Walker 1962). Other cyclical patterns of sound production in insects throughout the year relate to the phenological life cycle of the species. Annual cicadas (Tibicen spp.), for instance, will sing during hot days, late in the summer after they emerge from the ground, with the timing of emergence being a function of accumulated heating degree days (Williams and Simon 1995). Sounds produced from wing beats from flies, bees, and wasps could contribute significantly to the soundscapes if these insects are present in large numbers.

Amphibians such as frogs and toads rely primarily on vocalizations to attract mates (Gerhardt 1994). In the northern temperate regions of eastern North America, spring peepers (Pseudacris crucifer) are common singers at night in wetlands and ponds. Calls are intense during the breeding seasons, which extend from late winter (February) to early spring (May) in the northern United States and from late fall (October) to early spring (March) in more southern locations. Frequencies of frog and toad choruses range from 2 to 5 kHz.

Almost all birds use sound to attract mates, defend territories, sound alarms, and communicate other types of information. Many of the passerines are especially known for producing elaborate songs (Kroodsma 2005). Most songs and calls produced by birds occur in the 2 to 6 kHz range. The acoustic frequency of a bird’s song relates to its body size (large-bodied birds produce sounds as low as 1 kHz) and habitat type and structure; for example, some tropical birds use protracted pure tones in environments with persistent geophonic sounds of wind and rain, and some vocalizations reach frequencies in the 10 to 12 kHz range (Kerry Rabenold, Department of Biological Sciences, Purdue University, West Lafayette, Indiana, personal communication, 5 October 2010).

A variety of terrestrial mammals also produce sounds (McComb and Reby 2005). Groups that are frequent contributors of sound produced in landscapes include primates (e.g., monkeys, baboons), elephants, canines (e.g., wolves and coyotes), rodents (e.g., squirrels, chipmunks), and felines (e.g., lions), among others. Bats generally produce two types of sound; the first, referred to as “echolocation,” is emitted as ultrasonic frequencies (above human hearing ability) and is used to locate prey. The second, communication calls, are more readily audible to humans and are used to identify individuals.

Recently, considerable evidence has emerged showing that anthrophony can influence animal communication in a variety of ways. For example, American robins (Turdus migratorius) shift the timing of their singing in urban environments to the night (Fuller et al. 2007). In song sparrows (Melospiza melodia), the lowest-frequency notes were higher in environments with high ambient noise (Wood and Yezerinac 2006). Brumm (2004) found that free-ranging nightingales (Luscinia megarhynchos) in noisier environments sing more loudly than those in quieter environments, and Slabbekoorn and Peet (2003) determined that the great tit (Parus major) sings at higher pitches in urban noise conditions.

Rhythms of nature. The sounds of nature contain numerous rhythms or cycles. Many recognized temporal cycles of communication occur in terrestrial animals, the most well studied being those of birds, amphibians, and insects. Collectively, we refer to these periodic acoustic patterns as “the rhythms of nature.” Most songbirds are known to begin singing at the same time each year (Saunders 1947), and these birds sing most intensely early in the morning (Kacelnik and Krebs 1982) and late evening (referred to as the dawn and dusk chorus, respectively). Dawn chorus in birds is thought to occur when individuals, arriving back to their territory, use songs to advertise their presence (Staicer et al. 1996). This circadian pattern of singing in birds, the timing of which is largely affected by weather and climatic conditions, strongly correlates with sunrise and sunset and becomes more pronounced with the onset of breeding and migration.

A research agenda for soundscape ecology

We believe that we are now well poised to place soundscape ecology into a more research and application focus. Research is needed in several new areas, organized around the following main themes: measurement and quantification, spatial-temporal dynamics, environmental covariates, human impacts on soundscapes, soundscape impacts on humans, and soundscape impacts on wildlife.

Theme 1: Improve the measurement and quantification of sounds. Acoustic sensors are needed that can automate the recording of sounds, that are inexpensive, and that can be placed in large networks in hostile environments. Research is required that can automatically differentiate all sounds emanating from landscapes. For example, researchers need tools that can classify biological, geophysical, and anthropogenic sources of sounds. Scientists also need a better understanding of how these sources of sounds differ in their composition. How do anthrophonic sounds differ in composition (acoustic frequency, time interval) from biophonic sounds? Is the presence of certain kinds of sounds indicative of a healthy or deteriorating landscape? In situ measurements of biodiversity need to be compared with soundscape measures to determine how well vocal organisms provide a proxy for biodiversity in general. Research in this area can also advance our ability to use soundscape measures for natural resource management and biological conservation.

Theme 2: Improve our understanding of spatial-temporal dynamics across different scales. Research is needed on how soundscapes vary with landscape patterns and processes (figure 1, arrows 1 and 2). How do soundscapes differ with land-use patterns? Comparisons of soundscape dynamics should be made of various natural ecosystems around the world but also across areas that differ in the amount of human disturbance within an ecosystem. Vertebrate species richness has been shown to vary with vegetation structure (canopy height, density). Is soundscape diversity greatest where vegetation structure is most complex? More research is needed that attempts to characterize the different types of the temporal patterns of soundscapes. How do soundscapes vary over different time frames (seconds, minutes, hours, diurnally, annually) (figure 1, arrows 4 and 5) in different landscapes? How are the dawn and dusk choruses affected by human activities?

Theme 3: Improve our understanding of how important environmental covariates impact sound. Biophonic and geophonic sounds very likely vary according to many environmental factors, such as weather, plant phenology, and elevation. Specific research is needed on how soundscapes vary by temperature (air, soil, and water), solar radiation, lunar radiation, relative humidity, heating degree days, and moisture budgets (figure 1, arrow 4). Knowledge of these covariates will be necessary as researchers attempt to understand how human activity impacts natural soundscape dynamics. Studies on how geophonic sounds of wind, running water, and rain affect biophonic patterns will help us to understand the plasticity of biological communication as it relates to human-generated sounds.

Theme 4: Assess the impact of soundscapes on wildlife. There is a need for more research on how certain soundscape qualities (e.g., noise, ambient sounds like running water and wind) affect individual wildlife species and populations (figure 1, arrow 6). Research is required on the ways anthrophony affects wildlife behavior, such as breeding, predator-prey relationships, and physiology. As soundscape patterns such as signal composition, sound diversity, and temporal cycles change, what are the impacts to species’ life-history patterns?

Theme 5: Assess the impacts of humans on soundscapes. Humans create many objects that produce sounds (figure 1, arrow 5). How do engines, road noises, bells, sirens, and other machines affect soundscape composition? As new technologies emerge, how do these affect the soundscape? What policies are needed to protect soundscapes in various settings such as national parks or our cities and neighborhoods? How can land-use planners and policymakers determine future soundscapes?

Theme 6: Assess soundscape impacts on humans. Humans are surrounded by sounds that emanate from the environment and these sensory connections to nature are from the soundscape (figure 1, arrow 7). Research is needed on how natural sounds influence the development of individuals’ sense of place, place attachment, and connection to nature. More specifically, how do human demographic variables such as culture, place of residence, or age affect the strength of human values associated with soundscapes? What factors affect human (in)tolerance of soundscape changes, especially where those changes increase noise?

Soundscape ecology case studies

We present four case studies that illustrate various aspects of soundscape ecology. These studies also exemplify the kinds of research that can be conducted across the six research themes posed above. The first case study, which is not a separate study in itself as are the three others, represents selected recordings from the massive Krause 40-year-old soundscape archive. Krause, a musician and recording engineer, has recorded natural sounds for use in the entertainment industry. The second focuses on characterizing the “rhythms of nature” in midlatitude landscapes that vary across a human disturbance gradient. A third study, conducted in Sequoia National Park in the United States, attempts to determine whether organisms are partitioning their sounds and the extent to which geophonic sounds, such as rivers and wind, interfere with animal communication. The final study, conducted in montane forests in Tuscany, Italy, centers on mapping dynamic soundscapes.

Krause ambient sounds soundscape archive. We use several field recordings that are part of the massive Krause soundscape archive to illustrate how sounds reflect certain characteristics of landscapes and the organisms that live within them. A 1-minute-28-second recording of a tropical forest in Madagascar in 1996 (sound file 6) represents an excellent example of the ANH, exemplifying that sounds produced by animals are separated in space, time, and frequency. Here, dozens of birds vocalize with little frequency or temporal overlap. One bird (probably a sickle-billed vanga, Falculea palliata) produces four rapid calls followed by a brief pause at 1 kHz, much below the frequency of other bird vocalizations. This recording most likely represents some of the greatest acoustic niche separation in the world.

The nighttime recording of organisms producing sounds in a bai in the Central African Republic (sound file 7) illustrates how unique landscapes can create unique soundscapes. Here, the normal synchronous production of nighttime sounds by insects and frogs is interwoven with the loud trumpeting, bellowing, and grunting of forest elephants (Loxodonta cyclotis). A bai is a special landscape where forest elephants go (areas have been cleared by elephants) because of the high salt content of the mud surrounding ponds created by groundwater upwelling; thus, landscape structure and the specific animals occupying these areas can create a unique soundscape.

A recording (sound file 8) of the dawn chorus in Zimbabwe illustrates not only the complexity of sounds produced in the morning but also animals’ use of special landforms to propagate calls. The first minute contains a typical chorusing of about 30 different species of birds (see supplementary online materials at www.jstor.org/stable/10.1525/bio.2011.61.3.6). At 1:13 into this recording, however, baboons (Papio cynocephalus) begin to bark. Note how the echo decay of the baboons (> 4 seconds) differs from the echo decay of the birds (approximately one-third of a second), such as the black-eyed bulbul (Pycnonotus barbatus) in the dry forest. The landform is thus exploited by these animals to propagate their voices. Many animals, such as African lions (Panthera leo), forest and plains elephants, and hyenas, choose the time and place to make their voices echo.

Wiens and Milne (1989), among others, have emphasized the need to understand landscapes from the perspective of the size of an organism; they found that from a beetle’s point of view, the very fine structure of a landscape influences movement patterns. Additionally, many insects produce sounds that aid in breeding or communication that may not be audible to humans or to other organisms in the landscape. Ant stridulations (sound file 9) are not audible to the human ear but are known to occur continuously in ant colonies. Recent research (Hickling and Brown 2000) has also shown that only sounds produced in a near field on the order of 100 millimeters or less are detected by ants, and ambient sounds produced farther away are ignored.

Sounds produced by many organisms may also reflect the animals’ complex social structure. The recording (sound file 10) of gray wolves (Canis lupus) in Canada’s Algonquin Provincial Park in 2008 captures the vocalizations of wolves as the normal foreground biophony progresses. This recording may also elicit a strong sense of wildness, triggering many human senses and values. The entire context of wolves howling among the tapestry of boreal sounds can be a memorable experience (sensuFisher 1998), emphasizing the importance of our auditory connection with nature.

Tippecanoe Rhythms of Nature study. Several of this article’s authors (LJV, BCP, and BMN) conducted a yearlong study to measure near-continuous sounds in a variety of landscapes in northwestern Tippecanoe County, Indiana (see online supplementary material at www.jstor.org/stable/10.1525/bio.2011.61.3.6), in order to characterize different rhythms of nature and the impacts of humans on them. We deployed automated Wildlife Acoustics Songmeters in eight locations that varied in land-use characteristics, spanning old growth forest to agricultural fields (figure 5). The proportion of urban and agriculture within 100 meters of the recorder was used as a measure of human disturbance. We collected and analyzed more than 34,000 15-minute recordings. We were also interested in applying metrics traditionally used by ecologists, such as diversity, evenness, richness, and dominance. To accomplish this, we discretized the spectrogram into 10 frequency bands and calculated the amount of sound occurring in each band. We used these values to calculate (a) diversity (using Shannon’s index) and (b) evenness (using the Gini coefficient). We also determined the most dominant frequency band occurring in each 15-minute recording. The total amount of acoustic activity in each recording was used as a surrogate for sound sources, which in some cases will be correlated to species richness. These metrics were examined across landscapes and over two time periods.

Activity, diversity, and evenness were greatest for the natural landscapes (forests and wetlands), and all values decreased as human disturbance increased (figure 6). A plot of mean monthly Shannon’s diversity index values by site (figure 7a) shows that a peak in entropy occurs during the late summer. Late summer soundscapes are composed of birds and insects (mostly cicadas and crickets). Comparing these same sites across time of day (figure 7b) aggregated from May through September, a 7:00 a.m. (i.e., dawn chorus) and 10:00 p.m. peak (i.e., dusk chorus) are evident in all but the agricultural sites. Nighttime entropy values are twice that of midday values in all sites except the cornfield site. Sound files 11–34 contain a full day of recording from our wetland site. In May, all sites were dominated by low-frequency sounds (figure 8), but by late summer (August and September) bands 3 through 8 became prominent, especially in natural landscapes.

Sequoia National Park acoustic niche hypothesis study. Four relatively pristine habitats located in the Sequoia National Park were selected by BK and SHG (see figure 9) for a yearlong study to determine whether (a) sounds from animals occurred with any acoustic niche separation, and (b) geophony affected biophony patterns. A forest riparian zone (near a relatively noisy stream), an oak savanna, a dry savanna chaparral (with high winds), and an old-growth forest site were monitored daily at dawn, midday, dusk, and midnight (for 60 minutes during the period of September 2001 through October 2002) using digital acoustic recorders (see supplementary online materials for details). Randomly selected 11.5-second segments were analyzed by examining spectrograms and listening to the recordings. A total of 190 spectrograms were produced, and the vocal niches in these spectrograms were analyzed (a) qualitatively, by describing biophonic and geophonic patterns; and (b) quantitatively, by calculating the acoustic activity occurring at each site.

The vocalizations of American robins and the American dippers (Cinclus mexicanus) in the riparian zone (Buckeye Flats) location were evident, with frequencies of songs occurring in a manner that avoided masking by the nearby noisy stream. Insects produced sounds that were higher in pitch than birds, demonstrating niche partitioning. Only 57% of the spectrograms contained sounds. Within the oak savanna site (Sycamore Creek), vocalizations by birds ranged from 500 Hz (mourning dove, Zenaida macroura) to more than 20 kHz (unidentified bird); approximately 94% of the spectrogram was occupied by at least one vocal organism. The dry savanna chaparral (Shepard’s Saddle) site possessed the greatest diversity of sounds, from flies (200 Hz) to birds (8.7 kHz from one unidentified bird). Finally, in the old-growth site (Crescent Meadow), animals produced sounds from 200 Hz (flies) to around 9 kHz (birds); about 82% of the spectrogram was occupied by vocalizations. Frogs chorused between 600 Hz to 2 kHz, just below the acoustic frequency of the robin, which sings in the 2 to 3.3 kHz range. Sound files 35–38 contain sample dawn chorus recordings from this study. The amount of acoustic activity for each site (figure 9a) shows that the Buckeye Flats site contained more than 10 times the amount of acoustic activity, mostly from the geophonic sounds of the stream. Within each site (figure 9b), acoustic activity was highly variable over a season; fall, in half of the cases, was the most acoustically active season (see sample sound files 35–38).

Mapping the soundscapes in the Tuscany study. A two-month study was conducted from June to July of 2008 in a secondary montane beech forest in the Italian Apennine National Park, located along the northern slopes of Mount La Nuda. The study was conducted to determine how spatially variable soundscapes are in a relatively homogenous forest. Twenty digital recorders (Handy Recorder, H4) were placed in a 5 × 4 grid with 100-meter spacing. Eleven three-hour recordings (0600 to 0900) were collected under ideal meteorological conditions. Approximately 13 species of birds, such as the European robin (Erithacus rubecula), the chaffinch (Fringilla coelebs), and the blackcap (Sylvia atricapilla), vocalize in this forest (sample in sound file 39). An acoustic complexity index (see supplementary online materials) was used to quantify spectral complexity, and interpolation software was used to create soundscape maps.

Data from the acoustic recorders were used to construct soundtopes (Farina 2006)—a three-dimensional map of acoustic complexity (y axis) plotted across the landscape (plotted across the x and z axes). The 11 daily soundscape maps for this landscape (figure 10) indicate that large interseasonal changes of the soundscape occur. We anticipated that the soundscape maps would be similar throughout the year, reflecting static territorial boundaries. The breeding period of every species has a different phenological time and for each time requires specific resources (food, shelter, singing spots, etc.), these resources are spatially and temporally variable as well. The soundtope shifts across the environment consequently.

Summary of case studies. The above case studies illustrate various ways that data can be collected, analyzed, and interpreted. These case studies highlight many of the research themes described above. The Krause archive demonstrates the complex composition of a community of organism vocalizations, the interaction of landscape features and sound propagation, and the importance of an organism’s perception of scale in the landscape in which it lives. The Tippecanoe study shows that temporal patterns of soundscapes exhibit strong dawn and dusk chorus peaks that diminish with increasing human disturbance on the landscape. The Sequoia study attempts to quantify the effects of geophony on biophonic patterns, and shows that animals that communicate in each habitat do so at different frequencies to avoid overlap. Lastly, the Tuscany soundscape mapping study illustrates that soundtopes constructed from acoustic arrays could be used to quantify the spatial dynamics of soundscapes.

The way forward

The study of soundscapes can yield valuable information about the dynamics of a variety of landscapes. Given that technological advances are occurring rapidly and theories about the interplay of patterns and processes occurring within landscapes are maturing, we believe that soundscape ecology can enhance our understanding of how humans affect ecosystems. Indeed, we are at a critical juncture in our history, and there is a need for transformative approaches that help us to more thoroughly elucidate how humans affect our planet (Vitousek et al. 1997, Chapin et al. 2000).

At present, there is a renewed interest in studying ecosystems at large, continental scales. Automated acoustic recordings could provide a means to collect information at fine temporal resolutions (Porter et al. 2005). Initiatives such as the National Science Foundation’s NEON (National Ecological Observatory Network) project are being built to study ecosystems at subcontinental scales (Keller et al. 2008). Furthermore, recordings made today will become tomorrow’s “acoustic fossils,” possibly preserving the only evidence we have of ecosystems that may vanish in the future because of a lack of desire or ability to protect them.

We also argue that society should value natural soundscapes as it does other aspects of nature. Soundscapes represent the heritage of our planet’s acoustic biodiversity, and reflect Earth’s natural assemblage of organisms—soundscapes are an ecosystem service (MA 2005) that provides cultural and other services. Natural sounds are our auditory link to nature, and the trends toward increasing society’s “nature deficient disorder” (Louv 2008) are likely to continue as we replace natural sounds with those made by humans. This research reflects again on Rachel Carson’s call made in Silent Spring, in which she highlighted the dangers of pesticides and their potential threat to wildlife and the environment. The unintended silencing of organisms by a myriad of human activities provides yet another indication of our impact on the planet’s ecosystems.

Acknowledgments

This article emerged from two symposia held at the US International Association of Landscape Ecology (IALE) meetings, one in 2009 in Snowbird, Utah, and the second in 2010 in Athens, Georgia. Funding for the Tippecanoe Soundscape Study was obtained by a grant from the Lilly Foundation for the establishment of a Center for the Environment at Purdue. Development of acoustic metrics for the Tippecanoe Rhythms of Nature Study was made possible from a National Science Foundation III-XT grant to BP. Funds to support LV, BN, and SD were provided by the Department of Forestry and Natural Resources at Purdue University, and travel funds for LV and BN to attend the US IALE meetings in 2009 and 2008 respectively, were provided by the NASA Michigan State University US IALE travel fund. The authors greatly acknowledge the input on an earlier version of this manuscript by Barny Dunning, Lori Ivans, Kerry Rabenold, Jeff Holland, Jeff Dukes, Jim Plourde, Stuart Bolton, Kimberly Robinson, Camille Washington-Ottombre, Alex Pijanowski, Burak Pekin, and Christian Perry. Jarrod Doucette assisted with the graphics.

References cited