Average Correct Classification Rate Research Articles

Assessment of the 16S‐23S rDNA Intergenic Spacer Region in Enterococcus spp. for Microbial Source Tracking

A new library-based microbial source tracking (MST) approach intended for initial application in the coastal waters of Virginia was evaluated. Host-origin isolates of Enterococcus spp. were collected from beaches and the surrounding tidewater region of Virginia and used to construct a library based on the pattern of DNA band lengths produced by the amplification of the 16S-23S rDNA intergenic spacer (IGS) region, and subsequent digestion with MboI. Initial results from small host-origin libraries (64 and 200 total isolates) with discriminant analysis (DA) and logistic regression (LR) yielded high average rates of correct classification (ARCC) for a four-source classification split (birds, dogs, sewage, and wildlife), with ARCCs ranging from 83 to 100%. However, the poor results obtained when classification was attempted on a non-library validation set (VS, ARCCs of 47 and 48%, respectively, using DA and LR) demonstrated that a library of 200 isolates was insufficient to adequately represent the diversity of the enterococci in the sampled region. An increase in the library size to 1029 total isolates was accompanied by a reduction in the ARCC of the library to 42.7% with DA and 45.7% with LR, plus similarly poor results obtained from the VS. The low correct classification rates generated by the larger known-source library were unsuitable for field application. Many reported MST methods have been based on results obtained using small host-origin libraries without external validation. Our results indicate that such an approach can be very misleading, and that larger libraries and external validation is essential for the confirmation of preliminary results.

Size-invariant descriptors for detecting regions of abnormal growth in cervical vertebrae

Digitized spinal X-ray images exhibiting specific pathological conditions such as osteophytes can be retrieved from large databases using Content Based Image Retrieval (CBIR) techniques. For efficient image retrieval, it is important that the pathological features of interest be detected with high accuracy. In this study, new size-invariant features were investigated for the detection of anterior osteophytes, including claw and traction in cervical vertebrae. Using a K-means clustering and nearest neighbor classification approach, average correct classification rates of 85.80%, 86.04% and 84.44% were obtained for claw, traction and anterior osteophytes, respectively.

Using nutrient utilization patterns to determine the source of Escherichia coli found in surface water

Identifying the sources of fecal contaminants in surface water bodies such as rivers, lakes and beaches is of importance for environmental safety, public health safety, food safety and regulatory purposes. Nutrient utilization patterns (NUPs) were used as a bacterial source tracking technique to identify the possible sources of fecal coliform bacteria, Escherichia coli in Silver Lake, Delaware County, Iowa. A total of three hundred (300) E. coli isolates collected from different sources (water, birds, geese, cattle, hogs and soil contaminated by feces) were analyzed. A database was built from these isolates by using discriminant analysis to identify the nutrient utilization patterns that best classify all 300 isolates by source. The average rate of correct classification by source was 89.5% when applying the nutrient utilization patterns database. After this verification, the NUP for E. coli isolates from Silver Lake water were measured. Based on the NUPs of the Silver Lake isolates, 73.1% were found to originate from cattle and hogs. Smaller percentages were predicted to be coming from birds and geese. None of the isolates were predicted to be originating from the human source. The results indicate that livestock are the primary contributors to fecal pollution in this hypereutrophic Iowa lake. Keywords: Nutrient utilization patterns, Escherichia coli , bacterial source tracking, Silver Lake, Delaware, Iowa

Neural nets versus conventional techniques in credit scoring in Egyptian banking

Neural nets have become one of the most important tools using in credit scoring. Credit scoring is regarded as a core appraised tool of commercial banks during the last few decades. The purpose of this paper is to investigate the ability of neural nets, such as probabilistic neural nets and multi-layer feed-forward nets, and conventional techniques such as, discriminant analysis, probit analysis and logistic regression, in evaluating credit risk in Egyptian banks applying credit scoring models. The credit scoring task is performed on one bank's personal loans' data-set. The results so far revealed that the neural nets-models gave a better average correct classification rate than the other techniques. A one-way analysis of variance and other tests have been applied, demonstrating that there are some significant differences amongst the means of the correct classification rates, pertaining to different techniques.

Evaluation of Repetitive Extragenic Palindromic‐PCR for Discrimination of Fecal Escherichia coli from Humans, and Different Domestic‐ and Wild‐Animals

The objective of this study was to investigate the potential of repetitive extragenic palindromic anchored polymerase chain reaction (rep-PCR) in differentiating fecal Escherichia coli isolates of human, domestic- and wild-animal origin that might be used as a molecular tool to identify the possible source(s) of fecal pollution of source water. A total of 625 fecal E. coli isolates of human, 3 domestic- (cow, dog and horse) and 7 wild-animal (black bear, coyote, elk, marmot, mule deer, raccoon and wolf) species were characterized by rep-PCR DNA fingerprinting technique coupled with BOX A1R primer and discriminant analysis. Discriminant analysis of rep-PCR DNA fingerprints of fecal E. coli isolates from 11 host sources revealed an average rate of correct classification of 79.89%, and 84.6%, 83.8%, 83.3%, 82.5%, 81.6%, 80.8%, 79.8%, 79.3%, 77.4%, 73.2% and 63.6% of elk, human, marmot, mule deer, cow, coyote, raccoon, horse, dog, wolf and black bear fecal E. coli isolates were assigned to the correct host source. These results suggest that rep-PCR DNA fingerprinting procedures can be used as a source tracking tool for detection of human- as well as animal-derived fecal contamination of water.

Fecal source tracking by antibiotic resistance analysis on a watershed exhibiting low resistance

The ongoing development of microbial source tracking has made it possible to identify contamination sources with varying accuracy, depending on the method used. The purpose of this study was to test the efficiency of the antibiotic resistance analysis (ARA) method under low resistance by tracking the fecal sources at Turkey Creek, Oklahoma exhibiting this condition. The resistance patterns of 772 water-isolates, tested with nine antibiotics, were analyzed by discriminant analysis (DA) utilizing a five-source library containing 2250 isolates. The library passed various representativeness tests; however, two of the pulled-sample tests suggested insufficient sampling. The resubstitution test of the library individual sources showed significant isolate misclassification with an average rate of correct classification (ARCC) of 58%. These misclassifications were explained by low antibiotic resistance (Wilcoxon test P < 0.0001). Seasonal DA of stream E. coli isolates for the pooled sources human/livestock/deer indicated that in fall, the human source dominated (P < 0.0001) at a rate of 56%, and that human and livestock respective contributions in winter (35 and 39%), spring (43 and 40%), and summer (37 and 35%) were similar. Deer scored lower (17-28%) than human and livestock at every season. The DA was revised using results from a misclassification analysis to provide a perspective of the effect caused by low antibiotic resistance and a more realistic determination of the fecal source rates at Turkey Creek. The revision increased livestock rates by 13-14% (0.04 <or= P <or= 0.06), and decreased human and deer by 6-7%. Negative misclassification into livestock was significant (0.04 <or= P <or= 0.06). Low antibiotic resistance showed the greatest effect in this category.

Microbial source tracking in a rural watershed dominated by cattle

Antibiotic resistance analysis (ARA), frequency of sampling, and seasonality were evaluated in a rural Virginia watershed dominated by cattle. The selected watershed (Mill Creek) was 3767 ha in size, included two small communities (one sewered and one unsewered), and several farms that when combined contained over 3800 beef and dairy cattle. Monthly monitoring of fecal coliforms at two sampling sites in Mill Creek from January to December, 2001, revealed that the recreational standard (1000 colony forming units, CFUs/100 ml) was exceeded a total of eight times for a 33% violation rate at each site. In addition, stream samples were collected weekly for 4 consecutive weeks during seasonal high flows (March) and seasonal low flows (September–October), plus daily for 7 consecutive days within the weekly schedules for a combined total of 60 stream samples (30 at each of two sites). The recreational standard was exceeded once during seasonal high flow and nine times during seasonal low flow. Microbial source tracking (MST) was performed by ARA to assess the impact of cattle on water quality within the different sampling routines. The resistance patterns of 2880 water isolates and 1158 known source (host-origin) isolates were determined with seven antibiotics at 28 different concentrations. The 1158 isolate database was reduced to 562 unique isolates when clonal ARA patterns were removed. This database of 562 unique isolates had an average rate of correct classification (ARCC) of 95.4%, and several statistical procedures confirmed the library as accurate and representative. Sixty-five percent of 50 challenge-set isolates from sources, but not samples, used in the library were correctly identified. The 562 unique pattern database was used to classify Escherichia coli isolates from water samples into six host source categories. The ARA results showed that cattle were the major source of pollution in the stream and cattle were the dominant source in over 60% of the water samples. Sampling frequency and seasonality had no effect on the MST results, as cattle dominated both seasons and samplings. Deer were a minor contributor in the summer (high water demand), and geese were a minor contributor in the winter when migratory flocks were observed moving through the watershed. An unexpected human allocation was found, especially under seasonal high flow conditions. The exact origin of this human allocation is not known. This project demonstrated that a host-origin library, based on a phenotypic method, could be developed for a well-defined watershed and was both representative of the sources in the watershed and performed reasonably well against a challenge set.

Considerations When Using Discriminant Function Analysis of Antimicrobial Resistance Profiles To Identify Sources of Fecal Contamination of Surface Water in Michigan

The goals of this study were to (i) identify issues that affect the ability of discriminant function analysis (DA) of antimicrobial resistance profiles to differentiate sources of fecal contamination, (ii) test the accuracy of DA from a known-source library of fecal Escherichia coli isolates with isolates from environmental samples, and (iii) apply this DA to classify E. coli from surface water. A repeated cross-sectional study was used to collect fecal and environmental samples from Michigan livestock, wild geese, and surface water for bacterial isolation, identification, and antimicrobial susceptibility testing using disk diffusion for 12 agents chosen for their importance in treating E. coli infections or for their use as animal feed additives. Nonparametric DA was used to classify E. coli by source species individually and by groups according to antimicrobial exposure. A modified backwards model-building approach was applied to create the best decision rules for isolate differentiation with the smallest number of antimicrobial agents. Decision rules were generated from fecal isolates and applied to environmental isolates to determine the effectiveness of DA for identifying sources of contamination. Principal component analysis was applied to describe differences in resistance patterns between species groups. The average rate of correct classification by DA was improved by reducing the numbers of species classifications and antimicrobial agents. DA was able to correctly classify environmental isolates when fewer than four classifications were used. Water sample isolates were classified by livestock type. An evaluation of the performance of DA must take into consideration relative contributions of random chance and the true discriminatory power of the decision rules.

Phenotypic characterization of Escherichia coli through whole-cell fatty acid profiling to investigate host specificity

The objective of the study was to investigate whole-cell fatty acid methyl ester (FAME) profiles of 605 Escherichia coli isolates to determine their host specificity. The isolates were cultured from six possible sources of fecal pollution; 180 isolates from sewage, 85 from dairy cow, 98 from chicken, 76 from swine, 94 from deer, and 72 from waterfowl, mostly geese and ducks. The FAME profiles were presented as the relative masses of 12 FAMEs identified in the isolates and it was found that none of the six hosts carried a “signature” FAME, a FAME that is uniquely associated with a particular host category. However, two-sample t-test analyses indicated that the mean relative masses of seven FAMEs out of the 12 identified showed statistically significant differences (95% confidence interval) between isolates of human and non-human origins. In addition, a linear discriminant function based on mean relative mass variations in individual FAMEs classified the known-source isolates into their respective host categories with a 47.6% average rate of correct classification (ARCC) in a six-way discriminant analysis. The ARCC increased to 61.3% when the individual hosts were pooled into larger categories of human, livestock, and wildlife. The accuracy was 75.5% when isolates of human origin were discriminated against those of non-human origins. Random cluster formation analysis indicated that the library size was sufficient to prevent random grouping among the isolates.

Small-vocabulary speech recognition using surface electromyography

We present results of electromyographic (EMG) speech recognition on a small vocabulary of 15 English words. EMG speech recognition holds promise for mitigating the effects of high acoustic noise on speech intelligibility in communication systems, including those used by first responders (a focus of this work). We collected 150 examples per word of single-channel EMG data from a male subject, speaking normally while wearing a firefighter’s self-contained breathing apparatus. The signal processing consisted of an activity detector, a feature extractor, and a neural network classifier. Testing produced an overall average correct classification rate on the 15 words of 74% with a 95% confidence interval of (71%, 77%). Once trained, the subject used a classifier as part of a real-time system to communicate to a cellular phone and to control a robotic device. These tasks were performed under an ambient noise level of approximately 95 decibels. We also describe ongoing work on phoneme-level EMG speech recognition.

Development and Application of Rapid Antibiotic Resistance Analysis for Microbial Source Tracking in the Black River Watershed, Michigan

Microbial source tracking (MST) is often a desirable component of water quality monitoring. To enable local health departments to utilize MST regularly, rapid methods need to be developed that can be performed without costly equipment, increased laboratory space, or advanced technical requirements. This paper describes the development, validation, and application of a rapid, small-scale MST method based on antibiotic resistance analysis. An automated plate reader allows for earlier and repeated measurements and can distinguish treatment effects after only six hours of incubation, compared with the overnight incubation for the conventional assay. Using this rapid MST assay, a known-source Escherichia coli (E. coli) antibiotic resistance pattern database was constructed for the Black River watershed in southeastern Michigan. The average rate of correct classification for the known-source database was 70.4% when comparing each of seven specific animal sources, 72.5% when comparing three generic animal groups (human, waterfowl, and livestock), and 88.9% when comparing human to non-human animal sources. The rapid assay and database were then used to identify the sources of E. coli at a public recreational beach as human (34%), goose (34%), and cattle (20%).

Evaluation of denaturing gradient gel electrophoresis to differentiate Escherichia coli populations in secondary environments

The development of methodology to differentiate mixed populations of Escherichia coli in the secondary habitat might improve monitoring of fecal pollution indicators and facilitate the development of strategies to mitigate bacterial pollution. The objective of this study was to determine the ability of denaturing gradient gel electrophoresis (DGGE) to differentiate mixed assemblages of E. coli in the natural environment. After confirming the identity of 184 environmental bacterial isolates as E. coli, each was subjected to polymerase chain reaction (PCR) of the beta-glucuronidase gene (uidA) followed by DGGE fingerprinting. The ability of DGGE to discriminate individual isolates at the strain level was determined by comparing fingerprints to those resulting from a standard, library-dependent fingerprinting method, BOX-PCR. Computerized analysis of fingerprints indicated that DGGE and BOX-PCR identified 15 and 21 unique phylotypes respectively. Rank-abundance plots comparing the numerical distribution of unique E. coli phylotypes detected by both methods revealed no difference in resolution at the population level. In water and sediment samples from two beaches, DGGE effectively distinguished indigenous E. coli populations with an average rate of correct classification (site-based) of 83%. Denaturing gradient gel electrophoresis of uidA genes isolated and PCR-amplified from environmental samples appears to be an effective tool to differentiate unique E. coli populations and should be useful to characterize E. coli dynamics in the secondary environment.

Adaptive certainty-based classification for decomposition of EMG signals

An adaptive certainty-based supervised classification approach for electromyographic (EMG) signal decomposition is presented and evaluated. Similarity criterion used for grouping motor unit potentials (MUPs) is based on a combination of MUP shapes and two modes of use of motor unit (MU) firing pattern information: passive and active. Performance of the developed classifier was evaluated using synthetic signals of known properties and real signals and compared with the performance of the certainty classifier (CC). Across the sets of simulated and real EMG signals used for comparison, the adaptive certainty classifier (ACC) had both better average performance and lower performance variability. For simulated signals of varying intensity, the ACC had an average correct classification rate (CCr ) of 83.7% with a mean absolute deviation (MAD) of 5.8% compared to 78.3 and 8.7%, respectively, for the CC. For simulated signals with varying amounts of shape and/or firing pattern variability, the ACC had a CCr of 79.7% with a MAD of 4.7% compared to 76.6 and 6.9%, respectively, for the CC. For real signals, the ACC had a CCr of 70.0% with a MAD of 6.3% compared to 64.9 and 6.4%, respectively, for the CC. The test results demonstrate that the ACC can manage both MUP shape variability as well as MU firing pattern variability. The ACC adapts to EMG signal characteristics to create dynamic data driven classification criteria so that the number of MUP assignments made reflects the signal complexity and the number of erroneous assignments is kept sufficiently low. The ability of the ACC to adjust to specific signal characteristics suggests that it can be successfully applied to a wide variety of EMG signals.

Discriminant analysis of fecal bacterial species composition for use as a phenotypic microbial source tracking method

A rapidly growing method to identify origins of nonpoint source (NPS) pollution is microbial source tracking (MST). Current MST research utilizes either an organism's genetic or physiological traits to establish source identification. To determine if an MST method based on fecal bacterial species composition can be used to determine sources of NPS pollution, samples from known NPS contributors (human, cattle, poultry, and swine) were collected and analyzed for fecal coliform (FC) and fecal streptococci (FS). Five colonies from each bacterial type were randomly selected, isolated and identified using phenotypic profiles. The species composition was calculated from these data and analyzed statistically via discriminant analysis. The rates of correct classification (RCC) for FC species composition patterns were 64, 71, 47 and 70% for cattle, human, poultry and swine, respectively. The RCC for FS species composition patterns were 87, 86, 74, and 83% for cattle, human, poultry, and swine, respectively. The average rate of correct classification for samples from all known sources was significantly higher ( P = 0.05 ) for FS species composition data (82%) than for FC (63%). The average rate of correct classification was increased when the FC and FS species composition data was combined (93%). The results from this study indicate that a phenotypic MST methodology based on species composition of dominant fecal bacteria may be useful in determining major contributors to NPS pollution. Based on the average rates of correct classification, the use of FS species composition patterns appears to be more useful in identifying source than the use of FC patterns.

Differentiation of faecal Escherichia coli from humans and animals by multiple antibiotic resistance analysis

Multiple antibiotic resistance (MAR) was performed on 128 Escherichia coli isolates, recovered from faecal samples of humans and animals (cattle, goat, sheep) to determine and compare their antibiotic resistance patterns and to evaluate them statistically in order to specify the source of the faecal material. Disk diffusion method was applied with a selection of antibiotics. Statistical approach was performed with hierarchical cluster analysis (CA), discriminant analysis (DA) and principal component analysis. Comparing human and animal isolates there was significant difference in levels of resistance to all antibiotics tested (P<0.05) with 46 and 24 distinct resistance patterns for human and animal isolates respectively. CA and DA separated human and animal isolates with a high average rate of correct classification (99.2%), when all animal isolates were pooled together. MAR analysis compared with appropriate statistical evaluation may provide a useful tool for differentiating the human or animal origin of E. coli isolates derived from environmental samples. Subsequently, determination of the source of faecal pollution becomes possible. Determining the source of faecal pollution enables the prediction of possible risk for public health and the application of appropriate management plans for prevention of further contamination.

Evaluation of antibiotic resistance analysis and ribotyping for identification of faecal pollution sources in an urban watershed

The accuracy of ribotyping and antibiotic resistance analysis (ARA) for prediction of sources of faecal bacterial pollution in an urban southern California watershed was determined using blinded proficiency samples. Antibiotic resistance patterns and HindIII ribotypes of Escherichia coli (n = 997), and antibiotic resistance patterns of Enterococcus spp. (n = 3657) were used to construct libraries from sewage samples and from faeces of seagulls, dogs, cats, horses and humans within the watershed. The three libraries were analysed to determine the accuracy of host source prediction. The internal accuracy of the libraries (average rate of correct classification, ARCC) with six source categories was 44% for E. coli ARA, 69% for E. coli ribotyping and 48% for Enterococcus ARA. Each library's predictive ability towards isolates that were not part of the library was determined using a blinded proficiency panel of 97 E. coli and 99 Enterococcus isolates. Twenty-eight per cent (by ARA) and 27% (by ribotyping) of the E. coli proficiency isolates were assigned to the correct source category. Sixteen per cent were assigned to the same source category by both methods, and 6% were assigned to the correct category. Addition of 2480 E. coli isolates to the ARA library did not improve the ARCC or proficiency accuracy. In contrast, 45% of Enterococcus proficiency isolates were correctly identified by ARA. None of the methods performed well enough on the proficiency panel to be judged ready for application to environmental samples. Most microbial source tracking (MST) studies published have demonstrated library accuracy solely by the internal ARCC measurement. Low rates of correct classification for E. coli proficiency isolates compared with the ARCCs of the libraries indicate that testing of bacteria from samples that are not represented in the library, such as blinded proficiency samples, is necessary to accurately measure predictive ability. The library-based MST methods used in this study may not be suited for determination of the source(s) of faecal pollution in large, urban watersheds.

Occurrence of antibiotic resistance inEscherichia colifrom surface waters and fecal pollution sources near Hamilton, Ontario

Antibiotic resistance was examined in 462 Escherichia coli isolates from surface waters and fecal pollution sources around Hamilton, Ontario. Escherichia coli were resistant to the highest concentrations of each of the 14 antibiotics studied, although the prevalence of high resistance was mostly low. Two of 12 E. coli isolates from sewage in a CSO tank had multiple resistance to ampicillin, ciprofloxacin, gentamicin, and tetracycline above their clinical breakpoints. Antibiotic resistance was less prevalent in E. coli from bird feces than from municipal wastewater sources. A discriminant function calculated from antibiotic resistance data provided an average rate of correct classification of 68% for discriminating E. coli from bird and wastewater fecal pollution sources. The preliminary microbial source tracking results suggest that, at times, bird feces might be a more prominent contributor of E. coli to Bayfront Park beach waters than municipal wastewater sources.

The semi-automated classification of sedimentary organic matter in palynological preparations

The capture, analysis and classification of sedimentary organic matter in palynological preparations have been semi-automated. First, the morphological and textural discriminatory features used in classification schemes are measured using a computer-controlled stage and a digital camera mounted on a microscope in combination with Halcon image analysis algorithms. Second, the Exhaustive CHi-square Automatic Interaction Detector classification tree algorithm is applied to all feature measurements to establish their saliency as classification discriminators. Thirdly, the results of the classification tree algorithm are used to determine the inputs used by the actual classifier, which consists of a series of artificial neural networks (ANNs). The Gamma test (GT) is introduced as a tool to help facilitate the best use of limited data and to ensure that the ANNs are not over trained. The results show that the system developed is able to achieve an average correct classification rate of 87%. This is encouraging enough to prompt further research that could result in a commercially viable system. In the future, work will concentrate on refining the image capture component of the system and increasing the size of those databases that have been shown both empirically and by the GT to be too small to facilitate the construction of accurate classifiers.

Simultaneous extraction of geometry and surface properties of targets using simple infrared sensors

We investigate the use of low-cost infrared (IR) sensors for the simultaneous extraction of geometry and surface properties of com- monly encountered features or targets in indoor environments, such as planes, corners, and edges. The intensity measurements obtained from such sensors are highly dependent on the location, geometry, and sur- face properties of the reflecting target in a way that cannot be repre- sented by a simple analytical relationship, therefore complicating the lo- calization and recognition process. We propose the use of angular intensity scans and present an algorithm to process them to determine the geometry and the surface type of the target and estimate its position. The method is verified experimentally with planes, 90-deg corners, and 90-deg edges covered with aluminum, white cloth, and Styrofoam pack- aging material. An average correct classification rate of 80% of both geometry and surface over all target types is achieved and targets are localized within absolute range and azimuth errors of 1.5 cm and 1.1 deg, respectively. Taken separately, the geometry and surface type of targets can be correctly classified with rates of 99 and 81%, respectively, which shows that the geometrical properties of the targets are more distinctive than their surface properties, and surface determination is the limiting factor. The method demonstrated shows that simple IR sensors, when coupled with appropriate processing, can be used to extract sub- stantially more information than that for which such devices are com- monly employed. © 2004 Society of Photo-Optical Instrumentation Engineers.

A Multiscale Landscape Approach to Predicting Bird and Moth Rarity Hotspots in a Threatened Pitch Pine–Scrub Oak Community

Abstract: In the northeastern United States, pitch pine ( Pinus rigida Mill.)–scrub oak (Quercus ilicifolia Wang.) communities are increasingly threatened by development and fire suppression, and prioritization of these habitats for conservation is of critical importance. As a basis for local conservation planning in a pitch pine–scrub oak community in southeastern Massachusetts, we developed logistic‐regression models based on multiscale landscape and patch variables to predict hotspots of rare and declining bird and moth species. We compared predicted moth distributions with observed species‐occurrence records to validate the models. We then quantified the amount of overlap between hotspots to assess the utility of rare birds and moths as indicator taxa. Species representation in hotspots and the current level of hotspot protection were also assessed. Predictive models included variables at all measured scales and resulted in average correct classification rates (optimal cut point) of 85.6% and 89.2% for bird and moth models, respectively. The majority of moth occurrence records were within 100 m of predicted habitat. Only 13% of all bird hotspots and 10% of all moth hotspots overlapped, and only a few small patches in and around Myles Standish State Forest were predicted to be hotspots for both taxa. There was no correlation between the bird and moth species‐richness maps across all levels of richness (r =−0.03, p = 0.62). Species representation in hotspots was high, but most hotspots had limited or no protection. Given the lack of correspondence between bird and moth hotspots, our results suggest that use of species‐richness indicators for conservation planning may be ineffective at local scales. Based on these results, we suggest that local‐level conservation planning in pitch pine–scrub oak communities be based on multitaxa, multiscale approaches.