Skip to main content

Modelling E. coli for the Hauraki Plains and Coromandel Peninsula

TR 2016/39

Report: TR 2016/39

Author: Annette Semadeni-Davies, Michelle Greenwood and Sandy Elliott (NIWA)

About this report

“E. coli” bacteria live naturally in the guts of warm-blooded animals, including people and farm animals, and are excreted from the body in dung. Following excretion, E. coli can be carried into rivers and streams. The level of E. coli in a waterbody is a widely-used measure of the suitability of the water for swimming (and other recreation use). Previous work has shown that E. coli concentrations in rivers and streams in the Waikato region – including the Hauraki Plains and the Coromandel Peninsula – are typically higher than those suitable for swimming waters. E. coli concentrations tend to be highest in rivers draining areas used for farming. E. coli are also present in the wastewaters discharged from sewage treatment plants and from certain industries (e.g. meatworks).

In this study, data from routine river water quality monitoring was used to develop mathematical models of the amount of E. coli in rivers and streams on the Hauraki Plains and the Coromandel Peninsula. These models identified the factors contributing to the observed E. coli levels. Models like this can be used to identify the changes that would be needed in a given catchment to reduce E. coli levels to those that were suitable for swimming. Models like this will therefore be needed should the community decide it wishes to be able to swim safely in Hauraki and Coromandel rivers (e.g. in the proposed Plan Change #2).

Read or download the report

Modelling E. coli for the Hauraki Plains and Coromandel Peninsula [PDF, 3.5 MB]


1 Introduction
1.1 Scope
2 Input data
2.1 Drainage network
2.2 Catchment characteristics
2.3 Land use
2.4 Point sources and farm dairy effluent inputs
3 Calibration data
3.1 E. coli concentrations
3.2 Flow data
3.3 Mean annual loads
4 Load modelling
4.1 Model description
4.2 Calibration and performance assessment methods
4.3 Load model results
5 Concentration models
5.1 Concentration model results
6 Model uncertainty and error
7 Recommendations for further work
8 Summary and conclusions