AquaPark

1
AquaPark Norad funded project

• Planning and management of aquaculture parks for
  sustainable development of cage farms in the
  Philippines

2
TROPOMOD modelling of aquaculture zones

• Chris Cromey
• Map and Marine Ltd
• Out put from AquaPark
• Aquapark Project Partners
• Bureau of Fisheries and Aquatic Resources
• Akvaplan-niva AS
• Map and Marine Ltd

3
Model assumptions
4
Simulation
5

• TROPOMOD modelling approach
• 3 important aspects
• How severe is the impact what is the maximum
  impact underneath cages?
• How far to the boundary of the impact? (Allowable
  Zone of Effect)
• How can husbandry practices be optimised to use
  the zone most productively?
• Objectives
• Predict if impact is SEVERE underneath cages
• as shown by this deposition footprint
• Zone colour
• Predict distance to boundary of MODERATE impact
• Zone colour

Edge of Park
6
TROPOMOD modelling Approach
Maintain enough spacing between cage rows so that
remediation of sediments can take place impact
should be LOW between rows in each
zone Maintain enough space between cage
rows to prevent reduction of currents by high
aggregation of cages
Zone colour
7
TROPOMOD modelling Approach
Encourage careful feeding, so that there is less
waste feed and less wastage of money Encourage
better quality feed Feed digestibility is
increased Less feed is needed Better quality
feed also breaks up less, so more goes to
growth Distance between parks Prevent overlap
of impact by predicting the extent of the zones
e.g. 600 m between zones for this site
8
Panabo sediment trap model validation studies
Chris Cromey EMMA2 project modeller using data
from Akvaplan and BFAR survey teams
Objective to measure flux (deposition) of waste
feed and faeces under cages and compare with
model predictions this helps us to verify the
model predictions and test different
scenarios Sediment traps were deployed in
Spring 2010 survey One transect for Milkfish (0
metres from cage, 10 m and reference) One for
Grouper (0 metres from cage, 10 m and reference)
9
Sediment trap model validation study method for
deploying sediment traps
Cage
Current
Sea bed
1. Deploy traps 2. Retrieve, filter, dry
solids 3. Calculate observed flux (total waste
feed and faeces in traps grams per m2 bed per
day) 4. Compare with TROPOMOD model
75 cm HD 51 ratio
10
TROPOMOD model method

• Cage positions were obtained from recent Google
  Earth image
• Feed ration for cages obtained from AquaPark
  production survey
• Production survey results summarised to include
  modelling of different stages of Milkfish
  monoculture (starter, grower and finisher), as
  well as other species (Milkfish-Siganid
  polyculture, Grouper, etc.)

11
TROPOMOD model method

• Settling velocity of feed obtained from EMMA2
  project pellet break up experiments and PHILMINAQ
  project experiments
• Settling velocity of Milkfish faeces from
  PHILMINQ project
• Depth obtained from BFAR surveys
• Uneaten feed (27), digestibiity (49) and water
  content (9) data from PHILMINAQ project

12
TROPOMOD model method

• Current meter data from BFAR surveys used
• TROPOMOD set up with information and used to
  predict flux of waste feed and faeces (grams per
  m2 bed per day) for Panabo Aquaculture park

13
Model predictions
TROPOMOD predicted flux (deposition) of waste
feed and faeces for Panabo The colours show
different amounts of flux
Sediment trap transects Grouper (traps are in 6
m depth) Milkfish transect (traps are in 27 m
depth) Reference traps
Higher flux and predicted impact in the inshore
areas
14
Milkfish sediment trap transect Problem cage
locations taken from recent Google Earth image
and trap locations taken from survey do not match
as station M0m should be at cage edge, and M10m
should be 10 m from cage Solution make minor
adjustments to cage/station positions in model so
they match
Two stations (three traps at each) Feed ration
50 kg per cage per day
15
Grouper sediment trap transect These cages are
in shallower water and feed ration is 10 kg per
cage per day, which is less than for Milkfish
monoculture
Two stations (three traps at each)
16
Milkfish comparison of observed flux and
predicted
For both the 0 m and 10 m stations, the model
slightly over-predicted flux Note the variation
between traps at the same station Traps are
adjusted for background sedimentation
17
Grouper comparison of observed flux and
predicted
Model predictions of flux were satisfactory for
Grouper Note the variation between traps at the
same station Traps are adjusted for background
sedimentation
18
All sediment trap data collected by Cromey et al.
at SAMS Panabo measurements were less than 20 g
m-2 d-1, which is at the low end of measurements
taken
19
Conclusions and next steps TROPOMOD predicted
flux satisfactorily for Grouper, but
over-predicted for Milkfish we will look at
this more carefully after refinements Modelling
of the whole park showed higher impact in the
inshore areas, where the depth is shallower and
dispersion is lower Modelling the whole park
requires care, as cage locations and feed inputs
vary widely across the park depending on the
source of data for management scenarios, some
randomness can be introduced to account for this
20
Conclusions and next steps Next steps are to
model the park by organising species into
zones Identify areas for oysters and
seaweed Further refinements of the modelling are
needed, incorporating as much information from
the Spring 2010 surveys as possible