Ecological dimensions in the SAF

1
Ecological dimensions in the SAF
SPICOSA Training Support Pack
Presentation outline

• Why do we need the ecological dimension?
• Methodology and data inputs
• Outcomes of a study site application (SSA)
• Management implications

Material produced by Jakob Walve
(Jakob.walve_at_ecology.su.se)
2
Why do we need the ecological dimension?
SPICOSA Training Support Pack

• - There is an environmental problem to be solved!
  In SPICOSA, problems related to e.g.
  eutrophication and declining fish stocks are
  addressed.
• - We want to achieve sustainable development,
  with acceptable environmental impact and status,
  e.g. according to Water Framework Directive. In
  SPICOSA, e.g. the development of fisheries and
  mussel farming are explored.

Material produced by ltnamegt ltemail gt
ltorganisation logogt
3
Methodology Systems thinking is the key
SPICOSA Training Support Pack

• Issue resolution What is the problem? What are
  the objectives?
• System definition System boundaries? Key
  ecological processes?
• Data needs / availability What data are
  available or can be made available? What will
  available data allow or restrict? What new data
  can be collected? -The main idea is to make
  better use of existing data.
• Conceptual Model description of relationships
  between system components, from expert knowledge.
  Forms the basis for the problem solving through
  numerical modelling. Often has to be simplified
  problem scaling.
• Formulation and Appraisal, i.e. Mathematical and
  Numerical modelling Inputs from data are
  modulated by ecological transformation processes,
  described by mechanistic (process) or empirical
  (relational) knowledge, in a modelling software.
  Usually the objective is to determine the
  response of a few system properties to certain
  management options. Important steps are
  Calibration and Validation of the model using
  data from the studied system.

4
SPICOSA Training Support Pack

• Issue resolution and System Definition
• Example of Policy Issue and important ecological
  processes illustrated with Overall Conceptual
  model

• Loch Fyne, a 60 km long fiordic sea loch on the
  west coast of Scotland
• Example Policy Issue Managing Loch Fyne so as to
  Maximize the Value of Ecosystem Goods and
  Services to the Local Economy

From SPICOSA Deliverable D.3.2. SAF Protocol on
CZ System Design (fig. 3.1)
5
Methodology Systems thinking is the key
SPICOSA Training Support Pack

• Conceptual Model an example

Example of fuzzy conceptual model of
phytoplankton growth
The fuzzy model has to be made more precise by
identifying and distinguishing state variables
(stocks), fluxes and their controls
From SPICOSA Deliverable D.3.2. SAF Protocol on
CZ System Design
6
Methodology Systems thinking is the key
SPICOSA Training Support Pack

• Conceptual Model example expanded to include a
  feedback loop

The conceptual model has been expanded to include
a feedback loop the recirculation of nutrients
from grazers
From SPICOSA Deliverable D.3.2. SAF Protocol on
CZ System Design
7
Methodology Systems thinking is the key
SPICOSA Training Support Pack

• Conceptual Model expanded to include boundary
  conditions

The conceptual model is expanded to include
boundary conditions The boundary condition will
be quantified inside a convertor rather than as
a stock because boundary conditions are outside
the dynamic model and should not be changed as a
result of what happens inside the model. Notice
also that there has to be one exchange flow for
every stock in the model.
From SPICOSA Deliverable D.3.2. SAF Protocol on
CZ System Design
8
Methodology Systems thinking is the key
SPICOSA Training Support Pack

• Data needs/availability identifying the data
  needs in the conceptual model for phytoplankton

Identifying data needs in the conceptual model
for phytoplankton. In addition to this, there is
need for calibration and validation data
From SPICOSA Deliverable D.3.2. SAF Protocol on
CZ System Design
9
Methodology Systems thinking is the key
SPICOSA Training Support Pack

• Conceptual Model simplified version

The model often has to be or can be
simplified. The model should be made as simple
as possible, but still has to capture the
essential dynamics of the system in relation to
the objectives of the model
From SPICOSA Deliverable D.3.2. SAF Protocol on
CZ System Design (fig. 8.2)
10
Methodology applied in a case study
SPICOSA Training Support Pack

• Issue resolution Example Himmerfjärden SSA,
  Sweden

Problems Reduced value due to relatively low
water transparency, and loss of macrophytes.
There is risk for cyanobacterial blooms if
nitrogen loads are reduced. There is a general
need to meet WFD requirements (ecological statsus
is moderate to poor according to present
classification). Export of nutrients to the
Baltic Sea. Specific ecological questions What
can be achieved by different measures (STP,
agriculture). What will be the effect of
improvements in the open Baltic Sea? What is the
effect of moved STP discharge point (potentially
affecting availability of nutrients in different
parts of the estuary)?
Himmerfjärden A brackish estuary in the Baltic
Sea.

• Water area 232 km2
• Mean depth 17 m
• Maximum depth 52 m
• Salinity 4-7
• Freshwater from 9 brooks and diffuse runoff 10 m3
  s-1
• From Lake Mälaren 7 m3 s-1
• From large sewage treatment plant (STP) 1.5 m3 s-1

10 km
Open Baltic Sea
11

• System Definition Example Himmerfjärden SSA,
  Sweden

The model boundaries were defined according to
the existing divisions into drainage basins.
We chose this area to simplify the model. This is
the most impacted area. The black dot indicate
sewage treatment plant (STP) discharge. The
boundaries for the ecological model are the
shorelines, i.e we do not model drainage basin
processes, but estimate load reductions according
to various scenarios of land use
The main fresh water input comes from Lake
Mälaren through the city of Södertälje (most of
Lake Mälaren discharge is through Stockholm). The
large drainage basin of L.Mälaren is not shown
here.
The sea area was divided into three sub-basins
according to natural dividers (sills) and
available data (sampling stations shown as red
dots)
12

• Conceptual Model Himmerfjärden example, water
  exchange

The water exchange conceptual model for the three
sub-basins was first divided into only two depth
layers, but was later developed into a
three-layer model. This gave a more realistic
model reflecting the actual sill depths between
the basins. Still, of course, it is a
simplification of the real world.
Legend for model
Box name
Volume (Mm3)
Salinity (avg. for 1997-2000)
depth
The numerical water-exchange model is heavily
data-dependent Salinity data is used to
calculate flows according to mass-balance.
13
SPICOSA Training Support Pack
Conceptual Model Himmerfjärden example, Ecology
Version 2
Version 1
This version was the initial ecological
conceptual model of the System Design step The
link to water transparency Secchi depth was not
shown, but was thought to be linked by empirical
relationship with chlorophyll
This is how the model was actually developed as a
first version. The first operational version was
however further simplified (next slide...)
14
SPICOSA Training Support Pack
Conceptual Model version3
Nitrogen loading
Total Nitrogen concentration
Water exchange
Secchi depth (water transparency)
Nitrogen retention
Secchi depth is estimated according to empirical
relationship between nitrogen concentration and
Secchi depth
15
SPICOSA Training Support Pack
Conceptual Model version3 with main links to
socioeconomic model shown
Nitrogen loading
Total Nitrogen concentration
Water exchange
Secchi depth (water transparency)
Nitrogen retention
Cost estimation of load reductions
Secchi depth is estimated according to empirical
relationship between nitrogen concentration and
Secchi depth
Economic valuation of gains
16
SPICOSA Training Support Pack
In practise Overview of ExtendSim layout for
Himmerfjärden Ecological model
WE Input data
FW inflow calculation
WE Const.
WE Water exchange model FW Fresh water
Basin 1
Basin 2
Basin 3
Flow calculation
Water balance
Basin1 Salt balance
Salt balance
Example
Salinity error calculation
Nitrogen input data
Nitrogen calibration data
Nitrogen-data export function
Nitrogen balance
17
SPICOSA Training Support Pack
Extend layout for Ecological model
18
SPICOSA Training Support Pack
Hindcast (validation) results Himmerfjärden
example
Loss of N during spring phytoplankton blooms
Variations in boundary conditions, nitrogen input
and water exchange explain most of the variations
in total nitrogen concentration. The biology
added is a loss of nitrogen during the spring
bloom, seen as a sudden drop in modeled nitrogen
concentration (blue line) in spring. This model
serves mainly one purpose to calculate total
nitrogen concentrations and from these the water
transpareny (Secchi depth)
19
SPICOSA Training Support Pack
Results of two scenario runs in the preliminary
version of the water-exchange/ecological model
1. Reference scenario 10 mg nitrogen per liter
from sewage treatment plant (STP) 2. Improved
sewage treatment scenario 4 mg N/ L from STP
Nitrogen loads to the model basin Himmerfjärden
Scenario 1
Scenario 2
STP
To Socio-economic model
Nitrogen load
Total nitrogen concentration
Secchi depth
Total nitrogen reduced from 390 to 320 µg/l in
Himmerfjärden, the largest basin
Load from STP reduced from 10 mg/l to 4 mg/l
Secchi depth increased from 2.8 to 3.5m
The model will have to be developed according to
the Conceptual model version 2 (or some other
idea with a simpler model) to anwer questions
about nitrogen fixing cyanobacteria and
chlorophyll concentrations
20
SPICOSA Training Support Pack

• Numerical modelling Lessons learnt
• Start simple construct Ball-park model that
  works (is possible to run) and that is
  successively developed to a more advanced stage
  with tests at each stage
• Save new versions, and document the changes (at
  least briefly)

21
Management implications
SPICOSA Training Support Pack

• The model can be an important tool, but since it
  is a simplification, and has certain objectives,
  it cannot answer everything. It may be more or
  less uncertain depending on how far scenarios are
  taken.
• The model will most likely be one decision
  support tool among others! The most important
  tool is a good general and expert knowledge of
  the system! The model will not replace this!
• The model may highlight certain data needs. The
  model may reduce data needs, but more likely it
  will be helpful in prioritizing which data to
  collect.
• Model may give results that the model does not
  itself answer how to handle, e.g the costs for a
  Secchi depth improvement are higher than
  calculated gains, but may partly result from the
  fact that qualitative benefits may be difficult
  to value. Or that measures reducing
  eutrophication also decrease yield of fisheries.
  Or that banning of commercial fisheries in favour
  of tourist fishery may result in higher profits,
  but may be politically difficult.