## Overview

In this exercise you’ll be surveying ambiant sounds in and around Leiden. We’re interested in how ambiant conditions acosutic conditions change with respect to:

• Water Depth
• Environmental conditions (sediment type, nearby surroundings, underwater vegetation, etc).

There are two groups, and each group will focus on one of the two goals. The objectives of each group are:

• Group A: Find locations where there is variable water depth, but all other site charactersitics are very similar. This will allow us to examine how water depth affects the cutoff frequencies for ambiant sound.
• Group B: Find locations with similar water depths, but all other conditions are different (such as pond in a park versus a canal of similar water depth). This will allow us to see how different habitat types vary in their ambiant acoustic characteristics.

For each group, you’ll need to take calibrated measurements with a hydrophone. For information regarding your calibration values, how to use them, check this pdf document. This document will let you know what recorder settings you need to use and the calibration values associated with it.

Each group should make some 30 second recordings at approximately mid-depth. This should be long enough to get a sense of the ambiant acoustic conditions in the area.

## Analysis

Once you’ve loaded your sound into R (following the guidelines in the pdf), you can now make a PSD to show the resulting spectrum. To do this, we’ll need to install some packages:

# Install devtools (this will be used to load ezPSD)
install.packages("devtools")
require(devtools)

# Install ezPSD
install_github("RTbecard/ezPSD")

Now we can make some Power Spectral Density Plots (PSDs). In this example I’ll use simulated data.

#Load ezPSD
require(ezPSD)
## Loading required package: ezPSD
# Create simulated ambiant sound
wav <- rnorm(n = 44100*10, mean = 0, sd = 1e6)

# Calculate SPL
20*log10(rms(wav))
## [1] 120.0007
# Make PSD
psd <- ezWelch(wav, fs = 44100, wl = 512)
ggpsd(psd)
## Loading required package: ggplot2

The term fs refers to the sampling rate of the recording device (typically 44100Hz). wl is the window length.
A higher window length results in a higher resolution plot, with the tradeoff of being more variable. The black line in the plot is what we want to pay attention to. This is the mean power spectral density for a given frequency. In shallow water, we may expect to see a sharp decline in acoustic power below a certain cutoff frequency. Also, anthropogenic noise tends to be focued in the lower frequencies, so we may notice more power in lower frequencies in sites heavily affected by anthropogenic activity.