. . . Predicting and Minimizing Environmental Impacts . . .


Marine Wildlife Behavior Database


Section 7: Diving Speed

Section 1: Reference

Section 8: Dive Angle

Section 2: Identification of Subject

Section 9: Surfacing Angle

Section 3: Dive Depth

Section 10: Surface Behavior

Section 4: Dive Time

Section 11: Time At-Sea/On-Land

Section 5: Bottom Behavior

Section 12: Movement

Section 6: Vertical Rate

Literature Cited



It is important that clear instructions describe exactly how data should be entered into a database. This will ensure that data are entered in a consistent manner, making the database a fully functional element. This protocol is designed to define the Marine Wildlife Behavior Database (MWBD). It provides details on the sections of the database and the data that should be entered into each field (i.e., column). The first part includes an overview of the twelve sections of the database, including the type of data each section contains and any clarifications on data fields within a section. The second part includes a detailed description of each of the data fields within the sections, including the format (i.e., units) in which the data should be entered.

The MWBD is designed for the ideal situation in which each entry (i.e., row) corresponds to an individual animal that was studied. In reality, very few publications present results in such a discrete format. In most instances, data were pooled for the study, eliminating knowledge of the variation individual animals exhibit in a given behavioral state. There is a field in the database to indicate whether data on individual animals are available or whether the entry corresponds to data that have been summarized over a study. If data have been summarized over a specific behavioral state, age class, time of day, etc., then that should be noted and those data entered as individual rows in the database. In addition, ideally each study would include data for each field in the MWBD. However, this is rarely the case. If data are not available for a field, then “N/A” corresponding to “Not Available” is entered into the cell.

Finally, a few comments about how dives are defined, how dive depths are defined, and the use of the gamma distribution are necessary. In the MWBD, there are two types of dives, a “true dive” and a “surface dive.” A true dive has also been termed an inter-bout dive, whereas a surface dive is an intra-bout dive. True dives and surface dives are distinguished quantitatively for each row of data using a log survivorship plot in which a breakpoint between the duration of true and surface dives was determined by a significant change in the slope of the curve (CETAP, 1982; Fagen and Young, 1978). This method objectively defines the maximum duration of a surface dive and the minimum duration of a true dive, effectively categorizing every dive into these two types.

Secondly, the depth of a true dive may be defined relative to the sea surface (depth within the water column) or the sea floor (depth off the bottom). Unless an animal is diving to a particular feature, the former reference (i.e., the sea surface) should be used. However, if an animal is diving to a particular feature (a field in Section 2), the location of the animal in the water column relative to that feature must be identified (another field in Section 2). The location of the feature should be a given distance relative to either the sea surface or the sea floor. In either case, the location of the animal in the water column relative to the feature should take precedence over the actual recorded dive depths.

Thirdly, the gamma distribution is used to describe the distribution of the behavior data in this database. The gamma distribution is most appropriate for data that are not normally distributed, but have a skewed frequency distribution. Whenever possible, the raw data for a variable are fit to a gamma distribution, yielding an alpha parameter and a beta parameter. The alpha parameter is the shape parameter and the beta parameter is the rate parameter (the inverse of theta, the scale parameter). The distribution of data varies as the alpha and beta parameters vary (Figure 1).

design 2
Figure 1. Examples of distributions given different alpha and beta parameters, respectively.