Automatic classification of field-collected dinoflagellates by artificial neural network

Automatic taxonomic categorisation of 23 species of dinoflagellates was demonstrated using field-collected specimens. These dinoflagellates have been responsible for the majority of toxic and noxious phytoplankton blooms which have occurred in the coastal waters of the European Union in recent years and make severe impact on the aquaculture industry. The performance by human 'expert' ecologists/taxonomists in identifying these species was compared to that achieved by 2 artificial neural network classifiers (multilayer perceptron and radial basis function networks) and 2 other statistical techniques, k-Nearest Neighbour and Quadratic Discriminant Analysis. The neural network classifiers outperform the classical statistical techniques. Over extended trials, the human experts averaged 85% while the radial basis network achieved a best performance of 83%, the multilayer perceptron 66%, k-Nearest Neighbour 60%, and the Quadratic Discriminant Analysis 56%.

Analysis of classification accuracy for pre-filtered multichannel remote sensing data

Noise is one of the main factors degrading the quality of original multichannel remote sensing data and its presence influences classification efficiency, object detection, etc. Thus, pre-filtering is often used to remove noise and improve the solving of final tasks of multichannel remote sensing. Recent studies indicate that a classical model of additive noise is not adequate enough for images formed by modern multichannel sensors operating in visible and infrared bands. However, this fact is often ignored by researchers designing noise removal methods and algorithms. Because of this, we focus on the classification of multichannel remote sensing images in the case of signal-dependent noise present in component images. Three approaches to filtering of multichannel images for the considered noise model are analysed, all based on discrete cosine transform in blocks. The study is carried out not only in terms of conventional efficiency metrics used in filtering (MSE) but also in terms of multichannel data classification accuracy (probability of correct classification, confusion matrix). The proposed classification system combines the pre-processing stage where a DCT-based filter processes the blocks of the multichannel remote sensing image and the classification stage. Two modern classifiers are employed, radial basis function neural network and support vector machines. Simulations are carried out for three-channel image of Landsat TM sensor. Different cases of learning are considered: using noise-free samples of the test multichannel image, the noisy multichannel image and the pre-filtered one. It is shown that the use of the pre-filtered image for training produces better classification in comparison to the case of learning for the noisy image. It is demonstrated that the best results for both groups of quantitative criteria are provided if a proposed 3D discrete cosine transform filter equipped by variance stabilizing transform is applied. The classification results obtained for data pre-filtered in different ways are in agreement for both considered classifiers. Comparison of classifier performance is carried out as well. The radial basis neural network classifier is less sensitive to noise in original images, but after pre-filtering the performance of both classifiers is approximately the same.

A new procedure to optimize the selection of groups in a classification tree

Agglomerative cluster analyses encompass many techniques, which have been widely used in various fields of science. In biology, and specifically ecology, datasets are generally highly variable and may contain outliers, which increase the difficulty to identify the number of clusters. Here we present a new criterion to determine statistically the optimal level of partition in a classification tree. The criterion robustness is tested against perturbated data (outliers) using an observation or variable with values randomly generated. The technique, called Random Simulation Test (RST), is tested on (1) the well-known Iris dataset [Fisher, R.A., 1936. The use of multiple measurements in taxonomic problems. Ann. Eugenic. 7, 179–188], (2) simulated data with predetermined numbers of clusters following Milligan and Cooper [Milligan, G.W., Cooper, M.C., 1985. An examination of procedures for determining the number of clusters in a data set. Psychometrika 50, 159–179] and finally (3) is applied on real copepod communities data previously analyzed in Beaugrand et al. [Beaugrand, G., Ibanez, F., Lindley, J.A., Reid, P.C., 2002. Diversity of calanoid copepods in the North Atlantic and adjacent seas: species associations and biogeography. Mar. Ecol. Prog. Ser. 232, 179–195]. The technique is compared to several standard techniques. RST performed generally better than existing algorithms on simulated data and proved to be especially efficient with highly variable datasets.

Satellite discrimination of Karenia mikimotoi and Phaeocystis harmful algal blooms in European coastal waters: Merged classification of ocean colour data

The detection of dense harmful algal blooms (HABs) by satellite remote sensing is usually based on analysis of chlorophyll-a as a proxy. However, this approach does not provide information about the potential harm of bloom, nor can it identify the dominant species. The developed HAB risk classification method employs a fully automatic data-driven approach to identify key characteristics of water leaving radiances and derived quantities, and to classify pixels into “harmful”, “non-harmful” and “no bloom” categories using Linear Discriminant Analysis (LDA). Discrimination accuracy is increased through the use of spectral ratios of water leaving radiances, absorption and backscattering. To reduce the false alarm rate the data that cannot be reliably classified are automatically labelled as “unknown”. This method can be trained on different HAB species or extended to new sensors and then applied to generate independent HAB risk maps; these can be fused with other sensors to fill gaps or improve spatial or temporal resolution. The HAB discrimination technique has obtained accurate results on MODIS and MERIS data, correctly identifying 89% of Phaeocystis globosa HABs in the southern North Sea and 88% of Karenia mikimotoi blooms in the Western English Channel. A linear transformation of the ocean colour discriminants is used to estimate harmful cell counts, demonstrating greater accuracy than if based on chlorophyll-a; this will facilitate its integration into a HAB early warning system operating in the southern North Sea.

Marine ecosystem services: Linking indicators to their classification

There is a multitude of ecosystem service classifications available within the literature, each with its own advantages and drawbacks. Elements of them have been used to tailor a generic ecosystem service classification for the marine environment and then for a case study site within the North Sea: the Dogger Bank. Indicators for each of the ecosystem services, deemed relevant to the case study site, were identified. Each indicator was then assessed against a set of agreed criteria to ensure its relevance and applicability to environmental management. This paper identifies the need to distinguish between indicators of ecosystem services that are entirely ecological in nature (and largely reveal the potential of an ecosystem to provide ecosystem services), indicators for the ecological processes contributing to the delivery of these services, and indicators of benefits that reveal the realized human use or enjoyment of an ecosystem service. It highlights some of the difficulties faced in selecting meaningful indicators, such as problems of specificity, spatial disconnect and the considerable uncertainty about marine species, habitats and the processes, functions and services they contribute to.

Towards the Industrial Production of Omega-3 Long Chain Polyunsaturated Fatty Acids from a Genetically Modified Diatom Phaeodactylum tricornutum.

The marine diatom Phaeodactylum tricornutum can accumulate up to 30% of the omega-3 long chain polyunsaturated fatty acid (LC-PUFA) eicosapentaenoic acid (EPA) and, as such, is considered a good source for the industrial production of EPA. However, P. tricornutum does not naturally accumulate significant levels of the more valuable omega-3 LC-PUFA docosahexaenoic acid (DHA). Previously, we have engineered P. tricornutum to accumulate elevated levels of DHA and docosapentaenoic acid (DPA) by overexpressing heterologous genes encoding enzyme activities of the LC-PUFA biosynthetic pathway. Here, the transgenic strain Pt_Elo5 has been investigated for the scalable production of EPA and DHA. Studies have been performed at the laboratory scale on the cultures growing in up to 1 L flasks a 3.5 L bubble column, a 550 L closed photobioreactor and a 1250 L raceway pond with artificial illumination. Detailed studies were carried out on the effect of different media, carbon sources and illumination on omega-3 LC-PUFAs production by transgenic strain Pt_Elo5 and wild type P. tricornutum grown in 3.5 L bubble columns. The highest content of DHA (7.5% of total fatty acids, TFA) in transgenic strain was achieved in cultures grown in seawater salts, Instant Ocean (IO), supplemented with F/2 nutrients (F2N) under continuous light. After identifying the optimal conditions for omega-3 LC-PUFA accumulation in the small-scale experiments we compared EPA and DHA levels of the transgenic strain grown in a larger fence-style tubular photobioreactor and a raceway pond. We observed a significant production of DHA over EPA, generating an EPA/DPA/DHA profile of 8.7%/4.5%/12.3% of TFA in cells grown in a photobioreactor, equivalent to 6.4 μg/mg dry weight DHA in a mid-exponentially growing algal culture. Omega-3 LC-PUFAs production in a raceway pond at ambient temperature but supplemented with artificial illumination (110 μmol photons m-2s-1) on a 16:8h light:dark cycle, in natural seawater and F/2 nutrients was 24.8% EPA and 10.3% DHA. Transgenic strain grown in RP produced the highest levels of EPA (12.8%) incorporated in neutral lipids. However, the highest partitioning of DHA in neutral lipids was observed in cultures grown in PBR (7.1%). Our results clearly demonstrate the potential for the development of the transgenic Pt_Elo5 as a platform for the commercial production of EPA and DHA.