Monitoring distributions using call surveys: estimating site occupancy, detection probabilities and inferring absence

Monitoring programs serve to track changes in the distribution and abundance of species. A major problem with most monitoring programs is that species detection is imperfect and some populations are inevitably missed. Therefore, in most monitoring programs the true distribution of a species will be underestimated. Here, we report a field test of the reliability and performance of a commonly used method to monitor the distribution of amphibians (anuran call surveys). We surveyed the distribution of four anuran species in western Switzerland, and estimated detection probabilities to account for imperfect species detection and used these estimates to adjust our estimate of site occupancy (i.e., distribution). Next, we assessed how detection probabilities were affected by weather and how site occupancy was affected by site specific covariates. For one species (Hyla arborea), call surveys proved efficient in determining the regional distribution with only few site visits because detection probabilities were relatively high. The call surveys apparently missed many populations of another common species (Bufo calamita) because detection probabilities were lower. Two other species (Bombina variegata and Alytes obstetricans) were uncommon and strong inference from the analysis is not possible. Thus, multispecies surveys may be inefficient for rare species. Estimates of detection probabilities were used to calculate how many site visits are necessary to infer the absence of a species with some predetermined statistical certainty. The implications of “false absences” are important in ecology as they are known to bias usual habitat suitability models and overestimate extinction/colonization events in metapopulations. Large-scale monitoring programs would benefit from the application of an estimation-based approach to monitoring the distribution of species.     

Estimating occupancy of a data deficient mammalian species living in tropical rainforests: Sun bears in the Kerinci Seblat region, Sumatra

Tropical mammals represent some of the most threatened species, but also the least known because they tend to be difficult to study. To objectively evaluate the conservation status of these species, standardized methods are urgently required. The sun bear Helarctos malayanus is a case in point: it is cryptic, difficult to detect and consequently classified on the IUCN Red List as Data Deficient, and the highest priority for bear conservation research. In this study, we apply a detection/non-detection sampling technique using camera trap data with environmental covariates to estimate sun bear occupancy from three tropical forest study areas with different levels of degradation and protection status in Sumatra. Sun bear detections, and encounter rates, were highest in one of the primary forest study areas, but sun bear occupancy was highest in the degraded forest study area. Whilst, sun bears were recorded at a greater proportion of camera placements in degraded forest, these records were often on only one occasion at each placement, which greatly increased the final occupancy estimate. Primary forests with their large fruiting trees undoubtedly represent good sun bear habitat, but our results indicate that degraded forest can also represent important habitat. These forests should therefore not be considered as having limited conservation value and assigned to other uses, such as oil palm production, as has previously happened in Sumatra. Estimating occupancy between years will yield information on the population trends of sun bears and other tropical mammals, which can be used to provide more reliable conservation assessments.     

Modeling co-occurrence of northern spotted and barred owls: Accounting for detection probability differences

Barred owls (Strix varia) have recently expanded their range and now encompass the entire range of the northern spotted owl (Strix occidentalis caurina). This expansion has led to two important issues of concern for management of northern spotted owls: (1) possible competitive interactions between the two species that could contribute to population declines of northern spotted owls, and (2) possible changes in vocalization behavior and detection probabilities of northern spotted owls induced by presence of barred owls. We used a two-species occupancy model to investigate whether there was evidence of competitive exclusion between the two species at study locations in Oregon, USA. We simultaneously estimated detection probabilities for both species and determined if the presence of one species influenced the detection of the other species. Model selection results and associated parameter estimates provided no evidence that barred owls excluded spotted owls from territories. We found strong evidence that detection probabilities differed for the two species, with higher probabilities for northern spotted owls that are the object of current surveys. Non-detection of barred owls is very common in surveys for northern spotted owls, and detection of both owl species was negatively influenced by the presence of the congeneric species. Our results suggest that analyses directed at hypotheses of barred owl effects on demographic or occupancy vital rates of northern spotted owls need to deal adequately with imperfect and variable detection probabilities for both species.     

Small mammals, habitat patches and PVA models: a field test of model predictive ability

The results are described of comparisons between actual values for patch occupancy for two species of Australian small mammals (Bush Rat Rattus fuscipes and Agile Antechinus Antechinus agilis) determined from field sampling and predictions of patch occupancy made using VORTEX, a generic simulation model for Population Viability Analysis (PVA). The work focussed on a fragmented forest in south-eastern Australia comprised of a network of 39 patches of native eucalypt forest surrounded by extensive stands of exotic softwood Radiata Pine (Pinus radiata) plantation. A range of modelling scenarios were completed in which four broad factors were varied: (1) inter-patch variation in habitat quality; (2) the pattern of inter-patch dispersal; (3) the rate of inter-patch dispersal; and (4) the population sink effects of the Radiata Pine matrix that surrounded the eucalypt patches. Model predictions were made for the total number of animals, the distribution of animal density among patches, the total number of occupied patches, and the probability of patch occupancy. Predictions were then compared with observed values for these same measures based on extensive field surveys of small mammals in the patch system. For most models for the Bush Rat, the predicted relative density of animals per patch correlated well with the values estimated from field surveys. Predictions of patch occupancy were not significantly different from the actual value for the number of occupied patches in half the models tested. The better models explained 10-16% of the log-likelihood of the probability of patch occupancy. While some of the models gave reasonable forecasts of the number of occupied patches, even in these cases, they had only moderate ability to predict which patches were occupied. Field surveys revealed there was no relationship between patch area and population density for the Agile Antechinus—an outcome correctly predicted by only a few models. Five of the 18 scenarios completed for the Agile Antechinus gave predicted numbers of occupied patches not significantly different from the observed number. In each of these five cases, large standard deviations around the mean predicted value meant uncertainty generated by the simulation model limited the predictive power of the PVA. Some of the models gave reasonable predictions for the number of occupied patches, but those models were unable to predict which ones were actually occupied. The results of our study suggest that key processes influencing which specific patches would be occupied were not modelled appropriately. High levels of variability and fecundity drive the population dynamics of the Bush Rat and Agile Antechinus, making the patch system unpredictable and difficult to model accurately. Despite the fact that both the Bush Rat and the Agile Antechinus are two of the most studied mammals in Australia, there are attributes of their biology that are presently poorly understood (which were not included in the VORTEX model), but which could strongly influence patch occupancy. For example, local landscape features may be important determinants of inter-patch movement and habitat utilisation in the patch system. Further empirical studies are needed to explore this aspect of small mammal biology.     

A rapid approach to modeling species-habitat relationships

A basic element in the success of managing species of conservation concern is knowledge of the species’ habitat occupancy. Often, predictive species-habitat models are developed from GIS data sources that were intended for purposes other than predicting species habitat occupancy and are of inappropriate scale. In addition, the techniques used to quantify predictor variables from such data sources are often time consuming and cannot be repeated efficiently to reflect changing conditions. We used digital orthophotos and a grid cell classification scheme to develop an efficient technique to quantify predictor variables to model Florida scrub-jay habitat occupancy. We combined our classification scheme with a priori hypothesis development using expert knowledge and a previously published habitat suitability index and used an objective model selection procedure to choose candidate models. We classified a large area (43,000 ha) in a fraction of the time that would be required to map vegetation classes and were able to test models at varying scales using a grid-cell windowing process. Interpretation of the selected models confirmed existing knowledge of factors important to Florida scrub-jay habitat occupancy. The potential uses and advantages of using a grid cell classification scheme in conjunction with expert knowledge and an objective model selection procedure are discussed.     

Integrating occupancy modeling and interview data for corridor identification: A case study for jaguars in Nicaragua

Corridors are critical elements in the long-term conservation of wide-ranging species like the jaguar (Panthera onca). Jaguar corridors across the range of the species were initially identified using a GIS-based least-cost corridor model. However, due to inherent errors in remotely sensed data and model uncertainties, these corridors warrant field verification before conservation efforts can begin. We developed a novel corridor assessment protocol based on interview data and site occupancy modeling. We divided our pilot study area, in southeastern Nicaragua, into 71, 6 × 6 km sampling units and conducted 160 structured interviews with local residents. Interviews were designed to collect data on jaguar and seven prey species so that detection/non-detection matrices could be constructed for each sampling unit. Jaguars were reportedly detected in 57% of the sampling units and had a detection probability of 28%. With the exception of white-lipped peccary, prey species were reportedly detected in 82–100% of the sampling units. Though the use of interview data may violate some assumptions of the occupancy modeling approach for determining ‘proportion of area occupied’, we countered these shortcomings through study design and interpreting the occupancy parameter, psi, as ‘probability of habitat used’. Probability of habitat use was modeled for each target species using single state or multistate models. A combination of the estimated probabilities of habitat use for jaguar and prey was selected to identify the final jaguar corridor. This protocol provides an efficient field methodology for identifying corridors for easily-identifiable species, across large study areas comprised of unprotected, private lands.     

Grid-based monitoring methods for detecting population declines: Sensitivity to spatial scale and consequences of scale correction

A decrease in the area of occupancy of a species is often criterion by which a species is classified as threatened (i.e. IUCN Red List). However, it is unclear how to accurately measure change in area of occupancy. Area of occupancy is a measure of distribution and the spatial scale at which the distribution is measured will affect the ability to detect a decline. To overcome errors introduced by measuring distribution at different spatial scales, scale correction methods are often applied. It is known that scale correction reliably estimates area of occupancy; however, its suitability to estimate trends in area of occupancy has not been assessed. We investigate the effect of spatial scale and implementation of scale correction when estimating two different forms of decline in area of occupancy: spatially correlated and spatially uncorrelated declines. We explore these issues using simulations of three declining species and a grid-based monitoring method designed to detect the declines. Our results suggest that current grid-based methods are inadequate for detecting uncorrelated local extinctions, even if the total decline is substantial (e.g. 65% of the original range). We demonstrate that scale correction will lead to misleading conclusions in some situations. We suggest an alternative survey method to accurately estimate changes in area of occupancy over time. We provide guidelines for designing grid surveys to measure changes in area of occupancy to assess a species threat status using the IUCN Red List criterion.     

Viability and patch occupancy of a swamp rabbit metapopulation at the northern edge of its distribution

Swamp rabbits (Sylvilagus aquaticus) are state-endangered in Indiana, USA, and population decline has been attributed to habitat loss. We conducted pellet surveys as part of a long-term survey effort that has been conducted at approximate 10-year intervals over the last 40 years. We modeled patch occupancy and conducted a spatially-explicit population viability analysis (PVA). Although occupancy of individual patches varied over time, occupancy rate has been constant for the last 30 years, and Indiana swamp rabbits exist as a metapopulation that appears to be stable. Metapopulation dynamics were best characterized as being stationary, but area was an important factor in extinction rates; occupied patches (142 ± 37 ha) were significantly larger (P = 0.01) than unoccupied patches (79 ± 20 ha). We did not find strong support for models with colonization rates as a function of distance to neighboring patches, nor was distance to contiguous patches of habitat significantly different (P = 0.12) for occupied and unoccupied sites. Population viability analysis corroborated our findings based on occupancy modeling, and evaluation of the PVA model using occupancy data for the period 1985–2006 resulted in predictions that nearly matched our field observations (33% observed patch occupancy vs. 25% predicted patch occupancy). Population viability was most sensitive to reductions in survival and fecundity rates, but was otherwise robust to changes in parameters such as initial abundance and carrying capacity. Our findings provide novel insights into a poorly studied member of Sylvilagus and into species metapopulation dynamics at the edge of the range.     

Patterns of spatiotemporal change in large mammal distribution and abundance in the southern Western Ghats, India

Large mammals face high risks of anthropogenic extinction owing to their larger body mass and associated life history traits. Recent worldwide mammal declines have highlighted the conservation importance of effective assessments of trends in distribution and abundance of species. Yet reliable data depicting the nature and extent of changes in population parameters is sparse, primarily due to logistical problems in covering large areas and difficulties in obtaining reliable information at large spatial scales, particularly over time. We used key informant surveys to generate detection histories for 18 species of large mammals (body mass > 2 kg) at two points in time (present and 30 years ago) in the Southern subregion of the Western Ghats global biodiversity hotspot. Multiple-season occupancy models were used to assess temporal trends in occupancy, detectability and vital rates of extinction and colonization for each species. Our results show significant declines in distribution for large carnivores, the Asian elephant and endemic ungulates and primates. There is a significant decline in detectability for 16 species, which suggests a decline in their abundance. These patterns of change in distribution and abundance repeat in our assessments of spatial variation in occupancy dynamics between the three contiguous forest complexes and two human-dominated landscapes into which the southern Western Ghats has been fragmented. Extinction rates are highest in the human-dominated landscapes. Declines in abundance for several species suggest the presence of extinction debts, which may soon be repaid with imminent range contractions and subsequent species extinctions unless immediate remedial conservation measures are taken. Detection/non-detection surveys of key informants used in an occupancy modeling framework provide potential for rapid conservation status assessments of multiple species across large spatial scales over time.     

Patch-occupancy models indicate human activity as major determinant of forest elephant Loxodonta cyclotis seasonal distribution in an industrial corridor in Gabon

The importance of human activity and ecological features in influencing African forest elephant ranging behaviour was investigated in the Rabi-Ndogo corridor of the Gamba Complex of Protected Areas in southwest Gabon. Locations in a wide geographical area with a range of environmental variables were selected for patch-occupancy surveys using elephant dung to assess seasonal presence and absence of elephants. Patch-occupancy procedures allowed for covariate modelling evaluating hypotheses for both occupancy in relation to human activity and ecological features, and detection probability in relation to vegetation density. The best fitting models for old and fresh dung data sets indicate that (1) detection probability for elephant dung is negatively related to the relative density of the vegetation, and (2) human activity, such as presence and infrastructure, are more closely associated with elephant distribution patterns than are ecological features, such as the presence of wetlands and preferred fresh fruit. Our findings emphasize the sensitivity of elephants to human disturbance, in this case infrastructure development associated with gas and oil production. Patch-occupancy methodology offers a viable alternative to current transect protocols for monitoring programs with multiple covariates.     

Hierarchical models of animal abundance and occurrence

Much of animal ecology is devoted to studies of abundance and occurrence of species, based on surveys of spatially referenced sample units. These surveys frequently yield sparse counts that are contaminated by imperfect detection, making direct inference about abundance or occurrence based on observational data infeasible. This article describes a flexible hierarchical modeling framework for estimation and inference about animal abundance and occurrence from survey data that are subject to imperfect detection. Within this framework, we specify models of abundance and detectability of animals at the level of the local populations defined by the sample units. Information at the level of the local population is aggregated by specifying models that describe variation in abundance and detection among sites. We describe likelihood-based and Bayesian methods for estimation and inference under the resulting hierarchical model. We provide two examples of the application of hierarchical models to animal survey data, the first based on removal counts of stream fish and the second based on avian quadrat counts. For both examples, we provide a Bayesian analysis of the models using the software WinBUGS.     

Predictive modelling and ground validation of the spatial distribution of the New Zealand long-tailed bat (Chalinolobus tuberculatus)

The use of predictive models is continually increasing, but few models are subsequently field-checked and evaluated. This study evaluates the statistical strength and usefulness for conservation purposes of a predictive habitat use model developed for Chalinolobus tuberculatus, a threatened microchiropteran bat species found in the temperate rainforests of New Zealand. The relationship between various environmental variables and the presence/absence of the species was investigated using generalised linear modelling. The model developed was coupled with GIS data to develop maps of predicted occurrence within the West Coast region of New Zealand’s South Island. It was found that distance to forest boundary, slope, presence of Nothofagus, general land cover, variability in mean annual solar radiation, and mean ambient winter minimum temperature were significantly associated with the occurrence of the species. Evaluation of the statistical strength of the distribution model with independent data of species’ occurrence collected at 152 sites found that the C. tuberculatus model showed a moderate ability to predict both species presence and absence (τ(b) coefficient = 0.37). The field detection rate (0.45) using this model was significantly higher than that of historical surveys (0.12). The value of the species habitat model and the need to evaluate its utility in the development of conservation strategies is discussed.     

Optimising biodiversity assessments by volunteers: The application of occupancy modelling to large-scale amphibian surveys

Mobilising volunteers to carry out biodiversity assessments can help to identify priorities for conservation across broad geographical scales. However, even when volunteers carry out simple presence-absence surveys, there can be significant issues over false absences and subsequent data interpretation. Simple but scientifically robust protocols are therefore required for these programmes. Here we evaluate amphibian survey protocols for the National Amphibian and Reptile Recording Scheme (NARRS) in Britain, which aims to assess the status of five widespread amphibian species. Surveys were undertaken by trained volunteers and researchers in two contrasting landscapes over 2 years, and occupancy modelling was used to determine covariates of detection, and to optimise the number of surveys and number of methods required. Although surveys need to take into account seasonal and annual changes in the detectability of different species, there were also landscape effects. Frogs and toads were generally harder to detect in ponds in Kent than in Wales, while the converse was true of newts. Adding bottle-trapping to the suite of methods increased the detection of smooth and palmate newts in both areas, and of great crested newts in Wales. Overall, reliable assessment of the presence or absence of all five species at a site required four separate surveys, each using four different methods (visual encounter surveys during both day and night, dip netting and bottle-trapping). Our approach may prove useful for finding the best compromises between rigor and simplicity when volunteers are used in large-scale surveys.     

A gamma-shaped detection function for line-transect surveys with mark-recapture and covariate data

We have developed a procedure for estimating animal population size from aerial survey data collected simultaneously by two observers on the same sighting platform. We used a line transect sample design where transects follow elevation contours in mountainous terrain. Because our 10 data sets from aerial line transect surveys, conducted over a terrestrial environment, consistently show unimodal detection shapes, we chose a gamma-shaped detection function that is unimodal and admits covariates. We fit models separately to data from each observer, and then used all of the data to estimate the probabilities at the apex of the detection curves. We used a Horvitz-Thompson estimator to estimate the population size. We illustrate our procedure on a recently collected brown bear data set.     

Evaluating persistence and its predictors in a West African carnivore community

Carnivore extinctions frequently have cascading impacts through an ecosystem, so effective management of ecological communities requires an understanding of carnivore vulnerability. This has been hindered by the elusive nature of many carnivores, as well as a disproportionate focus on large-bodied species and particular geographic regions. We use multiple survey methods and a hierarchical multi-species occupancy model accounting for imperfect detection to assess extinction risk across the entire carnivore community in Ghana’s Mole National Park, a poorly studied West African savanna ecosystem. Only 9 of 16 historically occurring carnivore species were detected in a camera-trap survey covering 253 stations deployed for 5469 trap days between October 2006 and January 2009, and our occupancy model indicated low overall likelihoods of false absence despite low per-survey probabilities of detection. Concurrent sign, call-in, and village surveys, as well as long-term law enforcement patrol records, provided more equivocal evidence of carnivore occurrence but supported the conclusion that many carnivores have declined and are likely functionally or fully extirpated from the park, including the top predator, lion (Panthera leo). Contrary to expectation, variation in carnivore persistence was not explained by ecological or life-history traits such as body size, home range size or fecundity, thus raising questions about the predictability of carnivore community disassembly. Our results imply an urgent need for new initiatives to better protect and restore West Africa’s embattled carnivore populations, and they highlight a broader need for more empirical study of the response of entire carnivore communities to anthropogenic impact.     

Brown hyaenas on roads: Estimating carnivore occupancy and abundance using spatially auto-correlated sign survey replicates

Carnivore survey protocols that properly address spatial sampling and detectability issues are seldom feasible at a landscape-scale. This limits knowledge of large-scale patterns in distribution, abundance and their underlying determinants, hindering conservation of globally threatened carnivore populations. Occupancy analysis of data from logistically efficient sign surveys along consecutive road segments (spatially auto-correlated replicates) offers a potential solution. We adapted and applied this newly-developed method over 62,979 km² of human-modified land in South Africa. Our aims were to (1) generate unbiased estimates of brown hyaena occupancy and abundance (2) investigate two suspected determinants of occupancy using a combination of biological and socio-economic sampling techniques, and (3) use simulations to evaluate the effort required for abundance and occupancy estimates with acceptable bias, precision and power. Brown hyaena occupancy was estimated at 0.748 (±SE 0.1), and estimated overall density in agricultural land (0.15/100 km², ±SE 0.08) was an order of magnitude lower than in protected areas. Positive attitudes to carnivores and presence of wildlife farms exerted strong positive effects on occupancy, so changes in these factors may well exert monotonic impacts on local metapopulation status. Producing reliable occupancy and abundance estimates would require ?6 replicates and ?12 replicates per site respectively. Detecting 50% and 30% declines in brown hyaena occupancy with adequate power would require five annual surveys at ?65 sites and ?125 sites respectively. Our results suggest that protocols based on spatially auto-correlated sign survey replicates could be used to monitor carnivore populations at large, and possibly even country-wide spatial scales.     

Relating bat species presence to simple habitat measures in a central Appalachian forest

We actively sampled the bat community at 63 sites using detection and non-detection metrics on the Fernow Experimental Forest (FEF) in the central Appalachians of West Virginia using Anabat acoustical equipment May-June 2001-2003 to relate species presence to simple habitat measures such as proximity to riparian areas, forest canopy cover, forest canopy gap width, and forest canopy height. We acoustically detected eight species on the FEF, including the endangered Myotis sodalis. The presence of Lasiurus cinereus, M. lucifugus, M. sodalis, and Pipistrellus subflavus was associated more with riparian areas than upland areas. Both univariate comparisons and multiple logistic regression modeling showed that the probability that clutter-adapted foraging species such as M. septentrionalis and M. sodalis would be detected was greater as forest canopy cover increased or forest canopy gap size decreased, whereas the opposite was true for open-adapted foraging species such as Eptesicus fuscus and L. cinereus. The overall proportion of unidentifiable bat echolocation sequences to those identified to species was related to upland sites with increasing forest canopy cover indicating some sampling bias between cluttered and uncluttered habitats. However, given sufficient sample points, bat community surveys using acoustical detection show the ability to quickly develop generalized habitat associations for rugged areas such as the central Appalachians where traditional mist-net survey efforts often are logistically difficult and are lacking in scope. Moreover, these acoustical surveys also could lend themselves to species-specific predictive mapping of foraging habitat as well as allowing researchers to formulate testable hypotheses about detailed bat habitat relationships to be definitively tested with radio-telemetry techniques.     

Metapopulation dynamics and threatened species management: Why does the broad-toothed rat (Mastacomys fuscus) persist?

Many species that exist patchily across the landscape are declining due to incremental losses of their constituent sub-populations and increasing isolation of those that remain. For threatened species with such patchy metapopulation structure, it is particularly important to identify key habitat patches and understand what factors govern their occupancy so that their management can be targeted and effective. In this paper, we describe the spatial and temporal distribution of an endangered population of broad-toothed rats (Mastacomys fuscus) at Barrington Tops, New South Wales, and model its dynamics using metapopulation theory. The study population occurs patchily in swamps on the Barrington Plateau. Using faecal pellet searches, live-trapping and data from previous surveys, we identified 12 swamps where M. fuscus persists and 13 where there has been a history of colonisation and extinction. The species now appears to be entirely absent from seven of these latter swamps. Using logistic regression and model selection procedures, we found the strongest predictor of the presence of M. fuscus to be proximity to the nearest occupied swamp. Persistence declined strongly with swamp isolation, probably due to low success of individuals dispersing through the intervening habitat. These patterns support the interpretation that swamp patches at Barrington Tops contain a functioning but fragile metapopulation of M. fuscus. We predict continued loss of remaining sub-populations in peripheral swamps if current dynamics continue, and recommend research to identify the factors that are limiting dispersal and re-colonisation so that the species’ decline can be slowed and reversed.     

Acanthaster planci on the great barrier reef: Detailed surveys of four transects between 19° and 20°S

In 1973 and 1974 two surveys were conducted of all reefs falling within four transects in the region of the Great Barrier Reef thought to represent the southern section of the area in which large populations of Crown-of-thorns starfish (Acanthaster planci (L.)) were active. The surveys were carried out by towed or free-swimming snorkel divers who covered the entire perimeter and the majority of back reef and lagoonal areas of reefs visited. Of 25 reefs surveyed ten were found to carry large populations of A. planci during one or both surveys. The interval between surveys was 11 months for transects 1–3 and six months for transect 4. During the interval A. planci populations increased on two reefs but decreased on five although the overall number of A. planci sighted was similar for the two surveys.The surveys revealed considerable changes in the density and distribution of A. planci populations on individual reefs but did not produce evidence of migration of adult A. planci onto previously unpopulated reefs.Dense populations of A. planci were studied by SCUBA divers; the highest population density recorded during the surveys was considerably less than that reported during earlier surveys of starfish populations on the Great Barrier Reef.Although there were areas on a number of reefs in which considerable coral mortality appeared to have been caused by the predation of A. planci they were generally restricted and interspersed with areas with a moderate cover of live corals.     

Methods to detect species range size change from biological atlas data: A comparison using the South African Frog Atlas Project

Comparing historical species distribution data with recent biological surveys has the potential to quantify changes to species geographic ranges, and to detect declining species. The primary concern in making species range size comparisons between dissimilar datasets is the problem of incomparable sampling intensities. For South African amphibian data, an increase in sampling intensity over time was associated with a move from ad hoc to more coordinated sampling. The consequent increased geographic coverage and higher detection probabilities resulted in false range expansions. We assessed several approaches for managing variable sampling intensity: categorical comparisons, sub-sampling and correction factors. These methods were first applied to hypothetical ranges for evaluation, and then to the observed data of the South African Frog Atlas Project. Categorical comparisons oversimplified quantitative data and correctly measured the sign of change for fewer than 56% of the hypothetical ranges. Sub-sampling methods correctly estimated the sign of change for up to 78% of the hypothetical ranges, but it is essential to balance data removal with a reduction of false range expansions. The best measure of hypothetical range change was a mathematical correction factor, which achieved 83% accuracy in detecting the correct sign of change and 50% for the magnitude of change. According to this method, 68 (60.2%) South African frog species are likely to have undergone range contractions. The careful use of methods such as those assessed here allows for ad hoc data collections and biological surveys to be valuable in quantifying species range change, and hence for identifying potentially declining species.