Prediciting Algal Blooms in the Berowra Estuary, NSW, Australia

Publisher:
TuTech Verlag - TuTech Innovation GmbH
Publication Type:
Conference Proceeding
Citation:
Proceedings of the 10th International Conference on Hydroinformatics "Understanding Changing Climate and Environment and Finding Solutions", 2012, pp. 1 - 8
Issue Date:
2012-01
Full metadata record
Files in This Item:
Filename Description Size
Thumbnail2011005829OK.pdf Published version2.74 MB
Adobe PDF
The Berowra Creek estuary and its catchment are situated entirely within the Hornsby Shire Local Government Area on the northern outskirts of the Sydney metropolitan area. Berowra Creek is a major tributary of the lower Hawkesbury River, entering the Hawkesbury River some 25 kilometres from the ocean. The estuary itself extends for over 23 kilometres in a southerly direction from the Hawkesbury River to the tidal limit at Rocky Fall Rapids. Algal blooms are prevalent within the Berowra Estuary. When algal species are present in high numbers they pose serious problems for commercial and recreational users of the estuary. Management authorities require an understanding of the relationship between the incidence of algal blooms and the environmental conditions required to initiate and promote these populations. Artificial Neural Networks (ANNs) have been developed to predict the occurrence and risk of algal blooms within the Berowra Estuary. The ANNs developed were of a multilayer perceptron architecture and used the Broyden-Fletcher-Goldfarb-Shanno training algorithm. To enable the prediction of algal blooms, an instrumented buoy was deployed at Calabash Bay within the estuary. This buoy records temperature, salinity, photosynthetically available radiation and chlorophyll-a (CHLa). The ANNs were used to predict CHLa concentrations from one to seven days in advance. Presented in the paper will be an assessment of the prediction reliability together with an assessment of how the variables and the information pre-processing influences the reliability of the predictions obtained.
Please use this identifier to cite or link to this item: