Abstract. Obtaining correct abundance estimates to model population dynamics and promptly adjust management actions is a key goal of threatened species conservation. For elusive aquatic mammals like river dolphins, however, gathering such basic data can be extremely challenging. River dolphins in shallow waters are surveyed from canoes, with the (untested) assumption that visual counts are exhaustive estimates of animal population. These counts, however, should take into consideration that river dolphins usually emerge for very short periods and show very little of their body when surfacing. Recently, aerial survey methods have been introduced as new technology aiming to improve the reliability of wildlife censuses. Herein, we developed and tested a new survey method using a digital camera mounted on a blimp (non-rigid airship system) for aerial survey of river dolphins. We conducted our experiment by surveying the recently described Araguaian river dolphin (Inia araguaiaensis) during the low water season in the Parque Estadual do Cantão, a seasonally flooded lake system in the Amazonian forest, Brazil. Our survey, which was designed to monitor the river dolphin in 56 survey units within six main survey sectors, implied deploying two canoes navigating simultaneously: the first one with a single operator holding the blimp and recording data from its camera and the second one with three observers performing standard visual counts of dolphins. In contrast to common expectations, visual counts detected on average <75% of the dolphins recorded by the blimp's camera. Compared to the visual survey, the aerial method was able to detect more individuals, more dolphin groups, and to record more calves, which are usually difficult to detect. The differences in performance between the two methods were remarkable when surveyed sectors were larger. We showed that visual counts from canoes are less exhaustive surveys than aerial surveys, and provided a less expensive method (one operator needed to handle the blimp, compared to three visual observers) to count aquatic mammals. Our results can be used to correct current visual count estimates. With further development of blimps and drones in wildlife surveys, we foresee a promising future for the conservation and management of threatened species.

<p>In 2010, the Instituto Araguaia, a local NGO, started a program to protect and monitor giant otters, and to use them as an umbrella and flagship species for the conservation of the Cantão ecosystem. As part of the program, a census was carried out to determine the giant otter population of a group of 16 lakes in the park, and this population was monitored over a 28-month period spanning three breeding seasons, between September 2010 and December 2012.</p>

Village to flat, downstream areas of Long Beleh Village. The elevation range of the surveyed areas was - m. Despite these searches, which included the likely type locality described on herbarium specimens, we did not locate V. pentandra. We believe that extensive loss of forests, to oil palm plantations and coal mining, is the most likely reason for our failure to relocate V. pentandra. In addition, forestry companies are extracting timber from forests along the Belayan River. With these forests greatly reduced and fragmented, the areas we surveyed are some of the last remaining forested areas along this river. Although we recommend further surveys for V. pentandra, especially in upstream areas of Belayan River, in the northern Tabang Regency where forests are still in relatively good condition, we are able to update the conservation assessment of this species. Based on the findings of our survey, we reassess V. pentandra as Critically Endangered based on criteria Acd; i.e. with more than % suspected population reduction in the last three generations (A) based on a decline in area of occupancy, extent of occurrence and/or habitat quality (c), and potential level of exploitation (d). The forests of Kalimantan continue to be affected by conversion and degradation, and our updated assessment of this endemic tree species is an urgent call for the conservation of this and the other species of these forests.

We carried out monthly surveys of the giant otter population between 2010 and 2020 in a study area comprised of 1,500 hectares of igapo flooded forest with oxbow lakes in the Cantao region of central Brazil. We recorded 16-32 resident adults in the study area each year, distributed in 4-8 groups. Resident groups exhibited extensive home range overlap, with each group using several lakes and larger lakes used in rotation by up to six groups. Dens and campsites were also shared by multiple groups, but lakes were used by only one group at a time, and encounters between groups were very rare. 24 adult otters were observed to join an existing group. Some individuals changed groups multiple times. Resident adult turnover was high. Each year an average of 36% of resident adults were new immigrants, and 72% of groups left the area within two years. Resident groups had, on average, one litter every three years, and annual cub production showed high variability and a negative correlation to the number of new immigrants in the area. No pairs of giant otters reproduced successfully during the study. Groups of three otters formed through the recruitment of an adult individual by an existing pair and reproduced as successfully as larger groups. Group dynamics and territorial behavior in the Cantao flooded forest ecosystem, where optimal giant otter habitat is continuous in all directions, were found to be different from that reported in areas composed of patchy (isolated oxbow lakes) or linear (rivers) habitat. This suggest that giant otter social and territorial behavior is plastic and adapts to the spatial characteristics of the habitat.