Search

The increasing penetration of renewable energy sources tends to redirect the power systems community’s interest from the traditional power grid model towards the smart grid framework. During this transition, load forecasting for various time horizons constitutes an essential electric utility task in network planning, operation, and management. This paper presents a novel mixed power-load forecasting scheme for multiple prediction horizons ranging from 15 min to 24 h ahead. The proposed approach makes use of a pool of models trained by several machine-learning methods with different characteristics, namely neural networks, linear regression, support vector regression, random forests, and sparse regression. The final prediction values are calculated using an online decision mechanism based on weighting the individual models according to their past performance. The proposed scheme is evaluated on real electrical load data sensed from a high voltage/medium voltage substation and is shown to be highly effective, as it results in R2 coefficient values ranging from 0.99 to 0.79 for prediction horizons ranging from 15 min to 24 h ahead, respectively. The method is compared to several state-of-the-art machine-learning approaches, as well as a different ensemble method, producing highly competitive results in terms of prediction accuracy.

Natural control of invertebrate crop pests has the potential to complement or replace conventional insecticide based practices, but its mainstream application is hampered by predictive unreliability across agroecosystems. Inconsistent responses of natural pest control to changes in landscape characteristics have been attributed to ecological complexity and system-specific conditions. Here, we review agroecological models and their potential to provide predictions of natural pest control across agricultural landscapes. Existing models have used a multitude of techniques to represent specific crop-pest-enemy systems at various spatiotemporal scales, but less wealthy regions of the world are underrepresented. A realistic representation of natural pest control across systems appears to be hindered by a practical trade-off between generality and realism. Nonetheless, observations of context-sensitive, trait-mediated responses of natural pest control to land-use gradients indicate the potential of ecological models that explicitly represent the underlying mechanisms. We conclude that modelling natural pest control across agroecosystems should exploit existing mechanistic techniques towards a framework of contextually bound generalizations. Observed similarities in causal relationships can inform the functional grouping of diverse agroecosystems worldwide and the development of the respective models based on general, but context-sensitive, ecological mechanisms. The combined use of qualitative and quantitative techniques should allow the flexible integration of empirical evidence and ecological theory for robust predictions of natural pest control across a wide range of agroecological contexts and levels of knowledge availability. We highlight challenges and promising directions towards developing such a general modelling framework.

Complexity and context-dependence in ecological and socioecological phenomena often cause inconsistent or seemingly idiosyncratic responses. This apparent lack of generality presents a challenge to the implementation of ecological principles in environmental management. We mostly rely for prediction on data-hungry correlative models that offer little mechanistic understanding. The alternative of process-based modelling is knowledge-intensive and has limited applicability across systems, due to a trade-off between generality and realism. Here, we present an archetype approach, combining trait-mediated mechanistic understanding into ecological models of intermediate generality, as a way to overcome these limitations. We demonstrate its potential in the case of natural pest control across crop-pest-enemy systems. After reviewing the current modelling approaches and their shortcomings, we show that similar responses of natural pest control in cases that share key characteristics indicate the potential for context-sensitive generalizations. Example archetypes show great promise for improved understanding and prediction. We outline how statistical analysis of available data and rule sets for model development will allow systematic definition of archetypes representing the key processes of all major crop-pest-enemy systems of the world. In this and other applied cases, the use of archetype approaches is a major step forward in facilitating both scientific synthesis and uptake of ecological knowledge in environmental management. Archetype approaches can enhance the application of ecological principles not only to crop pest control, but also to the management of a wide range of natural resources that have been affected by fragmentation of ecological research.

WOS:000435510100008; International audience; The complexity and scales of the processes that shape communities of marine benthic macroinvertebrates has limited our understanding of their assembly mechanisms and the potential to make projections of their spatial and temporal dynamics. Individual-based models can shed light on community assembly mechanisms, by allowing observed spatiotemporal patterns to emerge from first principles about the modeled organisms. Previous work in the Rance estuary (Brittany, France) revealed the principal functional components of its benthic macroinvertebrate communities and derived a set of functional relationships between them. These elements were combined here for the development of a dynamic and spatially explicit model that operates at two spatial scales. At the fine scale, modeling each individual's life cycle allowed the representation of recruitment, inter-and intra-group competition, biogenic habitat modification and predation mortality. Larval dispersal and environmental filtering due to the tidal characteristics of the Rance estuary were represented at the coarse scale. The two scales were dynamically linked and the model was parameterized on the basis of theoretical expectations and expert knowledge. The model was able to reproduce some patterns of alpha-and beta-diversity that were observed in the Rance estuary in 1995. Model analysis demonstrated the role of local and regional processes, particularly early post-settlement mortality and spatially restricted dispersal, in shaping marine benthos. It also indicated biogenic habitat modification as a promising area for future research. The combination of this mechanism with different substrate types, along with the representation of physical disturbances and more trophic categories, could increase the model's realism. The precise parameterization and validation of the model is expected to extend its scope from the exploration of community assembly mechanisms to the formulation of predictions about the responses of community structure and functioning to environmental change.

Benthic macroinvertebrates are part of a complex network of interactions. The spatial and temporal scales of the processes that form the basis for these interactions have traditionally restricted their empirical investigation. The first chapter of the manuscript attempts a review of the modelling tools that have been employed for the study of the marine benthos. The implementation of a mechanistic modelling framework seems fitting, but it requires the derivation of a few model entities with a clear functional role. The second chapter of the manuscript employs the emergent group hypothesis to do that in a way that is objective and testable. The resulting grouping is tested against theoretical expectations and the results support its ability to represent functional diversity in the Rance estuary. The lack of knowledge for the attribution of relationships among functional components is still important. The third chapter of the manuscript addresses this issue based on ecological theories that predict the existence of functional trade-offs operating at both large and small spatial scales. In a first inception of the system, these elements are incorporated in the form of general rules of interaction into qualitative models of the functional groups. In spite of the interest in developing and analysing qualitative models, the goal of studying the dynamic and spatially explicit behaviour of benthic biodiversity can only be reached by a model with the same characteristics. The fourth chapter of the manuscript presents the architecture of an individual-based model, primarily transferring the rules of interaction from the qualitative models to a dynamic and spatially explicit framework., Les macroinvertébrés benthiques entretiennent un ensemble complexe d’interactions. Les échelles spatiales et temporelles des processus formant la base de ces interactions ont traditionnellement limité leur étude empirique. Le premier chapitre du manuscrit tente une revue des outils de modélisation utilisés dans l’étude du benthos marin. Même si l’implémentation d’un modèle mécaniste semble s’ajuster aux communautés benthiques, son utilisation nécessite la création d’un nombre limité d’entités avec un rôle fonctionnel clair. Le second chapitre du manuscrit utilise l’hypothèse des groupes émergents, afin de faire cela via une procédure objective et testable. Le groupement est testé face aux postulats théoriques et les résultats supportent sa capacité à reproduire la diversité fonctionnelle dans l’estuaire de la Rance. Le manque de connaissances dans l’attribution des relations entre les composantes fonctionnelles reste important. Le troisième chapitre du manuscrit s’inscrit dans ce besoin, basé sur des théories écologiques qui prévoient l’existence de trade-offs fonctionnels opérant à grande et petite échelle. Dans un premier temps, ces éléments sont incorporés dans des modèles qualitatifs des groupes fonctionnels. Malgré l’intérêt du développement et de l’analyse de modèles qualitatifs, le but d’étudier la dynamique et le comportement spatialement explicite de la biodiversité ne peut être atteint que par un modèle avec ces mêmes caractéristiques. Le quatrième chapitre du manuscrit présente l’architecture d’un modèle individu-centré, en mettant l’accent sur le transfert des règles d’interactions des modèles qualitatifs vers un cadre dynamique et spatialement explicite

Die Braunalge Fucus vesiculosus wird betrachtet als eine Schlüsselart für die Küstensysteme der Ostsee. Aufgrund der beständigen Verringerung ihrer Eindringtiefe ist heutzutage eine ausgedehnte Forschung nach den Faktoren hervorgerufen, die ihr Wachstum kontrollieren. In dieser Studie wurde ein Modell entwickelt, das versucht die komplexen Interaktionen um F. vesiculosus herum darzustellen, um ihre Wachstumsrate während des Jahres und die größte Tiefe ihrer vertikalen Verbreitung vorherzusehen. Gemäß dem Modell befindet sich die Spitze der Wachstumsrate im Frühsommer und die Tiefgrenze der Alge in einer Tiefe von 9,28 m, übereinstimmend mit Beobachtungen. Das Licht und sein Vorhandensein von Phytoplankton stellen die wichtigsten Faktoren dar, die das Braunalgenwachstum kontrollieren können. Das Fehlen von Elementen von Rekrutierungsprozessen und der geringe Fortschritt der Forschung im Bereich der biotischen Interaktionen, die die Fitness von F. vesiculosus beeinflussen, beschränken das Potenzial des Modells. Die weitere Entwicklung entsprechender Modelle verspricht es neue Herausforderungen zu der Wissenschaft zu stellen.