Modelling harmful algal blooms

The Earth is surrounded by water and a huge percentage of our livelihood dependent on or related to the waters which surrounds our planet Earth. For example, fisheries livelihood depends on the water with which they catch their stock, other than the weather, market, and resources. The tourism indust...

Full description

Bibliographic Details
Main Author: Tan, Yi Kang
Other Authors: Shu Jian Jun
Format: Final Year Project (FYP)
Language:English
Published: 2017
Subjects:
Online Access:http://hdl.handle.net/10356/72309
Description
Summary:The Earth is surrounded by water and a huge percentage of our livelihood dependent on or related to the waters which surrounds our planet Earth. For example, fisheries livelihood depends on the water with which they catch their stock, other than the weather, market, and resources. The tourism industry also has a component which depends entirely on beaches or beach destinations. And also, agricultural lands often uses irrigation waters from nearby rivers or lakes etc. In this respect, should there be compromises to water quality, like an algal bloom, it directly or indirectly impacts the wide network of human interaction it is linked to. There have been many efforts around the globe to identify, predict, and monitor algal blooms of different areas. Identification of algae can a sampling and analysis of the species of the phytoplankton species in the area. In this project, it is more related to the prediction part, which usually involves modelling and simulating the algal bloom. The models which have been explored here are both traditional and modern. Traditional models may be known as biological, physical models, or known as the Nutrient-Phytoplankton-Zooplankton-Detritus (NPZD) models as well as models which takes into account the physical phenomenon of physical forcing and ocean currents.. More models were later included for use and explored which are the generalized additive model (GAM), global circulation model (GCM), neural network models, and models which combine different logics and models together in used, also known as coupled model. Examples of coupled models are the wavelet neural network model (WNN) and the wavelet-based autoregressive fuzzy model. It is found that the when comparisons were made, it can be seen that the WNN performs the best among the other models in terms of ease of use, computational resources required and the its nominal accuracy. Further to it, in a country like Singapore, water is a scarce resource for its people. The water in which is forms a source for consumers are the multiple reservoirs around the island, as well as the few rivers we have. In addition, these water catchment areas serves multiple functions such as leisure activities venues like the marina barrage and the kallang river. In terms of our food chain, the waters in fish farms are of critical in terms of quality and the presence of harmful algal blooms may lead to negative health impacts for our people. We can therefore see that monitoring, modelling, and forecasting harmful algal blooms becomes an important step in negating and / or avoiding the negative impacts, ecologically, biologically, and financially for our country.