Abundance and trends

Why is estimating abundance and trends important for conservation and management of river dolphins?

Understanding how many river dolphins inhabit a certain stretch of river is one of the most fundamental questions to answer for resource managers and scientists studying river dolphins. Surveys to document the number of dolphins are usually one of the first activities undertaken when exploring or documenting dolphins in new areas. Understanding the size of a river dolphin population is important because it is closely linked to extinction risk, with smaller populations being less resilient and more prone to extinction. Small population size is one of the key parameters used to identify declining or threatened species on the IUCN Red List. Baseline data on the abundance, relative abundance and distribution of river dolphins is important to be able to evaluate changes in these parameters over time and to know whether populations are increasing or decreasing in abundance. In order to detect changes in abundance over time the key thing is for surveys to be repeated using the same methodology so that results are comparable.

Definitions

• Abundance: The total or ‘absolute’ number of dolphins in a defined portion of a river. Normally determined by sampling the population (using visual direct counts, acoustics, or photographic identification of individuals) and correcting that sample for missed detections and to account for bias (correction can be conducted using a variety of methods, including distance sampling, mark recapture) to generate a point estimate with a range and measure of precision (e.g. 102 river dolphins (95% Confidence Interval: 88 – 125, CV=15%).
• Direct count: A dolphin count that is uncorrected for bias and is therefore generally an underestimate of absolute abundance.
• Relative abundance or index of abundance: An index of abundance provides insight into relative population size, even when absolute numbers are too difficult to generate. This might be a direct count, an encounter rate (number of dolphins observed per km searched). In the absence of absolute abundance estimates, these indices, if collected in a consistent manner, allow researchers to compare values from one point in time to another (e.g. between seasons or between years), and detect trends over time.
• Trends in abundance: Changes in the number of dolphins present in a section of river over time. These are statistically significant where the confidence intervals of two estimates do not overlap.
• Density: The number of dolphins per area (usually km²) of river.
• Encounter rate: The number of dolphins observed per unit of time or distance searched on a river.

Estimating abundance of river dolphins

Estimating the abundance of animals that spend a significant portion of their time underwater is challenging under any circumstances, and can be particularly challenging in rivers. In each river and for each population a different set of constraints exists, which means it is necessary to evaluate what is the most feasible way of designing and implementing a survey to estimate abundance. Different approaches to abundance estimation that have been used for river dolphins include:

• Single platform direct counts: Basic direct dolphin counts are conducted from a single vessel without correction for animals that are missed because they were underwater. This survey method is relatively simple to design and implement, and generates a minimum abundance estimate¹.
• Double platform corrected counts: These can provide more accurate abundance estimates, because they help to account for animals that may be missed by one observer team because they are underwater as the boat passes, or an observer is momentarily inattentive. These can be conducted in three ways: 1) using a single large vessel where two teams can conduct visual observations independently of each other – referred to as ‘concurrent counts from double platform surveys’²; 2) where rivers are too narrow or shallow, or large survey vessels are not available, two boats can follow each other closely and the sighting locations from each independent platform can be compared using the exact geographic location of sightings and mark-recapture models used to estimate abundance – referred to as ‘concurrent counts from tandem surveys’ ³; and 3) a research vessel can tow an acoustic array to detect the vocalisations of dolphins. The visual and acoustic detections can be compared and analysed using mark-recapture models to estimate abundance.⁴⁶
• Distance sampling: Often referred to as line transect sampling, this is based on the principle that observers systematically cover a representative portion of the target population’s habitat, and can extrapolate the counts or observations of animals made in that representative portion to the wider area believed to be inhabited by the target population⁷. This can be difficult to implement effectively in rivers, where it is difficult to design transects that meet all the assumptions of being able to cover a representative portion of the habitat in a systematic and unbiased manner ^{1, 3, 4}. However the method has been used to obtain abundance estimates in the Yangtze River ⁵ and the the Amazon ^{8, 9}.
• Photo-identification and mark-recapture: This method can only be used for river dolphin species that have dorsal fins or other areas of their bodies that are visible above the water surface for long enough to be photographed, and for species where long-lasting marks can be used to recognise individual animals over time. Individuals are catalogued and monitored over time. Some researchers use the total number of uniquely identified individuals as an abundance estimate for the population they are studying. However, applying mark-recapture models allows the proportion of ‘known’ or re-sighted individuals observed on each subsequent survey versus ‘newly identified’ animals to be used to calculate the overall population size, also accounting for animals that are in the population but have never been available for photographing¹⁰. One advantage of photo-ID is that it can also provide information on survival and recruitment rates, and the movement of individuals as well as abundance. Mark-recapture abundance estimates have been generated for Irrawaddy dolphins in the Mahakam¹¹ and Mekong Rivers ¹².
• Acoustic monitoring: In some cases, acoustic monitoring can be used to obtain measures of relative abundance in different sections of rivers, as in the Ayeyarwady ¹³.

Trends in abundance / long-term monitoring

Often the most important thing for managers to understand is whether a population is increasing or decreasing in abundance. To estimate trends in abundance it is not necessary to estimate absolute abundance, which can be very challenging to determine, involving large inputs of time, money, and technical expertise to generate an estimate with low levels of bias and high levels of precision. Instead, an index of abundance is adequate for population monitoring and is often easier and cheaper to generate. An index of abundance is an estimate that is a proportion of the population size. If surveys to generate an index of abundance are conducted repeatedly over time, and they use identical methods, this generates a dataset in which trends in abundance may be detected ¹⁴.

Collecting data on strandings and mortality can also form a critical component of long-term monitoring and assessing trends. Where rates of deaths in a population are high, there is a good chance a river dolphin population is in decline, even if there are no reliable abundance estimates or trend data ¹⁵¹⁷.

Some of the key aspects that have to be considered when selecting the best methods to estimate abundance include:

• Animal behaviour – is it possible to photograph animals reliably in the field? If yes, abundance estimation with mark-recapture may be feasible. If not, photo-identification is unlikely to be a useful tool.
• Morphology and unique markings – are animals uniquely marked in some way, particularly on the dorsal fin? If yes, photo-identification and mark-recapture might be the most appropriate tool. If not a visual or acoustic survey will need to be selected.
• Animal distribution – are the animals uniformly distributed throughout the river channel or is their distribution biased towards the bank, or the centre of the channel or some other habitat feature?  If there is bias in distribution then a stratified survey design may be needed.
• Navigability of river – is the river deep from bank to bank? Or is it shallow?  If it is deep then pre-arranged systematic transect lines running from bank to bank may be possible. If it is shallow and boats can only travel along a winding transect in the middle of the river a different survey design will be needed¹.
• Availability of boats – are there large vessels with multiple platforms that are available and that can navigate the channel without becoming stuck in the shallows? If yes, there are options for double platform surveys with independent observers and more options for towing hydrophones.  If the river is small and shallow and the available vessels are also small, double platform surveys are not possible or will be more difficult to organize, and correction factors will need to be determined using tandem surveys, acoustics, drones, or some alternative option.
• Shipping and other safety concerns – are there many ships in the channel that could make navigation along predefined transects dangerous and limit the ability to follow a predefined survey path? If yes, then the survey design will need to account for this in order to reduce danger to survey teams (e.g. surveys for the Yangtze finless porpoise and Yangtze River dolphin).
• Acoustic considerations – is the river velocity very rapid with a lot of submerged vegetation meaning anchoring acoustic devices in fixed locations is difficult? Does the channel configuration, depth and vessel traffic mean towing a hydrophone array is safe and possible? Do the available boats have internal power sources? If yes, towing a powered acoustic array might be feasible. If not, a towed archival recorder might be better.
• Seasonality – most rivers with river dolphins have huge seasonal variations in discharge throughout the year. Local experience is important to know the correct timing for a survey. For example, surveying during the lowest water season may be optimal if the dolphins are concentrated and therefore easier to count. However, if the river shrinks so much that it becomes unnavigable then surveying in mid-water might be most appropriate. In most places, conducting surveys to estimate absolute abundance during high waters is unlikely to be optimal because the habitat available to the dolphins expands dramatically into side-channels, flooded forests, or additional inundated habitat, and dolphins will disperse and be more challenging to detect.