1 / 15

HABs & Neural Networks

HABs & Neural Networks. SESSION 5. Fisheries , marine protected areas , population outbursts , biodiversity shifts. Artificial neural network approach to population dynamics of Harmful Algal Blooms in Alfacs Bay (NW Mediterranean): Case studies of Karlodinium and Pseudo- nitzschia .

gelsey
Download Presentation

HABs & Neural Networks

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. HABs & Neural Networks SESSION 5. Fisheries, marine protectedareas, populationoutbursts, biodiversityshifts Artificial neural network approach to population dynamics of Harmful Algal Blooms in Alfacs Bay (NW Mediterranean): Case studies of Karlodinium and Pseudo-nitzschia. Carles Guallar, Margarita Fernández-Tejedor, Maximino Delgado and Jorge Diogène carlesguallar@gmail.com Barcelona, 29 November 2013

  2. HABs & Neural Networks Alfacs Bay (Ebro Delta) Karlodiniumspp. Pseudo-nitzschiaspp.

  3. HABs & Neural Networks Input layerHiddenlayer Output layer Variable 1 Variable 2 Forecast Variable 3 Variable 4 Variable 5 Characteristics: - Feedforward neural network - Sigmoidfunction - Backpropagationwithmomentumterm and flat spot elimination

  4. HABs & Neural Networks Environmental &Phytoplankton 40.65 Latitude N 40.70 40.75 Meteorological Ebro Riverflowrates 0.55 0.60 0.65 0.70 0.75 Longitude E

  5. HABs & Neural Networks Unique data set

  6. HABs & Neural Networks Quantitativedetectionlimit 3.1 Presence > 3.1 Prediction Cells L-1 Phytoplankton counts Classification < 3.1 Absence

  7. HABs & Neural Networks Log10 (Karlodiniumspp.) - Deepwatertemperature (5thprev. week) - Windgust (3rdprev. week) - Irradiance (8th prev. week) - Atmosfericpressure (Log10, 5thprev. week) - Ebro Riverflowrate (Log10, 5thprev. week) 5 previousweeks Lag (weeks) Log10 (Pseudo-nitzschiaspp.) - Deepwatertemperature (14thprev. week) - Windvelocity (10thprev. week) - Watercolumnsalinity (6th prev. week) - Atmosfericpressure (Log10, 13thprev. week) - Ebro Riverflowrate (Log10, 1stprev. week) 5 previousweeks Lag (weeks)

  8. HABs & Neural Networks One-stepweekAbsence-Presencemodels Karlodinium Pseudo-nitzschia Misclassification error (%) Error characteristics Absence Error characteristics Presence

  9. HABs & Neural Networks One-stepweekPredictionmodels Karlodinium Pseudo-nitzschia Coefficient of determination (R2)

  10. HABs & Neural Networks Neural InterpretationDiagram Absence-Presencemodels Karlodiniummodel Presence Absence Pseudo-nitzschiamodel Presence Absence

  11. HABs & Neural Networks Neural InterpretationDiagram Predictionmodels Karlodiniummodel Log10(Cells L-1) Pseudo-nitzschiamodel Log10(Cells L-1)

  12. HABs & Neural Networks ConnectionWeightApproach Absence-Presencemodels Predictionmodels Karlodiniummodels Pseudo-nitzschiamodels

  13. HABs & Neural Networks ConnectionWeightApproach Biological vs Environmental variables Absence-Presence Prediction Karlodinium Karlodinium Pseudo-nitzschia Pseudo-nitzschia

  14. HABs & Neural Networks Conclusions: Neural networkmodelsweredevelopedtopredictPseudo-nitzschiaspp. and Karlodiniumspp. ThepopulationdynamicsforPseudo-nitzschiaspp. and Karlodiniumspp. were similar forthewholeecosystem. Thebigsize of the neural networkmodelshighlightsthecomplexity of thephytoplanktondynamicsin AlfacsBay. Environmental variables are importantfactorsto drive phytoplanktondynamics in AlfacsBay.

  15. HABs & Neural Networks Thankyouverymuch. • Aknowledgments: • Sistema de Observación y Alerta de Proliferación de Microalgas Nocivas en Zonas de Producción Acuícola Marina (PURGADEMAR; IPT-2011-1707-310000). • Programa de seguiment de la qualitat de les aigües, mol·luscs i fitoplanctontòxic a les zones de producció de mariscdellitoralcatalà de la DGPiAM.

More Related