Drone-based effective counting and ageing of hippopotamus (Hippopotamus amphibius) in the Okavango Delta in Botswana

Abstract

Accurately estimating hippopotamus (Hippopotamus amphibius) numbers is difficult due to their aggressive nature, amphibious lifestyle, and habit of diving and surfacing. Traditionally, hippos are counted using aerial surveys and land/boat surveys. We compared estimates of numbers of hippos in a lagoon in the Okavango Delta, counted from land and video taken from a DJI Phantom 4 TM drone, testing for effectiveness at three heights (40 m, 80 m, and 120 m) and four times of day (early morning, late morning, early afternoon, and late afternoon). In addition, we determined effectiveness for differentiating age classes (juvenile, subadult, and adult), based on visual assessment and measurements from drone images, at different times and heights.Estimates in the pool averaged 9.18 (± 0.25SE, range 1 -14, n = 112 counts). Drone counts at 40 m produced the highest counts of hippos, 10.6% higher than land counts and drone counts at 80 m, and 17.6% higher than drone counts at 120 m. Fewer hippos were counted in the early morning, when the hippos were active and most likely submerged, compared to all other times of day, when they tended to rest in shallow water with their bodies exposed. We were able to assign age classes to similar numbers of hippos from land counts and counts at 40 m, although land counts were better at identifying juveniles and subadults. Early morning was the least effective time to age hippos given their active behaviour, increasingly problematic with increasing height. Use of a relatively low-cost drone provided a rigorous and repeatable method for estimating numbers and ages of hippos, but not in the early morning. Introduction1 Hippopotamus or hippo (Hippopotamus amphibius) are Vulnerable on the IUCN Red List 2 of Threatened Species, numbering 115,000-130,000 [1]. Habitat loss and hunting for meat and 3 ivory are driving declines [1], but much of the population data originates from aerial surveys, 4 which can be inaccurate [2-6]. Reliable and accurate spatial and temporal data on abundances 5 and demographics of hippo populations are essential for effective conservation management 6 [1,7,8] but hippos are inherently difficult to count because individuals regularly submerge and 7 surface throughout the day and have uniform appearance. They are also among the more 8 dangerous animals in Africa [9-11], limiting effectiveness of on-land and water methods of 9 counting [12]. 10 Hippos are usually surveyed from the air [9-12], but also from boats and land [9,13,14]; 11 each method has advantages and disadvantages. Aerial surveys cover large areas [4] but with 12 limited time to scan waterbodies and count hippos, given their speed. Also, aircraft noise may 13 cause hippos to submerge [19], contributing to underestimation [20]. Aerial surveys are costly 14 and logistically difficult, resulting in long intervals between surveys [21-24]. Slow, low-flying 15 microlight aircraft or helicopters capturing images may overcome some of these challenges [25] 16 but remain costly and potentially logistically difficult, ...

Hidden Hippos: Using Photogrammetry and Multiple Imputation to Determine the Age, Sex, and Body Condition of an Animal Often Partially Submerged

“…The percentage of hippos in each age class, for both the measured and imputed data, fall within the range of values reported elsewhere [30,36,63,[84][85][86][87][88][89][90]. Further discussion of hippo demographic structure based on these data can be found in Inman [57].…”

“…Drones are generally well regarded as a valuable tool for non-invasive monitoring of wildlife [95][96][97], including hippos [21,49,57,[98][99][100][101]. This paper expands on the current use of drones generally just to count hippos, by demonstrating that it can also provide accurate estimates of hippo demographic structure and body condition, even under difficult conditions when hippos are partially submerged.…”

“…In this study, the smallest lagoon occupied by a hippo (and only one solitary hippo once) was 0.0016 km 2 , with the next smallest occupied lagoon comprising 0.0034 km 2 , providing support for the threshold value. Suitable lagoons were identified from high-resolution drone imagery orthomosaics (refer to [57] for more information). We flew over the lagoons a total of 12 times (four times each season), each flight being two weeks (±1 day) after the previous.…”

Indigenous-led responsible innovation: lessons from co-developed protocols to guide the use of drones to monitor a biocultural landscape in Kakadu National Park, Australia

“…Drones have the capacity to produce increasingly accurate and precise wildlife and habitat monitoring data, allowing for more informed and proactive environmental management (Gonzalez et al 2016;Hodgson et al 2018). Additionally, drones can produce accurate and replicable data at larger scales than on-ground assessments (Lyons et al 2019); access remote or previously inaccessible areas (Inman et al 2019); capture dynamic visual footage (Rees et al 2018); decentralise data acquisition (Koh and Wich 2012); and provide employment benefits and training for local people (Vargas-Ramirez and Paneque-Gálvez 2019). Their application for environmental management is profound, particularly over large areas where the vehicle or personnel access is difficult, dangerous or detrimental (Nowak, Dziób, and Bogawski 2019).…”

Evaluation of low‐cost consumer‐grade UAVs for conducting comprehensive high‐frequency population censuses of hippopotamus populations

“…Studies conducted in the Democratic Republic of Congo outlined census parameters like optimum flight altitude, the impact of environmental conditions, and the importance of observer bias in calculating hippo population estimates from UAV surveys, in addition to the utilization of algorithms for automatic detection and counting of hippos from infrared UAV imagery (Lhoest et al, 2015; Linchant et al, 2018). Further research in Botswana evaluated the capabilities of a low‐cost UAV for collecting census data, including population demographics, under an experimental setting (Inman, Kingsford, Chase, & Leggett, 2019). However, the abovementioned studies have all been conducted under experimental conditions in closed lake or pond systems, without taking into account changing environmental conditions and habitat types, the variability of pod size and number, and have therefore not tested the capabilities of a UAV census conducted in a real‐world scenario.…”