AUTOMATIC FEATURE EXTRACTION AND CLASSIFICATION OF CROSSBILL (LOXIA SPP.) FLIGHT CALLS

ABSTRACT

In this paper a new method for the automatic classification of bird sounds is presented. Our method is based on acoustic parameters (features) taken from the first harmonic component computed from the sound spectrogram. The features are based on a line segment approximation of the first harmonic component. The final feature vectors, consisting of 16 real numbers, are then classified using a self-organizing map (SOM) neural network. Flight calls of four crossbill species (Loxia spp.) are used as a test example. In the first phase, an unsupervised network was trained and tested using common crossbill L. curvirostra flight calls recorded mainly in the Netherlands. The network was tested using two-barred L. leucoptera, Scottish L. scotica and parrot L. pytyopsittacus crossbill flight calls in the second phase. Finally, the results were validated applying the same network to flight calls of common crossbills and parrot crossbills recorded in Finland. The method automatically separated common crossbill flight calls from those of parrot crossbills. The classification accuracy of the Dutch recordings was 58% in the first phase and 54% in the second phase. The Finnish recordings were classified with 54% accuracy.     

Cladoceran community responses to biomanipulation and re‐oligotrophication in Lake Vesijärvi,Finland, as inferred from remains in annually laminated sediment

1. We studied the role of zooplankton in biomanipulation and the subsequent recovery phase in the Enonselkä basin of Lake Vesijärvi, using subfossil cladocerans in annually laminated sediment. Measures to restore the Enonselkä basin included reduction in external nutrient loading and mass removal of plankti‐ and benthivorous fish. Water clarity increased and the lake changed from a eutrophic to a mesotrophic state. However, some signs of increased turbidity were observed after 5–10 years of successful recovery. 2. Annual laminae in a freeze core sample were identified and sliced, based on the seasonal succession of diatoms. Cladoceran remains and rotifer eggs were counted, and Daphnia ephippia and Eubosmina and Bosmina ephippia and carapaces were measured. Annual changes in pelagic species composition were studied with principal component analysis. Individual species abundance, size measurements and various cladoceran‐based indices or ratios (commonly used to reconstruct changes in trophic state and fish predation) were tested for change between four distinct periods: I (1985–1988) dense fish stocks, poor water quality; II (1989–1992) fish removal; III (1993–1997) low fish density, improved water quality; IV (1998–2002) slightly increased fish density and poorer water quality. 3. After the removal of fish, the mean size of Daphnia ephippia and Eubosmina crassicornis ephippia and carapaces increased significantly. In contrast, the percentage of Daphnia did not increase. When based on ephippia, the ratio Daphnia/(Daphnia + E. crassicornis) increased, but the interpretation was obscured by the tolerance of fish predation by small Daphnia and by the fact that bosminids were the preferred food of roach. Moreover, ephippial production by E. crassicornis decreased in recent years. 4. The abundance of Diaphanosoma brachyurum and Limnosida frontosa increased significantly after the fish population was reduced, while that of Ceriodaphnia and rotifers decreased. 5. The expanding littoral vegetation along with improved water clarity was clearly reflected in the concentration of littoral species in the deep sediment core. The species diversity index for the entire subfossil community also increased. 6. The period of faltering recovery was characterised by greater interannual variability and an increased percentage of rotifers. Nevertheless, the mean sizes of Daphnia ephippia and E. crassicornis ephippia and carapaces indicated a low density of fish. The deteriorating water quality was apparently related to multiple stressors in the catchment after rehabilitation, such as intensified lakeshore building, as well as to exceptional weather conditions, challenging the management methods in use.     

Indirect carbon dioxide emissions from producing bioenergy from forest harvest residues

Forest harvest residues are important raw materials for bioenergy in regions practicing forestry. Removing these residues from a harvest site reduces the carbon stock of the forest compared with conventional stem‐only harvest because less litter in left on the site. The indirect carbon dioxide (CO₂) emission from producing bioenergy occur when carbon in the logging residues is emitted into the atmosphere at once through combustion, instead of being released little by little as a result of decomposition at the harvest sites. In this study (1) we introduce an approach to calculate this indirect emission from using logging residues for bioenergy production, and (2) estimate this emission at a typical target of harvest residue removal, i.e. boreal Norway spruce forest in Finland. The removal of stumps caused a larger indirect emission per unit of energy produced than the removal of branches because of a lower decomposition rate of the stumps. The indirect emission per unit of energy produced decreased with time since starting to collect the harvest residues as a result of decomposition at older harvest sites. During the 100 years of conducting this practice, the indirect emission from average‐sized branches (diameter 2 cm) decreased from 340 to 70 kg CO₂ eq. MWh^?1 and that from stumps (diameter 26 cm) from 340 to 160 kg CO₂ eq. MWh^?1. These emissions are an order of magnitude larger than the other emissions (collecting, transporting, etc.) from the bioenergy production chain. When the bioenergy production was started, the total emissions were comparable to fossil fuels. The practice had to be carried out for 22 (stumps) or four (branches) years until the total emissions dropped below the emissions of natural gas. Our results emphasize the importance of accounting for land‐use‐related indirect emissions to correctly estimate the efficiency of bioenergy in reducing CO₂ emission into the atmosphere.     

Increased nest predation in a declining and threatened Temminck's Stint Calidris temminckii population

We measured nesting success of the Temminck's Stint Calidris temminckii along the Finnish Bothnian Bay coast during 19 breeding seasons (1983–2001) and conducted a population census (1999–2002). We found 105 pairs, showing a marked decline from the previous survey (170 pairs 1987–95). Of the 424 'known-fate' nests, 47% hatched. Depredation caused 79.9% of the nest losses. Nesting failures increased from 1983–91 to 1992–2001 owing to a rise in nest predation. The proportion of failed nests that failed because of predation rose from 48.9 to 87.7%. When only depredated nests were considered as losses, Mayfield nest survival probability over the incubation period dropped from 69 to 31% (461 nests). This pattern emerged both in man-made and in natural habitats. Survival probability was independent of habitat type (natural vs. man-made). In an experiment involving videotaping of dummy nests, Common Gull Larus canus and Ruddy Turnstone Arenaria interpres were found to be the most important egg predators.     

Species traits explain recent range shifts of Finnish butterflies

This study provides a novel systematic comparative analysis of the species characteristics affecting the range margin shifts in butterflies towards higher latitudes, while taking phylogenetic relatedness among species into account. We related observed changes in the northern range margins of 48 butterfly species in Finland between two time periods (1992–1996 and 2000–2004) to 11 species traits. Species with positive records in at least ten 10 km × 10 km grid squares (in the Finnish National Butterfly Recording Scheme, NAFI) in both periods were included in the study. When corrected for range size change, the 48 butterfly species had shifted their range margins northwards on average by 59.9 km between the study periods, with maximum shifts of over 300 km for three species. This rate of range shifts exceeds all previously reported records worldwide. Our findings may be explained by two factors: the study region is situated in higher latitudes than in most previous studies and it focuses on the period of most prominent warming during the last 10–15 years. Several species traits exhibited a significant univariate relationship with the range margin shift according to generalized estimation equations (GEE) taking into account the phylogenetic relatedness among species. Nonthreatened butterflies had on average expanded their ranges strongly northwards (84.5 km), whereas the distributions of threatened species were stationary (−2.1 km). Hierarchical partitioning (HP) analysis indicated that mobile butterflies living in forest edges and using woody plants as their larval hosts exhibited largest range shifts towards the north. Thus, habitat availability and dispersal capacity of butterfly species are likely to determine whether they will be successful in shifting their ranges in response to the warming climate.     

Should ecological factors affect the evolution of age at maturity in freshwater clams?

We studied the evolution of age at maturity in freshwater clams of the genus Anodonta in relation to their ecology. We analysed an age-structured density-dependent population dynamics model, which we developed for freshwater clams, using several different options for density dependence. As evolutionary optimality criteria we applied both the maximization of a fitness measure (either intrinsic rate of increase or expected lifetime fecundity) and the concept of evolutionarily stable strategies (ESSs). All three evolutionary criteria yielded estimates which were too high for the optimal age at maturity in a deterministic model with a constant survival rate. The predictions are improved when size-selective predation is included in the model. Mature clams also face a risk of infection by castrating parasites, which would select for delayed maturity. Variable newborn survival selects for earlier reproduction, though the observed levels of stochasticity probably have a negligible effect.     

The fragmented Atlantic rain forest of Brazil: size, shape and distribution of forest fragments

The geographical characteristics of a total of 1839 forest fragments surrounded by sugar cane fields were studied in the Brazilian Atlantic rain forest region of the northeastern state of Pernambuco. The size and shape of the fragments as well as inter-fragment distances and the effects of varying edge width were examined using a geographical information system. The analyses show that the fragments are relatively small and close to each other. Approximately 48% of the rain forest fragments are <10 hectares, while only about 7% are >100 hectares. Forest fragments are close to each other, as fragments located 50m or less apart formed groups that included ca. 50% of the total forest area. At 350m inter-fragment distance, 98% of the rain forest area was included in groups of fragments. Due to the small size and irregular shape of the fragments, the total area of edge zone exceeds that of the interior habitat when the edge width is ca. 60m. At an edge width of 300m ca. 94% of the total fragment area is edge zone. For conservation purposes, ways of establishing networks of forest fragments connected by corridors and stepping stone fragments are demonstrated using GIS. Simulations using these techniques show that reforestation of sugar cane fields between the forest fragments would considerably increase the area of interior forest habitat and connectivity between fragments.     

Cell-specific mechanisms and systemic signalling as emerging themes in light acclimation of C3 plants

Chloroplasts perform essential signalling functions in light acclimation and various stress responses in plants. Research on chloroplast signalling has provided fundamental information concerning the diversity of cellular responses to changing environmental conditions. Evidence has also accumulated indicating that different cell types possess specialized roles in regulation of leaf development and stress acclimation when challenged by environmental cues. Leaf veins are flanked by a layer of elongated chloroplast-containing bundle sheath cells, which due to their central position hold the potential to control the flux of information inside the leaves. Indeed, a specific role for bundle sheath cells in plant acclimation to various light regimes is currently emerging. Moreover, perception of light stress initiates systemic signals that spread through the vasculature to confer stress resistance in non-exposed parts of the plant. Such long-distance signalling functions are related to unique characteristics of reactive oxygen species and their detoxification in bundle sheath cells. Novel techniques for analysis of distinct tissue types, together with Arabidopsis thaliana mutants with vasculature-specific phenotypes, have proven instrumental in dissection of structural hierarchy among regulatory processes in leaves. This review emphasizes the current knowledge concerning the role of vascular bundle sheath cells in light-dependent acclimation processes of C3 plants.