To Main Proceedings Document
A decision-tool for adaptive
renewable resources management
Erling Moxnes, SNF, Norway
Oje Danell, University of Uppsala, Sweden
Eldar Gaare, NINA, Norway
Jouko Kumpula, Finnish Game and
Fisheries Research Institute, Finland
Corresponding author: Erling Moxnes
Breiviksveien 40, 5045 Bergen
+47 55959526 (fax -439)
Erling.Moxnes@ snf.no
How to deal with complexity?
¢ Experiments show that management of
renewable stock resources is complicated,
Moxnes (98) and Moxnes (98).
dynamics (stocks and flows)
non-linearities
uncertainty and risk
learning
¢ What to offer clients with limited budgets?
Simulation studies
Optimization studies
Group model building
Simulators
Simplification and client involvement
(based on previous four)
Adaptation of reindeer herd size to
lichen availability
* Key problem found in experiments: An
unknown, non-linear, growth relationship
dL /dt =g(L) —c(N)
Consumption and growth (mm/year)
10
oN FD &
0 10 20 30 40 50 60
Lichen thickness (mm)
Simplifications in decision-tool
From complex optimization under
uncertainty, Moxnes et al.(99):
- Appropriate to aim for maximum
sustainable yield for lichen
From complex Bayesian estimation:
- Adjust afew key parameters in the priors
for lichen growth to fit available data
From complex adaptive management
(optimal management and learning)
- Manipulate the herd size to get growth data
for different levels of lichen (i.e. deviate
from maximum sustainable yield)
Working of tool:
Adaptive management
/ Estimation Nae
Uncertainty Goal
[
Herd size
Lichen
~~
Bayesian calibration
Testing/Use
¢ Simulator generates data for testing, and
can be used for training sessions
¢ Test with 10 students, each with 6 trials:
- key parameters were found with an
average accuracy of 15 to 20 percent.
* To be tested in all Nordic countries
200 é +44
150 a $38 ; 3 ; ie
100
50
0 {
0 5 10 15 20 25
¢ Hopefully, the tool will:
- generate client involvement in the
production and analysis of crucial data
- structure data such that proper policies
follow naturally
