Comparative density estimates of South Africa’s forest birds: integrating distance sampling, fixed radius counts, and trait-based predictive models

Distance sampling and its statistically improved variant, T-square sampling, are important sampling methods in plant ecology. They have often been applied in the context of plant density estimations and are comparatively easy to implement, since they intuitively follow the nearest-neighbour principle and thus do not require the layout of sample plots. Previous research studying distance sampling suggested that T-square sampling may also lead to an improved estimation of spatial tree diversity indices. We simulated distance and T-square sampling in six large fully mapped forest areas for seven tree diversity indices of which some competed for the same diversity aspect, i.e. tree location (dispersion), tree species and tree size diversity. Our results demonstrated that both distance and T-square sampling are indeed robust methods for sampling spatial measures of tree diversity. The sample size required for a sampling error of 10% does not exceed 20% of the total number of trees in a sampling area. T-square sampling has the ability to adapt to different spatial patterns of tree locations and this ability is key to the way the method controls estimation bias. The sample size required for species mingling and size differentiation clearly depends on the underlying spatial tree pattern in the sampling area. With most diversity indices, sample size reductions between 0.06% and 40% could be achieved by the application of T-square sampling compared to traditional distance sampling. All other conditions being equal, we could identify the uniform angle index, the species mingling index and the size differentiation index as those diversity indices achieving lower sampling error values than their competitors. For tree density estimations the Diggle and Byth estimators performed best. Based on our results, T-square sampling can be considered a robust sampling method for spatial tree diversity indices that is easy to apply in the field.

Although wildlife biologists need reliable estimates of white‐tailed deer (Odocoileus virginianus) density to facilitate management, few studies have examined distance sampling as a density estimation technique for this species. We compared direct (i.e., spotlighting from road transects) and indirect (i.e., counting pellets on randomly placed transects) distance‐sampling techniques for estimating deer densities in east‐central Illinois, southern Illinois, and northern Michigan (USA) during 2007–2008. Density estimates (95% CI) from indirect distance sampling for northern Michigan, east‐central Illinois, and southern Illinois were 6.1–12.7 deer/km2, 11.2–15.8 deer/km2, and 15.4 deer/km2, respectively. Density estimates from direct distance sampling for northern Michigan, east‐central Illinois, and southern Illinois were 18.3–25.2 deer/km2, 14.4–18.1 deer/km2, and 19.0 deer/km2, respectively. Although density estimates did not differ between techniques in east‐central Illinois and southern Illinois, density estimates derived by direct sampling were slightly higher than those derived by indirect sampling. Estimates of density from direct distance sampling were higher than indirect distance sampling in northern Michigan. The difference in estimates among study areas may be due to landscape‐specific differences in the behavioral response of deer to roads and the representativeness of road transects. In landscapes containing more agriculture, roads tend to be systematically distributed and forest edges are independent of road placement, which may explain why both distance‐sampling methods provided similar results in Illinois. However, in more forested landscapes such as Michigan, roads tend to follow streams and may provide forest edges that are relatively scarce on the landscape. Deer in forested landscapes may be attracted to roadsides, resulting in higher density estimates not indicative of surrounding forested areas. Therefore, use of road transects for direct distance sampling may be more applicable in non‐forested landscapes. © The Wildlife Society, 2012

Distance sampling methods to estimate population densities are in wide use; however, this method may not be suitable for certain species or in certain habitats. Although validation of population estimates derived from distance sampling is necessary to determine the reliability of population estimates, validation is lacking in most studies. We measured densities of six lizard species, with particular emphasis on the endemic Dunes Sagebrush Lizard (Sceloporus arenicolus) at 14 sites throughout the range of this species in the Mescalero Sands ecosystem in southeastern New Mexico. We tested the accuracy of distance sampling by comparing results from 238 distance line transects to densities measured in 20 total removal plots. Density estimates from the distance sampling method (N= 238 transects) for S. arenicolus and all lizard species combined were 4.6 lizards/ha and 26.14 lizards/ha, respectively. Density estimates from the total removal plots (N = 20) were 30.0 lizards/ha for S. arenicolus and 85.0 lizards/ha for all lizard species combined. It is clear that, even in the relatively open shinnery oak sand dune habitat, distance sampling methods were not reliable and underestimated the densities of lizards. The disparity in density estimates from distance sampling versus total removal plots was caused by violation of the assumption of perfect detection of individuals on the transect line. Individuals that were unavailable for detection greatly influenced the density estimates. Because of the difficulty in correcting for biases, we suggest that distance sampling is not an appropriate sampling method for estimating densities of lizards.

With animal species disappearing at unprecedented rates, we need an efficient monitoring method providing reliable estimates of population density and abundance, critical for the assessment of population status and trend.We deployed 160 camera traps (CTs) systematically over 743 locations covering 17,127 km2of evergreen lowland rainforest of Salonga National Park, block South, Democratic Republic of the Congo. We evaluated the applicability of CT distance sampling (CTDS) to species different in size and behaviour. To improve precision of estimates, we evaluated two methods estimating species' availability (‘A’) for detection by CTs.We recorded 16,700 video clips, revealing 43 different animal taxa. We estimated densities of 14 species differing in physical, behavioural and ecological traits, and extracted species‐specific availability from available video footage using two methods (a)‘ACa’(Cappelle et al. [2019]Am.J.Primatol., 81, e22962) and (b)‘ARo’(Rowcliffe et al. [2014]MethodsEcol.Evol. 5, 1170). With sample sizes being large enough, we found minor differences betweenACaandARoin estimated densities. In contrast, low detectability and reactivity to the camera were main sources of bias. CTDS proved efficient for estimating density of homogenously rather than patchily distributed species.Synthesis and applications.Our application of camera trap distance sampling (CTDS) to a diverse vertebrate community demonstrates the enormous potential of this methodology for surveys of terrestrial wildlife, allowing rapid assessments of species' status and trends that can translate into effective conservation strategies. By providing the first estimates of understudied species such as the Congo peafowl, the giant ground pangolin and the cusimanses, CTDS may be used as a tool to revise these species' conservation status in the IUCN Red List of Threatened Species. Based on the constraints we encountered, we identify improvements to the current application, enhancing the general applicability of this method.

We evaluated distance sampling and dung plots as cost‐effective methods of estimating the density of mountain haresLepus timiduson moorland in the Scottish Highlands. We compared density estimates derived from these techniques to those derived from labour‐intensive capture‐recapture techniques. Distance sampling and capture‐recapture techniques produced comparable density estimates at medium and low hare densities. Density estimates derived from distance sampling were higher than those derived from capture‐recapture in high‐density hare populations. Both distance sampling and capture‐recapture techniques gave wide confidence intervals at high hare density. Histograms of perpendicular sighting distances showed that a large proportion of hares were seen on or close to the transect line and that there was a rapid fall off in detection rates with distance. This finding indicated that hare behaviour may lead to problematic survey design and may reduce the precision of density estimates. The collection of accurate distance sampling data was particularly problematic when hare density was high. In contrast, in low‐density hare populations, considerable sampling effort was required to obtain sufficient sightings of hares to reliably estimate density. Dung plots provided a relative index of abundance that successfully ranked populations of mountain hares in order of increasing density as determined by distance sampling and capture‐recapture techniques. With careful study design, distance sampling provides a good compromise between accuracy, precision and effort in estimating the density of mountain hares. The use of dung plots is a rapid alternative when only estimates of relative abundance are required.

Wildlife biologists require density estimates for white‐tailed deer (Odocoileus virginianus) to facilitate management. Aerial surveys are often used to obtain density estimates, but are subject to problems necessitating the consideration of novel techniques. During winters 2008 and 2009, we estimated deer density on 6 forest preserves near Chicago, Illinois, USA, using aerial surveys and pellet‐based distance sampling (PBDS) methods to provide a comparison of these 2 density‐estimation techniques. Density estimates from aerial surveys were obtained by dividing both the raw count of deer observed on each preserve (unadjusted aerial density) and the raw count divided by 0.75 (i.e., assuming a 75% detection rate; adjusted aerial density) by the area of the preserve. We calculated deer densities from PBDS methods using Program DISTANCE 5.0 (PBDS density) and used paired t‐tests to compare density estimates between PBDS and aerial survey techniques. Unadjusted aerial density (10–29 deer/km2) and adjusted aerial density (13–39 deer/km2) estimates did not differ (t11 = −1.99–0.44, P = 0.071–0.666) from PBDS density estimates (12–36 deer/km2). We also compared costs and found PBDS (US$85/survey) was 88% cheaper than aerial surveys (US$722/survey). Problems with bias and precision exist with both methods, and managers should give them serious consideration when choosing which method to use to estimate deer densities. Given accurate pellet decay and deposition rates and a large sample size of pellet groups, PBDS may be advantageous due to less bias in density estimates, no dependence on continuous snow cover, cheaper survey costs, and no need for elaborate equipment or for professional biologists to conduct surveys. However, future research needs to address how to reduce coefficient of variations and confidence intervals for PBDS so that differences among years can be better differentiated. © 2012 The Wildlife Society.

Estimating population sizes in the heavily traded grey parrots of West and Central Africa would provide insights into conservation status and sustainability of harvests. Ideally, density estimates would be derived from a standardized method such as distance sampling, but survey efforts are hampered by the extensive ranges, patchy distribution, variable abundance, cryptic habits and high mobility of the parrots as well as by logistical difficulties and limited resources. We carried out line transect distance sampling alongside a simpler encounter rate method at 10 sites across five West and Central African countries. Density estimates were variable across sites, from 0–0.5 individuals km−2in Côte d'Ivoire and central Democratic Republic of the Congo to c. 30 km−2in Cameroon and > 70 km−2on the island of Príncipe. Most significantly, we identified the relationship between densities estimated from distance sampling and simple encounter rates, which has important applications in monitoring grey parrots: (1) to convert records of parrot groups encountered in a day's activities by anti-poaching patrols within protected areas into indicative density estimates, (2) to confirm low density in areas where parrots are so rare that distance sampling is not feasible, and (3) to provide a link between anecdotal records and local density estimates. Encounter rates of less than one parrot group per day of walking are a reality in most forests within the species’ ranges. Densities in these areas are expected to be one individual km−2or lower, and local harvest should be disallowed on this basis.

Summary1. Distance sampling is a widely used technique for estimating the size or density of biological populations. Many distance sampling designs and most analyses use the software Distance.2. We briefly review distance sampling and its assumptions, outline the history, structure and capabilities of Distance, and provide hints on its use.3. Good survey design is a crucial prerequisite for obtaining reliable results. Distance has a survey design engine, with a built‐in geographic information system, that allows properties of different proposed designs to be examined via simulation, and survey plans to be generated.4. A first step in analysis of distance sampling data is modelling the probability of detection. Distance contains three increasingly sophisticated analysis engines for this: conventional distance sampling, which models detection probability as a function of distance from the transect and assumes all objects at zero distance are detected; multiple‐covariate distance sampling, which allows covariates in addition to distance; and mark–recapture distance sampling, which relaxes the assumption of certain detection at zero distance.5. All three engines allow estimation of density or abundance, stratified if required, with associated measures of precision calculated either analytically or via the bootstrap.6. Advanced analysis topics covered include the use of multipliers to allow analysis of indirect surveys (such as dung or nest surveys), the density surface modelling analysis engine for spatial and habitat modelling, and information about accessing the analysis engines directly from other software.7.Synthesis and applications. Distance sampling is a key method for producing abundance and density estimates in challenging field conditions. The theory underlying the methods continues to expand to cope with realistic estimation situations. In step with theoretical developments, state‐of‐the‐art software that implements these methods is described that makes the methods accessible to practising ecologists.

Quantifying the bias in density estimated from distance sampling and camera trapping of unmarked individuals

Monitoring wild cervid populations have become a priority for management. However, accurate and reliable estimates of densities are difficult to achieve since they may be affected by environmental variation, species behaviour or observational issues. Therefore, to obtain unbiased estimates of densities it is necessary to adopt sampling methods that quantify the probability to detect the target species. In this study, we compare the results of roe deer sampling based on distance detection performed by two techniques: surveys on foot in the evening and nocturnal surveys by car. Estimates of roe deer population densities were conducted in Sierra de Guadarrama (Madrid, Spain). Distance sampling was conducted along tracks in 10 pine forests in October. Observations from the surveys done on foot were better fitted with detection functions, although this technique required more days and more observers for its realization, hence increasing field effort. Nocturnal surveys by car were also a proper technique and decreased distance sampling costs, since only three people were needed for 6 days to carry them out. However, observations obtained with this technique showed an imbalance in the detection function in the first few metres. This model was limited by the small number of roe deer observed in or near the line of progression. This is a handicap because functions used by the Distance software assume that the highest probability of detecting specimens is in the line of progression, causing an imbalance in the detection function at zero distance. To compensate for this, data were lefttruncated at 20 m. Therefore, when it is necessary to estimate absolute densities of roe deer populations, nocturnal distance sampling by car seems to be the most appropriate method due to its low cost, yet the influence of the vehicle on the distribution of roe deer and, therefore, on the estimated density, must be taken into account when carrying out such studies.

Distance sampling (DS) and territory mapping (TM) are globally applied bird survey techniques. However, specifically designed studies comparing results of both methods in different habitats in the framework of a scientific experiment have rarely been conducted. To provide a more generalized guidance for the field surveyor, here we evaluated estimates of bird abundances and number of bird species in four different habitats (broad-leaved forest, coniferous forest, open woodland and farmland) in central Germany. Abundances were estimated in parallel by TM and DS in 2006 and 2008, following standard protocols. Detection probability differed significantly among habitats and species. Density estimates by DS were in total 24% lower than those estimated by standardized TM. While the number of bird species detected with both methods was approximately the same, the estimated abundances of 15 bird species showed significant differences. Increasing the number from two to four and five registrations to count a territory by using TM decreased the density on average about 28 and 42%, respectively. Using standardized TM resulted in an overestimation of abundances of species showing a high detection probability. In contrast, DS estimated very high densities for species that had a very low detection probability. In fact, a highly negative correlation was found between the density estimated by DS and the detection probability. Using standardized TM and setting a fixed number of registrations before a location qualifies for a bird territory cannot compensate for the large differences in species detectability. Instead, the number of registrations required to count a territory should be adjusted to differences in detection probabilities and seasonal activity. From our results we can recommend a mean of four registrations if eight visits were conducted to count a territory. However, the lack of any statistically-based quality assessment reduces the serious usability of TM for estimating densities for science-based management application, whereas, the clear advantage of DS is that it provides error estimates and considers differences in species detectability.

Line transects have been widely applied for the estimation of animal abundance because they are regarded as simple, economical, and relatively precise. The recent development of automated techniques for the estimation of animal density from distance sampling data allows greater potential for field biologists and wildlife managers to become involved in the analytical summary of their research. An assessment was made of the ability of program distance to produce unbiased estimates of density in spite of potential sources of error from the estimation of transect and population density. Populations were simulated to investigate the robustness of program distance to changes in the density, distribution, and detection of animals across sampling areas and transects. It is concluded that if distance sampling data is collected reliably from a random sample of possible primary sampling units (PSUs) it can be expected that estimates of density will be presented accurately and with correct estimates of variance. If the proportion of the study area surveyed by transects is large however, then the presence of large between PSU variation will cause the variance estimates from program distance to be a sizeable overestimate. Copyright © 1999 John Wiley & Sons, Ltd.

Line‐transect distance sampling is widely used to estimate population densities using distances of observed targets from transect lines to model detectability. When the target taxa are high density, the frequent measuring of distances may make the method seem impractical. We present a method that improves the efficiency of distance sampling when the target species occurs at high density. Only a proportion of targets are measured to model the detection function, and the time saved on the survey is then used to cover a longer total length of transect and accrue a larger ‘count only’ sample. This approach can improve the precision of the population density estimate when the cost of measuring the distance to a detected target is more than half the cost of walking to the next target.We find the optimal proportion of distances to measure that minimises the variance of the density estimate for a fixed survey budget. We quantify how much this optimised strategy increases the precision of the density estimate compared with conventional line‐transect distance sampling. We then use simulated distance sampling data to test our expressions, and illustrate circumstances under which the optimised approach would be beneficial using distance sampling data on high‐density plants.The simulations indicate that the optimised method delivers benefits in precision, but the magnitude of the benefit is lower than predicted from our expressions, which are based on an asymptotic approximation of the variance. We apply an adjustment to the predicted benefit equation to account for this difference, and show that, in all three plant case studies, the optimised approach could improve the precision gained from a distance sampling survey between 20% and 50%.This new approach could broaden the ecological contexts in which distance sampling is applied, to include estimation of densities of abundant taxa where plots are conventionally used. The method may have interesting applications for other survey types, including multispecies surveys or those using cues or signs that occur at high density.

Estimating population density and population dynamics is essential for understanding primate ecology and relies on robust methods. While distance sampling theory provides a robust framework for estimating animal abundance, implementing a constrained, non-systematic transect design could bias density estimates. Here, we assessed potential bias associated with line distance sampling surveys along roads based on a case study with olive baboons (Papio anubis) in Lake Manyara National Park (Tanzania). This was achieved by comparing density estimates of olive baboons derived from road transect surveys with density estimates derived from estimating the maximum number of social groups (via sleeping site counts) and multiplying this metric with the estimated average size of social groups. From 2011 to 2019, we counted olive baboons along road transects, estimated survey-specific densities in a distance sampling framework, and assessed temporal population trends. Based on the fitted half-normal detection function, the mean density was 132.5 baboons km-2 (95% CI: 110.4–159.2), however, detection models did not fit well due to heaping of sightings on and near the transects. Density estimates were associated with relatively wide confidence intervals that were mostly caused by encounter rate variance. Based on a generalized additive model, baboon densities were greater during the rainy seasons compared to the dry seasons but did not show marked annual trends. Compared to estimates derived from the alternative method (sleeping site survey), distance sampling along road transects overestimated the abundance of baboons more than threefold. Possibly, this overestimation was caused by the preferred use of roads by baboons. While being a frequently used technique (due to its relative ease of implementation compared to spatially randomized survey techniques), inferring population density of baboons (and possibly other species) based on road transects should be treated with caution. Beyond these methodological concerns and considering only the most conservative estimates, baboon densities in LMNP are among the highest across their geographic distribution range.

Comparison of density estimates derived from strip transect and distance sampling for underwater visual censuses: a case study of Chaetodontidae and Pomacanthidae