Catálogo de publicaciones - libros

The sensitivity analyses by means of validated ANN models can contribute to improved understanding of the ecology of streams and rivers. The interpretation of resulting sensitivity curves may reveal impacts of environmental conditions on the occurrence of macroinvertebrate taxa. Such additional knowledge can be useful for the bioindication of stream habitats by means of macroinvertebrate assemblages, and enhance our capacity to monitor and mitigate stream ecosystems. The shape of the sensitivity curves of taxa would indicate how important it is to manage disturbances within certain bounds in order to maintain healthy aquatic ecosystems. Taxa with a threshold response to a disturbance appear to be eliminated at a stream site that proves to be beyond a certain disturbance level. Taxa with ramp responses would gradually become rarer as disturbance intensified. The identification of such threshold conditions would provide catchment and water resource managers with a powerful tool.

Overall it can be concluded that ANN provide a powerful tool for stream modelling allowing the user not only to achieve highly accurate predictions but discover information on general trends in the data. Therefore, this methodology can efficiently be applied to determine ecological requirements of stream organisms that are not fully understood.

Artificial neural networks were applied to the prediction and elucidation of two bloom forming algal species in the Nakdong river-reservoir system. The lower Nakdong River, which has characteristics of both rivers and reservoirs, represents a complicated system for algal bloom modeling. Yet, RNN proved capable not only to predict the distinct seasonal abundance and succession of Microcystis aeruginosa and Stephanodiscus hantzschii but elucidate key driving variables by means of sensitivity analyses. Findings of the sensitivity analysis corresponded very well with existing theories on the ecology of these two algae species.

This study yields promising results for the application of machine learning to complex ecosystems such as regulated rivers. It encourages inter-disciplinary research between ecologists, modelers and computer scientists in the newly emerging area of ecological informatics in order to better understand and predict ecological phenomena at different levels of organization.

The use of AFC pulse shape information does improve discrimination of microalgal taxa, and is likely to be even more useful when species that form chains are to be discriminated. The use of RBF ANNs was again shown to be a rapid and useful tool for analysing large sets of high dimensional data.

The results obtained with both methods match closely with the previous results. The predictive power of ANNs has often been demonstrated, and this new study puts to the fore their explicative power which is very interesting in ecological research.

This article paves the way forward for broad research concerning the contribution of the input variables in ANN’s, firstly by the use of other databases to test the methods, secondly by the discovery of new methods and finally by the investigation of other existing methods.

The software LAGRAMGE for computational assemblage and adaptation of ODE by using the expert knowledge and measured data has been applied for the simulation of chl- in Lake Kasumigaura. As a result two types of chl- models were discovered: (1) chl- equations without considering zooplankton grazing assembled and trained by data of consecutive years were data of the last year was used for testing, and (2) chl- equations considering zooplankton grazing assembled and trained by data of the years 1986 to 1989. The test results of the different models have demonstrated that LAGRAMGE can discover ODE that allow to simulate chl- in Lake Kasumigaura for a variety of years. However the generalisation of discovered equations for unseen data of consecutive years was unsatisfactory, and the accuracy of calculated trajectories with regards to timing and magnitudes of peak events was moderate. The results have highlighted the importance of nutrients as growth limiting factors, and the need for considering functional algae groups in order to appropriately represent their selective grazing by zooplankton.

The current study has demonstrated that complex limnological time-series data can beneficially be processed by ANN in order to provide: (1) one-week-ahead forecasting of outbreaks of harmful algae or water quality changes by recurrent supervised ANN, and (2) clusters to unravel ecological relationships regarding seasons, water quality ranges and long-term environmental changes by non-supervised ANN. It has also been shown that these methods provide a useful framework for comparative studies between largely different lakes. Future work will focus on the integration of super- and non-supervised ANN into a representative lake data warehouse archiving long-term time-series of a broad range of lakes and rivers reflecting diverse climate, morphometric and eutrophic conditions. It will further facilitate “basic research on complex interactions (that) will lead to explanations for the variability and unpredictability that presently hamper lake management efforts...” Carpenter (1988).

A hybrid evolutionary algorithm (HEA) has been developed to discover predictive rule sets in complex ecological data. It has been designed to evolve the structure of rule sets by using genetic programming and to optimise the random parameters in the rule sets by means of a genetic algorithm.

HEA was successfully applied to long-term monitoring data of the shallow, eutrophic Lake Kasumigaura (Japan) and the deep, mesotrophic Lake Soyang (Korea). The results have demonstrated that HEA is able to discover rule sets, which can forecast for 7-days-ahead seasonal abundances of blue-green algae and diatom populations in the two lakes with relatively high accuracy but are also explanatory for relationships between physical, chemical variables and the abundances of algal populations. The explanations and the sensitivity analysis for the best rule sets correspond well with theoretical hypotheses and experimental findings in previous studies.

Artificial neural networks were applied to the prediction and elucidation of two bloom forming algal species in the Nakdong river-reservoir system. The lower Nakdong River, which has characteristics of both rivers and reservoirs, represents a complicated system for algal bloom modeling. Yet, RNN proved capable not only to predict the distinct seasonal abundance and succession of Microcystis aeruginosa and Stephanodiscus hantzschii but elucidate key driving variables by means of sensitivity analyses. Findings of the sensitivity analysis corresponded very well with existing theories on the ecology of these two algae species.

This study yields promising results for the application of machine learning to complex ecosystems such as regulated rivers. It encourages inter-disciplinary research between ecologists, modelers and computer scientists in the newly emerging area of ecological informatics in order to better understand and predict ecological phenomena at different levels of organization.

The use of AFC pulse shape information does improve discrimination of microalgal taxa, and is likely to be even more useful when species that form chains are to be discriminated. The use of RBF ANNs was again shown to be a rapid and useful tool for analysing large sets of high dimensional data.

The software LAGRAMGE for computational assemblage and adaptation of ODE by using the expert knowledge and measured data has been applied for the simulation of chl- in Lake Kasumigaura. As a result two types of chl- models were discovered: (1) chl- equations without considering zooplankton grazing assembled and trained by data of consecutive years were data of the last year was used for testing, and (2) chl- equations considering zooplankton grazing assembled and trained by data of the years 1986 to 1989. The test results of the different models have demonstrated that LAGRAMGE can discover ODE that allow to simulate chl- in Lake Kasumigaura for a variety of years. However the generalisation of discovered equations for unseen data of consecutive years was unsatisfactory, and the accuracy of calculated trajectories with regards to timing and magnitudes of peak events was moderate. The results have highlighted the importance of nutrients as growth limiting factors, and the need for considering functional algae groups in order to appropriately represent their selective grazing by zooplankton.