 1988, GIS applications for Golden-cheeked Warbler habitat description. Proceedings of the 1988 GIS/LIS conference, San Antonio, TX (Falls Church, VA: American Society for Photogrammetry and Remote Sensing), pp. 401–406. SHRINER, S.A., SIMONS, T.R. and FARNSWORTH, G.L., 2002, A GIS-based habitat model for Wood Thrush Hylocichla mustelina in Great Smoky Mountains National Park. In Predicting Species Occurrence: Issues of Accuracy and Scale, J.M. Scott, P.J. Heglund and M.L. Morrison (Eds), pp. 529–535 (Washington, DC: Island Press). STEELE, B.B., BAYN, R.L. and GRANT, C.V., 1984, Environmental monitoring using populations of birds and small mammals: analyses of sampling effort. Biological Conservation, 30, pp. 157–172. STOCKWELL, D.R.B. and PETERS, D.P., 1999, The GARP modeling system: problems and solutions to automated spatial prediction. International Journal of Geographic Information Systems, 13, pp. 143–158. 2656 Avian habitat relationships STOMS, D.M., DAVIS, F.W. and COGAN, C.B., 1992, Sensitivity of wildlife habitat models to uncertainties in GIS data. Photogrammetric Engineering and Remote Sensing, 58, pp. 843–850. SUAREZ-SEOANE, S., OSBORNE, P.E. and ALONSO, J.C., 2002, Large-scale habitat selection by agricultural steppe birds in Spain: identifying species-habitat responses using generalized additive models. Journal of Applied Ecology, 39, pp. 755–771. THIBAULT, D., CHALIFOUX, S. and LAPERLE, M., 1998, Using satellite imagery as a planning tool for Harlequin duck inventory. International Journal of Remote Sensing, 19, pp. 5–9. TUCKER, K., RUSHTON, S.P., SANDERSON, R.A., MARTIN, E.B. and BLAIKLOCK, J., 1997, Modelling bird distributions—a combined GIS and Bayesian rule-based approach. Landscape Ecology, 12, pp. 77–93. VANDER HAEGEN, W.M., DOBLER, F.C. and PIERCE, D.J., 2000, Shrubsteppe bird response to habitat and landscape variables in eastern Washington, USA. Conservation Biology, 14, pp. 1145–1160. VENIER, L.A., PEARCE, J., MCKEE, J.E., MCKENNEY, D.W. and NIEMI, G.J., 2004, Climate and satellite-derived land cover for predicting breeding bird distribution in the Great Lakes Basin. Journal of Biogeography, 31, pp. 315–331. VERLINDEN, A. and MASOGO, R., 1997, Satellite remote sensing of habitat suitability for ungulates and ostrich in the Kalahari in Botswana. Journal of Arid Environments, 35, pp. 563–574. WALLIN, D.O., ELLIOTT, C.C.H., SHUGART, H.H., TUCKER, C.J. and WILHELMI, F., 1992, Satellite remote sensing of breeding habitat for an African weaver-bird. Landscape Ecology, 7, pp. 87–99. WARKENTIN, I.G., GREENBERG, R. and ORTIZ, J.S., 1995, Songbird use of gallery woodlands in recently cleared and older settled landscapes of the Selva Lacandona, Chiapas, Mexico. Conservation Biology, 9, pp. 1095–1106. WATZKE, H., LITZBARSKI, H., OPARINA, O.S. and OPARIN, M.L., 2001, The migration of Great Bustards Otis tarda from the Saratov region (Russia)—first results of a satellite tracking study. Vogelwelt, 122, pp. 89–94. WHITE, D., MINOTTI, P.G., BARCZAK, M.J., SIFNEOS, J.C., FREEMARK, K.E., SANTELMANN, M.V., STEINITZ, C.F., KIESTER, A.R. and PRESTON, E.M., 1997, Assessing risks to biodiversity from future landscape change. Conservation Biology, 11, pp. 349–360. YATES, M.G., GOSS-CUSTARD, J.D. and RISPIN, W.E., 1996, Towards predicting the effect of loss of intertidal feeding areas on overwintering shorebirds (Charadrii) and shelduck (Tadorna tadorna): refinements and tests of a model developed for the Wash, east England. Journal of Applied Ecology, 33, pp. 944–954. YEN, P.P.W., HUETTMANN, F. and COOKE, F., 2004, A large-scale model for the at-sea distribution and abundance of Marbled Murrelets (Brachyramphus marmoratus) during the breeding season in coastal British Columbia, Canada. Ecological Modelling, 171, pp. 395–413. YOUNG, T.N., EBY, J.R., ALLEN, H.L., HEWITT, M.J. and DIXON, K.R., 1987, Wildlife habitat analysis using Landsat and radiotelemetry in a GIS with application to Spotted Owl preference for old growth. Proceedings of the 1987 GIS/LIS conference, San Francisco, CA (Falls Church, VA: American Society for Photogrammetry and Remote Sensing), pp. 595–600.