1,721,154 research outputs found
Marine Strategy, una sfida ed un'opportunità per la Biologia Marina italiana
Full text linkThe Marine Strategy Framework Directive, which came into force in 2008, can be
regarded as the environmental pillar for the Integrated European Maritime Policy. In the first phase of
its implementation EU member Countries carried out an initial assessment of the ecological status, set
environmental targets and defined the concept of Good Ecological Status. While marine biologists from
Italian Universities and other research Institutions actively participated in this process, new challenges
will be brought by its next phases, requiring a deeper involvement of the scientific community and a truly
holistic approach
Modeling Posidonia oceanica shoot density and rhizome primary production
Full text linkPosidonia oceanicameadows rank among the most important and most productive ecosystems in
the Mediterranean basin, due to their ecological role and to the goods and services they provide.
estimations of crucial ecological process such as meadows productivity could play a major role in an
environmental management perspective and in the assessment of P. oceanicaecosystem services.
In this study, a Machine Learning approach, i.e. Random Forest, was aimed at modeling P. oceanica
shoot density and rhizome primary production using as predictive variables only environmental
factors retrieved from indirect measurements, such as maps. Our predictive models showed a good
level of accuracy in modeling both shoot density and rhizome productivity (R
2
= 0.761 and R
2
= 0.736,
respectively). Furthermore, as shoot density is an essential parameter in the estimation of P. oceanica
productivity, we proposed a cascaded approach aimed at estimating the latter using predicted values
of shoot density rather than observed measurements. In spite of the complexity of the problem, the
cascaded Random forest performed quite well (R2 = 0.637). While direct measurements will always
play a fundamental role, our estimates could support large scale assessment of the expected condition
of P. oceanica meadows, providing valuable information about the way this crucial ecosystem works
Collection and analysis of a global marine phytoplankton primary-production dataset
Full text link660285405
A Machine Learning approach to the assessment of the vulnerability of Posidonia oceanica meadows
No full textPosidonia oceanica is an endemic Mediterranean seagrass that ranks among the most important and valuable species, with regard to both its ecological role and the services it provides. Despite this species is one of the main targets of conservation actions, the current regression trend of P. oceanica is alarming, underlying the urgent need for reliable methods capable of assessing meadows vulnerability. To address this need, we developed a Habitat Suitability Model (HSM) aimed at assessing the vulnerability of P. oceanica meadows in the Italian marine coastal waters using the Random Forest (RF) Machine Learning technique. Building on the current knowledge on both spatial distribution and condition of meadows in the Italian seas, the RF was used as a classifier aimed at modeling the habitat suitability for P. oceanica, rather than for predictive purposes. The assessment of the potentially most vulnerable P. oceanica meadows at increasing risk of regression was performed through the analysis of the RF output. The HSM showed a good level of accuracy, i.e. Cohen’s K = 0.685. The proposed approach provided valuable information regarding the vulnerability of P. oceanica meadows over the Italian marine coastal waters. In addition, an evaluation of the relative importance of the predictors was carried out using the permutation measure. The developed HSM can support conservation and monitoring programs regarding this species playing a crucial role in the marine ecosystems of the Mediterranean Sea
Embedding ecological knowledge into artificial neural network training: A marine phytoplankton primary production model case study
No full textEnhancing the understanding of marine phytoplankton primary production is paramount due to the relationships
with oceanic food webs, energy fluxes, carbon cycle and Earth's climate.
As field measurements of this process are both expensive and time consuming, indirect approaches, which can
estimate primary production from remotely sensed imagery, are the only viable large-scale solution. We boosted
the quality of phytoplankton primary production estimates, with respect to a previously developed model, by
embedding ecological knowledge into the training of an artificial neural network. In order to achieve this goal,
we drove the training procedure on the basis of both theoretical and data-derived ecological knowledge about
phytoplankton primary production. A “single peak” constraint exploits the theoretical knowledge about the
vertical shape of the production profile; a “depth-weighted error” procedure was based on available information
about the production magnitude along the water column; a variable learning rate and momentum approach
allowed to better exploit the available data to train our artificial neural network.
Thanks to this customized procedure, we improved the quality of the primary production estimates from both
a theoretical and a numerical point of view. Accordingly, the new artificial neural networks not only provided
ecologically sounder estimates, but also explained up to 4% more variance with respect to the traditional error
back-propagation solution. This result was achieved exclusively through an ecologically-driven customization of
the basic algorithm, since the dataset and predictive variables were the same utilized in the conventional
counterpart, trained with a classic error back-propagation algorithm.
We suggest that the proposed rationale could lead to improved performances in similar modelling applications
A Machine Learning Approach to Chlorophyll a Time Series Analysis in the Mediterranean Sea
No full textUnderstanding the dynamics of natural system is a crucial task in ecology especially when climate change is taken into account. In this context, assessing the evolution of marine ecosystems is pivotal since they cover a large portion of the biosphere. For these reasons, we decided to develop an approach aimed at evaluating temporal and spatial dynamics of remotely-sensed chlorophyll a concentration. The concentrations of this pigment are linked with phytoplankton biomass and production, which in turn play a central role in marine environment. Machine learning techniques proved to be valuable tools in dealing with satellite data since they need neither assumptions on data distribution nor explicit mathematical formulations. Accordingly, we exploited the Self Organizing Map (SOM) algorithm firstly to reconstruct missing data from satellite time series of chlorophyll a and secondly to classify them. The missing data reconstruction task was performed using a large SOM and allowed to enhance the available information filling the gaps caused by cloud coverage. The second part of the procedure involved a much smaller SOM used as a classification tool. This dimensionality reduction enabled the analysis and visualization of over 37 000 chlorophyll a time series. The proposed approach provided insights into both temporal and spatial chlorophyll a dynamics in the Mediterranean Basin
A depth-resolved artificial neural network model of marine phytoplankton primary production
No full textMarine phytoplankton primary production is an extremely important process and its estimates play a major role not only in biological oceanography, but also in a broader context, due to its relationship with oceanic food webs, energy fluxes, carbon cycle and Earth's climate.The measurement of this process is both expensive and time consuming. Therefore, indirect methods, which can estimate phytoplankton primary production using only remotely sensed predictive information, have many advantages. We describe the development of a depth-resolved model based on an Artificial Neural Network for estimating global phytoplankton primary production. Furthermore, we applied two different approaches, based on input perturbation analysis and on connection weights, to assess the relative importance of the predictive variables. Finally, we compared the results of our depth-resolved model with a previous depth-integrated solution, showing that through the depth-resolution we gained not only useful information on the vertical distribution of the estimated primary production, but also an enhanced accuracy in its depth-integrated estimates
Un metodo basato sulla fauna ittica e su tecniche di Intelligenza Artificiale per la valutazione dello stato ecologico dei fiumi ai sensi della Direttiva 2000/60/CE
Full text link- …
