Rookognise: Acoustic detection and identification of individual rooks in field recordings using multi-task neural networks

Individual-level monitoring is essential in many behavioural and bioacoustics studies. Collecting and annotating those data is costly in terms of human effort, but necessary prior to conducting analysis. In particular, many studies on bird vocalisations also involve manipulating the animals or human presence during observations, which may bias vocal production. Autonomous recording units can be used to collect large amounts of data without human supervision, largely removing those sources of bias. Deep learning can further facilitate the annotation of large amounts of data, for instance to detect vocalisations, identify the species, or recognise the vocalisation types in recordings. Acoustic individual identification, however, has so far largely remained limited to a single vocalisation type for a given species. This has limited the use of those techniques for automated data collection on raw recordings, where many individuals can produce vocalisations of varying complexity, potentially overlapping one another, with the additional presence of unknown and varying background noise. This paper aims at bridging this gap by developing a system to identify individual animals in those difficult conditions. Our system leverages a combination of multi-scale information integration, multi-channel audio and multi-task learning. The multi-task learning paradigm is based the overall task into four sub-tasks, three of which are auxiliary tasks: the detection and segmentation of vocalisations against other noises, the classification of individuals vocalising at any point during a sample, and the sexing of detected vocalisations. The fourth task is the overall identification of individuals. To test our approach, we recorded a captive group of rooks, a Eurasian social corvid with a diverse vocal repertoire. We used a multi-microphone array and collected a large scale dataset of time-stamped and identified vocalisations recorded, and found the system to work reliably for the defined tasks. To our knowledge, the system is the first to acoustically identify individuals regardless of the vocalisation produced. Our system can readily assist data collection and individual monitoring of groups of animals in both outdoor and indoor settings, even across long periods of time, and regardless of a species’ vocal complexity. All data and code used in this article is available online.     

Presence-absence versus presence-only modelling methods for predicting bird habitat suitability

Habitat suitability models can be generated using methods requiring information on species presence or species presence and absence. Knowledge of the predictive performance of such methods becomes a critical issue to establish their optimal scope of application for mapping current species distributions under different constraints. Here, we use breeding bird atlas data in Catalonia as a working example and attempt to analyse the relative performance of two methods: the Ecological Niche factor Analysis (ENFA) using presence data only and Generalised Linear Models (GLM) using presence/absence data. Models were run on a set of forest species with similar habitat requirements, but with varying occurrence rates (prevalence) and niche positions (marginality). Our results support the idea that GLM predictions are more accurate than those obtained with ENFA. This was particularly true when species were using available habitats proportionally to their suitability, making absence data reliable and useful to enhance model calibration. Species marginality in niche space was also correlated to predictive accuracy, i.e. species with less restricted ecological requirements were modelled less accurately than species with more restricted requirements. This pattern was irrespective of the method employed. Models for wide‐ranging and tolerant species were more sensitive to absence data, suggesting that presence/absence methods may be particularly important for predicting distributions of this type of species. We conclude that modellers should consider that species ecological characteristics are critical in determining the accuracy of models and that it is difficult to predict generalist species distributions accurately and this is independent of the method used. Being based on distinct approaches regarding adjustment to data and data quality, habitat distribution modelling methods cover different application areas, making it difficult to identify one that should be universally applicable. Our results suggest however, that if absence data is available, methods using this information should be preferably used in most situations.     

Effectiveness of mobile apps in teaching field-based identification skills

It has been suggested that few students graduate with the skills required for many ecological careers, as field-based learning is said to be in decline in academic institutions. Here, we asked if mobile technology could improve field-based learning, using ability to identify birds as the study metric. We divided a class of ninety-one undergraduate students into two groups for field-based sessions where they were taught bird identification skills. The first group has access to a traditional identification book and the second group were provided with an identification app. We found no difference between the groups in the ability of students to identify birds after three field sessions. Furthermore, we found that students using the traditional book were significantly more likely to identify novel species. Therefore, we find no evidence that mobile technology improved students’ ability to retain what they experienced in the field; indeed, there is evidence that traditional field guides were more useful to students as they attempted to identify new species. Nevertheless, students felt positively about using their own smartphone devices for learning, highlighting that while apps did not lead to an improvement in bird identification ability, they gave greater accessibility to relevant information outside allocated teaching times.     

Species-specific features affect the ability of census-derived models to map winter avian distribution

We identify autoecological traits of bird species that influence the accuracy of predictive models of species distribution based on census data obtained from stratified sampling. These models would serve as a complementary approach to the development of regional bird atlases. We model the winter bird abundance of 64 terrestrial bird species in 77 census plots in Central Spain (Madrid province), using regression tree analyses. The predicted distribution of species density derived from statistical models (birds/10 ha) was compared with the published relative abundances depicted by a very accurate regional atlas of wintering birds (birds observed per 10 h). Statistical models explained an average of 41.7% of the original deviance observed in the local bird distribution (range 19.6–79.3%). Significant associations between observed relative abundances (atlas data) and predicted average densities in 1×1 km squares within 10×10 km UTMs were attained for 44 out of 64 species. Interspecific discrepancies between predicted and observed distribution maps decreased with between-year constancy in regional bird distribution and the degree of ecological specialization of species. Therefore, statistical modeling using census localities allowed us to depict geographical variations in bird abundance that were similar to those in the quantitative atlas maps. Nevertheless, bird distributions derived from statistical models are less reproducible in some species than in others, depending on their autoecological traits.     

A comparison of algorithms for the identification of specimens using DNA barcodes: examples from gymnosperms

In order to use DNA sequences for specimen identification (e.g., barcoding, fingerprinting) an algorithm to compare query sequences with a reference database is needed. Precision and accuracy of query sequence identification was estimated for hierarchical clustering (parsimony and neighbor joining), similarity methods (BLAST, BLAT and megaBLAST), combined clustering/similarity methods (BLAST/parsimony and BLAST/neighbor joining), diagnostic methods (DNA–BAR and DOME ID), and a new method (ATIM). We offer two novel alignment‐free algorithmic solutions (DOME ID and ATIM) to identify query sequences for the purposes of DNA barcoding. Publicly available gymnosperm nrITS 2 and plastid matK sequences were used as test data sets. On the test data sets, almost all of the methods were able to accurately identify sequences to genus; however, no method was able to accurately identify query sequences to species at a frequency that would be considered useful for routine specimen identification (42–71% unambiguously correct). Clustering methods performed the worst (perhaps due to alignment issues). Similarity methods, ATIM, DNA–BAR, and DOME ID all performed at approximately the same level. Given the relative precision of the algorithms (median = 67% unambiguous), the low accuracy of species‐level identification observed could be ascribed to the lack of correspondence between patterns of allelic similarity and species delimitations. Application of DNA barcoding to sequences of CITES listed cycads (Cycadopsida) provides an example of the potential application of DNA barcoding to enforcement of conservation laws. © The Willi Hennig Society 2006.     

Climate change is creating a mismatch between protected areas and suitable habitats for frogs and birds in Puerto Rico

Climate change is altering the spatial distribution of many species around the world. In response, we need to identify and protect suitable areas for a large proportion of the fauna so that they persist through time. This exercise must also evaluate the ability of existing protected areas to provide safe havens for species in the context of climate change. Here, we combined passive acoustic monitoring, semi-automatic species identification models, and species distribution models of 21 bird and frog species based on past (1980–1989), present (2005–2014), and future (2040–2060) climate scenarios to determine how species distributions relate to the current distribution of protected areas in Puerto Rico. Species detection/non-detection data were acquired across?~?700 sampling sites. We developed always-suitable maps that characterized suitable habitats in all three time periods for each species and overlaid these maps to identify regions with high species co-occurrence. These distributions were then compared with the distribution of existing protected areas. We show that Puerto Rico is projected to become dryer by 2040–2060, and precipitation in the warmest quarter was among the most important variables affecting bird and frog distributions. A large portion of always-suitable areas (ASA) is outside of protected areas (>?80%), and the percent of protected areas that overlaps with always-suitable areas is larger for bird (75%) than frog (39%) species. Our results indicate that present protected areas will not suffice to safeguard bird and frog species under climate change; however, the establishment of larger protected areas, buffer zones, and connectivity between protected areas may allow species to find suitable niches to withstand environmental changes.

     

Characterization of moorland vegetation and the prediction of bird abundance using remote sensing

Aims To characterize and identify upland vegetation composition and height from a satellite image, and assess whether the resulting vegetation maps are accurate enough for predictions of bird abundance. Location South‐east Scotland, UK. Methods Fine‐taxa vegetation data collected using point samples were used for a supervised classification of a Landsat 7 image, while linear regression was used to model vegetation height over the same image. Generalized linear models describing bird abundance were developed using field‐collected bird and vegetation data. The satellite‐derived vegetation data were substituted into these models and efficacy was examined. Results The accuracy of the classification was tested over both the training and a set of test plots, and showed that more common vegetation types could be predicted accurately. Attempts to estimate the heights of both dwarf shrub and graminoid vegetation from satellite data produced significant, but weak, correlations between observed and predicted height. When these outputs were used in bird abundance–habitat models, bird abundance predicted using satellite‐derived vegetation data was very similar to that obtained when the field‐collected data were used for one bird species, but poor estimates of vegetation height produced from the satellite data resulted in a poor abundance prediction for another. Conclusions This pilot study suggests that it is possible to identify moorland vegetation to a fine‐taxa level using point samples, and that it may be possible to derive information on vegetation height, although more appropriate field‐collected data are needed to examine this further. While remote sensing may have limitations compared with relatively fine‐scale fieldwork, when used at relatively large scales and in conjunction with robust bird abundance–habitat association models, it may facilitate the mapping of moorland bird abundance across large areas.     

What Makes Vocalisation Frequency an Unreliable Signal of Body Size in Birds? A Study on Black Swans

In many animal species, the frequency (pitch) of vocalisations correlates negatively with body size and may thus signal competitive ability. However, this relationship is absent in other species. Understanding why this difference exists across species may help to explain some of the diversity of vocal communication systems. We assessed whether vocalisation frequency signals body size in black swans (Cygnus atratus), and how this is affected by (i) variation in frequency within individuals and (ii) size variation across individuals. Frequency was correlated with body size and mass, with slopes close to the allometry expected if the birds were maximising sound radiation, but the explained variation in frequency was low. Within‐individual variation in vocalisation frequency was greater in male than female swans, and the reliability of frequency as a signal of size in males was correspondingly lower. A review of the literature on the relationship between the frequency of avian vocalisations and body size also showed smaller effect sizes for more variable vocalisations (birdsongs), than for simpler vocalisations. Vocalisation frequency was more reliably correlated with body size when the sexes were pooled (creating a larger range of variation in size) than when the relationship was examined for either sex separately, although male and female data followed the same allometric line. These results show that variation in frequency within individuals and low variation in size across individuals reduce the reliability of vocalisation frequency as a signal of body size, which helps to understand differences among species in the signal value of vocalisation frequency.     

Enhanced migratory waterfowl distribution modeling by inclusion of depth to water table data

In addition to being used as a tool for ecological understanding, management and conservation of migratory waterfowl rely heavily on distribution models; yet these models have poor accuracy when compared to models of other bird groups. The goal of this study is to offer methods to enhance our ability to accurately model the spatial distributions of six migratory waterfowl species. This goal is accomplished by creating models based on species-specific annual cycles and introducing a depth to water table (DWT) data set. The DWT data set, a wetland proxy, is a simulated long-term measure of the point either at or below the surface where climate and geological/topographic water fluxes balance. For species occurrences, the USGS' banding bird data for six relatively common species was used. Distribution models are constructed using Random Forest and MaxEnt. Random Forest classification of habitat and non-habitat provided a measure of DWT variable importance, which indicated that DWT is as important, and often more important, to model accuracy as temperature, precipitation, elevation, and an alternative wetland measure. MaxEnt models that included DWT in addition to traditional predictor variables had a considerable increase in classification accuracy. Also, MaxEnt models created with DWT often had higher accuracy when compared with models created with an alternative measure of wetland habitat. By comparing maps of predicted probability of occurrence and response curves, it is possible to explore how different species respond to water table depth and how a species responds in different seasons. The results of this analysis also illustrate that, as expected, all waterfowl species are tightly affiliated with shallow water table habitat. However, this study illustrates that the intensity of affiliation is not constant between seasons for a species, nor is it consistent between species.     

Assessing hunters’ ability to identify shot geese: implications for hunting bag accuracy

Reliable hunting bag statistics are a prerequisite for sustainable harvest management. Recently, Internet-based hunting bag reporting systems have been introduced in some European countries, e.g. Denmark, which may enable faster and more detailed reporting. However, reporting of waterfowl bags on a species-specific level may be biased from the individual hunters’ ability to correctly identify species, particularly because juvenile birds can only be identified from subtle differences. We assessed hunters’ ability to identify the five goose species huntable in Denmark. Identifications were made from a line-up of ten full-bodied geese including adults and juveniles. From a total of 2160 identifications made by active hunters, 85.5% were correct while 14.5% were assigned to a wrong species. Active hunters had on average an identification accuracy of 76.0%, highest for Canada goose (99.1%) and lowest for white-fronted goose (74.6%) and bean goose (73.7%). Identification accuracy was significantly lower for juvenile than for adult individuals of white-fronted and bean geese. Correcting the official Danish Bag Record (2013/2014) for identification accuracy, the bags of white-fronted and bean geese increase by 56.5 and 104.4%, respectively, while the bags of greylag and pink-footed geese decrease by 6.7 and 9.0%; the bag for Canada goose remains unchanged. Although identification accuracy is probably higher under field conditions, the study documents that inaccurate species identification is a source of bias in national bag statistics. Hence, improving identification skills by hunters is important to improve bag data accuracy when based on Internet reporting.     

Mapping from heterogeneous biodiversity monitoring data sources

Field monitoring can vary from simple volunteer opportunistic observations to professional standardised monitoring surveys, leading to a trade-off between data quality and data collection costs. Such variability in data quality may result in biased predictions obtained from species distribution models (SDMs). We aimed to identify the limitations of different monitoring data sources for developing species distribution maps and to evaluate their potential for spatial data integration in a conservation context. Using Maxent, SDMs were generated from three different bird data sources in Catalonia, which differ in the degree of standardisation and available sample size. In addition, an alternative approach for modelling species distributions was applied, which combined the three data sources at a large spatial scale, but then downscaling to the required resolution. Finally, SDM predictions were used to identify species richness and high quality areas (hotspots) from different treatments. Models were evaluated by using high quality Atlas information. We show that both sample size and survey methodology used to collect the data are important in delivering robust information on species distributions. Models based on standardized monitoring provided higher accuracy with a lower sample size, especially when modelling common species. Accuracy of models from opportunistic observations substantially increased when modelling uncommon species, giving similar accuracy to a more standardized survey. Although downscaling data through a SDM approach appears to be a useful tool in cases of data shortage or low data quality and heterogeneity, it will tend to overestimate species distributions. In order to identify distributions of species, data with different quality may be appropriate. However, to identify biodiversity hotspots high quality information is needed.     

ITS序列分析与MALDI-TOF MS质谱技术在丝状真菌鉴定中的应用

丝状真菌常用的鉴定方法为形态方法和基因鉴定方法,前者限于检验人员的知识和技能,后者操作繁琐,费用略昂贵,不适合常规开展。因此,寻找丝状真菌快速鉴定方法势在必行。本文采用VITEK MALDI-TOF MS（基质辅助激光解析电离时间飞行质谱）IVD数据库（3.0版本）对临床分离的254株丝状真菌进行鉴定,并以ITS（internal transcribed spacer 内转录间隔区）序列分析为标准,验证MALDI-TOF MS质谱技术鉴定丝状真菌的准确性。结果表明MALDI-TOF MS质谱技术可以对大部分丝状真菌实现快速、准确的鉴定,其中对毛癣菌属（100%）、毛孢子菌属（100%）、毛霉菌属（100%）、曲霉菌属（96.5%）准确率很高,对犬小孢子菌（75%）、镰刀菌属（50%）、新月弯孢霉（46.2%）准确率较低,对丝状真菌鉴定的总体准确率为86.36%,与ITS测序分析符合率为83.97%。     

Rapid molecular identification of Listeria species by use of real-time PCR and high-resolution melting analysis

Identification of Listeria species via a molecular method is critical for food safety and clinical diagnosis. In this study, an assay integrating real-time quantitative PCR (Q-PCR) with high-resolution melting (HRM) curve analysis was developed and assessed for rapid identification of six Listeria species. The ssrA gene, which encodes a transfer-messenger RNA (tmRNA) is conserved and common to all bacterial phyla, contains a variable domain in Listeria spp. Therefore, Q-PCR and a HRM profile were applied to characterize this gene. Fifty-three Listeria species and 45 non-Listeria species were detected using one primer set, with an accuracy of 100% in reference to conventional methods. There was a 93.3% correction rate to 30 artificially contaminated samples. Thus, Q-PCR with melting profiling analysis proved able to identify Listeria species accurately. Consequently, this study demonstrates that the assay we developed is a functional tool for rapidly identifying six Listeria species, and has the potential for discriminating novel species food safety and epidemiological research.     

Sequencing improves our ability to study threatened migratory species: Genetic population assignment in California's Central Valley Chinook salmon

Effective conservation and management of migratory species requires accurate identification of unique populations, even as they mix along their migratory corridors. While telemetry has historically been used to study migratory animal movement and habitat use patterns, genomic tools are emerging as a superior alternative in many ways, allowing large‐scale application at reduced costs. Here, we demonstrate the usefulness of genomic resources for identifying single‐nucleotide polymorphisms (SNPs) that allow fast and accurate identification of the imperiled Chinook salmon in the Great Central Valley of California. We show that 80 well‐chosen loci, drawn from a pool of over 11,500 SNPs developed from restriction site‐associated DNA sequencing, can accurately identify Chinook salmon runs and select populations within run. No other SNP panel for Central Valley Chinook salmon has been able to achieve the high accuracy of assignment we show here. This panel will greatly improve our ability to study and manage this ecologically, economically, and socially important species and demonstrates the great utility of using genomics to study migratory species.     

Vocal individuality,but not stability,in wild palm cockatoos (Probosciger aterrimus)

The ability to identify individuals within a population is often essential for a detailed understanding of the ecology and conservation of a species. However, some species, including large parrots, are notoriously difficult to catch and mark for individual identification. Palm cockatoos (Probosciger aterrimus) are a large, poorly understood species of parrot which are likely in severe decline within the eastern part – and possibly the western part – of their range on Cape York Peninsula, Australia. Here, we investigated whether three different palm cockatoo call types are sufficiently individually distinctive to function as a non-invasive “marker” for identifying individuals over time. Using Discriminant Function Analysis, overall identification accuracy among 12 putative individuals for all call types was 81% (i.e. 148 out of 183 calls were assigned to the correct individual) on the basis of multiple temporal, energy (amplitude) and frequency measurements on the spectrogram. For three different call types, individual identification accuracy among males and females ranged from 69 to 95%. However, based on a limited sample sizes of five putative individuals between years, our data suggest that individual call structure, as quantified by call parameters, was not stable between years. We discuss the applicability of these results for future studies of palm cockatoos and other parrot species.     

A DNA microarray for identification of selected Korean birds based on mitochondrial cytochrome c oxidase I gene sequences

DNA barcoding with the gene encoding cytochrome c oxidase I (COI) in the mitochondrial genome has been proposed as a standard marker to identify and discover animal species. Some migratory wild birds are suspected of transmitting avian influenza and pose a threat to aircraft safety because of bird strikes. We have previously reported the COI gene sequences of 92 Korean bird species. In the present study, we developed a DNA microarray to identify 17 selected bird species on the basis of nucleotide diversity. We designed and synthesized 19 specific oligonucleotide probes; these probes were arrayed on a silylated glass slide. The length of the probes was 19-24 bps. The COI sequences amplified from the tissues of the selected birds were labeled with a fluorescent probe for microarray hybridization, and unique hybridization patterns were detected for each selected species. These patterns may be considered diagnostic patterns for species identification. This microarray system will provide a sensitive and a high-throughput method for identification of Korean birds.     

Updating bird species distribution at large spatial scales: applications of habitat modelling to data from long-term monitoring programs

Mapping of species distributions at large spatial scales has been often based on the representation of gathered observations in a general grid atlas framework. More recently, subsampling and subsequent interpolation or habitat spatial modelling techniques have been incorporated in these projects to allow more detailed species mapping. Here, we explore the usefulness of data from long-term monitoring (LTM) projects, primarily aimed at estimating trends in species abundance and collected at shorter time intervals (usually yearly) than atlas data, to develop predictive habitat models. We modelled habitat occupancy for 99 species using a bird LTM program and evaluated the predictive accuracy of these models using independent data from a contemporary and comprehensive breeding bird atlas project from the same region. Habitat models from LTM data using generalized linear modelling were significant for all the species and generally showed a high predictive power, albeit lower than that from atlas models. Sample size and species range size and niche breadth were the most important factors behind variability in model predictive accuracy, whereas the spatial distribution of sampling units at a given sample size had minor effects. Although predictive accuracy of habitat modelling was strongly species dependent, increases in sample size and, secondarily, a better spatial distribution of sampling units should lead to more powerful predictive distribution models. We suggest that data from LTM programs, now established in a large number of countries, has the potential for being a major source of good quality data suitable for the estimation and regularly update of distributions at large spatial scales for a number of species.     

Birdstrikes and barcoding: can DNA methods help make the airways safer?

While flying remains one of the safest means of travel, reported birdstrikes on aircraft have risen. This is a result of increased aircraft flight movements, changes in agricultural methods and greater environmental awareness contributing to growing populations of hazardous bird species, as well as more diligent reporting of incidents. Measures to mitigate this hazard require accurate data about the species involved; however, the remains of birds from these incidents are often not easy to identify. Reported birdstrikes include a substantial number where the species cannot be determined from morphology alone. DNA barcoding offers a reliable method of identifying species from very small amounts of organic material such as blood, muscle and feathers. We compare species identification based on morphological criteria and identifications based on mitochondrial cytochrome c oxidase subunit I DNA barcoding methods for New Zealand species. Our data suggest that DNA-based identification can substantially add to the accuracy of species identifications, and these methods represent an important addition to existing procedures to improve air safety. In addition, we outline simple and effective protocols for the recovery and processing of samples for DNA barcoding.     

DNA barcoding and minibarcoding as a powerful tool for feather mite studies

Feather mites (Astigmata: Analgoidea and Pterolichoidea) are among the most abundant and commonly occurring bird ectosymbionts. Basic questions on the ecology and evolution of feather mites remain unanswered because feather mite species identification is often only possible for adult males, and it is laborious even for specialized taxonomists, thus precluding large‐scale identifications. Here, we tested DNA barcoding as a useful molecular tool to identify feather mites from passerine birds. Three hundred and sixty‐one specimens of 72 species of feather mites from 68 species of European passerine birds from Russia and Spain were barcoded. The accuracy of barcoding and minibarcoding was tested. Moreover, threshold choice (a controversial issue in barcoding studies) was also explored in a new way, by calculating through simulations the effect of sampling effort (in species number and species composition) on threshold calculations. We found one 200‐bp minibarcode region that showed the same accuracy as the full‐length barcode (602 bp) and was surrounded by conserved regions potentially useful for group‐specific degenerate primers. Species identification accuracy was perfect (100%) but decreased when singletons or species of the Proctophyllodes pinnatus group were included. In fact, barcoding confirmed previous taxonomic issues within the P. pinnatus group. Following an integrative taxonomy approach, we compared our barcode study with previous taxonomic knowledge on feather mites, discovering three new putative cryptic species and validating three previous morphologically different (but still undescribed) new species.     

Identifying ‘climate keystone species’ as a tool for conserving ecological communities under climate change

Aim

Climate change affects ecological communities via impacts on species. The community's response to climate change can be represented as the temporal trend in a climate-related functional property that is quantified using a relevant functional trait. Noteworthy, some species influence this response in the community more strongly than others.

Innovation

Leveraging on the concept of keystone species, we propose that species with a strong effect on the community's functional response to climate change beyond their relative abundance can be considered as ‘climate keystone species’. We develop a stepwise tool to determine species' effects on a community's climate response and identify climate keystone species. We quantify the species-specific effect by measuring the difference in the community's climate response with and without the species. Next, we identify climate keystone species as those with a strong residual effect after weighting with their relative abundances in the community.

Main Conclusions

To illustrate the use of the stepwise tool with empirical data, we identify climate keystone species that have a strong effect on the change in the average temperature niche in North American bird communities over time and find the identification tool ecologically relevant. Identification of climate keystone species can serve as an additional conservation method to efficiently protect ecological communities and, in turn, the ecosystem functions they provide.