Abundance Surveys – Counting Whales
Updated: September 2016
In order to effectively manage human impacts on marine mammals, we need to know how many there are in the management area. We also need to know if this number is changing over time: is the population going up or down? The best way to get this information is to carry out surveys to estimate the abundance of whales or seals in specific areas. This must be done repeatedly over time to determine if the population is rising or falling.
Why do we count whales?
Surveys are carried out to generate estimates of absolute abundance, which is the number of animals in a specific area at a specific time. In some cases estimates of relative abundance (a fraction of absolute abundance that is assumed to be constant) are also useful. If surveys are repeated over time, trends in abundance (whether the number of animals is increasing or decreasing or stable) can be estimated. The main use of these estimates is for the management of whale and seal populations. They are also used for general environmental monitoring and ecosystem research.
Marine Mammal Management
“Management” does not mean telling whales what to do! It is actually the management of human impacts on whale populations, mainly direct catch (hunting) and indirect catch (by-catch, ship strikes), but also other impacts such as pollution and climate change. The main goal of management is usually to ensure that human impacts on marine mammal populations are sustainable, meaning that they do not cause the populations to decrease below a pre-defined threshold. To do this, estimates of abundance are combined with past, present and projected future catch levels in a population model, a mathematical model that mimics the response of the population to catch. This enables managers to set allowable direct and/or indirect catch levels that will not endanger the population.
How do we count whales?
The abundance of whales and seals is usually (though not always) estimated using surveys. Surveys are conducted at sea using either ships or aircraft. Generally ships are used for large offshore areas, whereas aircraft are more useful for smaller nearshore areas.
The survey area is usually divided up into smaller areas, called strata. Survey effort—the amount of time or distance the survey vessel will spend in the area—is divided between strata based on how common the target species is expected to be. Generally speaking, more effort is given to areas with high densities, because this produces a more precise estimate. The survey vessel sails or flies a series of predefined transects that cover each stratum in the survey area. The transects are designed in advance of the survey to cover the area evenly.
Visual surveys, which use human observers to count animals, are the most common type. Observers concentrate on the area ahead of and beside the vessel. When an observer sees an animal or group of animals, she immediately records the observation, noting such things as species identity, group size and the number of young in the group. The observer then takes a measurement of the perpendicular distance – the distance at 90 degrees to the direction of travel – to the animal or group. These distances are later used in a statistical analysis to estimate the width of the strip that is being surveyed by the vessel.
Aerial photographic surveys are often used for seals, which haul out on land or ice. An area of known size is photographed, and the animals are later identified and counted on the photos.
The collection of perpendicular distances in visual surveys allows the analyst to estimate the width of the strip that is surveyed for a particular species. Knowing the area that was surveyed, and the number of animals that were seen, the density of animals in the area can then be calculated as:
No. Animals / Area surveyed = No. Animals per unit area
For example, if 2 whales are seen in a surveyed area of 2 square kilometres, there is a density of 1 whale per square kilometre. This density is then applied to the entire stratum to estimate the number of animals there. Another example – if there is 1 whale per square kilometre, and the stratum has an area of 10 square kilometres, we would estimate that there are 10 whales in the stratum. Of course the calculations are much more complex than this in reality. The estimate of density has variance- a measure of the uncertainty of the estimate. Abundance estimates are therefore always expressed with a measure of uncertainty, usually a 95% confidence limit. Instead of saying that “there are 10 whales in the stratum”, we would instead say something like “There are about 10 whales in the stratum, maybe as many as 15, maybe as few as 5”.
North Atlantic Sightings Surveys – NASS
North Atlantic Sightings Surveys are a series of internationally coordinated cetacean surveys that have been conducted in 1987, 1989, 1995, 2001, 2007 (Trans-NASS (T-NASS)) and 2015. The main purpose of the surveys was and is to get quantitative information on the distribution and abundance of all cetacean species in the survey area, which encompasses much of the northern North Atlantic between Norway and North America. The NASS are perhaps the largest-scale wildlife surveys ever attempted, covering a maximum area of around 2 million square nautical miles, about the same size as Western Europe, and involving as many as 15 ships, 4 aircraft and 100 observers in a single survey. The distance covered by ships and planes while surveying has been as high as 40,000 nautical miles, a distance equivalent to about twice around the world. As many as 5,000 whales have been seen in a single survey. The long time span of the NASS make them ideal for determining abundance trends. Since 1995, the NASS have been planned and coordinated by the NAMMCO Scientific Committee.
The earliest surveys (1987 and 1989) were the largest, involving the participation of 5 countries: the Faroes, Greenland, Iceland, Norway and Spain. By 1995, Spain and Greenland were no longer participating and the survey area was consequently smaller. After 1995, Norway began surveying a portion of their area annually on a 6 year rotation. As a result, the area covered by Norway in 2001 and later surveys was smaller than in previous years. In 2007, Greenland and Canada participated in the survey, and it was coordinated with simultaneous surveys in offshore European waters (CODA survey) and off the US eastern seaboard (SNESSA survey). This made the T-NASS 2007 one of the largest ever and the first to achieve trans-Atlantic coverage.
All species encountered are recorded on these surveys. However, each participating country has one or more target species: species which are of highest priority for the survey. This may be because the species is harvested or subject to indirect take, or because of other conservation concerns. For example, the common minke whale is a target species for Norway because it is harvested there, while minke and fin whales are target species for Iceland for the same reason. The identification of target species influences survey design, in that the survey will cover only areas where the target species is expected to occur, and the stratification and effort allocation will be optimized for the target species.
Because of the huge area covered by the NASS, large ships are used to cover offshore areas. Ships can stay in the field for weeks at a time; the entire area takes as long as a month to complete. Some coastal areas, notably Iceland, West Greenland and eastern Canada, have been covered by aircraft.
Estimates of abundance have been produced for all target species and many non-target species from the NASS and T-NASS. In 2009, NAMMCO published Volume 7: North Atlantic Sightings Surveys: Counting whales in the North Atlantic, 1987-2001 of NAMMCO Scientific Publications devoted entirely to the NASS up to and including the 2001 survey. In addition, many estimates have been published elsewhere or produced as non-published working papers. These estimates have been used by the NAMMCO Scientific Committee, the Scientific Committee of the International Whaling Commission, and other agencies to provide advice on the conservation and management of many marine mammal stocks. Abundance estimates from the NASS are provided under the individual species pages, where applicable.
The NASS2015 surveys resulted in 21,277 nm of survey effort covering an area of approximately 1,200,000 nm2. Additional surveys were conducted in summer 2016 in the Jan Mayen area, Norwegian Sea and coastal Iceland. The analyses of cetacean distribution and abundance are ongoing, but many estimates are now available in the Abundance Estimates Working Group report, and further discussions on the results can be found in the report from a Workshop held in October 2017. More estimates will be available in 2018.
In 2015, Norway surveyed the Norwegian Sea and part of the Jan Mayen area by ship. These areas will also be re-surveyed in 2016 to ensure full coverage of both areas. The abundance estimates generated by these surveys will be combined with the estimates from the other Norwegian ‘mosaic’ surveys to get an estimate for the total areas that are surveyed by the Norwegians.
The survey was conducted on the vessel Fisktrans, and took take place from 22 June to 30 August.
The aerial survey was conducted on board a Partenavia, and focussed on coastal areas. The observers used a newly-constructed device to help measure accurate angles, which is important for calculating the density of whales on the survey area.
The Icelandic ship-based surveys were conducted on two ships sailing at about the same time.
The ship-based survey began in early July 2015 on the vessel Høglkettur, and continued until early August. The focus of this survey was counting pilot whales, but all marine mammals were be documented.
An important part of the pilot whale abundance estimate is getting accurate group size estimates. It can often be difficult to see how many whales are in a group from the vantage point of the ship. To help out with this problem, in 2015 the Faroes survey used a drone to fly over groups of whales that they encountered and took aerial video of the group. Researchers will then use this video to see how many whales were in the group and compare this to how many the observers said was in the group.
The survey team also towed a hydrophone to record any vocalizations.
The Greenlandic aerial surveys were conducted during August and September 2015 using a Twin Otter aircraft.
Nations: 5 (Canada, Faroe Islands, Greenland, Iceland, Norway) and coordination with European CODA and American SNESSA surveys
Area (nm2): 1,298,088 nm2
Effort (nm): 12,440 nm
Nations: 3 (Faroe Islands, Iceland, Norway)
Area: 742,401 nm2
Effort: 10,715 nm
Nations: 3 (Faroe Islands, Iceland, Norway)
Area: 1,609,332 nm2
Effort: 21,309 nm
Nations: 4 (Faroe Islands, Iceland, Norway, Spain)
Area: 1,987,570 nm2
Effort: 28,965 nm
Nations: 5 (Faroe Islands, Greenland, Iceland, Norway, Spain)
Area: 1,311,071 nm2
Effort: 27,633 nm
Trends in Abundance
Populations of animals my rise or fall in abundance for many reasons. For example, if many animals are taken by hunters, and they are not replaced quickly enough by natural reproduction, a population will fall. A population might rise if it had previously been reduced to a low level but hunting had since ceased. The detection of trends in abundance is of great importance to management, because we need to know if the management measures taken are having the desired effect on the population.
Many species of whales live a long time and don’t have calves very frequently. As a result, whale populations tend to change slowly, neither rising nor falling very quickly. This makes the detection of changes in whale populations difficult, because only small changes can be expected from year to year. In addition, abundance estimates are approximations, with variance expressed as a 95% confidence interval. A change in abundance between two or more surveys can only be detected if the difference in abundance is falls outside of the overlap between the confidence intervals.
The detection of trends in abundance requires two or more surveys, usually separated by several years so that some change can be expected. The surveys must cover the same area which ideally must include most of the range of the population being surveyed, because animal distribution can change from year to year. Finally, the surveys should be conducted at the same time of year, because many whales migrate seasonally.
Five NASS have been carried out over a 20 year period (1987 to 2007). These surveys therefore provide a good opportunity for detecting trends in abundance over that period. However, the survey area has changed over that period, and one survey (1989) was conducted later in the year than the others. Therefore not all the survey data can be used to look for trends.
Detecting trends: Pilot whales as an example
Pilot whales are taken every year in a drive hunt in the Faroe Islands. Therefore trends in abundance are of great interest to management: we need to know if the population is rising or falling over time. Although five NASS have been carried out between 1987 and 2007, only a relatively small proportion of each survey area, the common area, has been covered every time. To look for trends, the abundance of pilot whales was estimated in the E and W common areas for each survey. While there appears to have been some decline in abundance estimates over the period, the high degree of uncertainty within survey estimates means that this apparent decline is not statistically significant. In contrast, highly significant trends have been detected for fin and humpback whales.
Frequently asked questions
1. Question: Do surveys count all the whales in the survey area?
Answer: No they don’t. The survey design usually allows only a small proportion of the total survey area to be viewed by observers. Therefore they are likely to see only a similar proportion of the total number of whales in the area. Total abundance is estimated using the density of animals on the transects.
2. Question: It is easier to see whales that are close to the observer than those that are farther away. How do you account for that in estimating abundance?
Answer: Most modern surveys use a method called “Line Transect”. In such a survey, the observers measure or estimate the perpendicular distance to every group of whales they see. As expected, in most cases more whales are seen nearby than farther away, simply because it is more difficult to see things that are far away. But some species of whales are large and easy to see, whereas others are small and hard to spot. For example, a fin whale is much easier to see, and can be seen from farther away, than a minke whale. As a result, the distribution of perpendicular distances is different for each species. Analysts use this distribution to estimate the “effective strip width” for each species. The effective strip width corresponds to the width of the strip that is actually surveyed by the observers on the ship or plane. As you would expect, effective strip width tends to be narrow for species like minke whales that are small and hard to see, and wider for species like fin whales which are large and easier to see from a distance.
Effective Strip Width
3. Question: Observers are not perfect, and even the best observer is going to miss seeing some whales. How do you account for whales that are missed by observers?
Answer: The line transect method assumes that all animals close to the transect are seen by observers, but human observers do miss whales, even when they are nearby. This problem is usually addressed by having two separate teams of observers operating from the same ship or plane. The two teams usually operate independently, meaning they don’t communicate with each other about sightings. Often the teams see the same whales, but sometimes one team records a sighting that is missed by the other. The proportion of missed sightings for each team, along with the distance data gathered by the teams, enables the analyst to estimate the proportion of visible whales that were missed during the survey. This is used to correct the estimate of abundance.
4. Question: Some whales stay underwater for long periods of time, and would not be visible to observers when the ship or aircraft passes by. How do you account for these “missing” whales?
Answer: This is indeed a very difficult problem for whale surveys. Whales that are diving when the ship or plane passes by are invisible to observers and cannot be counted. This problem is dealt with in two ways:
For species that don’t stay under water for long periods of time, like minke whales and harbour porpoises, the issue can be addressed by having two sets of observers with different viewing fields. On a ship survey, there might be one set of observers using powerful binoculars to search far ahead of the ship, and another set of observers using only their eyes to search relatively nearby. On an aerial survey, the plane might circle around after a whale is sighted and survey the same part of the transect again: a so-called “racetrack” survey. The sightings of these two sets of observers (or one set of observers two times in the case of racetrack) are compared and sightings of the same group of whales (duplicate sightings) identified. The analyst uses these data to estimate the proportion of whales that would have been missed if there had been only one set of observers searching close to the ship, and this is used to correct the estimate of abundance.
An alternate approach uses information on the diving habits of the target species, such as the proportion of time the whale spends at or near the surface, and the frequency of dives. These data are usually obtained in external studies, often involving the application of satellite or radio tags which record and transmit information on the activities of the animal. This information is combined with data from the survey to estimate the probability that an animal would have been visible (or not) during the passage of the ship or plane, and this is used to correct the estimate of abundance.
5. Question: Don’t ships and airplanes scare the whales away?
Answer: This is more of an issue for ships than for aircraft, because aircraft move so quickly that a whale does not have time to flee even if it does hear the plane. Whales can have differing responses to ships. For example, some species of dolphins are attracted to ships and approach them, even surfing on the ship’s bow wave. Other species do tend to leave the area when they sense the approach of a ship. Still others, such as fin whales, do not seem to react to ships much at all. Either type of responsive movement – aversive (fleeing) or attractive (approaching) – can be problematic because it can lead to underestimation of density with aversive movement or overestimation with attractive movement. The problem is addressed in the same way as that of diving whales, by having two sets of observers with different viewing fields. In this case, the distribution of perpendicular distances for each team is compared to determine if responsive movement has occurred. For example, if attractive movement is occurring, the team using naked eyes only will tend to see a greater proportion of animals close to the transect, because the whales are approaching the ship after being sighted by the team using binoculars. In some cases, the team using binoculars actually tracks individual pods of whales to determine how they respond to the ship. The analyst uses these data to adjust the estimate of effective strip width to account for the responsive movement observed.
6. Question: Whales move around, and could move between transect lines during the survey. How do you know you are not counting the same whales again and again?
Answer: It could happen that the same group of whales is counted twice during a survey. But it is assumed that the whales move in such a way that they are equally likely to be not counted at all. For example a whale might move near to a transect immediately after the ship or plane has passed. Therefore the movement of whales, if it is random, should have no overall effect on the number of whales seen. Whale movement can be an issue if a survey is conducted during a time when the whales are migrating. In such a case, the survey proceeds in a direction opposite to the movement of the whales, so that the survey does not count the same animals multiple times.