FWC’s Second Update to the Statewide Florida Manatee Abundance Estimate

In 2015 the Florida Fish and Wildlife Conservation Commission (FWC) accomplished a key goal of its Manatee Management Plan (MMP) with the publication of results from the first statewide abundance estimate of the Florida manatee for the years 2011 and 2012 (Martin et al 2015). A primary conservation goal of the plan was to “implement peer-reviewed and statistically sound methods to estimate the manatee population and monitor trends.” Those findings were a significant improvement over the traditional synoptic survey method, which provides only a minimum number of manatees known to be alive using warm water and winter habitats on a particular survey day. Our newer abundance survey method differs from the older synoptic survey method in several key ways: 1) it takes place over the course of a week or more for each coast (the two coasts are flown in consecutive years); 2) it is scheduled for a time of year when all Florida manatees should be in the state but are more spread out instead of congregated at warm water and winter habitats; 3) instead of a prescribed flight path, we sample thousands of locations that are randomly selected (within guidelines) by a computer; and 4) two observers on each flight each independently count the number of manatees they see at each location. This method is preferable because: 1) it samples all likely manatee habitat in the State of Florida including the Florida Panhandle; 2) it accounts for important sources of error, such as manatees that are not seen and detected by observers during the survey; and 3) it provides a statistically sound framework for estimating statewide manatee abundance.

In December of 2015 and 2016, FWC conducted its second statewide Florida manatee abundance survey. Surveys were conducted December 1–4th, the 7th and 9th, 2015 on Florida’s west coast and December 5–8th and 12th, 2016 on the east coast. Our best estimate of statewide abundance for the 2015-2016 period is 8,810 with 95% probability (Bayesian credible interval or CI) that the real abundance is between 7,520 and 10,280 manatees. The best estimate for Florida’s west coast was 4,810 (CI: 3,820–6,010) and 4,000 (CI: 3,240–4,910) for the east coast. These ranges are important to consider as they describe the uncertainty surrounding the estimates. For example, the statewide point estimate of 8,810 by itself is of limited usefulness because it does not make known the uncertainty associated with this estimate. We do not have a good sense for how far the estimate of 8,810 is from the true population without the perspective of the range (credible interval) we present. If our statistical assumptions are correct, this credible interval tells us that we are 95% confident that the true population size lies between the lower and upper bounds. Given the amount of uncertainty in the estimates, because of issues like manatees being submerged where they cannot be seen by observers, observers missing manatees that are available to be seen, and observers not being able to sample all locations where manatees could be, our credible intervals are broad. It is important to consider these limitations and others described below when using these numbers to assess the status of the manatee population.

Designing a new method for estimating manatee abundance has been challenging because manatees occur over large, diverse habitats which makes it difficult to apply traditional statistically sound survey methods. To meet this challenge, our innovative approach was designed, tested and vetted by experts. However, there are limitations which, as explained in Martin et al. (2015) and the new report, we believe may have led to an underestimate of abundance. We addressed several of those issues in our later survey, which may explain some of the increase in our 2015–2016 abundance estimates. With every new survey we have an opportunity to improve our methods and estimates.

Some of the changes to the survey and model included (for more details see report):

1. Moving the survey window of our second survey from March to December—a time of year with more favorable weather conditions—to limit the negative effects of poor weather. This could account for some of the changes in distribution and numbers over the 5-year period since manatee movements change with season.
2. Accounting for observer’s ability to detect a manatee based on water clarity (turbidity) and sea state or sea surface condition. We included the effects of higher sea states on an observer’s ability to see manatees.
3. Developing a simplified version of the model for regions in which manatees were not detected during the survey (e.g., the Florida Panhandle), instead of leaving them out. We used data from other flights and telemetry to set minimum abundance for those regions.

Our abundance surveys are a large undertaking. Not only do they require almost four times the resources of the traditional synoptic survey, they are logistical challenging on many levels. They require many trained observers and pilots; up to seven aircraft and 14 observers per day in some cases, sometimes on short notice, and the weather needs to be favorable over an entire coastline for more than a week. Because of these challenges, we fly only one coast each year. The new estimates are presented on both a statewide and coastal scale. However, biologically, there is little movement between coasts, which supports an east-west division for management purposes. Genetic information shows some variability between the two populations, but not complete genetic isolation.

How to Interpret the Results:

There are several considerations that should be taken into account when putting these results into context. Both the survey methods and models have their limitations. The estimates of abundance for both survey periods indicate a reasonably high degree of uncertainty; therefore, it is important to consider the range of possibilities. With only two statewide point estimates, each with some degree of uncertainty, we recommend against using these values to infer population trends or population growth rates for the state or any region within it. However, we anticipate as more surveys are conducted and more information becomes available, reliability will improve and our confidence in the estimates will increase. In future efforts, we plan to further improve our survey and estimation methods. The frequency of conducting future surveys using the new method will be assessed in light of statutory requirements and information needs. The population estimate obtained from the abundance survey method can be combined with other information (e.g., survival rates, reproductive rates) to create population models (mathematical models of how and why population abundances change over time), such as the Core Biological Model and integrated population model (currently for southwest Florida only), that evaluate the status of the manatee population. By combining survival estimates, reproductive estimates, abundance estimates, carcass recovery data, and possibly other data streams, researchers are able to increase accuracy of estimates of abundance, population growth rates, survival rates, reproductive rates and importantly, the effects of unusual mortality events on the Florida manatee populations. These models will help us reduce the costs of monitoring manatees and help us assess the long-term viability and persistence of the Florida manatee population.