Here at AEI, one of the fun tasks on my plate is writing lay summaries of new scientific research.  Usually.  Early in 2008, a dense volume of the journal Aquatic Mammals was published, which featured the results of a multi-year effort by an all-star team of American ocean noise researchers, who were attempting to distill all the current research on ocean noise, and to recommend Exposure Criteria for marine mammals.  Suffice to say, I read it several times, highlighting madly, but kept putting it aside, reticent to attempt a coherent narrative summary.  

Well, I finally followed through, and what follows (below the fold) is a pretty decent summation of what they came up with.  The headline news is twofold: in addressing noise that may cause physical injury (defined as permanent hearing loss), the authors present a dizzying array of extrapolations and assumptions (largely precautionary but sometimes pure leaps of faith) in order to try to assess the impact of extremely loud sound on marine mammals, given that there is very very little direct data to work with.  They conclude that safety limits could be modestly increased without deafening more whales.  On the behavioral side of the ledger, things are not that much clearer, but much more fascinating.  A series of charts that compile results from all known behavioral response observations highlight the wide range of responses that a given level of sound may cause, but also provide some solid evidence that many marine mammals show fairly dramatic behavioral change when encountering fairly modest sound levels, far below those that current regulations consider necessary to monitor.  With that, if you want to know more, I invite you to click on through….

US Ocean Noise Panel Recommends New Limits to Avoid Injury; Behavioral Responses Too Varied to Set Limits, but Show Significant Responses at Moderate Noise Levels
A Special Issue of the journal Aquatic Mammals
$12 on CD; $138 printed. [WEBSITE]
Southall, Bowles, Ellison, Finneran, Gentry, Green Jr, Kastak, Ketten, James Miller, Nachtigall, Richardson, Thomas, Tyack. Marine Mammal Noise Exposure Criteria: Initial Scientific Recommendations. Aquatic Mammals, Volume 33, Number 4, 2007
At the heart of continuing controversies over ocean noise is the question of what levels of noise animals can experience before injury occurs or their behavior is significantly changed. Environmental advocates tend to push for lower limits on sound exposure, urging a precautionary approach (to limit noise until we have more understanding of their effects), while agencies generally base their standards on the most recent scientific results. For several years, an all-star team of ocean noise researchers met and attempted to distill the many different studies that had been done through the early 2007; the result of their efforts is this dense volume. Summarizing its contents is perhaps the most challenging lay summary AEI has yet undertaken—it’s been in the office for the past year, read and highlighted several times, then set aside before attempting a narrative summary. But the time must come, and is now. For those who want to participate in the continuing discussion over noise standards, reading this volume is essential; for those who simply want a sense of where research and regulatory thinking stands at this time, this summary will likely suffice

Lay Summary Abstract:
This document presents a matrix approach to help organize diverse studies that address the effects of noise on marine mammals. One section proposes specific criteria for preventing physical injury (Level A harassment), and another section collects available (and largely ambiguous) data on behavioral responses to sound. Using a number of assumptions (some being informed leaps of faith and others designed to be precautionary) and extrapolating from improved but still relatively few studies of TTS, the authors propose to increase the Level A criterion from today’s 180-190dB widely used limits to 198-215db (SEL) and 230dB (SPL peak) for whales and dolphins, and somewhat lower levels for pinnipeds. The behavioral response matrices are especially useful as a way of organizing widely divergent data, since behavioral responses tend to vary greatly between species and even among species, depending on what behaviors individuals are engaged in. However, a general pattern is evident, suggesting that many species are quite sensitive to noise at levels much lower than currently regulated (110-140dB RMS SPL). The paper concludes by stressing the need to expand study of noise impacts from marine mammals, to consider effects on fish, invertebrates, and polar bears, and to consider the likelihood that effects on any of these groups could cascade through the food chain to affect entire ecosystems.
(Ed. note: a recent paper by a number of other ocean noise researchers took exception to parts of this report’s conclusions.  To read our summary of that paper, click here)

Introduction
In many ways, the approach taken by this team follows on that sketched out in a related volume published in 2005 by the National Academies of Science, which attempted to create a functional framework for assessing the biological impact of repeated behavioral disruption. In both cases, the resulting publications suggest the use of a “matrix,” within which we can organize existing (and more importantly, future) data. These matrices allow us to set different noise criteria for animals with different hearing and vocal ranges (e.g. large whales/low frequency and dolphins/higher frequency), as well as for different types of sound sources (e.g., single pulse like an explosion, multiple pulses like airguns or pile driving, and nonpulses like shipping, sonars, or seafloor processing of oil and gas). The resulting matrices look at the ways that a given sound will be heard by and affect animals in each of the different “functional hearing groups.”

The result can seem to be a complex jumble of information, raising fears of leading to a confusing diversity of noise standards, but the researchers are clearly making a rather herculean effort to create a practical way to organize what is, at this point, a sketchy but diverse set of data (it is not uncommon to have only a handful of studies, if that, of any particular noise source on any particular hearing group).

The report separates acoustic impacts into two classes, roughly equivalent to those used by US regulatory agencies: physical injury (Level A Harassment) and behavioral change (Level B Harassment). In this first attempt at assessing the data on physical injury, the collaborators also lean toward what they consider a precautionary standard, lumping together several matrices into proposed exposure limits that match what they consider to be the most acoustically sensitive of the groups or physiologically dangerous of the sounds that are being considered. Moving into the behavioral effects section of the study, things get much less definite. While the authors were willing to make fairly universal recommendations about protecting animals from physical injury (by presuming that fundamental physiology does not differ markedly between species), there are clearly wide variations in behavioral responses between species, and between individuals of any species, along with clear differences in sensitivity depending on what activity the animals are engaged in. While the “most precautionary” standard for physical injury leads to a noise limit that is sufficiently high (200dB or more) to be applied at sea without changing our behavior, to set behavioral thresholds based on the most sensitive individuals, species, or situations would require limits so low (120-140dB) as to likely cause major disruptions in shipping, naval, and oil and gas operations at sea. (Not to imply a need for such extreme precautionary limits: in most situations, animals are far enough away or less sensitive than the extreme cases, so such limits would likely be overkill).

Physical Injury
As a criteria for injury, they use permanent shifts in hearing sensitivity (PTS). They also consider two ways of measuring sound exposure: peak Sound Pressure Level (loudest sound level experienced) and Sound Exposure Level (cumulative sound energy experienced), and suggest setting exposure criteria such that the lowest of the two thresholds becomes the de facto standard. Bottom line: for all species of whales and dolphins, they suggest an upper limit of 230dB re 1uPa (peak, flat) or 198dB re 1uPa2-s (SEL, weighted to hearing groups’ auditory curve) for single and multiple pulses, with the SEL limit raised to 215dB for non-pulses. These represent modest increases over today’s widely used de facto standards of 180-190dB for Level A Harassment; however, such sound levels only occur very close to sound sources (even loud sound sources such as airguns and military sonar), so very few animals are generally exposed to these levels of sound.

While the researchers are completely transparent with their method of coming to these figures, we should highlight a few key points made in passing in their narrative. First, PTS thresholds are based on measured TTS (temporary hearing changes) in lab settings; there is only good TTS data on two species (both mid-frequency dolphins), and this is assumed to apply to all cetacean species. Second, the difference between measured TTS onset and assumed PTS onset is based on studies of very different animals, both terrestrial mammals (researchers do not want to deafen dolphins to discover the true values). In some cases, studies on chinchillas are used, augmented with some arbitrary (and likely precautionary) changed assumptions about the ways that tissues respond to increasing sound to create tissue damage and TTS. In other cases, the assumptions are based on human studies of the ways that hearing loss compounds above the first onset of TTS and gradually causes PTS.

It should also be noted that this study covers not only dolphins and whales, but also pinnipeds (seals and sea lions), in both air and water, where again, the most precautionary studies become the basis of recommended exposure criteria (in one case, the proposed criteria are “based on the results from a single individual”). Finally, the researchers suggest using lower criteria for beaked whales, at least for non-pulsed sounds including mid-frequency tactical sonars, but refrain from making a specific recommendation.

Behavioral Effects
Because of this wide variation, the authors refrained from making any specific recommendations about noise limits to minimize behavioral disruption. Instead, they created a set of matrices that are designed to be used in coming years to both organize new data and to design studies to fill in the many gaps in the data that they reveal. These matrices make for fascinating study: there is one for each of the sets of “functional hearing groups” and sound type (e.g., one for low-frequency cetaceans and single pulses, one for low-frequency cetaceans and multiple pulses, and one for low-frequency cetaceans and nonpulses; with the same set for each of the 5 hearing groups). In each matrix, the authors compile results from existing studies of that sound type and that hearing group, arranged with increasing received sound level across the top (in 10dB steps, from 80-200dB RMS), and increasing severity of behavioral change up the left side (a 0 to 9 scale, from brief or minor change, through moderate changes in swimming or vocalizing, to obvious and extended changes in behavior). Each chart shows the cumulative total of individuals that responded at each level of severity to each level of sound exposure. The fascination comes in seeing the wide range of responses shown in virtually every chart; in some cases, a given dB level of sound exposure includes observed reactions at every level of the scale, from no response at all to very obvious disruption, and similarly, a given severity of response (say, obvious flight) can be observed at low sound levels, but not at higher levels. The challenges to setting an absolute limit are clear.

At the same time, though, the reasons that many environmentalists (and increasing numbers of scientists) are raising concerns about widespread and systemic noise impacts is also made crystal clear. The chart of mid-frequency cetacean (including most dolphin species) responses to non-pulses (primarily studies of boat noise) is striking: over half the extreme reactions observed (rated at Level 8: obvious or long-term avoidance of the area, prolonged separation of females and calves, prolonged disruption of mating behavior) occurred at sound levels of 100-130dB re 1uPa RMS, those these results are somewhat confounded by the fact that at the same time, a majority of the animals who were exposed to 100-120dB showed little or no response.

Also striking, and less ambiguous, are the low-frequency cetacean (baleen whales) responses to nonpulses (which include sonar, shipping, and enduring oil and gas noise such as drilling or seafloor processing). Here, a clear threshold appears in the data: at below 110dB, virtually all observations show little or no response, but at both 110-120dB and 120-130dB, the majority of individuals observed respond at “level 6” (avoidance, brief separation of females and calves, extended changes or cessation in vocalizing, visible startle response, brief cessation of reproductive behavior). Low-frequency cetaceans also showed quite a threshold in response to multiple pulsed sounds (e.g. airguns): 60 out of 75 responses at 110-120dB were rated Level 0 (no response), while 47 out of 72 at 120-130dB were at Level 6; this clarity is somewhat muddied by a wide range of responses to sound at 150-180dB, with clusters at both Level 0 and Level 6. Another stark, though less severe, threshold is seen with pinnipeds in water: the vast majority of responses to 120-130dB sound are at Level 0, and then jump to Level 4 at 130-140dB.

This section of the report is also enhanced by accompanying tables that detail all the studies that are compiled to create the numerical averages across each specific chart. These tables highlight whether a given study saw consistent or varied response to the sound levels in particular situations. Sometimes, the responses were all the same, say 6, while in other cases, a single study observed reactions ranging from 0 to 6 to a particular sound exposure and situation.

Further Research Recommendations
In addition to the obvious need for direct data on TTS in more species, the authors suggest that future research could be designed to fill in some of the gaps, or areas of extremely little data, in the behavioral response charts. This ongoing compilation of data will also help communicate new findings to those engaged in related research. In addition, they encourage greatly expanded study of ambient noise levels and patterns of biological sound in different ambient noise conditions, specifically mentioning the potential of remotely deployed passive acoustic monitoring systems to “become the new standard.”

The paper concludes by stressing that we have very little data on the impacts of noise on ocean life other than marine mammals, and that both more basic research and noise criteria standards “are perhaps as urgently (or more urgently) needed for some other groups” of marine life, including fish, invertebrates, sea turtles, sea otters, and polar bears. And finally, the authors address the need for noise criteria that consider ecosystemic effects: “…The effects of noise exposure on some elements of the local food webs may have a cascade effect to other elements within the web…such ecological effects should be anticipated.”