Skip to main content

Bayesian Network Analysis Exploring the Benthic Carrying Capacity for Finfish Farming Within the Firth of Thames

TR 2007/50

Report: TR 2007/50

Author: Hilke Giles, NIWA

Abstract

Environment Waikato is currently scoping a plan change to allow for the diversification of aquaculture within existing aquaculture management areas in the region that will potentially allow for the cultivation of species other than mussels, including finfish. Currently no information about environmental effects of finfish aquaculture is available for this region. This report presents an exploration of the benthic carrying capacity of the Firth of Thames for finfish farming to provide some background information for the aquaculture diversification plan change.

An estimate of the scale of expected benthic effects associated with fish farming in the Firth of Thames Wilson Bay Marine Farming Zone was obtained by carrying out a Bayesian network analysis supported by literature studies. In addition, estimates of the likely spatial extent of benthic effects were provided based on previous published studies as a first step to predicting the footprint of finfish cages.

The Bayesian network used in this study is a modified and re-parameterised version of an existing Bayesian network that was developed to quantitatively assess the relationships between benthic fish farm impact parameters and site and farm characteristics based on data published in peer-reviewed international journals from finfish farms located in temperate zones. The scale of benthic effects expected from establishing finfish farms in the Wilson Bay Marine Farming Zone was examined using selected case studies that represent realistic farming scenarios for this area. Case studies were created by examining different combinations of fish stocking density and water depth, which are expected to influence the scale of benthic impact. The focus of this report was to highlight the differences in benthic impacts predicted from the case studies and the trends generated from varying the input parameters, rather than making predictions of absolute variable values.

The Bayesian network analysis suggested that of the examined input parameters the free water depth below fish cages has the largest effect on the severity of benthic impacts and it is recommended that the minimum free water depth below cages should be 10 m. It is general practice in finfish aquaculture to use cages between 10 and 15 m depth and allow for a similar depth of water below cages to promote the dispersal of faeces and uneaten food, consequently the most suitable areas for the installation of fish cages are those with water depths of 20 m or more. Changes in stocking density only resulted in small changes in the probability distributions of most variables.

Selected literature studies were reviewed to examine the usefulness of monitoring parameters for the assessment of benthic impacts in the Firth of Thames that were not included in the Bayesian network. Video surveys and sediment trap deployments are not recommended following problems experienced during earlier work caused by unfavourable environmental conditions. The examination of opportunistic macrofauna species was generally accepted as a good indicator of benthic impact; however, it was suggested that measurements of biogeochemical parameters may reveal earlier signs of impact and allow remedial measures to be taken if necessary to prevent severe impacts.

It is recommended that pre-impact studies should be carried out in locations chosen for fish farming to gain an understanding of these processes prior to the additional organic enrichment. This would allow changes in biogeochemical processes to be identified and limits of acceptable sediment modification to be chosen based on sound data. Additional parameters considered potentially useful for the detection of severity and spatial extent of benthic impacts once fish farms are operating are trace metals and stable isotopes.

Estimates of the spatial extent of expected benthic impacts were derived from a review of peer-reviewed literature and monitoring data from New Zealand fish farms and it was concluded that 100 m was a cautious estimate. Since the largest change of most examined parameters took place within about 50 m of the farm and the gap between farm blocks in Area A is 75 m a 50 m buffer zone between the outermost cages inside a farm block and the perimeter of the block was considered an adequate estimate of the buffer zone for initial applications before measurements are available to make realistic assessments of spatial effects. Especially if cages larger than those examined in this study (>15 m diameter) are to be installed in the Firth of Thames, it is strongly recommended that benthic impacts should be measured at high spatial and temporal resolution until sufficient information on their severity and spatial extent has been gathered to make sound recommendations on minimum buffer zones for farm blocks.

To enable a reliable detection of farm footprints, it is also recommended that the natural variability of parameters used for future monitoring is measured prior to any farming activity. This will enable the identification of changes caused by the farms and minimise the problem of separating natural from farm induced changes observed in the farm area.

Bayesian Network Analysis Exploring the Benthic Carrying Capacity for Finfish Farming Within the Firth of Thames [PDF, 545 KB]

Contents
  Executive Summary iv
  Glossary 1
1. Introduction 1
1.1 Background 1
1.2 Brief overview of benthic effects of finfish farming 2
2. Scope and stipulations of this report 4
2.1 Finfish species 5
2.2 Wilson Bay hydrographic conditions 6
2.2.1 Water depth 6
2.2.2 Current speed 7
2.3 Case studies 8
3. Bayesian network analysis 10
3.1 Bayesian networks 10
3.2 Bayesian network for assessment of potential benthic effects of finfish farming 11
3.3 Predicted benthic effects of finfish farming in the Firth of Thames 13
4. Sensitivity to increases in FCR 23
5. Review of selected literature studies 25
5.1 Trace metals and video surveys 25
5.2 Biogeochemical processes and sediment trap deployments 26
5.3 Opportunistic macrofauna species 27
5.4 Stable isotopes 28
6. Estimates of farm footprints in the Firth of Thames 29
6.1 Spatial extent of benthic impacts 29
6.1.1 Estimates based on peer-reviewed literature 29
6.1.2 Estimates based on New Zealand monitoring data 29
6.2 Implications for buffer zones of farm blocks in the Wilson Bay Marine Farming Zone 32
7. Summary and recommendations 34
7.1 Bayesian network analysis 34
7.2 Reviewed studies 34
7.3 Spatial extent of benthic effects 35
8. References 36
9. Appendix 40