Analysis of habitat use based on radio-tagged animals presents difficulties inadequately addressed by current methods. Areas of concern are sampling level, data pooling across individuals, non-independence of habitat proportions, differential habitat use by groups of animals, and arbitrary definition of habitat availability. We advocate proportional habitat use by individual animals as a basis for analysis. Hypothesis testing of such nonstandard multivariate data is done by compositional analysis, which encom- passes all MANOVA/MANCOVA-type linear models. The applications to habitat use range from testing for age class effects or seasonal differences, to examining relationships with food abundance or home range size. We take as an example the comparison of habitat use and availability. The concepts are explained and demonstrated on two data sets, illustrating different methods of treating missing values. We compare utilized with available habitats in two stages, examining home range selection within the overall study area first, then habitat use within the home range. At each stage, assuming that use differs from random, habitats can be ranked according to relative use, and significant between-rank differences located. Compositional analysis is also suited to the analysis of time budgets or diets.

Compositional Analysis of Habitat Use From Animal Radio-Tracking Data

Kernel methods for estimating home range are being used increasingly in wildlife research, but the effect of sample size on their accuracy is not known. We used computer simulations of 10-200 points/ home range and compared accuracy of home range estimates produced by fixed and adaptive kernels with the reference (REF) and least-squares cross-validation (LSCV) methods for determining the amount of smoothing. Simulated home ranges varied from simple to complex shapes created by mixing bivariate normal distributions. We used the size of the 95% home range area and the relative mean squared error of the surface fit to assess the accuracy of the kernel home range estimates. For both measures, the bias and variance approached an asymptote at about 50 observations/home range. The fixed kernel with smoothing selected by LSCV provided the least-biased estimates of the 95% home range area. All kernel methods produced similar surface fit for most simulations, but the fixed kernel with LSCV had the lowest frequency and magnitude of very poor estimates. We reviewed 101 papers published in The Journal of Wildlife Management (JWM) between 1980 and 1997 that estimated animal home ranges. A minority of these papers used nonparametric utilization distribution (UD) estimators, and most did not adequately report sample sizes. We recommend that home range studies using kernel estimates use LSCV to determine the amount of smoothing, obtain a minimum of 30 observations per animal (but preferably >50), and report sample sizes in published results.

Effects of sample size on kernel home range estimates

1.  Dispersal is of critical ecological and evolutionary importance for several issues of population biology, particularly population synchrony, colonization and range expansion, metapopulation and source–sink dynamics, and population genetic structure, but it has not previously been possible to compare dispersal patterns across a wide range of species or to study movement outside the confines of local study areas.



2.  Using resampling methods, we verified that statistically unbiased estimates of average dispersal distance and of intraspecific variance in dispersal distance could be extracted from the bird ringing data of the British Trust for Ornithology.



3.  Using data on 75 terrestrial bird species, we tested whether natal and breeding dispersal were influenced by a species’ habitat requirements, diet, geographical range, abundance, morphology, social system, life history or migratory status. We used allometric techniques to ascertain whether these relationships were independent of body size, and used the method of phylogenetically independent contrasts to ascertain whether they were independent of phylogeny.



4.  Both natal and breeding dispersal distances were lower among abundant species and among species with large geographical ranges. Dispersal distances and life-history variables were correlated independent of phylogeny, but these relationships did not persist after controlling for body size. All morphometrical variables (wing length, tarsus length and bill length) were not significantly correlated with dispersal distances after correcting for body size or phylogenetic relatedness.



5.  Migrant species disperse further than resident ones, this relation was independent of body size but not of phylogeny. A significant positive relation was observed between breeding dispersal distance and long-term population decline among migrants, but not among residents.



6.  The species living in wet habitats disperse further than those living in dry habitats, which could be explained by the greater patchiness of wet habitats in space and/or time. This relationship was observed only for breeding dispersal, suggesting that this habitat variable does not impose the same constraint on natal dispersal.

https://www.jstor.org/stable/pdfplus/2647274.pdf

Patterns of natal and breeding dispersal in birds

1. Although the home range is a fundamental ecological concept, there is considerable debate over how it is best measured. There is a substantial literature concerning the precision and accuracy of all commonly used home range estimation methods; however, there has been considerably less work concerning how estimates vary with sampling regime, and how this affects statistical inferences. 2. We propose a new procedure, based on a variance components analysis using generalized mixed effects models to examine how estimates vary with sampling regime. 3. To demonstrate the method we analyse data from one study of 32 individually marked roe deer and another study of 21 individually marked kestrels. We subsampled these data to simulate increasingly less intense sampling regimes, and compared the performance of two kernel density estimation (KDE) methods, of the minimum convex polygon (MCP) and of the bivariate ellipse methods. 4. Variation between individuals and study areas contributed most to the total variance in home range size. Contrary to recent concerns over reliability, both KDE methods were remarkably efficient, robust and unbiased: 10 fixes per month, if collected over a standardized number of days, were sufficient for accurate estimates of home range size. However, the commonly used 95% isopleth should be avoided; we recommend using isopleths between 90 and 50%. 5. Using the same number of fixes does not guarantee unbiased home range estimates: statistical inferences differ with the number of days sampled, even if using KDE methods. 6. The MCP method was highly inefficient and results were subject to considerable and unpredictable biases. The bivariate ellipse was not the most reliable method at low sample sizes. 7. We conclude that effort should be directed at marking more individuals monitored over long periods at the expense of the sampling rate per individual. Statistical results are reliable only if the whole sampling regime is standardized. We derive practical guidelines for field studies and data analysis.

https://besjournals.onlinelibrary.wiley.com/doi/epdf/10.1111/j.1365-2656.2006.01164.x

Effects of sampling regime on the mean and variance of home range size estimates

Kernel density estimators have been used to estimate home range size but little is known of their statistical properties. I applied kernel-based estimators of home range size, calculated from 95% probability contours of nonparametric density estimators, to computer-simulated radiolocation data. Four hypothetical models of home range suggested by Boulanger and White (1990) were used to evaluate bias and precision of these estimators in estimating known home range sizes. Kernel methods compared well with the best methods that are available for home range size estimation provided the appropriate level of smoothing was selected. I used brush rabbit (Sylvilagus bachmani) telemetry data to illustrate how Monte Carlo methods may also be used to assess estimator performance from field radiolocation data. A kernel estimator is preferred to a harmonic mean estimator in this example because it is less biased (i.e., the harmonic mean method has an inherent problem).

http://www.jstor.org/stable/pdfplus/3801959.pdf

Using Monte Carlo simulation to evaluate kernel-based home range estimators

Ninety-three papers on home-range analysis using radio-tracking data were reviewed; these papers were found in a literature search of 18 of the major journals likely to include such papers, published in the 5-year period to the end of 1988. The review showed that even 25 years after the first radio-tracking studies, in the majority of papers there was still insufficient attention given to accurate and sufficient data collection, and to using appropriate and sophisticated analytical techniques to assess home-range size and configuration. This paper is designed to help people undertaking a radio-tracking study to avoid some of the most common pitfalls. It is based on some of the problems we have experienced studying several species of larger mammals. We use our collective experience to produce a guide on how to plan a radio-tracking study, to highlight some of the potential problems in designing the study and collecting the data, and to identify some of the difficulties that may be encountered during the analytical stages. The advantages and disadvantages of the most frequently used methods of home-range analysis are discussed and methods for determining the minimum number of radio-fixes and techniques for adjusting inadequate sample sizes are described, as are the problems that may be caused by autocorrelated data.

Home‐range analysis using radio‐tracking data–a review of problems and techniques particularly as applied to the study of mammals

(1) Fox capture-mark-recapture studies were reviewed to obtain standardized information on straight-line recovery and dispersal distances. The term recovery distance is used to include all recoveries, but dispersal distance only includes those animals known to have dispersed from their natal range. Data on fox home-range size and population density from these studies are also summarized. (2) The relationship between fox home-range size and population density was investigated, and a regression equation with an inverse relationship was considered to be the most realistic to explain the observed data. (3) Distance moved was correlated positively with home-range size and negatively with population density. Regression equations relating the mean distance moved by male and female foxes to home-range size are presented. (4) The distributions of observed fox recovery distances were examined and modelled. Most distributions showed that the majority of foxes do not move far, but that a small proportion move much greater distances. (5) The relationships between fox population density, home-range size and dispersal and recovery distance are discussed, as is the application of the quantitative relationships to spatial models of fox and fox/rabies population dynamics.

https://www.jstor.org/stable/pdfplus/2403834.pdf

Dispersal distance, home-range size and population density in the red fox (Vulpes vulpes): a quantitative analysis

SUMMARY (1) Field trials to evaluate bait uptake by urban foxes were performed in three areas in Bristol, U.K. Seasonal uptake was investigated by laying and monitoring uptake of bait at 2-month intervals throughout the year in the principal experimental area. A biomarker (iophenoxic acid) was incorporated in one of the summer and one of the winter trials, and subsequently foxes were captured in that area to examine for elevated levels of protein-bound iodine in the blood. Bait uptake at different bait densities per fox was investigated in the other two areas. (2) There was a steady increase in the proportion of baits investigated and taken from January/February to September/October, with a slight decrease at the end of the year. There was considerable variation in the proportion of baits taken in different habitat types. Of the foxes sampled, in the summer 35% of adults and 17% of cubs, and in the winter 29% of adults and 23% of sub-adults, were positive for the biomarker. When the number of baits available per fox was doubled, the rate of bait uptake per fox increased in the same proportion. (3) The results obtained were compared with bait uptake from rural foxes in Europe and North America. Bait uptake in Bristol was generally lower than in other studies, as was the proportion of biomarked individuals. The differences in bait uptake in Bristol, throughout the season and at different bait densities per fox, are discussed in relation to fox foraging behaviour and ecology. Mathematical models predict that at least 75% control levels are necessary to prevent rabies outbreaks from spreading, with even higher levels of control needed in some urban areas, and the results presented here show that a control strategy based on a baiting campaign using the protocol currently favoured in Britain is unlikely to reach a large enough proportion of the fox population to control a rabies outbreak.

W. J. Trewhella

Papers

Home‐range analysis using radio‐tracking data–a review of problems and techniques particularly as applied to the study of mammals

Dispersal distance, home-range size and population density in the red fox (Vulpes vulpes): a quantitative analysis

A field trial evaluating bait uptake by an urban fox (Vulpes vulpes) population