The map protocol follows CEE guidelines [25] and complies with the ROSES reporting standards (Additional file 2).
Searching for articles
Languages
All searches will be conducted using English search terms, but no restriction will be placed on result language. We will aim to find translators for any relevant article (or report) written in a language other than English that will pass through all screening stages (the authors themselves are able to translate results from Hebrew, Spanish, Swedish, Norwegian, Danish, Finnish, German, and French).
Search string
We conducted a scoping exercise in Web of Science in which we tested the impact of adding or modifying relevant keywords in terms of the number of papers added to the search results and their relevance. We also chose three key reviews [18, 20, 21] on the subject and checked whether the keywords successfully returned the reviews and a selection of their key references (Additional file 3). We will use the search string using Boolean operators which will be modified in accordance with the specific database we will be searching. The search string keywords represent the exposure (anthropogenic noise) and the outcome of the primary question (change in vocalization). Population and comparator will be screened from the results (i.e. studies focusing on the effects of noise on humans or that have no comparator will be excluded). All searches will be conducted using Ben-Gurion University’s institutional subscription. In cases when this subscription will not provide access to the full text of an article, we will use one of the other authors’ university subscriptions, or contact the paper’s author directly. The search string is as follows (in a Web of Science format):
Exposure: ts = (noise OR acoust* OR soundscape*) AND (pollut* OR artificial* OR anthropogen* OR urban OR disturb* OR *sonar OR “air gun*” OR vehicle* OR road* OR transportation OR *plane* OR turbine* OR ship* OR broadcast OR playback OR “white noise”)
AND
Outcome: ts = (signal* NEAR/3 modif*) OR (signal* NEAR/3 change*) OR (signal* NEAR/3 alter*) OR (signal* NEAR/3 shift*) OR (signal* NEAR/3 modulat*) OR (signal* NEAR/3 increas*) OR (signal* NEAR/3 decreas*) OR (call* NEAR/3 modif*) OR (call* NEAR/3 change*) OR (call* NEAR/3 alter*) OR (call* NEAR/3 shift*) OR (call* NEAR/3 modulat*) OR (call* NEAR/3 increas*) OR (call* NEAR/3 decreas*) OR (song* NEAR/3 modif*) OR (song* NEAR/3 change*) OR (song* NEAR/3 alter*) OR (song* NEAR/3 shift*) OR (song* NEAR/3 modulat*) OR (song* NEAR/3 increas*) OR (song* NEAR/3 decreas*) OR (vocal* NEAR/3 modif*) OR (vocal* NEAR/3 change*) OR (vocal* NEAR/3 alter*) OR (vocal* NEAR/3 shift*) OR (vocal* NEAR/3 modulat*) OR (vocal* NEAR/3 increas*) OR (vocal* NEAR/3 decreas*) OR (whistle* NEAR/3 modif*) OR (whistle* NEAR/3 change*) OR (whistle* NEAR/3 alter*) OR (whistle* NEAR/3 shift*) OR (whistle* NEAR/3 modulat*) OR (whistle* NEAR/3 increas*) OR (whistle* NEAR/3 decreas*) OR (frequency NEAR/3 modif*) OR (frequency NEAR/3 change*) OR (frequency NEAR/3 alter*) OR (frequency NEAR/3 shift*) OR (frequency NEAR/3 modulat*) OR (frequency NEAR/3 increas*) OR (frequency NEAR/3 decreas*) OR (amplitude NEAR/3 modif*) OR (amplitude NEAR/3 change*) OR (amplitude NEAR/3 alter*) OR (amplitude NEAR/3 shift*) OR (amplitude NEAR/3 modulat*) OR (amplitude NEAR/3 increas*) OR (amplitude NEAR/3 decreas*) OR “signal* plasticity” OR “call* plasticity” OR “song* plasticity” OR “vocal* plasticity” OR lombard OR “acoustic adaptation” OR “frequenc* shift*” OR phenolog* OR “nocturnal sing*” OR “nocturnal call*” OR “evoked vocal response*”)
When possible, the search will be limited to journals on the broad topics of biology (including ecology and zoology), environmental sciences, urban studies, acoustic, and multidisciplinary studies.
Abbreviated search
Where a complex search string is not accepted, searches will aim to include at least one term for exposure and one for outcomes. All modified search strings will be recorded.
Estimating comprehensiveness of search
To evaluate the performance of the search strategy, we compared the results of our search with a test list of 42 articles compiled from three recent published reviews on different specific signal modifications in response to anthropogenic noise (i.e., changes in amplitude, changes in frequency) [18, 20, 21]. The search string was then amended and tweaked to maximize the inclusion of relevant papers. Following the search, six papers were found to be not relevant for the map’s topic. Out of the remaining 36 articles, the final search yielded 27. The missing nine were articles without abstracts (an inherent bias of the search in ISI); however going over these papers revealed that in all of them, our keywords were found in the text, and thus they are expected to come up in a one of the searches in the other databases.
Publication databases
The search will be conducted using the following online databases:
-
ISI Web of Science Core Collection
-
Scopus
-
JSTOR
-
PubMed
-
Agricola
-
Environmental Sciences and pollution management
-
Sustainability science abstracts
-
Wiley Online Library
-
BioOne
-
WildPro Electronic Library
-
ProQuest
Internet searches
The following search engine will also be used for finding studies (we will include only the first 200 search results):
Specialist sources
Websites of the following organizations will be searched for any relevant literature.
All methods of retrieval for grey literature will be documented and presented in a supplementary file for the final publication.
Article screening and study eligibility criteria
Study screening process
We will conduct the screening process through the online open source platform of CADIMA (https://www.cadima.info/). CADIMA’s platform facilitates extraction of articles, removal of duplicate studies, a three-stage screening process (at the title, abstract, and full text levels), and measuring inter-screener reliability. Review articles will not be included, but their reference lists will be checked to ensure that all relevant literature they cover is included in our searches.
We will use CADIMA’s three-stage screening process to evaluate all articles found using our search string. First, we will evaluate the articles’ relevance based on their title, with those deemed outside the scope of the review excluded from further evaluation stages. To check consistency, a subset of 200 article titles or 10% (whichever is greater) will be assessed by an additional reviewer (i.e. by a total of two reviewers chosen from the authors of this protocol). Cohen’s kappa coefficient will be calculated to determine the level of agreement between the two reviewers and a value above 0.6 will indicate adequate consistency. In case of inconsistency, discrepancies will be discussed and the inclusion criteria will be clarified or modified. Second, we will repeat the process for all articles that have passed the title stage, reading their abstract to assess their relevance. Lastly, we will evaluate the relevance of all studies identified as relevant at the abstract and title stages by studying their full text. Full texts will be retrieved via open source platforms and institutional access where possible. If the full text of a given study cannot be found through those means, authors will be contacted to request a copy. Any reviewer who is the author of a study will not decide on the inclusion of that study and, in cases of uncertainty, the reviewer will err towards inclusion.
At the third stage, a subset of 20 articles or 10% (whichever is greater) will be assessed by two different reviewers. As in stage one, Cohen’s kappa coefficient will be calculated to determine the level of agreement between reviewers and a value above 0.6 will indicate adequate consistency. In case of inconsistency, discrepancies will be discussed and the inclusion criteria will be clarified or modified. In cases where there is uncertainty about the relevance of a study, the reviewers will lean towards inclusion. All articles screened at full text that are excluded will be recorded with the reason as to why they are excluded.
Eligibility criteria
For an article to be included it must fulfil the following criteria:
-
Eligible subjects Any animal or group of animals (excluding humans).
-
Eligible exposure Anthropogenically produced noise, regardless of its source (e.g., roads, urban environment, industry, etc.)
-
Eligible outcomes The study must report changes to the amplitude, frequency, temporal structure, or phenology of any type of vocalizations by the subject animals in response to the anthropogenic noise.
-
Eligible comparator The study must compare between the vocal behavior of individuals that are exposed to anthropogenic noise and individuals that are not exposed to anthropogenic noise.
Study validity assessment
All studies that pass the full-text screening stage will undergo a critical appraisal of the quality and validity of the study on a 1–5 scale (1 being the lowest quality, and 5 being the highest) for the following parameters:
-
Study design (a score of 1 representing observational study, and a score of 5 representing a Before-After-Control-Impact (BACI) study design).
-
Replication (a score of 1 representing no replication, and a score of 5 representing over 6 replications).
-
Study setting (a score of 1 representing a closed laboratory setting, a score of 5 representing a field study, and intermediate scores representing semi-natural laboratory conditions).
-
The temporal extent of the study (a score of 1 representing a study conducted over several days only, and a score of 5 representing a study taking place over an entire season or even multiple seasons).
Whenever information on the study design or results is unclear from the text, we will contact the authors directly and ask them for the information. The overall quality of the study will be determined by the average score across the parameters. To ensure repeatable estimation of quality, a subset of 10% of studies will be assessed by two reviewers (chosen, as before, from the authors of this protocol). Any reviewer who is the author of a study will not decide on the ranking of that study and any disagreement between the reviewers will be solved through discussion and, if necessary, consultation with a third reviewer.
Data coding and extraction strategy
We will extract descriptive data from all studies meeting our inclusion criteria. All data extracted will be coded and recorded in an Excel spreadsheet.
Extracted information will include:
-
Date of data extraction
-
Animal species
-
Sex composition
-
Age composition
-
Life history stage
-
Study location
-
Study date (season, year)
-
Study setting (lab/field)
-
Study design (observational, BA, CI, BACI)
-
Source of anthropogenic noise
-
Properties of anthropogenic noise (when available)– intensity (dBA), frequency (Hz), and temporal structure
-
Acoustic metric used to measure vocalizations
-
Sample size
-
Type of change(s) in vocalization (Amplitude, frequency, temporal structure)
-
Call type (e.g., territorial call, mating call, alarm call)
-
Mean level of recorded change
-
Measures of variation such as standard error, standard deviation and confidence intervals
-
Fitness outcomes of signal modification (if available)
-
Any other recorded behavioral changes in response to anthropogenic noise (if available)
If relevant data are not easily decipherable, we will seek the data in online data repositories or, when needed, contact the lead and corresponding authors via e-mail and ask for the missing data or information.
To ensure that data are extracted in a way that is repeatable and consistent, two reviewers will extract information from a subset of 10% of articles. Any inconsistencies in the extracted information will be discussed and the extraction methodology refined.
Study mapping and presentation
We will synthesize the data from all included studies into a narrative review. The narrative review will be accompanied by at least four heat maps in which we will quantify the relative number of studies on changes to the vocalization behavior of animals in response to noise pollution as a function of the taxa of the subject animals, the geographic location of the study (i.e., creating an atlas of the available evidence on the subject), the noise pollution source, and the vocalization type (i.e. whether it is a mating call, and alarm call, or some other call). These heat maps will serve us to recognize knowledge gaps and knowledge clusters. We will further synthesize each knowledge cluster to analyze the vocalization modification types that have been studied (and whether more than one type of modification has been studied in the same species or system), and the quality of the analysis (in terms of study design and statistical robustness). In addition to the narrative synthesis and the heat maps, the full Excel spreadsheet with all the extracted data coded will be made available for download.