The Science of Sound- Soundscape Monitoring at Homestead National Monument

Many people are familiar with the concept of a landscape, and yet they have given no thought to the soundscape. The soundscape of an area is important as the landscape, who wants to go into the wild and see an amazing view, only to hear constant traffic and man-made noises. The National Park Service created the Natural Sounds and Night Skies division to help parks manage their soundscapes.  I was lucky enough to attend a training with some of the NPS Natural Sounds staff in Denver back in May of this year. The training explained the mission of the Natural Sounds program, trained us in the set-up and take-down of acoustical monitoring equipment, and the analysis of all the data collected.

The National Park Service Management Policy from 2006 states in section 4.9 : Park natural soundscape resources encompass all the natural sounds that occur in parks, including the physical capacity for transmitting those natural sounds and the interrelationships among park natural sounds of different frequencies and volumes. Natural sounds occur within and beyond the range of sounds that humans can perceive, and they can be transmitted through air, water, or solid materials. The National Park Service will preserve, to the greatest extent possible, the natural soundscapes of parks. The Service will restore to the natural condition wherever possible those park soundscapes that have become degraded by unnatural sounds (noise), and will protect natural soundscapes from unacceptable impacts. (http://www.nps.gov/policy/mp/policies.html#_Toc157232745) All of the data collected allows parks to establish base line data and help them to create management plans to protect soundscape and help to restore soundscapes in areas where they have been degraded. 

Soundscapes are important for many reasons, among them allowing animals to hunt and visitors to appreciate the sights and sounds of the parks. If you are interested in more information about the importance of sound please visit the NPS Natural Sounds and Night Skies website at: http://www.nature.nps.gov/sound/

Homestead National Monument of America realized the importance of sounds on the monument and reached out to the NPS Natural Sounds and Night Skies program. The monument purchased the necessary equipment to begin gathering acoustical data. The equipment included a SPL (Sound Pressure Level) meter, microphone, mp3 recorder, anemometer (wind), battery and tripods. In August of 2011, a technician from the Natural sounds program came to the monument to teach Natural Resource Management Specialist Jesse Bolli how to set up/take down the equipment and transfer the data to the computer. 

Jesse began to collect data from various sites during the fall and winter. The equipment normally gathers data for approximately 20 days at each site. Jesse set up the equipment 5 times before I arrived on site and we have set it up together once since I received my training. This left me with 6 sets of data to analyze. The NPS Natural Sounds program has developed a program call the Acoustical Monitoring Toolbox. I use many of the features of this program to turn the raw data into something more understandable. The next four graphics are spectrograms for one day at two different sites. I created these images using the AMT (Acoustical Monitoring Toolbox). The first image is of a site close to the road. The more yellow in an image indicates that it is loud, blue colors indicate quieter sounds. All of the white/yellow spikes indicate when a vehicle passed. The second graphic is for the same day at the some site, but is corrected to show only the frequencies that the human ear can pick up. The last two graphics are for a site that is in the woods and much further from the road (site 4). The fact that there is much more blue than yellow shows that the second site is much quieter than the first site. It is also a good example of how it is louder during the day time hours than in the nighttime hours.

Spectrogram- Site 1

Spectrogram (A-weighted) adjusted for human hearing, Site 1

Spectrogram, Site 4

Spectrogram (A-weighted) adjusted for human hearing, Site 4

This graphic was created using the AMT and a program called 'R' running a script created by the Natural Sounds program. It compares how loud a noise is and at what frequency it is at. These graphs are interesting to look at as they split up day and night and show ranges where transportation, conversation, and song birds fall.

The biggest part of data analysis is listening to the sounds that were recorded. For each site 8 days are analyzed. The AMT has a program that splits the audio into 10 seconds clips for every 2 minutes, meaning that you listened to 2 hours of audio per day analyzed. The audio is then listened to in a program call the Listening Center, which allows the listener to assign a number to each individual sound. Once numbers are assigned, the data can be ran through the 'R' program to create graphic that can show what percentage of time trucks or birds or any specific sound was heard.

Screen shot of Listening Center.

 The anemometer gathers wind speed data, which is important in choosing what days are analyzed. Days with little wind are much easier to listen to. It also allows us to see what days or times may not be analyzed by AMT software, since days with wind speeds above 5 m/s (meter per second) are not analyzed.

Wind Data

The following two tables were created by the Natural Sounds program as a base to compare site specific numbers to.

Sound pressure level examples
Park Sound Sources	Common Sound Sources	dBA
Volcano crater (HALE)	Human breathing at 3m	10
Leaves rustling (CANY)	Whispering	20
Crickets at 5m (ZION)	Residential area at night	40
Conversation at 5m (WHMI)	Busy restaurant	60
Snowcoach at 30m (YELL)	Curbside of busy street	80
Thunder (ARCH)	Jackhammer at 2m	100
Military jet at 100m AGL(YUCH)	Train horn at 1m	120

Effects of sound pressure levels on humans
SPL (dBA)	Relevance
35	Blood pressure and heart rate increase in sleeping humans (Haralabidis et al., 2008)
45	World Health Organization’s recommendation for maximum noise levels inside bedrooms (Berglund, Lindvall, and Schwela, 1999)
52	Speech interference for interpretive programs (U.S. Environmental Protection Agency, 1974)
60	Speech interruption for normal conversation (U.S. Environmental Protection Agency, 1974)

I created the following two tables using the data collected by the SPL (Sound Pressure Level) meter. The 20-1,250 range is the lower frequencies, which consists of mainly traffic, while the 12.5-20,000 includes all of the frequencies that were collected. The first table shows the percent time that noise was above a certain level, it is broken up by day an nigh hours and shows data for two different seasons at the same site.

Percent time above for summer and winter
Site	Frequency (Hz)	% Time above sound level: 0700 to 1900				% Time above sound level: 1900to 0700
		35dBA	45dBA	52dBA	60dBA	35dBa	45dBA	52dBA	60dBA
HOME001 Winter	20-1,250	88.2	15.84	1.73	0.04	83.40	9.33	0.49	0.01
	12.5-20,000	92.02	18.77	2.21	0.05	85.15	11.09	0.58	0.01
HOME001 Summer	20-1,250	93.87	10.71	1.56	0.03	69.13	2.78	0.15	0.00
	12.5-20,000	100	99.93	73.81	0.32	100	100	91.76	28.49

Exceedence levels (Lx) represent the sound pressure levels exceeded x percent during the  given measurement period (e.g. L90 is the SPL that has been exceeded 90% of the time). For the following table, the top value reports the Lx for the 20 – 1,250 Hz subset of the frequency range, and the bottom Lx value is calculated from the 12.5 – 20,000 Hz spectrum.  Exceedence levels for existing conditions
Site	Frequency (Hz)	Exceedence levels (dBA): 0700 to 1900			Exceedence levels (dBA): 1900 to 0700
		L90	L50	L10	L90	L50	L10
HOME001 Winter	20-1,250	35.2	40.1	45.5	35.7	38.8	43.3
	12.5-20,000	35.8	40.7	46.4	35.9	39.2	43.7
HOME001 Summer	20-1,250	36.0	39.1	44.6	34.8	36.6	40.3
	12.5-20,000	51.9	53.6	55.2	57.9	59.0	60.0