The counting of animals requires a lot of work, but it is an important tool for aiding decision making in wildlife conservation and management.
Why count animals?
Certain parameters can be estimated from aerial census and other counts. These include:
Population estimates – the corrected numbers after counts
Population demography – age and sex structure (of rhino populations and also rare antelope)
Population trends - identify species that are declining in numbers
Population density - by combining the number of animals observed during counts with information about the size of the area surveyed, Kruger management can estimate the population density of mammals in an area.
These are used for:
GIS spatial analysis to model habitat use
Which informs game offtakes, dehorning requirements, and reintroductions.
Provide information towards drafting of species management plans e.g., elephant and rhinos
Assess habitat quality - the distribution of mammals across a landscape can provide valuable information about the quality of their habitat. For example, if a species is found only in a small area, it may indicate that the remaining habitat is fragmented, degraded or not suited.
Accounting for census bias and error
Biases
Detectability bias - animals further away are more difficult to see
Availability bias - animal under tree is hard to see(up to 30% in Kruger, higher when more vegetation cover)
Observer bias - animal missed in the open (about 4% in Kruger)
Sampling error
Animal counted and then runs in the direction of where counting will still take place will be counted twice = higher numbers than actual
Animal running away from where counting will take place may be missed = lower numbers than actual
Precision and detecting change
Precision is represented by a range of numbers with a 95% chance that the 'actual' number is within the range; the narrower the range, the more precise
X - Hard to detect change when high uncertainty because of high bias and sampling errors - we think we have 100 animals, with a wide spread of 80-120, the previous year the estimated number was 120 (100-140): Is this decrease the result of error?
√ - Easier to detect change when bias and sampling errors low, e.g., we think we have 100 animals, with a narrow spread of 95-105, the previous year the estimated number was 120 (115-125): Certainly seems more probable that decrease is real
When counting from the air & Correct for:
Crocodile and hippos - No correction, total count
Rare antelope - No correction, minimum number alive
Plains game - Detection bias and sampling error
Elephant and buffalo - Sampling error
Rhinoceros - All three biases and sampling error
Helicopter census
Count crocodile and hippos (yearly) Fly along river, speed of 50 knots, altitude 100m. Record: crocodiles in four size classes: <1m (small) to >4m (very large), all hippopotamus, water birds such as fish eagle, saddle-billed stork and Goliath heron. Includes all perennial rivers as well as relevant man-made dams
Count rhinos, elephants, buffalo and rare antelope (yearly if budget allows) Blocks are 6km x 6km. Within each block, transects are 400m apart with a 200m observation strip on each side of the helicopter. The survey team consists of a pilot, a data recorder and two observers. Helicopter circles back to allow aging and sexing of rhino where possible and better counting of elephant breeding herds. Two helicopters are used to complete the census. Blocks cover the entire southern half of the park, sample blocks are used north of the Olifants River and counts are then adjusted to the total area.
Fixed-wing census
Initiated in the early 1970s.
The census includes impala, kudu, giraffe, warthog, waterbuck, wildebeest, zebra.
Transects every 3.0 km (north of Olifants River) and every 3.75 km (south of Olifants River)
About every 2 years but depends on funding
109 east-west transects cover 27% of total area
The aircraft with 6 observers flies at a height of 250 ft above ground level and airspeed of 90- 100 knotts.
Observations are recorded in four distance categories away from the aircraft on either side: Strip A 0-100m, Strip B 100- 200m, Strip C 200-300m and Strip D 300-400m. (refer to the PDF at the bottom of the page for diagram)
Call-up method
Lions and hyenas (about every 5 years or as needed)
Calls of buffalo calf in distress played from speaker at night at up to 232 sites across Kruger
Lions within 4.5 km will respond
If there are cubs, only 0.28 probability that pride will respond. With no cubs, 0.78 chance that pride will respond
If pride shows up, 95% of pride members will be included
Hyenas within 2.1 km will respond
68% of hyenas will respond
Park is divided into 6 regions based on prey biomass
Numbers are corrected for the area in the 6 regions and summed for entire park
Mark-recapture method
Cheetah and wild dog (every 5 years): citizen science project where the public takes photos, ID using coat patterns
Leopard: camera trap photos, ID using coat patterns
Black rhino: census photos, ID using ear notches
Estimates for animal abundances in Kruger National Park. 95% confidence interval ranges are provided where available, as well as the method and the year of most recent census.
Produced by: Corli Coetsee, Cathy Greaver and Sam Ferreira (Scientific Services)
Released by: SANParks - South African National Parks 01/24
Official SANParks Census - Click Document Below to Download