A statistical modeling framework is described for estimating the abundances of spatially distinct subpopulations of animals surveyed using removal sampling. To illustrate this framework, hierarchical models are developed using the Poisson and negative-binomial distributions to model variation in abundance among subpopulations and using the beta distribution to model variation in capture probabilities. These models are fitted to the removal counts observed in a survey of a federally endangered fish species. The resulting estimates of abundance have similar or better precision than those computed using the conventional approach of analyzing the removal counts of each subpopulation separately. Extension of the hierarchical models to include spatial covariates of abundance is straightforward and may be used to identify important features of an animal's habitat or to predict the abundance of animals at unsampled locations.

1.Drought is a natural disturbance that can cause widespread mortality of aquatic organisms in wetlands. We hypothesized that seasonal drying of marsh surfaces (i.e. hydrological disturbance) shapes spatio-temporal patterns of fish populations. 2. We tested whether population dynamics of fishes were synchronized by hydrological disturbance (Moran effect) or distance separating study sites (dispersal). Spatio-temporal patterns were examined in local populations of five abundant species at 17 sites (sampled five times per year from 1996 to 2001) in a large oligotrophic wetland. 3. Fish densities differed significantly across spatio-temporal scales for all species. For all species except eastern mosquitofish ( Gambusia holbrooki ), a significant portion of spatio-temporal variation in density was attributed to drying events (used as a covariate). 4. We observed three patterns of response to hydrological disturbance. Densities of bluefin killifish ( Lucania goodei ), least killifish ( Heterandria formosa ), and golden topminnow ( Fundulus chrysotus ) were usually lowest after a dry down and recovered slowly. Eastern mosquitofish showed no distinct response to marsh drying (i.e. they recovered quickly). Flagfish ( Jordanella floridae ) density was often highest after a dry down and then declined. Population growth after a dry down was often asymptotic for bluefin killifish and golden topminnow, with greatest asymptotic density and longest time to recovery at sites that dried infrequently. 5. Fish population dynamics were synchronized by hydrological disturbance (independent of distance) and distance separating study sites (independent of hydrological disturbance). Our ability to separate the relative importance of the Moran effect from dispersal was strengthened by a weak association between hydrological synchrony and distance among study sites. Dispersal was the primary mechanism for synchronous population dynamics of flagfish, whereas hydrological disturbance was the primary mechanism for synchronous population dynamics of the other species examined. 6. Species varied in the relative role of the Moran effect and dispersal in homogenizing their population dynamics, probably as a function of life history and ability to exploit dry-season refugia.

Enclosure traps that quickly surround well-defined areas of habitat are perhaps the most widely used method for sampling fishes in vegetated habitats. However, relatively few data are available to evaluate the effects of habitat structure on sampling characteristics of enclosure traps. In this study, we determined how clearing efficiency and accuracy of 1-m 2 throw traps varied across a range of environmental conditions in the Florida Everglades by sampling within enclosed areas of marsh habitat. Throw trap clearing efficiency and sampling accuracy did not differ among two widely separated locations and appeared to be unaffected by variation in water depth, canopy height, plant cover, plant stem density, and periphyton volume. Sampling accuracy averaged 63% of fishes present after correcting for clearing efficiencies. On average, 83% of the fishes present in a throw trap were recovered. Therefore, it appeared that about 17% of the missing fishes may have burrowed into the substrate or been discarded with sorted detritus. In contrast, the remaining 20% of fishes probably avoided the throw trap. This is the first study to differentiate between potential sources of throw trap sampling errors. Importantly, density estimates obtained by throw traps were positively correlated (r = 0.82) with actual population densities. Mean fish lengths and fish size distributions obtained by throw trapping usually did not differ from actual mean lengths or fish size distributions. Finally, high concordance of fish species ranks indicated that throw traps accurately described fish community structure. Throw traps appeared to provide relatively accurate estimates of fish density, fish size, and community structure across a range of environmental conditions.

We present a novel framework for analyzing univariate time series data. At the heart of the approach is a versatile algorithm for measuring the similarity of two segments of time series called geometric template matching (GeTeM). First, we use GeTeM to compute a similarity measure for clustering and nearest-neighbor classification. Next, we present a semi-supervised learning algorithm that uses the similarity measure with hierarchical clustering in order to improve classification performance when unlabeled training data are available. Finally, we present a boosting framework called TDEBOOST, which uses an ensemble of GeTeM classifiers. TDEBOOST augments the traditional boosting approach with an additional step in which the features used as inputs to the classifier are adapted at each step to improve the training error. We empirically evaluate the proposed approaches on several datasets, such as accelerometer data collected from wearable sensors and ECG data.

In surveys of natural populations of animals, a sampling protocol is often spatially replicated to collect a representative sample of the population. In these surveys, differences in abundance of animals among sample locations may induce spatial heterogeneity in the counts associated with a particular sampling protocol. For some species, the sources of heterogeneity in abundance may be unknown or unmeasurable, leading one to specify the variation in abundance among sample locations stochastically. However, choosing a parametric model for the distribution of unmeasured heterogeneity is potentially subject to error and can have profound effects on predictions of abundance at unsampled locations. In this article, we develop an alternative approach wherein a Dirichlet process prior is assumed for the distribution of latent abundances. This approach allows for uncertainty in model specification and for natural clustering in the distribution of abundances in a data-adaptive way. We apply this approach in an analysis of counts based on removal samples of an endangered fish species, the Okaloosa darter. Results of our data analysis and simulation studies suggest that our implementation of the Dirichlet process prior has several attractive features not shared by conventional, fully parametric alternatives.