Model-predicting the effect of freshwater inflow on saltwater layers, migration and life history of zooplankton in the Arctic Ocean: Towards scenarios and future trends

• Main Author: Schmid, Moritz
• Other Authors: Huettmann, Falk PhD, Mühlenberg, Michael Dr.
• Format: Master Thesis
• Language: English
• Published: 2015
• Subjects: 570; Machine learning; climate change; Arctic Ocean; Zooplankton; Geographic Information System (GIS); Calanus sp; Metridia sp; Future prediction; Canadian Earth System Model 2 (CanESM2); TreeNet; Random Forests; Biologie (PPN619462639); Arctic; Fairbanks; Arktis*; Climate change; Alaska; Copepods
• Online Access: http://hdl.handle.net/11858/00-1735-0000-0022-5F98-2; https://doi.org/10.53846/goediss-4983; https://nbn-resolving.org/urn:nbn:de:gbv:7-11858/00-1735-0000-0022-5F98-2-8

Dt. Titel: Vorhersagemodelle für den Einfluss von Süßwasser Einstrom auf Salzwasser Schichten, Wanderbewegung und Lebenszyklen des Zooplankton im Arktischen Ozean: Szenarien und Trends in der Zukunft The Arctic Ocean is warming up and an increasing freshwater inflow is triggering major changes in ocean layers. This model study aims at creating a baseline, and analyzing the effect of freshwater content changes, subsequent freshwater sealing as well as related parameters in the Arctic Ocean on migration and life history of zooplankton such as copepods and euphausiids. Copepods and euphausiids make for a major part of the zooplankton biomass in the Arctic Ocean, and are an important part of the food chain. Analyses are carried out using an ecosystem-based, spatial modeling approach with machine learning algorithms (Salford Systems TreeNet®, Random Forests® and R implementations). The underlying data consists of over 100 predictors including a globally unique data set of physical oceanography. Raw data that was used in this project is available as metadata from the Core Science Metadata Clearinghouse (former National Biological Information Infrastructure) and available at http://mercury.ornl.gov/clearinghouse/ and on servers from the University of Alaska Fairbanks. The Canadian Earth System Model 2 (CanESM2) was utilized to model the effect of changing climate on zooplankton for the next 100 years and for a low emission (RCP26) and a high emission scenario (RCP85). The results consist of spatially explicit (where every point in the layer is geo referenced) and predicted layers for Geographic Information Systems (GIS) that show predicted plankton presence/random absence as well as the relative index of depth and life stage distribution where the zooplankton is most likely to occur. The models show a clear trend towards an increasing relative index of depth where zooplankton is most likely to be found for the year 2100. Moreover, a trend towards a diminishing ecological niche for adult life stages of zooplankton was ...